*Review* **Involvement of Kynurenine Pathway in Hepatocellular Carcinoma**

**Shivani Krishnamurthy <sup>1</sup> , David Gilot <sup>2</sup> , Seong Beom Ahn <sup>1</sup> , Vincent Lam <sup>1</sup> , Joo-Shik Shin <sup>3</sup> , Gilles Jackie Guillemin 1,† and Benjamin Heng 1,\* ,†**


**Simple Summary:** The kynurenine pathway (KP) is a biochemical pathway that synthesizes the vital coenzyme, nicotinamide adenine dinucleotide (NAD<sup>+</sup> ). In cancer, the KP is significantly activated, leading to tryptophan depletion and the production of downstream metabolites, which skews the immune response towards tumour tolerance. More specifically, advanced stage cancers that readily metastasize evidence the most dysregulation in KP enzymes, providing a clear link between the KP and cancer morbidity. Consequently, this provides the rationale for an attractive new drug discovery opportunity for adjuvant therapeutics targeting KP-mediated immune tolerance, which would greatly complement current pharmacological interventions. In this review, we summarize recent developments in the roles of the KP and clinical trials examining KP inhibition in liver cancer.

**Abstract:** As the second and third leading cancer-related death in men and the world, respectively, primary liver cancer remains a major concern to human health. Despite advances in diagnostic technology, patients with primary liver cancer are often diagnosed at an advanced stage. Treatment options for patients with advanced hepatocarcinoma (HCC) are limited to systemic treatment with multikinase inhibitors and immunotherapy. Furthermore, the 5-year survival rate for these late-stage HCC patients is approximately 12% worldwide. There is an unmet need to identify novel treatment options and/or sensitive blood-based biomarker(s) to detect this cancer at an early stage. Given that the liver harbours the largest proportion of immune cells in the human body, understanding the tumour–immune microenvironment has gained increasing attention as a potential target to treat cancer. The kynurenine pathway (KP) has been proposed to be one of the key mechanisms used by the tumour cells to escape immune surveillance for proliferation and metastasis. In an inflammatory environment such as cancer, the KP is elevated, suppressing local immune cell populations and enhancing tumour growth. In this review, we collectively describe the roles of the KP in cancer and provide information on the latest research into the KP in primary liver cancer.

**Keywords:** primary liver cancer; kynurenine pathway; immune evasion; indoleamine 2,3 dioxygenase 1; tryptophan 2,3 dioxygenase 2; IDO inhibitor

#### **1. Primary Liver Cancer**

Primary liver cancer is the second leading cause of cancer mortality in men and the sixth most commonly occurring cancer worldwide, with an estimated 905,677 cases and 830,180 deaths in 2020 [1]. It is a tumour that develops in the liver and is known to be highly invasive and spread to other organs such as the lungs, bone marrow, lymph nodes, and brain [2–4]. Hepatocellular carcinoma (HCC), accounting for more than 75% of all

**Citation:** Krishnamurthy, S.; Gilot, D.; Ahn, S.B.; Lam, V.; Shin, J.-S.; Guillemin, G.J.; Heng, B. Involvement of Kynurenine Pathway in Hepatocellular Carcinoma. *Cancers* **2021**, *13*, 5180. https://doi.org/ 10.3390/cancers13205180

Academic Editor: Dan G. Duda

Received: 15 September 2021 Accepted: 13 October 2021 Published: 15 October 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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/).

primary liver cancer cases, and intrahepatic cholangiocarcinoma (ICC), which accounts for a lesser proportion, approximately 12–15% of all liver cancer cases, are the two main histological types of this malignancy [5]. HCC arises from hepatocytes in the liver and is the most common cause of death in people with a history of chronic liver disease [6] or cirrhosis [7].

The global burden of liver cancer-related mortality is increasing worldwide, with an estimation of >1 million diagnosed with this cancer annually by 2025 [8,9]. The highest HCC incidence and mortality rates are observed in Africa and East Asia, although a growing trend in incidence rates has been observed in western countries, including the USA and parts of Europe [10]. In Australia, the incidence rate of primary liver cancer has increased 5-fold from 2003 to 2011. According to the Australian Institute of Health and Welfare's burden of cancer report, this cancer is a significant health threat and a burden to the Australian community [11]. A recent study showed that the age-adjusted incidence of HCC increased from 1.38/100,000 persons in 1982 to 4.96/100,000 in 2014 [12]. Incidence of HCC is up to four times higher in men compared to women and is projected to be the fifth and sixth most common cause of cancer death in Australian men and women, respectively in 2020. The gender discrepancy in primary liver cancer incidence can be attributed to biological and behavioural risk factors [13].

Important risk factors are chronic hepatitis B virus (HBV) or hepatitis C virus (HCV) infections, liver cirrhosis, chronic alcohol consumption, metabolic-associated fatty liver disease (MAFLD), and non-alcoholic steatohepatitis (NASH) [14]. HCC predominantly develops in the setting of cirrhosis and chronic liver diseases. Cirrhosis of the liver caused by any liver disease is a major risk factor, and HCC is the primary cause of death in hepatic cirrhosis patients [15]. The most common risk factor is chronic viral hepatitis [16–18], with HBV infection accounting for approximately 50% of the HCC cases. However, HBV vaccinations have reduced the risk associated with HBV-induced HCC [19,20]. Chronic HCV patients with cirrhosis or chronic liver damage are at higher risk of developing HCC [21]. However, a significant decrease in the risk of HCC attributed to HCV infections has been observed because of effective antiviral drugs [22]. Additionally, higher prevalence of obesity- or diabetes-related MAFLD and NASH (the most severe form of MAFLD) is also driving the increase in HCC incidence rates [23–26]. Studies suggest that older age is another important risk factor that increases the risk of developing primary liver cancer [27–30]. Statistical epidemiology shows that primary liver cancer patients mostly comprise individuals above 50 years, with mean onset age increasing from 58.2 years in 1990 to 62.5 years in 2017 [31].

#### *HCC Stages and Its Prognosis*

Overall survival for HCC patients is poor, with a 5-year relative survival rate of 34% for patients diagnosed with localized tumour mass, 12% for patients with regional cancer that has spread outside the liver to surrounding tissues or lymph nodes, and 3% for patients diagnosed with distant or metastasized liver cancer [32]. The Barcelona Clinical Liver Cancer (BCLC) staging system is widely accepted and used to identify the stage of HCC based on tumour characteristics and burden, the Child–Pugh score of hepatic function, and patient performance status [33]. The median survival time for HCC patients can vary according to the stage of cancer diagnosed. Based on the BCLC staging system, these values are more than 6 years for early stage (0 and A), 26 to 30 months for intermediate stage (B), 12 to 19 months for advanced stage (C), and nearly 3 months for end-stage (D) HCC after receiving treatment (Figure 1) [8].

Surgical resection or partial hepatectomy [34], laparoscopic liver resection [35], and liver transplantation [36] are the most common treatments used for early stage HCC patients (i.e., BCLC stage A), when the tumour mass is more than 2 cm but less than 5 cm in size and is confined to the liver, with no evidence of vascular invasion. Radiofrequency ablation is the primary treatment of choice for single tumours less than 2 cm in size (BCLC

stage 0) and is also an alternative for early stage HCC patients unsuitable for surgery or liver transplantation due to the presence of multiple tumour nodules and liver dysfunction [37].

**Figure 1.** Classification of HCC and its characteristics: Based on the BCLC staging system, HCC can be classified as stages 0, A, B, C, and D. Stage A has the highest median survival time of more than 60 months while stage D has less than 4 months. Localised surgery and radiotherapy are the choice of treatments for stage 0 to B, while systemic treatment with palliative care is usually recommended for stages C to D.

> Unfortunately, patients with HCC are often asymptomatic in the early stages; hence, detecting early stages of cancer in patients remains a challenge. A combined diagnostic approach consisting of ultrasound imaging, magnetic resonance imaging, computed tomography, and detecting alpha-fetoprotein (AFP) levels in patient sera is used to diagnose cancer and predict HCC prognosis [38]. AFP is a type of glycoprotein that is produced by embryonic endoderm tissue cells and is usually in high concentrations in maternal serum during foetal development [39,40]. This concentration of AFP drops during adulthood due to the inability of mature hepatocytes to synthesize this glycoprotein [40]. Transformed cancer cells including hepatocytes can regain this ability to synthesize AFP and have therefore been used as blood-based biomarkers for HCC diagnosis [41]. However, this biomarker is not effective in detecting patients with a low concentration of AFP (AFP <20 ng/mL), such as during early stage HCC, and a portion of advanced HCC, where AFP remains low throughout disease progression [42]. A promising alternate blood biomarker is glypican-3 (GPC-3). GPC-3 is a cell-surface proteoglycan that is highly expressed in embryonic tissues and is involved in cell proliferation and survival during foetal development [43]. In adults, GPC-3 expression is only limited to lung, ovary, mesothelium, mammary glands, and kidney [44,45]. However, high levels of GPC-3 expression are observed in HCC tissues but not in healthy adult liver, and it is a commonly used immunohistochemical marker to detect the degree of HCC tumour differentiation [46,47]. Although studies have shown 83.4% sensitivity in HCC [48], the diagnostic use of GPC-3 as an HCC biomarker remains controversial due to conflicting results [49–51]. A delay of as little as three months in diagnosis can result in the cancer progressing to later stages and, more importantly, it

reduces patient survival rate. Focusing on early diagnosis is important to increase patients' survival rate rather than treatment options. [52]. Other locoregional treatment strategies for some early and intermediate HCC patients (BCLC stage B) who are not fit to undergo surgery or transplantation include trans-arterial chemoembolization (TACE) [53], local radiotherapy, or a combination approach of laparoscopy with TACE or radiotherapy is used to prevent from further cancer progression [34].

Most HCC cases are diagnosed at advanced stages (BCLC stage C and D) when the tumours are too aggressive for surgical resection and have metastasized to other organ sites. Systemic treatment, which includes molecular-targeted therapy, remains a recommended treatment for locally advanced or metastatic unresectable HCC tumours [34]. To date, the first-line drug treatments for advanced HCC patients include sorafenib [54], lenvatinib [55], and atezolizumab (anti-PDL1 antibody) in combination with bevacizumab (anti-VEGF antibody) [56]. The recent IMbrave150 trial reported that patients treated with the combination regimen of atezolizumab and bevacizumab showed improved overall survival and progression-free survival compared to sorafenib. The most common treatment-related adverse events observed with combination immunotherapy are fatigue, pain, loss of appetite, and diarrhoea [57]. On the basis of these positive findings from the trial, the Therapeutic Goods Administration (TGA)-approved regimen has now been extensively used to treat patients with unresectable HCC and was added to the Australian Pharmaceutical Benefits Scheme (PBS) program in 2020 [58]. While there has been significant improvement in treatment opportunities over the last decade, this malignancy is associated with a high recurrence rate and poor overall survival. Clinical trials evaluating the efficacy and safety of immune-therapeutic drugs such as pembrolizumab or nivolumab for advanced liver cancer treatment failed to improve overall survival of patients and significant immune-related adverse side effects were observed, resulting in failure of the clinical trials [59,60].

Although the understanding of the disease and treatment opportunities for HCC have drastically improved over the last decade, this malignancy remains a fatal disease worldwide. There is an urgent need to identify a specific set of biomarkers to (1) detect early stage HCC with high accuracy in patients and (2) to effectively allow the assessment of response to treatment to rapidly estimate whether a patient responds to treatment. Identification of novel and specific diagnostic set of biomarkers to detect patients who may be at risk and with early stage HCC, prognostic predictors that can effectively distinguish between patients with favourable or unfavourable prognosis in the same tumour stage, and more specific treatment targets are all critical. An important aspect to consider is the unique relationship between the liver and the immune system. The liver is a critical immunological frontline of the body, where complex immunological activity occurs to prevent infection in the body [61,62]. Interestingly, some biochemical pathways promote tumour tolerance by decreasing the recognition of cancer antigen, inducing immune suppression and chronic inflammation. Notably, an interesting biochemical pathway that mediates tumour tolerance is the kynurenine pathway (KP) of tryptophan (TRP) metabolism. Elevation of KP activity by tumour cells suppresses the local immune response and enhances tumour survival and invasion [63,64]. This review will examine the role of KP in HCC progression. Understanding how HCC manipulates immune-suppressive KP may lead to the identification of potential therapeutic targets for HCC.

#### **2. The KP**

TRP is one of the eight essential amino acids that are only obtainable through the diet [65]. TRP and its metabolites play a critical role in various cellular growth and maintenance processes. Up to 90% of the TRP is catabolized by the KP to produce nicotinamide adenine dinucleotide (NAD<sup>+</sup> ), an important enzyme co-factor involved in the regulation of important cellular processes (Figure 2) [66]. KP is tightly regulated under a healthy physiological state and produces various metabolites with immunesuppressive and redox activity. These metabolites include kynurenine (KYN), kynurenic acid (KYNA), 3-hydroxykynurenine (3-HK), anthranilic acid, 3-hydroxyanthranilic acid

(3-HAA), picolinic acid, and quinolinic acid (QUIN) [67]. The pathway begins with three rate-limiting enzymes, indoleamine 2,3-dioxygenase (IDO1) [68], indoleamine 2,3 dioxygenase 2 (IDO2) [69,70], and tryptophan 2,3-dioxygenase (TDO2) [71] that catabolise the substrate TRP to KYN.

**Figure 2.** A simplified diagram of the KP: majority of TRP is catabolized through the KP to synthesize the vital energy cofactor, NAD<sup>+</sup> .

γ Although the three rate-limiting enzymes catabolise the same substrate, TRP, they each have different inducers and regions of expression. In normal physiological conditions, IDO1 enzyme expression is limited to endothelial cells in the lungs and placenta, epithelial cells scattered in the female genital tract and mature dendritic cells in secondary lymphoid organs, and is known to be induced by interferon-gamma (IFN-γ) [72]. Compared to IDO1, IDO2 enzyme expression is restricted and confined to hepatocytes, bile duct, neuronal cells of the cerebral cortex, and kidneys [73]. While IDO1 and IDO2 share 43% gene similarity, IDO1 remains the dominant enzyme [69]. Interestingly, the activity of IDO2 elevates when the IDO1 gene is deleted [74]. The third rate-limiting enzyme, TDO2, is primarily expressed in liver, and is the major enzyme to regulate systemic TRP levels in the liver [75,76]. TDO2 enzyme expression is known to be induced partly by glucocorticoids and its substrate TRP [77]. Though these rate-limiting enzymes are cytosolic, their enzymatic activity induces TRP metabolism and accumulation of KP metabolites in the extracellular space, which is facilitated by specific amino-acid transporters [78]. In an inflammatory environment such as cancer, KP is highly activated, resulting in depletion of local TRP in the tumour micro-environment. This process facilitates tumour cells to evade immune detection by reducing the proliferation of effector T lymphocytes and favouring the differentiation of regulatory T (Tregs) cells [79].

#### *Involvement of the KP in Cancer*

After the discovery that placental IDO1 was the key enzyme mediating immune suppression in maternal–foetal tolerance in 1998, the research focus was expanded to examine whether the KP was involved in immune evasion and cancer [80,81]. Indeed, the KP is frequently dysregulated in cancer and suppresses tumour surveillance in two different mechanisms. The first mechanism involves the overexpression of the rate-limiting enzymes IDO1 and TDO2 to deplete TRP within the tumour microenvironment. TRP is one of the amino acids required for the survival and proliferation of immune T-cells such as T helper (Th) and cytotoxic T-cells (Tc). Therefore, immune surveillance will be strongly suppressed in a TRP-deprived tumour microenvironment driven by an overactive IDO1/TDO2 tumour [82]. A study by Uyttenhove et al. confirmed overexpression of IDO1 in various human cancer tissues and cell lines, suggesting that was involved in protecting tumours from immune detection [83]. The overexpression of IDO1 in tumours has been suggested to be induced by the IFN-γ generated by tumour-infiltrating T-cells as an adaptive resistance mechanism [84]. Syngeneic animal studies showed that treatment of the IDO1 inhibitor 1-methyltryptophan (1-MT) limited the growth of IDO1-overexpressed tumours [83,85]. A subsequent breast cancer animal model study by Muller et al. demonstrated that combined treatment with 1-MT and cancer chemotherapeutic drug paclitaxel slowed down the tumour growth progression by 30% [86]. Importantly, they observed that the efficacy of this combination therapy was highly dependent on the presence of T-cells, and the inhibition of IDO1 could potentiate the efficacy of chemotherapy.

Apart from IDO1, overexpression of TDO2 in tumour cells has been shown to facilitate immune escape. TDO2 mRNA expression was detected in different types of tumours including hepatocarcinoma [87], glioblastoma [88], breast cancer [89], and colorectal cancer [90,91]. These studies also demonstrated that TDO2 was responsible for the depletion of TRP in IDO-negative tumours to evade immune surveillance [63,88,92]. This notion was supported by an animal model study by Pilotte et al., who showed that treatment using TDO2 inhibitor in an animal model reversed the TDO2-mediated immune evasion mechanism and prevented the growth of TDO2-overexpressing tumours [92]. Consequently, this led to further studies exploring new TDO2 inhibitors for use in the treatment of TDO2-overexpressing cancer [93–95].

Though the role of the IDO2 enzyme in cancer remains less understood, studies have shown that IDO2 expression is upregulated in certain malignancies such as colon cancer, gastric and renal cancer [96], pancreatic cancer [97], non-small cell lung cancer [98], and may have roles in tumour immune escape, facilitating cancer cell proliferation and metastasis. Sorensen et al. described the immunogenic role of IDO2 by demonstrating the presence of spontaneous T<sup>c</sup> reactivity against IDO2 in healthy and cancer patient blood samples, and reported that IDO2 supported Tregs cells generation that was induced by human dendritic cells [99].

The second mechanism of KP-mediated tumour evasion involves the bioactive KP metabolites KYN, 3HK, 3-HAA, and QUIN. Studies have shown that these metabolites can promote tumour proliferation and modulate the immune cell population. KYN, the first metabolite of KP, can function as an endogenous ligand to activate the aryl hydrocarbon receptor (AhR) in an autocrine/paracrine fashion, and emerging evidence points toward the tumour-promoting role of KYN-mediated activation of the AhR [100,101]. AhR is a ligand-activated transcription factor of the basic helix–loop–helix (bHLH) Per–Arnt– Sim (PAS) family [102]. It is expressed in many immune cells and plays a vital role in regulating various immune functions in a wide range of physical and pathological processes [103–105]. Activation of AhR may facilitate cancer cell proliferation, tissue invasion, metastasis, and angiogenesis [106]. The KYN-AhR signalling pathway can suppress the differentiation and activity of immune cells, resulting in an impaired immune response against tumours, leading to tumour immune tolerance [107]. Various studies have demonstrated the importance of KYN-AhR activation in IDO1- or TDO2-expressing tumour cells and its role in enhancing cancer cell survival and motility. These studies

suggested that TDO2-expressing cancer cells escape immune surveillance by activating AhR in various immune cells including dendritic cells, macrophages, natural killer cells, innate lymphoid cells, T<sup>c</sup> cells, and Tregs cells [108,109]. Opitz et al., found that murine tumours in AhR-proficient mice expressing high AhR and TDO2 expression levels had an enhanced tumour growth rate by suppressing the infiltration of antitumour immune cells, increasing levels of inflammatory cytokines. Furthermore, the study suggested that the TDO2-Kyn-AhR signalling pathway might also be involved in other malignancies, including sarcoma, bladder cancer, cervix cancer, colorectal cancer, lung, and ovarian cancer [88]. Moreover, Ulrike et al. revealed that IDO1 enzyme expression was induced by inflammatory cytokines such as Interleukin 6 (IL-6), and could activate an autocrinepositive inflammatory feedback loop (IDO-AhR-IL-6-STAT3 signalling pathway) that could promote tumour growth and survival [110].

In addition to KYN, kynurenic acid (KYNA) is also an endogenous AhR ligand [111]. In the presence of IL-1β, KYNA binds to AhR and induces production of IL-6, which may also contribute to the IDO-AhR-IL-6-STAT3 autocrine-positive inflammatory feedback loop mentioned earlier. Interestingly, the production of KYNA may not be limited to just via KP but rather through an alternate TRP metabolism mediated by Interleukin-4-induced gene 1 (IL4I1) in a cancer setting. Sadik et al. revealed that IL4I1 was elevated in cancers such as melanoma. An IL4I1-driven AhR activity though KYNA increases tumour cell motility and T-cell proliferation [112]. Given that the activity of IL4I1 is independent of the KP and can limit antitumor immune cell response [113], inhibiting the formation of KYNA metabolite either via the KP or through IL4I1 gene reaction may be necessary to block the activation of AhR in cancer.

The KP metabolites downstream of KYN, including 3-HK, 3-HAA, and QUIN have been shown to inhibit T-cell proliferation and activation. A study by Fallarino et al. showed that 3-3-HAA and QUIN could induce selective apoptosis in Th1 cells and thymocytes of effector T-cell population in vitro by the activation of caspase-8 activity and the release of cytochrome c from mitochondria [114]. The 3-HAA also significantly inhibits CD8<sup>+</sup> T-cell proliferation stimulated through cytokines by driving the T-cells to a proliferative arrest and directly inhibiting the phosphorylation of phosphoinositide-dependent kinase 1 and preventing the activation of nuclear factors after T-cell receptor stimulation [115]. A study by Favre et al. showed that 3HAA also disturbed the balance between T<sup>h</sup> and Treg cell populations, driving them towards an immunosuppressive Treg pathway in vitro [116]. Furthermore, a later study by Zaher et al. confirmed that 3HK and 3HAA suppressed CD4<sup>+</sup> T-cell proliferation along with significant T-cell death [117].

#### **3. Involvement of the KP in Chronic Liver Disease and HCC**

The role of KP in liver diseases has been gaining interest in the recent years. A number of studies have measured high KP activity in chronic liver diseases such as primary biliary cirrhosis, HCV-associated chronic hepatitis, and liver cirrhosis [118,119]. Claria et al. [120] reported that KP activity was elevated in patients with acute decompensation and acuteon-chronic liver failure, and was associated with pathogenesis and mortality in cirrhotic patients. The study concluded that elevated KP activity may be used as an independent prognostic predictor of poor clinical outcomes in cirrhotic patients. In contrast, elevated IDO1 activity during early stages of the HBV infection in hepatocytes was reported to significantly reduce viral replication and enhance the protective immune response [121].

Although the liver is a site of robust immunological activity, liver cancer cells can remain undetected and proliferate. This suggests that these cancer cells can evade local immune surveillance, possibly by using the KP, as observed in various malignancies. Although the research on KP and HCC is limited, the activity of the three upstream enzymes of the pathway, including IDO1, TDO2, and KMO enzymes, has been extensively studied in HCC cells and tissue specimens. These study findings revealed that IDO1, TDO2, and KMO enzyme activity was upregulated in HCC (Table 1).


#### **Table 1.** Summary of all KP research carried out on HCC.

\* HCC patients with chronic hepatitis B (HBV) or hepatitis C (HCV) virus infection. \*\* Patients with chronic hepatitis C virus infection without HCC. RT-PCR: reverse transcription-polymerase chain reaction, WB: Western blot, IHC: immunohistochemistry, IF: immunofluorescence, HPLC: high-performance liquid chromatography, shRNA: short hairpin or small hairpin RNA, siRNA: small interfering RNA.

#### *3.1. IDO1*

The immunological and prognostic roles of IDO1 in HCC were first investigated by Ishio et al. in 2004 [122]. The results showed that IDO1 mRNA expression was strongly induced in tumour-infiltrating cells of the HCC tumour, which might facilitate an antitumour immune reaction and the expression of IDO in tissue specimens of HCC patients significantly correlated with better recurrence-free survival rates. A later study by Ke Pan et al. observed elevated IDO1 enzyme mRNA and protein expressions in liver tumour and its adjacent normal tissues compared to distant non-involved normal tissues, suggesting that IDO1 overexpression was confined to the tumour microenvironment [123]. A potential explanation for the confined IDO1 expression could be due to the presence of inflammatory cytokine(s) in the tumour microenvironment that activate IDO1 activity. Indeed, a later study by Li et al. demonstrated that IDO1 enzyme expression was observed only in IFNγ-stimulated HCC cells through the IFN-γ-JAK2-STAT1-signalling pathway. Moreover, high IDO1 expression in HCC positively correlated with abundance of CD8+ T-cells, thus reflecting an antitumour immune response and suggesting that IDO1 could be used as a favourable prognostic indicator for HCC patients [124]. Lastly, Brown et al. suggested that IDO1 enzyme inhibitors in combination with immune checkpoint inhibitors could be a novel treatment approach for liver cancer treatment [125].

#### *3.2. TDO2*

A recent study conducted by Hoffman et al., showed that the majority of the tumour cells in HCC tissues expressed TDO2 in HCC [126]. This study demonstrated the immune-regulatory role of the TDO2 enzyme in HCC tumour cells, and suggested that the TDO2 enzyme was a promising immunotherapy treatment target for HCC. Another study by Li et al. characterized the overexpression of TDO2 enzyme in HCC cancer cells and suggested that it might play a vital role in promoting HCC cancer cell growth, migration, and invasion in vitro and in vivo [127]. Additionally, TDO2 expression was correlated with the development of the tumour, such as size, tumour differentiation, and vascular invasion. Based on these strong correlation data, the authors suggested that TDO2 expression could be used as an effective biomarker to predict overall or disease-free survival of HCC patients. Activation of AhR is associated with the loss of cell contact inhibition and changes to the extracellular matrix, and extensive studies have demonstrated that this activation induces epithelial to mesenchymal transition (EMT) in various cancers [130–132]. Overexpression of AhR in HCC has been shown to be associated with its tumour proliferation and invasion [133,134]. A recent study by Lei Li et al. showed that upregulated expression of the TDO2 enzyme promotes the migration and invasion capabilities of HCC cells by the KYN-AhR-mediated induction of epithelial to mesenchymal transition, a process that is vital for cancer metastasis [87].

#### *3.3. KYN Levels in Patient Sera*

A recent retrospective study on a cohort of HCC patients with chronic HCV infection revealed that KYN levels were elevated in HCV-mediated HCC patient sera in comparison to healthy controls (non-HCC patients). Bekki et al. observed that KYN production gradually increased when chronic HCV progressed to HCC, and suggested the potential of using serum KYN levels as a biomarker for predicting survival and prognosis in early stage HCV-mediated HCC patients [128].

#### *3.4. KMO*

Kynurenine 3-monooxygenase (KMO) is the immediate KP enzyme after the ratelimiting step, and it is widely distributed in the peripheral tissues of the liver and kidney, astrocytes and microglial cells situated in the brain, central nervous system [135,136], and phagocytes, including macrophages and monocytes [137]. KMO localizes to the outer membrane of mitochondria and catabolizes KYN to 3-HK. The role of KMO enzyme expression in cancer has rarely been studied in comparison to IDO and TDO2 enzymes. Liu et al. identified the oncogenic role of KMO in triple-negative breast cancer progression [138]. Moreover, high surface expression of KMO was detected in cytosol and on the cell membranes of breast cancer tissue specimens, indicating its potential as a treatment target for TNBC [139]. A recent study investigated the correlation between upregulated KMO activity and poor clinical outcomes in colorectal cancer (CRC) patients and demonstrated that KMO inhibition suppressed CRC cell proliferation in vitro [140]. On analysing KMO enzyme expression in 120 matched HCC tissue samples, Jin et al. showed that the expression of the KMO enzyme is significantly elevated in HCC tumour tissue compared to adjacent normal liver tissue. High KMO expression correlated with poor patient outcomes, which indicates that the KMO enzyme may be a significant prognostic marker in HCC patients [129]. Results from the in vitro experiment comparing KMO enzyme levels in human normal liver cells and HCC cell lines showed that KMO enzyme was upregulated in HCC cells and might play a role in promoting tumour proliferation, metastasis, and invasion. The study also demonstrated that KMO knockdown in HCC cell lines by small interfering RNA (siRNA) transfection decreased cancer cell proliferation, thus suggesting that KMO could be a novel target for HCC treatment.

#### *3.5. Clinical Trials: IDO1 Inhibitors as HCC Treatment*

IDO1 inhibitors are small molecule drugs that competitively block the activity of the IDO1 enzyme without inhibiting IDO2 or TDO2 [141]; several of these drugs are in clinical development. The safety and efficacy of many IDO1 inhibitors, including Indoximod, Epacadostat, Navoximod, BMS-986205, and others, have been tested in combination with other immunotherapy drugs such as pembrolizumab and nivolumab for the treatment of various metastatic cancers. Currently, two small molecule IDO1 inhibitors, BMS-986205/NCT03695250 and INCB024360 (Epacadostat)/NCT02178722, are in phase I/II clinical trial to evaluate their safety and efficacy in HCC patients [142,143]. The clinical trial NCT03695250 is a single-group assignment that examines the safety, tolerability, and efficacy of BMS-986205 with nivolumab in unresectable/metastatic HCC. It is still active but not recruiting patients; hence, the results have not been published yet. The expected treatment-related adverse events of BMS-986205 would be at grade 1–2 such as fatigue and nausea, as reported in the other trials examining the efficacy of BMS-986205 in cancer patients. Clinical trial NCT02178722 evaluated the safety, tolerability, and efficacy of Epacadostat in combination with pembrolizumab. This trial concluded that the combination regime has an acceptable safety profile in patients with advanced cancers, achieving an objective response rate in 12 of 22 cancer patients [144,145]. Treatment-related adverse events observed in 84% of the patients enrolled were of grade 1–2. The most common events were fatigue, rash, arthralgia pruritus, and nausea. This result supports additional phase 3 studies in other malignancies but not in HCC.

#### **4. Conclusions**

HCC is one of the few malignancies for which the risk factors have been wellestablished. Although patients with early stage HCC have the best median survival time and can usually be cured by resection, liver transplant, or ablation, they are often asymptomatic. Hence, most patients present with late-stage HCC and have a poor prognosis. The approved first-line treatment of late-stage HCC is multikinase inhibitors such as sorafenib, which confers a slightly longer survival time. However, this treatment is associated with substantial side effects that have a negative impact on quality of life. This therefore changes the treatment focus by combining current antitumoral drugs with immunotherapy, and this approach has significantly benefited HCC patients. A recently concluded trial examining combination therapy of atezolizumab with bevacizumab showed a significant improvement in overall survival and progression-free survival as compared to sorafenib. Since this study, it has been adopted as the first-line treatment for late-stage HCC. Considering the strong evidence of its ability to mediate immune suppression, the KP might be an alternative immunotherapy target and play a role in the progression of liver cancer, as summarized in Figure 3.

This notion is supported by clinical studies that showed an elevated KP enzyme profile in HCC cells and tumour tissue specimens, with elevated expressions associated with disease aggressiveness. Although current IDO1 inhibitor clinical trials are still in phase I/II evaluation, it is possible to suggest that the use of KP inhibitors in combination regimens may improve the survival mark of early and advanced HCC.

**Figure 3.** The KP-mediated immune tolerance and cancer invasion: KP promotes immune tolerance by two different mechanisms. Firstly, elevated IDO1/TDO2 enzyme activity in either tumour or immune cell depletes TRP concentration in its local tumour environment. A TRP-stripped environment induces cell arrest in T-cells while inducing differentiation and proliferation of Treg cells. Secondly, downstream KP metabolites induce cell arrest in T-cells and Treg proliferation by either interaction with AhR or by direct interaction with immune cells themselves. In addition to KP-mediated immune tolerance, elevated KP promotes cancer cell motility and proliferation by either overproduction of NAD<sup>+</sup> for cellular repair or byactivation of AhR.

> — — **Author Contributions:** Conceptualization, S.K., G.J.G. and B.H.; writing—original draft preparation, S.K., G.J.G. and B.H.; Writing—review and editing, D.G., S.B.A., V.L. and J.-S.S.; funding acquisition, G.J.G. and B.H. All authors have read and agreed to the published version of the manuscript.

> — **Funding:** S.K. is supported by International Macquarie University Research Excellence Scholarship— Master of Research scholarship; D.G. is supported by Aviesan Plan Cancer and Cancer pole Grand Ouest; S.B.A. is supported by Cancer Council NSW funding RG19-04; G.J.G. is supported by the National Health and Medical Research Council (NHMRC) APP1176660, PANDIS.org and Macquarie University; B.H. is supported by NHMRC.

**Acknowledgments:** The figures were created with BioRender.com.

**Conflicts of Interest:** The authors declare that they have no competing interests.

#### **Abbreviations**


#### **References**


## *Review* **Tumor Microenvironment: Key Players in Triple Negative Breast Cancer Immunomodulation**

**Hongmei Zheng 1,2,\* , Sumit Siddharth <sup>2</sup> , Sheetal Parida <sup>2</sup> , Xinhong Wu 1,\* and Dipali Sharma <sup>2</sup>**


**Simple Summary:** The tumor microenvironment (TME) is a complicated network composed of various cells, signaling molecules, and extra cellular matrix. TME plays a crucial role in triple negative breast cancer (TNBC) immunomodulation and tumor progression, paradoxically, acting as an immunosuppressive as well as immunoreactive factor. Research regarding tumor immune microenvironment has contributed to a better understanding of TNBC subtype classification. Shall we treat patients precisely according to specific subtype classification? Moving beyond traditional chemotherapy, multiple clinical trials have recently implied the potential benefits of immunotherapy combined with chemotherapy. In this review, we aimed to elucidate the paradoxical role of TME in TNBC immunomodulation, summarize the subtype classification methods for TNBC, and explore the synergistic mechanism of chemotherapy plus immunotherapy. Our study may provide a new direction for the development of combined treatment strategies for TNBC.

**Abstract:** Triple negative breast cancer (TNBC) is a heterogeneous disease and is highly related to immunomodulation. As we know, the most effective approach to treat TNBC so far is still chemotherapy. Chemotherapy can induce immunogenic cell death, release of damage-associated molecular patterns (DAMPs), and tumor microenvironment (TME) remodeling; therefore, it will be interesting to investigate the relationship between chemotherapy-induced TME changes and TNBC immunomodulation. In this review, we focus on the immunosuppressive and immunoreactive role of TME in TNBC immunomodulation and the contribution of TME constituents to TNBC subtype classification. Further, we also discuss the role of chemotherapy-induced TME remodeling in modulating TNBC immune response and tumor progression with emphasis on DAMPs-associated molecules including high mobility group box1 (HMGB1), exosomes, and sphingosine-1-phosphate receptor 1 (S1PR1), which may provide us with new clues to explore effective combined treatment options for TNBC.

**Keywords:** triple negative breast cancer; tumor microenvironment; immunomodulation

#### **1. Introduction**

Triple negative breast cancer (TNBC), characterized by the absence of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) expression, comprises 10–20% of all breast cancers [1]. Owing to the lack of ER/PR/Her2 protein expression/amplification, TNBCs do not respond to existing endocrine and Her2 targeted therapies and exhibit poor prognosis [2]. It has been proposed that TNBCs with a higher involvement of immune cells termed as 'hot tumors' have better prognosis and a greater response to immunotherapy while TNBCs with a lower involvement of immune cells termed as 'cold tumors' are marked with poor prognosis and poor response to immunotherapy [3]. From this point of view, TNBC patients have been further segregated

**Citation:** Zheng, H.; Siddharth, S.; Parida, S.; Wu, X.; Sharma, D. Tumor Microenvironment: Key Players in Triple Negative Breast Cancer Immunomodulation. *Cancers* **2021**, *13*, 3357. https://doi.org/10.3390/ cancers13133357

Academic Editor: Charles Theillet

Received: 16 June 2021 Accepted: 1 July 2021 Published: 4 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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/).

into different subgroups [4–8]. The tumor microenvironment (TME) is an ensemble of endothelial cells, cells of the immune system, adipocytes, and fibroblasts, in addition to the soluble factors released from all the cellular components (including cancer cells) [9,10]. TME can be classified from different perspectives such as host and non-host origin, cellular origin and constituents [9,11–13]. TME presents a complex network that plays a crucial role in TNBC immunomodulation and tumor progression.

Cancer initiation and development is not just a biological process triggered by cancer cells in isolation; in fact, it has to be evaluated along with the complicated TME with an emphasis on the interaction between cancer cells and their surrounding extra-cellular matrix. Indeed, considering alterations in microenvironment as active players during cancer progression brings another dimension of complexity [14]. During TNBC progression, tumor immune microenvironment remodeling including the change of the ratio of immune cells and release of multiple immune inhibitory and reactive cytokines is a critical feature [15,16]. Based on the constituents of TME, TNBCs have been stratified into 'tumor immune microenvironment (TIME) subtypes' aiding in predicting outcomes and proposing potential treatments guided by the distinct phenotypes of TNBC [16,17]. Chemotherapy, the foremost treatment for TNBC, could induce immunogenic cell death (ICD) and promote the release of damage-associated molecular patterns (DAMPs) [18] including high mobility group box1 (HMGB1), exosomes and sphingosine-1-phosphate receptor 1 (S1PR1) by damaged or activated cells via the activation of TLR4 signal pathway [19] and stimulate the release of various immune molecules such as TGF-β, IK12p7, and IFN-γ [20].

In this review, we focus on immune TME and summarize its immunosuppressive and immunoreactive roles, discuss constituent immune cells involved in TNBC immunomodulation, and the contribution of TIME in stratification of TNBC. Further, we discuss the role of chemotherapy-induced TME changes in modulating TNBC immune response and tumor progression, with a focus on HMGB1, exosomes, and sphingosine-1-phosphate (S1P)/sphingosine kinase 1 (SPHK1)/S1PR1, an axis whose therapeutic modulation may result in neoteric combination therapy for TNBC patients.

#### **2. Two Roles of TME in TNBC Immunomodulation**

According to the contribution to immune response, the tumor microenvironment (TME) can be classified as immunosuppressive and immunoreactive. Tumor infiltrating lymphocytes (TILs), the major cell types in the microenvironment, are heterogeneous and mainly composed of lymphocytes in tumor nests and tumor stroma. TILs can be classified into several different subtypes, mainly CD3<sup>+</sup> T cells and CD20<sup>+</sup> B cells in solid tumors, though CD20<sup>+</sup> B cell infiltration is relatively less. CD3<sup>+</sup> T cells include CD8<sup>+</sup> cytotoxic T lymphocytes (CD8<sup>+</sup> TILs), CD4<sup>+</sup> helping T lymphocytes, and Foxp3<sup>+</sup> regulatory T lymphocytes (Foxp3<sup>+</sup> Tregs) [21,22]. Different subtypes of TILs take part in immunomodulation with distinct mechanisms and play various roles in breast cancer immunomodulation [22]. Figure 1 pictorially represents immunosuppressive and immunoreactive TMEs (Figure 1).

#### *2.1. Immunosuppressive TME in TNBC*

#### 2.1.1. PD-1/PD-L1 Axis

Programmed death-ligand 1 (PD-L1) and programmed cell death protein-1 (PD-1) are important negative co-stimulating signaling molecules in immunoglobulin superfamily (IgSF) and play an important role in host immunomodulation [23]. PD-L1 is expressed in many solid tumors including breast cancer and is a negative prognosis indicator [24,25]. PD-1 is expressed in TILs [26]. Theoretically, PD-L1 expression on tumor cells combined with PD-1 expression on TILs play a negative role in immunomodulation, which inhibits the activation of TILs, causing the tumor cell to survive through immune escape.

**Figure 1.** Immunosuppressive and immunoreactive TME. Immunosuppressive tumor microenvironment (TME) is mainly constituted of M2 macrophages, forkhead box P3<sup>+</sup> (Foxp3<sup>+</sup> ) regulatory T lymphocytes (Tregs), myeloid-derived suppressor cells (MDSCs), and PD-1/PD-L1 axis. Immunoreactive TME is mainly constituted of CD8<sup>+</sup> T cells, natural killer (NK) cells, and M1 macrophages. PD-1/PD-L1 axis becomes immunoreactive in response to anti-PD1 or anti-PD-L1 monoclonal antibody (aPD-1/PD-L1 mAb) owing to the activation of CD8<sup>+</sup> T cells. (Foxp3, forkhead box P3; Tregs, regulatory T lymphocytes; MDSC, myeloid-derived suppressor cell; NK, natural killer; PD-1, programmed cell death protein-1; PD-L1, programmed death-ligand 1; aPD-1 mAb, anti-PD-1 monoclonal antibody; ECM, extra cellular matrix; TME, tumor microenvironment).

κ α The TME involves immune suppressing factors to support the progression of tumors which have escaped host immune surveillance [27–31]. Various immune check-point inhibitors have been developed that have shown efficacy in TNBC patients [32,33]. Clinical studies have shown a paradoxical role of PD-L1 regarding its prognostic value in patients with TNBC owing to the heterogeneity of PD-L1 expression in different tumor sites, non-standard detection methods, and distinct antibodies [31,34–41]. In the impassion 130 clinical trial, compared to TNBC patients receiving nab-paclitaxel plus placebo, a better median overall survival (OS) was observed in patients receiving atezolizumab (PD-L1 inhibitor) combined with nab-paclitaxel and most benefit was observed in PD-L1 positive subgroup [42]. However, in a phase 1b clinical trial (ClinicalTrials.gov Identifier: NCT01848834) which evaluated the safety and effectiveness of PD-1 inhibitor (pembrolizumab) in PD-L1 positive TNBC patients, the overall response rate was only 18.5% and the expression level of PD-L1 was not significantly related to the clinical response [43]. These disparate results might be related to multiple TME-related factors that can modulate the therapeutic effects of PD-1/PD-L1 inhibitors in TNBC. Preclinical studies have shown that PD-L1 expression is modulated by multiple signaling pathways including microRNA-200/ZEB1 axis, WNT, loss of PTEN, PI3K, and MUC1-C/MYC/NF-κB axis [31,44–46]. Voorwerk and colleagues reported that doxorubicin and cisplatin treatment caused an upregulation of inflammation-related genes JAK-STAT and TNF-α signaling, immunerelated genes associated with PD-1/PD-L1, and T cell cytotoxicity pathways. Short-term and low-dose doxorubicin and cisplatin may create an immunoreactive TME and increase the response to PD-1 inhibitor in TNBC [47]. In conclusion, specifically designed clinical trials are needed to interrogate the involvement of various TME-related factors in order to enhance the efficacy of PD-1/PD-L1 inhibitors in TNBC.

#### 2.1.2. Foxp3<sup>+</sup> Tregs

In TME, different classes of TILs exist, which have shown great prognostic value in patients with TNBC. Regulatory T lymphocytes (Tregs) are a lineage of lymphocytes involved in immunosuppression that are characterized by the expression of the forkhead box P3 (Foxp3) transcription factor [48,49]. Foxp3<sup>+</sup> Tregs are the major constituent of the TILs in claudin-low TNBC tumors and it has been speculated that the recruitment of Foxp3<sup>+</sup> Tregs to the TME inhibits an effective anti-tumor immune response of checkpoint inhibitors [50]. Jamiyan and colleagues detected the expression of stromal Foxp3<sup>+</sup> Tregs in 107 TNBC samples using IHC and found that a low stromal Foxp3<sup>+</sup> Tregs level was significantly associated with favorable recurrence free survival (RFS) and OS [51]. In contrast, high Foxp3<sup>+</sup> TILs expression in 43 TNBC tissues by IHC and Foxp3+/CD25<sup>+</sup> TILs were positively correlated with better OS [52]. High densities of intra-tumoral Tregs and CD20<sup>+</sup> B cells represented a good prognostic panel in TNBCs [53]. However, mRNA expression of Foxp3 by qRT-PCR in 826 breast tumor tissue samples including 84 TNBC samples, was not significantly related to disease free survival (DFS), while none of the markers studied including CD3, CD8, and Foxp3 were of prognostic value for OS [54]. This phenomenon is somewhat explained by a study showing that activation of tumor antigen-specific Tregs in the bone marrow caused the accumulation of Tregs in breast cancer tissue leading to both antitumor immunity and local immune suppression in breast cancer [55]. The mechanisms underlying pro-tumor role of Foxp3<sup>+</sup> Tregs included (i) down-regulation of Notch pathway [56]; (ii) direct suppression via cell-cell contact and indirect suppression via secretion of antiinflammatory mediators such as interleukins (IL-4, IL-5 and IL-10) [57–59]; (iii) decreased secretion of cytokine IFN-γ and IL-17 and activation of STAT1/STAT3 [59]. The prognostic significance of Tregs in TNBCs, therefore, remains controversial and warrants more careful investigations.

#### 2.1.3. M2 Macrophages

M2 macrophages, the main tumor-associated macrophages, (TAMs), can promote breast cancer initiation, angiogenesis, invasion, and metastasis by generating an immunosuppressive TME via releasing cytokines, chemokines, and growth factors [60]. TAMs expressing CD163<sup>+</sup> (marker of M2 macrophages) positively correlate with tumor associated fibroblasts and epithelial-mesenchymal transition, which in turn are associated with aggressive behaviors and short DFS in 278 patients with histologically confirmed TNBC [61,62]. Another clinical study showed that high CD68<sup>+</sup> (marker of M2 macrophage) TAMs expression associates with poor distant metastasis free survival (DMFS), DFS and OS in 287 patients with TNBC [63]. Mechanistically, in vivo and in vitro studies showed that the presence of CD11b+F4/80+CD206<sup>+</sup> TAMs significantly associate with proliferating tumor cells in a TNBC mouse model. RNA sequencing analysis revealed that TAMs promote MAPK pathway activation in 4T1 cells [64]. Reactive oxygen species (ROS)-induced macrophages produce an immunosuppressive subtype (M2) and increase the expression of PD-L1 via activating NF-κB signaling, as well as release immunosuppressive chemokines such as interleukin-10 (IL-10), IL-17, IL-4, IL-1β, insulin-like growth factor-binding protein 3 (IGFBP-3), and chemokine (C-X-C motif) ligand 1 (CXCL1) [65]. The JAK2/STAT3 signaling pathway can up-regulate the expression of PD-L1 in CD169<sup>+</sup> macrophages, but cannot up-regulate the expression of PD-L1 in breast cancer cells, thus avoiding immune surveillance [66]. Metastasis- and inflammation-associated microenvironmental factor S100A4 activates the basal-like subtype of breast cancer cells to trigger monocyte-to-macrophage (M2) differentiation and polarization, and elevates secretion of pro-inflammatory cytokines

such as IL-8, IL-6, CXCL10, CCL2 and CCL5 [67]. Further, macrophage colony-stimulating factor (M-CSF), the main stimulator of macrophage migration, caused aggregation of M2 macrophages through an increased elongation of pseudopodia [68]. Inhibitors of differentiation (ID) 4 significantly associates with M2 macrophage marker CD68 protein expression in a series of TNBC tissues. ID4 activates an angiogenic procedure at the molecular level in the macrophages through paracrine signaling including the decrease of constituents of the anti-angiogenic miR-15b/107 group and an increase of angiogenesis-associated mRNAs [69,70]. GM-CSF BRCA1-IRIS overexpressing TNBC cells secrete high quantities of GM-CSF in an NF-κB and a HIF-1α-dependent manner to induce macrophages to IRIS overexpressing cells and polarize them to pro-tumor TAMs (M2). GM-CSF triggers TGF-β1 expression on TAMs through activating STAT5, NF-κB and/or ERK signaling [71].

#### 2.1.4. MDSCs

Myeloid-derived suppressor cells (MDSCs) are an important part of immunosuppressive network [72]. CD33<sup>+</sup> MDSCs are a risk factor for progressive disease (PD) plus stable disease (SD) in breast cancer tissues prior to neoadjuvant chemotherapy [73]. Higher expression of MDSCs has been noted in TNBCs in comparison to non-TNBCs with their recruitment to the primary cancer and metastasis occurring via ∆Np63-dependent activation of the chemokines CCL22 and CXCL2 [74]. Glycolysis restriction reduces MDSCs through inhibiting cancer granulocyte G-CSF and GM-CSF expression [75] while hypoxia enhances the expansion of MDSCs and upregulates the expression of PD-L1 in the hypoxic TME of 4T1 tumor-bearing mice [76]. Studies have shown that the monoclonal antibody that neutralizes IL-8 (HuMax-IL8) and the traditional Chinese medicine Prim-O-glucosylcimifugin (POG) can inhibit the recruitment, proliferation, metabolism and immunosuppressive ability of MDSCs [77,78]. The 4T1 TNBC model effectively exhibits induction of immunosuppressive MDSCs accumulation by releasing inflammatory cytokines that produce permissive pro-metastatic TME [79]. Monocytic MDSCs (M-MDSC) and granulocytic MDSCs (G-MDSC) are two types of MDSCs in circulating peripheral blood. G-MDSC levels increase sharply and M-MDSCs decrease significantly after doxorubicin and cyclophosphamide treatment [80]. Investigations have shown that CCL5 is a key modulator of Rb1 activation and is associated with the immunosuppressive activity of MDSCs, especially the G-MDSC subset [81,82].

#### *2.2. Immunoreactive TME in TNBC*

#### 2.2.1. NK Cells

Natural killer (NK) cells, a type of cytotoxic lymphocytes, are crucial constituents of the innate immune system whose function in enhancing the anti-tumor immunity in TNBC has been studied extensively. NK cells are abundant in early cancer tissue in human solid tumors; however, they dwindle in metastatic human cancers [83]. These findings show that NK cells play a key role in immune surveillance, but once tumorigenesis occurs, TME is suppressive for NK cells. Evasion of active immune suppression in the TME is an important consideration for enhancing the anti-tumor ability of tumor-infiltrating NK cells. Zhang and colleagues detected the expression of NKp46, Foxp3, CD8, CD163 or Gas6 in 278 TNBC tissues using IHC with an aim to develop a prognostic risk model for TNBC. Multivariate analysis showed that TNM stage, Foxp3 positive lymphocytes along with prognostic risk scores can be used as independent indicators of OS and DFS in TNBC [84]. Tumor-derived IL-18 upregulates PD-1 expression on CD56dimCD16dim/<sup>−</sup> NK cells and relates to the bad/ prognosis of TNBC [85]. McArdle and colleagues examined the abundance of NK cells, MDSCs, monocyte subsets and Foxp3<sup>+</sup> Tregs in the peripheral blood of 85 breast cancer patients and they found that chemotherapy had no effect on the percentage of these immune cells, but peripheral blood cells could distinguish TNBC patients that are at high risk of relapse after chemotherapy [86]. Tissue-infiltrating NK cells in solid tumors appear to have a less robust activity compared with circulating NK cells [87–90]. NK cells isolated from either breast cancer patients or healthy donors show

high cytotoxicity against patient-derived tumor cells in vitro and prevent tumor initiation and growth in immunocompromised mice in vivo [91]. Expanded cord blood-NK cells show cytotoxicity towards primary breast tumor cells derived from TNBC and estrogen receptor-positive/progesterone receptor-positive breast cancer [92]. Baseline circulating tumor cells (CTCs) status is positively associated with peripheral NK cells among those receiving first-line treatment in 75 patients with TNBC. Baseline CTCs combined with peripheral NK enumeration (CTC-NK) can predict PFS of TNBC patients more precisely [93]. NK cells are the major effectors of antibody (Ab)-dependent cell-mediated cytotoxicity (ADCC) and thus play an important role in Ab-based therapies. In vivo and in vitro studies revealed that tissue factor (TF)-targeting antibody-like immunoconjugate (called L-ICON)- CAR-NK cells have direct killing effects against TNBC cells and also mediate L-ICON ADCC to acquire a stronger effect [94]. Avelumab, a human IgG anti-PD-L1 mAb, triggers ADCC against a panel of TNBC cells and enhances NK-cell mediated cytotoxicity, which is independent of the blockade of the PD-1/PD-L1 pathway but is involved with IL-2 and IL-15 [95]. CD85j, an inhibitory receptor which can recognize both classical and non-classical HLA-I molecules, is highly expressed in TNBC, and can impair the function of cetuximab through NK-cell functional deficiency even when stimulatory cytokines IL-2 or IL-15 are abundantly present [96]. More interestingly, NK cell infiltration and recruitment can be mediated by a bispecific Ab (MesobsFab) whose anti-tumor activity depend on mesothelin expression on the target cells and it can be a potential antibody-based immunotherapeutic for TNBC patients [97]. NK cell function is regulated by molecules from promoting and suppressing receptors interacting with ligands on target cells. Lectin-like Transcript-1 (OCIL, CLEC2D, LLT1) is a ligand that interacts with NK cell receptor NKRP1A and prevents NK cell activation. Inhibiting LLT1 on TNBCs with antibodies hinders the interaction with NKRP1A and increases lysis of TNBCs by primary NK cells [98].

#### 2.2.2. CD8<sup>+</sup> TILs

CD8<sup>+</sup> TILs are the main kind of cytolytic lymphocytes in tumors. Kronqvist and group detected the expression of stromal TILs and CD8<sup>+</sup> TILs in 179 patients with TNBC using IHC and observed that the prognostic value of CD8<sup>+</sup> TILs and TILs varied when detected in various cancer compartments [99]. Presence of CD8<sup>+</sup> TILs in a large cohort of 12,439 breast cancer patients correlated with a significant decrease in the relative hazard of death in both the ER- positive and the ER- negative HER2-positive subtypes [100]. Ishida and colleagues assessed the CD8<sup>+</sup> TILs and Foxp3<sup>+</sup> Tregs status of the residual tumors in 131 patients with TNBC who received neoadjuvant chemotherapy (NAC) at three institutions and the rates of their changes before and after NAC were evaluated. They found that TNBC patients with a high CD8<sup>+</sup> TILs level or high CD8/Foxp3 ratio in residual tumors exhibit significantly favorable recurrence-free survival (RFS) and breast cancer-specific survival (BCSS) [101]. Another study also showed that CD8<sup>+</sup> TILs were related to favorable DMFS, DFS, and BCSS in the entire 207 breast cancer group and in 56 TNBC group [102]. BRCA1-IRIS overexpressing (IRISOE) TNBC carcinomas had more CD25+/Foxp3<sup>+</sup> Tregs and few CD8+/PD-1<sup>+</sup> cytotoxic T-cells, which showed that the interaction between macrophages and IRISOE cells initiated an immunosuppressive TME within TNBC tumors [71]. TOPOIIα and CD4<sup>+</sup> TILs were significantly positively associated with CD8<sup>+</sup> TILs and they exhibited a significantly good 5-year DFS but only a high infiltration of CD8<sup>+</sup> TILs showed significantly better 5-year OS in 52 TNBC patients that received taxane-anthracycline-based NAC [103,104]. Calcium/calmodulin-dependent kinase (CaMKK2), expressed in tumor-related stromal cells, could promote tumor growth. The inhibition of CaMKK2 within myeloid cells suppresses tumor growth by increasing immune-stimulatory myeloid subsets and intra-tumoral accumulation of CD8<sup>+</sup> T cells in TNBC [105]. PARP inhibitor Olaparib induced CD8<sup>+</sup> T cell activation and infiltration via activation of the cGAS/STING pathway, which provided rationale for combining the PARP inhibitors with immunotherapies for TNBC [106]. A recent study reported that CD8<sup>+</sup> TILs were crucial for infected cell vaccine (ICV) efficacy, which was composed of autolo-

gous tumor cells infected with an oncolytic Maraba MG1 virus in vitro in the BALB/c-4T1 model. Increased migration and proliferation ability of human CD8<sup>+</sup> TILs were observed following exposure to ICV [107]. A series of studies illuminated the mechanisms of different infiltration levels of CD8<sup>+</sup> TILs in immunomodulation and anti-tumor response of TNBC. By spatially modulating the diffusion/chemotactic coefficients of T cells via partial differential equations, Almohanad et al. found that a type of chemorepellent inside cancer cell clusters but not dense collagen fibers, prevents the infiltration of CD8<sup>+</sup> TILs into cancers and cancer cell clusters, which may imply a poor prognosis in TNBC [108]. Intra tumoral CD8<sup>+</sup> TILs enhance the efficacy of treatment through triple combined inhibition of PDGFRβ/ MEK1/2/JAK2 signal pathway in vivo in TNBC [109]. Gruosso et al., found that there were many different kinds of CD8<sup>+</sup> TILs localization profiles with distinct meta-signatures, which were prognostic indicators in a cohort of TNBC [17]. Dong et al. investigated the genome-scale CD8<sup>+</sup> TILs CRISPR screen in the context of immunotherapy in vivo and in vitro and found that DHX37 interacts with PDCD11 and affects NF-κB activity to modulate CD8<sup>+</sup> TILs activation, cytokine production, and cytotoxicity [110].

#### 2.2.3. M1 Macrophages

M1 phenotype macrophages, also called classical macrophages, are pro-inflammatory, and can activate the immune response and oppose tumorigenesis [111]. In vitro and in vivo studies have shown that M1 macrophage polarization decreases the expression of nuclear REST corepressor 1 (CoREST), LSD1 and the zinc finger protein SNAIL, and LSD1 inhibitors can target both CoREST and flavin adenine dinucleotide (FAD) binding domains of LSD1 to initiate macrophages toward M1 phenotype in TNBC successfully [112]. Another study revealed that exposure to infected cell vaccine (ICV) could induce the polarization of monocytes to M1 subtype [107].

Using the 4T1 TNBC murine model, Meyer and colleagues showed that in the early stages of disease, higher M1-related cytokines are released and decreased M2 macrophages infiltrate in the TME, while upon metastasis a dramatic enhancement in M2-related cytokine expression levels are detected and more immunosuppressive cells such as M2 macrophages infiltrate in the TME [113]. High level of CCL5 is related to recruitment of M1 macrophages, CD8<sup>+</sup> TILs, CD4 activated T lymphocytes, and NK activated cells in TNBC using CIBERSORT analysis [114]. The clinical significance and involved mechanisms of each constituent in TNBC microenvironment are included in Table 1.


**Table 1.** Clinical significance and involved mechanisms of immune cells and markers.


**Table 1.** *Cont.*

#### **3. The Composition of TME Contributes to TNBC Subtype Classification**

During TNBC progression, TME reconstruction including the ratio of immune cells and release of various immune cytokines play crucial roles, and the research focusing on stromal and immune composition of TME has contributed significantly to different subtype classification of TNBC [17]. Lehmann and colleagues distinguished six TNBC subtypes showing unique gene expression profiles and ontologies, comprised of two basal-like (BL1 and BL2), a mesenchymal stem-like (MSL), a mesenchymal (M), an immunomodulatory (IM), and a luminal androgen receptor (LAR) subtype. Interestingly, immune genes in IM subtype overlap with gene signatures in medullary breast cancer which is correlated with good prognosis despite its high-grade scores [118]. Park and colleagues distinguished four stromal axes abundant for T cells, B cells, epithelial markers and desmoplasia and assigned a score along with each marker and associated it with different TNBC subtypes. This classification method better depicted tumor heterogeneity and led to a superior evaluation of benefit from therapeutics and prognosis [119].

In addition, three subtypes of TNBC have been identified: an apocrine cluster (C1), which is more related to luminal, PIK3CA-mutated hallmarks and shows intermediate biological aggressiveness; and two basal-like clusters (C2 and C3), which show a major biological discrepancy related to immune response and are sensitive to drugs combating immunosuppression or stimulate adaptive immune response respectively [120]. Shao and colleagues analyzed genomic, clinical, and transcriptomic data of 465 primary TNBC patients, and also identified four subtypes of TNBC, including basal-like immune-suppressed (BLIS), immunomodulatory (IM), luminal androgen receptor (LAR) and mesenchymallike (MES). They also showed that IM subtype is related to immune response and there are elevated immune cell signaling, TILs, high mRNA expression quantities of immune checkpoint blocking genes such as PD-L1, PD-1, CTLA4, and IDO1 [121]. Using the data of 465 Taiwanese with breast cancer, five TNBC subtypes were classified, namely, basallike (BL), mesenchymal stem like (MSL), immunomodulatory (IM), mesenchymal (M), and luminal androgen receptor (LAR), and they observed the interaction between IM subtype and MSL subtype, which also implied the involvement of TME in TNBC subtype classification [122]. Distinguishing a four-gene decision tree signature (*TP53BP2, EXO1, RSU1* and *FOXM1*) using transcriptomic and genomic data analysis established six subtypes of TNBC, named MC1 to MC6, comprised by five of varying sizes (MC1- MC5) and one large subtype MC6. Further study showed high level of CD8<sup>+</sup> and CD4<sup>+</sup> immune signatures and decreased expression of MAPK pathway related genes in MC6 subtype [123]. Another group identified three TNBC subtypes including Immunity High (Immunity H), Immunity Medium (Immunity M), and Immunity Low (Immunity L) based on the immunogenomic profiling of 29 immune signatures. In Immunity H subtype, greater anti-tumor immune response and immune cell infiltration, as well as favorable prognosis were detected compared to the other subtypes, which showed the close relationship between tumor immune microenvironment and TNBC classification [124]. TNBC tumors were classified into four subgroups (luminal-androgen receptor expressing, basal, claudin-high and claudin-low), in addition to two subgroups associated with immune activity using gene expression and clinical data and the latter two immune subgroups were defined as correlated to immune activity closely. Meanwhile, claudin-high subgroup had low response to neoadjuvant chemotherapy, and luminal immune-positive subgroup had favorable survival prognoses [125]. A recent study identified four TNBC epitopes,

named as Epi-CL-A, Epi-CL-B, Epi-CLC, and Epi-CL-D using genome-wide DNA methylation properties and clinical and demographic variables, as well as gene mutation and gene expression data. Intriguingly, subtype Epi-CL-D showed a positive regulation of T lymphocyte-mediated cytotoxicity and associated molecules such as IL15RA and CCL18, which partially explained the favorable outcome and a positive immune response in this subtype [126]. Furthermore, a research group classified TNBC tumors into immune subtype A and B by the density of monocytes, γδ T cells, stromal CD4<sup>+</sup> T cells, M1 macrophages and M2 macrophages using CIBERSORT or IHC method and they proved that enriched immune-related pathways and higher levels of immune checkpoint cytokines such as PD-1, PD-L1 and CTLA-4 could be detected in phenotype A [127]. Romero-Cordoba et al. also identified three immuno-clusters in TNBC tumors using clustering analysis based on immune-related gene expression signatures and found that platelet to lymphocyte ratio (PLR) was associated with tumor immune infiltration [128].

We have included all the classification methods and the clinical significance (Table 2). Classification of TNBC has been developed extensively implying that a precision-treatment era has come in TNBC. Chemotherapy still remains the key treatment for TNBC but other targeted therapies including immunotherapy can be combined for better tailored treatments and are the focus of ongoing research efforts.


**Table 2.** TNBC subtype classification.

#### **4. Chemotherapy-Induced TME Remodeling Modulates TNBC Immune Response**

It has been reported that cytotoxic drugs such as anthracycline and platinum agents, could induce immunogenic cell death (ICD), and stimulate anti-tumor immune response of T lymphocytes [18,129]. Damage-associated molecular patterns (DAMPs) are cytokines that are released by damaged or activated cells; have great immune stimulating response,

and cause ICD [18]. ICD involves the cell surface exposure of calreticulin (CRT), release of DAMPs-related high mobility group box1 (HMGB1) and autophagy-dependent ATP release, which together, leads to the antigen uptake and presentation of DC cell, and then activates the CD8<sup>+</sup> TILs to play the anti-tumor role [130,131]. Carboplatin or paclitaxel combined with radiation generates both chemotherapeutic enhancement of ICD and a dose-dependent induction of ICD in TSA mammary carcinoma cells [132]. Doxorubicin and paclitaxel treatment results in the recruitment of innate immune cells and CSF1Rdependent macrophages infiltration in PyMT-MMTV mammary carcinoma through an increase of CCL2, CXCL2, CSF-1, interleukin-34 and vascular permeability [133,134]. Docetaxel polarizes MDSCs toward M1-like phenotype and upregulates macrophages markers (CD86, MHC class II, and CD11c) in vivo and in vitro partly through an inhibition of STAT-3 in 4T1-Neu mammary cancer implants [135]. All these studies emphasize that chemotherapy can induce TME remodeling through distinct signaling pathways. In this part, we have focused on three crucial factors related to chemotherapy-induced TME remodeling, which are HMGB1, exosomes and S1PR1. The clinical significance of HMGB1, exosomes and S1P/SPHK1/S1PR1 as well as their involvement in TNBC immunomodulation and tumor progression is shown in Figure 2.

**Figure 2.** Chemotherapy-induced immunogenic cell death and immunomodulation in TNBC. Chemotherapy induces immunogenic cell death (ICD), and then promotes the release of damage-associated molecular patterns (DAMPs) including high mobility group box1 (HMGB1), exosomes and sphingosine-1-phosphate receptor 1 (S1PR1) by damaged or activated cells. Chemotherapy combined with targeted therapy could enhance anti-tumor immunity through promoting function of immunoreactive lymphocytes and blocking or reversing function of immunosuppressive cells. (ICD, immunogenic cell death; DAMPs, damage-associated molecular patterns; HMGB1, high mobility group box1; S1P, sphingosine-1-phosphate; SPHK1, sphingosine kinase 1; S1PR1, sphingosine-1-phosphate receptor 1; TNBC, triple negative breast cancer; TME, tumor microenvironment).

#### *4.1. Chemotherapy-Induced HMGB1 Release Participates in TNBC Immunomodulation* 4.1.1. Chemotherapy-Induced HMGB1 Enhances Anti-Tumor Immune Response

High mobility group box1 (HMGB1) is a highly conserved DNA-binding nuclear protein, involved in many kinds of diseases, including cancer, arthritis, and sepsis [136]. Extracellular HMGB1 in response to inflammation activates the host immune system. HMGB1 can combine with TLR-2, TLR-4, and TLR-9, and recruit the inflammatory cells to microenvironment. This activates the DCs, enhances the antigen presentation ability and anti-tumor immune response [137].

4.1.2. HMGB1 Is Related to High Recurrence Risk and Progressive Disease after Neoadjuvant Chemotherapy

A study indicated that the nuclear expression of HMGB1 in breast cancer cells negatively correlates with Tregs and TAMs [138], and could predict the recurrence risk of residual tumor [139]. HMGB1 expression in cytoplasm is higher in HER2-positive and TNBCs tumors than in hormone receptor (HR)-positive tumors. High cytoplasmic HMGB1 significantly correlates with advanced histologic grade, abundant TILs, and high expression of CD8<sup>+</sup> TILs but shows no prognostic significance in TNBC [140]. Intracellular HMGB1 expression has been detected in fibroblasts conditioned medium (CM) treated breast cancer cells and in doxorubicin-treated cells. Extracellular HMGB1 is upregulated in CM after doxorubicin-induced MDA-MB-231 cell death, which show the potential of fibroblasts in stroma to contribute to chemo-resistance partly by fibroblast-induced HMGB1 production [141]. It has been shown that low cytoplasmic HMGB1-positive breast tumor cells and high ASMA-positive fibroblasts predict adverse prognosis in TNBC [142]. Tanabe and colleagues reported that positive HMGB1 expressions are higher in the clinical progressive disease (cPD) than in control group during neoadjuvant chemotherapy in TNBC patients [143]. Some of HMGB1 single nucleotide polymorphisms (SNPs) have been related to tumor progression in T2 tumor, pathologic grade 3 disease, and distant metastasis in TNBC and HER2-enriched tumors compared with luminal tumors [144]. By targeting HMGB1- RAGE signaling pathway, miR-205 impairs the viability and epithelial-to-mesenchymal transition in TNBC cells [145]. HMGB1 released by breast cancer cells is N-glycosylated at Asn37, which promotes the transition from monocytes to MDSC-like cells and contributes to M-MDSC differentiation from bone marrow through the p38/NFκB/Erk1/2 signaling pathway [146].

#### *4.2. Chemotherapy-Induced Exosomes Secretion Interconnects TME and TNBC Immune Response* 4.2.1. Chemotherapy-Induced Exosomes Are Released to TME

Exosomes are tiny membrane vesicles (30–100 nm in diameter) synthesized in late endosomes and secreted into the extracellular milieu by various cells. They contain functional molecules (lipids, proteins, DNA, and RNA) that can be transferred to recipient cells, playing a key role in intercellular communication [147]. Apoptosis exosome vesicles (AEVs) are special exosomes overexpressing S1PR1 and S1PR3 released by the tumor cells in response to certain chemicals. These AEVs induce the expression of inflammatory chemokines and cytokines which participate in the pathological and physiological process of DAMPs [147].

#### 4.2.2. Exosomes Are Related to TNBC Tumor Progression and Provide Therapy Options

Some investigations have explored connections between exosomes and TNBCs [148]. Hypoxia induces the production of exosomes and microvesicles (MVs) in breast cancer cells through HIF-dependent RAB22A expression, which can stimulate ECM invasion, focal adhesion formation, lung colonization and is associated with decreased OS and MFS in the mouse models [149]. Stevic and colleagues determined miRNA expression profiles of exosomes originated from the plasma of TNBC and HER2-positive breast cancer patients before neoadjuvant therapy. They found that exosomal miRNAs (miR-155 and miR-301) correlate with the risk factors and clinicopathological factors significantly and can predict pCR rate [150]. Extracellular vesicles (EVs) from HCC1806 but not from MDA-MB-231 cells exhibit enhanced drug resistance and alter the levels of genes involved in cell apoptosis and proliferation pathways in MCF10A cells [151]. Ni and colleagues quantified the levels of miRNAs expression in exosomes from plasma of 8 ductal carcinoma in situ (DCIS) patients, 32 breast cancer (BC) patients and 8 healthy women; they found that different levels of exosomal miRNAs had distinct prognostic value in different subtypes of BC and the expression of miR-16 was lower in TNBC than HR-positive counterparts [152]. Exosomes from TNBC tissues regulate cell apoptosis and TME changes. MiR-770 played its multi-functional role in TNBC by down-regulating gene STMN1 as follows: (i) was

associated with favorable prognosis of TNBC, (ii) increased the sensitivity of TNBC cells to doxorubicin through induction of apoptosis, (iii) regulated TAMs-induced chemotherapy resistance, and (iv) inhibited invasion and migration ability of TNBC cells via EMT pathway [153]. Intriguingly, chemotherapy-induced senescent cells secreted more extracellular vesicles than non-senescent cells in TNBC [154]. Exosomes could facilitate co-delivery of cholesterol-modified miR-159 and therapeutic quantities of doxorubicin to TNBC cells both in vitro and in vivo [155]. A formulation of erastin (a low molecular weight chemotherapy drug that induces ferroptosis)-loaded exosome was labeled with special chemicals to target TNBC cells, which enhanced the uptake efficiency of drugs into MDA-MB-231 cells and had a better preventing effect on the migration and proliferation, revealing that the exosome-based therapy might serve as a novel and powerful delivery method for anti-cancer therapy [156].

#### *4.3. S1P/SPHK1/S1PR1 Link TME Changes to TNBC Immunomodulation* 4.3.1. S1P/SPHK1/S1PR1 Is Associated with TME Changes

Sphingosine-1-phosphate (S1P), a novel lipid signaling mediator with both intracellular and extracellular functions, is generated by sphingosine kinase 1 (SPHK1), an enzyme catalyzing phosphorylation of sphingosine. S1P/SPHK1 interacts with constituents in TME and modulate the progression and metastasis of breast cancer. Binding of S1P to sphingosine-1-phosphate receptor (S1PRs) on cell surface activates cytokines in the cytoplasm and gene activation in the nucleus in an autocrine and paracrine manner [157,158]. S1P, S1PRs, and SPHK1 expression are related to metastatic progression in breast cancers in vivo [159]. An investigation in melanoma suggested that S1PR1 causes immune functional change of T lymphocytes via PPARγ signal pathway [160]. A recent investigation in breast cancer showed that S1PR1 causes the change of TAMs phenotype, promotes neolymph vascularization, and the change of TME via activating inflammatory factors such as Nlrp3 and IL-1β [161]. Another team also showed that S1PR1 phosphorylates the complex of vasculogenic mimicry (VM), and the inhibition of S1PR1 decreases endotheliumdependent vessel (EDV), but causes the production of VM, invasion, and metastasis in vitro and in vivo [162]. Kim and colleagues showed that IL-22 induces S1PR1 and IL22R1 expression in myeloid cells and macrophages, and induce MCP1 expression in myeloid stem cells (MSCs), and then facilitate macrophage infiltration, implying a potential effect of IL-22 on promoting bone metastasis of breast cancers via IL22R1/S1PR1 pathway [158]. S1P1 is expressed in tumor antigen-specific bone marrow (BM) Tregs selectively in breast cancer, and can be induced by BM-resident antigen-presenting cells in conjunction with T cell receptor stimulation [163].

#### 4.3.2. S1P/SPHK1/S1PR1 Is Associated with TNBC Tumor Progression

A preclinical study detected the function of S1PR1-antibody on the growth of breast cancer cell lines MDA-MB-231 and SK-BR-3. They found that S1PR1-antibody not only increases the cytotoxicity of carboplatin on MDA-MB-231 cells but also enhances the antiproliferative outcome of S1P on SK-BR-3 cells [164]. It has been reported that apoptotic tumor cells release S1P, and then stimulate the generation of lipocalin 2 (LCN2) in TAMs and is associated with breast cancer metastasis [165]. As the key kinase of S1P combination, SPHK1 has been found to be overexpressed in TNBC compared with other breast cancer subtypes, and promotes tumor metastasis. By targeting SPHK1 or its downstream signaling pathway (NF-κB pathway) with available inhibitors, TNBC metastasis is effectively inhibited [166]. Maiti and colleagues found that SPHKs/S1P axis is a crucial constituent of survival and growth of LM2-4 cells compared to parental MDA-MB-231 cells, and nuclear SPHK2 (in MDA-MB-231 cells) is also indispensable for LM2-4 cells survival and growth [167]. Obesity and high-fat diet are the main cause for increased expression of the S1P and SPHK1, and targeting the SPHK1/S1P/S1PR1 decreases key proinflammatory cytokines, macrophage infiltration, and tumor progression [168]. However, Lei and colleagues found that S1PRs expression inhibits tumor progression in breast cancer patients [169]. The clinical significance of DAMPs-associated molecules (HMGB1, exosomes, and S1P/SPHK1/S1PR1) and the mechanisms involved in TNBC immunomodulation and tumor progression are included in Table 3.

**Table 3.** Clinical significance and involved mechanisms of DAMPs-associated molecules.


#### **5. Conclusions**

The role of the tumor microenvironment (TME) in triple negative breast cancer (TNBC) immunomodulation is vitally important. The deeper understanding of immunosuppressive and immunoreactive TME has contributed to specific subtype classification of TNBC. In future, we may be able to treat TNBC patients with more precision according to their subtype. Agents that remodel TME, promote function of immunoreactive lymphocytes, block function of immunosuppressive cells, and prevent inhibitory signaling pathways can all be considered. Furthermore, therapies targeting HMGB1, exosomal microRNAs, and S1P/SPHK1/S1PR1, can also be considered in combination with chemotherapy. In conclusion, immunosuppressive and immunoreactive role of TME, the contribution of TME in TNBC subtype classification, chemotherapy-induced TME changes and its role in TNBC immunomodulation are crucial for TNBC management. TME has provided a new direction to explore novel and effective combination regimens for precision treatment of TNBC.

**Author Contributions:** Conceptualization, H.Z., S.P., S.S., X.W. and D.S.; methodology, H.Z. and S.P.; software, S.S.; validation, S.S., D.S. and S.P.; formal analysis, D.S.; investigation, H.Z.; resources, H.Z., D.S.; data curation, H.Z.; writing—original draft preparation, H.Z.; writing—review and editing, D.S.; visualization, S.P.; supervision, D.S.; project administration, X.W.; funding acquisition, H.Z. and X.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Natural Science Foundation of Hubei Province, grant number 2020CFB874.

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

#### **References**

