The Emerging Role of miRNAs for the Radiation Treatment of Pancreatic Cancer

Simple Summary Pancreatic cancer is an aggressive disease with a high mortality rate. Radiotherapy is one treatment option within a multimodal therapy approach for patients with locally advanced, non-resectable pancreatic tumors. However, radiotherapy is only effective in about one-third of the patients. Therefore, biomarkers that can predict the response to radiotherapy are of utmost importance. Recently, microRNAs, small non-coding RNAs regulating gene expression, have come into focus as there is growing evidence that microRNAs could serve as diagnostic, predictive and prognostic biomarkers in various cancer entities, including pancreatic cancer. Moreover, their high stability in body fluids such as serum and plasma render them attractive candidates for non-invasive biomarkers. This article describes the role of microRNAs as suitable blood biomarkers and outlines an overview of radiation-induced microRNAs changes and the association with radioresistance in pancreatic cancer. Abstract Today, pancreatic cancer is the seventh leading cause of cancer-related deaths worldwide with a five-year overall survival rate of less than 7%. Only 15–20% of patients are eligible for curative intent surgery at the time of diagnosis. Therefore, neoadjuvant treatment regimens have been introduced in order to downsize the tumor by chemotherapy and radiotherapy. To further increase the efficacy of radiotherapy, novel molecular biomarkers are urgently needed to define the subgroup of pancreatic cancer patients who would benefit most from radiotherapy. MicroRNAs (miRNAs) could have the potential to serve as novel predictive and prognostic biomarkers in patients with pancreatic cancer. In the present article, the role of miRNAs as blood biomarkers, which are associated with either radioresistance or radiation-induced changes of miRNAs in pancreatic cancer, is discussed. Furthermore, the manuscript provides own data of miRNAs identified in a pancreatic cancer mouse model as well as radiation-induced miRNA changes in the plasma of tumor-bearing mice.

Hereinafter, the miRNAs upregulated in plasma samples of tumor-bearing mice are discussed in the context of the recent literature and summarized in Table 1.
miR-339-3p was found to be deregulated in Vater's papilla adenocarcinoma [38]. In addition, miR-339-3p was downregulated in colorectal cancer (CRC) and its low-level expression was associated with lymph node metastasis in patients with CRC [39].
In relationship to pancreatic disorders, miR-320d was identified as a potential marker for late chronic pancreatitis [40]. miR-320d was recognized as a promising biomarker for early diagnosis of CRC because miR-320d expression could discriminate adenoma and CRC patients from healthy controls [41]. In serum samples of hepatocellular carcinoma (HCC) patients, the expression level of exosomal serum miR-320d was remarkably reduced compared to the respective controls [42].
A study from Long et al. described reduced miR-92b-3p expression levels in human PDAC tissues and a correlation with advanced tumor/node/metastasis (TNM) stages. Moreover, it was postulated that miR-92b-3p might act as a tumor suppressor in PDAC [43]. In gastric cancer, the serum level of exosomal miR-92b-3p was found to be higher than in healthy individuals, thus serving as a potential non-invasive biomarker for early diagnosis of gastric cancer [44]. miR-584-5p was negatively associated with tumor size in gastric cancer and therefore highlighted as a tumor suppressor and potential therapeutic target [45]. In lung adenocarcinoma, miR-584-5p was one of six upregulated miRNAs which might serve as circulating biomarkers for the early tumor detection [46].
The analysis from Ni et al. showed that miR-197-3p was downregulated in HCC tissues and low levels were correlated with larger tumor size and enhanced invasion, indicating an aggressive subtype [47].
Interestingly, miR-1307-3p and miR-1246, which were also deregulated in our study ( Figure 1B), are two of five miRNAs (miR-1246, miR-1307-3p, miR4634, miR-6861-5p, and miR-6875-5p) that are suitable for early detection of breast cancer [48]. In addition, the authors claim that the combination Analysis of all miRNAs in the dataset separates plasma samples derived from control mice (no tumor) to tumor-bearing (MIA PaCa-2) mice (PC1) and PC2 shows intra-group variability. (B). The upregulated miRNAs in the plasma of tumor-bearing (MIA PaCa-2) mice compared to the plasma of non-tumor-bearing mice are shown. Significant miRNAs were selected based on a log2 fold change ≥ |1| and an adjusted p-value of ≤0.05. Only transcripts with a base mean ≥50 were included.
Hereinafter, the miRNAs upregulated in plasma samples of tumor-bearing mice are discussed in the context of the recent literature and summarized in Table 1.  [53] cervical cancer upregulated serum [54] Cancers 2020, 12, 3703 5 of 20 miR-339-3p was found to be deregulated in Vater's papilla adenocarcinoma [38]. In addition, miR-339-3p was downregulated in colorectal cancer (CRC) and its low-level expression was associated with lymph node metastasis in patients with CRC [39].
In relationship to pancreatic disorders, miR-320d was identified as a potential marker for late chronic pancreatitis [40]. miR-320d was recognized as a promising biomarker for early diagnosis of CRC because miR-320d expression could discriminate adenoma and CRC patients from healthy controls [41]. In serum samples of hepatocellular carcinoma (HCC) patients, the expression level of exosomal serum miR-320d was remarkably reduced compared to the respective controls [42].
A study from Long et al. described reduced miR-92b-3p expression levels in human PDAC tissues and a correlation with advanced tumor/node/metastasis (TNM) stages. Moreover, it was postulated that miR-92b-3p might act as a tumor suppressor in PDAC [43]. In gastric cancer, the serum level of exosomal miR-92b-3p was found to be higher than in healthy individuals, thus serving as a potential non-invasive biomarker for early diagnosis of gastric cancer [44].
miR-584-5p was negatively associated with tumor size in gastric cancer and therefore highlighted as a tumor suppressor and potential therapeutic target [45]. In lung adenocarcinoma, miR-584-5p was one of six upregulated miRNAs which might serve as circulating biomarkers for the early tumor detection [46].
The analysis from Ni et al. showed that miR-197-3p was downregulated in HCC tissues and low levels were correlated with larger tumor size and enhanced invasion, indicating an aggressive subtype [47].
Interestingly, miR-1307-3p and miR-1246, which were also deregulated in our study ( Figure 1B), are two of five miRNAs (miR-1246, miR-1307-3p, miR4634, miR-6861-5p, and miR-6875-5p) that are suitable for early detection of breast cancer [48]. In addition, the authors claim that the combination of these five serum miRNAs can be used to differentiate breast cancer from benign diseases of the pancreas, biliary tract, and prostate, or from other cancers.
Additionally, the upregulation of miR-1246 in combination with miR-196a in plasma exosomes was found to be a potential indicator for localized pancreatic cancer. Therefore, these miRNAs might serve as circulating biomarkers for the early detection of this severe disease [49]. The study from Wei et al. showed that the expression level of serum miR-1246 was upregulated in PDAC patients compared to healthy controls and strongly reduced after tumor resection [50]. In addition, miR-1246 was part of a four-miRNA panel that was significantly upregulated in the serum of pancreatic cancer patients [51]. Further studies described elevated serum exosomal miR-1246 as a potential biomarker for the early diagnosis of gastric cancer [52] as well as esophageal squamous cell carcinoma (SCC) [53] and cervical cancer [54].
The listed publications demonstrate that circulating miRNAs represent a rich source of potential biomarkers for the early diagnosis, classification, therapeutic success, and recurrence monitoring in various human cancers, especially PDAC. Therefore, miRNA-based liquid biopsy holds promising impact for further implementation in the clinical routine.

miRNA Response to Ionizing Radiation
Radiation therapy uses ionizing radiation (IR) to induce double-strand breaks in the genomic DNA of tumor cells. Failure to restore genomic integrity before mitosis can lead to cell death or malignant transformation. Therefore, cells trigger the DNA double-strand repair. In this complex process, first sensor proteins recognize the DNA damage. Subsequently, transducer proteins recruit effector proteins responsible for cell cycle arrest, apoptosis, transcription arrest, and DNA repair [55].
Increasing evidence demonstrates that miRNAs play a critical role in the cellular response to IR, as multiple examples of miRNA expression changes in response to IR have been reported [56,57]. It was shown that miRNAs are responsible for regulating almost every cellular pathway, including the DNA damage response (DDR), after IR [55]. Whereas some miRNAs involved in the DDR are upregulated by IR (e.g., miR-34a, miR-100, and miR-143), others (e.g., miR-15b, miR-222, and miR-181a) are downregulated [55]. Interestingly, several miRNAs participating in the DDR (e.g., let-7 family, miR-15a, miR-16, miR-21, miR-24, miR-155, miR-182, and miR-302 cluster) have been reported to be both down-and upregulated, depending on the cell type, radiation dose, and time point of measurement post-RT [55].
Conversely, different members of the DDR pathway are involved in radiation-induced miRNA regulation. Ataxia telangiectasia mutated (ATM), a DDR pathway member, plays an important role in the regulation of miRNA biogenesis [58]. ATM-dependent activation of KH-type splicing regulatory protein (KSRP) upon DNA damage increases miRNA processing and expression of a specific class of miRNAs. miR-21 and miR-16, both playing a role in DDR, belong to the miRNAs that are upregulated by ATM-KSRP. Activation of the transcription factor p53 by IR leads to enhanced expression of the miR-34 family and also regulates other miRNAs, including let-7a and let-7b [59,60]. IR also affects the activity of other transcription factors (e.g., NF-kB, Myc, and E2F), which modulate the expression of several miRNAs in the DDR [57].
A very recent systematic review and meta-analysis summarize the effects of RT on circulating miRNAs in humans, nonhuman primates, and mice [63]. Radiation treatment in the included studies was very heterogeneous, including total body irradiation and local tumor-specific irradiation with different fractionation schemes and irradiation doses. Nevertheless, the meta-analysis identified 28 miRNAs with significant radiation-induced changes (18 miRNAs upregulated and 10 miRNAs downregulated). Whereas in this meta-analysis nine publications analyzed the effect of total body irradiation on the miRNA levels in plasma/serum of healthy mice, only one study used a tumor xenograft model (breast cancer) and investigated the impact of localized radiotherapy on tumor-specific miRNA plasma levels [64]. In this xenograft mammary tumor mouse model, decreased plasma levels of miR-155, miR-10b, and miR-21 have been detected after RT [64]. Specifically, for pancreatic cancer, only one study analyzed the impact of RT on the expression level of one specific miRNA and found enhanced miR-194-5p levels in exosomes derived from the plasma of irradiated PDX (patient derived xenograft) mice [62]. No in vivo studies describing the effect of RT on the expression levels of different miRNAs in pancreatic cancer were found in the literature.
Therefore, we investigated the impact of radiation on circulating miRNA levels in a pancreatic cancer xenograft mouse model (supplementary methods and Table S2). MIA PaCa-2 tumors were irradiated with a single dose of 5 Gy. Sham-irradiated (0 Gy) tumor-bearing mice served as control. Mice were sacrificed 24 h after irradiation and the plasma was collected. Small RNA sequencing revealed 21 miRNAs that were significantly modified ( Figure 2). Interestingly, 20 miRNAs were down-regulated after irradiation with 5 Gy, and only one miRNA (miR-184) was upregulated. The predicted targets of these miRNAs are summarized in Table S2.
From an earlier study, it is known that miR-374b functions as an oncogene by targeting PTEN, resulting in the activation of the PI3K/Akt signaling cascade in human gastrointestinal stromal tumor cells [65]. However, there are no reports concerning the role of miR-374b in pancreatic cancer yet. It is known that the miR-15b family is involved in cell cycle regulation, proliferation, and apoptosis [66]. A recent study showed that the expression of SMURF2, a tumor suppressor gene, is inhibited by miR-15b in pancreatic cancer [67]. There is also evidence of the downregulation of miR-652 in various cancerous diseases, but its role in the radiation response is still unknown [68].
by the respective miRNA (miR-15b, miR-451, miR421, miR-98, miR-16) leads to increased radiosensitivity. However, for other miRNAs the interrelation between target gene expression and radioresistance/radiosensitivity is not evident. Therefore, more research in this field is essential to understand the function and regulation of these miRNAs and their targets in mediating radioresistance, and finally to implement the data in clinical routine to improve the effectiveness of RT.  In PDAC, the overexpression of miR-144-3p reduces cell migration, proliferation, and invasion [69]. Currently, an increasing number of miR-93 and miR-106b target genes have been identified, suggesting miR-93 and miR-106b may differentially affect the behavior of tumors. A study found that the upregulation of miR-93/106b enhances PD-L1 in response to irradiation [70]. It is known that the PD-1/PD-L1 pathway serves as a mechanism for tumors to evade an antigen-specific T cell immunologic response. However, pancreatic cancer patients have been shown to be mostly resistant to PD-1 inhibition [71].
The downregulation of the tumor-suppressive miR-451a leads to enhanced cancer cell migration and invasion in hypopharyngeal SCC [72]. miR-186 overexpression has been observed in PDAC and was shown to contribute to its invasive potential [73]. Different studies revealed upregulation of miR-17 in pancreatic cancer, leading to increased proliferation and invasion [74,75]. Furthermore, inhibition of miR-17 increased the sensitivity of pancreatic cancer cell lines to chemotherapy by upregulation of Bim [76]. miR-421 has been found to be upregulated in pancreatic cancer as an oncogene and potential regulator of DPC4/Smad4 [77]. miR-98, a member of the let-7 family, regulates many cellular biological responses, including cell migration and apoptosis, after irradiation. The radiation-induced inhibition of miR-98 is closely related to both the p53 and ATM signaling pathways [78,79]. In a recent publication, it was shown that downregulation of miR-98-5p and other members of the let-7 family leads to increased PDAC proliferation and metastasis by reversely regulating the MAP4K4 pathway [80]. In the study of Morimura and colleagues, miR-20a levels in the plasma of pancreatic cancer patients were increased compared with healthy donors [81]. It was shown that the expression of miR-20a promotes radioresistance in nasopharyngeal cancer cells [82]. Higher levels of miR-20a activate the PTEN/PI3K/Akt signaling pathway and induce radioresistance of HCC [83]. A recent study identified miR-103 as one of the most important miRNAs in a functional screen for DNA damage disrupting agents. The authors identified Rad51 and Rad51D as key targets of miR-103 in tumor cells [84].
Another in vitro study demonstrated that miR-142 has a regulatory relationship with HIF-1α. miR-142 inhibits pancreatic cancer cell proliferation and invasion, partly by targeting HIF-1α at its binding site [85]. An earlier study analyzed the role of miR-26a in pancreatic tissue by quantifying its expression levels in 106 PDAC tissue samples [86]. The authors found that miR-26a expression was downregulated in PDAC compared to adjacent normal tissue. miR-140 inhibits cell viability, proliferation, and invasion in PDAC [87]. One study showed that miR-16 can enhance radiation sensitivity by regulating the TLR1/NF-κB signaling pathway and act as a potential therapeutic approach to overcome radioresistance for lung cancer treatment [88]. Jiao et al. compared the miRNA expression profile in PDAC with benign cystic tumors to identify miRNAs deregulated during PDAC development [89]. Amongst others, let-7i and let-7d were downregulated in tissue samples of PDAC [89].
The only miRNA that was significantly upregulated after irradiation in our study was miR-184. In recent years, miR-184 has been extensively explored in various cancer types. miR-184 was not only found to be upregulated in pancreatic cancer but can also facilitate the proliferation and invasion of tumor cells while suppressing apoptosis [90].
In radioresistant pancreatic cancer cell lines that were established by repeated exposure to radiation, miR-23b expression was reduced compared to the parental cell lines [123]. Overexpression of miR-23b rendered the radioresistant cells more sensitive to radiation. Furthermore, the expression of ATG12 (Ubiquitin-like protein), a target of miR-23b, was increased in the radioresistant cells. As ATG12 is involved in vesicle formation during autophagy, enhanced ATG12 expression increased autophagy and, subsequently, radioresistance. Furthermore, an association between decreased ATG12 expression and elevated miR-23b levels was observed in human pancreatic cancer tissue, indicating that miR-23b might affect autophagy activity in pancreatic cancer cells [123]. Another study showed that miR-216a can render radioresistant pancreatic cancer cells radiosensitive by inhibiting beclin-1-mediated autophagy [124]. Additionally, in a xenograft mouse model, miR-216a expression reduced the tumor growth after irradiation.
Tomihara et al. observed that high c-Met expression is associated with lower PFS and OS in pancreatic cancer patients receiving preoperative radiochemotherapy [125]. Further investigations revealed that miR-181b is downregulated in radioresistant pancreatic cancer cell lines leading to the upregulation of the transcription factor ETS1 and the c-Met pathway [125]. mTOR expression and subsequent mTOR signaling pathways have been upregulated after RT in human pancreatic cancer biopsies [61]. In vitro, mTOR was also upregulated upon radiation and mTOR expression has been shown to contribute to radioresistance. As mTOR is a miR-99b target, down-regulation of miR-99b enhanced mTOR expression and radioresistance in vitro [61].
Increased miR-193a levels have been found in irradiated dying pancreatic cancer cells, leading to elevated miR-193a levels in surviving cells [127]. Furthermore, miR-193a promotes proliferation and repopulation of the surviving cells via TGF-β2/TGF-βRIII signaling [127]. In patient-derived xenograft mouse models, radiation in combination with miR-193 antagonist inhibited cell repopulation and metastasis, and improved the survival. These data suggest that inhibition of miR-193 might increase radiosensitivity.
Transfection with a miR-620 mimic increased the radioresistance of MIA PaCa-2 cells and also of breast and prostate cancer cell lines [128]. The tumor suppressor gene HPGD (15-hydroxyprostaglandin dehydrogenase) was identified as a target of miR-620. miR-620 overexpression leads to degradation of HPGD, enhanced prostaglandin E2 (PGE2) levels and signaling through the EP2 receptor promoting the survival and radioresistance of tumor cells.
Pancreatic cancer cell lines have reduced miR-34a expression compared to normal pancreatic ductal epithelial cells [131]. p53 has been shown to directly regulate miR-34 family members, and subsequently, the expression of miR-34 is strongly reduced in p53-mutant cancer cells [132]. miR-34 restoration in p53-mutant pancreatic cancer cells, which have initially low miR-34 levels leads to the downregulation of anti-apoptotic Bcl-2 and the Notch signaling pathway, which is involved in the maintenance of cancer stem cells [126,132,133]. Furthermore, miR-34 restoration reduces clonogenic cell growth and enhances radio-and chemosensitivity [126]. Therefore, miR-34 may restore the tumor-suppressing function of p53 in p53-deficient human pancreatic cancer cells and might constitute a new approach to treating p53-mutated pancreatic cancer.
Inhibition of Lin28 by siRNA abrogates posttranscriptional degradation of let-7a and increases radiosensitivity of AsPc-1 cells presumably through down-regulation of Kras expression [129].
Baek et al. investigated both the influence of miRNA expression on the radiosensitivity to gamma-rays and to carbon ion beam RT [130]. Overexpression of miR-374 did not affect the sensitivity to conventional gamma-ray radiation, but the sensitivity to carbon ion beam radiation was enhanced. These data suggest that miR-374 could be a potential radiosensitizer for carbon ion RT.
In summary, only a few miRNAs associated with radioresistance in pancreatic cancer have been identified so far. Therefore, comprehensive studies analyzing miRNA expression profiles are necessary to identify miRNAs, which can predict the individual radiosensitivity or constitute targets for radiosensitization.

Discussion
Nowadays, early diagnosis is essential for the successful treatment of pancreatic cancer, and surgical resection is the only potentially curative therapy [3]. Since pancreatic cancer is a locally invasive as well as a systemic disease, the disease recurs in most patients, which leads to a five-year survival rate of only 30% after complete resection [134]. miRNAs could help to detect pancreatic cancer in an earlier stage, leading to increased curative treatments. In addition, more research into neoadjuvant treatment options, including RT or combined CRT, is urgently needed. The goal of RT is mainly to downsize the tumor to enable a secondary resection. Therefore, neoadjuvant RT may lead to improved long-term prognosis in patients with borderline and primary non-resectable, locally advanced pancreatic cancer [4]. While RT was historically a central treatment component, its role has been called into question based on the publication of clinical trials with conflicting results and due to the effectiveness of combined chemotherapeutic regimes (e.g., FOLFIRINOX) [135].
Novel molecular biomarkers are required to increase the efficacy of RT and to define which patients benefit most from it [21]. Especially, miRNAs could have the potential as novel biomarkers for diagnosis, monitoring of recurrence, and predicting prognosis and survival of patients with pancreatic cancer [22,136]. Our present study provides an identification of miRNAs in an in vivo pancreatic cancer mouse model. A panel of seven miRNAs was significantly upregulated in plasma of tumor-bearing mice (Figure 1). In the next step, this biomarker panel should be validated in plasma of pancreatic cancer patients. Especially, circulating miRNAs represent a valuable source and are easily accessible [26].
Several miRNAs have been shown to play a role in radioresistance in different tumor entities [56,91]. Furthermore, significant changes in the expression levels of miRNAs can be observed after exposure to IR. It is known that miRNAs are responsible for regulating almost every cellular pathway, including the DNA damage response after IR [57]. Understanding the regulation and function of miRNAs is essential to improve the effectiveness of RT. To determine radiation-induced miRNA changes in the plasma of tumor-bearing mice, local irradiation of subcutaneous MIA PaCa-2 tumors was performed. Small RNA sequencing revealed 21 miRNAs that were significantly regulated ( Figure 2). Interestingly, 20 out of these 21 miRNAs were downregulated. Although a mechanistic basis is not yet available, the reduced levels can be assumed to promote translation of specific miRNA target proteins [137]. A current literature search reveals an association between the expression of these miRNAs, their targets (e.g., Chk1, c-Myc, ATM and Cyclins) and either radioresistance or radiosensitivity (Table 2) [98,107,113,120]. Most of these miRNAs (miR-15b, miR-451, miR-186, miR-421, miR-98, miR-142, miR-26b, miR-16, let-7 family) mediate radiosensitivity [98,[105][106][107][108][109][112][113][114][116][117][118]120,121] suggesting that the radiation-induced downregulation of these miRNAs might enhance radioresistance.
Specifically, for pancreatic cancer, only very few publications about the interplay between miRNAs and radioresistance exist, which were reviewed in this manuscript (Table 3).
Although radioresistance is rather complex, emerging evidence has demonstrated that epigenetic alterations, including miRNA changes, play important roles in resistance to RT. Therefore, more comprehensive preclinical and clinical studies analyzing miRNA expression profiles in context with radioresistance are urgently needed to identify miRNAs that can predict the individual radiosensitivity. Moreover, further studies are required to confirm the molecular mechanisms of the deregulated circulating miRNAs as diagnostic biomarkers for pancreatic cancer. In addition, miRNAs can help to develop innovative cancer therapies, which is of great significance for improving the life expectancy of pancreatic cancer patients [138].
The possibility to detect miRNAs in blood samples makes them a promising tool for rapid PDAC diagnosis, as well as patient stratification, to allow personalized therapy [139]. However, miRNAs are not yet routinely used for cancer management, and we should consider the challenges behind this strategy. One of them is the need to standardize protocols [140]. From sample collection and preparation to miRNA detection, inter-laboratory reproducibility must be ensured. In this regard, sampling tubes, patient-related factors, sample storage, and sample processing can affect RNA yield and purity [141].
The high heterogeneity of PDAC should also be regarded as an additional challenge, as it makes it harder to select reliable biomarkers [139]. Concerning miRNAs applications in RT, it should also be considered that there is still insufficient knowledge about their functions in different RT concepts such as dose fractionation. Information is also lacking regarding the use of alternative radiation qualities such as protons or carbon ions and the tumor microenvironment such as hypoxia. Additionally, the translation of preclinical data from in vitro or xenograft models into clinical routine using patient materials is still poorly realized [137]. Preclinical miRNA candidates will have to be investigated in large prospective trials of different PDAC patient populations to gain confidence in the selected biomarkers [142]. Future trials should be designed to evaluate miRNA biomarkers in a temporal manner, by collecting samples before and after RT, leading us towards personalized RT for PDAC [142].

Conclusions
In the last decade, miRNAs have become promising tools as prognostic and diagnostic biomarkers as well as therapeutic targets for innovative and personalized cancer treatment. Several miRNAs have been found differentially expressed and to be predictive for the treatment outcome in multiple cancer entities. With a five-year survival rate of less than 7%, pancreatic cancer presents an urgent need to identify the subset of patients which could benefit from treatment intensification and to establish novel individualized CRT treatment options. In this article, we focused on the role of miRNAs in pancreatic cancer, radioresistance, and radiation-induced changes, which could lead the way to personalized treatment in the future.