Noncoding RNAs Associated with Therapeutic Resistance in Pancreatic Cancer

Therapeutic resistance is an inevitable impediment towards effective cancer therapies. Evidence accumulated has shown that the signaling pathways and related factors are fundamentally responsible for therapeutic resistance via regulating diverse cellular events, such as epithelial-to-mesenchymal transition (EMT), stemness, cell survival/apoptosis, autophagy, etcetera. Noncoding RNAs (ncRNAs) have been identified as essential cellular components in gene regulation. The expression of ncRNAs is altered in cancer, and dysregulated ncRNAs participate in gene regulatory networks in pathological contexts. An in-depth understanding of molecular mechanisms underlying the modulation of therapeutic resistance is required to refine therapeutic benefits. This review presents an overview of the recent evidence concerning the role of human ncRNAs in therapeutic resistance, together with the feasibility of ncRNAs as therapeutic targets in pancreatic cancer.


Introduction
Most pancreatic cancer (PaC) patients are diagnosed at an advanced stage owing to the lack of early detections; therefore, surgical management is unavailable for over 80% of patients [1,2]. Moreover, PaC is resistant to treatment options, such as radiotherapy, chemotherapy, and targeted therapy [1,3,4]. These features underline the requirement of developing more effective treatments for PaC. Noncoding RNAs (ncRNAs) are differentially expressed in cancer and control diverse signaling pathways involved in the regulation of therapeutic resistance [5][6][7][8]. An improved understanding of the relationship between therapeutic resistance and ncRNAs can provide meaningful insights to develop new treatment strategies for PaC. This review highlights the role of human ncRNAs in modulating the effectiveness of treatments in PaC.

Noncoding RNAs
A large number of studies have provided evidence that microRNAs (miRNAs), in general, repress the translation and induce the degradation of their target messenger RNAs (mRNAs) via binding to the 3 untranslated region (3 UTR) [9]. Long noncoding RNAs (lncRNAs) play critical roles in gene regulation [10]. They can regulate chromatin structure, gene transcription, and pre-mRNA splicing [11]. Furthermore, the stability of proteins is affected by lncRNAs [12]. Another functional competency of lncRNAs is to sponge miRNAs, thus constraining the abundance and activity of miRNAs. For example, a recent study demonstrated that lncRNA-ADPGK-AS1 inhibits miR-205-5p, thereby promoting the progression of PaC via activating epithelial-to-mesenchymal transition (EMT) [13]. Moreover, circular RNAs (circRNAs) can control gene transcription via interaction with RNA-binding proteins [8,14]. They also regulate the signaling pathways through the sequestration of miRNAs [8,15].

Mechanisms of Therapeutic Resistance
Therapeutic resistance is related to EMT, cancer stem cells (CSCs), and efflux transporters. PaC cells expressing high levels of EMT markers are resistant to gemcitabine, 5-fluorouracil , and cisplatin. In fact, the efficacy of these anti-cancer agents is restored by an inhibition of zinc finger E-box-binding homeobox (ZEB1) [16][17][18]. Another study also showed that maintenance of the EMT program mediates radioresistance in PaC [19]. In addition, pancreatic CSCs are resistant to currently available therapies owing to their hallmarks, including the intense expression of anti-apoptotic factors and drug efflux transporters [20]. The treatment of gemcitabine promotes cancer stemness, thus reinforcing chemoresistance in PaC [21]. Thus, the inhibition of cancer stemness has been attempted to increase therapeutic efficacy against PaC [22,23]. In particular, cancer growth and metastasis are remarkably suppressed by the combination of gemcitabine with afatinib, a cancer stemness inhibitor [23].
Moreover, cellular factors related to survival and apoptosis are linked to therapeutic resistance. A recent study showed that gemcitabine resistance is aggravated by an activation of AKT serine/threonine kinase (AKT) signaling; therefore, AKT inhibition augments the efficacy of gemcitabine by activating apoptotic cell death in vitro and in vivo [24]. In addition, extracellular signal-regulated kinase (ERK) positively regulates the level of anti-apoptosis factors such as B-cell CLL/lymphoma 2 (BCL2), impeding caspase activations [25]. Activated ERK is involved in therapeutic resistance to several agents, such as gemcitabine, paclitaxel, and 5-FU [26][27][28].
Accumulating evidence has shown that autophagy has a cytoprotective activity against anti-cancer therapies [29,30]. In PaC, the sensitivity of cells to doxorubicin is enhanced by the pharmacological suppression of autophagy [31]. The silencing of autophagy-related 5 (ATG5) increases doxorubicin-induced apoptosis as well [31]. In addition, autophagy is induced by several agents, including gemcitabine, 5-FU, and salinomycin. The inhibition of autophagy augments the cytotoxicity of these agents in PaC [32][33][34]. It suggests that cancer cells withstand stressful conditions via the compensatory activation of autophagy. It has been reported that miR-10a-5p can act as a tumor-suppressive miRNA or an oncogenic miRNA, depending on cancer types. The overexpression of miR-10a-5p suppresses cell cycle progression and metastasis in cervical and colorectal cancer, respectively [35,36]. By contrast, a recent study demonstrated that miR-10a-5p confers gemcitabine resistance by targeting transcription factor-activating enhancer-binding protein 2C (TFAP2C) in PaC [37]. In this study, it was observed that the overexpression of miR-10a-5p or TFAP2C increases or decreases the expression of EMT-related genes such as snail family transcriptional repressor 1 (SNAI1), respectively ( Figure 1 and Table 1). In line with this, the administration of gemcitabine inefficiently reduces the growth of miR-10a-5p-overexpressing PaC cells in a mouse xenograft model [37]. However, another study showed that TFAP2C triggers tumorigenesis and EMT by upregulating the level of transforming growth factor-β receptor 1 (TGFBR1) in lung cancer [38]. These findings suggest that the function of TFAP2C is dissimilar in a cellular context-dependent manner.    Several studies demonstrated that tumor necrosis factor alpha-induced protein 3 (TNFAIP3, also known as A20) inhibits EMT. The knockdown of TNFAIP3 facilitates the migration and invasion of nasopharyngeal cancer cells [61]. Furthermore, TNF-induced motility is suppressed by TNFAIP3 in hepatocellular carcinoma cells [62]. Moreover, TN-FAIP3 diminishes the level of EMT markers such as ZEB1 via inactivating nuclear factor kappa B (NF-κB) signaling, thereby negatively modulating the migration and invasion capacities of lung cancer cells [63]. These results suggest that miRNAs targeting TNFAIP3 can regulate the sensitivity of cells to anti-cancer therapies. In PaC, it was found that TN-FAIP3 is targeted by miR-125a-5p and that both miR-125a-5p overexpression and TNFAIP3 knockdown desensitize cells to gemcitabine [43] ( Figure 1 and Table 1).

MiR-223-3p
It was demonstrated that miR-223-3p is upregulated in gemcitabine-resistant PaC cells [49,50]. Further evidence has shown that miR-223-3p is capable of targeting F-box and WD repeat domain-containing 7 (FBXW7) and induces gemcitabine resistance via activating Notch signaling-mediated EMT [49,50] ( Figure 1 and Table 1). Moreover, it was revealed that the level of miR-223-3p is downregulated by genistein and that the combination of genistein and miR-223-3p inhibitors synergistically sensitizes resistant cells to gemcitabine in vitro and in vivo [50]. However, miR-223-3p can repress the migration and invasion of osteosarcoma cells [66], implying that the role of miR-223-3p is disparate in a cell-type-dependent manner.

MiR-21-5p and MiR-221-3p
Cancer stemness is enhanced by miR-21-5p, which is capable of targeting TGFBR2 in colorectal cancer. Furthermore, it was identified that miR-221-3p intensifies cancer stemness by targeting DNA methyltransferase-3 beta (DNMT3B) in breast cancer [70,71]. Both miRNAs are upregulated in stem-like PaC cells compared to non-stem cancer cells [40], suggesting that these miRNAs can play an essential role in stemness regulation, probably via targeting TGFBR2 and DNMT3B in PaC. Notably, the knockdown of miR-21-5p and miR-221-3p suppresses the population of stem-like PaC cells, as well as increasing the effects of 5-FU and gemcitabine in vitro. Moreover, the in vivo growth of stem-like PaC cells is significantly reduced by the combined knockdown of miR-21-5p and miR-221-3p [40]. These results suggest that the inhibition of these miRNAs can be a potential therapeutic strategy for PaC ( Figure 1 and Table 1).

MiR-181c-5p
The induction of apoptosis can be impeded by miR-181c-5p owing to its potentiality to target Fas cell surface death receptor (FAS) in Ewing's sarcoma [79]. Moreover, it was unveiled that miR-181c-5p renders PaC cells resistant to gemcitabine, 5-FU, and paclitaxel by reducing the level of drug-induced apoptosis [45]. In this study, the Hippo signaling pathway was found to be restrained by miR-181c-5p, which targets multiple genes such as mammalian STE20-like protein kinase 1 (MST1) [45] ( Figure 1 and Table 1). The Hippo pathway has been proven to inactivate Yes-associated protein 1 (YAP1), resulting in the downregulation of anti-apoptotic factors such as BCL2 [80,81]. However, miR-181c-5p can inhibit tumorigenesis and stemness in cervical cancer and glioblastoma [82,83], pointing out that the function of miR-181c-5p is highly diverse depending on the cancer type.
2.3.6. MiR-296-5p BCL2-related ovarian killer (BOK) is a non-canonical member of the BCL2 family and serves as a tumor suppressor by triggering cell death [84,85]. A recent study unveiled that the overexpression of miR-296-5p contributes to resistance to 5-FU and gemcitabine by directly targeting BOK [52] (Figure 1 and Table 1). In addition, miR-296-5p enhances the invasion and EMT process, suggesting that miR-296-5p can act as an EMT-stimulating miRNA [52].

MiR-125a-3p
Although miR-125a-5p is an EMT-promoting factor (see Section 2.1.2), the EMT process can be subdued by miR-125a-3p that is generated from the same miRNA precursor. It was indicated that the effect of gemcitabine is increased by miR-125a-3p, which represses EMT by targeting proto-oncogene C-Fyn (FYN) [113] (Figure 2 and Table 2). The expression of miR-125a-3p and miR-125a-5p is downregulated and upregulated, respectively, in PaC tissues [140,141]. These findings suggest that the differential stability of miR-125a-3p and miR-125a-5p is regulated by undiscovered specific degradation factors, contributing to therapeutic resistance.

MiR-138-5p and MiR-153
It has been noted that miR-138-5p performs a tumor-suppressive function by regulating migration, invasion, and EMT in breast and colorectal cancer [142,143]. Furthermore, miR-153 is recognized to suppress EMT and metastasis in oral cancer, breast cancer, as well as hepatocellular carcinoma [144,145]. In addition, both miR-138-5p and miR-153 have been proposed to inhibit the progression of PaC through regulating proliferation, migration, and invasion [146,147]. Moreover, it was validated that miR-138-5p targets vimentin (VIM) and increases the anti-proliferative effect of 5-FU in vitro. Moreover, miR-153, which targets SNAI1, reinforces the inhibitory effects of gemcitabine on cell viability in vitro and the growth of PaC cells in vivo [116,120] (Figure 2 and Table 2). These findings demonstrate the role of them as bona fide EMT-and therapeutic resistance-suppressing miRNAs.

MiR-3656
Ras homolog family member F (RHOF) exerts oncogenic effects through promoting EMT and metastasis [150]. In PaC, RHOF knockdown leads to an increase in EMTpromoting factors, such as VIM and TWIST1 [138]. RHOF is targeted by miR-3656, and the cytotoxicity of gemcitabine is ameliorated in miR-3656-overexpressing cells. Further, TWIST1 overexpression interferes with the chemosensitization effect of miR-3656 in vitro. It was also confirmed that miR-3656 enhances gemcitabine-induced growth inhibition, along with a decrease in VIM and TWIST1 levels in vivo [138] (Figure 2 and Table 2). These observations suggest that the miR-3656/RHOF/EMT axis notably modulates the responsiveness of cancer cells to gemcitabine.

MiR-30a-5p
As mentioned in Section 3.1.1, miR-30a-5p can modulate the effect of gemcitabine on cancer cells. In addition to this, miR-30a-5p is able to target forkhead box D1 (FOXD1), an upstream activator of ERK signaling. As a consequence, the overexpression of miR-30a-5p can induce apoptosis in vitro and potentiate the anti-cancer activity of gemcitabine in vivo [108] (Figure 2 and Table 2).

MiR-374-5p
In breast cancer, miR-374-5p promotes cell proliferation, survival, migration, and invasion [164]. By contrast, miR-374-5p performs a tumor-suppressive function in lung and bladder cancer and is associated with overall patient survival [165,166]. In PaC, miR-374-5p attenuates therapeutic resistance. PaC cells transfected with miR-374-5p exhibit a high degree of apoptosis following treatments with gemcitabine in vitro [127]. In this study, it was noticed that miR-374-5p potentiates gemcitabine efficacy, thereby extending survival in a xenograft mouse model of PaC. Moreover, the effect of cisplatin tends to be increased by miR-374-5p in resistant cells [128]. Such resistance-alleviating effects of miR-374-5p can be due to the direct inhibition of several anti-apoptotic genes, such as BCL2, X-linked inhibitor of apoptosis (XIAP), and baculoviral IAP repeat-containing 3 (BIRC3) [127] (Figure 2 and Table 2).

MiR-455-3p and MiR-1285
Tafazzin (TAZ), a YAP homolog, is responsible for therapeutic resistance and is inactivated by the Hippo pathway. The blocking of YAP/TAZ signaling is expected to reduce the development of therapeutic resistance [167,168]. Gemcitabine efficacy can be augmented by atorvastatin, which suppresses YAP/TAZ signaling [169] (also see Section 2.3.4 describing the Hippo pathway and YAP1). In PaC, the downregulation of miR-455-3p and miR-1285 aggravates gemcitabine resistance. On the other hand, the overexpression of these miRNAs leads to the improvement of gemcitabine efficacy [130,137]. In their study, it was confirmed that TAZ and YAP1 are directly modulated by miR-455-3p and miR-1285, respectively ( Figure 2 and Table 2).

MiR-494-3p
Both proto-oncogene c-Myc (MYC) and sirtuin 1 (SIRT1) are highly expressed in PaC [170,171]. The silencing of either MYC or SIRT1 can stimulate apoptosis induction, thus increasing the anti-cancer activity of several agents, such as 5-FU and gemcitabine [172,173]. Further, it was shown that both MYC and SIRT1 can be targeted by miR-494-3p. Accordingly, PaC cells are sensitized to 5-FU and gemcitabine by miR-494-3p restoration [131] ( Figure 2 and Table 2). It is noteworthy that the metastasis of hepatocellular carcinoma is accelerated by miR-494-3p [174], indicating that miR-494-3p plays context-specific functions.

MiR-23-3p and MiR-137-3p
Lipidation of LC3I to LC3II is necessary for autophagosome formation and is known to be facilitated by ATG5 and ATG12 [179]. In connection with therapeutic resistance, the inhibition of either ATG5 or ATG12 can sensitize cells to therapeutic agents [180,181]. Further, it was reported that the effectiveness of radiotherapy and doxorubicin is advanced by miR-23-3p and miR-137-3p, respectively [105,114]. These miRNAs inhibit overall cell viability in vitro and enhance the ability of anti-cancer therapies to impede the in vivo growth of PaC. Such improvement of therapeutic responses is due to the fact that ATG12 and ATG5 are targeted by miR-23-3p and miR-137-3p, respectively [105,114] (Figure 2 and Table 2).

MiR-29a-3p
ATG9A functions as one of the essential components for the autophagy process by controlling the generation of phosphatidylinositol-4-phosphate, which promotes autophagosomelysosome fusions [182]. In addition, transcription factor EB (TFEB) induces autophagy via regulating the level of autophagy and lysosomal genes [183]. Both ATG9A and TFEB facilitate the process of autophagy, and they were validated as miR-29a-3p target genes in PaC. Furthermore, the sensitivity of cells to gemcitabine is increased by miR-29a-3p [106] ( Figure 2 and Table 2).

MiR-29c-5p
Ubiquitin-specific-processing protease 22 (USP22) has been recognized to promote EMT process and metastasis via activating FAK and repressing anti-cancer immunity in PaC [184,185]. USP22 also increases LC3II and autophagosome levels so that USP22 can enhance gemcitabine resistance through activating autophagy [186]. Moreover, it was revealed that miR-29c-5p increases the cytotoxic potency of gemcitabine through inhibiting USP22-mediated autophagy in vitro. In a xenograft mouse model, the overexpression of miR-29c-5p also suppresses autophagy, sensitizing PaC cells to gemcitabine [107] (Figure 2 and Table 2).

MiR-410-3p
High-mobility group box 1 (HMGB1) is capable of promoting autophagy by disengaging BCL2 from beclin-1, an autophagy factor [187]. In PaC, it was confirmed that HMGB1 promotes metastasis and that its expression is escalated in gemcitabine-resistant cells [188,189]. Furthermore, a recent study denoted that miR-410-3p targets HMGB1 to exert negative effects on gemcitabine resistance in PaC [129] (Figure 2 and Table 2).

LncRNA-AB209630
Although it is necessary to uncover the precise mechanism, it has been reported that lncRNA-AB209630 can perform tumor-suppressive functions. In hepatocellular carcinoma, the level of lncRNA-AB209630 is low, and the overexpression of lncRNA-AB209630 restrains the migration and invasion of cells [192]. Moreover, lncRNA-AB209630 significantly induces apoptotic cell death and inhibits cell proliferation, as well as invasion in hypopharyngeal cancer [193]. In this study, it was also noticed that the low expression of lncRNA-AB209630 is correlated with poor prognosis. Furthermore, it was observed that lncRNA-AB209630 suppresses proliferation, colony formation, and PI3K/AKT activities in gemcitabine-resistant PaC cells [194]. These results suggest that lncRNA-AB209630 can reverse gemcitabine resistance, at least partly via modulating pro-survival signaling ( Figure 3 and Table 3).  Table 3. CircRNA, lncRNA, and therapeutic resistance in PaC.

LncRNA Expression In Vivo Experiment and/or Clinical Relevance
Ref.

LncRNA Expression In Vivo Experiment and/or Clinical Relevance
Ref.
Increased in cancer tissues compared to adjacent normal tissues - [118] LncRNA-AB209630 Reduced in cancer tissues compared to adjacent normal controls Poor patient prognosis is associated with low lncRNA-AB209630 levels [194]

LncRNA Expression In Vivo Experiment and/or Clinical Relevance Ref.
LncRNA-GAS5 LncRNA-HCP5 High expression is detected in gemcitabine-resistant SW1990 and PANC-1 cells. Upregulated in cancer tissues compared to normal tissues Poor survival rate of patients is associated with high expression of lncRNA-HCP5 [125] LncRNA-HOTTIP Increased in cisplatin-resistant PANC-1, HS766T, and AsPC-1 cells - [115] LncRNA-PVT1 Overexpressed in cancer tissues compared to adjacent pancreatic tissues Intraperitoneal injections of gemcitabine (50 mg/kg) in mice bearing xenografts of PANC-1 cells stably expressing lncRNA-PVT1. Correlated with vascular infiltration and distant metastasis. Poor overall survival of patients with high lncRNA-PVT1 expression [135,196] LncRNA-SBF2-AS1 Abundantly expressed in gemcitabine-resistant AsPC-1 and PANC-1 cells. High expression is detected in cancer tissues compared to adjacent normal tissues High expression is correlated with lymph node metastasis and poor overall survival of patients [117] LncRNA-SLC7A11-AS1 Highly expressed in gemcitabine-resistant BxPC-3 cells.

LINC00346
LINC00346 plays a critical role in several aspects of cancer progression. LINC00346 is responsible for glioma angiogenesis by stimulating the migration and tube formation of glioma-associated endothelial cells [205]. Furthermore, LINC00346 is upregulated in colorectal cancer tissues, inhibits apoptotic cell death, and triggers cell proliferation, migration, as well as invasion [206]. In addition, LINC00346 promotes cisplatin resistance in nasopharyngeal cancer partly via sponging miR-342-5p, a tumor-suppressive miRNA [207]. In PaC, the depletion of LINC00346 renders cells susceptible to gemcitabine by increasing the level of miR-188-3p and caspase-3 activities in vitro. The inhibitory effect of gemcitabine on PaC growth is augmented by LINC00346 silencing in xenografts [121] (Figure 3, Tables 2 and 3). In support of this finding, it was observed that miR-188-3p exerts a gemcitabine-sensitizing activity by targeting bromodomain-containing 4 (BRD4) [121], which can facilitate NF-κB-dependent EMT [208].

LncRNA-SLC7A11-AS1
Nuclear factor erythroid 2-related factor 2 (NFE2L2, also called NRF2) has antioxidant properties through transcriptionally stimulating the expression of antioxidant genes such as glutathione S-transferases [218]. The expression of NFE2L2 is controlled by proteasomal degradation via the SKP1-CUL1-F-box protein (SCF) complex [219]. In addition, cancer stemness is known to be suppressed by beta-transducin repeat containing E3 ubiquitinprotein ligase (BTRC, also known as β-TrCP), one of the SCF components [220]. Recently, it was ascertained that lncRNA-SLC7A11-AS1 promotes cancer stemness via scavenging reactive oxygen species (ROS) and that the silencing of lncRNA-SLC7A11-AS1 re-sensitizes resistant cells to gemcitabine [197]. The knockdown of lncRNA-SLC7A11-AS1 strengthens the suppressive effect of gemcitabine on colony formation in vitro and the growth of PaC in vivo (Figure 3 and Table 3). Mechanically, it was proven that lncRNA-SLC7A11-AS1 binds to BTRC proteins and prevents BTRC-mediated degradation of NFE2L2 [197].

Conclusions
Efforts have been made to discover possible and efficacious combination strategies for subjugating therapeutic resistance, a prevalent and severe problem for curing cancer. In addition, it has been suggested that combination therapy using mechanistically diverse agents is beneficial for cancer treatment [223][224][225][226]. For example, ERK inhibition induces the compensatory activation of PI3K/AKT, and simultaneous PI3K inhibition synergistically augments the anti-cancer efficacy of an ERK inhibitor [226]. In this respect, targeting ncRNAs is fascinating since a single ncRNA is capable of controlling multiple signaling pathways in cells. Moreover, ncRNAs can regulate the cancer microenvironment, contributing to disease progression and therapeutic resistance [227]. A ncRNA-based therapy through the depletion or restoration of ncRNAs has been perceived to strikingly boost the effects of anti-cancer treatments in cancer [228,229]. Moreover, experimental evidence presented here demonstrated that ncRNA-based therapy is a potential strategy to surmount the therapeutic resistance of currently available treatments, such as chemotherapy and radiation therapy, in PaC.
A recent investigation indicated that miRNAs selectively advance the efficacy of drugs. For example, miR-326 strengthens the anti-cancer effect of gefitinib but not that of doxorubicin. Moreover, the effect of a miRNA on drug efficacy is different between breast cancer cell lines [230]. In addition, the application of miRNA primary/precursor forms for cancer treatments requires a concern about the opposite function of miR-3p and -5p (Section 3.1.3). Therefore, more investigations on the function of miRNAs and the relationship between miRNAs and anti-cancer agents are warranted to find highly effective combination pairs. Further, even though lncRNA-SNHG14 acts as a gemcitabine resistance factor in PaC (Section 4.2.9), lncRNA-SNHG14 is able to suppress invasion and promote apoptotic cell death via sponging miR-92a-3p in glioblastoma [231], showing its dual role in cancer (also see Section 4.2.1 about the dual role of circ-HIPK3). Regarding miR-92a-3p, it was reported that this miRNA serves as an oncogenic miRNA by accelerating cell proliferation and metastasis in PaC [232]. Additionally, circ-HIPK3 and lncRNA-TUG1 can interact with miR-421 and miR-197-3p, respectively [233,234], and both miRNAs are also ascertained as oncogenic factors in PaC [235,236]. These findings demonstrate a possibility of the sequestration of oncogenic miRNAs in other oncogenic ncRNAs. Are some oncogenic miRNAs reactivated, contributing to compensatory activation of signaling pathways in oncogenic ncRNA-depleted cells? More experimental and bioinformatic approaches for comprehensive analyses of circRNA/lncRNA-miRNA networks are necessary. Ongoing endeavors to understand the detailed feature of ncRNAs will provide unique opportunities to invent better ncRNA-based therapeutic strategies for PaC.