*Article* **A DSC Test for the Early Detection of Neoplastic Gastric Lesions in a Medium-Risk Gastric Cancer Area**

**Valli De Re 1,\* , Stefano Realdon <sup>2</sup> , Roberto Vettori <sup>1</sup> , Alice Zaramella 3,4, Stefania Maiero <sup>2</sup> , Ombretta Repetto <sup>1</sup> , Vincenzo Canzonieri 5,6 , Agostino Steffan <sup>1</sup> and Renato Cannizzaro 2,6**

	- <sup>4</sup> Department of Surgery, Oncology, and Gastroenterology (DiSCOG), University of Padua, Via Giustiniani 2, 35128 Padua, Italy

**Abstract:** In this study, we aimed to assess the accuracy of the proposed novel, noninvasive serum DSC test in predicting the risk of gastric cancer before the use of upper endoscopy. To validate the DSC test, we enrolled two series of individuals living in Veneto and Friuli-Venezia Giulia, Italy (n = 53 and n = 113, respectively), who were referred for an endoscopy. The classification used for the DSC test to predict gastric cancer risk combines the coefficient of the patient's age and sex and serum pepsinogen I and II, gastrin 17, and anti-*Helicobacter pylori* immunoglobulin G concentrations in two equations: Y1 and Y2. The coefficient of variables and the Y1 and Y2 cutoff points (>0.385 and >0.294, respectively) were extrapolated using regression analysis and an ROC curve analysis of two retrospective datasets (300 cases for the Y1 equation and 200 cases for the Y2 equation). The first dataset included individuals with autoimmune atrophic gastritis and first-degree relatives with gastric cancer; the second dataset included blood donors. Demographic data were collected; serum pepsinogen, gastrin G17, and anti-*Helicobacter pylori* IgG concentrations were assayed using an automatic Maglumi system. Gastroscopies were performed by gastroenterologists using an Olympus video endoscope with detailed photographic documentation during examinations. Biopsies were taken at five standardized mucosa sites and were assessed by a pathologist for diagnosis. The accuracy of the DSC test in predicting neoplastic gastric lesions was estimated to be 74.657% (65% CI; 67.333% to 81.079%). The DSC test was found to be a useful, noninvasive, and simple approach to predicting gastric cancer risk in a population with a medium risk of developing gastric cancer.

**Keywords:** gastric cancer; pepsinogen; gastrin G17; *Helicobacter pylori*; screening

#### **1. Introduction**

Gastroscopy is the standard procedure for gastric cancer (GC) diagnosis with a falsenegative rate of about 19% [1]. This procedure is invasive, time-consuming, and uncomfortable.

The 5-year GC survival rate is poor, reaching approximately 30% in Europe [2]. Studies in countries such as Japan and Korea have shown a high incidence of GC, with a 30% reduction in GC mortality due to gastroscopy screening programs [3].

Overall, a GC diagnosis at an early/asymptomatic stage not only improves clinical outcomes [4] but is associated with endoscopic resections in most cases, resulting in the now-standard treatment for early gastrointestinal cancers without regional lymph node metastasis [5].

A population with an age-standardized incidence rate (ASIR) of 10 to 20 per 100,000 is considered to have an intermediate risk of developing GC. According to published data, in

**Citation:** De Re, V.; Realdon, S.; Vettori, R.; Zaramella, A.; Maiero, S.; Repetto, O.; Canzonieri, V.; Steffan, A.; Cannizzaro, R. A DSC Test for the Early Detection of Neoplastic Gastric Lesions in a Medium-Risk Gastric Cancer Area. *Int. J. Mol. Sci.* **2023**, *24*, 3290. https://doi.org/10.3390/ ijms24043290

Academic Editor: Monica Currò

Received: 27 September 2022 Revised: 30 January 2023 Accepted: 3 February 2023 Published: 7 February 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

northern Italy, the ASIR was calculated as 12 to <14 in 2017, on average twofold higher in males (33.9) than in females (17.0) [6,7]; thus, Italy is a geographic area whose population has an intermediate risk of developing GC.

At present, the screening for GC is only performed in countries with an elevated risk of GC, such as Japan and Korea [8]. In a country with an intermediate risk of GC, using gastroscopy as first-line testing alone is not considered feasible due to its invasiveness and expensive cost [9]. On the other hand, the inappropriateness of upper endoscopies leads to decreased diagnostic yield [10,11]. Thus, it is necessary to consider a less invasive and more cost-effective solution to find subjects at risk of developing GC. Moreover, it is necessary to take changes in GC epidemiology that have occurred over the last few decades into consideration. GC epidemiology has changed concomitantly with a reduction in *Helicobacter pylori* (*H. pylori*) infections, but also with an increased incidence of cardia GC and an increase in GC diagnoses and mortality in younger adults [12].

The available pepsinogen (PG) test (PG test) is based on a combination of the serum PG-I level and the PG-I/PG-II ratio, which is used as a marker for chronic atrophic gastritis (CAG), and it has been widely proposed for the selection of patients at risk for GC [13,14].

However, while the PG test is accurate for CAG diagnosis, it suffers from an unsatisfactory specificity in predicting GC due to using the widely standardized cutoff points of PGI < 70 ng/mL and PGI/PGII < 3.0 [15–17]. Therefore, the ability to predict GC risk needs to be improved.

Recently, the addition of serum gastrin G17 (G17) and anti-*H. Pylori* immunoglobulin (IgG) to the PG test was proposed to improve the diagnosis of atrophic gastritis and GC [18]. PGI is only secreted by the fundic glands; PGII is secreted by the fundic glands, the pylorus, and the Brunner glands; and gastrin G-17 is only secreted by gastric antral G cells. Accordingly, serum PG and G17 levels can be used to localize morphology and detect the extension of gastric lesions [19]. Gastrin G17 might also sustain the proliferation and migration of epithelial gastric cells [20]. *H. pylori* is pathogenetic for GC and a subtype of autoimmune gastritis [21]. Age ≥ 60 and male sex have been reported to be independent risk factors for GC in several studies [22,23].

Based on our earlier study [24] and two new datasets in this work, first (the discovery part), we developed a model—herein called the DeRe–Steffan–Cannizzaro (DSC) test—to discriminate patients at risk of developing GC. Briefly, the model is based on regression analyses to calculate the coefficient of the patient's age and sex and their PG, anti-*H. Pylori* immunoglobulin (IgG), and serum gastrin G17 levels.

Second (the validation part), we validated the DSC test. Two monocenter studies were conducted involving 53 retrospectively selected (based on their diagnosis) individuals who presented at the Gastroenterology Unit of the Veneto Institute of Oncology IOV-IRCCS, Padua, Italy (validation cohort 1), and 113 consecutive individuals who were referred by physicians to the oncological gastroenterology unit of the Centro di Riferimento Oncologico of Aviano, Italy (validation cohort 2). Physician indications included dyspepsia, preneoplastic lesions, and/or suspected GC. Dyspepsia was defined as upper gastrointestinal symptoms, such as gastric pain and/or burning, without a typical disease and with no clear cause [10,11]. The preneoplastic lesions were diagnosed via gastroscopy and histologically examining the biopsies [25–29]. The GC diagnoses were confirmed via histopathological examinations [30,31]. Data on each patient's age and sex were collected, and expression profiles of serum PG, gastrin G17, and *H. Pylori* IgG were performed in order to group patients based on GC risk classes into negative, neutral, and positive results, respectively. The accuracy of the model was then demonstrated with gastroendoscopy and a histological diagnosis.

This is the first study on GC risk prediction adopting the DSC model. The model showed an overall 74.66% accuracy rate for GC diagnosis, notably improving the detection sensitivity from 15.00% to 70.00% and retaining a good specificity (74.66%) compared with the standard cutoff for PG tests. This study could help physicians make better decisions and allocate proper resources to improve GC prevention, for example, by removing high-grade preneoplastic lesions and selecting patients for endoscopic surveillance. In addition, by diagnosing opportunistic GC at an early stage, the GC survival rate can be improved [2,3].

#### **2. Results**

#### *2.1. Study Design*

Based on our previous observations [24] and the two new cohorts in this work (Figure 1, discovery cohort 1 and discovery cohort 2), we developed a model based on the coefficients of the patient's age, sex, and serum PG, *H. Pylori* IgG, and gastrin G17 levels to discriminate patients at an elevated risk of GC. The results of the Y1 and Y2 equations obtained from discovery cohort 1 and discovery cohort 2 are reported in Figure 1. Data from all the study's participants are reported in Table 1. The median G17 level in the female category in discovery cohort 1 was higher than in the other groups due to the prevalence of females affected by autoimmune atrophic gastritis [24]. In discovery and validation cohorts, *H. pylori* IgG and PG II levels were higher in individuals >65 years old, in accordance with the decreased incidence of *H. pylori* infection in the global population [12] and increases in serum PG II levels often found in individuals infected with *H. pylori* [24].

Next, we recruited two validation cohorts, the first consisting of a retrospective series of selected participants from the Veneto region (validation cohort 1, n = 53) and the second consisting of a prospective series of consecutive participants who were referred to gastroenterologists for gastroscopy from the Friuli geographic area (validation cohort 2, n = 113). The study participants' data are reported in Table 1.

#### *2.2. DSC Classifications for GC Risk*

Two validation cohorts were enrolled in the study. The first cohort consisted of a retrospective series of 53 participants (validation cohort 1); the second cohort consisted of a prospective series of 113 participants consecutively enrolled since May 2020 (validation cohort 2).

The DSC test measures the serum biomarkers and demographic data of the participants. The patients' data are reported in Table 1. Based on the results of both the DSC Y1 and Y2 equations, Y1 < 0.385 and Y2 < 0.294 (as reported in Figure 1 and described in the methods section), we classified participants receiving the serological test as having negative, neutral, or positive DSC results.

The results of the DSC test are reported in Table 2 according to the calculated Y1 and Y2 equations for each participant. Based on these criteria, 19 (35.8%) and 81 (71.7%) participants were classified as negative; 5 (9.4%) and 10 (8.8%) were classified as neutral; and 29 (54.7%) and 22 (19.5%) were classified as positive for GC risk in validation cohort 1 and validation cohort 2, respectively.

#### *2.3. Gastroscopy and Histopathological Diagnosis*

The same participants were classified after using gastroscopy and histological examinations to diagnose them with GC, dysplasia, severe atrophy (OLGA stages III–IV), or no/moderate-grade atrophy (OLGA stages 0–II). The diagnosis results are summarized in Table 3.

In the first cohort, consisting of a retrospective series of participants (validation cohort 1), the diagnosis was GC in nine patients, dysplasia in five, and severe atrophy in 14. The remaining participants (n = 25) showed no atrophy or mild–moderate atrophy (OLGA stages 0–II).

Of the 113 consecutive participants in the second cohort (validation cohort 2), a neoplastic lesion was diagnosed in two patients (i.e., one advanced, differentiated, diffusetype GC with middle tumor infiltration; one early antrum GC), dysplasia was diagnosed in three patients (one adenoma with high-grade dysplasia in the corpus; one low-grade dysplasia with intestinal metaplasia and neuroendocrine hyperplasia in the antrum; one dysplasia with intestinal metaplasia in the corpus), a preneoplastic adenoma was diagnosed

in one case, and severe atrophy was diagnosed in 15 cases; the remaining participants (n = 92) showed no atrophy or mild–moderate atrophy (OLGA stages 0–II).

**Figure 1.** Schematic of the study design and main results.

**Figure 1.** Schematic of the study design and main results.



**Table 1.** Distribution of demographic and serological data in the discovery and validation cohorts according to sex and age categories.


**Table 2.** DSC GC risk classifications for participants in the validation cohorts.

**Table 3.** Aggregation of participants according to histopathological diagnosis.


In the patient with an early GC and the four patients with dysplasia, mini-invasive gastric resections were conducted via endoscopy.

#### *2.4. DSC Classification Accuracy*

A comparison between the DSC classification results and a histological classification was used to calculate the accuracy of the DSC test. A subgroup analysis was performed separately for validation cohort 1 and validation cohort 2, and then for all cases.

In both validation cohorts, we observed a substantially larger risk of GC in individuals classified as DSC-positive (validation cohort 1: 7/9 (77.78%); validation cohort 2: 2/2 (100%)) than those classified as DSC-neutral (validation cohort 1: 1/5 (20.0%); validation cohort 2: 0/10) or DSC-negative (validation cohort 1: 1/19 (5.3%); validation cohort 2: 0/81). The relative risk of GC diagnosis in individuals classified with a positive DSC score was RR 2.90 (95% CI; 0.67–12.67) and RR 20 (95% CI; 0.99–402.45) in validation cohort 1 and validation cohort 2, respectively.

Figure 2 shows the distribution of DSC-positive, -neutral, and -negative results in the validation datasets according to the individual diagnosis of each participant.

Based on the DSC test, 60.2% of the overall participants in the validation cohorts were classified as negative, 9.0% as neutral, and 30.7% as positive for GC risk (Table 2).

The predictive value of the DSC test for the risk of GC was calculated using a diagnostic test; the AUC value was 0.723 considering all participants in the validation datasets (0.614 in validation cohort 1 and 0.749 in validation cohort 2). The overall accuracy was 74.657%, with a calculated disease prevalence of about 0.01% in the general Italian population [6] (see details in Table 4).

#### *2.5. Reproducibility of the DSC Method*

In 26 participants (17 negative, three neutral, and six positives in the DSC test), the test was repeated after a median interval of 15.5 months (IQR, 10 to 19 months) (Table 5). We found a change in DSC classifications from negative to neutral in two participants with a diagnosis of mild–moderate atrophy (OLGA 0-II category). In the remaining 24 cases, both the classification and the diagnosis remained the same (Table 5).

**Figure 2.** Distribution of DSC results taken from the validation 1 (**A**) and validation 2 (**B**) datasets according to the diagnostic categories. **Figure 2.** Distribution of DSC results taken from the validation 1 (**A**) and validation 2 (**B**) datasets according to the diagnostic categories.


LR, likelihood ratio; PV, predictive value.


**Table 5.** Reproducibility of the DSC test.

*2.6. Comparison of the Overall Validation Process (n = 166 Cases) Using the DSC Test and the Standardized Pepsinogen Test*

The standard cutoff for the PG test is PG I < 70 ng/mL and a PG I/PG II ratio of ≤3.0, resulting in an area under the curve (AUC) of 0.470 (Table 6). For the DSC test, the AUC was 0.723, a better AUC, mostly due to the increase in the sensitivity value from 15.00% to 70.00%. Because the predictive value depends on the prevalence of the disease (0.01% in the general population of Italy), the predictive value for positive cases was 0.03% when using the DSC test and 0.01% when using the PG test, while the overall accuracy remained similar between tests.

**Table 6.** Comparison between the predictive values of the DSC test and pepsinogen tests for GC risk.


LR, likelihood ratio; PV, predictive value.

#### **3. Discussion**

In this work, we proposed a screening strategy to select individuals at risk for GC based on our newly developed DSC test. The DSC test showed good accuracy (74.66%) with an increase in sensitivity when compared with the PG test, meaning that it could be helpful in identifying gastroenterological patients for opportunistic GC screening in medium-risk areas, e.g., Italy. The DSC test is noninvasive, reproducible, and has high specificity (i.e., a true negative rate). In individuals with a positive DSC classification, gastroscopy and surveillance should be recommended to detect GC at an early stage and improve prevention rates, for example, by using mini-invasive cures.

Earlier studies have confirmed the high risk of GC in patients with atrophic and/or metaplastic gastritis. In particular, two pathological classifications developed from the Sydney System to grade long-standing gastritis and metaplasia (the OLGA and OLGIM classifications) have been shown to be informative in determining severe atrophy/metaplasia (stage III-IV) and an increased risk of GC development [32–34]. Based on these results, gastroenterology guidelines recommend an endoscopic follow-up every 3 years in individuals with a diagnosis of severe atrophy/metaplasia [35]. Performing the DSC test on these patients could be a useful approach for clinical practices better detect individuals at higher risk, who would then be examined using a stricter endoscopic approach in the follow-up.

The results obtained from using the DSC test on the prospective, non-selected, individual cohort (validation cohort 2) are comparable to those achieved using a similar approach in an area of high GC incidence (GC prevalence, 2.84%; ROC curve predicting GC, 0.79). We propose using this test as an opportunistic screening method in individuals attending the gastroenterology attention to improve the detection rate of GC at an early stage [36].

To the best of our knowledge, there has only been one prospective study that combined the pepsinogen test in a population with a medium GC risk [37], but the study was focused on surveilling patients with precancerous lesions (atrophic gastritis, intestinal metaplasia, dysplasia). The authors showed that only patients with extended precancerous lesions with a low PGI/II ≤ 3 ratio and/or OGLIM stage (III-IV) developed high-grade dysplasia or neoplasia at follow-ups after about 57 months. However, it is noteworthy that atrophy is usually associated with *H. pylori* infection and intestinal-type GC. The proposed DSC model introduces the possibility of selecting individuals at high risk of opportunistic neoplastic lesions for further endoscopic examinations. In our study, at a median follow-up of 15.5 months, two individuals out of 17 showed an increase in DSC classification from the negative category to the neutral category, however the histological diagnosis remained at moderate atrophy (OLGA stage 0-II).

GC is a disease of old age, with about one-half of patients with GC being over age 65. A primary characteristic of aging is a progressive loss of physiological gastric tissue integrity that, in turn, leads to impaired function and increased alteration of the PG-I/PG-II ratio, which leads to a decline in gastric acidity [37] and a low circulating concentration of vitamin B12 [38,39]. This could justify a potential decrease in the accuracy of the DSC classification in older subjects (>75 years old); as such, it is necessary to take this aspect into consideration regarding the idea of screening the general population. Therefore, it may be helpful to add other biomarkers to increase the DSC accuracy by reducing the number of false positives in particular in individuals who are >75 years old.

Several studies have reported the serum indicators of GC. Some of them have been proposed in combination with pepsinogen to improve the accurate diagnosis of GC, e.g., sugar carbohydrate antigen 72-4 (CA72-4) [40], CEA, CA12-5, and CA19-9 [41]; metabolites such as hydroxylated sphingomyelins (SM(OH)) and acylcarnitine derivatives (C2, C16, and C18:1) [42]; alcohol dehydrogenase (ADH) activity [43]; interleukin-6 (IL-6); human epididymal protein 4 (HE-4); adiponectin; ferritin and Krebs von den Lungen (KL-6) [44]; soluble T cell immunoglobulin; and mucin domain molecule 3 (sTim-3) [45]. Overall, these results are interesting, but they are all preliminary and need to be evaluated in large prospective studies in combination with the DSC test.

In the last few years, new, increasingly complex technological strategies have emerged, such as the single-cell RNA sequencing (scRNA) transcriptome, which can characterize cellular and molecular networks present in a tissue at the same time [46]. scRNA offers the possibility of finding new GC diagnostic markers by reducing the complexity of gene RNA expression patterns in distinct cell populations. This approach is also useful in deciphering GC pathogenesis and detecting rare, tumor-specific cells at the onset of early GC stages. Currently, only one study has reported potential diagnostic markers for GC using this technology [46], but candidates have also been found in other tissues. Moreover, it is necessary to screen candidates such as secretory proteins, which are abundantly expressed in GC cells and could be evaluated in the serum alone or in addition to the DSC test for GC prediction. Furthermore, their sensitivity and specificity in comparison to the DSC test should be assessed.

Our study had limitations due to the limited number of cases tested and its application in a unique laboratory. Its use in other medium-risk GC populations is necessary to support its validity.

#### **4. Materials and Methods**

#### *4.1. Study Cohorts*

This study was based on four cohorts of participants: two discovery cohorts (discovery cohort 1 and discovery cohort 2) and two validation cohorts (validation cohort 1 and validation cohort 2). Demographic and serological data were collected from all participants. Subjects who initially declined consent were excluded from the study.

Discovery cohort 1 included 300 retrospectively selected subjects at risk of GC, i.e., subjects with a family history of GC, autoimmune chronic atrophy gastritis, severe precancerous lesions, or a diagnosis of GC.

Discovery cohort 2 included 200 subjects: blood donors and patients with a GC diagnosis.

Validation cohort 1 consisted of 53 selected individuals from a retrospective series of cases with no atrophy, mild–moderate atrophy, dysplasia, or GC enrolled in the Veneto region, Italy.

Validation cohort 2 consisted of prospective participants recruited from May 2020 to September 2022 (n = 113). Cases were consecutive participants who were referred to a gastroenterologist by a physician for gastroscopy.

Participants from the four cohorts received serological tests to determine their serum PG, G17, and *H. pylori* IgG values. Participant data according to the cohorts are detailed in Table 1.

The study was conducted following the Declaration of Helsinki and approved by the Unique Regional Ethics Committee for Friuli-Venezia Giulia (approval number CRO-2019-46). Written informed consent was obtained from all participants.

#### *4.2. Construction of the DSC Model*

Data from each participant enrolled in discovery cohort 1 and discovery cohort 2 were used to perform a regression analysis that discriminated patients with GC and determined the coefficients of each variable (i.e., PG, age, sex, *H. pylori* IgG, and gastrin G17)

The resulting Y1 and Y2 equations discriminated patients affected by GC in discovery cohort 1 and discovery cohort 2, respectively.

Y1 = −7.49 − 0.0025 × PG I + 0.097 × PG II − 0.017 × G17 − 0.0007 × HP IgG + 0.1 × age + 1.18 × if male.

Y2 = −9.28 + 0.0015 × PG I + 0.0824 × PG II − 0.009 × G17 + 0.0001 × HP IgG +0.1337 × age + 0.423 × if male.

To establish the best cutoff for the discrimination, we used an ROC curve analysis. The combination of the Y1 and Y2 results, based on cutoffs of Y1 < 0.385 and Y2 < 0.294, determined the DSC classification model.

In detail, subjects positive for Y1 > 0.385 and Y2 > 0.294 were classified as having high GC risk; subjects positive for only one (Y1 or Y2) were classified as having medium risk; and subjects were deemed low risk when any of the Y1 or Y2 criteria were satisfactory.

#### *4.3. DSC Model Assessment in the Validation Cohorts*

Participants in the validation cohorts were classified using the DSC model. The DSC classification of each participant was recorded in a database. For each participant, the data were correlated with the diagnosis obtained via gastroscopy and the histological examination of the biopsies.

#### *4.4. Serological Data*

Blood samples of approximately 5 mL each were obtained from all participants after 10 h of fasting. The tubes were centrifugated for 10 min at ≥10,000 rpm. The serum was stored immediately at −20 ◦C until an assay was performed. Serologic testing for *H. pylori*-IgG, PGI, PGII, and gastrin G17 was performed using an automated chemiluminescence immunoassay (CLIA) Maglumi analyzer (Medical Systems). Recommended cutoff points, as reported by the manufacturer, were PGI: 70–240 ng/mL, PGII: <13 ng/mL, G17: 2–10 pmol/L, H.p IgG titer: <30 EIU. A combination of PGI < 70 ng/mL and a PGI/PGII ratio of <3 is informative for atrophic gastritis (AG).

#### *4.5. Diagnosis*

A gastroendoscopy was performed by gastroenterologists on all participants, and biopsies were collected (two biopsies from the antrum, two from the body, and one from the incisure). Gastric mucosae were examined with high-definition (HD) white-light endoscopy and narrow-band imaging (NBI) to improve the visibility of blood vessels and mucosal structures.

Pathological examinations of biopsy samples fixed in buffered formalin (10%) were conducted by an expert pathologist, and the results were reported according to updated OLGA stages to define gastritis [29], and Lauren [30] and Who [31] for GC classifications of gastric tumors and dysplasia. According to the pathological examinations, participants were classified into three groups: without atrophy or metaplasia, with mild–moderate preneoplastic gastric lesions (atrophy or metaplasia), with dysplasia, and with neoplastic lesions.

#### *4.6. Statistical Analyses*

All analyses were conducted using MedCalc Statistical Software, version 19.0.4 (Med-Calc Software bvba, Ostend, Belgium). The results were summarized using age intervals and descriptive statistics. Comparisons between groups were made using a one-way ANOVA test. The predicted DSC scores for each patient were categorized into low-risk, medium-risk, and high-risk groups based on results of the Y1 and Y2 equations as follows: high-risk: Y1 > 0.385 and Y2 => 0.294; medium-risk: Y1 > 0.385 or Y2 => 0.294; low-risk: neither Y1 > 0.385 nor Y2 => 0.294. Considering the histopathology of the gastric lesions, the subjects were categorized into three groups: without lesions, with mild–moderate gastric lesions (chronic and/or autoimmune atrophic gastritis, intestinal metaplasia), and with severe gastric lesions (dysplasia, early GC, and advanced GC). Categorical data were entered into a two-way table by counting the number of observations that fell into each group of variables. A Spearman's rank correlation was used to test the relationship between the DSC and histological categories. To test the accuracy of the high-risk DSC model in predicting the diagnosis of dysplasia and GC, we used a diagnostic test; *p* < 0.05 was considered statistically significant.

#### **5. Conclusions**

We developed the DSC test model and assessed its accuracy for the classification of people at a high risk of developing GC in two validation datasets. The DSC test achieved a good accuracy of 74.66% and a sensitivity increase of 70.00% compared with 15.00% for the PG test, which supports its potential utility in clinical practice for opportunistic GC identification and the selection of patients at elevated risk for strict follow-ups. It remains to be seen whether the DSC test is effective in follow-ups, since gastric lesions may progress or develop later.

**Author Contributions:** Conceptualization: V.D.R. and R.C.; methodology: V.D.R.; software: R.V. and A.S.; validation: V.D.R., R.C., S.R. and V.C.; formal analysis: V.D.R., S.M., V.C. and A.Z.; resources: R.C. and A.S.; investigation: V.D.R., R.C. and V.C.; data curation: V.D.R.; writing—original draft preparation: V.D.R.; funding acquisition: R.C. and A.S.; writing—review and editing: V.D.R., O.R., R.C. and V.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** We would like to thank the Italian Ministry of Health—Ricerca Corrente for its financial support for this research.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board Unique Regional Ethics Committee for Friuli-Venezia Giulia (approval number CRO-2019-46; date of approval, 14 January 2020). Written informed consent was obtained from all participants.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available in this article.

**Acknowledgments:** We would like to thank Cinzia Crepaldi and Chiara Corso for their technical support in this research.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Article* **Discovery of Small Molecule COX-1 and Akt Inhibitors as Anti-NSCLC Agents Endowed with Anti-Inflammatory Action**

**Mehlika Dilek Altıntop 1,\* , Gül¸sen Akalın Çiftçi 2,3, Nalan Yılmaz Sava¸s <sup>3</sup> , ˙Ipek Ertorun <sup>4</sup> , Betül Can <sup>4</sup> , Belgin Sever <sup>1</sup> , Halide Edip Temel <sup>2</sup> , Özkan Alata¸s <sup>4</sup> and Ahmet Özdemir 1,\***

	- <sup>3</sup> Graduate School of Health Sciences, Anadolu University, 26470 Eski¸sehir, Turkey
	- <sup>4</sup> Department of Medical Biochemistry, Faculty of Medicine, Eskisehir Osmangazi University, 26480 Eski¸sehir, Turkey
	- **\*** Correspondence: mdaltintop@anadolu.edu.tr (M.D.A.); ahmeto@anadolu.edu.tr (A.Ö.); Tel.: +90-222-335-0580 (ext. 3772) (M.D.A); +90-222-335-0580 (ext. 3780) (A.Ö.)

**Abstract:** Targeted therapies have come into prominence in the ongoing battle against non-small cell lung cancer (NSCLC) because of the shortcomings of traditional chemotherapy. In this context, indolebased small molecules, which were synthesized efficiently, were subjected to an in vitro colorimetric assay to evaluate their cyclooxygenase (COX) inhibitory profiles. Compounds **3b** and **4a** were found to be the most selective COX-1 inhibitors in this series with IC<sup>50</sup> values of 8.90 µM and 10.00 µM, respectively. In vitro and in vivo assays were performed to evaluate their anti-NSCLC and anti-inflammatory action, respectively. 2-(1*H*-Indol-3-yl)-*N*0 -(4-morpholinobenzylidene)acetohydrazide (**3b**) showed selective cytotoxic activity against A549 human lung adenocarcinoma cells through apoptosis induction and Akt inhibition. The in vivo experimental data revealed that compound **3b** decreased the serum myeloperoxidase and nitric oxide levels, pointing out its anti-inflammatory action. Moreover, compound **3b** diminished the serum aminotransferase (particularly aspartate aminotransferase) levels. Based on the in vitro and in vivo experimental data, compound **3b** stands out as a lead anti-NSCLC agent endowed with in vivo anti-inflammatory action, acting as a dual COX-1 and Akt inhibitor.

**Keywords:** Akt; anti-inflammatory action; COX-1; hydrazones; non-small cell lung cancer; thiosemicarbazides

#### **1. Introduction**

Non-small cell lung cancer (NSCLC), which accounts for the majority (~85%) of lung cancer cases, is by far the primary cause of cancer-related death throughout the world [1]. Despite significant advances in both diagnosis and treatment, the prognosis for patients with NSCLC still remains poor and the 5-year survival rates of the patients are very low [2]. Surgery, chemotherapy, radiotherapy, immunotherapy, and targeted therapy are existing treatment modalities for NSCLC [3]. Clinical outcomes of patients with NSCLC depend on the cancer stage at the time of diagnosis [4]. The early stages of NSCLC carry the maximum potential for therapeutic intervention and, therefore, its early detection is critical for managing the disease and improving the survival rate [4]. However, there are many challenges in the diagnosis of NSCLC, as it is often asymptomatic early in its course [5,6].

The best treatment option for early stage NSCLC continues to be surgical resection. When the disease is diagnosed at an advanced stage, surgical intervention is no longer an option [7]. In this case, radiotherapy and chemotherapy (e.g., platinum-based chemotherapy) become major therapeutic approaches for unresectable NSCLC [1]. Despite their benefits in NSCLC therapy, conventional chemotherapeutic agents destroy normal cells along with cancer cells and, therefore, these drugs cause severe toxicity and adverse effects [8,9].

**Citation:** Altıntop, M.D.; Akalın Çiftçi, G.; Yılmaz Sava¸s, N.; Ertorun, ˙I.; Can, B.; Sever, B.; Temel, H.E.; Alata¸s, Ö.; Özdemir, A. Discovery of Small Molecule COX-1 and Akt Inhibitors as Anti-NSCLC Agents Endowed with Anti-Inflammatory Action. *Int. J. Mol. Sci.* **2023**, *24*, 2648. https://doi.org/10.3390/ ijms24032648

Academic Editor: Laura Paleari

Received: 5 December 2022 Revised: 29 December 2022 Accepted: 30 December 2022 Published: 31 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Two major barriers to NSCLC management are resistance to radio(chemo)therapy and metastasis [1,9], both of which are the main causes of NSCLC-related mortality [10,11].

The above-mentioned drawbacks have shifted the paradigm of cancer therapy from traditional chemotherapy to targeted therapy, a milestone approach that aims to maximize therapeutic benefits with negligible side effects [3].

The lungs are particularly prone to injury and inflammation since the lungs are continuously exposed to the external environment [12]. Mounting evidence has demonstrated the causal link between chronic inflammation and lung cancer. According to epidemiological data, approximately 20% of cancer-related deaths are associated with unabated inflammation [13]. Chronic inflammation plays a multifaceted role in carcinogenesis; conversely, cancer can also lead to inflammation [12]. Inflammation predisposes to the development of lung cancer [14] and can contribute to tumor initiation, promotion, progression, and metastasis [15]. Targeting inflammation stands out as a rational strategy not only for cancer therapy but also for cancer prevention [16]. Nonsteroidal anti-inflammatory drugs (NSAIDs) significantly diminish the risk of developing certain types of cancer (e.g., colon, lung, breast, and prostate cancer) by reducing tumor-related inflammation [13]. Long-term aspirin use has been reported to reduce the incidence and mortality associated with several cancer types. Several possible mechanisms have been suggested to explain the link between NSAID use and cancer prevention. One of those is cyclooxygenase (COX) inhibition, which reduces the production of inflammatory mediators, particularly prostaglandins (PGs) [16].

COX-1 expression has been reported to be up-regulated in tumorigenesis [17] and implicated in multiple aspects of cancer pathophysiology and, therefore, the inhibition of COX-1, by a variety of selective and nonselective inhibitors, is an emerging approach for pharmacologic intervention in cancer. However, there is only one selective COX-1 inhibitor currently prescribed as an NSAID (mofezolac), just in Japan, for the management of pain and inflammation [17–20].

Akt, also known as protein kinase B (PKB), is one of the most frequently hyperactivated protein kinases in a variety of human cancers including NSCLC [21–23]. Akt overactivation affects several downstream effectors and mediates multiple pathways that promote tumorigenesis (e.g., cell survival, growth, and proliferation) [21]. Furthermore, the hyperactivation of Akt intrinsically up-regulates the nuclear factor-κB (NF-κB) pathway, which transcriptionally initiates pro-inflammatory networks to build up the inflammatory tumor microenvironment [24]. Although diverse small molecule Akt inhibitors have been entered in clinical trials, none of them have been approved [25].

Hydrazides-hydrazones are not only versatile intermediates for the synthesis of various heterocyclic compounds but also commonly occurring motifs in drug-like molecules because of their unique features (e.g., serving as both H-bond donors and acceptors) and diverse pharmacological applications for the management of microbial infections, cancer, and inflammation [26–29]. Hydrazones exert pronounced antitumor action through diverse mechanisms including apoptosis induction, cell cycle arrest, angiogenesis inhibition, and inhibition of a plethora of biological targets related to the pathogenesis of cancer, including Akt [29–36]. Moreover, mounting evidence has demonstrated the anti-inflammatory and/or COX inhibitory potential of hydrazones [37–41].

Thiosemicarbazides are sulfur and nitrogen-containing ligands distinguished by their capability to form complexes with transition metals (e.g., iron, zinc, and copper) [42]. Thiosemicarbazides have aroused great interest not only as intermediates for the synthesis of biologically active heterocycles but also privileged motifs in many bioactive pharmaceutical products [42–45]. Thiosemicarbazides/thiosemicarbazones show a wide range of pharmacological activities ranging from anticancer activity to anti-inflammatory potency due to their unique structural features, allowing them to interact with the pivotal residues of biological targets associated with the pathogenesis of many diseases, particularly cancer and inflammation [42–56]. Triapine, a synthetic thiosemicarbazone, is a small molecule antineoplastic agent endowed with ribonucleotide reductase (RNR) inhibitory activity [42–45].

The indole ranks among the top 25 most common nitrogen heterocycles in U.S. Food and Drug Administration (FDA)-approved drugs. It is also a key structural component of an essential amino acid (tryptophan), a monoamine neurotransmitter (serotonin), and countless natural products (e.g., vinca alkaloids) [57]. The diverse applications of the indole core in challenging diseases (e.g., lung cancer, inflammatory diseases) make it one of the most privileged heterocyclic scaffolds for drug discovery [57,58]. Among vinca alkaloids, vinorelbine is the most frequently used antimitotic drug to treat lung cancer and vinblastine, in combination with cisplatin, is used in the management of NSCLC. Nintedanib (in combination with docetaxel), alectinib, osimertinib, anlotinib, and sunitinib are indole-based anti-NSCLC agents (Figure 1) [59]. In general, indole derivatives have been reported to exert marked anti-NSCLC action through diverse mechanisms including the induction of apoptosis, the inhibition of crucial biological targets such as microtubule, topoisomerases, protein kinases (e.g., Akt), and histone deacetylases (HDACs) [58–62]. The indole is also considered to be one of the most eligible scaffolds for anti-inflammatory drug discovery [62–67]. Indomethacin (Figure 2) is one of the most commonly prescribed NSAIDs exerting its action through the inhibition of COXs. Moreover, several experimental studies have revealed that indomethacin shows significant antiproliferative activity against a broad array of cancer (e.g., colorectal, lung) cell lines [68–72]. *Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 4 of 23

**Figure 1.** Indole-based anti-NSCLC agents. **Figure 1.** Indole-based anti-NSCLC agents.

**Figure 2.** Indomethacin.

targeted therapy of NSCLC.

Taken together, the aforementioned data [26–72] prompted us to design two classes of indole-based small molecules (**3a-j**, **4a-g**) for the targeted therapy of NSCLC. In this context, we performed the synthesis of new hydrazones (**3a-j**) and thiosemicarbazides (**4ag**) efficiently and conducted in vitro and in vivo assays to assess their potential for the

**Figure 1.** Indole-based anti-NSCLC agents.

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 4 of 23

**Figure 2.** Indomethacin. **Figure 2.** Indomethacin.

Taken together, the aforementioned data [26–72] prompted us to design two classes of indole-based small molecules (**3a-j**, **4a-g**) for the targeted therapy of NSCLC. In this context, we performed the synthesis of new hydrazones (**3a-j**) and thiosemicarbazides (**4ag**) efficiently and conducted in vitro and in vivo assays to assess their potential for the targeted therapy of NSCLC. Taken together, the aforementioned data [26–72] prompted us to design two classes of indole-based small molecules (**3a-j**, **4a-g**) for the targeted therapy of NSCLC. In this context, we performed the synthesis of new hydrazones (**3a-j**) and thiosemicarbazides (**4a-g**) efficiently and conducted in vitro and in vivo assays to assess their potential for the targeted therapy of NSCLC. *Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 5 of 23

#### **2. Results**

**2. Results**

tively.

value of 0.13.

The reaction sequence for the preparation of the hitherto unreported small molecules (**3a-j**, **4a-g)** is depicted in Figure 3, starting from 2-(1*H*-indol-3-yl)acetic acid. The convenient and efficient reaction of compound **2** with aromatic aldehydes or ketones and aryl isothiocyanates yielded new hydrazones (**3a-j**), and thiosemicarbazides (**4a-g**), respectively. The reaction sequence for the preparation of the hitherto unreported small molecules (**3a-j**, **4a-g**) is depicted in Figure 3, starting from 2-(1*H*-indol-3-yl)acetic acid. The convenient and efficient reaction of compound **2** with aromatic aldehydes or ketones and aryl isothiocyanates yielded new hydrazones (**3a-j**), and thiosemicarbazides (**4a-g**), respec-

**Figure 3.** The synthetic route for the preparation of compounds **3a-j** and **4a-g**. Reagents and conditions: (i) EtOH, H2SO4, reflux, 12 h; (ii) NH2NH2.H2O, EtOH, reflux, 4 h; (iii) RCHO or RCOR′, EtOH, reflux, 15 h; (iv) R″C6H4NCS, EtOH, rt, 8 h. **Figure 3.** The synthetic route for the preparation of compounds **3a-j** and **4a-g**. Reagents and conditions: (i) EtOH, H2SO<sup>4</sup> , reflux, 12 h; (ii) NH2NH<sup>2</sup> .H2O, EtOH, reflux, 4 h; (iii) RCHO or RCOR<sup>0</sup> , EtOH, reflux, 15 h; (iv) R"C6H4NCS, EtOH, rt, 8 h.

New hydrazones (**3a-j**) and thiosemicarbazides (**4a-g**) were subjected to in vitro assays to determine their COX inhibitory profiles. Among compounds **3a-j**, compound **3a** was found to be a nonselective COX inhibitor with IC50 values of 10.35 µM and 12.50 µM for COX-1 and COX-2, respectively (Table 1). On the other hand, compound **3b** was the most selective COX-1 inhibitor (IC50 = 8.90 µM) in this series with a selectivity index (SI) New hydrazones (**3a-j**) and thiosemicarbazides (**4a-g**) were subjected to in vitro assays to determine their COX inhibitory profiles. Among compounds **3a-j**, compound **3a** was found to be a nonselective COX inhibitor with IC<sup>50</sup> values of 10.35 µM and 12.50 µM for COX-1 and COX-2, respectively (Table 1). On the other hand, compound **3b** was the most selective COX-1 inhibitor (IC<sup>50</sup> = 8.90 µM) in this series with a selectivity index (SI) value of 0.13.

Indomethacin - - 0.12 ± 0.01 0.58 ± 0.08 0.21 Celecoxib - - 8.88 ± 0.38 2.75 ± 0.05 3.23

**3a** 4-(Pyrrolidin-1-yl)phenyl H 10.35 ± 0.35 12.50 ± 0.71 0.83 **3b** 4-Morpholinophenyl H 8.90 ± 0.14 71.00 ± 1.41 0.13 **3c** 4-(Piperidin-1-yl)phenyl H >100 >100 - **3d** 4-(4-Methylpiperazin-1-yl)phenyl H 78.50 ± 8.50 >100 <0.79 **3e** 4-(Methylsulfonyl)phenyl H 83.75 ± 6.25 35.00 ± 9.90 2.39 **3f** 4-(Methylsulfonyl)phenyl CH3 93.75 ± 6.25 51.00 ± 12.73 1.84 **3g** 4-Morpholinophenyl CH3 51.00 ± 1.00 52.50 ± 0.71 0.97 **3h** 4-(2-Morpholinoethoxy)phenyl H 38.50 ± 1.5 50.50 ± 0.70 0.76 **3i** 1-Methyl-1*H*-indol-3-yl H 31.25 ± 1.25 >100 <0.31 **3j** 5-Methoxy-1*H*-indol-3-yl H >100 44.50 ± 6.36 >2.25

**Table 1.** COX inhibitory profiles of compounds **3a-j** and positive controls.


**Figure 3.** The synthetic route for the preparation of compounds **3a-j** and **4a-g**. Reagents and conditions: (i) EtOH, H2SO4, reflux, 12 h; (ii) NH2NH2.H2O, EtOH, reflux, 4 h; (iii) RCHO or RCOR′, EtOH,

New hydrazones (**3a-j**) and thiosemicarbazides (**4a-g**) were subjected to in vitro assays to determine their COX inhibitory profiles. Among compounds **3a-j**, compound **3a** was found to be a nonselective COX inhibitor with IC50 values of 10.35 µM and 12.50 µM for COX-1 and COX-2, respectively (Table 1). On the other hand, compound **3b** was the most selective COX-1 inhibitor (IC50 = 8.90 µM) in this series with a selectivity index (SI)

The reaction sequence for the preparation of the hitherto unreported small molecules (**3a-j**, **4a-g**) is depicted in Figure 3, starting from 2-(1*H*-indol-3-yl)acetic acid. The convenient and efficient reaction of compound **2** with aromatic aldehydes or ketones and aryl isothiocyanates yielded new hydrazones (**3a-j**), and thiosemicarbazides (**4a-g**), respec-

**Table 1.** COX inhibitory profiles of compounds **3a-j** and positive controls. *Int. J. Mol. Sci.* **2023**, *<sup>24</sup>*, x FOR PEER REVIEW <sup>5</sup> of 23 **Table 1.** COX inhibitory profiles of compounds **3a-j** and positive controls.

reflux, 15 h; (iv) R″C6H4NCS, EtOH, rt, 8 h.

value of 0.13.

\* IC<sup>50</sup> for COX-1/IC<sup>50</sup> for COX-2. \* IC50 for COX-1/IC50 for COX-2.

**2. Results**

tively.

Among compounds **4a-g**, compound **4a** was the most selective COX-1 inhibitor (IC<sup>50</sup> = 10.00 µM) (Table 2). Other compounds did not show any inhibitory potency on COX-1 at the tested concentrations. Among compounds **4a-g**, compound **4a** was the most selective COX-1 inhibitor (IC50 = 10.00 µM) (Table 2). Other compounds did not show any inhibitory potency on COX-1 at the tested concentrations.

**Table 2.** COX inhibitory profiles of compounds **4a-g** and positive controls. **Table 2.** COX inhibitory profiles of compounds **4a-g** and positive controls.


\* IC50 for COX-1/IC50 for COX-2. \* IC<sup>50</sup> for COX-1/IC<sup>50</sup> for COX-2.

All compounds were examined for their cytotoxic effects on L929 mouse fibroblast (normal) cells using the MTT test. Based on the in vitro experimental data, compound **3a**, the nonselective COX inhibitor, showed cytotoxicity toward L929 cells with an IC50 value of 17.33 µM (Table 3), which is close to its IC50 values indicated in Table 1. On the other hand, compounds **3b** and **4a** did not show cytotoxicity against L929 cells at their effective concentrations. As a result, compounds **3b** and **4a** (Figure 4), the selective COX-1 inhibitors in this series, were chosen for further studies. All compounds were examined for their cytotoxic effects on L929 mouse fibroblast (normal) cells using the MTT test. Based on the in vitro experimental data, compound **3a**, the nonselective COX inhibitor, showed cytotoxicity toward L929 cells with an IC<sup>50</sup> value of 17.33 µM (Table 3), which is close to its IC<sup>50</sup> values indicated in Table 1. On the other hand, compounds **3b** and **4a** did not show cytotoxicity against L929 cells at their effective concentrations. As a result, compounds **3b** and **4a** (Figure 4), the selective COX-1 inhibitors in this series, were chosen for further studies.

> **Compound IC50** (**µM**) **3a** 17.33 ± 2.08 **3b** 176.67 ± 5.77 **3c** 21.33 ± 0.58 **3d** 20.00 ± 1.73 **3e** 85.00 ± 17.32 **3f** <3.90 **3g** <3.90

**Figure 4.** Selective COX-1 inhibitors in this series.

**Table 3.** IC50 values of all compounds for L929 cells.


Among compounds **4a-g**, compound **4a** was the most selective COX-1 inhibitor (IC50 = 10.00 µM) (Table 2). Other compounds did not show any inhibitory potency on COX-1

**Table 3.** IC<sup>50</sup> values of all compounds for L929 cells. **Compound R″ IC50** (**µM**) **SI \* COX-1 COX-2**

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 6 of 23

**Table 2.** COX inhibitory profiles of compounds **4a-g** and positive controls.

\* IC50 for COX-1/IC50 for COX-2.

at the tested concentrations.

**Figure 4.** Selective COX-1 inhibitors in this series. **Figure 4.** Selective COX-1 inhibitors in this series.

tors in this series, were chosen for further studies.

**Table 3.** IC50 values of all compounds for L929 cells. **Compound IC50** (**µM**) **3a** 17.33 ± 2.08 **3b** 176.67 ± 5.77 **3c** 21.33 ± 0.58 Compounds **3b** and **4a** were also subjected to the MTT assay to assess their cytotoxicity toward A549 human lung adenocarcinoma cell line. Based on the data presented in Table 4, compound **3b** was found to be the most potent anticancer agent on A549 cells with an IC<sup>50</sup> value of 89.67 µM compared to cisplatin (IC<sup>50</sup> = 22.67 µM). On the other hand, compound **4a** showed cytotoxic activity against A549 cells with an IC<sup>50</sup> value of 179.33 µM.



After 24 h incubation of A549 cells treated with compounds **3b** and **4a** in this series and cisplatin, flow cytometry-based apoptosis detection assay was performed to identify early and late apoptotic cells using Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) staining. The percentages of A549 cells undergoing apoptosis (early and late) exposed to compounds **3b** and cisplatin at their IC50/2 concentrations were found to be 11.67% and 6.57%, respectively. On the other hand, the percentages of A549 cells undergoing apoptosis (early and late) exposed to compounds **3b** and cisplatin at their IC<sup>50</sup> concentrations were 12.85% and 4.46%, respectively (Table 5, Figure 5). The percentages of A549 cells undergoing early and late apoptosis exposed to compound **4a** at its IC50/4 concentration were 7.34% and 5.19%, respectively. On the other hand, the percentages of A549 cells undergoing early and late apoptosis exposed to compound **4a** at its IC50/2 concentration were found to be 7.07% and 6.62%, respectively (Table 5, Figure 5).


**Table 5.** Percentages of typical quadrant analysis of Annexin V FITC/PI flow cytometry of A549 cells treated with compounds **3b**, **4a,** and cisplatin.

A549 cells were cultured for 24 h in medium with compound **4a** (at its IC50/4 and IC50/2 concentrations), compound **3b,** and cisplatin (at their IC50/2 and IC<sup>50</sup> concentrations). At least 10,000 cells were analyzed per sample, and quadrant analysis was performed. *Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 8 of 23

**Figure 5.** *Cont*.

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 9 of 23

**Figure 5.** Flow cytometric analysis of A549 cells treated with IC50/2 and IC50 concentrations of compounds **3b**, **4a,** and cisplatin. At least 10,000 cells were analyzed per sample, and quadrant analysis was performed. Q1-LR, Q1-UR, Q1-LL, and Q1-UL quadrants represent early apoptosis, late apoptosis, viability, and necrosis, respectively. (**a**) Control; (**b**) Control; (**c**) Compound **3b** at IC50/2 concentration; (**d**) Compound **3b** at IC50 concentration; (**e**) Compound **4a** at IC50/4 concentration; (**f**) Compound **4a** at IC50/2 concentration; (**g**) Cisplatin at IC50/2 concentration; (**h**) Cisplatin at IC50 concentration. Akt inhibition caused by compounds **3b** and **4a** in A549 cells was examined using a colorimetric assay. Compounds **3b** and **4a** caused Akt inhibition in A549 cell line with IC50 **Figure 5.** Flow cytometric analysis of A549 cells treated with IC50/2 and IC<sup>50</sup> concentrations of compounds **3b**, **4a,** and cisplatin. At least 10,000 cells were analyzed per sample, and quadrant analysis was performed. Q1-LR, Q1-UR, Q1-LL, and Q1-UL quadrants represent early apoptosis, late apoptosis, viability, and necrosis, respectively. (**a**) Control; (**b**) Control; (**c**) Compound **3b** at IC50/2 concentration; (**d**) Compound **3b** at IC<sup>50</sup> concentration; (**e**) Compound **4a** at IC50/4 concentration; (**f**) Compound **4a** at IC50/2 concentration; (**g**) Cisplatin at IC50/2 concentration; (**h**) Cisplatin at IC<sup>50</sup> concentration.

values of 32.50 and 45.33 µM as compared to GSK690693 (IC50= 5.93 µM) (Table 6). **Table 6.** Akt inhibitory effects of compounds **3b**, **4a**, GSK690693, and cisplatin in A549 cells. **Compound IC50** (**µM**) Akt inhibition caused by compounds **3b** and **4a** in A549 cells was examined using a colorimetric assay. Compounds **3b** and **4a** caused Akt inhibition in A549 cell line with IC<sup>50</sup> values of 32.50 and 45.33 µM as compared to GSK690693 (IC50= 5.93 µM) (Table 6).

3b 32.50 ± 4.95

the LPS + compound **4a** group significantly decreased the MPO activity compared to the


4a 45.33 ± 6.51 **Table 6.** Akt inhibitory effects of compounds **3b**, **4a**, GSK690693, and cisplatin in A549 cells.

LPS group (*p* < 0.05). The decrease in the MPO activity caused by compound **4a** was higher than that caused by the indomethacin therapy. **Table 7.** Effects of compounds **3b**, **4a,** and indomethacin on MPO levels. **Groups MPO** (**U/L**) Control 1.03 ± 0.51 LPS 1.79 ± 0.27 LPS + Compound 3b 1.50 ± 0.81 LPS + Compound 4a 0.84 ± 0.26 # LPS + Indomethacin 1.06 ± 0.73 The lipopolysaccharide (LPS)-induced sepsis model was used to assess the in vivo anti-inflammatory activities of compounds **3b** and **4a**. According to the data indicated in Table 7, the myeloperoxidase (MPO) activity of the LPS group increased as compared to the control group. However, this increase is not statistically significant. The LPS + compound **3b** group slightly decreased the MPO activity compared to the LPS group, while the LPS + compound **4a** group significantly decreased the MPO activity compared to the LPS group (*p* < 0.05). The decrease in the MPO activity caused by compound **4a** was higher than that caused by the indomethacin therapy.

**Table 7.** Effects of compounds **3b**, **4a,** and indomethacin on MPO levels.


Values are given as mean <sup>±</sup> standard deviation (SD). Significance according to LPS values, # : *p* < 0.05. One-way ANOVA, post-hoc Tukey test *n* = 8.

As presented in Table 8, there was a significant increase in the nitric oxide (NO) level after LPS administration compared to the control (*p* < 0.001). LPS + compound **3b**, LPS + compound **4a**, and LPS + indomethacin caused a significant decrease in the serum NO levels. However, this decrease in LPS + compound **3b** was similar to the control. The NO level was significantly higher in the LPS group than in the control group, while it was markedly lower in the compound **4a** pre-treatment group compared to the LPS group (*p* < 0.05).


**Table 8.** Effects of compounds **3b**, **4a,** and indomethacin on NO levels (µmol/L).

Significance relative to control values, \*: *p* < 0.05, \*\*\*: *p* < 0.001. One-way ANOVA, Kruskal–Wallis test, *n* = 8.

According to the in vivo experiments, the alanine aminotransferase (ALT) level decreased in all groups compared to the LPS group (Table 9). This decrease was greater in the group treated with compound **4a** compared to the group treated with compound **3b**. Likewise, the aspartate aminotransferase (AST) level decreased in all groups compared to the LPS group. This decrease was greater in the group treated with compound **3b** compared to the group treated with compound **4a** and the group treated with indomethacin. However, the decrease caused by the compounds in the ALT and AST levels was not statistically significant compared to the LPS group.

**Table 9.** Effects of compounds **3b**, **4a,** and indomethacin on ALT and AST levels.


Values are given as mean ± SD. One-way ANOVA, post-hoc Tukey test, *n* = 8.

#### **3. Discussion**

Experimental studies have demonstrated that hydrazones show marked antitumor action through various mechanisms, including the inhibition of Akt [31] or the phosphatidylinositol 3 kinase (PI3K)/Akt signaling pathway [32,36]. *N*0 -benzylidene-2-[(4-(4-methoxyphenyl)pyrimid in-2-yl)thio]acetohydrazide was previously reported to exert marked anticancer activity against the 5RP7 H-*ras* oncogene transformed rat embryonic fibroblast cell line via the induction of apoptosis and the inhibition of Akt (IC<sup>50</sup> = 0.50 µg/mL) [31]. According to western blot data reported by Han et al., (*S*)-2-{[5-[1-(6-methoxynaphtalene-2-yl)ethyl]-4-(4-fluorophenyl)- 4*H*-1,2,4-triazole-3-yl]thio}-*N*0 -[(5-nitrofuran-2-yl)methylidene]acetohydrazide caused a significant decrease in the epidermal growth factor receptor (EGFR), PI3K, and Akt phosphorylation in PC3 human prostate cancer cells [32]. Bak et al. indicated that 5-hydroxy-7,40 diacetyloxyflavanone-*N*-phenyl hydrazone (N101-43) induced apoptosis via the up-regulation of Fas/FasL expression, the activation of caspase cascade, and the inhibition of the PI3K/Akt signaling pathway in NSCLC cells [36].

The anti-inflammatory and/or COX inhibitory potential of hydrazones was demonstrated by in vitro and in vivo studies [37–41]. In our previous work, 2-[(1-methyl-1*H*tetrazol-5-yl)thio]-*N*0 -(4-(piperidin-1-yl)benzylidene)acetohydrazide and 2-[(1-methyl-1*H*tetrazol-5-yl)thio]-*N*0 -(4-(morpholin-4-yl)benzylidene)acetohydrazide caused selective COX-1 inhibition [37].

Thiosemicarbazides show pronounced antiproliferative activity toward a variety of tumor cells through diverse mechanisms [42–52]. Our research team reported that 4-(1,3 benzodioxol-5-yl)-1-([1,10 -biphenyl]-4-ylmethylene)thiosemicarbazide showed remarkable anticancer activity against A549 human lung adenocarcinoma and C6 rat glioma cells through apoptosis induction mediated by the disruption of ∆Ψm [52].

In vitro and in vivo experimental data revealed that thiosemicarbazides exert marked anti-inflammatory action through several mechanisms including COX inhibition [53–56]. In our recent work [53], 4-[4-(piperidin-1-ylsulfonyl)phenyl]-1-[4-(4-cyanophenoxy)benzylidene]t hiosemicarbazide was found to be a selective COX-1 inhibitor with an IC<sup>50</sup> value of 1.89 µM. On the other hand, 4-[4-(piperidin-1-ylsulfonyl)phenyl]-1-[4-(4-nitrophenoxy)benzylidene]thio semicarbazide was determined to be a nonselective COX inhibitor (COX-1 IC<sup>50</sup> = 13.44 µM, COX-2 IC<sup>50</sup> = 12.60 µM). Based on the LPS-induced sepsis model, these agents diminished the MPO, NO, high-sensitivity C-reactive protein (hsCRP), malondialdehyde (MDA), ALT, and AST levels. Both compounds were identified as potential anti-inflammatory agents [53].

Indole-based small molecules exert a notable anti-NSCLC action through multiple mechanisms such as the induction of apoptosis and the inhibition of crucial biological targets including protein kinases (e.g., Akt) [58–62]. Furthermore, mounting evidence has demonstrated that indole derivatives show marked anti-inflammatory action via COX inhibition [62–67]. In our previous study [65], 3-(5-bromo-1*H*-indol-3-yl)-1- (4-cyanophenyl)prop-2-en-1-one was found to be a nonselective COX inhibitor (COX-1 IC<sup>50</sup> = 8.10 µg/mL, COX-2 IC<sup>50</sup> = 9.50 µg/mL), while 3-(5-methoxy-1*H*-indol-3-yl)-1-(4- (methylsulfonyl)phenyl)prop-2-en-1-one inhibited COX-1 (IC<sup>50</sup> = 8.60 µg/mL). According to the LPS-induced sepsis model, both compounds markedly decreased the MPO, NO, hsCRP, MDA, ALT, and AST levels. Both indole derivatives were identified as potential anti-inflammatory agents [65].

Based on the aforementioned studies [26–72], two classes of indole-based small molecules (**3a-j**, **4a-g**) for the targeted therapy of NSCLC were designed. In this context, we carried out the synthesis of new hydrazones (**3a-j**) and thiosemicarbazides (**4a-g**) efficiently and performed in vitro and in vivo experiments to assess their potential for the targeted therapy of NSCLC.

Among compounds **3a-j**, compound **3a** was determined to be a nonselective COX inhibitor with IC<sup>50</sup> values of 10.35 µM and 12.50 µM for COX-1 and COX-2, respectively, while compound **3b** was found to be a selective COX-1 inhibitor (IC<sup>50</sup> = 8.90 µM). Compound **3b** exhibited COX-2 inhibitory activity with an IC<sup>50</sup> value of 71.00 µM. The SI values of compounds **3a** and **3b** were determined to be 0.83 and 0.13, respectively. In particular, the pyrrolidine ring enhanced the inhibitory effects on both COXs, whereas the morpholine substitution caused selective COX-1 inhibitory potency. The replacement of the morpholine ring (compound **3b**) with the piperidine ring (compound **3c**) or the piperazine ring (compound **3d**) led to a substantial drop in COX-1 inhibitory activity. Compound **3c**, carrying a piperidine ring, showed the lowest COX inhibition (>100 µM) in this series.

According to the in vitro data related to the inhibitory effects of compounds **3b** and **3h** on COXs, it can be concluded that the ethoxy linker between the morpholine and the benzene rings diminishes COX-1 inhibition, while it enhances COX-2 inhibition. Taking into account the inhibitory effects of compounds **3b** and **3g** on COXs, it is important to note that the methyl branching decreases COX-1 inhibition and increases COX-2 inhibition.

The SI values of compounds **3e** and **3f** were found to be 2.39 and 1.84, respectively, indicating that the methylsulfonyl group significantly enhances COX-2 selectivity.

Based on the experimental results related to the inhibitory effects of compounds **3i** and **3j** on COXs, the methyl substituent at the 1st position of the indole scaffold enhances COX-1 selectivity, while the methoxy substituent at the 5th position of the indole core enhances COX-2 selectivity.

Among compounds **4a-g**, the most selective COX-1 inhibitor was found to be compound **4a** (IC<sup>50</sup> = 10.00 µM, SI = 0.13). It can be concluded that the bromo substituent at the 4th position of the phenyl moiety significantly enhanced the COX-1 inhibitory potency. Other compounds did not exhibit any inhibitory activity towards COX-1 at the tested concentrations. Thiosemicarbazides tested in this work, except for compound **4a**, were found to have a tendency to inhibit the COX-2 enzyme.

Among the indole-based small molecules (**3a-j**, **4a-g**), compounds **3b** and **4a**, selective COX-1 inhibitors in this series, were chosen for further studies since both compounds did not exert cytotoxicity toward L929 (normal) cells at their effective concentrations reported for their COX-1 inhibitory activity.

To investigate their potential as anti-NSCLC agents, their cytotoxic effects on A549 cells were evaluated by means of the MTT assay protocol. Based on the experimental data, compound **3b** was the most potent anticancer agent on A549 cell line with an IC<sup>50</sup> value of 89.67 µM. It can be concluded that the anticancer activity of compound **3b** against A549 cells is selective since the IC<sup>50</sup> value of compound **3b** for L929 cells is 176.67 µM. On the other hand, compound **4a** showed cytotoxic activity against A549 and L929 cells with IC<sup>50</sup> values of 179.33 µM and 84.00 µM, respectively. The cytotoxic activity of compound **4a** toward A549 cells was found to be nonselective at its IC<sup>50</sup> value. For this reason, the IC50/4 and IC50/2 concentrations of compound **4a** were applied in the flow cytometry analyses of apoptosis and the Akt inhibition assay.

In cancer, cells lose their ability to undergo apoptosis, resulting in uncontrolled proliferation [73]. The induction of apoptosis is reported to be an intriguing modality for the management of all types of cancer since apoptosis evasion is a hallmark of cancer and is nonspecific to the cause or the type of the cancer [74]. Based on the flow cytometry-based apoptosis detection assay performed in this work, A549 cells treated with compounds **3b** and **4a** underwent apoptosis, pointing out their apoptosis-inducing efficacy.

Akt participates in the pathogenesis of NSCLC and, therefore, the inhibition of Akt by natural and synthetic agents stands out as a rational strategy for cancer therapy [21–25]. The colorimetric assay conducted in this study revealed that the Akt inhibitory activity of compound **3b** (IC<sup>50</sup> = 32.50 µM) was more notable than that of compound **4a** (IC<sup>50</sup> = 45.33 µM) in A549 cells.

Sepsis is described as a life-threatening organ dysfunction provoked by a dysregulated host response to infection [53]. Inflammatory imbalance plays a fundamental role in the pathogenesis of sepsis and occurs throughout the whole process of sepsis [75]. The parameters related to inflammation are crucial for evaluating a sepsis case [76]. In this work, the LPS-induced sepsis model was used to evaluate the in vivo anti-inflammatory activities of compounds **3b** and **4a**.

MPO is linked to several diseases, particularly those in which strong infiltration of polymorphonuclear cells (PMNs) and acute or chronic inflammation are involved. MPO contributes to the pathophysiology of diverse diseases such as rheumatoid arthritis, atherosclerosis, pulmonary fibrosis, renal glomerular injury, multiple sclerosis, Huntington's disease, Alzheimer's disease, Parkinson's disease, liver diseases, diabetes, obesity, and cancer. MPO is reported to promote tumor initiation and progression. MPO participates in the regulation of tumor growth, apoptosis, migration, and metastasis [77]. In this work, compound **3b** caused a slight decrease in the MPO activity compared to the LPS group, whereas compound **4a** significantly diminished the MPO activity compared to the LPS group (*p* < 0.05).

Sepsis is characterized by a robust rise in NO production throughout the body that is driven by inducible NO synthase (iNOS) [78]. Due to the key role of NO in the pathogenesis of inflammation as a signaling molecule [78–80], herein the effects of compounds **3b** and **4a** on the serum NO levels were evaluated. The in vivo experimental data revealed that compounds **3b** and **4a** diminished the serum NO levels.

Aminotransferases, also referred to as transaminases, are commonly used as markers of hepatocellular injury in nonclinical toxicology studies and clinical trials. In general, aminotransferase activity in blood (serum or plasma) is elevated in the hepatocellular damage induced by diseases or drugs such as anti-inflammatory drugs [81–84]. Based on the in vivo experimental data performed in this work, both compounds caused a decrease

in the serum aminotransferase levels. In particular, compound **3b** diminished the serum AST level more than indomethacin.

Taking into account the knowledge obtained from the in vitro and in vivo assays, compound **3b** can be considered as a lead compound for the targeted therapy of NSCLC due to its direct cytotoxic effects on A549 cells as well as its possible effects on the tumor microenvironment (e.g., tumor-related inflammation).

#### **4. Materials and Methods**

#### *4.1. Chemistry*

The chemicals were procured from commercial suppliers and were used without further purification. Melting points (M.p.) were determined on the Electrothermal IA9200 digital melting point apparatus (Staffordshire, UK) and were uncorrected. Thin Layer Chromatography (TLC) was performed on TLC Silica gel 60 F254 aluminum sheets (Merck, Darmstadt, Germany) using petroleum ether:ethyl acetate solvent system (1:1). IR spectra were recorded on the IRPrestige-21 Fourier Transform Infrared spectrophotometer (Shimadzu, Tokyo, Japan). <sup>1</sup>H and <sup>13</sup>C NMR spectra were recorded on the Varian Mercury 400 NMR spectrometer (Agilent, Palo Alto, CA, USA). HRMS spectra were recorded on the LC/MS IT-TOF system (Shimadzu, Tokyo, Japan) using the electrospray ionization (ESI) technique.

#### 4.1.1. Preparation of ethyl 2-(1*H*-indol-3-yl)acetate (**1**)

Compound **1** was synthesized starting from 2-(1*H*-indol-3-yl)acetic acid according to a previous work [85].

#### 4.1.2. Preparation of 2-(1*H*-indol-3-yl)acetohydrazide (**2**)

Compound **2** was obtained by the reaction of compound **1** with hydrazine hydrate according to a previous work [85].

4.1.3. General Method for the Preparation of *N*0 -benzylidene/(1-arylethylidene)-2-(1*H*indol-3-yl)acetohydrazide Derivatives (**3a-j**)

A mixture of compound **2** and aromatic aldehyde or ketone in ethanol was heated under reflux for 15 h. At the end of this period, the precipitate was filtered off and dried. The product was crystallized from ethanol.

2-(1*H*-Indol-3-yl)-*N*0 -[4-(pyrrolidin-1-yl)benzylidene]acetohydrazide (**3a**)

Yield: 78%. M.p.: 302–303 ◦C. IR νmax (cm−<sup>1</sup> ): 3196.05, 3074.53, 3043.67, 2966.52, 2914.44, 2873.94, 2848.86, 1668.43, 1595.13, 1546.91, 1521.84, 1487.12, 1460.11, 1431.18, 1386.82, 1350.17, 1323.17, 1292.31, 1249.87, 1224.80, 1174.65, 1163.08, 1118.71, 1047.35, 1001.06, 983.70, 958.62, 929.69, 914.26, 856.39, 804.32, 719.45, 682.80. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 2.02–2.05 (m, 4H), 3.40–3.42 (m, 4H), 3.60 and 4.02 (2s, 2H), 6.92–7.07 (m, 4H), 7.21 (dd, *J =* 2.4 Hz, 12.8 Hz, 1H), 7.31–7.35 (m, 1H), 7.48–7.59 (m, 3H), 7.88 and 8.08 (2s, 1H), 10.85 and 10.88 (2s, 1H), 11.05 and 11.28 (2s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 25.41 (2CH2), 32.13 (CH2), 47.70 (2CH2), 108.79 (C), 111.74 (CH), 115.13 (2CH), 118.74 (CH), 119.19 (CH), 121.33 (CH), 124.32 (CH), 124.75 (C), 127.90 (C), 128.27 (2CH), 136.45 (C), 146.90 (CH), 152.26 (C), 172.72 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C21H22N4O: 347.1866, found: 347.1864.

2-(1*H*-Indol-3-yl)-*N*0 -(4-morpholinobenzylidene)acetohydrazide (**3b**)

Yield: 85%. M.p.: 306–307 ◦C. IR νmax (cm−<sup>1</sup> ): 3275.13, 3178.69, 3055.24, 2964.59, 2922.16, 2870.08, 2825.72, 1660.71, 1604.77, 1558.48, 1541.12, 1519.91, 1506.41, 1489.05, 1456.26, 1446.61, 1425.40, 1392.61, 1375.25, 1338.60, 1313.52, 1301.95, 1259.52, 1224.80, 1186.22, 1176.58, 1159.22, 1109.07, 1095.57, 1062.78, 1045.42, 1006.84, 958.62, 921.97, 875.68, 858.32, 846.75, 823.60, 798.53, 786.96, 742.59, 682.80. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.19 (t, *J =* 4.41 Hz, 4.62 Hz, 4H), 3.72–3.75 (m, 4H), 3.61 and 4.02 (2s, 2H), 6.92–7.08 (m, 4H), 7.21 (dd, *J =* 2.4 Hz, 12.8 Hz, 1H), 7.31–7.35 (m, 1H), 7.49–7.60 (m, 3H), 7.88 and 8.07 (2s, 1H), 10.85 and 10.89 (2s, 1H), 11.05 and 11.28 (2s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 32.13 (CH2), 53.79 (2CH2), 66.40 (2CH2), 108.79 (C), 111.74 (CH), 115.13 (2CH), 118.74 (CH), 119.19 (CH), 121.33 (CH), 124.32 (CH), 124.75 (C), 127.91 (C), 128.27 (2CH), 136.45 (C), 146.90 (CH), 152.29 (C), 172.73 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C21H22N4O2: 363.1816, found: 363.1824.

2-(1*H*-Indol-3-yl)-*N*0 -[4-(piperidin-1-yl)benzylidene]acetohydrazide (**3c**)

Yield: 80%. M.p.: 265–266 ◦C. IR νmax (cm−<sup>1</sup> ): 3203.76, 3082.25, 3034.03, 2972.31, 2935.66, 2856.58, 2825.72, 1668.43, 1598.99, 1552.70, 1514.12, 1448.54, 1427.32, 1384.89, 1350.17, 1282.66, 1247.94, 1220.94, 1182.36, 1124.50, 1024.20, 962.48, 914.26, 858.32, 804.32, 721.38, 651.94. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 1.58 (brs, 6H), 3.32 (brs, 4H), 3.60 and 4.02 (2s, 2H), 6.92–7.08 (m, 4H), 7.21 (dd, *J =* 2.4 Hz, 12.8 Hz, 1H), 7.31–7.35 (m, 1H), 7.48–7.59 (m, 3H), 7.88 and 8.08 (2s, 1H), 10.85 and 10.89 (2s, 1H), 11.05 and 11.28 (2s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 24.40 (CH2), 25.42 (2CH2), 32.12 (CH2), 48.95 (2CH2), 108.79 (C), 111.74 (CH), 115.13 (2CH), 118.74 (CH), 119.19 (CH), 121.33 (CH), 124.32 (CH), 124.75 (C), 127.91 (C), 128.27 (2CH), 136.45 (C), 146.90 (CH), 152.26 (C), 172.73 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C22H24N4O: 361.2023, found: 361.2031.

2-(1*H*-Indol-3-yl)-*N*0 -[4-(4-methylpiperazin-1-yl)benzylidene]acetohydrazide (**3d**)

Yield: 81%. M.p.: 218–220 ◦C. IR νmax (cm−<sup>1</sup> ): 3398.57, 3205.69, 3165.19, 3111.18, 3043.67, 2939.52, 2883.58, 2831.50, 1649.49, 1602.85, 1517.98, 1446.61, 1427.32, 1409.96, 1377.17, 1340.53, 1286.52, 1232.51, 1184.29, 1159.22, 1141.86, 1124.50, 1105.21, 1080.14, 1001.06, 956.69, 943.19, 921.97, 806.25, 794.67, 742.59, 686.66. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 2.21 (s, 3H), 2.40–2.44 (m, 4H), 3.20–3.22 (m, 4H), 3.59 and 4.02 (2s, 2H), 6.92–7.08 (m, 4H), 7.21 (dd, *J =* 2.4 Hz, 12.8 Hz, 1H), 7.31–7.35 (m, 1H), 7.48–7.59 (m, 3H), 7.88 and 8.08 (2s, 1H), 10.85 and 10.89 (2s, 1H), 11.05 and 11.28 (2s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 32.13 (CH2), 46.21 (CH3), 47.69 (2CH2), 54.89 (2CH2), 108.79 (C), 111.74 (CH), 115.13 (2CH), 118.74 (CH), 119.19 (CH), 121.33 (CH), 124.32 (CH), 124.75 (C), 127.91 (C), 128.27 (2CH), 136.45 (C), 146.90 (CH), 152.28 (C), 172.73 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C22H25N5O: 376.2132, found: 376.2148.

2-(1*H*-Indol-3-yl)-*N*0 -(4-methylsulfonylbenzylidene)acetohydrazide (**3e**)

Yield: 86%. M.p.: 264–265 ◦C. IR νmax (cm−<sup>1</sup> ): 3344.57, 3205.69, 3055.24, 2929.87, 2897.08, 1668.43, 1604.77, 1556.55, 1489.05, 1454.33, 1408.04, 1365.60, 1328.95, 1313.52, 1290.38, 1242.16, 1222.87, 1199.72, 1145.72, 1089.78, 1055.06, 1018.41, 983.70, 972.12, 956.69, 943.19, 869.90, 835.18, 792.74, 769.60, 750.31, 729.09, 686.66, 651.94. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.24 (s, 3H), 3.67 and 4.09 (2s, 2H), 6.94–7.09 (m, 2H), 7.25 (dd, *J =* 2.4 Hz, 9.6 Hz, 1H), 7.34 (t, *J =* 8.0 Hz, 8.4 Hz, 1H), 7.56–7.60 (m, 1H), 7.90–7.96 (m, 4H), 8.07 and 8.31 (2s, 1H), 10.87 and 10.93 (2s, 1H), 11.52 and 11.76 (2s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 31.87 (CH2), 43.64 (CH3), 108.10 (C), 111.51 (CH), 118.53 (CH), 118.87 (CH), 121.12 (CH), 124.18 (CH), 127.31 (C), 127.50 (2CH), 127.70 (2CH), 136.19 (C), 139.35 (C), 141.21 (C), 144.37 (CH), 173.17 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C18H17N3O3S: 356.1063, found: 356.1071.

2-(1*H*-Indol-3-yl)-*N*0 -[1-(4-methylsulfonylphenyl)ethylidene]acetohydrazide (**3f**)

Yield: 81%. M.p.: 204–205 ◦C. IR νmax (cm−<sup>1</sup> ): 3342.64, 3190.26, 3088.03, 3032.10, 3005.10, 2924.09, 2848.86, 1668.43, 1585.49, 1562.34, 1489.05, 1456.26, 1417.68, 1394.53, 1338.60, 1296.16, 1280.73, 1226.73, 1188.15, 1145.72, 1093.64, 1070.49, 1008.77, 977.91, 964.41, 852.54, 839.03, 788.89, 758.02, 740.67, 717.52, 700.16, 688.59. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 2.30 (s, 3H), 3.26 (s, 3H), 3.81 and 4.12 (2s, 2H), 6.95–7.07 (m, 2H), 7.20–7.36 (m, 2H), 7.54–7.61 (m, 1H), 8.01 (d, *J =* 8.0 Hz, 2H), 8.15 (d, *J =* 8.8 Hz, 2H), 10.63 and 10.66 (2s, 1H), 10.84 and 10.89 (2s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 14.96 (CH3), 31.87 (CH2), 43.42 (CH3), 108.10 (C), 111.26 (CH), 118.25 (CH), 118.64 (CH), 121.12 (CH), 123.84 (CH), 126.70 (C), 127.11 (2CH), 127.39 (2CH), 141.55 (C), 142.06 (C), 142.97 (C), 156.20 (C), 173.17 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C19H19N3O3S: 370.1220, found: 370.1202.

2-(1*H*-Indol-3-yl)-*N*0 -[1-(4-morpholinophenyl)ethylidene]acetohydrazide (**3g**)

Yield: 79%. M.p.: 198–199 ◦C. IR νmax (cm−<sup>1</sup> ): 3269.34, 3080.32, 3047.53, 2966.52, 2916.37, 2848.86, 1668.43, 1608.63, 1593.20, 1546.91, 1516.05, 1454.33, 1444.68, 1417.68, 1379.10, 1361.74, 1340.53, 1301.95, 1263.37, 1236.37, 1197.79, 1118.71, 1068.56, 1051.20, 1026.13, 937.40, 923.90, 864.11, 821.68, 798.53, 742.59, 729.09, 648.08. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 2.31 (s, 3H), 3.19 (t, *J =* 4.41 Hz, 4.62 Hz, 4H), 3.72–3.75 (m, 4H), 3.60 and 4.02 (2s, 2H), 6.92–7.08 (m, 4H), 7.21 (dd, *J =* 2.4 Hz, 12.8 Hz, 1H), 7.31–7.35 (m, 1H), 7.48–7.59 (m, 3H), 10.85 and 10.89 (2s, 1H), 11.05 and 11.27 (2s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 14.96 (CH3), 32.13 (CH2), 53.79 (2CH2), 66.40 (2CH2), 108.79 (C), 111.74 (CH), 115.13 (2CH), 118.74 (CH), 119.19 (CH), 121.33 (CH), 124.32 (CH), 124.75 (C), 127.91 (C), 128.27 (2CH), 136.45 (C), 143.30 (C), 156.18 (C), 172.73 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C22H24N4O2: 377.1972, found: 377.1982.

2-(1*H*-Indol-3-yl)-*N*0 -[4-(2-morpholinoethoxy)benzylidene]acetohydrazide (**3h**)

Yield: 83%. M.p.: 132–135 ◦C. IR νmax (cm−<sup>1</sup> ): 3383.14, 3319.49, 3196.05, 3045.60, 2958.80, 2918.30, 2850.79, 1664.57, 1604.77, 1548.84, 1510.26, 1456.26, 1421.54, 1355.96, 1340.53, 1303.88, 1240.23, 1201.65, 1170.79, 1116.78, 1049.28, 1010.70, 983.70, 952.84, 925.83, 860.25, 831.32, 742.59, 646.15. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 2.45–2.47 (m, 4H), 2.66–2.70 (m, 2H), 3.55–3.58 (m, 4H), 4.04 (s, 2H), 4.08–4.12 (m, 2H), 6.95–7.07 (m, 4H), 7.23 (dd, *J =* 2.4 Hz, 12.4 Hz, 1H), 7.34 (t, *J =* 8.0 Hz, 1H), 7.57–7.65 (m, 3H), 7.94 and 8.16 (2s, 1H), 10.86 and 10.91 (2s, 1H), 11.15 and 11.39 (2s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 31.85 (CH2), 53.79 (2CH2), 57.12 (CH2), 65.64 (CH2), 66.35 (2CH2), 108.44 (C), 111.48 (CH), 115.05 (2CH), 118.47 (CH), 118.91 (CH), 121.07 (CH), 124.07 (CH), 127.18 (C), 127.63 (C), 128.41 (2CH), 136.19 (C), 146.18 (CH), 159.90 (C), 172.64 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C23H26N4O3: 407.2078, found: 407.2071.

2-(1*H*-Indol-3-yl)-*N*0 -[(1-methyl-1*H*-indol-3-yl)methylene]acetohydrazide (**3i**)

Yield: 84%. M.p.: 221–224 ◦C. IR νmax (cm−<sup>1</sup> ): 3414.00, 3147.83, 3101.54, 3061.03, 2980.02, 2945.30, 2908.65, 2819.93, 1651.07, 1612.49, 1570.06, 1539.20, 1502.55, 1462.04, 1452.40, 1421.54, 1404.18, 1377.17, 1346.31, 1332.81, 1321.24, 1253.73, 1244.09, 1197.79, 1157.29, 1139.93, 1120.64, 1087.85, 1072.42, 1045.42, 1008.77, 948.98, 933.55, 900.76, 856.39, 808.17, 785.03, 744.52, 734.88, 673.16. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.64 and 4.13 (2s, 2H), 3.79 (s, 3H), 6.94–7.18 (m, 3H), 7.22–7.28 (m, 2H), 7.36 (t, *J =* 8.4 Hz, 8.8 Hz, 1H), 7.47 (t, *J =* 8.4 Hz, 9.2 Hz, 1H), 7.64 (t, *J =* 7.6 Hz, 1H), 7.73 (d, *J =* 2.4 Hz, 1H), 8.20–8.38 (m, 2H), 10.86 and 10.92 (2s, 1H), 10.99 and 11.19 (2s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 31.94 (CH2), 32.90 (CH3), 108.74 (C), 110.42 (CH), 110.83 (CH), 111.49 (C), 118.50 (CH), 118.93 (CH), 120.89 (CH), 121.10 (CH), 121.86 (CH), 122.80 (CH), 123.96 (CH), 124.70 (C), 127.70 (C), 133.81 (CH), 136.23 (C), 137.76 (C), 142.98 (CH), 172.09 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C20H18N4O: 331.1553, found: 331.1538.

2-(1*H*-Indol-3-yl)-*N*0 -[(5-methoxy-1*H*-indol-3-yl)methylene]acetohydrazide (**3j**)

Yield: 82%. M.p.: 231–233 ◦C. IR νmax (cm−<sup>1</sup> ): 3415.93, 3373.50, 3049.46, 3012.81, 2958.80, 2931.80, 2877.79, 2829.57, 1654.92, 1614.42, 1577.77, 1539.20, 1487.12, 1456.26, 1421.54, 1396.46, 1354.03, 1342.46, 1307.74, 1292.31, 1261.45, 1213.23, 1182.36, 1176.58, 1130.29, 1105.21, 1087.85, 1072.42, 1049.28, 1022.27, 1006.84, 950.91, 923.90, 856.39, 810.10, 744.52, 725.23, 671.23, 651.94. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.59 (s, 3H), 3.74 and 4.16 (2s, 2H), 6.83 (dd, *J =* 2.4 Hz, 8.8 Hz, 1H), 6.93–7.10 (m, 2H), 7.27–7.38 (m, 3H), 7.64 (t, *J =* 8.8 Hz, 9.2 Hz, 1H), 7.72–7.80 (m, 2H), 8.23 and 8.41 (2s, 1H), 10.85 and 10.91 (2s, 1H), 11.02 and 11.18 (2s, 1H), 11.40 (s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 31.90 (CH2), 55.01 (CH3), 103.47 (CH), 108.71 (C), 111.50 (C), 111.59 (CH), 112.39 (CH), 112.64 (CH), 118.52 (CH), 118.83 (CH), 121.14 (CH), 123.88 (CH), 124.83 (C), 127.73 (C), 130.56 (C), 132.19 (CH), 136.19 (C), 143.76 (CH), 154.55 (C), 172.04 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C20H18N4O2: 347.1503, found: 347.1505.

4.1.4. General Method for the Preparation of 4-aryl-1-[2-(1*H*-indol-3-yl)acetyl]thiosemicar bazide Derivatives (**4a-g**)

A mixture of compound **2** and aryl isothiocyanate in ethanol was stirred at room temperature for 8 h. The precipitate was filtered off. The product was crystallized from ethanol.

4-(4-Bromophenyl)-1-[2-(1*H*-indol-3-yl)acetyl]thiosemicarbazide (**4a**)

Yield: 87%. M.p.: 187–189 ◦C. IR νmax (cm−<sup>1</sup> ): 3390.86, 3311.78, 3286.70, 3207.62, 3143.97, 3057.17, 2997.38, 2927.94, 1680.00, 1647.21, 1620.21, 1589.34, 1544.98, 1506.41, 1485.19, 1452.40, 1419.61, 1352.10, 1309.67, 1282.66, 1247.94, 1207.44, 1138.00, 1087.85, 1074.35, 1049.28, 1004.91, 987.55, 871.82, 823.60, 792.74, 736.81, 715.59, 669.30. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.63 (s, 2H), 6.97 (t, *J =* 6.8 Hz, 1H), 7.07 (t, *J =* 6.8 Hz, 1H), 7.25 (d, *J =* 2.4 Hz, 1H), 7.34 (d, *J =* 8.0 Hz, 1H), 7.42 (d, *J =* 8.0 Hz, 2H), 7.51 (d, *J =* 8.4 Hz, 2H), 7.59 (d, *J =* 7.6 Hz, 1H), 9.59 (brs, 1H), 9.73 (s, 1H), 10.10 (brs, 1H), 10.89 (s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 31.18 (CH2), 108.39 (C), 111.75 (CH), 118.80 (CH), 119.26 (CH), 121.46 (CH), 122.70 (C), 124.44 (CH), 127.70 (C), 129.40 (2CH), 131.37 (2CH), 136.51 (C), 139.05 (C), 170.35 (C), 181.10 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C17H15BrN4OS: 403.0223, found: 403.0204.

4-(4-Trifluoromethylphenyl)-1-[2-(1*H*-indol-3-yl)acetyl]thiosemicarbazide (**4b**)

Yield: 80%. M.p.: 184–186 ◦C. IR νmax (cm−<sup>1</sup> ): 3392.79, 3315.63, 3292.49, 3223.05, 3163.26, 3070.68, 2995.45, 2927.94, 1681.93, 1649.14, 1616.35, 1568.13, 1544.98, 1504.48, 1454.33, 1419.61, 1357.89, 1321.24, 1246.02, 1224.80, 1209.37, 1184.29, 1163.08, 1132.21, 1120.64, 1112.93, 1085.92, 1070.49, 1012.63, 985.62, 846.75, 788.89, 736.81, 711.73, 665.44. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.64 (s, 2H), 6.97 (t, *J =* 6.8 Hz, 1H), 7.07 (t, *J =* 6.8 Hz, 1H), 7.26 (d, *J =* 2.0 Hz, 1H), 7.34 (d, *J =* 8.0 Hz, 1H), 7.60 (d, *J =* 7.6 Hz, 1H), 7.67–7.75 (m, 4H), 9.75 (brs, 1H), 9.88 (s, 1H), 10.14 (brs, 1H), 10.89 (s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 31.16 (CH2), 108.36 (C), 111.76 (CH), 118.80 (CH), 119.25 (CH), 121.45 (CH), 123.45 (CH), 124.46 (2CH), 125.64 (C), 126.15 (C), 127.69 (2CH), 132.50 (C), 136.51 (C), 143.45 (C), 170.35 (C), 181.10 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C18H15F3N4OS: 393.0991, found: 393.0989.

4-(4-Cyanophenyl)-1-[2-(1*H*-indol-3-yl)acetyl]thiosemicarbazide (**4c**)

Yield: 89%. M.p.: 180–182 ◦C. IR νmax (cm−<sup>1</sup> ): 3425.58, 3313.71, 3284.77, 3201.83, 3145.90, 3059.10, 2995.45, 2956.87, 2914.44, 2223.92, 1680.00, 1651.07, 1620.21, 1602.85, 1541.12, 1510.26, 1475.54, 1454.33, 1409.96, 1334.74, 1290.38, 1244.09, 1226.73, 1203.58, 1174.65, 1136.07, 1093.64, 1060.85, 1012.63, 975.98, 837.11, 790.81, 769.60, 734.88, 692.44. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.64 (s, 2H), 6.97 (t, *J =* 7.2 Hz, 7.6 Hz, 1H), 7.07 (t, *J =* 7.2 Hz, 7.6 Hz, 1H), 7.26 (s, 1H), 7.35 (d, *J =* 7.6 Hz, 1H), 7.60 (d, *J =* 7.2 Hz, 1H), 7.78 (s, 4H), 9.76 (brs, 1H), 9.97 (s, 1H), 10.16 (brs, 1H), 10.90 (s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 31.17 (CH2), 108.04 (C), 109.58 (C), 111.49 (CH), 118.53 (C), 118.97 (CH), 119.16 (CH), 121.20 (CH), 124.20 (CH), 127.40 (C), 129.32 (2CH), 132.55 (2CH), 136.24 (C), 143.87 (C), 170.35 (C), 181.10 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C18H15N5OS: 350.1070, found: 350.1063.

4-[4-(Piperidin-1-ylsulfonyl)phenyl]-1-[2-(1*H*-indol-3-yl)acetyl]thiosemicarbazide (**4d**)

Yield: 85%. M.p.: 182–184 ◦C. IR νmax (cm−<sup>1</sup> ): 3390.86, 3288.63, 3197.98, 3089.96, 2939.52, 2850.79, 1645.28, 1595.13, 1550.77, 1496.76, 1467.83, 1404.18, 1336.67, 1315.45, 1276.88, 1244.09, 1226.73, 1215.15, 1149.57, 1093.64, 1053.13, 1028.06, 1012.63, 983.70, 929.69, 860.25, 839.03, 819.75, 777.31, 752.24, 738.74, 719.45, 698.23, 667.37. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 1.33–1.36 (m, 2H), 1.50–1.54 (m, 4H), 2.87 (t, *J =* 4.8 Hz, 5.2 Hz, 4H), 3.66 (s, 2H), 6.98 (t, *J =* 7.2 Hz, 1H), 7.08 (t, *J =* 7.2 Hz, 1H), 7.27 (d, *J =* 2.0 Hz, 1H), 7.36 (d, *J =* 7.6 Hz, 1H), 7.61 (d, *J =* 8.0 Hz, 1H), 7.67 (d, *J =* 8.4 Hz, 2H), 7.83 (d, *J =* 8.4 Hz, 2H), 9.75 (brs, 1H), 9.93 (s, 1H), 10.16 (brs, 1H), 10.90 (s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 23.37 (CH2), 25.14 (2CH2), 31.18 (CH2), 47.08 (2CH2), 108.35 (C), 111.78 (CH), 118.83 (CH), 119.26 (CH), 121.48 (CH), 124.49 (CH), 125.05 (2CH), 127.68 (C), 128.11 (2CH), 135.47 (C), 136.53 (C), 143.89 (C), 170.34 (C), 181.11 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C22H25N5O3S2: 472.1472, found: 472.1452.

4-[4-(1*H*-Pyrazol-1-yl)phenyl]-1-[2-(1*H*-indol-3-yl)acetyl]thiosemicarbazide (**4e**)

Yield: 85%. M.p.: 196–198 ◦C. IR νmax (cm−<sup>1</sup> ): 3305.99, 3223.05, 3167.12, 3134.33, 3095.75, 3061.03, 2999.31, 2933.73, 1680.00, 1647.21, 1622.13, 1573.91, 1546.91, 1523.76, 1454.33, 1421.54, 1396.46, 1359.82, 1332.81, 1317.38, 1305.81, 1249.87, 1222.87, 1199.72, 1159.22, 1136.07, 1124.50, 1089.78, 1043.49, 1033.85, 1008.77, 985.62, 935.48, 840.96, 792.74, 758.02, 744.52, 717.52, 667.37. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.65 (s, 2H), 6.53 (t, *J =* 2.4 Hz, 1H), 6.99 (t, *J =* 7.2 Hz, 1H), 7.08 (t, *J =* 7.2 Hz, 1H), 7.27 (d, *J =* 2.0 Hz, 1H), 7.36 (d, *J =* 8.0 Hz, 1H), 7.55 (d, *J =* 8.0 Hz, 1H), 7.62 (d, *J =* 8.0 Hz, 2H), 7.74 (d, *J =* 1.6 Hz, 1H), 7.80 (d, *J =* 8.8 Hz, 2H), 8.46 (d, *J =* 2.4 Hz, 1H), 9.65 (brs, 1H), 9.71 (s, 1H), 10.13 (s, 1H), 10.90 (s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 31.20 (CH2), 108.23 (C), 108.45 (CH), 111.77 (CH), 118.56 (CH), 118.82 (CH), 119.28 (2CH), 121.48 (CH), 124.47 (CH), 126.80 (CH), 127.73 (C), 128.09 (2CH), 136.53 (C), 137.64 (2C), 141.30 (CH), 170.32 (C), 181.10 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C20H18N6OS: 391.1336, found: 391.1334.

4-(1,3-Benzodioxol-5-yl)-1-[2-(1*H*-indol-3-yl)acetyl]thiosemicarbazide (**4f**)

Yield: 81%. M.p.: 168–170 ◦C. IR νmax (cm−<sup>1</sup> ): 3300.20, 3209.55, 3149.76, 3057.17, 2929.87, 2897.08, 1678.07, 1643.35, 1591.27, 1539.20, 1500.62, 1481.33, 1454.33, 1419.61, 1334.74, 1282.66, 1240.23, 1197.79, 1122.57, 1089.78, 1037.70, 981.77, 923.90, 850.61, 815.89, 808.17, 790.81, 731.02, 698.23. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.62 (s, 2H), 6.02 (s, 2H), 6.72 (d, *J =* 8.4 Hz, 1H), 6.87 (d, *J =* 8.0 Hz, 1H), 6.96–7.09 (m, 3H), 7.26 (s, 1H), 7.34 (d, *J =* 8.4 Hz, 1H), 7.60 (d, *J =* 8.0 Hz, 1H), 9.47 (brs, 1H), 9.55 (s, 1H), 10.06 (s, 1H), 10.89 (s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 30.65 (CH2), 101.21 (CH2), 107.38 (C), 107.97 (CH), 111.25 (CH), 118.29 (2CH), 118.78 (CH), 120.95 (CH), 123.95 (CH), 127.24 (CH), 133.08 (C), 136.02 (2C), 144.59 (C), 146.56 (C), 170.35 (C), 181.10 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C18H16N4O3S: 369.1016, found: 369.0998.

4-[4-(Benzyloxy)phenyl]-1-[2-(1*H*-indol-3-yl)acetyl]thiosemicarbazide (**4g**)

Yield: 88%. M.p.: 198–200 ◦C. IR νmax (cm−<sup>1</sup> ): 3394.72, 3290.56, 3213.41, 3155.54, 3059.10, 3032.10, 2939.52, 2873.94, 1681.93, 1649.14, 1618.28, 1564.27, 1546.91, 1504.48, 1456.26, 1417.68, 1381.03, 1359.82, 1294.24, 1244.09, 1219.01, 1170.79, 1138.00, 1089.78, 1051.20, 999.13, 912.33, 879.54, 829.39, 790.81, 734.88, 702.09, 646.15. <sup>1</sup>H NMR (400 MHz, DMSO-*d6*): 3.64 (s, 2H), 5.10 (s, 2H), 6.97–7.01 (m, 3H), 7.06–7.10 (m, 1H), 7.27–7.29 (m, 3H), 7.33–7.47 (m, 6H), 7.61 (d, *J =* 8.0 Hz, 1H), 9.47 (brs, 1H), 9.53 (s, 1H), 10.07 (s, 1H), 10.90 (s, 1H). <sup>13</sup>C NMR (100 MHz, DMSO-*d6*): 30.68 (CH2), 69.34 (CH2), 107.99 (C), 111.27 (CH), 114.22 (2CH), 118.31 (2CH), 118.79 (CH), 120.96 (CH), 123.95 (CH), 127.25 (C), 127.67 (2CH), 127.80 (2CH), 128.41 (2CH), 132.15 (C), 136.03 (C), 137.08 (C), 155.77 (C), 170.39 (C), 181.17 (C). HRMS (ESI) (*m*/*z*): [M + H]<sup>+</sup> calcd. for C24H22N4O2S: 431.1536, found: 431.1554.

#### *4.2. Biochemistry*

4.2.1. In Vitro COX Inhibition Assay

COX (ovine) Colorimetric Inhibitor Screening Assay (Cayman, Ann Arbor, MI, USA) was conducted to detect the peroxidase component of COX-1 and COX-2 according to the manufacturer's instructions [53]. The assay was performed in triplicate. Half maximal inhibitory concentration (IC50) data (µM) were expressed as mean ± SD.

#### 4.2.2. Cell Culture and Drug Treatment

A549 human lung adenocarcinoma and L929 mouse fibroblast cell lines were obtained from American Type Culture Collection (ATCC) (Manassas, VA, USA). Both cell lines were cultured, and drug treatments were carried out as previously reported [31,86].

#### 4.2.3. MTT Assay

MTT assay was conducted as previously explained in the literature [87] with small modifications [86]. Cisplatin was used as a positive control. The assay was performed in triplicate. IC<sup>50</sup> data (µM) were expressed as mean ± SD.

#### 4.2.4. Flow Cytometry-Based Apoptosis Detection

FITC Annexin V Apoptosis Detection kit (BD Pharmingen, San Jose, CA, USA) was applied based on the manufacturer's instructions after the incubation of A549 cells with compound **4a** (at its IC50/4 and IC50/2 concentrations), compound **3b**, and cisplatin (at their IC50/2 and IC<sup>50</sup> concentrations) for 24 h [87].

#### 4.2.5. Determination of Akt Inhibition

After A549 cells were incubated with compounds **3b** (22.42 µM, 44.84 µM, 89.67 µM), **4a** (22.42 µM, 44.84 µM, 89.67 µM), Akt inhibitor GSK690693 (3.61 µM, 7.23 µM, 14.45 µM), and cisplatin (5.67 µM, 11.34 µM, 22.67 µM) for 24 h, Akt Colorimetric In-Cell ELISA Kit (Thermo Fisher Scientific, Waltham, MA, USA) was used according to the manufacturer's instructions [87]. The assay was performed in triplicate. IC<sup>50</sup> data (µM) were expressed as mean ± SD.

#### 4.2.6. Experimental Animals

Male albino Sprague Dawley rats (~250–300 g) were procured from the Medical and Surgical Experimental Animals Application and Research Center of Eskisehir Osmangazi University (ESOGU). In the animal house, the rats were housed in stainless steel cages under standard atmospheric conditions at 22 ± 1 ◦C and exposed to 12 h/12 h light/dark cycle [53]. Food and water were given ad libitum. All experiments and protocols reported in this work were approved by ESOGU Animal Experiments Local Ethics Committee (10 December 2018/700).

#### 4.2.7. Chemicals and Drug Administrations

Compounds **3b**, **4a**, and indomethacin (Sigma-Aldrich, St. Louis, MO, USA) were dissolved in 5% dimethyl sulfoxide (DMSO) and then diluted. The final DMSO concentration in the solution was 0.5% (*v*/*v*). The agents were administered by gastric intubation. LPS (Sigma-Aldrich, St. Louis, MO, USA) (1 mg/kg) dissolved in 0.9% sodium chloride solution was intraperitoneally injected only once on the 7th day for the experimentally induced sepsis model [53].

#### 4.2.8. In vivo Experimental Design

Rats were randomly divided into five groups (*n* = 8) as control group, LPS group, test groups (**3b** and **4a**), and reference group. 0.5% DMSO was used as control solution for LPS group. Indomethacin (5 mg/kg) was used as a reference agent. Control group (Group I) was fed with basal rat chow throughout the experimental period. LPS group (Group II) was fed with basal rat chow for six days (only 0.5% DMSO was administered by gastric intubation) and LPS was injected intraperitoneally in 0.9% sodium chloride solution only once on the 7th day. Groups III, IV, and V were fed with basal rat chow and compound **3b** (10 mg/kg/day), compound **4a** (10 mg/kg/day), and indomethacin were administered, respectively, by gastric intubation for six days. Then, LPS was injected intraperitoneally in 0.9% sodium chloride solution only once on the 7th day for three groups as well. After 24 h of LPS injection, all rats were sacrificed by ketamine (80 mg/kg) ve xylazine (10 mg/kg) anesthesia via intraperitoneal route. Blood samples were collected via cardiac puncture in tubes containing gel for obtaining serum [53].

Serum ALT and AST levels were determined using enzyme-based Roche Diagnostics kit in Roche Modular Systems analyzer by photometric assay [53] based on the manufacturer's instructions. The other serum samples were stored at −80 ◦C (Thermo Electron, Waltham, MA, USA) for subsequent analyses of MPO and NO levels.

#### 4.2.9. Determination of MPO Levels

Suzuki's assay [88] was performed with slight modifications [53]. The rate of MPOcatalyzed oxidation of 3,30 ,5,50 -tetramethylbenzidine (TMB) was followed by recording the absorbance increase at 655 nm for 5 min. Taking into account the linear phase of the reaction, the absorbance change was measured per minute. The enzyme activity was expressed as the amount of the enzyme producing one absorbance change per minute under assay conditions [53].

#### 4.2.10. Determination of NO Levels

Nitrate and nitrite, which represent the best index of the entire NO production, are the stable end products of NO *in vivo*. Nitrate in serum was assayed by a slight modification of the Cd-reduction method as reported by Cortas and Wakid [89].

#### 4.2.11. Statistical Analyses

The data used in statistical analyses were obtained from eight animals for each group and statistically evaluated by means of Statistical Package for the Social Sciences (SPSS) for Windows 17.0. Comparisons were performed by one-way ANOVA (Tukey for post-hoc analyses) test. Differences between groups were considered statistically significant at a level of *p* < 0.05.

#### **5. Conclusions**

In this paper, two classes of indole-based small molecules (**3a-j**, **4a-g**) were designed and synthesized for the targeted therapy of NSCLC. Based on the data gathered from the COX colorimetric inhibitor screening assay, compounds **3b** and **4a** were found to be the selective COX-1 inhibitors in this series with IC<sup>50</sup> values of 8.90 and 10.00 µM, respectively. In vitro and in vivo assays were conducted to assess their potential for the targeted therapy of NSCLC. The experimental data demonstrate that compound **3b** exerts selective anticancer activity against A549 cells through apoptosis induction and Akt inhibition. Compound **3b** also caused a substantial drop in the serum MPO and NO levels, pointing out its potential as an anti-inflammatory agent. Moreover, compound **3b** decreased the serum aminotransferase (particularly AST) levels. Taken together, compound **3b** stands out as a lead anti-NSCLC agent endowed with in vivo anti-inflammatory action acting as a dual COX-1 and Akt inhibitor. In the view of this work, a new generation of indole-based small molecules with enhanced antitumor potency could be designed through the molecular modification of compound **3b** for the targeted therapy of NSCLC.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijms24032648/s1.

**Author Contributions:** Conceptualization, M.D.A., G.A.Ç. and A.Ö.; methodology, M.D.A., G.A.Ç., N.Y.S., ˙I.E., B.C., H.E.T., Ö.A. and A.Ö.; software, M.D.A., G.A.Ç., ˙I.E. and B.C.; validation, M.D.A., G.A.Ç., ˙I.E. and B.C.; formal analysis, M.D.A., G.A.Ç., ˙I.E. and B.C.; investigation, M.D.A.; resources, M.D.A., G.A.Ç., Ö.A. and A.Ö.; writing—original draft preparation, M.D.A.; writing—review and editing, M.D.A., G.A.Ç., N.Y.S., ˙I.E., B.C., B.S., H.E.T., Ö.A. and A.Ö.; visualization, M.D.A.; project administration, M.D.A. and A.Ö.; funding acquisition, M.D.A. and A.Ö. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by Anadolu University Scientific Research Projects Commission under the grant no: 1902S013. The APC was funded by Anadolu University Scientific Research Projects Commission under the grant no: 2107S205.

**Institutional Review Board Statement:** The animal study protocol was approved by the Animal Experiments Local Ethics Committee of Eskisehir Osmangazi University (protocol no: 700 and date of approval: 10.12.2018).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data are contained within the article or Supplementary Material.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

### *Review* **Modulating Glycolysis to Improve Cancer Therapy**

**Chaithanya Chelakkot 1,†, Vipin Shankar Chelakkot 2,† , Youngkee Shin 3,4 and Kyoung Song 5,\***


**Abstract:** Cancer cells undergo metabolic reprogramming and switch to a 'glycolysis-dominant' metabolic profile to promote their survival and meet their requirements for energy and macromolecules. This phenomenon, also known as the 'Warburg effect,' provides a survival advantage to the cancer cells and make the tumor environment more pro-cancerous. Additionally, the increased glycolytic dependence also promotes chemo/radio resistance. A similar switch to a glycolytic metabolic profile is also shown by the immune cells in the tumor microenvironment, inducing a competition between the cancer cells and the tumor-infiltrating cells over nutrients. Several recent studies have shown that targeting the enhanced glycolysis in cancer cells is a promising strategy to make them more susceptible to treatment with other conventional treatment modalities, including chemotherapy, radiotherapy, hormonal therapy, immunotherapy, and photodynamic therapy. Although several targeting strategies have been developed and several of them are in different stages of pre-clinical and clinical evaluation, there is still a lack of effective strategies to specifically target cancer cell glycolysis to improve treatment efficacy. Herein, we have reviewed our current understanding of the role of metabolic reprogramming in cancer cells and how targeting this phenomenon could be a potential strategy to improve the efficacy of conventional cancer therapy.

**Keywords:** glycolysis; cancer metabolism; combination therapy

#### **1. Introduction**

Cancer cells reprogram their metabolism to promote growth, metastasis, and survival. They exhibit an increased glycolytic dependency and show an elevated glucose uptake and fermentation of glucose to lactate to meet the heightened anabolic needs for cancer cell proliferation [1]. Increased glycolysis is not only important for meeting the energy needs of the cells but is also crucial for the generation of metabolic intermediates necessary for macromolecule synthesis in cancer cells [2,3]. This phenomenon, often referred to as the 'Warburg effect,' is observed even in the presence of completely functional mitochondria [4]. The Warburg effect has been studied for over 90 years, and several studies have explored the mechanisms governing the increased glycolytic dependency of cancer cells. Several oncogenic proteins and tumor suppressors, including the hypoxia-inducible factor (HIF-1), Myc, p53, and PI3K/Akt/mTOR pathway, have been implicated in regulating this cancer cell-specific metabolic reprogramming.

The altered metabolism in the cancer cells provides an avenue for developing cancer cell-specific therapeutic targets and anti-cancer agents. Indeed, therapeutic strategies that target glycolysis and cancer cell-specific biosynthetic pathways are a major focus area in cancer research. Although the increased glycolytic dependency of neoplastic cells suggests the potential therapeutic efficacy of glycolytic inhibitors in cancer treatment,

**Citation:** Chelakkot, C.; Chelakkot, V.S.; Shin, Y.; Song, K. Modulating Glycolysis to Improve Cancer Therapy. *Int. J. Mol. Sci.* **2023**, *24*, 2606. https://doi.org/10.3390/ ijms24032606

Academic Editors: Laura Paleari and Ana Cristina Gonçalves

Received: 15 November 2022 Revised: 18 January 2023 Accepted: 20 January 2023 Published: 30 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

glycolytic inhibition alone is not effective in a clinical setting [5]. Targeting metabolism, especially in combination with chemotherapy, is expected to improve therapy responses and may help overcome drug resistance [6]. Elevated glycolysis in cancer cells and the resulting lactic acidosis modulate the tumor stroma to a pro-tumorigenic microenvironment. Targeting glycolytic changes in the tumor microenvironment has been shown to be a safe and effective strategy to enhance therapeutic efficacy [7,8]. In this review, we explore and discuss glycolytic modulation in cancer cells and how it could aid as a therapeutic strategy in combination therapies with chemotherapy and radiotherapy, immunotherapy, hormonal therapy, and photodynamic therapy (PDT).

#### **2. Modulating Glycolysis to Improve Chemotherapy and Radiotherapy**

Metabolic modulation has been shown to sensitize cancer cells toward chemotherapy and radiotherapy. Increased glycolysis, facilitated by an increased glucose uptake, is the major energy source in cancer cells apart from being the major source of macromolecules for cell proliferation and survival [9]. Recently it was demonstrated that glycolysis-addicted cancer cells show metabolic rewiring via mTORC1 activation [10,11]. Sustained mTORC1 activation bypasses glycolysis by directing the glucose flux toward the pentose phosphate pathway. Metabolic rewiring, including dysregulated glycolysis, elevated ATP production, and cell-death escape mechanisms, are the major culprits for therapeutic resistance in cancer cells. The intracellular ATP level in cancer cells is also associated with metastasis and stemness. Thus, targeting glycolysis or intracellular/extracellular ATP levels [12–14] is a promising strategy to sensitize cancer cells toward chemotherapy. Several studies have reported that glycolytic inhibitors improve the efficacy of cancer treatment and that glycolytic inhibition is a promising strategy when used as a combination therapy with other treatment modalities. In line with this, inhibiting glycolytic enzymes, hexokinase (HK) [15], pyruvate kinase (PK) [16], and lactate dehydrogenase (LDH) [17], have shown sensitizing effects with several chemotherapeutic agents [5].

#### *2.1. Targeting Glucose Transporters and Glucose Uptake to Improve Chemotherapy*

The inhibition of glucose transporters (GLUTs), a critical rate-limiting step in glucose metabolism, modulates the therapeutic efficacy of several drugs. GLUT expression is elevated in several types of cancers and is associated with a poor prognosis suggesting their key role in cancer cell metabolism [18–22]. Glycolysis inhibition using inhibitors of GLUT1, when combined with routine cancer therapy, has proven to be relevant in potentiating their effects in a synergistic manner in pre-clinical studies for several cancers [19,23–26]. GLUT1 inhibition curbs the self-renewing capacity and tumor-initiating potential of cancer stem cells and has a substantial significance from a therapeutic perspective [27].

A widely studied small molecule inhibitor of GLUT1, WZB117 synergistically inhibits breast cancer cells by inducing DNA damage when treated in combination with an allosteric AKT inhibitor [24,28]. WZB117 also sensitizes breast cancer cells toward treatment with adriamycin [29] and radioresistant breast cancer cells to radiotherapy [27,30]. Previous studies have reported the synergistic effects of GLUT1 inhibitor #43 in melanoma cells and have shown that GLUT1 inhibition induces apoptosis, intracellular reactive oxygen species (ROS) generation, and the loss of mitochondrial membrane potential. Combination therapy with GLUT1 inhibitor #43 enhances the DNA-damaging effects of cisplatin by regulating the AKT/mTOR pathways [24]. Another GLUT1 inhibitor, BAY-876, enhances the cisplatin-mediated inhibition of esophageal squamous cell carcinoma [23]. siRNAmediated GLUT1 inhibition also showed similar results and improved the efficacy of low-dose cisplatin treatment [23]. In vivo studies in uterine cancer, patient-derived models have shown that glycolytic activation contributes to the stemness of uterine endometrial cancer, and BAY-876-mediated GLUT inhibition synergistically suppressed endometrial cancer cell proliferation when used in combination with paclitaxel [31].

Combination therapy with GLUT modulators can also improve the bioavailability of chemotherapeutic drugs. The co-treatment of paclitaxel with silybin (a GLUT modu-

lator) significantly improved the oral bioavailability of the drug in several in vitro and in vivo studies and overcame the major drawback of the limited oral bioavailability of paclitaxel [32,33]. A nanomedicine-based combination therapy using GLUT1 inhibitor and chemotherapeutic agent, curcumin, deprived cancer cells of glucose and sensitized cancer cells to chemotherapy, induced apoptosis, improved anti-tumor effects, and alleviated side-effects in vitro and in vivo [34]. Thus, combination therapy with GLUT1 inhibitors might be a rational therapeutic strategy and could also allow for low-dose treatment with chemotherapy drugs providing a paradigm for high-efficacy, low-toxicity therapeutic options [35].

Another straightforward and interesting strategy to improve the effectiveness of chemotherapy is to deprive cancer cells of glucose [36]. Icard et al. proposed that the modulation of glucose intake in combination with chemotherapy could improve the efficacy of the drug via the deprivation of ATP to cancer cells. A recent study showed that intermittent fasting throughout chemotherapy was well tolerated in patients and reduced chemotherapy-induced toxicity as measured by hematologic, metabolic, and inflammatory parameters [37]. Contrastingly, an opposite effect was reported in a pre-clinical model of pancreatic ductal adenocarcinoma (PDAC), wherein a relative glucose abundance sensitized PDAC cells to chemotherapy. Hyperglycemic patients with stage IV PDAC showed an enhanced response to chemotherapy, possibly via impaired glutathione biosynthesis [38]. A case report on non-small cell lung cancer (NSCLC) with bone and brain metastasis reported the efficacy of glucose uptake inhibition in combination with chemotherapy as a palliative treatment strategy. Fasting-induced hypoglycemia or insulin-induced hypoglycemia combined with low-dose chemotherapy could benefit cancer patients, particularly those who do not tolerate the conventional dosage of drugs [39].

#### *2.2. Targeting Glycolysis Enzymes to Improve Chemotherapy and Radiotherapy*

The enhanced glycolysis in the cancer cells correlates with an upregulation and activation of critical glycolytic enzymes. Targeting the key glycolytic enzymes is a promising strategy to rewire the altered tumor metabolism, to sensitize (or re-sensitize, when resistance develops) cancers to chemotherapy.

HK catalyzes the first step in glucose metabolism and converts glucose to glucose-6 phosphate. Four HK isoforms, HK1-4, with different cellular distributions and glucose affinity have been identified. HK1 and HK2 are located on the outer mitochondrial membrane and are associated with AKT-mediated cell survival [40,41]. Further, HK2 is associated with the recurrence and poor prognosis of breast cancer (BC) [42]. HK2 expression is also elevated in lung cancer, and shows significant association with the tumor stage. The deletion of the Hk2 gene in lung cancer cells ameliorated glucose-derived ribonucleotides and glutamine-derived carbon utilization in anaplerosis [43]. Targeting HK2 inhibits cell proliferation and shifts the metabolic profile of cancer cells from glycolytic to oxidative phosphorylation (OXPHOS) [43,44].

2-Deoxy glucose (2-DG), an HK inhibitor, has been shown to sensitize cancer cells to chemotherapy and radiotherapy and is being investigated in clinical trials. 2-DG is a glucose analog that triggers the intracellular accumulation of 2-deoxy-d-glucose-6-phosphate (2-DG6P), inhibiting the function of HK and glucose-6-phosphate isomerase [45]. Glycolysis inhibition with 2-DG can improve the therapeutic efficacy of trastuzumab in treating HER2+ BC [46]. Similarly, the therapeutic efficacy of paclitaxel was enhanced when treated in combination with 2-DG in in vivo studies in NSCLC and osteosarcoma models [47]. A recent study showed that 2-DG could sensitize glioblastoma cells to chloroethyl nitrosourea by regulating glycolysis, intracellular ROS generation, and endoplasmic reticulum stress induction [48]. Another study reported that combining 2-DG with autophagy inhibiting drug hydroxychloroquine enhances apoptosis in BC cells. The inhibition of autophagy combined with 2-DG induced the accumulation of misfolded proteins in the endoplasmic reticulum and resulted in sustained endoplasmic reticulum stress, induced through the pERK-eIF2α-ATF4-CHOP axis to enhance apoptosis in BC cells [49]. Although a few clinical

trials (NCT05314933, NCT00096707) are investigating the toxicity, tolerability, pharmacokinetics, and recommended dose of 2-DG in advanced tumors [50], additional studies and trials are required to characterize the mechanism of action and treatment benefits of 2-DG in cancer.

Curcumin, another compound with a proven anti-tumor effect, is also known to inhibit HK expression by inhibiting transcriptional repressor SLUG. Combination therapy of curcumin with docetaxel demonstrated a high response rate, low-toxicity, and improved patient tolerance in prostate cancer [51].

3-Bromopyruvate acid (3BrPA) is another classic glycolytic inhibitor, which inhibits several enzymes in the glycolytic pathway, including HK and LDH, and is a potent inhibitor of cancer cell growth [52–55]. Combining 3BrPA with rapamycin enhanced the anti-tumor efficacy through the dual inhibition of mTOR signaling and glycolysis in LC and neuroblastoma [56,57]. In BC cells, 3BrPA enhanced the expression of thioredoxin interacting protein (TXNIP) and inhibited HK2 expression via c-Myc downregulation [58], and enhanced tamoxifen-induced cytotoxicity in vitro [59]. The combination regimen with tamoxifen also enhanced oxidative stress and reduced glutathione levels in cells, and affected tumor angiogenesis and metastasis in animal models [59]. Further, the intra-cranial delivery of 3BrPA with temozolomide showed synergistic effects and increased survival in animal models of glioma [60]. Moreover, enhanced therapeutic efficacy was demonstrated when 3BrPA was combined with sorafenib in murine models of liver cancer [61]. 3BrPA can also enhance the anti-tumor effect of low-dose radiation via the reprogramming of mitochondrial metabolism and hindering of ATP generation [62]. 3BrPA also inhibits monocarboxylate transporter 1 (MCT1) expression, which mediates the bidirectional transport of lactate in cancer cells, and sensitizes cancer cells to ionizing radiation [63].

Elevated glycolysis in cancer cells results in the conversion of pyruvate to lactate, even under aerobic conditions. Lactate is excreted at high levels from tumor cells and acts as a metabolic fuel and oncometabolite with signaling properties. Lactate utilization by tumor cells depends on the expression of monocarboxylic transporters (MCTs), which are upregulated in cancer cells [64,65]. MCTs facilitate the shuttle of lactate from cancer cells to neighboring cells, tumor stroma, and tumor-associated endothelial cells and induce metabolic rewiring. Lactate is involved in several tumorigenic functions of cancer cells, including tumor microenvironment modulation and tumor angiogenesis [66]. Targeting MCTs has been shown to suppress tumor growth in cancer cells [67]. MCT1 and MCT4 inhibitors impair leukemia cell proliferation and enhance their sensitivity toward chemotherapy [68]. Currently, a phase 1 clinical trial is evaluating the toxicity and pharmacokinetic profile of AZD3965, an MCT inhibitor in cancer therapy for B-cell lymphoma [69,70].

LDH is a key glycolytic enzyme that is elevated in aggressive cancers and is essential for tumor maintenance [71–73]. LDH-A is regulated by numerous oncogenic transcription factors, including c-Myc and HIF-1, and is closely associated with malignant phenotypes of cancer cells [74]. LDH-A overexpression upregulated AKT phosphorylation and PI3K, which upregulated cyclin D1 and c-Myc expression in LC cells [75–78]. LDH overexpression is also associated with epithelial–mesenchymal transition (EMT)-related genes, SNAIL and SLUG, and is thus involved in regulating the metastatic progression of cancer cells [79]. LDH-A levels could thus serve as a biomarker for cancer diagnosis and prognosis [80–82]. LDH inhibition induces oxidative stress, impacting cancer stem cells' renewable capacity. The overburden of mitochondrial complex II is speculated to account for the increased ROS production in LDH-A-inhibited cancer stem cells [83,84]. The widely studied LDH-A inhibitors include pyruvate analog, oxamate (OXM), and the NADH competitive inhibitor, gossypol. LDH-A inhibition with OXM triggers a specific tumor reduction in brain tumors by reducing ATP levels, increasing ROS production, and inducing apoptosis [85]. OXM also induces autophagy via the AKT-mTOR signaling pathway in certain cancer cell types [86]. Recent pre-clinical and clinical studies have supported the combined use of LDH inhibitors with concurrent treatments as a promising strategy in cancer therapy [85]. Combination therapy of OXM with other chemotherapeutic drugs, an including mTORC1 inhibitor

and phenformin (phenethylbiguanidine; an anti-diabetic agent), have shown synergistic effects suggesting their implications in combination therapies [87,88]. A triple combination therapy of doxorubicin with metformin and OXM induced autophagy and apoptosis in colorectal cancers by downregulating hypoxia-induced HIF-1 expression [85].

Enolase (ENO-1), which converts 2-phosphoglycerate (2PG) to phosphoenol pyruvate (PEP), is emerging as a promising target for cancer therapy, partially owing to its diverse functions apart from being a major enzyme in glycolysis [89–91]. The overexpression of ENO-1 is associated with disease progression, metastasis-free survival, and overall survival in colorectal cancer, BC, gastric cancer, gliomas, head and neck cancer, and leukemia. ENO-1 promotes tumor cell progression via a plethora of mechanisms, including inducing angiogenesis, evading immune suppression and growth suppressors, and resisting cell death [92–94]. Small molecule inhibitors of ENO-1 have been shown to inhibit cancer cell growth [95–97]. POMHEX, a selective enolase inhibitor, has been shown to selectively inhibit the tumor cell progression of glioma cells in vivo by triggering apoptosis, showing a favorable safety profile and tolerance in non-human primates [98]. A previous study identified a potent inhibitor of ENO1, macrosphelide A, which demonstrates anti-cancer effects by simultaneously inactivating ENO1, aldolase, and fumarase [95].

6-phosphofructokinase/fructose-2,6-bisphophatase (PFKFBs) catalyzes the first irreversible step in glycolysis, which is the conversion of fructose-6-phosphate to fructose-1,6 bis-phosphate. As with most other glycolytic enzymes, PFKFB activity and expression are enhanced in many cancers. The selective inhibition of PFKFB3 displays broad anti-tumor activity in syngenic pre-clinical models and early human studies by inducing necroptotic cell death, apoptosis, cell cycle arrest, and inhibiting invasion [99,100]. A phase-1 dose escalation study for PFK-158, a first-in-human, first-in-class, small molecule inhibitor of PFKFB3, showed commendable tolerance and tumor burden reduction in pancreatic cancer, renal cell carcinoma, and adenocystic carcinoma patients [101–103]. The synergistic effect of PFK-158 with other FDA-approved targeted-chemotherapy agents can potentially improve their chemotherapy efficacy and is being validated. In gynecologic cancers, it was shown that PFK-158 improves lipophagy and sensitizes platinum therapy-resistant cells to carboplatin/oxaliplatin therapy [104]. Combining PFKFB3 inhibition with standard chemotherapy can thus be a novel strategy to improve the outcome in gynecologic and endometrial cancer patients who are resistant to therapy or have advanced, recurrent diseases [89].

Besides its glycolytic function, PFKFB3 is a crucial player in regulating endothelial cells, tumor angiogenesis, and tumor vascularization [105,106]. A transient inhibition of PFKFB3 in endothelial cells using 3-(3-pyridinyl)-1-(4-pyridinyl)-2-propen-1-one (3PO) induced tumor vessel normalization, impaired metastasis, and improved chemotherapy [107,108]. The PFKFB3 inhibitor, AZ67, inhibited angiogenesis in vivo, independent of glycolysis regulation [109]. A combination therapy with PFKFB3 inhibitor and VEGF inhibitor, bevacizumab, improved tumor vasculature, alleviated tumor hypoxia, normalized lactate production, and improved the efficacy and delivery of doxorubicin in glioblastoma [110].

Pyruvate kinase M2 (PKM2), which catalyzes the conversion of PEP to pyruvate, is upregulated in numerous cancers and has emerged as a critical regulator of cancer cell metabolism [111,112]. Apart from being a key enzyme in glycolysis, nuclear PKM2 regulates the expression of GLUT1 and LDH-A through positive feedback to further support glycolytic metabolism [113]. PKM2 expression is upregulated under hypoxic conditions and induces tumor angiogenesis and metastasis [114]. The association of PKM2 expression with poor prognosis and overall survival indicates that PKM2 level could serve as a prognostic or diagnostic marker for cancers [115–117]. PKM2 inhibition induces apoptosis and tumor regression in xenograft models of different cancer types [89] and plays a role in maintaining redox homeostasis and glutathione turnover. PKM2 inhibition also increases the efficacy of docetaxel treatment in vitro and xenograft models of LC [118,119]. In NSCLC patients who received platinum therapy in a first-line setting, tumors with low PKM2 expression showed significantly longer progression-free survival and overall survival [120]. In tumor xenograft

models of NSCLC, combination therapy with PKM2 siRNA and chemotherapeutic agents increased apoptosis and inhibited tumor growth [121].

Targeting Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) is being explored as an alternative approach for inhibiting glycolysis [122]. GAPDH catalyzes the first step in which energy is derived from NADH in the 'pay-off-phase' of glycolysis. NADH, the first molecule generated in this phase, is critical for regulating intracellular ROS and redox balance. Targeting GAPDH triggers the accumulation of glucotrioses such as glyceraldehyde-3-phosphate and dihydroxy acetone phosphate in the cells, the partial degradation of which results in the formation of cytotoxic methylglyoxal [123]. Thus, inhibiting GAPDH not only depletes ATP but also triggers cytotoxicity through the upregulation of ROS and the accumulation of methylglyoxal [122]. 3-BrPA, discussed above, is a potent inhibitor of GAPDH and was shown to deplete intracellular ATP. Additionally, 3-BrPA showed high specificity and selectivity for GAPDH both in vitro and in vivo [122,124,125].

Although targeting glycolytic enzymes can improve the efficacy of chemotherapy and radiotherapy, the ubiquitous nature of glycolysis and glycolytic enzymes presents the challenge of the systemic toxicity of glycolysis inhibition. The selective targeting of cancer-specific enzymes or enzyme isoforms and the targeted delivery of therapeutic agents could circumvent this challenge.

#### *2.3. Modulating Glycolysis to Overcome Drug Resistance*

Aberrant glycolysis is a major contributor to drug resistance in cancer [126,127]. The mechanism underlying glucose metabolism reprogramming-induced drug resistance is not clearly understood. Increased glucose uptake induces gemcitabine resistance in pancreatic cancer, doxorubicin resistance in BC, and cisplatin resistance in genitourinary cancers [6,128]. It is thought that glucose metabolism reprogramming in cancer cells induces DNA repair and immune suppression in the tumor microenvironment, contributing to drug resistance. Anabolic alterations could account for the increased nucleotide demand required for the efficient repair of chemotherapy/radiation-induced DNA damage. The DNA repair pathways in reprogrammed cancer cells induces the activation of several protumorigenic signaling pathways, including Wnt, PI3K/AKT, NF-κB, and MAPK, triggering prolonged cancer cell survival and apoptosis resistance [89,129]. Aberrant glycolysis can also promote DNA repair by increasing nucleotide turn over by enhancing the hexosamine biosynthetic pathway (HBP) and pentose phosphate pathway (PPP) [130,131]. By limiting pyruvate flux into OXPHOS, upregulated glycolysis also enables cancer cells to reduce the ROS accumulation in cells, another mechanism by which metabolic reprogramming contributes to resistance to therapy. Increasing evidence also suggests that the activation of Wnt, PI3K/AKT, and Notch pathways activate autophagy which also contributes to cancer cell survival and resistance to therapy, whereas inhibiting autophagy sensitizes cancer cells to therapy [89,132,133]. Autophagy, thus, downregulates cell metabolism leading to cancer cell quiescence and survival, inducing radio-resistance [134].

Metabolic changes in the tumor cells happen hand-in-hand with similar reprogramming of the tumor microenvironment. This metabolic reprogramming induces immunosuppression and immune escape of cancer cells and contributes to the development of resistance to chemotherapy and radiotherapy [135,136]. The upregulation of glycolytic enzyme HK2 suppresses the mTOR-S6K signaling pathway and blocks chemotherapyinduced apoptosis by binding to voltage-dependent anion channels, and suppresses the formation of mitochondrial permeability transition pores, contributing to chemoresistance. Aberrant glycolytic pathways in cancer stem cells also play critical roles in contributing to resistance to therapy via enhancing cancer cell stemness by activating the PI3K/AKT pathway and upregulating the stem cell-like properties [89]. Enhanced exosomal secretion from cancer stem cells also activates neighboring cancer cells toward stemness and promotes chemo/radio-resistance [137,138]. The overexpression of ENO-1 in cancer cells can also contribute to cisplatin resistance in different cancer types [139] and is considered a biomarker to predict prognosis and drug resistance in cancers [85,140].

Glycolytic inhibitors are reported to sensitize cancer cells to chemotherapy and radiotherapy, thereby overcoming resistance to therapy. 3BrPA aids in dissociating HK2 from the mitochondrial complex and improves therapy response to daunorubicin [15]. A combination therapy of curcumin and docetaxel has demonstrated improved drug response and tolerance in a clinical study in prostate cancer patients [141]. 2-DG, a glycolytic inhibitor that modulates several glycolytic enzymes, restores sensitivity to adriamycin in ER+ BC cells. In HER2+ BC, trastuzumab inhibits tumor growth by downregulating heat shock factor 1 (HSF1) and LDH-A, thereby inhibiting glycolysis. A combination therapy of trastuzumab with LDH-A siRNA-mediated glycolysis inhibition synergistically inhibited tumor growth in trastuzumab-resistant breast cancer cells, suggesting their usefulness in overcoming drug resistance. Further, the combination therapy of trastuzumab with glycolytic inhibitor 2-DG and oxamate increased the sensitivity of ErbB2-positive cancer cells to therapy. An allosteric inhibitor of phosphoglycerate mutase (PGAM), a glycolytic enzyme that converts 3-phosphoglycerate to 2-phosphoglycerate, has been shown to overcome erlotinib resistance in NSCLC. PGAM inhibition alters the ERK and AKT signaling pathways and induces oxidative stress and ROS production to overcome erlotinib resistance [142].

ENO-1 overexpression has been associated with chemoresistance in prostate and pancreatic cancer cells [102,143–145]. Cisplatin-resistant gastric cancer cells also exhibit enhanced glycolysis by upregulating ENO-1. ENO1 inhibition in cisplatin-resistant cells increased sensitivity to the therapy by activating apoptotic pathways or inducing autophagy [96]. In ovarian cancer cells, inhibiting ENO-1 expression increased cell senescence and improved cisplatin resistance [146]. Hypoxia-induced resistance to gemcitabine is a critical issue in PDAC treatment. A recent study demonstrated that the shRNA-based downregulation of ENO-1 modulated redox homeostasis, increased intracellular ROS concentration, and sensitized resistant PDAC cells to gemcitabine treatment. In ovarian cancer models, PFKFB3 inhibitors, 3-PO and PFK-158, impaired metabolic reprogramminginduced stemness and chemoresistance, possibly by modulating apoptosis via the NF-κB pathway [147,148].

PKM2 can contribute to chemoresistance against cisplatin and gemcitabine treatment in different cancer types. PKM2 overexpression has been reported to be a biomarker for cancer resistance. PKM2 regulates the DNA repair mechanism in addition to glucose metabolism and induces resistance to genotoxic damage, driving treatment resistance [149,150]. Targeting PKM2 sensitizes cancer cells to treatment. In NSCLC, shRNA-based silencing of PKM2 enhanced radiation-induced autophagy in vitro and in vivo [149] and increased the sensitivity to docetaxel treatment [119]. PKM2 expression also correlated with a resistance to platinum-based therapy in colorectal cancer [151].

A few studies, however, have reported contradicting results, where PKM2 activation was shown to act as a chemosensitizer in some cancer types. In a study by Anastasiou et al., an increase in intracellular ROS concentration in response to therapy was shown to inhibit PKM2, which in turn diverted the glucose flux into PPP generating redox potential for the detoxification of ROS. These regulatory properties of PKM2 confer an additional advantage to cancer cells to tolerate therapy-induced oxidative stress. The endogenous expression of oxidation-resistant-PKM2 mutants increased oxidative stress and impaired tumor progression [152]. PKM2 activation could thus be an attractive strategy in cancer therapy. High levels of PKM2 activate the mTOR-HIF1α pathway and are associated with a positive chemotherapy response in cervical cancer patients treated with cisplatinneoadjuvant chemotherapy [153,154]. The high expression of PKM2 has been shown to enhance drug response to epirubicin and 5-fluorouracil in BC [155], whereas a decrease in PKM2 levels/activity contributes to cisplatin/oxaliplatin resistance in cervical cancer, colorectal cancer, and gastric cancer cells [153,156,157].

The development of resistance is frequently encountered in cancer treatment, and the link between cancer cell metabolism and the development of resistance is becoming more apparent. Targeting metabolic enzymes is an efficient strategy to re-sensitize the resistant cells to chemotherapy and radiotherapy. However, the clinical application of glycolysis inhibition to overcome drug resistance has remained limited. Future studies should identify the key metabolic shifts that contribute to the development of drug resistance and explore their potential as drug targets to improve the sensitivity of cancers to chemotherapeutic agents and radiotherapy and to overcome the development of resistance. Figure 1 summarizes how targeting glycolysis can be used to modulate cancer therapy (Figure 1). apparent. Targeting metabolic enzymes is an efficient strategy to re-sensitize the resistant cells to chemotherapy and radiotherapy. However, the clinical application of glycolysis inhibition to overcome drug resistance has remained limited. Future studies should identify the key metabolic shifts that contribute to the development of drug resistance and explore their potential as drug targets to improve the sensitivity of cancers to chemotherapeutic agents and radiotherapy and to overcome the development of resistance. Figure 1 summarizes how targeting glycolysis can be used to modulate cancer therapy (Figure 1).

The development of resistance is frequently encountered in cancer treatment, and the link between cancer cell metabolism and the development of resistance is becoming more

inhibit PKM2, which in turn diverted the glucose flux into PPP generating redox potential for the detoxification of ROS. These regulatory properties of PKM2 confer an additional advantage to cancer cells to tolerate therapy-induced oxidative stress. The endogenous expression of oxidation-resistant-PKM2 mutants increased oxidative stress and impaired tumor progression [152]. PKM2 activation could thus be an attractive strategy in cancer therapy. High levels of PKM2 activate the mTOR-HIF1α pathway and are associated with a positive chemotherapy response in cervical cancer patients treated with cisplatin-neoadjuvant chemotherapy [153,154]. The high expression of PKM2 has been shown to enhance drug response to epirubicin and 5-fluorouracil in BC [155], whereas a decrease in PKM2 levels/activity contributes to cisplatin/oxaliplatin resistance in cervical cancer, col-

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 8 of 35

orectal cancer, and gastric cancer cells [153,156,157].

**Figure 1.** Targeting glycolysis to improve cancer therapy. The cancer cells show enhanced dependency on glycolysis that may be targeted to improve the treatment efficacy of conventional cancer therapy modalities, including chemotherapy, radiotherapy, immunotherapy, hormonal therapy, and photodynamic therapy. Glycolysis metabolism can be potentially targeted by limiting glucose uptake (targeting glucose transporters), targeting glycolysis enzymes, targeting glutaminolysis, targeting lactate synthesis, targeting MCT, or targeting mitochondrial complexes. The increased gly-**Figure 1.** Targeting glycolysis to improve cancer therapy. The cancer cells show enhanced dependency on glycolysis that may be targeted to improve the treatment efficacy of conventional cancer therapy modalities, including chemotherapy, radiotherapy, immunotherapy, hormonal therapy, and photodynamic therapy. Glycolysis metabolism can be potentially targeted by limiting glucose uptake (targeting glucose transporters), targeting glycolysis enzymes, targeting glutaminolysis, targeting lactate synthesis, targeting MCT, or targeting mitochondrial complexes. The increased glycolysis in the cancer cells increase the release of lactic acid to the tumor microenvironment, acidifying it and making it pro-cancer and immunosuppressive. The tumor-infiltrating immune cells also show a similar shift toward glucose metabolism increasing the competition for glucose in the tumor microenvironment. Modulating glycolysis in the immune cells can potentially improve immune therapy. This figure was created using the Biorender app.

#### **3. Targeting Glycolysis to Enhance Immunotherapy**

The advances in our understanding of the remarkable potential of the immune system to fight cancer have garnered tremendous attention on immunotherapy for cancer.

Arguably, immunotherapy is now being considered one of the most promising therapeutic strategies for several types of cancers, and immune checkpoint blockade (ICB)- and adoptive-cell therapy (ACT)-based therapeutic strategies have been approved for several cancers [158–160]. However, several reports have shown that a high percentage of patients initially fail to respond to these interventions or acquire resistance in the long run [161,162], limiting the application of these promising strategies. Several studies have reported that the metabolic reprogramming of the cancer cell that leads to the development of an immunosuppressive tumor microenvironment is one of the main contributors to the reduced efficacy of immunotherapy [163]. Further, it has been suggested that metabolic interventions can significantly enhance the efficacy of immunotherapy [164,165]. Therefore, understanding the challenges the immune cells face in the harsh immunosuppressive tumor microenvironment and identifying strategic interventions to overcome these challenges would contribute to improving immunotherapy.

#### *3.1. Glucose Metabolism in the Immune Cells of the Tumor Microenvironment*

It is known that cancer cells undergo a metabolic reprogramming called the 'Warburg effect' or aerobic glycolysis in response to hypoxia and oncogenic signals, such as Myc and PI3K [165]. This preference for glycolysis is also shared by other rapidly proliferating cells in the TME, including effector T-cells and M1-like macrophages, to satisfy their increased energy requirements [166]. In contrast, other cells in the TME, including memory T-cells, regulatory T-cells (Tregs), and M2-like macrophages, rely on fatty acid oxidation (FAO) to satisfy their energy needs.

The naïve T-cells utilize TCA-coupled OXPHOS as their primary energy source [167]. On MHC activation, the T-cell receptor (TCR) and CD28 activate the PI3K-AKT-mTORC1 and Myc signaling pathways and induce metabolic reprogramming [167]. The effector T-cells upregulate aerobic glycolysis and enhance their anabolic metabolism for cancerkilling and clone expansion. In addition, the glycolytic intermediates support effector T-cell activation and cytokine generation. Phosphoenolpyruvate (PEP), a glycolytic metabolite, blocks sarco/endoplasmic reticulum Ca2+-ATPase (SERCA)-mediated endoplasmic reticulum calcium uptake and the nuclear factor of activated T-cells (NFAT) signaling, enabling TCR signaling [168]. In addition, phosphoenolpyruvate carboxykinase 1 (PCK1) catalyzes the conversion of oxaloacetate (OAA) into PEP, and the overexpression of PCK1 enhances the cancer-killing functions of adoptive transferred CD4+ and CD8+T-cells [168]. However, although CD8+ T-cells experiencing continuous stimulation or hypoxia differentiated into functional effectors in vitro, it rapidly drove T-cell dysfunction and exhaustion [169].

Similar to the CD8+ T-cells, the functions of the CD4+ T-cells are also affected by specific metabolic reprogramming. Increased glycolysis promotes IL-2, TNFα, and IFN secretion in the CD4+ T-cells, and the inhibition of glycolysis drives the functional and metabolic exhaustion of the CD4+ T-cells [170,171]. In line with this, the inflammatory CD4+ T-cells (Th1 and Th17) show enhanced glycolysis. The Th17 cells exclusively express the pyruvate dehydrogenase (PDH) inhibitor, pyruvate dehydrogenase kinase isozyme 1 (PDHK1), which, when downregulated, leads to the selective reduction of Th17 cells [172]. On the other hand, Tregs show upregulated OXPHOS and FAO [172], and their expression of FOXP3 inhibits Myc and attenuates PI3K-AKT-mTORC1 axis-mediated activation of glycolysis and increases oxidation and catabolic metabolism, rendering a survival advantage in the TME [64,173]. Furthermore, the Tregs utilize lactate metabolism, and culturing them in high-glucose conditions decreases their stability [167]. The mechanisms by which the low levels of oxygen, high levels of lactate, and the high competition for glucose potentially contribute to T-cell dysfunction in the tumor microenvironment is been summarized in Figure 2.

marized in Figure 2.

of glycolysis and increases oxidation and catabolic metabolism, rendering a survival advantage in the TME [64,173]. Furthermore, the Tregs utilize lactate metabolism, and culturing them in high-glucose conditions decreases their stability [167]. The mechanisms by which the low levels of oxygen, high levels of lactate, and the high competition for glucose potentially contribute to T-cell dysfunction in the tumor microenvironment is been sum-

**Figure 2.** Naïve T-cells relay on oxidative metabolism. After activation, the effector T-cells increase glycolysis to support their function. On antigen clearance, the effector T-cells enter a memory state. On antigen persistence, as with long-term tumor elimination, inhibitory receptors such as PD-1 and CTLA4 reprogram the T-cell metabolism leading to metabolic impairments. Exhausted T-cells show reduced glycolysis and glutaminolysis and dependence of fatty acid oxidation. The low levels of oxygen, high levels of lactate, and the high competition for glucose potentially contribute to T-cell dysfunction in the tumor microenvironment. The image was created using Biorender app. **Figure 2.** Naïve T-cells relay on oxidative metabolism. After activation, the effector T-cells increase glycolysis to support their function. On antigen clearance, the effector T-cells enter a memory state. On antigen persistence, as with long-term tumor elimination, inhibitory receptors such as PD-1 and CTLA4 reprogram the T-cell metabolism leading to metabolic impairments. Exhausted T-cells show reduced glycolysis and glutaminolysis and dependence of fatty acid oxidation. The low levels of oxygen, high levels of lactate, and the high competition for glucose potentially contribute to T-cell dysfunction in the tumor microenvironment. The image was created using Biorender app.

The M1-like macrophages preferentially utilize glycolysis to sustain the inflammatory phenotype [174], while the M2-like macrophages depend on TCA and FAO to maintain an immunosuppressive phenotype [175]. The highly acidic environment of melanoma was found to induce tumor-associated macrophages (TAM) toward a cancer-promoting phenotype [176]. The tumor-associated neutrophils (TAN), on the other hand, exhibit proor anti-tumor effects in different cancer microenvironments. In pancreatic ductal adenocarcinoma (PDAC), it was shown that the TANs undergo an LDH-A-mediated glycolytic switch and exhibit a tumor-promoting phenotype [177]. Triple-negative BC (TNBC) cells with an accelerated glycolysis support myeloid-derived suppressor cell (MDSC) development and facilitate CD8+T-cell inhibition and cancer progression [178]. In addition, an increase in lactate generation enhances the tumor-promoting capacity of MDSCs [179]. The M1-like macrophages preferentially utilize glycolysis to sustain the inflammatory phenotype [174], while the M2-like macrophages depend on TCA and FAO to maintain an immunosuppressive phenotype [175]. The highly acidic environment of melanoma was found to induce tumor-associated macrophages (TAM) toward a cancer-promoting phenotype [176]. The tumor-associated neutrophils (TAN), on the other hand, exhibit pro- or anti-tumor effects in different cancer microenvironments. In pancreatic ductal adenocarcinoma (PDAC), it was shown that the TANs undergo an LDH-A-mediated glycolytic switch and exhibit a tumor-promoting phenotype [177]. Triple-negative BC (TNBC) cells with an accelerated glycolysis support myeloid-derived suppressor cell (MDSC) development and facilitate CD8+T-cell inhibition and cancer progression [178]. In addition, an increase in lactate generation enhances the tumor-promoting capacity of MDSCs [179].

The TME is characterized by a decrease in nutrients, insufficient vasculature, increased lactate accumulation, and hypoxia, conditions that affect the cancer-killing capacity of the T-cell. The cancer cells and the immune cells compete for glucose utilization, and the activated glycolysis of the cancer cells endows them with an advantage over the immune cells, impairing the function and survival of the effector T-cells. It has been shown that the expression of glycolysis-related genes, such as *ALDOA*, *ALDOC*, *ENO2*,

*GAPDH*, *GPI*, and *PFKM*, negatively correlates with T-cell infiltration in melanoma and NSCLC [180]. The lower availability of glucose in the TME affects the glycolytic capacity of the T-cells [168], inducing T-cell exhaustion [181]. Furthermore, the lack of glucose also affects mitochondrial functions, promoting terminal CD8+ T-cell exhaustion [182]. In addition to the competition for glucose, the increased production of lactate due to the increased rate of glycolysis contributes to the immunosuppressive character of the TME. Lactate metabolism has been shown to contribute to oncogenesis [183], and LDH has been reported to be a marker for poor prognosis and immunotherapy efficacy [184,185]. The low pH in the TME triggers reduced CD25 and TCR expression in the effector CD8+ T-cells and inactivated STAT5 and ERK signaling, affecting anti-tumor immunity [186]. Lactate also affects TCR signaling [187], and exposure to large amounts of lactate makes the Tregs switch to OXPHOS for regenerating NAD+; however, the Tregs fail to maintain the NAD+/NADH balance through this pathway [64]. Moreover, lactate in the TME promotes the expression of pro-inflammatory cytokines, including IL-23 and IL-17, supporting tumorigenesis and impairing anti-tumor immunity [188]. The poor vasculature and the increased metabolism of the cancer cells contribute to the formation of a hypoxic environment, further affecting anti-tumor immunity. Hypoxia triggers epigenetic reprogramming of effector T-cells, reducing their functional capabilities [189,190]. Although hypoxia-inducible factor-1α (HIF-1α), expressed in response to hypoxia, can induce Tregs and bind to the promoter region of the FOXP3 to promote transcription [191] (22988108), it contributes to the development of an immunosuppressive TME by enhancing cancer-promoting immune cell functions [192,193].

#### *3.2. Signaling Mechanisms Regulating Glycolysis*

The signaling pathways that modulate glycolysis in the immune cells can be potentially targeted to improve their anti-tumor functions, and drugs such as metformin and phenformin have been extensively tested in the clinical setting [194–196]. The liver kinase B1 (LKB1)-AMPK pathway and the PI3K-AKT-mTOR axis are the two primary signaling mechanisms that modulate glycolysis in the cell. The LKB1-dependent kinases regulate metabolic pathways by targeting several effectors, including AMPK [197]. LKB1 loss upregulates GLUT1 and hexokinase 2 (HK2), increasing T-cell glycolytic transcription and flux [198].

On the other hand, the deregulation of the PI3K-AKT-mTOR axis promotes HIF-1 activation and GLUT1 expression [199,200]. PI3K activation increases AKT phosphorylation and glycolytic flux via LDH-A in T-cells [201]. AKT is the primary glycolysis regulator in both cancer and immune cells. AKT activation induces GLUT1 and LDH-A expression [202], activates HK1 by promoting HK2 and PFK2 phosphorylation [203], and inhibits PDH by activating pyruvate dehydrogenase kinase-1 (PDK1) [204], resulting in activated glycolysis. mTOR regulates the expression of HIF-1α, the major transcription factor regulating several glycolytic enzymes and GLUT1 [199]. mTOR kinases determine effector and memory CD8+ T-cell fates, and blocking mTOR promotes T-cell effector functions [205,206].

#### *3.3. Targeting Glycolysis to Improve Immunotherapy*

Adoptive-cell transfer (ACT) and immune checkpoint blockade (ICB) are the two primary strategies for immunotherapy. The interplay between anti-tumor immunity and cancer metabolism suggests that combining immunotherapy with glycolysis-targeted therapy is a promising strategy to improve treatment efficacy. ACT utilizes therapeutic-modified immune cells to directly boost anti-tumor immunity. Ex vivo-expanded T-cells and T-cells engineered to express antigen-specific TCRs or chimeric antigen receptors (CARs) are primarily used for ACT [207]. ACT has shown promising efficacies, particularly with CD19-specific CAR-T cells, in the treatment of B-cell acute lymphoblastic leukemia and B-cell lymphomas [208]. However, responses in other cancers have been poor. It has been suggested that optimizing T-cell metabolism to support robust initial and durable T-cell responses for target cells and the TME would improve and broaden their applicability.

CAR-T cells can be engineered to express specific signaling domains, and endowing them with specific properties can tailor them for effector activity and long-lasting memory. The two FDA-approved CARs carry a CD19-targeting extracellular domain coupled with an intracellular signaling domain from CD3ζ and the co-stimulatory molecules CD28 or 4- 1BB [209]. Linking TCR signaling with the co-stimulatory signals enables the CARs to elicit effector functions in the absence of additional inflammatory or co-stimulatory stimuli [210]. CAR-T cells carrying the CD28 domain have increased glycolysis and show enhanced effector responses but are short-lived. On the other hand, CAR-T cells expressing the 4-1BB domain show elevated OXPHOS and FAO and display a memory phenotype [211]. This is in line with the normal physiological functions of the co-stimulatory molecules. CD28 stimulation activates PI3K/AKT/mTOR, promoting glycolysis and effector differentiation, whereas 4-1BB activates AMP-driven FAO and OXPHOS [211]. Based on this understanding, it has been proposed that introducing signaling mutations could improve CAR-T cell effector functions and survival. For example, it was shown that mutations of the YMNM signaling motif of CD28 increase CAR-T cell survival and reduce T-cell exhaustion, enabling enhanced tumor control [212]. Furthermore, additional co-stimulatory molecules modulating T-cell metabolism may be incorporated into CAR constructs to improve T-cell function. For example, ICOS, a CD28 family member, promotes glycolysis and mTORC1 activity in T follicular helper cells [213]; GITR agonists enhance cellular metabolism to support CD8+ T-cell proliferation and effector cytokine production [214]; OX40 is associated with enrichment of glycolysis and lipid metabolism transcripts; and OX40 agonists enhanced lipid uptake in Tregs [215]. Similar to the CAR-T cells, T-cells procured from tumors and expanded ex vivo, or modified to express engineered TCRs, can also be optimized through metabolic manipulations, and T-cells with low glycolysis rates can be generated or selected for longevity while retaining effector functions [216].

The in vitro stimulation and the T-cell engineering phases of the ACT strategy provides the added advantage of the opportunity to modify T-cell metabolism and mitochondria without affecting other cells and tissues and thus prevent any potential boosting of the cancer cell metabolism from the intervention. It was also shown that blocking glutamine metabolism increases T-cell function [217]. Furthermore, supplementing the culture medium with glutamine antagonist 6-Diazo-5-oxo-l-norleucine (DON) enhanced CAR-T cell FAO and reduced glycolysis, making the CAR-T cells remain in a more undifferentiated state [218]. Inhibiting glycolysis using the HK-2 inhibitor, 2-DG, before ACT induced a memory-like phenotype in the T-cells, enabling a more efficient control of tumors and prolonged survival in an animal model [219]. Similarly, CD19-CAR-T cells treated with AKT inhibitors showed reduced glycolysis and a memory-like phenotype and had robust tumor elimination potential [220]. Conversely, modulating mitochondrial OXPHOS and FAO would enable CAR-T cell longevity and the continued expression of a memory-like phenotype. Furthermore, a transient glucose restriction followed by glucose re-exposure could enhance the tumor-clearing efficacy of CD8+ T-cells [221].

Unlike ACT, which directs anti-tumor immunity through pharmaceutical interventions, ICB aims to modify inhibitory signals to activate endogenous anti-tumor-specific T-cells. Furthermore, ICB is primarily targeted against solid tumors, which show a greater influence of the TME in modulating T-cell metabolism. In addition to the initial challenge of T-cells infiltrating the tumors, ICB must overcome several obstacles, including tumorinfiltrating lymphocyte (TIL) exhaustion, the upregulation of inhibitory receptors and epigenetic modifications, metabolic adaptations resulting in nutrient deficits, impaired translocation of GLUT1 to the cell surface, the downregulation of glycolytic enzymes, GAPDH and ENO-1, and dysregulated and fragmented mitochondria with increased ROS generation [222–225]. It was reported that the TIL inflammatory function could be enhanced by rescuing TIL metabolism by expressing PCK to promote gluconeogenesis and replacing the intracellular glycolytic intermediates or by improving mitochondrial metabolism by treating with pyruvate or acetate [225]. The immunosuppressive environment of the TME and chronic antigen stress direct the T-cells to an exhausted state, characterized by the

expression of immune checkpoints and decreased cytotoxicity. The checkpoint molecules, including CD28, CD40L, and cytotoxic T lymphocyte-associated protein-4 (CTLA-4), along with TCR signaling, drive TIL exhaustion and contribute to the persistence of the exhausted state. The concurrent metabolic adaptations further enhance the immunosuppressive effects of the checkpoints. Thus, targeting glycolytic regulators and metabolites that support the immune checkpoint-directed T-cell inhibition is a potential strategy to improve the efficacy of ICB.

The most well-known and commonly targeted immune checkpoints in cancer are the programmed cell death protein-1 (PD-1) and CTLA-4. CTLA-4 is expressed primarily on Tcells and plays an immunosuppressive role during the initial phase of T-cell activation and downregulates T-cell activation-triggered glycolysis. PD-1 is activated after TCR activation and impedes glucose uptake and glycolysis while promoting FAO. Thus, these checkpoint signals prevent T-cell activation and inflammation. It was shown that PD-1–deficient T-cells maintain higher metabolic activity in chronic infection [226,227]. Thus, blocking PD-1 and CTLA-4 relieves PI3K/AKT/mTORC1 signaling and allows increased T-cell stimulation and metabolism, inducing an effector-like phenotype. In addition, this metabolic shift induces epigenetic reprogramming inducing effector functions and longevity. Although PD-1 and CTLA-4 are the most extensively targeted ICB candidates, several other co-inhibitory and co-stimulatory molecules modulate T-cell metabolism. Notably, T-cell immunoglobulin mucin receptor 3 (TIM-3) downregulates glycolysis and GLUT1 expression [228], while LAG3 downregulates OXPHOS [229]. Similarly, inhibiting 4-1BB and OX40 enhances T-cell OXPHOS and promotes effector function and longevity [230]. Taken together, immune checkpoint molecules induce metabolic dysfunction and thus affect anti-tumor immunity, suggesting that combining ICB with metabolic regulators is a potential strategy to improve treatment efficacy.

#### *3.4. Glycolysis-Targeting Therapies to Improve Immunotherapy Efficacy*

mTOR is an oncogenic molecule that contributes to the regulation of metabolism in TILs. The inhibition of the mTOR signaling axis downregulates the malignant phenotype of cancer cells. Owing to their inhibitory effects, several rapamycin analogs have been approved for treating cancers [231,232]. However, it was shown that mTOR inhibition could diminish anti-tumor immunity [194] as these inhibitors directly affect the lineage differentiation-determining glycolytic activity in T-cells. It was shown that rapamycin suppresses Th17 differentiation and promotes Treg differentiation under TGFβ induction [233,234]. Further, the activation of AKT-mTORC1 signaling was associated with T-cell function restoration and the reduced expression of PD-1 and TIM-3 [235]. Additionally, the over-activation of mTORC1 affects the immunosuppressive function of Tregs, while low mTORC1 levels enhance Treg activity [236]. These suggest that an optimized inhibition of the PI3K-mTORC1 signaling axis is critical for improving the efficacy of immunotherapies.

Metformin has shown promising effects in different cancers [195] and has been shown to regulate metabolism by interacting with AMPK, the PI3K-AKT-mTOR axis, and HIF-1α [237,238]. Additionally, metformin promotes the cancer-killing capacity of CD8+ T-cells by modulating glycolysis [239–241] and downregulates immune checkpoint expression and glycolytic flux through HIF-1α inhibition [242,243]. Furthermore, it was shown that metformin stops the cancer cells from using the lactate and ketone bodies produced by cancer-associated fibroblasts as nutrients and thus suppresses cancer progression [244]. A recent study combined 2-DG, an HK inhibitor [245], BAY-876, a GLUT-1 inhibitor, and chloroquine and developed the nano-drug, D/B/CQ@ZIF-8@CS, which inhibited glycolysis and improved anti-CTLA-4 immunotherapy by reducing Treg metabolic fitness [246]. Tregs pretreated with 2-DG showed enhanced inhibition of T-cell proliferation in ovarian cancer [247]. Furthermore, HK upregulates PD-L1 expression in cancer cells, and combining the HK inhibitor, Lonidamine, with anti-PD-1 therapy improved cancer cell elimination in a mouse model [248].

Knocking out glucose-6-phosphate isomerase (GPI), the enzyme that catalyzes the conversion of glucose-6-phosphate (G6P) to fructose-6-phosphate (F6P), upregulated OXPHOS and sustained the survival of cancer cells [245]. Additionally, GPI inhibition selectively eliminated inflammatory encephalitogenic and colitogenic Th17 cells without affecting the homeostatic microbiota-specific Th17 cells [249]. However, it remains unknown whether GPI-targeted therapies would improve the efficacy of immunotherapy.

The GAPDH inhibitor, dimethyl fumarate (DMF), promotes the oxidative PPP and inhibits glycolysis and OXPHOS in cancer cells. This reduces the competition between cancer cells and T-cells for glucose consumption and promotes the efficacy of ICB and IL-2 therapy [250]. Low-dose osimertinib was shown to inhibit GAPDH and tumor endothelial glycolysis and promote vascularization and immune cell infiltration and thus improve the efficacy of anti-PD-1 therapy [251]. Inhibiting fructose-2,6-bisphosphatase 3 (PFKFB3), which is upregulated in several cancers, repressed glycolysis and upregulated PD-L1 expression [193]. On the contrary, glucose deficiency upregulated PD-L1 through the EGFR/ERK/c-Jun pathway, leading to the upregulation of PFKFB3, and promoted glycolysis [252,253]. This suggests the existence of a positive feedback loop between metabolism and checkpoint molecules [254]. A dual-target drug comprising paclitaxel and the PFKFB3 inhibitor, PFK15, blocked cancer-associated fibroblast-mediated cancer cell growth and reduced the lactate concentration in the TME [193]. PFK15 was also shown to upregulate PD-1 and LAG-3 expression in the context of type 1 diabetes [171]. Pyruvate kinase isoform M2 (PKM2), the final rate-limiting enzyme in glycolysis, promotes PD-L1 expression in macrophages, DCs, and tumor cells and contributes toward accelerated tumor progression [255]. High PKM2 expression was associated with a poor prognosis of pancreatic ductal adenocarcinoma, and the knocking down of PKM2 improved the efficacy of anti-PD-1/PD-L1 therapy [256].

ENO-1, which catalyzes the conversion of 2-phosphoglycerate to PEP and also acts as a plasminogen receptor and a DNA-binding protein, was shown to be overexpressed in several cancers. [64]. A pan-cancer analysis showed that ENO-1 expression correlated with immune cell infiltration, including B cells, CD8+ and CD4+ T-cells, macrophages, neutrophils, and dendritic cells [257]. The presence of autoantibodies against ENO-1 correlated with a better prognosis in PDA, suggesting that ENO-1 was a good molecular candidate for improving immune cell response to cancers [258]. Antibodies against ENO-1 were detected in approximately 60% of patients with PDAC, and ENO-1-specific T-cell responses are observed in patients who have the anti-ENO-1 antibodies. In line with this, an ENO-1 DNA vaccine induced an antibody and cellular response and increased the median survival in mouse models of PDA [259]. ENO-1-targeting DNA vaccines have shown prophylactic and therapeutic potential in PDAC mouse models by inducing complementdependent cytotoxicity and immune cell response [144,259,260]. In a spontaneous mouse model of PDAC, co-treatment with gemcitabine and ENO-1 DNA vaccine enhanced CD4 anti-tumor activity and impaired tumor progression [261,262]. Additionally, a recent study showed that targeting ENO-1 using specific antibodies targets multiple TME niches involved in prostate cancer (PC) progression and bone metastasis via a plasmin-related mechanism [263]. These show the potential of ENO-1 targeted therapies in improving the efficacy of immunotherapies.

Optimizing T-cell metabolism is a promising strategy for improving cancer immunotherapy. Metabolic modification can potentially increase stemness and long-term memory, enhance effector functions, and reduce T-cell exhaustion. Future studies should identify key metabolic transitions and regulatory steps that are differently regulated in cancer cells and immune cells and develop effective targeting strategies to enhance the synergistic effects of metabolic modulation and cancer immunotherapy [264].

#### **4. Targeting Glycolysis to Enhance Hormonal Therapy**

Hormonal therapy has shown remarkable advancement as a therapeutic strategy for cancers dependent on hormones, especially in breast, prostate, and other gynecological cancers. Aromatase inhibitors (AI), estrogen receptor (ER) antagonists, ER modulators, anti-estrogens, and GnRH agonists are effective therapeutic drugs and have shown high success rates in patients with hormone-sensitive recurring or metastatic gynecologic malignancies [265]. Hormone therapy interferes with the hormone-dependency of cancer cells by limiting hormone production in the body [266]. While hormonal therapy has improved survival and reduced recurrence in different cancer types [266], de novo or acquired resistance to hormonal therapy is a major clinical problem that requires the development of innovative strategies [264]. Resistance to hormonal therapy invariably occurs in most patients with ER+ metastatic BC and castration-resistance PC (CRPC) [267]. Metabolic reprogramming is an inherent feature of endocrine-resistant cancer cells, implicating that combination therapy with metabolic regulators and conventional hormonal therapy might be beneficial in overcoming resistance [268]. However, it is unclear whether metabolic rewiring is a cause or consequence of endocrine resistance, and several studies are investigating the cross-talk between hormone signaling and cancer cell metabolism [269]. Somatic mutation in estrogen receptors is related to the clinical development of the resistance to hormone therapies [268,270–272]. The Y537S mutation in ER-α enhanced mitochondrial metabolism and glycolysis in BC cells. The Y537S mutation is also associated with poor clinical outcomes, suggesting that enhanced glucose metabolism is a highly conserved mechanism of endocrine resistance [268].

Elevated glucose levels resembling hyperglycemia in BC cells have been attributed to a reduced response to tamoxifen therapy and could act as a marker for responses to hormonal therapy [273,274]. An increased glycolytic rate is a characteristic feature of tamoxifen-resistant cells, and inhibiting glycolysis is expected to restore tamoxifen sensitivity [273,275]. Elevated glycolysis in BC cells is also associated with mitochondrial malfunction and upregulated AKT/mTOR and HIF-1α signaling pathways. Tamoxifenresistant BC cells escape cell death by increasing autophagy through the inactivation of TOR-S6K via the HK2 pathway [276]. Glycolytic inhibition by the knockdown of HK2 or 3BrPA treatment downregulated AKT/mTOR signaling and could be a therapeutic strategy to overcome tamoxifen resistance in BC [277]. A recent study investigated the potency of a combination therapy employing low-dose tamoxifen (ERα antagonist) and metabolism inhibitors, 2-DG and CB-839 (glutaminolysis inhibitor), in improving the antiproliferation effect in tamoxifen-resistant ERα-positive BC cells. The triple combination showed superior cell growth inhibition by inducing apoptosis and c-Myc downregulation; however, a combination of tamoxifen with 2-DG did not show significantly strong inhibition of cell viability [278,279]. The pharmacological inhibition of glycolysis with PFK-158, a PFKFB3 inhibitor, with tamoxifen or fulvestrant has been explored as a potential therapeutic intervention to overcome endocrine resistance. PFKFB3 upregulation, with an elevated basal expression of PFKFB3 mRNA, is observed in endocrine-therapy-resistant BC cells and is associated with adverse recurrence-free survival in BC patients. The anti-tumor effect of PFK-158 is exacerbated when combined with tamoxifen and fulvestrant treatment [280]. PFKFB3 inhibition activated necroptotic markers receptor-interacting kinase 1 (RIPK1) and mixed lineage kinase domain-like pseudokinase (MLKL), implicating the possible mechanism of PFK-158-induced cell death [281]. In a long-term estrogen deprivation model (LTED) of AI resistance, cancer cells were demonstrated to have increased glycolysis dependency. The inhibition of glycolysis with HK2 inhibitors, along with AI, and letrozole, reduced cell viability [282]. Dietary interventions that target metabolic rewiring have also been shown to improve the efficacy of endocrine therapy in liver metastatic BC patients. Metastatic burden in the liver increases with increasing carbohydrate percentage in the diet. A fasting-mimicking diet increased the efficacy of fulvestrant treatment and reduced the metastatic burden in BC liver metastatic models, providing a proof-of-concept for a more straightforward strategy to circumvent drug resistance, with potential applicability in other cancer types as well [283].

Metabolic reprogramming is emerging as a crucial mechanism contributing to resistance to endocrine therapy in PC [284]. The expression of key glycolytic enzymes, including

LDH-A, MCT-4, and GLUT1, is elevated in mCRPC patients [285]. Glycolysis inhibition by targeting GLUT1 plays an important role in drug response in prostate PC [286]. In PC, elevated androgen levels increase glucose uptake and upregulate the expression of GLUTs, implying a cross-talk between androgen signaling and glycolytic pathway, a mechanism that protects PC cells from glucose-deprivation-induced oxidative stress [287,288]. NF-κB-mediated GLUT1 overexpression and upregulated glucose metabolism are associated with enzalutamide resistance in PC [289,290]. In xenograft models of CRPC and enzalutamide-resistant PC patients, GLUT1 inhibition significantly reduced tumor volume and displayed synergistic effects with androgen receptor (AR)-targeted therapy [286]. Glycolytic inhibitors, gossypol (LDH-A inhibitor), and AZD3965 (MCT-1 inhibitor) are currently in clinical trials as potential glycolysis-targeting agents in mCRPC. Progesterone treatment was reported to have an anti-tumor effect in glioblastoma multiform (GBM) in vitro and in vivo and improved the efficacy of temozolomide [291]. Recently it was shown that high-dose progesterone treatment inhibits GBM growth by inhibiting the key modulators of glycolytic metabolism. This early observation highlighted the potential of progesterone in metabolic reprogramming; however, more direct evidence is essential to validate this, and future studies should determine the synergistic effect of direct glycolytic inhibitors and progesterone in GBM treatment [291].

Endocrine resistance remains a major clinical barrier that requires the development of novel strategies to circumvent the resistance. Several mechanisms that contribute to endocrine resistance have been identified. Metabolic rewiring is frequently observed in most cancer cells that exhibit resistance, and targeting glucose metabolism with well-established glycolytic inhibitors has shown to enhance the sensitivity to endocrine therapy in breast and PC models. The mutual interplay between glucose metabolism and androgen receptor/ER signaling implies that combination approaches of endocrine therapy with metabolic modulators could be a standard-of-care to overcome resistance. Dietary interventions that modulate glucose metabolism have also been demonstrated to be an interesting strategy for evading resistance to therapy. Well-designed clinical trials are urgently needed to elucidate the clinical utility of the strategies mentioned above and to develop metabolic drugs as routine standard-of-care in endocrine-resistant cancer patients in clinical settings.

#### **5. Targeting Glycolysis to Improve Photodynamic Therapy**

Photodynamic therapy (PDT) is a relatively new, minimally invasive therapeutic procedure that relies on the selective accumulation of a photosensitive compound in the cancer cells, which, on excitation with light of an appropriate wavelength, would generate ROS, predominantly singlet oxygen, within the cancer cells, and eventually kill the cancer cell, with minimal damage to the surrounding tissue [292–294]. Although PDT is widely used to treat several cancers, its efficacy is limited by several factors, including the effective irradiation of deep tissue. Therefore, several studies have attempted to improve the efficacy of PDT by combining it with other chemotherapeutic agents. It has been shown that glycolytic inhibitors disrupt cancer cell metabolism, elevate the cellular ROS level, and disrupt the mitochondria, resulting in cell death [295]. Therefore, when combined with PDT, glycolytic inhibitors could, in theory, enhance the levels of cellular ROS and thus trigger increased cancer cell death.

5-aminolevulinic acid (5-ALA) is one of the most commonly used photosensitizers for photodynamic therapy. 5-ALA is a naturally occurring non-proteinogenic δ-amino acid synthesized in the mitochondria by the condensation of glycine and succinyl-CoA by mitochondrial 5-ALA synthase (ALAS). This is the first committed step toward heme biosynthesis. The final precursor of heme is Protoporphyrin IX (PpIX), which is a highly potent photosensitizer. The exogenous supplementation of 5-ALA overrides the normal feedback inhibition of ALAS and results in the accumulation of PpIX selectively in cancer cells, owing to the differences in the heme biosynthesis pathway enzyme activities between the cancer cells and normal cells. This cancer-specific accumulation of PpIX is exploited for selectively purging cancer cells by PDT and for the visualization of tumor

tissue by photodynamic diagnosis (PD). 5-ALA was approved by the U.S FDA in 2017 as an adjunct for the visualization of malignant tissue in grade III and IV glioma (NDA 208630/SN0014) and is currently used in the clinic to guide the resection of malignant glioma and glioblastoma. In addition to its role as a precursor of PpIX, 5-ALA has been reported to enhance aerobic bioenergetics [296], promote mitochondrial protein expression, and stimulate heme-oxygenase-1, triggering heme degradation [297]. Figure 3 describes the heme biosynthetic pathway, and how it is correlated with glycolysis. 208630/SN0014) and is currently used in the clinic to guide the resection of malignant glioma and glioblastoma. In addition to its role as a precursor of PpIX, 5-ALA has been reported to enhance aerobic bioenergetics [296], promote mitochondrial protein expression, and stimulate heme-oxygenase-1, triggering heme degradation [297]. Figure 3 describes the heme biosynthetic pathway, and how it is correlated with glycolysis.

and disrupt the mitochondria, resulting in cell death [295]. Therefore, when combined with PDT, glycolytic inhibitors could, in theory, enhance the levels of cellular ROS and

5-aminolevulinic acid (5-ALA) is one of the most commonly used photosensitizers for photodynamic therapy. 5-ALA is a naturally occurring non-proteinogenic δ-amino acid synthesized in the mitochondria by the condensation of glycine and succinyl-CoA by mitochondrial 5-ALA synthase (ALAS). This is the first committed step toward heme biosynthesis. The final precursor of heme is Protoporphyrin IX (PpIX), which is a highly potent photosensitizer. The exogenous supplementation of 5-ALA overrides the normal feedback inhibition of ALAS and results in the accumulation of PpIX selectively in cancer cells, owing to the differences in the heme biosynthesis pathway enzyme activities between the cancer cells and normal cells. This cancer-specific accumulation of PpIX is exploited for selectively purging cancer cells by PDT and for the visualization of tumor tissue by photodynamic diagnosis (PD). 5-ALA was approved by the U.S FDA in 2017 as an adjunct for the visualization of malignant tissue in grade III and IV glioma (NDA

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 17 of 35

thus trigger increased cancer cell death.

**Figure 3.** The heme biosynthesis is connected with glucose and glutamine metabolism. Enhanced glycolysis and glutaminolysis in the cancer cell might contribute to the elevated rate of heme synthesis and the upregulation of several heme biosynthesis pathway enzymes in the cancers. Addition of exogenous 5-ALA bypasses the feedback inhibition of ALAS and increase the accumulation of protoporphyrin IX (PpIX) in the cancer cells. The elevated accumulation of PpIX is exploited for fluorescence-guided detection (photodynamic diagnosis) of cancers. Further, irradiating PpIX-accumulating cancer cells with light generated singlet oxygen and ROS that kills the cancer cell, with minimal damage to the surrounding tissue (photodynamic therapy; PDT). The interdependency of **Figure 3.** The heme biosynthesis is connected with glucose and glutamine metabolism. Enhanced glycolysis and glutaminolysis in the cancer cell might contribute to the elevated rate of heme synthesis and the upregulation of several heme biosynthesis pathway enzymes in the cancers. Addition of exogenous 5-ALA bypasses the feedback inhibition of ALAS and increase the accumulation of protoporphyrin IX (PpIX) in the cancer cells. The elevated accumulation of PpIX is exploited for fluorescence-guided detection (photodynamic diagnosis) of cancers. Further, irradiating PpIXaccumulating cancer cells with light generated singlet oxygen and ROS that kills the cancer cell, with minimal damage to the surrounding tissue (photodynamic therapy; PDT). The interdependency of glucose metabolism and the heme synthesis pathways suggest that targeting glycolysis could enable the modulation of PpIX accumulation in the cancer cells. The image was created using Biorender app.

the modulation of PpIX accumulation in the cancer cells. The image was created using Biorender app. A recent study showed that 5-ALA is a potent competitive inhibitor of LDH with efficacies comparable to oxamate (OXM) and tartronate (TART) [298]. They showed that treatment with 5-ALA induced glycolysis inhibition and triggered cell death in glioblastoma cell lines. Further, it was shown that up to 25% of the 5-ALA was used for glycolysis inhibition in these cells, leaving a lower amount of 5-ALA for conversion to PpIX and subsequent use as a photosensitizer for PDT and PD. Treating the glioblastoma cells with an LDH inhibitor before 5-ALA treatment enhanced the efficacy of PDT by 15%. Precise delineation of the tumor–normal tissue is of critical importance, especially in brain cancer surgeries, and a 15% increase in PD efficacy is a significant improvement.

glucose metabolism and the heme synthesis pathways suggest that targeting glycolysis could enable

Another study showed that exogenous 5-ALA suppressed oxidative metabolism and glycolysis and reduced cell proliferation in ovarian and BC cells [299]. Further, 5-ALA also destabilized Bach1 and inhibited cancer cell migration. The study also showed that 5-ALA-induced suppression of oxidative metabolism and glycolysis was mediated through different mechanisms in BC and ovarian cancer but involved Bach1 destabilization, AMPK activation, and the induction of oxidative stress. Additionally, an inverse relationship between oxidative metabolism and ALA sensitivity was revealed.

It was also shown that the administration of 5-ALA for 6 weeks reduced the plasma glucose levels in rats without affecting their plasm insulin levels and induced HO-1 expression in the liver and white adipose tissue [297]. An increase in HO-1 indicates an increase in heme in the liver, which promotes the formation of nuclear receptor subfamily 1 (Rev-Erbα) with its corepressor nuclear receptor co-repressor 1 (NCOR), which in turn downregulates the enzymes involved in gluconeogenesis such as PEPCK and G6Pase, resulting in reduced glucose production in the liver [300]. On the other hand, 5-ALA administration enhances glucose metabolism in the adipocytes by decreasing the total amount of adipose tissue or decreasing the number of mitochondria in the adipose tissue [301]. It has been shown that the induction of peroxisome proliferator-activated receptor γ coactivator 1-α (PGC-1α), a master transcription coactivator, increases heme synthesis by upregulating ALAS [302]. Glucose intake repressed PGC-1α mediated ALAS in this study, suggesting that nutrient stress could trigger increased heme synthesis. Nutrient stress is a characteristic feature of the tumor microenvironment, and earlier studies have shown that BC cells (MCF7) grown under glucose deprivation produced higher levels of PpIX than those cultured under standard conditions [66,303]. Furthermore, co-treatment with glycolytic inhibitors and 5-ALA reduced intracellular PpIX levels [304]. This has been attributed to the inactivation of ABC transporters induced by ATP depletion, which in turn decreased the flux of precursors into the cell [303,305]. Interestingly, other studies have reported an increase in cellular PpIX accumulation with the inhibition of ABC transporters [306,307]. On the other hand, a combined treatment with 5-ALA-PDT and dichloroacetate, an inhibitor of pyruvate dehydrogenase, showed improved efficacy in BC (MCF 7) cells [308].

In another study, an 18 h glucose deprivation prior to PDT reduced intracellular glutathione levels and increased the cytotoxicity of PDT [309]. A similar increase in PDT efficacy was observed when inhibitors of glutathione synthesis (buthionine-sulfoximine) or its regeneration (1,3-bis-(2-chlorethyl)-1-nitrosourea) were used for co-treatment with PDT [310]. Changes in the availability of glycolytic substrates affect NADPH availability in the cells. NADPH is a critical agent involved in the anti-oxidative defense mechanisms of the cell. As mentioned, PDT relies on increased ROS in the cancer cells to kill them. PDTbased ROS generation under conditions of impaired glucose and glutathione metabolism results in much higher intracellular ROS levels, contributing to increased efficacy. In line with this, ROS scavengers were shown to protect the cells from ALA-PDT-induced damage [311].

A recent study showed that the metabolic reprogramming toward aerobic glycolysis in cancer cells contributes to the development of resistance to 5-ALA-PDT. Further, they showed that treatment with metformin reduced aerobic glycolysis and increased OXPHOS in squamous cell carcinoma cells, and improved the cytotoxic effect of PDT by increasing PpIX production, ROS generation, and AMPK expression, and inhibiting the AKT/mTOR pathway [312]. In another study, combining glycolysis inhibitor 2-DG with 5-ALA induced an enhanced accumulation of PpIX in HepG2 liver cancer cells [313], contributing to improved treatment efficacy. Further, treatment with 2-DG and 3-bromopyruvate (3-BP) synergistically improved the efficacy of PDT in BC cells [314].

Oncogenic transformation has been reported to upregulate glycolytic enzymes and contribute to the increased exogenous ALA-treatment-induced PpIX accumulation in cells [315]. The increased PpIX accumulation in cancer cells is often attributed to their lower ferrochelatase (FECH) levels. FECH is the terminal enzyme in the heme synthesis pathway that catalyzes the chelation of PpIX with the ferrous ion to generate heme. Lower FECH levels have shown a significant association with the cancer grade, the TNM stage, unfavorable prognosis, and impaired immune cell infiltration in clear cell renal cell carcinoma [316]. However, oncogenic Ras/MEK has been shown to increase the conversion of

PpIX to heme by increasing HIF-1α expression and thereby promoting FECH activity [317]. Inhibiting MEK decreased HIF-1α expression and FECH activity and increased 5-ALAinduced PpIX accumulation in cancers [317,318]. Oncogenic Ras/MEK signaling-induced HIF-1α expression has also been implicated in increasing glycolytic flux and driving cancer progression [319,320]. Although 5-ALA-induced PpIX accumulation in cancer cells is a well-documented and well-studied concept, our knowledge of the underlying mechanisms remains largely vague. The inter-dependencies between oncogenic transformation, glycolytic flux, and heme metabolism should be further studied to identify the optimal targeting strategy to enhance 5-ALA-induced PpIX accumulation in cancer cells and to improve the efficacies of PDT and PD.

#### **6. Conclusions**

The metabolic signatures of tumor cells are different from normal cells, which allows the tumor cells to adapt to the increased energy and metabolite demands [321,322]. Several signaling mechanisms that regulate or hijack the canonical reactions in glucose metabolism have been identified; however, there is no universal mechanism underlying the reprogramming of glucose metabolism in all cancers. Elevated glucose metabolism, hypoxia-induced GLUT, LDH-A, and PFKFB3 overexpression, and the AKT- and c-Myc-mediated transcriptional activation of HK2 are observed in most cancer cells, and these metabolic changes may be exploited for developing effective therapeutic approaches. The tumor microenvironmental regulation and immune suppressive effects of metabolic rewiring are also crucial players in cancer cell progression, survival, and resistance. Apart from glycolytic modulation, mitochondrial dysfunction, elevated ROS production, and dysregulated TCA, cycle enzymes also participate in oncogenic signaling and tumor progression, which were not discussed in this review, and are elegantly reviewed elsewhere [323,324].

Elevated glucose metabolism and nutrient uptake have been exploited for tumor diagnosis in the clinic, and F-18 fluoro-2-deoxyglucose PET is widely used in the clinical staging of cancers [325,326]. However, glycolysis inhibition has not been exploited to its full potential in cancer therapy and it has not translated into the clinic. The inhibition of glycolysis may have undesirable consequences as normal cells also use the same glycolytic enzymes, and it is hence necessary to identify the enzymes or enzyme isoforms that are specifically upregulated or preferentially used by cancer cells. Additionally, several of the glycolysis inhibitors that have been developed for clinical use have been shown to have off-target effects, killing non-cancerous cells [327]. In addition, though the inhibition of glycolysis might inhibit cancer cell proliferation, cancer cells may adapt by upregulating OXPHOS or glutaminolysis, which could result in the development of resistance to therapy, in addition to co-morbidities such as cachexia in patients. This rewiring of metabolic pathways also poses challenges to precision therapies. In fact, previous studies have highlighted the dispensability of the Warburg effect in cancer cell growth and have shown that a complete disruption of glycolysis would require the deletion of both LDHA and LDHB genes. The study evaluated a potent LDH A/B dual inhibitor GNE-140 and demonstrated that the "glycolytic Warburg" phenotype of tumor cells depends on both LDH A and LDH B expression, and is a dispensable phenotype that can be replaced by OXPHOS [328]. This emphasizes the importance of evaluating the simultaneous inhibition of glycolysis and OXPHOS as a therapeutic strategy. In cell lines that are innately resistant to GNE-140 as they predominantly use OXPHOS, the inhibition of OXPHOS sensitized the cells to GNE-140 treatment [329]. Understanding the pathways that contribute to resistance in glycolysis targeting drugs is hence crucial [245,329–331].

Despite early indications, glycolytic inhibitors such as 2-DG have failed in the clinical setting due to their limited effects as a monotherapy agent. In addition, it might be necessary to combine glycolytic inhibitors with other agents that target alternate pathways that are activated in response to glycolytic inhibition. Studies have shown promising effects when metformin (that inhibits mitochondrial complex 1) was used along with glycolytic inhibitors. The simultaneous targeting of MCT1 inhibitor AZD3965 and the mitochondrial

respiratory complex 1 inhibitor IACS01759 is a more clinically relevant strategy compared with targeting only one pathway in B-cell lymphomas [332]. Using glycolytic inhibitors as adjuvant therapy with already approved conventional therapies, including chemotherapy, radiotherapy, immunotherapy, and photodynamic therapy, is an attractive strategy, and several early studies have shown promising results. Dietary interventions and lifestyle changes also affect the metabolic landscape of cancer cells, and studies should address this issue in detail to emphasize their clinical implications.

Though metabolic rewiring unquestionably affects cancer cell proliferation, the translation of metabolic reprogramming to the clinic must overcome several hurdles [333]. In-depth analytical and extensive pre-clinical studies should identify targetable metabolic enzymes/enzyme isoforms that are efficacious in different tumor types with minimal toxicity to normal cells. Another major challenge in the clinical development of cancer therapeutics is the need to identify patient groups that would benefit from the therapy. Dependency on metabolic pathways is not universal for all cancer types, which makes it all the more important to identify appropriate patient groups to avoid unwanted adverse effects and toxicity. Technical hurdles in measuring metabolism in vivo also add more complexity. It is essential to integrate genetic and metabolic biomarkers and tissue information to optimize the criteria for patient selection. An interesting study published recently showed the possibility of using glycolytic enzymes as a surrogate for glycolytic activity in cancer cells, using liquid biopsy. Cells expressing high levels of HK2 were identified in both cytokeratin (CK)-positive and CK-negative CTC populations isolated from lung adenocarcinoma patients [334]. The possibility of monitoring the glycolytic metabolic rewiring in real-time using liquid biopsy samples and downstream single-cell level molecular, genomic, and metabolic studies is likely to create a paradigm shift in the clinical utility of glycolytic modulations in cancer therapy. Well-designed clinical trials that unravel the metabolic dependencies of different cancer types and their association with genetic and histopathological information might answer several questions on the complex and sophisticated nature of cancer metabolism. Metabolomic studies, with powerful technological backing, would take the lead in the early identification of cancer risk factors and aid in improving cancer therapy and cancer screening, diagnosis, and therapy monitoring in the coming years.

**Author Contributions:** Conceptualization, C.C. and V.S.C. resources, C.C. and V.S.C.; writing—original draft preparation, C.C., and V.S.C.; writing—review and editing, C.C., V.S.C., and K.S.; supervision, Y.S.; project administration, K.S.; funding acquisition, Y.S. and K.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 202200070001, No. 202100400001).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Review* **Taste and Smell Disorders in Cancer Treatment: Results from an Integrative Rapid Systematic Review**

**Tania Buttiron Webber <sup>1</sup> , Irene Maria Briata <sup>1</sup> , Andrea DeCensi 1,2 , Isabella Cevasco <sup>3</sup> and Laura Paleari 4,\***


**Abstract:** Taste and smell disorders (TSDs) are common side effects in patients undergoing cancer treatments. Knowing which treatments specifically cause them is crucial to improve patients' quality of life. This review looked at the oncological treatments that cause taste and smell alterations and their time of onset. We performed an integrative rapid review. The PubMed, PROSPERO, and Web of Science databases were searched in November 2022. The article screening and study selection were conducted independently by two reviewers. Data were analyzed narratively. Fourteen studies met the inclusion criteria and were included. A high heterogeneity was detected. Taste disorders ranged between 17 and 86%, while dysosmia ranged between 8 and 45%. Docetaxel, paclitaxel, nab-paclitaxel, capecitabine, cyclophosphamide, epirubicin, anthracyclines, and oral 5-FU analogues were found to be the drugs most frequently associated with TSDs. This review identifies the cancer treatments that mainly lead to taste and smell changes and provides evidence for wider studies, including those focusing on prevention. Further studies are warranted to make conclusive indication possible.

**Keywords:** dysgeusia; dysosmia; taste; smell; cancer treatment; rapid review

**Citation:** Buttiron Webber, T.; Briata, I.M.; DeCensi, A.; Cevasco, I.; Paleari, L. Taste and Smell Disorders in Cancer Treatment: Results from an Integrative Rapid Systematic Review. *Int. J. Mol. Sci.* **2023**, *24*, 2538. https://doi.org/10.3390/ ijms24032538

Academic Editor: Takeo Nakanishi

Received: 23 December 2022 Revised: 18 January 2023 Accepted: 20 January 2023 Published: 28 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

Global cancer statistics estimate that around 19.3 million new cancer diagnoses occurred in 2020. Lung cancer remained the leading cause of death, with an estimated 1.8 million deaths (18%), followed by colorectal (9.4%), liver (8.3%), stomach (7.7%), and female breast (6.9%) cancers [1].

Taste and smell disorders (TSD) are common side effects in cancer patients undergoing chemotherapy (CT) treatments but often are described as single entities and patients may have difficulty in identifying them [2]. The reported prevalence of taste disturbance ranged from 20% to 86% [3], and its development occurs approximately 2–3 weeks after the start of cancer treatment and persists throughout the duration of the therapy [4]. The prevalence of patients with dysosmia is in the range of 5–60% [5].

Interestingly, in the literature, it is reported that only a few patients report taste and smell alterations spontaneously and these symptoms are often underestimated by oncologists and nurses [6]. A study by Gill et al. reported a discrepancy in the importance given to retaining a normal sense of taste and smell, as reported by patients and by the multidisciplinary team involved in their care (*p* < 0.013) [7]. We previously hypothesized that the hesitancy of physicians in approaching these disorders may be due to a "cultural aspect" where the physician tends to underestimate and leave untreated the adverse events (AEs) related to therapies that do not have a clinical implication [8]. However, it is important to consider these disorders as they can lead to reduced food enjoyment and, most importantly, an inappropriate nutrient intake, with a high impact on the nutritional status, quality of life, and possibly on the efficacy of therapy itself [9].

Literature describes five basic tastes: sweet, sour, bitter, salty, and umami [10,11]. The sense of taste starts with the activation of the taste receptors located on the microvilli or taste receptor cells. These cells are clustered and together they form taste buds. Taste receptor cells are modified epithelial cells that can detect and process gustatory, olfactory, and trigeminal stimulation [12]. Dysgeusia can be classified as follows: (1) ageusia, which is a complete lack of taste; (2) hypogeusia, which is a decreased taste sensitivity; (3) hypergeusia, which is a heightened taste sensitivity; and (4) phantageusia, which is the perception of an unpleasant taste in the absence of a corresponding stimulus in the environment [9]. The sense of smell is even more complex than the sense of taste. Over a trillion different smells can be identified. There are two ways for odors to reach the olfactory epithelium: via the ortho-nasal passage or via the retro-nasal passage [13]. In general, four categories of smell disorders are classified depending on how they impact odor perception. (1) Anosmia is the absence of smell perception; (2) hyposmia is a quantitatively reduced ability to perceive scent; (3) parosmia is a qualitative distortion of an ordinarily detected smell; and (4) phantosmia is the perception of odors when none are present [14].

In addition to cancer treatment, including radiotherapy and surgical treatments, other factors may contribute to taste and smell disorders, such as age, oral infections, smoking, alcohol abuse, chronic diseases such as diabetes, hypertension, and chronic rhinosinusitis, and the type of cancer [15–17]. In fact, the study by Dhuibhir and colleagues showed a high prevalence of taste and smell disorders in newly diagnosed cancer patients before treatment [18].

Currently, TSD can be assessed through clinical methods (objective) or self-reported by patients (subjective). The objective methods assess the oral sensitivity to taste agents through the thresholds of the five taste qualities. The numerical results facilitate the comparison of the taste perception abilities between populations [19] but do not reflect the 'real-world' taste experience [20] as they do not capture dimensions of taste that are important to patients, such as flavor, food enjoyment, or hedonic changes [21]. For this reason, patient-reported questionnaires and qualitative research methods that capture patients' individual experience are recommended [6,22].

A review by Enriquez-Fernandez et al. [23] reports a growing interest in the assessment of taste and smell changes in cancer patients but presents limitations in terms of the heterogeneity in the number of items, assessment range, and in the domains of taste changes. They suggest developing a standardized tool validated by patients to ensure that the terms associated with sensory changes are understood and reliably used by clinicians and researchers.

The aforementioned papers have mainly highlighted the pathophysiology, prevalence, clinical features, and assessment tools of chemosensory alterations. However, there are limited literature reviews highlighting the oncological therapies that lead to these alterations. The aim of this rapid review was to examine the existing and current literature on cancer treatments that can cause TSDs to develop prevention and education strategies in the future.

#### **2. Materials and Methods**

#### *2.1. Study Design*

Between October and November 2022, an integrative rapid review was conducted as a knowledge synthesis strategy to provide timely information [24]. Despite being thoroughly studied, the field of TSD is evolving due to novel cancer treatments. As a result, timely reviews can describe current research and report on clinical and organizational levels [25].

#### *2.2. Needs Assessment and Topic Selection*

The primary need was to map the most recent data on cancer drugs that cause TSD, to summarize the knowledge and enable nurses and oncologists to continuously improve the quality of care and patient management. Thus, the review question was: which oncological drugs cause taste and smell disorders? To this review, we decided to include only studies focused on cancer treatment for solid tumors in adults.

#### *2.3. Study Development*

According to the methodological process inspired firstly by Tricco and colleagues in 2017 [25], which was then further developed by Langlois et al. in 2019 [26], the following seven-stage process was implemented: (1) a needs assessment and topic selection; (2) study development; (3) literature search; (4) screening and study selection; (5) data extraction; and (6) risk-of-bias assessment. In addition, the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) guidelines were applied [27].

#### *2.4. Literature Search*

The literature search was performed independently by 2 reviewers (TBW and IMB) through the MEDLINE (via PubMed), PROSPERO, and Web of Science databases between October and November 2022. The inclusion criteria were: (a) studies on adult patients with solid tumors undergoing treatment with oncological drugs, (b) studies designed to detect the incidence and prevalence of TSD and/or assess the time of onset; (c) quantitative and qualitative primary studies; and (d) studies published in English within the past 10 years. The exclusion criteria were: (a) studies on patients with hematologic malignancies and (b) studies on patients undergoing radiation therapy. Therefore, the Medical Subject Headings [MeSH] and free-text words used were: "dysgeusia", "taste alteration", "anosmia", "olfaction disorders", "smell alteration", "therapeutics", "therapies", "treatments", "cancer", and "neoplasm". The research was limited to the last 10 years.

#### *2.5. Screening and Study Selection*

The screening of titles and abstracts was performed by three researchers (TBW, IMB, and LP) to identify the articles' eligibility in relation to the inclusion criteria. Then, an independent full-text evaluation was performed by the same researchers to determine if the studies fully met the inclusion criteria. When disagreements occurred, the final decision to include or exclude an article was made by a consensus. As reported in the PRISMA flow diagram (Figure 1), the flow of a study's inclusion is summarized together with the reasons for exclusion. At the end of the process, 14 papers were retrieved.

#### *2.6. Data Extraction*

The data were extracted and reported in a Microsoft Excel® spreadsheet. The following data were collected from each selected study and reported in the grid: the (a) author(s), year, country; (b) study design; (c) aims; (d) participants; (e) assessment tool used for TSD detection; (f) cancer treatment; and (g) key findings. The full grid is available as Table 1.

#### *2.7. Risk of Bias Assessment*

The review team shared in advance the decisions about inclusion and exclusion to prevent information and a selection bias. In addition, to ensure the consistency of the results, the following methodological requirements [28] were respected: (a) the verification of the study selection and data extraction were performed by three reviewers [TBW, IMB, LP]) and (b) an additional independent researcher (=a fourth reviewer [ADC]) contributed and reviewed the narrative synthesis and the summary table.

**Figure 1.** Flowchart of the data collection and selection process in accordance with PRISMA-ScR **Figure 1.** Flowchart of the data collection and selection process in accordance with PRISMA-ScR guidelines.







Abbreviations. CT: chemotherapy; HT: hormone therapy; 5-FU: 5-fluorouracil; TAs: taste alterations; FOLFOX: folinic acid, fluorouracil, oxaliplatin; FEC: epirubicin, fluorouracil, cyclophosphamide; EC: epirubicin, cyclophosphamide; FOLFIRI: folinic acid, fluorouracil, irinotecan; TA: taste alteration; TJ: carboplatin, paclitaxel; TPF: docetaxel, cisplatin, fluorouracil; GEMCARBO: gemcitabine, carboplatin; GEMOX: gemcitabine, oxaliplatin; CISGEM: cisplatin, gemcitabine; BC: breast cancer; CTCAE: Common Terminology Criteria for Adverse Events; Tx: taxane; CiTAS: Chemotherapy induced Taste Alteration Scale; PRO-CTCAE: patient-reported outcome; DTX: docetaxel; PTX: paclitaxel; nab-PTX: nab-paclitaxel; NSCLC: non-small cell lung cancer.

#### **3. Results**

#### *3.1. Study Characteristics*

A preliminary search aimed at exploring knowledge on cancer treatments causing TSDs was conducted by examining the MEDLINE database (via PubMed), Web of Science, and PROSPERO between October and November 2022. There were not recently published or ongoing reviews on this specific topic; meanwhile, a large number of articles on TSDs in cancer patients related to eating habits and quality of life have been published in recent years. The database searches, after the duplicates were removed, returned 703 articles, of which 113 were screened. Out of the 113 articles assessed for their eligibility, 14 studies met the inclusion criteria (Figure 1). The specific characteristics of each selected study that met the inclusion criteria are presented in Table 1. Most of the selected studies used a quantitative method [9,29–40], and only one was a qualitative study [41]. Nine studies focused on changes in the taste [30–32,34,36,37,39–41], 3 focused on smell [33,35,38], and 2 focused on a combination of both [9,29]. The study population among the studies was very different in terms of the cancer diagnosis, stage, treatment, line of therapy, and sample size. Moreover, a high degree of heterogeneity in the tools to assess TSD was observed, even within the study itself. In fact, some studies used a subjective evaluation [9,29,30] while others used validated questionnaires (e.g., CITAS) [30,36] or standardized measurement scales such as CTCAE [31,36,37]. Two studies used the Sniffin' Sticks test [33,38] while one study used taste recognition thresholds (TRTs) [36].

#### *3.2. Prevalence, Onset, Resolution of Taste and Smell Disorders*

Taste disorders were found in 17% to 86% of people and were linked to a poor appetite and a 10% weight loss [29,30]. Campagna et al. reported that 43% of participants complained of TSDs at the start of CT, and 75% reported it by the fourth week of treatment [30]. The worsening of symptoms occurs within 5–7 days immediately following the CT infusion and decreases by about 9% immediately before the next cycle [31]. The complete resolution of symptoms (e.g., from oxaliplatin) occurs within 6–8 weeks after the completion of treatment [31,41]. Xerostomia is strongly associated with a bad taste in the mouth (OR = 5.96; CI = 2.37–14.94; *p*-value = 0.000) and a loss of taste (OR = 5.96; CI = 2.37–14.94; *p*-value = 0.000) [29]. Salty

and sour were the most affected tastes (*p* = 0.001 and 0.05, respectively) [40]. Body surface, smokers, and people over the age of 70 had a significant negative impact on taste and smell [31,33,34]. The duration of CT (*p* = 0.001), female gender (*p* = 0.001), and location of the primary tumor in the uterus (*p* = 0.008), head and neck (*p* = 0.012), and testicles (*p* = 0.011) independently increased the risk of dysgeusia [37]. The prevalence of olfactory disorders ranged from 8% to 45%; in normosmics or hyposmics, the mean decrease in the threshold determination, discrimination, and identification (TDI score) was significant during the second cycle of cancer treatment and smoking and being over 50 years old were risk factors for smell alterations [9,29,33].

#### *3.3. Cancer Treatment*

Docetaxel is the main drug related to the occurrence of TSDs and is associated with a higher prevalence of more severe and longer taste alterations than paclitaxel or nab paclitaxel [36]. Anthracyclines, carboplatin, epirubicin, cyclophosphamide, capecitabine, and cisplatin/pemetrexed are also frequently related to TSDs [9,30]. Low levels of taste alterations were found in gemcitabine/carboplatin (GEMCARBO) and cisplatin/gemcitabine (CISGEM) combinations [30], as well as in cisplatin and gemcitabine administered individually [32]. Additionally, 5-fluorouracil (5-FU) or its oral analogues showed a high prevalence of TAs [34]. Docetaxel, previous CT, dry mouth, and peripheral neuropathy were significantly associated with taste alterations [36]. Regarding the effect of cisplatin on odor, detection, and identification abilities were unaffected by an administration of cisplatin; a decrease in pleasantness was observed only for food odors and not for non-food odors [35].

#### **4. Discussion**

In addition to oncology drugs, the literature reports hundreds of drugs from all major therapeutic classes that have been clinically reported to cause unpleasant and altered taste sensations when administered alone or in combination with other drugs. These unpleasant sensations include metallic and bitter tastes, a partial or complete loss of taste, and distortions and perversions of taste [42].

As suggested in the review by Schiffman et al., there are a number of topics which are useful for understanding the biological basis of drug-induced taste disorders: (1) the interaction of drugs with taste receptors on the apical side of the tongue in the oral cavity; (2) genetic differences among patients that affect the taste perception of drugs; (3) taste sensations caused by injectable drugs; (4) drug interactions caused by the use of multiple drugs; and (5) potential biochemical causes.

It is also important to recognize which groups are most vulnerable to this alteration. These include (1) the elderly, who use a disproportionate number of drugs [43], (2) people with certain genetic polymorphisms related to the perception of a bitter taste [44], (3) people with a reduced drug clearance [45], and (4) people with a drug metabolism [46].

TSDs in cancer patients are an often underestimated and underreported problem that may result from the disease and/or its treatment; this is probably because physicians and nurses do not regularly use standardized taste tests to verify and validate drug-related taste disorders in patients. Additionally, many disorders of taste cannot be categorized according to conventional tastes such as sweet (sucrose), sour (citric acid), salty (NaCl), bitter (quinine), or umami (monosodium glutamate). Dysgeusia and dysosmia alter the pleasure of eating and reduce appetite, which, especially in compromised patients, can lead to malnutrition, increased treatment-related toxicities, and a worsened quality of life. Therefore, the identification of risk factors, such as the use of a specific oncological treatment, that may promote the development of TSDs, is an important aspect to reduce the impact of this condition on these frail patients. Our research identified 14 articles published in the last 10 years that investigated cancer treatments leading to TSDs. In accordance with the existing literature, the range of taste alterations varies between 17% and 86% [29,30] and the severity of the symptoms varies during the cycle [30,31,35]. In fact, the symptoms

severity tends to worsen 5–7 days after the CT cycle and then diminish by about 10% before the following cycle [35]. Moreover, the taste alterations tend to persist for a long time [31,41], suggesting that the risk of malnutrition and a worsened quality of life may continue even after the end of the cancer treatment. Salty and sour tastes seem to be the tastes which are most affected by cancer treatments [40], so it might be useful to provide patients with specific nutritional guidance aimed to minimize the alterations. Dysosmia is less investigated but still its prevalence ranges from 8 to 45% of cancer patients [9,29]. According to the results of the current study, dysgeusia and dysosmia were more strongly associated with breast, gynecological, and colorectal cancer [32]. Docetaxel, paclitaxel, nab-paclitaxel, capecitabine, cyclophosphamide, epirubicin, anthracyclines, and oral 5-FU analogues were found to be the drugs most frequently associated with TSDs [9,30,34,36]. Other important risk factors for TSDs included the number of chemotherapy cycles, the female sex, the presence of distant metastases, and the primary tumor's location in the uterus, testicles, or head and neck [37]. An interesting correlation emerged between dysgeusia and peripheral neuropathy; numbness or tingling in the hands or feet (OR, 2.04; 95% CI, 1.25–3.57; *p* = 0.004) were significantly associated with TAs [36]. Knowing the factors most associated with TSDs is crucial for physicians and nurses to carefully monitor their occurrence and severity and to implement adequate prevention strategies. Sevryugin et al., in a recent review [3], summarized a wide range of therapy alternatives, including zinc and polaprezinc, radioprotectors, vitamins and supplements, anti-xerostomia agents, active swallowing exercises, nutritional interventions, delta-9-tetrahydrocannabinol, and photobiomodulation that can be used as a strategy to reduce TSDs.

The high heterogeneity among the selected studies in terms of the diagnosis, stage of disease, treatment, the instruments used to assess the TSDs, and the sample size makes it difficult to make firm conclusions. The limited number of studies exploring specifically which cancer therapies cause alterations in taste and smell leads us to hypothesize that these disorders have not yet been given due attention. In addition, none of the studies included in the present review considered new therapies, such as immunotherapy, suggesting that further studies are needed to investigate the impact of cancer therapies more comprehensively on TSDs. Most published studies relate taste and smell alterations to quality of life, so interventions in a preventive context would be necessary; although, there is no consensus on the prevention strategies to be used in this setting, so an algorithm for selecting the best treatment for TSDs was developed [3]. The algorithm can help the clinician to provide a therapeutic solution for chemosensory disorders or it can help the researcher to design an appropriate clinical trial to increase the knowledge on this underestimated problem.

#### *4.1. Study Limitations*

This rapid review aimed to highlight current cancer drugs that can cause changes in taste and smell; however, we know that there are no studies that take into consideration other therapies such as hormone therapy, target therapies, immunotherapies, and monoclonal therapy. Furthermore, the wide heterogeneity of the evaluation tools used, and the different moments of detection do not allow for an accurate generalizability of the results. Our results are not to be considered conclusive, as another limitation is that we explored a limited number of databases.

#### *4.2. Implications for Clinical Practice and Research*

Oncologists and nurses should be trained on treatments that induce taste and smell disorders to educate patients about proper nutrition and reduce the impact of these symptoms on their quality of life.

#### **5. Conclusions**

Taste and smell disorders are not life-threatening events for patients but have a significant impact on their quality of life. Oncologists, nurses, and nutritionists play an important role in the management of these chemotherapy-related symptoms, so further studies are

needed to provide specific information to patients on which oncology drugs cause dysgeusia or anosmia, the time of their onset and duration, and to support clinical governance strategies as well.

**Author Contributions:** T.B.W., I.M.B. and L.P. designed the literature search. T.B.W., I.M.B. and L.P. screened and reviewed articles, extracted and analyzed data. T.B.W. and I.M.B. wrote the first draft of the manuscript. L.P., I.C. and A.D. critically reviewed and revised the manuscript and approved the final draft. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Acknowledgments:** To my beloved son Rami.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
