**Involvement of Di**ff**erentially Expressed microRNAs in the PEGylated Liposome Encapsulated <sup>188</sup>Rhenium-Mediated Suppression of Orthotopic Hypopharyngeal Tumor**

#### **Bing-Ze Lin 1,**† **, Shen-Ying Wan 1,2,**† **, Min-Ying Lin <sup>1</sup> , Chih-Hsien Chang 1,3 , Ting-Wen Chen 4,5,6 , Muh-Hwa Yang 7,8,9 and Yi-Jang Lee 1,7,\***


Academic Editor: Krishan Kumar

Received: 30 June 2020; Accepted: 6 August 2020; Published: 8 August 2020

**Abstract:** Hypopharyngeal cancer (HPC) accounts for the lowest survival rate among all types of head and neck cancers (HNSCC). However, the therapeutic approach for HPC still needs to be investigated. In this study, a theranostic <sup>188</sup>Re-liposome was prepared to treat orthotopic HPC tumors and analyze the deregulated microRNA expressive profiles. The therapeutic efficacy of <sup>188</sup>Re-liposome on HPC tumors was evaluated using bioluminescent imaging followed by next generation sequencing (NGS) analysis, in order to address the deregulated microRNAs and associated signaling pathways. The differentially expressed microRNAs were also confirmed using clinical HNSCC samples and clinical information from The Cancer Genome Atlas (TCGA) database. Repeated doses of <sup>188</sup>Re-liposome were administrated to tumor-bearing mice, and the tumor growth was apparently suppressed after treatment. For NGS analysis, 13 and 9 microRNAs were respectively up-regulated and down-regulated when the cutoffs of fold change were set to 5. Additionally, miR-206-3p and miR-142-5p represented the highest fold of up-regulation and down-regulation by <sup>188</sup>Re-liposome, respectively. According to Differentially Expressed MiRNAs in human Cancers (dbDEMC) analysis, most of <sup>188</sup>Re-liposome up-regulated microRNAs were categorized as tumor suppressors, while down-regulated microRNAs were oncogenic. The KEGG pathway analysis showed that cancer-related pathways and olfactory and taste transduction accounted for the top pathways affected by <sup>188</sup>Re-liposome. <sup>188</sup>Re-liposome down-regulated microRNAs, including miR-143, miR-6723, miR-944, and miR-136 were associated with lower survival rates at a high expressive level. <sup>188</sup>Re-liposome could suppress the HPC tumors in vivo, and the therapeutic efficacy was associated with the deregulation of microRNAs that could be considered as a prognostic factor.

**Keywords:** hypopharyngeal cancer; <sup>188</sup>Re-liposome; repeated therapy; NGS; microRNA

#### **1. Introduction**

Hypopharyngeal cancer (HPC) represents malignant growth in the hypopharynx region and accounts for about 5% of all head and neck cancers (HNSCC) [1]. As HPC is a rare cancer type with a late occurrence of symptoms and tumor spreading, it is not uncommon for it to be detected at advanced stages with a high mortality rate and poor prognosis [2]. HPC can be treated by conventional surgery, radiotherapy and chemotherapy, while radiotherapy alone is usually used at an early stage [3]. On the other hand, a combination of different therapeutic modalities can improve the five-year survival of this disease [4]. Several lines of evidence have claimed that a combination of radiotherapy and chemotherapy would provide better control of locoregional recurrence compared to surgical procedures [4,5]. A radiopharmaceutical named 99mTc-MIBI (methoxy-isobutyl-isonitrile) has been reported to detect HPC with up to a 95% sensitivity using single photon emission computed tomography (SPECT) [6]. However, nuclear medicine has not been reported to have assessed or monitored the efficacy of HPC therapy as mentioned above.

Radiopharmaceuticals are not only used for diagnostic purposes but also therapeutic purposes, so-called radiotheranostics [7]. Rhenium-188 (188Re) belongs to this type of radionuclide as it emits 85% high-energy β-particles (2.12MeV) and 15% γ-rays (155keV) [8]. The average soft tissue penetration distance of β-particles is only around 3.8mm, suggesting that <sup>188</sup>Re is suitable for tumor ablation and will not have significant side effects on distant normal tissues [9,10]. <sup>188</sup>Re has been conjugated to hydroxyethylidene diphosphonate (HEDP) for bone pain palliation [11,12]. The antibody conjugated <sup>188</sup>Re has also been reported to treat different cancers [13]. Additionally, low immunogenic peptides such as somatostatin derivative conjugated <sup>188</sup>Re, have been investigated in clinics for patients with advanced pulmonary cancer [14]. <sup>188</sup>Re-labeled lipiodol and microspheres were used for the treatment of hepatocellular carcinoma [15–17]. <sup>188</sup>Re-labelled radiocolloids have also been developed to treat skin cancers within a brachytherapy device [18]. <sup>188</sup>Re-loaded lipid nanocapsules, also called <sup>188</sup>Re-liposome, have been demonstrated to be biocompatible and subjected to a phase 0 clinical study for patients with metastatic tumors [19]. <sup>188</sup>Re-liposome has shown a theranostic efficacy in various human cancers, including colorectal cancer, glioblastomas, lung cancer, ovarian cancer, esophageal cancer and head and neck cancer using xenograft tumor models [20–26]. The molecular mechanisms of rhenium-188 labelled radiopharmaceuticals are of interest for investigation aimed at interpreting the potent therapeutic efficacy.

Radiogenomics is defined as connecting radiomics and genetic profiles to apply the feature of medical imaging in radiation-mediated molecular responses [27]. The purpose of this discipline is to predict the association between gene expression and the radiotherapy-induced toxic effects of tumors [28]. Next-generation sequencing (NGS) analysis is the most cutting-edge technology that can decipher dramatic amounts of gene expressive alterations in tumors, with or without therapies [29,30]. NGS applications include RNA sequencing (RNA-Seq) that can analyze the expression of various RNA populations (mRNA, microRNA, long non-coding RNA, etc.) and their modification forms generated from alternative splicing, mutations or gene fusion [31]. Compared to the conventional cDNA expression microarray, RNA-Seq has a broader spectrum for finding novel and unidentified transcripts [32]. This method is important for providing more detailed and quantitative data for bioinformatics analysis of genetic profiling in novel drug development and predicting the signaling pathways of toxicity and therapeutic effects [33]. As a radioactive compound, <sup>188</sup>Re-liposome is believed to influence the genetic profile of tumors. However, related studies have rarely been reported.

In this study, we investigated the effects of <sup>188</sup>Re-liposome on the miRNA expressive profiles of HPC derived xenograft tumors using RNA-Seq technology. The changed miRNA profiles were analyzed using the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database and Ingenuity Pathway Analysis (IPA). The total expressive amount of miRNA from <sup>188</sup>Re-liposome-treated tumors was about 10% lower than that of untreated tumor. The changed miRNA profile led to 4498 differentially expressed genes (DEGs) that influenced 30 molecular pathways, including olfactory transduction, the cancer pathway, and taste transduction, which accounted for the top three pathways. We also found that <sup>188</sup>Re-liposome up-regulated several tumor suppressor microRNAs, such as miR-34-5p, miR-193a-5p, miR-125b-5p, miR-133a-5p, and miR-133b-5p. Concomitantly, several oncomirs, including miR-21-5p, miR-32-5p and miR-205-5p were down-regulated by <sup>188</sup>Re-liposome. An online Kaplan-Meier (K-M) plotter with The Cancer Genome Atlas (TCGA) database was also used to compare the expression of these miRNA and the survival of head and neck cancer patients. The results of <sup>188</sup>Re-liposome-induced miRNA dysregulation detected by RNA-Seq were discussed.

#### **2. Results**

#### *2.1. E*ff*ects of <sup>188</sup>Re-Liposome on HPC Derived Orthotopic Tumors Using the Repeated Dose Regime*

A flowchart of <sup>188</sup>Re-liposomal manufacturing and the chemical structures of the <sup>188</sup>Re-perrhenate precursor and BMEDA chelator are illustrated in Figure 1. The timeline was schemed for the establishment of HPC tumor-bearing mice using FaDu-3R cells (Supplementary Figure S2), the administration of <sup>188</sup>Re-liposome, the optical imaging of tumor responses, tumor resection for RNA extraction, and NGS analysis (Figure 2A). The growth of orthotopic tumors was significantly suppressed by <sup>188</sup>Re-liposome but not saline as detected using bioluminescent imaging (Figure 2B,C). The size of resected tumors also exhibited obvious differences between the saline control and <sup>188</sup>Re-liposome-treated mice after 30 days of initial implantation (Figure 2D). The body weights of tumor-bearing mice were not significantly affected by <sup>188</sup>Re-liposome (Figure 2E). Furthermore, we showed that DNA damage marker γ-H2AX was significantly up-regulated by <sup>188</sup>Re-liposome treatment (Figure 2F,G). These data suggest that the regime of <sup>188</sup>Re-liposome treatment with repeated doses exhibited therapeutic efficacy, with little systemic toxicity.

#### *2.2. Use of NGS Analysis to Investigate the microRNA Expressive Profile of HPC Tumor Treated with <sup>188</sup>Re-Liposome*

The resected FaDu HPC tumors treated with saline or repeated doses of <sup>188</sup>Re-liposome were subjected to RNA extraction to obtain a high quality of total RNA for RNA-seq analysis (Supplementary Figure S3). The quality of raw data obtained by this analysis was determined by the GC content, by which the raw data quality of the saline control and <sup>188</sup>Re-liposome-treated tumor was shown to be similar (Supplementary Figure S4). To confirm the expressive change in microRNA in tumors treated with or without <sup>188</sup>Re-liposome, the reads number of small RNA (15–55 mers) was extracted and counted using a Small RNA Analysis tool (see Materials and Methods). The average lengths of small RNA after extraction were 32.9 and 32 mers for the control and <sup>188</sup>Re-liposome treated tumor, respectively. Under this condition, the total reads number of the control and <sup>188</sup>Re-liposome treated tumor was 10,612,998 and 9,844,775, respectively. The reads were then subjected to the miRBase database (release 21) for the annotation of small RNA using the Annotate and Merge Count of Small RNA Analysis software. The annotated small RNAs of saline-treated tumors and <sup>188</sup>Re-liposome treated tumors were 639 and 572, respectively. The differentially expressed microRNAs were clustered in a heatmap for a comparison of saline-treated and <sup>188</sup>Re-liposome treated HPC tumor models (Supplementary Figure S5). Furthermore, we set the cutoff at five-fold change (log2) of microRNA and showed that 13 microRNA and 9 microRNA with mature forms were up-regulated and down-regulated by <sup>188</sup>Re-liposome normalized to the saline control, respectively (Table 1). The 13 up-regulated microRNAs ranked from highest to lowest fold change were miR-206-3p, mir668-3p, mit-485-3p, miR-382-5p, miR-1268b-5p, miR-193a-5p, miR-7-1-5p, miR-378a-5p, miR-1266-5p, miR-4510-5p, miR-370-3p, miR-34a-5p, and miR-342-5p. The nine down-regulated microRNAs ranked from highest to lowest fold change were miR-142-5p, miR-6723-5p, miR-944-3p, miR-142-3p, miR-136-3p, miR-151b-3p, miR-194-2-5p, miR-143-5p, and miR-3960-3p. According to the online analysis of dbDEMC, <sup>188</sup>Re-liposome-up-regulated microRNAs were mostly naturally down-regulated in HNSCC and were predicted as tumor suppressors (Table 2). MiR-193a, miR-7-1-5p and miR-342-5p were up-regulated in HNSCC, but could still be tumor suppressors in different types of cancer (see references in Table 2). For the nine <sup>188</sup>Re-liposome down-regulated microRNAs, only four of them (miR-136-3p, miR-142-3p, miR-944-3p, and miR-142-5p) were reported in the HNSCC of dbDEMC database. Interestingly, three out of these four microRNAs were found to be up-regulated in HNSCC and predicted as oncogenes (Table 3). However, most of the <sup>188</sup>Re-liposome-down-regulated microRNAs contain oncogenic properties (see references in Table 3). These results suggested that the therapeutic efficacy of <sup>188</sup>Re-liposome was associated with the deregulation of tumor-suppressive and/or oncogenic microRNAs.

**Figure 1.** The flowchart of Re-liposomal preparation. The chemical structures of Na ReO , BM **Figure 1.** The flowchart of <sup>188</sup>Re-liposomal preparation. The chemical structures of Na188ReO<sup>4</sup> , BMEDA and formed <sup>188</sup>Re-BMEDA were also illustrated.

γ **Figure 2.** Comparison of PEGylated <sup>188</sup>Re-liposomal accumulation in orthotopic hypopharyngeal cancer (HPC) tumors after repeated injections. (**A**) The experimental scheme for <sup>188</sup>Re-liposome treatment and analysis. (**B**) Reporter gene imaging of tumor growth responding to repeated doses of <sup>188</sup>Re-liposome, and the saline-treated control. (**C**) Quantification of bioluminescent imaging (BLI) signals. \*: *p* < 0.05. (**D**) Representative photos of excised orthotropic tumors with or without the treatment of <sup>188</sup>Re-liposomes. (**E**) Caliper measurement of tumor volumes. Data are represented as means ± S.D. \*\*: *p* < 0.01. (**F**) Measurement of body weights of mice. (**G**) Comparison of the γ-H2AX protein expression in tumors with or without the treatment of <sup>188</sup>Re-liposomes. (**H**) Densitometric quantification of Western blots. \*: *p* < 0.05.


**Table 1.** Expression of microRNAs in human hypopharyngeal tumor model treated with 188Re-liposome.

<sup>a</sup>—Fold change in microRNA over 5 or below −5 were selected. <sup>b</sup>—TMM: Trimmed mean of M values. <sup>c</sup>—RPM: Reads of exon model per million mapped reads. (ExonMappedReads × 10<sup>6</sup> /TotalMapped Reads).

#### *2.3. Validation of microRNA Identified in NGS Data Using qPCR*

We next used qPCR to validate up-regulated and down-regulated microRNA in the FaDu HPC tumor model treated with <sup>188</sup>Re-liposome. We selected microRNAs that displayed over five-fold deregulation with the highest RPM, including miR-206-3p, miR-382-5p, miR-378a-5p, miR-3960-3p, and miR-142-5p to be validated. The results showed that the expressive patterns of these microRNAs obtained by qPCR were consistent with the observations of NGS analysis (Figure 3).

#### *2.4. Investigation of Di*ff*erentially Expressed microRNAs in Clinical Samples*

As <sup>188</sup>Re-liposome could influence the expression of certain microRNAs that may correlate with the tumor-suppressive effect of the HPC model, we were interested in examining the status of these microRNAs in clinical HNSCC tumors. First, we obtained clinical HNSCC tissues from patients (n = 6) to investigate the differential expression of microRNAs in tumor tissues and adjacent normal tissues using qPCR analysis. We selected miR-206-3p, miR-378a-5p and miR-142-5p as they exhibited the highest differential deregulation by <sup>188</sup>Re-liposome (Figure 3). The results showed that the expression of miR-206-3p and miR-378a-5p was down-regulated (Figure 4A,B), while miR-142-5p was up-regulated in tumors compared to normal tissues (Figure 4C). Additionally, we employed the clinical information of HNSCC in the TCGA database to compare the differentially expressed microRNAs in HNSCC and normal tissues. Because the number of cases of hypopharynx cancer was too low to be analyzed, here we used clinical information on larynx cancer types instead. A heatmap was generated for the microRNAs displaying over five-fold change caused by <sup>188</sup>Re-liposome (Supplementary Figure S6). Accordingly, we found that miR-206 (equivalent to miR-206-3p) and miR-378a-5p were significantly down-regulated, while miR-143-5p, miR-142-3p, and miR-944 were significantly up-regulated in tumors (Figure 4D). MiR-142-5p also exhibited a trend of up-regulation, although the significance was marginal. This suggests that <sup>188</sup>Re-liposome can reverse the expression of the microRNAs that were originally deregulated in HNSCC.



a—The change of microRNA was determined by the dbDEMC online databases. <sup>b</sup>—The selected references may be not HPC or HNSCC related. c—The tumor suppressive function was concluded from clinical patients.


**Table 3.**Target prediction algorithm for microRNAs of human hypopharyngeal tumor model down-regulated by<sup>188</sup>Re-liposome.

a No direct evidence, but just an implication.

**Figure 3.** Validation of the next-generation sequencing (NGS) results of the HPC tumor model by qPCR analysis. The results were each microRNA of <sup>188</sup>Re-liposome-treated HPC normalized to that of untreated controls.

**Figure 4.** Comparison of the microRNA expression in clinical head and neck cancer (HNSCC) tissues and adjacent normal tissues by qPCR analysis. (**A**) MiR-206-3p. (**B**) MiR-378a-5p (**C**) MiR-142-5p. (**D**) Differentially expressed microRNA (DEmiRNA) in larynx tumor and normal tissues using the clinical information of The Cancer Genome Atlas (TCGA). \*: *p* < 0.05, \*\*: *p* < 0.01. \*\*\*: *p* < 0.001.

#### *2.5. Prediction of Genes Targeted by <sup>188</sup>Re-Liposome-Deregulated microRNAs*

We next examined the potent target genes that might be affected by <sup>188</sup>Re-liposome deregulated microRNAs. The miRDB online database was employed to rank the numbers of predicted genes targeted by microRNAs that were deregulated by <sup>188</sup>Re-liposome (see Materials and Methods). The ranking of target numbers affected by <sup>188</sup>Re-liposome up-regulated and down-regulated microRNAs was obtained, respectively (Figure 5A,B). Furthermore, miR-34a, miR-206-3p and miR-4510-5p were used to draw the Venn diagrams as their predicted target genes were ranked as the top three in <sup>188</sup>Re-liposome up-regulated microRNAs. Only one target, named *SLC44A2* encoded choline transporter-like protein 2 was co-regulated by these three microRNAs (Figure 5C). The same logic was used for <sup>188</sup>Re-liposome down-regulated microRNAs, and an only one target, named *U2SURP* encoded U2 snRNP-associated SURP motif-containing protein, was co-regulated by miR-142-5p, miR-194-5p and miR-944-3p (Figure 5D). Therefore, these bioinformatics analyses suggest that microRNAs and associated target genes of HPC tumors oppositely regulated by <sup>188</sup>Re-liposome are distinct.

**Figure 5.** Analysis of miRNA-target interactions. (**A**) The individual target number of microRNAs up-regulated by <sup>188</sup>Re-liposome. (**B**) The individual target number of microRNAs down-regulated by <sup>188</sup>Re-liposome. (**C**,**D**) The Venn diagram calculated and drawn by the three microRNAs with the most targets.

#### *2.6. Analysis of the Molecular Pathways Regulated by <sup>188</sup>Re-Liposome-A*ff*ected microRNA*

We next used the pathview R package to investigate the genes influenced by microRNAs regulated by <sup>188</sup>Re-liposome. MicroRNA samples exhibiting over two-fold change were selected, and the affected target genes were subjected to the KEGG pathway database for determining the potent molecular pathways disturbed by <sup>188</sup>Re-liposome. We found that thirty pathways in resected HPC tumors were significantly affected by <sup>188</sup>Re-liposome (*p* < 0.05). The top three pathways with the lowest p values were genes involved in olfactory transduction, pathways in cancer, and taste transduction (Table 4). Additionally, <sup>188</sup>Re-liposome-influenced pathways could be categorized as cancer and carcinogenesis, cell adhesion and cytoskeletal organization, drug metabolism via cytochrome P450, tumor suppression and oncogenes. An integrated cancer pathway involved in the KEGG database was shown to demonstrate the related genes that could be regulated by <sup>188</sup>Re-liposome-induced or -suppressed microRNA expression (Figure 5). This revealed that genes associated with cell cycle progression, proliferation, and apoptosis were affected by <sup>188</sup>Re-liposome, including the down-regulation of *cyclin D*, *cyclin E*, cyclin-dependent kinase (*CDK*) *4*/*6*, *E2F* transcription factor, and *bcl-2* anti-apoptotic factor, and the up-regulation of *p15*, *p16*, and *p27* cell cycle inhibitors and the *Rb* tumor suppressor gene (Figure 6). This pathway analysis provides a potent profile of the molecular mechanism for <sup>188</sup>Re-liposome regulated microRNA expression and tumor suppression of the HPC tumor model.

**Figure 6.** Use of the Kyoto Encyclopedia of Genes and Genomes (KEGG) database to display cancer pathway-associated genes affected by <sup>188</sup>Re-liposome mediated miRNA expression in HPC. Red: Up-regulated genes; green: down-regulated genes; yellow: unknown regulated genes; gray: unchanged genes.


**Table 4.**Thirty significant changed pathways after repeated doses of<sup>188</sup>Re-liposome treatment.

a DEG: Differentially expressed genes. b Only*p*<0.05 was counted.

#### *2.7. Association of <sup>188</sup>Re-Liposome-Regulated microRNA and Patients' Survival Rate*

As <sup>188</sup>Re-liposome could up-regulate tumor suppressive microRNA or down-regulate oncogenic microRNA of the HPC tumor model, we considered whether these microRNAs could be considered as prognostic factors for patients' survival rates. The microRNAs deregulated by <sup>188</sup>Re-liposome that displayed over five-fold change were examined. Notably, the online dataset only includes precursor forms of microRNAs for an analysis of patients' survival. For <sup>188</sup>Re-liposome up-regulated microRNAs in the HPC model, only miR-342 and miR-378a potentially exhibited an increase in the survival rate in HNSCC patients, with a marginal significance (Figure 7A,B). Interestingly, a high expression of miR-342 was significantly associated with a better survival rate in female patients (HR = 0.61, 95% CI = 0.38–0.98, *p* = 0.038). On the other hand, <sup>188</sup>Re-liposome down-regulated microRNAs, including miR-143, miR-6723, miR-944, and miR-136 were associated with a reduced survival rate in HNSCC patients when they were highly expressed (Figure 7C–F). Additionally, miR-3960 was also associated with reduced survival, yet the expression was too low for meaningful analysis (data not shown). These data suggest that specific miRNAs affected by <sup>188</sup>Re-liposome may be important for perspective clinical evaluation.

**Figure 7.** The association of <sup>188</sup>Re-liposome deregulated microRNAs with the survival rates of HNSCC patients. Kaplan–Meier (K-M) plot (**A**) hsa-miR-342; (**B**) hsa-miR-378a; (**C**) hsa-miR-143; (**D**) hsa-miR-6723; (**E**) hsa-miR-944; (**F**) hsa-miR-136.

#### **3. Discussion**

HPC mostly originates from mucosal squamous cells with a low incident rate. Because of unapparent early symptoms and a high metastatic ability, HPC accounts for the lowest survival rate of all head and neck cancers [72]. The FaDu cell line is a squamous cell carcinoma of the human hypopharynx commonly used for the study of molecular mechanisms of head and neck cancer in vitro and in vivo [73]. We have previously established an orthotopic tumor model using this cell line and found that *let-7* microRNA was associated with the therapeutic efficacy of <sup>188</sup>Re-liposome nanoparticles [24,74]. In this study, we used NGS analysis and found additional microRNAs that might be involved in the therapeutic efficacy of <sup>188</sup>Re-liposome after repeated administration. The expressions of the total RNA number and annotated small RNA were reduced in <sup>188</sup>Re-liposome treated tumors, suggesting that <sup>188</sup>Re-liposome would suppress gene transcription and expression.

As <sup>188</sup>Re is a high-energy β particles-emitter, it is expected to induce DNA damage. Indeed, the DNA damage marker γ H2AX was significantly induced by <sup>188</sup>Re-liposome compared to an untreated control. Interestingly, γ H2AX has been reported to be a tumor suppressor because of its role in the maintenance of genomic stability [75]. In our study, we found that <sup>188</sup>Re-liposome induced γ H2AX in HPC tumors. In the RNA-seq dataset, we also found that miR-138-2-5p, a potent inhibitor of H2AX [76], was down-regulated by over two-fold by <sup>188</sup>Re-liposome treatment. Together, these results are partially consistent with previous reports that might account for the therapeutic mechanisms of <sup>188</sup>Re-liposome from the viewpoint of γ-H2AX-mediated DNA damage responses.

According to the results of RNA-seq, <sup>188</sup>Re-liposome induced more than 200 microRNA to change their levels. To raise the selective criteria, we focused on the microRNA species exhibiting over five-fold up-regulation or down-regulation induced by <sup>188</sup>Re-liposome by comparing them to the untreated controls. The top three up-regulated microRNAs (miR-206-3p, miR-668-3p, and miR-485-3p) and down-regulated microRNAs (miR-142-5p, miR-944-3p, and miR-142-3p) displaying fold-change were categorized as tumor suppressors and oncogenes, respectively (Tables 2 and 3). Although miR-6723-5p was also highly suppressed by <sup>188</sup>Re-liposome, its role in cancer has not been interpreted in the dbDEMC database or microRNA Cancer association database (miRCancer) [77]. Most of the microRNAs deregulated by <sup>188</sup>Re-liposome were expressed oppositely in HNSCC, and they were reported to be potent tumor suppressors or oncogenes in different types of cancers. However, there were no available data for miR-4510-5p, miR-1268b-5p miR-6723-5p, miR-151b-3p, miR-143-5p, miR-194-2-5p and miR3960-3p in the database. As NGS is prominently used to find unknown genes that can be induced by treatment agents, the uncharacterized microRNAs shown to be significantly influenced by <sup>188</sup>Re-liposome treatment would be of interest for investigating their roles in the future.

The NGS analysis identified highly differentially expressed microRNAs that were also validated using FaDu tumors with or without <sup>188</sup>Re-liposome treatment, and clinical HNSCC samples using qPCR. The expression of analyzed microRNAs was consistent in these two different resources, although the case number of clinical samples was limited. The clinical information of the TCGA database could only be employed to analyze larynx cancer because the number of cases of HPC was too low to have normal tissues for analysis. Even so, we still found several microRNAs (e.g., miR-206-3p, miR-378-5p, and miR-142-5p) that were consistent with the results of NGS analysis and our clinical samples (Figure 4). In addition to <sup>188</sup>Re, <sup>177</sup>Lu is also a radionuclide that can emit 86% of β-particles and 14% of photons with a lower energy but longer half-life period. <sup>177</sup>Lu-octreotate has been reported to differentially regulate 57 specific microRNAs in mouse renal cortical tissue identified by the Mouse miRNA Oligo chip 4plex [78]. However, little microRNA was overlapped between their results and ours. Besides different types of radionuclides, the different animal models and microRNA mining methods may also account for the distinct observations in these two studies. Therefore, the differentially expressed microRNAs in the HPC model could be considered as specific prognostic factors for <sup>188</sup>Re-liposome treatment.

The predicted genes targeted by microRNAs were also analyzed by the miRDB public database. For the <sup>188</sup>Re-liposome up-regulated and down-regulated top three microRNAs, the *SLC44A2* gene

and *U2SURP* gene were the targets recognized by these oppositely regulated microRNAs, respectively. Hence, it is expected that *SLC44A2* would exhibit an oncogenic property and *U2SURP* should be a tumor suppressor gene. *SLC44A2*, first discovered in the inner ear, is a member of the choline transporter-like protein family of membrane transporter proteins [79,80]. *U2SURP* is involved in RNA splicing as it is part of spliceosomes [81]. However, little is known about the association of these two genes with human cancers. It would be an interesting target gene to investigate for its role in mediating the efficacy of <sup>188</sup>Re-liposome.

Using the KEGG pathway database, we found that 30 pathways in orthotopic HPC tumors were significantly influenced by <sup>188</sup>Re-liposome treatment. Although the cancer-suppression-related pathways were expected to be regulated by <sup>188</sup>Re-liposome, most of the affected genes in the annotated pathway displayed olfactory transduction. Radiation therapy has been reported to cause olfactory loss in head and neck cancer patients [82]. However, the conclusion is that radiation can damage olfactory cells. The position of hypopharyngeal cancer was not in the olfactory tract, yet FaDu cells were orthotopically injected into the buccal position of the mouse. This operation was based on the fact that FaDu cells are also buccal carcinoma cells [83]. Whether the microenvironmental difference influences the gene regulatory pathway of tumors is unclear. We believed that the excised tumor should not be contaminated by olfactory cells because the human tumor was very small after <sup>188</sup>Re-liposome treatment, that is, the tumor size was not big enough to reach the olfactory tract.

An assessment of the correlation between the gene expression and survival rate of patients is important for evaluating the clinical relevance of preclinical study for novel genes and drugs. Here we used the Kaplan–Meier plotter online tool that includes the datasets of TCGA program, the Gene Expression Omnibus (GEO) and the European Genome-Phenome Archive (EGA) to find the <sup>188</sup>Re-liposome regulated primary microRNAs and their association with patients' survival rates [84]. Although several potent tumor suppressive or oncogenic microRNAs were influenced by <sup>188</sup>Re-liposome, only part of these microRNAs exhibited the expected association with patients' survival probability in HNSCC. For instance, <sup>188</sup>Re-liposome suppressed miR-142 exhibited higher survival probabilities in HNSCC when they were expressed at higher levels, although they were expected to be oncogenic microRNAs [85]. A potent limitation is the sample size (523 cases) of HNSCC patients, which may be too small to draw conclusions on the role of microRNAs in patients' survival rates. Besides, the online K-M plotter only analyzes the effects of precursor microRNA on patients' survival rate. The association of mature miRNAs with the survival rate remains unknown. Whether different forms of the same microRNAs will differentially influence the results of the survival rate remains to be addressed.

In summary, current data suggest that the de-regulation of microRNAs correlates with the therapeutic efficacy of <sup>188</sup>Re-liposome on human HPC tumors. Using NGS, we also found several microRNAs that have not been fully characterized for their roles in cancer development and therapy. Whether these microRNAs are important for mediating the efficacy of <sup>188</sup>Re-liposome would be interesting to further investigate. Additionally, the KEGG pathway analysis showed that not only cancer pathways but also olfactory and taste transduction were significantly changed in HPC tumors after they were treated by <sup>188</sup>Re-liposome. Although no clinical evidence showing that patients treated with <sup>188</sup>Re-liposome will lose olfactory and gustatory sensation, olfactory sensory dysfunction and gustatory impairment often occur after patients are treated with radiotherapy in the head and neck area [82,86–89]. To the best of our knowledge, this is the first study uncovering the therapeutic mechanisms of <sup>188</sup>Re-liposome by an investigation of the pan-expression of microRNA. As <sup>188</sup>Re-liposome has entered the clinical trial stage, these data may further extend the concept of precise medicine using this radiotheranostic agent and allow the affected microRNAs to be prognostic factors after cancer treatment.

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

#### *4.1. Cell Lines and Plasmid*

Human FaDu HPC cells (American Type Culture Collection, Manassas, VA, USA) were cultured in RPMI-1640 (Life Technologies Inc., Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS) (Thermo Fisher Scientific Inc., Waltham, MA, USA), 1% penicillin (Sigma-Aldrich Co., St. Louis, MO, USA), and 1% L-glutamine (Sigma-Aldrich Co., St. Louis, MO, USA). FaDu-3R cells harboring multiple reporter genes were used and cultured as reported previously [24]. Cells were incubated at 37oC in a humidified incubator with 5% CO<sup>2</sup> and passaged every two days.

#### *4.2. Preparation of <sup>188</sup>Re-Liposome*

The procedure of <sup>188</sup>Re-liposomal preparation has been reported before [23]. In brief, <sup>188</sup>Re was milked from the <sup>188</sup>W/ <sup>188</sup>Re generator system (Institute National des Radioelements, Fleurus, Belgium) and conjugated with sodium perrhenate. Moreover, <sup>188</sup>Re was conjugated with *N*,*N*-bis(2-mercapatoethly)-*N*′ ,*N*′ -diethylenediamine (BMEDA, ABX GmbH, Radeberg, Germany), and the quality of <sup>188</sup>Re-BMEDA was validated by using the instant thin-layer chromatography (iTLC) followed by a radioactive scanner (Bioscan AR2000; Bioscan, TriFoil Imaging Inc., Chatsworth, CA, USA). Furthermore, PEGylated liposome (NanoX; Taiwan Liposome Co. Ltd., Taipei, Taiwan) was used to encapsulate <sup>188</sup>Re-BMEDA and eluted using the PD-10 column (GE Health BioSciences, Pittsburgh, PA, USA) (Supplementary Figure S1). The average molecular weight of polyethylene glycol (PEG) was 2000. The particle size (84.6 ± 4.12 nm) and surface charge (1.1 ± 1.9 mV) were measured by the dynamic light scattering apparatus (Zetasizer Nano ZS90, Malvern Panalytical Ltd., Malvern, UK). The in vitro stabilities of <sup>188</sup>Re-liposome in normal saline and rat plasma were, respectively, over 92% and 82% in 72 h as reported before [20].

#### *4.3. Establishment of HPC Orthotopic Tumor Model for Evaluating the Therapeutic E*ffi*cacy of <sup>188</sup>Re-Liposome*

Six-week-old male BALB/c nude mice (*N* = 5 for each experimental group) were purchased from National Laboratory Animal Center, Taipei, Taiwan and used for the establishment of orthotopic HPC tumor model. FaDu-3R cells (1 × 10<sup>6</sup> ) were resuspended in 50 µL of OPTI-MEM (Sigma-Aldrich, St. Louis, MO, USA) and then injected into the buccal position of each mouse at right side using a 27 G insulin needle. For intravenous injection of <sup>188</sup>Re-liposome, 23.68 MBq (640 µCi) corresponding to 80% maximum tolerated dose (MTD), as we mentioned before [23]. To evaluate the therapeutic efficacy, the tumor viability and growth rate were measured using the luciferase reporter gene imaging and caliper measurement. The luminescent signals were acquired by the In Vivo Imaging System (*Optima*, Biospace Lab Inc., Paris, France). The tumor volume was calculated by the formula: (width<sup>2</sup> × length)/2 after caliper measurement every three days [90]. The animal experiments were approved by the Institutional Animal Care and Utilization Committee (IACUC) of National Yang-Ming University (No. 1061010).

#### *4.4. Tumor Collection and Next-Generation Sequencing (NGS)*

Tumors were harvested from tumor-bearing mice after four weeks of <sup>188</sup>Re-liposome treatment. Total RNA of both saline control and <sup>188</sup>Re-liposome treated group were extracted using the QIAGEN RNA mini kit (Thermo Fisher Scientific Inc., Waltham, MA, USA) according to manufacturer's instructions. Furthermore, the quality of RNA was detected using the Nanodrop spectrophotometer (Nanodrop Technologies LLC, Wilmington, DE, USA). The integrity and concentration of RNA samples were determined using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) with RNA 6000 nano kit (Agilent Technologies, Santa Clara, CA, USA). The TruSeq Small RNA Library kit (Illumina, Inc., San Diego, CA, USA) was then used to ligate RNA with adapters followed by the reverse transcription-PCR to generate cDNA library. The library was then sequenced by the HiSeq

4000 Sequencing System (2 × 150 bp paired-end Sequencing) and the results were processed with the Illumina software (Illumina, Inc., San Diego, CA, USA). To qualify the results of small RNA sequencing, the sequences were applied to the CLC Genomics Workbench v10 to obtain the qualified reads [91]. CLC Genomics Workbench counts different types of small RNAs in the data and compares them to databases of microRNAs or other small RNAs [92].

#### *4.5. MicroRNA Expression Analysis*

The 'miRBase' online source was used for the annotation of small RNA [93]. In the next step, all of the miRNAs were normalized by TMM (trimmed mean of M values) method using edgeR (R package: v.3.10.5) (Bioconductor, New York City, NY, USA) [94]. For a two-group experiment, we used the 'Fold Change' to tell how many times bigger the mean expression value in <sup>188</sup>Re-liposome group is relative to that of saline-only group. If the mean expression value in <sup>188</sup>Re-liposome group is small than that in saline-only group, the value will present negative sign, and vice versa. The criteria for microRNAs selection were fold change >5.

#### *4.6. Western Blot Analysis*

Tumors were collected from the tumor-bearing mice after 4 weeks of <sup>188</sup>Re-liposome treatment, and lysed in T-PERTM Tissue Protein Extraction Reagent (Thermo Fisher Scientific, Waltham, MA, USA) containing 1% protease inhibitor cocktail (Sigma-Aldrich Co., St. Louis, MO, USA). Protein lysates (50 µg) were run on 8–12% SDS-PAGE, electro-transferred to nitrocellulose membrane, blocking and incubated with antibody as reported previously [95]. The primary antibodies were anti-γH2AX (GTX628789, GeneTex Inc., Irvine, CA, USA), and anti-glyceraldehyde3-phosphate dehydrogenase (GAPDH) (MA5-15738, Invitrogen Inc., Carlsbad, CA, USA).

#### *4.7. Validation of microRNA Expression Using qPCR*

To validate miRNA levels before and after HPC tumor treated with <sup>188</sup>Re-liposome and to compare normal tissue to HNSCC tissues using clinical samples, quantitative PCR (qPCR) of targeted miRNA was performed. Briefly, complementary DNA (cDNA) was generated from 2 µg total RNA using SuperScript II reverse transcriptase (Life-Technologies Co., Carlsbad, CA, USA). Then, the cDNA products were mixed with the Fast SYBR Green Master Mix (Life-Technologies Co., Carlsbad, CA, USA) and subjected to the StepOnePlus Real-time PCR System (Life-Technologies Co., Carlsbad, CA, USA) following the manufacturer's instructions. The sequences of stem loop primers, forward primers and universal primer used for miR-206-3p, miR-382-5p, miR-378a-5p, miR-3960-3p, and miR-142-5p amplification were summarized in Table 5. Use of human tissue samples was approved by the Institutional Review Board (No. 2019-01-010BC).

#### *4.8. Heatmap Analysis of NGS Data*

To gain the heatmap of total microRNAs' expression, 'TreeView' (v1.1.6r4) was used [96]. The heatmap analysis was based on the small RNA that has twice difference between individuals with or without the treatment of <sup>188</sup>Re-liposome, and then took Log2 value to draw out.

#### *4.9. Analysis of microRNA Using the Cancer Genome Atlas (TCGA)*

The expression levels of miRNAs and clinical information for TCGA Head-Neck Squamous Cell Carcinoma (HNSC) were downloaded from Broad GDAC Firehose [97]. The miRNA expression values for samples having anatomic subdivision labeled as normal larynx tissue (12 cases), larynx cancers (117 cases) or hypopharynx cancers (10 cases) were used. In house R scripts were used to parse and generate heatmap and boxplots [98]. The difference between normal and tumor samples were tested with Wilcoxon signed-rank test.


**Table 5.** The list of stem loop primers, forward primers, and reverse primer used for qPCR of microRNAs.

#### *4.10. Characterization of miRNAs*

The miRNA of interests were subjected to a Database of Differentially Expressed MiRNAs in human Cancer 2.0 (dbDEMC v2.0) that collects the data sets of Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) to exhibit differentially expressed miRNAs in human cancers detected by high throughput method [99]. The roles of miRNAs belonging to tumor suppressor genes or oncogenes were predicted accordingly. The miRDB online database for prediction of miRNA targets by a MirTarget bioinformatics tool was used to analyze putative downstream genes influenced by the miRNAs of interests [100,101]. Each microRNA was selected for target expression analysis for FaDu cells, and the targets expression level over 20 was counted as they were most relevant to FaDu cells. Target expression level was determined by RNA-Seq using the RPKM method (Reads Per Kilobase of transcript, per Million mapped reads). An online Venn diagrams drawing tool was exploited to calculate the intersections of list of miRNA targets [102].

#### *4.11. The Pathway Analysis*

The 'pathview' (R package: v1.4.2) software (Bioconductor, New York City, NY, USA) was used to draw pathways and find significant gene change (Fold change > 2) with the Kyoto Encyclopedia of Genes and Genomes (KEGG) [103,104]. Moreover, the pathways would be regarded significant if the *p*-value < 0.05.

#### *4.12. Statistical Analysis*

Statistical analysis was performed using GraphPad Prism 6.0 (GraphPad Software, San Diego, CA, USA). All data were represented as the means ± standard deviation (SD) with independent experiments. The Student's *t*-test was used for statistical analysis. Two-way Analysis of Variance (ANOVA) was used to compare the tumor growth curves. The Kaplan–Meier method with the log-rank test was used to analyze the association of miRNAs and patients' survival rates using the online K-M plotter with public datasets [84].

**Supplementary Materials:** The following are available online, Figure S1: Measurement of the combined efficiency between <sup>188</sup>Re and BMEDA chelator for embedding into liposome. Figure S2: Demonstration of reporter genes expression in FaDu-3R cells. Figure S3: Qualification of total RNA for NGS analysis. Figure S4: GC content analysis and sequencing qualification of RNA–Seq. Figure S5: A heatmap of differentially expressed microRNAs with or without <sup>188</sup>Re-liposomal treatment. Figure S6: A heatmap of selected differentially expressed microRNAs in normal tissues and larynx cancer tissues using the clinical information of TCGA.

**Author Contributions:** Conceptualization, Y.-J.L. and S.-Y.W.; methodology, B.-Z.L.; software, T.-W.C.; validation, B.-Z.L., M.-Y.L. and Y.-J.L.; formal analysis, B.-Z.L.; investigation, B.-Z.L.; resources, C.-H.C. and M.-H.Y.; data curation, M.-Y.L.; writing—original draft preparation, Y.-J.L.; writing—review and editing, Y.-J.L.; supervision, Y.-J.L. and S.-Y.W.; funding acquisition, Y.-J.L. and S.-Y.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by National Yang-Ming University-Far Eastern Memorial Hospital Joint Research Program, grant number 108DN04; Ministry of Science and Technology grant number 105-2623-E-010-001-NU, 106-2623-E-010-002-NU, and 108-2314-B-010-016) and "The APC was funded by National Yang-Ming University-Far Eastern Memorial Hospital Joint Research Program".

**Acknowledgments:** We thanked Shen-Nan Lo, Ming-Hsuan Lin for helping the <sup>188</sup>Re preparation, production, and quality assurance. We thank Liang-Ting Lin for discussion of references. We also thank the Taiwan Animal Consortium (MOST 106-2319-B-001-004)—Taiwan Mouse Clinic which is funded by the Ministry of Science and Technology (MOST) of Taiwan for technical support in in vivo imaging experiments. We also thank the support from the Cancer Progression Research Center, National Yang-Ming University, from the Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan.

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

#### **References**


**Sample Availability:** Samples of the compounds BMEDA and liposomes are available from the authors.

© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

*Review*

## **Pyrazoles as Key Sca**ff**olds for the Development of Fluorine-18-Labeled Radiotracers for Positron Emission Tomography (PET)**

## **Pedro M. O. Gomes , Artur M. S. Silva and Vera L. M. Silva \***

LAQV-REQUIMTE, Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal; pm.gomes@ua.pt (P.M.O.G.); artur.silva@ua.pt (A.M.S.S.) **\*** Correspondence: verasilva@ua.pt; Tel.: +351-234-370714

Academic Editor: Krishan Kumar

Received: 15 March 2020; Accepted: 8 April 2020; Published: 9 April 2020

**Abstract:** The need for increasingly personalized medicine solutions (precision medicine) and quality medical treatments, has led to a growing demand and research for image-guided therapeutic solutions. Positron emission tomography (PET) is a powerful imaging technique that can be established using complementary imaging systems and selective imaging agents—chemical probes or radiotracers—which are drugs labeled with a radionuclide, also called radiopharmaceuticals. PET has two complementary purposes: selective imaging for diagnosis and monitoring of disease progression and response to treatment. The development of selective imaging agents is a growing research area, with a high number of diverse drugs, labeled with different radionuclides, being reported nowadays. This review article is focused on the use of pyrazoles as suitable scaffolds for the development of <sup>18</sup>F-labeled radiotracers for PET imaging. A brief introduction to PET and pyrazoles, as key scaffolds in medicinal chemistry, is presented, followed by a description of the most important [ <sup>18</sup>F]pyrazole-derived radiotracers (PET tracers) that have been developed in the last 20 years for selective PET imaging, grouped according to their specific targets.

**Keywords:** positron emission tomography (PET); pyrazoles; fluorine-18; radionuclides; PET probes; imaging pharmaceuticals

#### **1. Introduction**

Positron emission tomography (PET) is a nuclear medicine functional imaging technique that uses gamma rays (formed as a result of the annihilation of the emitted positrons) to provide three-dimensional images that give information about the functioning of a person's specific organs. PET is based on the detection of tiny amounts (picomolar) of a biological substance labeled with a short-lived positron-emitting radionuclide (PET tracer) without disturbing the biological system and has the advantage of being a noninvasive, functional and highly sensitive technique which provides a great wealth of information [1]. The PET tracers are drugs or biomolecules labeled with radionuclides such as <sup>11</sup>C, <sup>13</sup>N, <sup>15</sup>O, <sup>18</sup>F, <sup>64</sup>Cu, <sup>68</sup>Ga, <sup>89</sup>Zr and <sup>124</sup>I. Since radionuclides can replace the stable analogues, the PET probes have the same chemical structure as the drugs and biomolecules without altering their biological activity. The choice of the PET radionuclide must follow some criteria: (i) physical and chemical characteristics, (ii) availability and (iii) timescale of the biological process in the study. For different biological processes, some radionuclides are better than others. If the biological process in the study gives results hours or days after the injection of PET probe, <sup>64</sup>Cu, <sup>89</sup>Zr or <sup>124</sup>I can be used, because these three radionuclides have a long half-life time (12.8 h, 78.4 h and 4.17 days, respectively). On the contrary, radionuclides as <sup>11</sup>C or <sup>18</sup>F are used for labeling small organic compounds for faster biological processes [1]. To avoid unnecessary radiation, it is important to choose a radionuclide

with a half-life that enable the reaction of the radionuclide with the carrier molecule and matches the timescale of the biological process in the study. <sup>18</sup>F is the ideal radionuclide for routine PET imaging, because it has a short half-life (109.8 min) but enough to allow all the process of synthesis, transport and imaging. Furthermore, it is a 97% positron-emitter, and the low positron energy of <sup>18</sup>F leads to a high resolution of PET imaging [1,2]. ‐ ‐ ‐

In the last three decades, the interest in PET has been growing, and nowadays, this technique is used in several areas of medicine, such as oncology [3], cardiology [4], neurology [5] and also in drug development and evaluation [6]. Besides the aforementioned advantages of PET, there are also some limitations; pregnant women should not undergo in PET imaging; the radioactivity of PET tracers limits the number of times one patient can undergo PET scans and is a very expensive treatment [7]. Moreover, the potential of PET strongly depends on the availability of suitable PET radiotracers. For instance, 2-[18F]fluoro-2-deoxyglucose ([18F]FDG) is one of the most used radiotracers for the clinical diagnosis and evaluation of cancer. However, it has inherent drawbacks, namely its high accumulation in inflamed and infected tissues giving false-positive results and its low uptake in tumors that grow slowly, which can lead to false-negative results [8]. Finally, the development of new PET-imaging probes is not trivial, and radiochemistry is a major limiting factor for this field. Hence, the research and development of new PET tracers is a challenging issue of major importance within the scientific community. ‐ ‐ ‐ ‐ ‐ ‐

Following our interest in the synthesis and biological evaluation of pyrazole-type compounds envisaging their use in medicine [9–14] and considering their huge potential for application in PET field, herein we present an overview of pyrazole-derived PET radiotracers developed in the last 20 years. Pyrazoles labeled with <sup>18</sup>F radionuclide will be the focus of this review article, given the aforementioned advantages of this radionuclide for routine PET imaging. ‐ ‐

#### **2. Pyrazoles**

This family of compounds is characterized by a simple heteroaromatic five-membered ring containing three carbon and two nitrogen atoms in adjacent positions [15]. The occurrence of the pyrazole core in nature is rare, probably due to the difficulty of living organisms to make a N-N bound [16]. Until now, there are only approximately 20 natural compounds with a pyrazole core on their structures [17]. Although scarce in nature, pyrazole (1*H*-pyrazole (**1**), Figure 1) and its reduced derivatives (pyrazolines **2**–**4**, Figure 1) are considered privileged structures in medicinal chemistry. Pyrazole scaffold is present in several synthetic drugs, being celecoxib (Celebrex®), sildenafil (Viagra®), rimonabant, fomepizole, penthiopyrad and sulfaphenazole the most notorious [15]. ‐ ‐ ‐

**Figure 1.** Chemical structures and numbering of pyrazole (**1**) and dihydropyrazole (pyrazoline) tautomers **2**–**4** [18].

‐ ‐ ‐ ‐ Pyrazole derivatives act on diverse and relevant biological targets, which make them attractive for the development of PET tracers, and possess a wide range of pharmaceutical activities. Abdel-Maksoud et al. have demonstrated antitumoral activity of several compounds with a pyrazole and sulfonamide moieties [19]; antibacterial and antifungal activity of pyrazole was demonstrated by Chowdary et al. [20]. Pyrazoles also showed good activity as monoamine oxidase inhibitors, antidepressant and anticonvulsant agents [21], as BRAF inhibitors [22] and DNA gyrase inhibitors [23] and present antileishmanial, anti-inflammatory, analgesic, antidiabetic and cannabinoid activity; cyclin-dependent kinase and tissue-nonspecific alkaline phosphatase inhibitory activity and moderate

‐

antihepatotoxic activity, among others [9,11,24–29]. In 2013, Prabhu et. al. reported the antioxidant activity of pyrazole [30]. They demonstrated that pyrazole (1,2-diazole) can be used to prevent nephrotoxicity caused by cisplatin, a drug used to treat several cancers. Cisplatin provokes renal damage because of its toxicity to proximal tubule cells and can reduce glomerular filtration, resulting in renal failure. One of the reasons of nephrotoxicity induced by cisplatin is the decreasing concentration of glutathione (GSH). Pyrazole prevents nephrotoxicity induced by cisplatin by counteracting this effect, increasing the concentration of this enzyme [30]. Recently, Silva et al. reported the antioxidant activity of around three hundred pyrazoles [18].

#### *2.1. Pyrazoles as Probes for PET Imaging*

The labeling of pyrazole-type compounds with <sup>18</sup>F for PET imaging is a research issue that has been growing significantly, as evidenced by the number of papers and citations in the last years, according to our search on Web of Science using the keywords (pyrazole) and (positron emission tomography) and (18F) (Figure 2). In this section are described the most important [18F]-labeled pyrazoles developed for use as PET radiotracers. The compounds are presented according to their specific targets, which are indicated by alphabetical order and not by their relevance in the PET field. Whenever possible, the most promising applications of these compounds in this field are discussed. ‐ ‐

**Figure 2.** Number of papers published and citations per year found using the keywords "pyrazoles" and "positron emission tomography" and "18F" in Web of Science in the period 2000–2020.

#### 2.1.1. Adenosine Receptors Ligands

‐ ‐ ‐ ‐ ‐ ‐ ‐ Adenosine, an endogenous-signaling substance, is a purine ribonucleoside composed of adenine (purine base) and ribose (sugar molecule), which is produced in response to metabolic stress and cell damage. It induces several physiopathological effects, regulating the central nervous, cardiovascular, peripheral, and immune systems due to the rapid generation of adenosine from cellular metabolism and the widespread distribution of its receptor subtypes in almost all organs and tissues [31]. There are four types of adenosine receptors: A1, A2A, A2B and A3. A2AR are abundant in dopamine-rich regions of the brain, being the striatum the region with a higher concentration of A2AR [31]. In preclinical studies, A2AR antagonists showed potential benefits in the treatment of some neurodegenerative diseases like Alzheimer's disease (AD) and Parkinson's disease (PD), neuroinflammation, ischemia, spinal cord injury, drug addiction and other conditions [31]. Khanapur et al. developed a pyrazolo[4,3-*e*]-1,2,4-triazolo[1,5-*c*]pyrimidine labeled with carbon-11, [11C]SCH442416 (**5**) and its [18F]fluoroethyl and [18F]fluoropropyl derivatives ([18F]FESCH (**6**) and [18F]FPSCH (**7**), respectively) (Figure 3), and both radioligands showed a distribution in the rat brain corresponding to the regional A2AR densities, as evidenced by in vitro autoradiography (ARG) experiments and binding assays [32]. These two tracers showed a similar ratio specific/nonspecific binding, using the striatum as the specific binding and cerebellum as the nonspecific binding (4.6 at 25 and 37 min after injection, respectively) and reversible binding in rat brains, although their kinetics were slightly different [32]. Recently, the same authors studied the full kinetics of radioligands **6** and **7** in rat brains. On the basis of the Akaike information criterion, they have found that 1TCM was the most appropriate model for describing [18F]FPSCH kinetics, whereas 2TCM was the most suitable model for [18F]FESCH kinetics. Using dynamic PET imaging, under baseline and full blocking conditions, they proved that **6** is the most suitable PET radioligand for quantifying A2AR receptor expression in the rat brain. However, before starting the clinical use of **6**, it will be necessary to reevaluate the brain uptake in humans due to possible interspecies differences in tracer kinetics and metabolisms [33].

**Figure 3.** Adenosine receptor (A2AR) radioantagonists **5**–**7** for PET imaging [32,33].

#### 2.1.2. Cannabinoid Receptors Ligands

‐ ‐ Products derived from *Cannabis sativa* are some of the oldest and widely used drugs in the world. These products are known as natural cannabinoids, but several synthetic cannabinoids have been developed as well. Cannabinoids have been used as analgesics for more than 100 years [34]. Additionally, they have been used as antiemetic agents to prevent chemotherapy-induced nausea and vomiting, because they can bind to opiate receptors in the forebrain, blocking the vomiting center in the medulla [35]. In 1998, Williams and Kirkham have demonstrated that anandamide, an endogenous cannabinoid, provokes hyperphagia in satiated rats [36]. Cannabinoid-type compounds bind to cannabinoid receptors, which can be divided in two groups—cannabinoid receptor type 1 (CB1), predominantly found in the brain, and the peripheral cannabinoid receptor type 2 (CB2), mainly expressed in immune tissues, and both are G-protein-coupled membrane receptors [37].

‐ ‐ The imaging of the CB<sup>1</sup> receptor is of great importance for studying its role in neuropsychiatric disorders, including depression, and in obesity, drug or alcohol addiction, and is an active target for in vivo imaging development [38]. The first selective CB<sup>1</sup> receptor antagonist was

‐ ‐ ‐

[*N*-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1*H*-pyrazole-3-carboxamide] SR141716 (**8**), known as rimonabant (Figure 4) [39]. This compound was approved in Europe, in 2006, to treat obesity by reducing the patient's appetite. Two years later, the European Medicines Agency (EMA) withdrew rimonabant from sale due to its evident secondary effects. Some studies demonstrated that SR141716 induced anxiety, depression, agitation, eating disorders, irritability, aggression and insomnia. Rimonabant was not approved in the USA by the Food and Drug Administration (FDA) [40]. However, some analogs of SR141716 (**8**) have been labeled with radionuclides for PET imaging. [ <sup>18</sup>F]SR144385 (**9**) and [18F]SR147963 (**10**) [37,41], [18F]NIDA-42033 (**11**) and its related ester derivative (**12**) [42], [18F]O-1302 (**13**) [43] and [18F]AM5144 (**14**) [44] are some examples of PET tracers labeled with <sup>18</sup>F (Figure 4), while [123I]AM251 (**15**) and [123I]AM281 (**16**) [44] are two examples of PET tracers labeled with <sup>123</sup>I (Figure 5).

**Figure 4.** Structure of rimonabant (**8**) and structurally related [18F]pyrazole PET tracers **9**–**14** [37,39,41–44].

‐

‐

‐ **Figure 5.** Pyrazole-derived PET tracers **15** and **16** labeled with <sup>123</sup>I [44].

‐ Studies were performed to evaluate the specificity of cannabinoids [18F]SR144385 (**9**) and [ <sup>18</sup>F]SR147963 (**10**) (Figure 4) for CB<sup>1</sup> receptors [37]. After 15 min of injection in male CD-1 mice (25–30 g), compound **10** showed a higher brain uptake compared to compound **9** (5.70% ID/g and 3.06% ID/g, respectively). The hippocampus was the brain region that exhibited the highest uptake of both tracers, followed by the striatum, cerebellum, frontal cortex, cortex, olfactory tubercles and hypothalamus. Both the brain stem and thalamus showed low uptakes of the tracers, and the thalamus showed the fastest decrease of %ID/g. These results are in accordance with the knowledge of CB<sup>1</sup> receptors' brain density in rats, which was found to be lower in the thalamus region. For evaluation of the target to nontarget ratio, the thalamus was used as indicator of nonspecific binding. Compounds **9** and **10** showed differences in the post-injection time they reached the maximum target to nontarget ratio (90 min for compound **9** with a ratio of 2.50 and 60 min for compound **10** with a ratio of 1.69) [37]. In vivo selectivity and specificity studies of [18F]SR144385 (**9**) showed that significant blocking of this tracer uptake was achieved when a 1-mg/kg dose of the structurally similar blocking agent SR141716 (**8**) was given 30 min prior to the radiotracer. Likewise, the uptake of [18F]SR147963 (**10**) at either 30 or 60 min post-injection was also blocked by a 1-mg/kg dose of SR141716 (**8**) given 30 min before the radiotracer.

In 2005, Li et al. synthetized the radioactive pyrazole [18F]AM5144 (**14**) (Figure 4) as a PET radioligand candidate for cannabinoid CB<sup>1</sup> receptors. Using baboons, they demonstrated that the highest radioactivity concentration of compound **14** was found on the cerebellum, and the lowest concentration was found in the thalamus. They concluded that there was a poor brain uptake of **14** and of other related [18F]pyrazoles because of their high lipophilicity, although this property is not the key factor in brain uptake. This study proved that there is no close relationship between the clogP value and brain uptake, indicating that there are other uptake key factors [44]. Despite the high in vitro-binding affinity and moderate lipophilicity of [18F]O-1302 (**13**) (Figure 4), it is not suitable for imaging the CB<sup>1</sup> receptor in the brain due to poor brain entry and high levels of nonspecific binding [43]. Although radiotracer uptake in the brain is often considered a function of the log P that peaks between a log P of 2 and 3, the investigation of the factors that affect brain uptake beyond lipophilicity is crucial to a better comprehension of this process. Some studies highlighted that there may be large species differences in brain penetration for a given PET radiotracer; for example, the brain uptake of [123I]AM251 (**15**) (Figure 5) is very different comparing mice and monkeys, being higher for mice [44,45].

Chang et al. observed that [18F]DBPR211 (**17**) (Figure 6) was distributed in the brain, liver, heart, thigh muscle and kidney when intravenously injected in mice. In the brain, it was found in a very low percentage (less than 1%) over 90-min scans. This result is evidence of its action as a peripherally restricted CB<sup>1</sup> antagonist [46]. Taking into account these observations, this radiotracer can be seen as a promising tool to study metabolic processes associated with peripheral CB<sup>1</sup> receptors.

‐ **Figure 6.** [ <sup>18</sup>F]pyrazole-derived peripheral CB<sup>1</sup> antagonist **17** [46].

#### 2.1.3. Cyclooxygenase Inhibitors

‐ ‐ ‐ ‐ ‐ ‐ μ ‐ ‐ μ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ Cyclooxygenase-2 (COX-2) is an enzyme with high-level expressions at sites of inflammation and cancer and is also a promising target for neuroinflammation imaging [47,48]. The development of agents capable of monitoring COX-2 levels is highly desirable for cancer prevention and therapy. To achieve this goal, Uddin et al. synthetized several indomethacin and celecoxib derivatives and evaluated their IC<sup>50</sup> values for the inhibition of COX-2 in vivo and in intact cells [49]. Indomethacin derivatives were effective COX-2 inhibitors in intact cells (0.09–0.26 µmol/L), but the synthesis of the respective radiotracer has some drawbacks, because the *p*-chlorobenzoyl group is not stable in the conditions of radiochemical synthesis. Among celecoxib derivatives, the more effective COX-2 inhibitor was compound **18** with an IC<sup>50</sup> value of 0.16 µmol/L. After the synthesis of <sup>18</sup>F-**18** (Figure 7), in vivo biodistribution was studied in the inflamed and contralateral footpad of male Sprague Dawley rats. The major advantage of this inflammation model is the possibility to image the inflamed footpad in comparison with the noninflamed contralateral footpad. Compound **18** had a higher accumulation in the inflamed footpad in comparison with the noninflamed. The selective binding between **18** and COX-2 was proved with an assay using celecoxib to block COX-2 before the injection of **18.** To confirm the COX-2 specificity of **18**, COX-2-null mice were injected with carrageenan to promote inflammation. COX-2-null feet demonstrated a 1.08 ± 0.09 ratio of inflamed/noninflamed feet. On other hand, the wild-type mice group, used as the control, had an uptake ratio of 1.48 ± 0.04. The ability of **18** to target COX-2 in human tumor xenografts was also demonstrated using 1483 HNSCC cells and HCT116 cells injected in the left and right hip of female nude mice. In the COX-2-null HCT116 tumor cells, the uptake of **18** was minimal, while, in the 1483 HNSCC cells, a significative uptake of the radiotracer was observed. Once again, to prove that the difference in the uptake of **18** was related with the expression of COX-2, mice bearing 1483 xenografts were pretreated with celecoxib to block the active site of COX-2. There was a significative lower uptake of **18** for mice pretreated with celecoxib when compared with the control group [49].

‐ ‐ ‐ ‐ ‐ Lebedev et al. developed a new, high-affinity <sup>18</sup>F-COX-2 inhibitor **19** (Figure 7) that is radiolabeled directly on a heteroaromatic ring with the purpose to increase biodistribution and metabolic stability [50]. In vitro studies demonstrated a clear correlation between COX-2 expression and uptake of this radiotracer. Moreover, pharmacokinetic studies in healthy mice revealed no bone retention or defluorination within 2 h of injection, significant blood clearance, since the molecule is excreted from blood within an hour mainly through the hepatobiliary excretion pathway, crossing of blood-brain barrier (BBB) and no significant metabolites in major organs. Although these properties make this probe ideal for PET imaging, some aspects related to radiochemical synthesis severely limit the application of this compound as a probe. In fact, the use of Et4NF+4HF has limited the specific activity to 3 Ci/mmol,

‐

and the current decay-corrected radiochemical yield of 2%, although enough for preclinical studies and the production of a single patient dose, need further improvement to achieve the successful use of this compound as a PET tracer.

‐ ‐ **Figure 7.** [ <sup>18</sup>F]-labeled COX-2 inhibitors **18** and **19** [49,50]. ‐ ‐

‐ ‐ ‐ ‐ ‐ ‐ ‐ Aiming to develop selective probes for the two COX enzyme subtypes, COX-1 and COX-2, McCharty et al. have prepared [18F]SC63217 (**20**) and [18F]SC58125 (**21**) (Figure 8), starting from the corresponding nonradioactive COX-1 and COX-2 selective inhibitors, SC63217 and SC58125, respectively [51]. Both compounds are structurally similar, presenting only a different substituent on only one aromatic ring. In vitro binding studies of both compounds, using J774 macrophages, revealed that compound **20** is a potential probe for COX-2, while **21** was not a good marker of COX-1 due to high nonspecific binding. In vivo studies showed that, for each tracer, rat biodistribution was well-matched with the known distribution of these enzymes [51]. ‐ ‐ ‐ ‐ ‐ ‐ ‐

‐ ‐ ‐ **Figure 8.** [ <sup>18</sup>F]-labeled pyrazoles as probes for COX-1 **20** and COX-2 **21** [51].

#### ‐ ‐ ‐ 2.1.4. Dopamine Receptors Ligands

‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ The subfamily of D2-like dopamine receptors includes the D2-, D3- and D4-receptor subtypes and mediates the action of dopamine in the brain by the inhibition of adenylate cyclase activity. While the distribution of D2- and D3- receptors is well-known, there are still some uncertainties regarding D4-receptor (D4R) expression. If D4R selective radioligands are available, PET can be suited to gain deeper knowledge into the distribution and pathophysiological role of D4R in humans.

‐ ‐ ‐ ‐ ‐ *‐* ‐ ‐ ‐ ‐ ‐ *‐* In 2008, the pyrazolo[1,5-*a*]pyridine derivatives, FAUC 113 (**22**) and FAUC 213 (**23**) (Figure 9), bearing a (4-chlorophenyl)piperazinylmethyl moiety in the 3 and 2 positions of the pyrazolo[1,5-*a*]pyridine core, were synthetized. After an evaluation of their binding affinities for D4-like dopamine receptors, **22** and **23** were chosen as lead compounds for the development of [18F]-labeled PET tracers. A novel compound with a fluoroethoxyphenyl substituent in the *para*-position of the phenylpiperazinyl moiety showed the highest specificity to the D<sup>4</sup> receptor. Due to these results, the next step was radiochemical synthesis to prepare the same labeled derivative, [18F]FAUC F41 (**24**) (Figure 9). Using coronal rat brains, in vivo AGR studies were performed to evaluate the specific binding of **24** to D4R. An increased uptake of this compound was detected in the hippocampus, hypothalamus, cortex, medial habenular nucleus and central medial thalamic nucleus. The observed

binding pattern was mainly consistent with the known D4R distribution in the rat brains. The log P value for this compound was found to be 2.9, which may indicate adequate BBB penetration. Moreover, this ligand revealed high stability in human serum, even after long-term incubation for up to 90 min [52]. These results demonstrate that [18F]FAUC F41 (**24**) represents a potential radioligand for studying the D4R in vivo by PET imaging. ‐ ‐

‐ ‐ **Figure 9.** [ <sup>18</sup>F]-labeled high-affinity dopamine receptor (D4R) tracers **22**–**24** [52].

‐ ‐

‐ β ‐ ‐ ‐ With the aim to develop a radiotracer with high selectivity and favorable lipophilicity for imaging of the D3-receptor in the brain, Stöβel et al. have synthetized compound **26** as the radioactive [ <sup>18</sup>F]-labeled analogue of the D<sup>3</sup> ligand FAUC 329 (**25**) (Figure 10) [53]. In vitro AGR studies showed that **26** successfully visualized D3-rich brain regions, including the islands of Calleja. However, in vivo PET imaging revealed that it does not significantly accumulate in the CNS structures of rat brains, probably due to a low BBB penetration. Instead, a significant uptake occurred in the brain ventricular system, due to a significant penetration of this compound in the blood-liquor barrier and, more noticeable, in the pituitary gland, outside the BBB [53]. The results of PET studies also suggest that the main mode of action of FAUC 329, in vivo, could be due to binding to the dopamine receptors in the pituitary gland. ‐ β ‐ ‐ ‐

**Figure 10.** Radiotracer **26** derived from the dopamine D<sup>3</sup> receptor ligand FAUC 329 (**25**) [53].

#### 2.1.5. Glucocorticoid Receptor Ligands

‐ ‐

‐ ‐

‐ ‐ ‐ In 1998, Hoyte et al. reported a series of aryl-pyrazolo steroids similar to the potent glucocorticoid cortivazol **27** and evaluated the affinity of these compounds for glucocorticoid receptors (GRs). Among them, the fluoro analog **28,** which showed good binding and was a very potent glucocorticoid, was labeled with <sup>18</sup>F to be used as a glucocorticoid receptor-mediated imaging agent for PET (Figure 11) [54].

‐

′‐ ‐ ‐ ‐ ‐ ‐ ‐ β α‐ ‐

′‐ ‐ ‐ ‐ ‐ ‐ ‐ β α‐ ‐

**Figure 11.** Cortivazol (**27**) and related [18F]arylpyrazolo steroids **28**–**29** as potential glucocorticoid receptor ligands for PET imaging [54,55].

‐ ‐ ‐ ‐ ‐ ‐ Later, Würst et al. used the 2′ -(4-fluorophenyl)-21-[18F]fluoro-20-oxo-11β,17α-dihydroxypregn-4-eno[3,2-*c*]pyrazole (**29**) (Figure 11) as a ligand for studying brain GRs. Biodistribution and radiopharmacological studies with male Wistar rats revealed promising brain uptakes and low in vivo radiodefluorinations in comparison with other PET radioligands for brain GRs [55].

‐

‐

‐

‐ ‐

#### 2.1.6. Insulin-Like Growth Factor-1 Receptor Ligands ‐ ‐

‐ ‐ ‐ ‐ Insulin-like growth factor-1 receptor (IGF-1R) is a potential therapeutic target, because it is overexpressed in many cancers, AD, traumatic brain injury, amyotrophic lateral sclerosis (ALS), Friedreich ataxia and aging. In 2013, Majo et al. synthetized and evaluated in vitro [ <sup>18</sup>F]BMS-754807 (**30**) (Figure 12), a potent and reversible small molecule inhibitor of the IGF-1R/IR kinases' family, currently in phase II clinical trials. ARG studies using surgically removed and pathologically identified grade IV glioblastoma, breast cancer and pancreatic tumors demonstrate that **30** binds to IGF-1R. The selectivity over other kinases, the presence of metabolically stable fluorine in the 2-substituted pyridine ring, which is amenable for radiolabeling by nucleophilic displacement with [18F]fluoride and a calculated lipophilicity (clogP) of 3.5, make this ligand a potential PET-imaging agent for in vivo monitoring of IGF-1R [56]. ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐

‐ **Figure 12.** Potential PET radiotracer **30** for IGF-IR imaging [56].

#### ‐ ‐ 2.1.7. Phosphodiester-10A Enzyme Inhibitors

‐

‐

‐ ‐ ‐ ‐ ‐ ′ ′‐ ‐ Phosphodiester-10A (PDE10A) is an enzyme that hydrolyzes adenosine and/or guanosine 3 ′ ,5′ -cyclic monophosphates (cAMP and cGMP, respectively). In the medium spiny neurons of the striatum, PDE10A messenger RNA and the corresponding protein are highly abundant. PDE10A inhibitors are a potential target for the diagnosis of schizophrenia, Huntington's disease, PD, obsessive-compulsive disorder and addiction.

‐

In 2010, Tu et al. made a first attempt to visualize PDE10A using a <sup>11</sup>C-radiolabeled PDE10A inhibitor named papaverine (**31**) (Figure 13). In vitro <sup>11</sup>C-papaverine showed selective PDE10A binding, but in vivo failed because of rapid washout of the tracer [57]. To overcome this problem, Celen et al. synthetized a specific and selective radioligand for PDE10A, the <sup>18</sup>F-quinoline-labeled [ <sup>18</sup>F]JNJ-41510417 (**32**) (Figure 13). They used rats and PDE10A knockout mice to show that **32** binds specifically and reversibly to PDE10A in the striatum, presenting high accumulation therein. This brain region was the only to show an increase of tracer concentration after the injection (1.6 SUV after 2 min vs. 2.6 SUV after 60 min). Other brain regions, the hippocampus, cortex and cerebellum showed a washout of the radiotracer. These results are in accordance with the distribution of PDE10A. Despite the [18F]JNJ-41510417 (**32**) good target specificity and signal-to-noise characteristics, slow brain kinetics due to its high potency is a limitation [58]. ‐ ‐ ‐ ‐

‐ ‐ ‐ ‐ ‐ ‐ ‐

‐

‐ ‐

‐ ‐ ‐

‐ ‐ ‐ ‐ ‐ ‐

‐ **Figure 13.** Radiotracers **31**–**33** for phosphodiester-10A (PDE10A) visualization in the brain by PET imaging [57–62].

‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ Then, Laere et al., in collaboration with Janssen Pharmaceuticals, synthetized and studied the human biodistribution of another radioligand [18F]JNJ-42259152 (**33**) [2-(4-(1-(2[18F]fluoroethyl)-4-(4-pyridinyl)-1*H*-pyrazol-3-yl)phenoxy)methyl)-3,5-dimethylpyridine] (Figure 13), which belongs to a series of structurally related analogues of MP10 in which the 2-quinolinyl heterocycle was replaced with substituted 2-pyridinyl moieties, resulting in a slightly lower potency compared to [18F]JNJ-41510417 (**32**) [59,60]. Six healthy male Caucasians (23–69 years; 3 younger than 40 years and 3 older than 60 years) have been volunteers in the study managed by Laere. Compound **33** showed a rapid uptake in the striatum, a few minutes after the injection, with a high clearance rate consistent with specific binding in this target region. The results of this pilot study are in accordance with the distribution of PDE10A in the human brain and show a promising kinetics and biodistribution of [18F]JNJ-42259152 (**33**) [60].

Laere et al. have demonstrated that PDE10A activity in the brain can be reliably quantified using [ <sup>18</sup>F]JNJ-42259152 (**33**) [61]. A relatively fast kinetics for the striatal region was observed, followed by a subsequent moderately fast washout. The regional in vivo distribution of this radiotracer was in agreement with the known distribution of PDE10A, being found predominantly in the putamen followed by the caudate nucleus, ventral striatum and substantia nigra. Compared with the activity in the striatum, the cortical and cerebellar activity were more than 10-fold lower. Later, Ooms et al. have investigated the effect of alterations in cAMP levels on [18F]JNJ-42259152 binding to PDE10A in the striatum homogenates. Increased affinity of this radiotracer for PDE10A was observed in the presence of cAMP, which seems to have an important role in the allosteric regulation of PDE10A [62].

Stepanov et al. described the synthesis of two [18F]-labeled PET radioligands to target PDE10A, the [18F]FM-T-773-d<sup>2</sup> (**34**) and [18F]FE-T-773-d<sup>4</sup> (**35**) (Figure 14), and their in vivo evaluation in nonhuman primates [63]. High brain uptake was measured for both radioligands and a fast washout. Specific binding reached the maximum after 30 min for [18F]FM-T-773-d<sup>2</sup> (**34**) and after 45 min for

[ <sup>18</sup>F]FE-T-773-d<sup>4</sup> (**35**). On account of brain uptake specific binding and kinetics, [18F]FM-T-773-d<sup>2</sup> (**34**) was considered the more promising PET radioligand for further clinical evaluation.

‐ **Figure 14.** PET radiotracers **34** and **35** for in vivo phosphodiester-10A (PDE10A) evaluation in nonhuman primates' brains [63].

#### 2.1.8. Translocator Protein Receptor Ligands

 α‐ The translocator protein (TSPO) receptor is an 18 kDa protein organized around five large transmembrane α-helices and located on the mitochondrial outer membrane [2]. This protein has a key role in the regulation of several cellular processes: steroid biosynthesis, cholesterol metabolism, apoptosis and cellular metabolism [64]. TSPO is highly expressed in organs involved in steroid synthesis as adrenal glands, testis, ovaries and pituitary glands [65]. In the central nervous system (CNS) and liver, TSPO expression is modest. However, in the case of acute or neurodegenerative pathologies associated with microglia or astrocytes, levels of TSPO in the brain are upregulated. The upregulation of this protein is directly correlated with the degree of neuronal damage. For these reasons, TSPO is considered a very promising target for the early imaging of neuroinflammation [66] and a possible indirect marker of neuronal loss progression, multiple sclerosis and AD [66–68] and has high relevance in neuroscience. The expression of this protein is also elevated in several cancers: colon, breast, glioma, prostate, colorectal, liver and ovary cancer, relating TSPO with disease progression and survival [2,64,65,69]. These evidences increased the interest in TSPO and led to the development of several radiolabeled ligands for the evaluation of the expression of this protein and detection of some of the aforementioned diseases by PET imaging.

‐ ‐ ‐ ‐ Since the discovery of the first nonbenzodiazepine ligand for TSPO, the isoquinoline carboxamide PK11195 (**36**) [70–72], several families of compounds were evaluated as TSPO ligands—among them, [11C]DPA-713 (**37**), [18F]DPA-714 (**38**), [18F]DPA (**39**), [18F]VUIIS1008 (**40**) and [18F]DPA-716 (or [18F]PBR146) (**41**), which are pyrazolo[1,5-*a*]pyrimidine acetamides (Figure 15) [72,73].

‐α‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ Chauveau et al. compared compound **38** with **36** and **37** using a rat model of acute neuroinflammation. (*R*,*S*)-α-amino-3-hydroxy-5-methyl-4-isoxazolopropionic acid (AMPA) was used to provoke neuroinflammation. Compounds **37** and **38** were specifically localized in the neuroinflammatory site with a similar signal-to-noise ratio in vitro. However, the fluorine-labeled tracer **38** achieved a higher bioavailability in the brain, higher uptake and higher binding potential than the other radiotracers. In the reference zone (contralateral area), a lower uptake of **37** and **38** were found when compared with **36**. These results showed that **37** and **38** have lower nonspecific bindings than **36** [74].

‐

‐

‐

‐

‐

‐

‐ ‐

‐

‐ ‐

‐

‐ **Figure 15.** [ <sup>11</sup>C]PK11195 (**36**) and pyrazolopyrimidine-derived radiotracers **37**–**41** for translocator protein (TSPO) PET imaging [72–82].

‐ ‐ ‐ ‐ ‐ Compound **38** was also used to monitor the TSPO levels after the injection of some antibiotic like minocycline. This drug inhibits the activation of microglial cells [75]. Furthermore, the radiotracer **38** was used on studies of rodent models of excitotoxicity, herpes encephalitis [76], astrocytic activation, excitotoxically lesioned nonhuman primate brains, abdominal aortic aneurysm [77], rheumatoid arthritis [78,79] and neuroinflammatory changes in the brain after morphine exposure [65,67]. It demonstrated a good uptake in the primate brain and an eight-fold higher uptake in the lesioned striatum of a quinolinic acid-lesioned rat model of activated microglia. Both **37** and **38** showed better imaging properties than **36** in the striatum of lesioned rats. A study using a rat model of herpes simplex encephalitis (HSE) suggested that **38**, as an agonist of TSPO, is potentially suitable for visualizing mild neuroinflammation [76].

Kuszpit et al. demonstrated that [18F]DPA-714 (**38**) is a sensitive tool for the detection of neuroinflammation, induced by Zika virus (ZIKV) infection, using a mice model of ZIKV neurological disease. Moreover, the evaluation of therapeutics being developed for the treatment of the disease is also possible by [18F]DPA-714 (**38**) imaging [80].

In 2017, Keller et al. compared the activity of **38** with its analogue [18F]F-DPA (**39**) (Figure 15), which presents the radionuclide directly linked to the phenyl ring. This study showed that compound **39** is metabolically more stable than its analogue **38** in rats' brains, being regarded as a promising TSPO radiotracer, because it shows a higher ratio between specific and nonspecific binding [81]. Later, in 2018, the same author described the potential of [18F]F-DPA **39** to visualize activated microglia at an early stage of AD pathology. The in vivo PET imaging and ex vivo brain AGR data indicated and increased uptake of this radiotracer **39** with age in the brains of transgenic animals (APP/PS1-21 mouse models of AD) in comparison with wild-type animals [68].

In 2012, Tang et al. used glioma-bearing rats to study the feasibility of using [18F]DPA-714 (**38**) for the visualization of TSPO expressing in brain tumors. This PET tracer showed to be highly specific to TSPO in glioma cell line homogenates. In vivo studies showed a higher uptake of **38** in tumor tissues than in other brain regions, suggesting that this compound could be used as a novel predictive cancer imaging tool [82]. In 2013, the same author carried out the synthesis and structure-activity relationship study of pyrazolopyrimidine-derived radioligands, presenting different substituents, namely at the 5, 6 and 7 positions, in order to find the more robust PET ligand. The best result was achieved with compound [18F]VUIIS1008 (**40**) (Figure 15), which showed a negligible binding in normal brains but a much higher binding in tumor tissues [64]. This result demonstrated that this radiotracer is a promising PET ligand for TSPO in tumors.

In 2014, Médran-Navarrete et al. evaluated a new <sup>18</sup>F-labeled analogue of [18F]DPA-714 (**38**), the [ <sup>18</sup>F]DPA-C5yne (**42**) (Figure 16), as a TSPO radioligand. In vitro studies revealed that **42** was stable in plasma at 37 ◦C for at least 90 min. AGR studies, using slices of AMPA-lesioned rat brains, showed a high specificity of binding and selectivity for TSPO, highlighting the potential of **42** as a radiotracer for TSPO [83]. One year later, Damont et al. synthetized a series of novel pyrazolo[1,5-*a*]pyrimidines and evaluated, in vitro, their affinity, lipophilicity and metabolism. Based on the results obtained, two of the synthetized compounds were chosen for <sup>18</sup>F-radiolabeling affording ligands [18F]DPA-C5yne (**42**) and (**43**) an analogue of [18F]DPA-714 (**38**), where the oxygen atom was replaced by a methylene group (Figure 16). Neuroinflammation PET images, using anesthetized Wistar rats seven days after AMPA-induced brain lesions in the right striatum, showed that both radiotracers **42** and **43** have a high in vivo-specific binding for TSPO. Compound **43** presented an ipsilateral-to-contralateral ratio of 3.57 ± 0.48 comparable to **38** (3.71 ± 0.39), and **42** showed an ipsilateral-to-contralateral ratio of 4.62 ± 0.44. These results suggest that both **42** and **43** are appropriate for neuroinflammation imaging [72].

‐ **Figure 16.** Pyrazolopyrimidine-derived radiotracers **42**–**45** for TSPO PET imaging [83–87].

‐ Recently, Tang et al. synthetized a new TSPO PET tracer [18F]VUIIS1018A (**44**), an analogue of [ <sup>18</sup>F]DPA-714 (**38**) where the 7-methyl group of the pyrazolopyrimidine ring was replaced by a *n*-butyl group (Figure 16), and evaluated its behavior in a preclinical model of neuroinflammation. These authors concluded that this structural modification increased the lipophilicity of **44** compared with **38** (3.7 vs. 2.4, respectively). After 60 min of injection of **44**, more than 85% of this probe was intact, indicating that it is very stable in the brain. It is noteworthy that, for **38**, just 50% of the probe was intact after 60 min of injection. In vivo blocking experiments and in vitro AGR assays confirmed a high binding specificity of **44** for TSPO, showing that it can be a promising TSPO PET tracer [84]. Tang et al.

‐

‐

‐

‐

‐ ‐ ‐

‐ ‐ developed a preclinical evaluation to image glioma. In this work, **44** exhibited a lower accumulation in healthy brains, what was regarded as an advantage to distinguish lower-grade gliomas. Compared with [18F]DPA-714 (**38**) and [18F]VUIIS1008 (**40**), **44** had an improved tumor-to-background ratio, a higher specific-to-nonspecific binding ratio and a higher tumor-binding potential. These results showed that **44** is a promising candidate to detect tumors with modest TSPO expression [85].

A series of novel 2-phenylpyrazolo[1,5-α]pyrimidin-3-ylacetamides were synthesized, and their in vitro binding affinities for TSPO and lipophilicity (log P7.5) were evaluated using [18F]DPA-714 (**38**) as the control. Based on the results of these assays, the tosylated precursor was selected for radiofluorination, affording **45** (Figure 16). Using LPS induced neuroinflammation rat models, a dynamic micro-PET study was performed and demonstrated a higher uptake of **45** in the ipsilateral region and a higher ratio of target-to-background than **38**. These results suggest that **45** can be a promising PET probe candidate for TSPO imaging [86].

In 2018, Verweij et al. measured the cellular response in patients after an acute coronary syndrome by PET imaging using [18F]DPA-714 (**38**) as a probe. TSPO receptor is highly expressed in myeloid cells. Using a PET tracer with a high affinity for the TSPO receptor as **38** was possible to determine the hematopoietic activity. In the acute phase, the treatment with **38** revealed a higher uptake in the bone marrow and the spleen. Three months after, the bone marrow uptake decreased to levels comparable to the healthy control. On the other hand, the spleen uptake remained elevated [87].

#### **3. Conclusions**

Considering that <sup>18</sup>F is the most suitable radionuclide for routine PET imaging and that pyrazoles are a key motif in medicinal chemistry and drug design, we made a compilation of [18F]pyrazole-derived imaging probes that have been developed and evaluated in the last 20 years. Regarding the examples presented herein, it is evident that pyrazoles are important scaffolds for the development of radiotracers for the diagnosis of several pathologies. In fact, pyrazoles interact with remarkable targets, namely with adenosine receptors, cannabinoid receptors, cyclooxygenase enzymes, dopamine receptors, glucocorticoid receptor, insulin-like growth factor-1 receptor, phosphodiester-10A enzyme and the translocator protein receptor. Among the probes presented herein, most are cannabinoid ligands and dopamine ligands, with a huge potential for brain imaging and TSPO ligands, namely pyrazolopyrimidine-derived radiotracers, that have shown important applications for the imaging of neuroinflammation and cancer.

The successful use of the described pyrazoles-imaging probes in clinics may help to increase the understanding of several diseases such as AD, PD, Huntington's disease, atherosclerosis, neuropsychiatric disorders, neuroinflammation, cardiovascular diseases and cancer, among others, and to identify ways to improve therapy. In this sense, the pyrazoles described herein can be regarded as potential theragnostic agents and as important templates for the development of novel radiotracers with improved properties for PET imaging.

**Author Contributions:** V.L.M.S. and P.M.O.G. idealized the manuscript. P.M.O.G. did the bibliographic research and prepared the original draft. V.L.M.S. and A.M.S.S. supervised the manuscript preparation, revised it and did the necessary corrections. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors would like to thank the University of Aveiro and FCT/MEC for the financial support to the QOPNA (FCT UID/QUI/00062/2019) and LAQV-REQUIMTE (UIDB/50006/2020) research projects, financed by national funds and, when appropriate, co-financed by FEDER under the PT2020 Partnership Agreement to the Portuguese NMR network and, also, to the assistant professor position of Vera L. M. Silva (within CEE-CINST/2018; since 01/09/2019). Vera L. M. Silva also thanks the funding from national funds through the FCT-I.P. in the framework of the execution of the program contract provided in paragraphs 4, 5 and 6 of art. 23 of Law no. 57/2016 of 29 August, as amended by Law no. 57/2017 of 19 July and the Integrated Programme of SR&TD "pAGE—Protein Aggregation Across the Lifespan" (reference CENTRO-01-0145FEDER-000003), co-funded by the Centro 2020 program, Portugal 2020 and European Union through the European Regional Development Fund.

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

#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Molecules* Editorial Office E-mail: molecules@mdpi.com www.mdpi.com/journal/molecules

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com ISBN 978-3-0365-2426-9