Involvement of Differentially Expressed microRNAs in the PEGylated Liposome Encapsulated 188Rhenium-Mediated Suppression of Orthotopic Hypopharyngeal Tumor

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 188Re-liposome was prepared to treat orthotopic HPC tumors and analyze the deregulated microRNA expressive profiles. The therapeutic efficacy of 188Re-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 188Re-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 188Re-liposome, respectively. According to Differentially Expressed MiRNAs in human Cancers (dbDEMC) analysis, most of 188Re-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 188Re-liposome. 188Re-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. 188Re-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.


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 99m Tc-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 ( 188 Re) 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 188 Re is suitable for tumor ablation and will not have significant side effects on distant normal tissues [9,10]. 188 Re has been conjugated to hydroxyethylidene diphosphonate (HEDP) for bone pain palliation [11,12]. The antibody conjugated 188 Re has also been reported to treat different cancers [13]. Additionally, low immunogenic peptides such as somatostatin derivative conjugated 188 Re, have been investigated in clinics for patients with advanced pulmonary cancer [14]. 188 Re-labeled lipiodol and microspheres were used for the treatment of hepatocellular carcinoma [15][16][17]. 188 Re-labelled radiocolloids have also been developed to treat skin cancers within a brachytherapy device [18]. 188 Re-loaded lipid nanocapsules, also called 188 Re-liposome, have been demonstrated to be biocompatible and subjected to a phase 0 clinical study for patients with metastatic tumors [19]. 188 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][21][22][23][24][25][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, 188 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 188 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 188 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 188 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 188 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 188 Re-liposome-induced miRNA dysregulation detected by RNA-Seq were discussed.

Effects of 188 Re-Liposome on HPC Derived Orthotopic Tumors Using the Repeated Dose Regime
A flowchart of 188 Re-liposomal manufacturing and the chemical structures of the 188 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 188 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 188 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 188 Re-liposome-treated mice after 30 days of initial implantation ( Figure 2D). The body weights of tumor-bearing mice were not significantly affected by 188 Re-liposome ( Figure 2E). Furthermore, we showed that DNA damage marker γ-H2AX was significantly up-regulated by 188 Re-liposome treatment ( Figure 2F,G). These data suggest that the regime of 188 Re-liposome treatment with repeated doses exhibited therapeutic efficacy, with little systemic toxicity.

Use of NGS Analysis to Investigate the microRNA Expressive Profile of HPC Tumor Treated with 188 Re-Liposome
The resected FaDu HPC tumors treated with saline or repeated doses of 188 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 188 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 188 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 188 Re-liposome treated tumor, respectively. Under this condition, the total reads number of the control and 188 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 188 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 188 Re-liposome treated HPC tumor models (Supplementary Figure S5). Furthermore, we set the cutoff at five-fold change (log 2 ) of microRNA and showed that 13 microRNA and 9 microRNA with mature forms were up-regulated and down-regulated by 188 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, 188 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 188 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 188 Re-liposome-down-regulated microRNAs contain oncogenic properties (see references in Table 3). These results suggested that the therapeutic efficacy of 188 Re-liposome was associated with the deregulation of tumor-suppressive and/or oncogenic microRNAs. 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, 188 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 188 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 188 Re-liposome-down-regulated microRNAs contain oncogenic properties (see references in Table 3). These results suggested that the therapeutic efficacy of 188 Re-liposome was associated with the deregulation of tumor-suppressive and/or oncogenic microRNAs.

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 188 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).

Investigation of Differentially Expressed microRNAs in Clinical Samples
As 188 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 188 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 188 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 188 Re-liposome can reverse the expression of the microRNAs that were originally deregulated in HNSCC. Hsa-miR-206-3p down-regulation Tumor suppressor Weaks cell proliferation, migration, invasion, promotes S phase cell arrest.

Prediction of Genes Targeted by 188 Re-Liposome-Deregulated microRNAs
We next examined the potent target genes that might be affected by 188 Re-liposome deregulated microRNAs. The miRDB online database was employed to rank the numbers of predicted genes targeted by microRNAs that were deregulated by 188 Re-liposome (see Materials and Methods). The

Prediction of Genes Targeted by 188 Re-Liposome-Deregulated microRNAs
We next examined the potent target genes that might be affected by 188 Re-liposome deregulated microRNAs. The miRDB online database was employed to rank the numbers of predicted genes targeted by microRNAs that were deregulated by 188 Re-liposome (see Materials and Methods). The

Prediction of Genes Targeted by 188 Re-Liposome-Deregulated microRNAs
We next examined the potent target genes that might be affected by 188 Re-liposome deregulated microRNAs. The miRDB online database was employed to rank the numbers of predicted genes targeted by microRNAs that were deregulated by 188 Re-liposome (see Materials and Methods). The ranking of target numbers affected by 188 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 188 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 188 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 188 Re-liposome are distinct.
Molecules 2020, 25, x FOR PEER REVIEW 10 of 25 ranking of target numbers affected by 188 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 188 Reliposome up-regulated microRNAs. Only one target, named SLC44A2 encoded choline transporterlike protein 2 was co-regulated by these three microRNAs ( Figure 5C). The same logic was used for 188 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 188 Re-liposome are distinct.

Analysis of the Molecular Pathways Regulated by 188 Re-Liposome-Affected microRNA
We next used the pathview R package to investigate the genes influenced by microRNAs regulated by 188 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 188 Re-liposome. We found that thirty pathways in resected HPC tumors were significantly affected by 188 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, 188 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 188 Reliposome-induced or -suppressed microRNA expression ( Figure 5). This revealed that genes associated with cell cycle progression, proliferation, and apoptosis were affected by 188 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 188 Re-liposome regulated microRNA expression and tumor suppression of the HPC tumor model.

Analysis of the Molecular Pathways Regulated by 188 Re-Liposome-Affected microRNA
We next used the pathview R package to investigate the genes influenced by microRNAs regulated by 188 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 188 Re-liposome. We found that thirty pathways in resected HPC tumors were significantly affected by 188 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, 188 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 188 Re-liposome-induced or -suppressed microRNA expression ( Figure 5). This revealed that genes associated with cell cycle progression, proliferation, and apoptosis were affected by 188 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 188 Re-liposome regulated microRNA expression and tumor suppression of the HPC tumor model.

Association of 188 Re-Liposome-Regulated microRNA and Patients' Survival Rate
As 188 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 188 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 188 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, 188 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 188 Re-liposome may be important for perspective clinical evaluation.

Association of 188 Re-Liposome-Regulated microRNA and Patients' Survival Rate
As 188 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 188 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 188 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, 188 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 188 Re-liposome may be important for perspective clinical evaluation.

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 188 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 188 Re-liposome after repeated administration. The expressions of the total RNA number and annotated small RNA were reduced in 188 Re-liposome treated tumors, suggesting that 188 Re-liposome would suppress gene transcription and expression.
As 188 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 188 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 188 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 188 Re-liposome treatment. Together, these results are partially consistent with previous reports that might account for the therapeutic mechanisms of 188 Re-liposome from the viewpoint of γ-H2AX-mediated DNA damage responses.
According to the results of RNA-seq, 188 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 188 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 188 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 188 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 188 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 188 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 188 Re, 177 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. 177 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 188 Re-liposome treatment.
The predicted genes targeted by microRNAs were also analyzed by the miRDB public database. For the 188 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 188 Re-liposome.
Using the KEGG pathway database, we found that 30 pathways in orthotopic HPC tumors were significantly influenced by 188 Re-liposome treatment. Although the cancer-suppression-related pathways were expected to be regulated by 188 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 188 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 188 Re-liposome regulated primary microRNAs and their association with patients' survival rates [84]. Although several potent tumor suppressive or oncogenic microRNAs were influenced by 188 Re-liposome, only part of these microRNAs exhibited the expected association with patients' survival probability in HNSCC. For instance, 188 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 188 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 188 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 188 Re-liposome. Although no clinical evidence showing that patients treated with 188 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][87][88][89]. To the best of our knowledge, this is the first study uncovering the therapeutic mechanisms of 188 Re-liposome by an investigation of the pan-expression of microRNA. As 188 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.

Establishment of HPC Orthotopic Tumor Model for Evaluating the Therapeutic Efficacy of 188 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 6 ) 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 188 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 2 × 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).

Tumor Collection and Next-Generation Sequencing (NGS)
Tumors were harvested from tumor-bearing mice after four weeks of 188 Re-liposome treatment. Total RNA of both saline control and 188 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].

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 188 Re-liposome group is relative to that of saline-only group. If the mean expression value in 188 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.

Validation of microRNA Expression Using qPCR
To validate miRNA levels before and after HPC tumor treated with 188 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).

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 188 Re-liposome, and then took Log2 value to draw out.

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.

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].

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.

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 188 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 188 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.