Loss of Interleukin-13-Receptor-Alpha-1 Induces Apoptosis and Promotes EMT in Pancreatic Cancer

In search of new therapies for pancreatic cancer, cytokine pathways have attracted increasing interest in recent years. Cytokines play a vital role in the crosstalk between tumour cells and the tumour microenvironment. The related inflammatory cytokines IL-4 and IL-13 can regularly be detected at increased levels in the microenvironment of pancreatic cancer. They share a receptor heterodimer consisting of IL-4Rα and IL-13Rα1. While IL-4Rα induces a more oncogenic phenotype, the role of IL-13Rα1 was yet to be determined. ShRNA-based knockdown of IL-13Rα1 was performed in Capan-1 and MIA PaCa-2. We assessed cell growth and migratory capacities under the influence of IL-13Rα1. Pathway alterations were detected by immunoblot analysis. We now have demonstrated that the loss of IL-13Rα1 induces apoptosis in pancreatic cancer cells. This was associated with an epithelial-to-mesenchymal transition. Loss of IL-13Rα1 also abolished the effects of exogenous IL-4 and IL-13 stimulation. Interestingly, in wild type cells, cytokine stimulation caused a similar increase in migratory capacities as after IL-13Rα1 knockdown. Overall, our results indicate the vital role of IL-13Rα1 in the progression of pancreatic cancer. The differential expression of IL-4Rα and IL-13Rα1 has to be taken into account when considering a cytokine-targeted therapy in pancreatic cancer.


Introduction
Pancreatic cancer (PC) remains one of the carcinomas with the worst prognosis [1]. Most patients diagnosed with pancreatic cancer show aggressive local growth combined with rapid development of distant metastases, where innovative surgical and medical treatments are urgently needed [2,3]. For promising approaches to detect and cure pancreatic cancer, it is crucial to understand the tumour development and progression, where the tumour microenvironment (TME) has received increasing attention [4][5][6]. The characteristic desmoplastic reaction of PC originates from a heterogeneous composition of the TME including mesenchymal and immune cells, as well as a dense collagen-based tumour stroma [7].
In our previous studies, the expression of IL-4 and IL-13 ligands, as well as IL-4Rα and IL-13Rα1 receptor chains, was shown in pancreatic cancer cell lines [24][25][26][27]. Exogenous IL-4 and IL-13 enhanced the growth of pancreatic cancer cells in a dose-dependent manner [25,26], which was inhibited by IL-4-/IL-13-neutralizing antibodies. Furthermore, overexpression of IL-13 in pancreatic cancer tissues and the high co-expression of IL-13 and IL-4Rα correlated with a higher risk of lymph node metastasis [26]. Additionally, the inhibition of IL-4Rα in Capan-1 reduced cell proliferation and migration [27].
Previous results are indicative of a contributing role of the IL-13/IL-13Rα1 axis to pancreatic cancer. However, the isolated effects of the IL-13Rα1-receptor chain on the malignant phenotype of pancreatic cancer cells and the underlying mechanisms were not studied yet.

Expression of IL-13Rα1, IL-4Rα and γc Chains in Cultured Human Pancreatic Cancer Cells
The protein levels of IL-13Rα1, IL-4Rα and γc in cultured human pancreatic cancer cell lines A818-6, AsPC-1, Capan-1, PANC-1 and MIA PaCa-2, were determined by Western blot (WB). All pancreatic cancer cell lines expressed IL-13Rα1 (47 kDa), IL-4Rα (140 kDa) and γc (64 kDa) at various levels ( Figure 1a). There, Capan-1 expressed the highest level of IL-13Rα1 and was thus chosen as target cell line for IL-13Rα1 knockdown (KD). MIA PaCa-2 expressed moderate level of IL-13Rα1 but the least levels of both IL-4Rα and γc, which indicated MIA PaCa-2 also as an attractive target cell line, potentially being more dependent of IL-13Rα1. ShRNA-based transfection targeting IL-13Rα1 was used to generate clones with reduced expression of the receptor chain. WB was performed to verify the downregulation of IL-13Rα1 with the highest efficacy in clones C-4-1 and C-4-2 (Figures 1b and S1a,b). Sham-transfected clones C-N-2 and C-N-3 showed no difference in IL-13Rα1 expression compared with Capan-1 wild type (C-WT). In MIA PaCa-2, immunoblotting revealed high efficacy of IL-13Rα1-downregulation in clones M-1-6, M-3-5, M-3-8, M-4-3 and M-4-4. Sham-transfected clones M-N-3 and M-N-4 showed no difference in IL-13Rα1 expression compared with MIA PaCa-2 wild type (M-WT) and were used as control clones in further experiments ( Figure 1c).
Furthermore, IL-4Rα expression showed no difference between C-WT, C-N-2, C-4-1, and C-4-2 (Figures 1b and S1c). Interestingly, the expression of γc was decreased in C-KD clones after transfection (Figures 1b and S1d). The effect of IL-13Rα1-downregulation on pancreatic cancer cell proliferation was investigated by cell viability assay (MTT assay) and colony formation assay. The results indicated an increasing difference in viable cells between control groups and C-KD clones in the MTT assay over time ( Figure 2a). After 72 h, significantly less viable cells were detected for clones C-4-1 and C-4-2 compared to clones C-N-2 and C-N-3 (p < 0.0001, Figures 2b and S6a). There was no significant difference among C-WT and C-N clones. Furthermore, as shown in the soft agar assay (Figures 2c,d and S6b,c), the number, as well as the size of the colonies formed by pancreatic cancer cells in soft agar after 21 days, was decreased in the IL-13Rα1-KD clones.
These findings were replicable in MIA PaCa-2 (Figures 3 and S6d-f). Thus, IL-13Rα1downregulation reduced pancreatic cancer cell survival in both anchorage-dependent and -independent assays.

Effect of IL-13Rα1-Downregulation on the Cell Cycle
Cell cycle analysis ( Figure 4a) showed no significant difference in cell cycle progression through G0/G1, S, and G2/M phase. However, C-4-1 and C-4-2 cells showed high fractions of cells in the sub-G1 phase. We can therefore conclude that the results seen before are not due to reduced proliferation but rather that the loss of IL-13Rα1 induces apoptosis of Capan-1 cells. We confirmed an increase in apoptotic cells by Annexin V staining ( Figure 4b). No consistent alterations were shown in alternative cell death pathways (S7)

Effect of IL-13Rα1-Downregulation on Cell Mobility and Migration
Migratory capacities were tested in the scratch assay first. As shown, wound healing rates of C-4-1 and C-4-2 were significantly higher than C-WT and C-N-2 ( Figure 5a). Wounds of IL-13Rα1-KD clones, unlike the control groups, were closed at 48h, which indicates that IL-13Rα1-downregulation enhances the mobility of pancreatic cancer cells ( Figures 5 and S8a). Wound healing rates of control groups and C-KD clones were further investigated after the treatment with recombinant IL-4 and IL-13. As shown, exogenous IL-4, but not IL-13 increased the wound healing rates of C-WT and C-N clones (Figure 5c,d). Wound closure of IL-13Rα1-KD clone C-4-1 was not affected by exogeneous IL-4 or IL-13. (±SEM) and (e) mean colony size of the largest 10 colonies µm 2 (±SEM) in one well (9.4 cm 2 ) of a 6-well plate. Colony number and size were automatically calculated using ImageJ 1.52a. Data shown are means of 3 independent experiments (* p < 0.05, ** p < 0.01). Abbreviations: C-IL-13-Rα1-WT: Biological replicates with normal IL-13Rα1 expression (WT and sham transfected Neo clones), C-IL-13-Rα1-KD: Biological replicates with reduced IL-13Rα1 expression (C-4-1 and C-4-2).   Furthermore, the directed migration of C-WT, C-N and C-KD clones was accessed in the Boyden chamber assay. Consistently, the migratory capacity of the IL-13Rα1-KD clones were increased (Figure 6a,b). Migration was increased 4.5-fold in C-4-1 and 2.5-fold in C-4-2 clones, respectively, compared to control cells (p < 0.0001) ( Figure S8b). Similar to above, the influence of exogenous IL-4 and IL-13-treatment was determined. Again, IL-4, but also IL-13 treatment, significantly increased the directed migration of cells with normal IL-13Rα1 expression but was without effect on C-KD clones (Figure 6c).

Effect of IL-13Rα1-Downregulation on Epithelial-to-Mesenchymal Transition (EMT)
With altered migratory capacities, the switch from an epithelial to a mesenchymal phenotype is a common finding. Morphologically, no changes were observed in Giemsa staining ( Figure S2). On a cellular level, E-cadherin and vimentin are critically involved markers in EMT [28]. The expression of E-cadherin and vimentin was compared in C-WT, C-N-2 and C-4-1 ( Figure 7). Interestingly, we found lower expression of E-cadherin and higher expression of vimentin in C-4-1 compared to C-WT and C-N-2, suggesting a more mesenchymal phenotype.

Effect of IL-13Rα1-Downregulation on IL-4 and IL-13 Signalling
Changes in the downstream signalling of the IL-4 and IL-13 axis in pancreatic cancer cells were investigated in Capan-1 (Figure 8). Baseline expression of relevant pathway components of IL-4 and IL-13 signalling (STAT3, STAT6, ERK1/2, Akt, and PI3K) were examined in Capan-1 WT cells as well as control clone C-N-2 and the IL-13Rα1 knockdown clone C-4-1. The respective pathway activation was determined by protein phosphorylation after treatment with IL-4 (0.4 nM for 30 min) and IL-13 (1 nM for 30 min). Control transfected C-N-2 showed comparable or even increased baseline expression of the analysed pathway components with C-WT cells. In the loading-control corrected expression of STAT6, STAT3, ERK and PI3K were comparable to control cells. However, C-4-1 showed a markedly reduced baseline expression of Akt (Table 1, Figure 8). Expectedly, exogenous cytokine stimulation resulted in strong pathway activation downstream of IL-4Rα and IL-13Rα1. The phosphorylation of STAT6 is the most prominent with very low baseline expression and strong activation after stimulation. Similarly, but not to the same extent, all other pathways showed increased phosphorylation.
The phosphorylation of STAT6 was markedly reduced after knockdown of IL-13Rα1, indicating the disrupted signalling through the Type II IL-4-receptor. With reduced baseline expression, phosphorylation levels of Akt were also reduced. STAT3, ERK and PI3K showed comparable activation levels after cytokine stimulation irrespective of IL-13Rα1 expression (Table 2, Figure 8)
In human samples, IL-13 was not expressed in the physiological pancreatic compartments (ductal, acinar, islets) but showed immunoreactivity in 43% of PDAC specimen and the high co-expression of IL-13 and IL-4Rα was associated with an increased risk for lymph node metastasis [26]. By analysing sequencing data from the International Cancer Genome Consortium (ICGC) from the study groups PACA-CA and -AU (Pancreatic Cancer Canadian and Australian), we found mutations in 44 of 659 patients. However, only in two cases, clinically significant single base substitutions were found, resulting in one missense mutation and one gained STOP-codon [34]. On the contrary, in differential gene expression datasets of human PDAC, IL-13Rα1 IL-4Rα and the common γc chain was consistently upregulated compared to normal controls. Dey and colleagues furthermore showed that upregulation of IL-13Rα1, IL-4Rα and γc was dependent on KRAS expression and loss of KRAS was associated with reduced receptor chain expression [35]. While the tumour promoting role of IL-4Rα on pancreatic cancer cells in vitro and in vivo has been shown before, the role and mechanism of action of IL-13Rα in pancreatic cancer is yet unclear.
In order to further investigate the role of IL-13Rα1, we screened the protein expression of IL-13Rα1 and IL-4Rα in 5 PC cells lines. Interestingly, exogenous IL-13 significantly enhanced the growth of AsPC-1 and Capan-1 cells, with high IL-13Rα1 expression, while PANC-1 and MIA PaCa-2, with low IL-13Rα1 levels, were less responsive [24,26], indicating a positive impact of the IL-13-IL-13Rα1 axis on cell survival and growth. This is contrary to IL-4, where IL-4-enhanced cell proliferation was independent of IL-4Rα expression, although IL-4-coupled toxin was more efficient in cell lines with high IL-4Rα expression [24]. Thus, not only the distinct expression levels of respective receptor chains, but also the ratio between IL-13Rα1 and IL-4Rα expression may be crucial for the effect of IL-4 and IL-13 on the cancer cell phenotype. This is of special interest as in both our previous [27] and current results, the expression of the unmodified receptor chain of the heterodimer complex IL-4Rα/IL-13Rα1 was unaltered by the downregulation of its partner. However, we were the first to show that the alternative receptor chain, the IL-2 common γ-chain was downregulated in parallel to IL-13Rα1 knockdown (KD).
IL-2 γc expression is gained in pancreatic cancer progression and ranks among the highest upregulate genes during pancreatic cancer progression [36,37]. The clinical implication of this upregulation is, however, not yet determined, as knockout of γc only moderately reduced tumour burden in vivo [35,37]. The reduced expression of γc after IL-13Rα1 knockdown in our study may be associated with a reduced oncogenic potential. However, its clinical significance warrants further studies.
In line with previous results for IL-4Rα, our data showed an inhibitory effect of IL-13Rα1-downregulation on cell viability/growth in two different pancreatic cancer cell lines. This was due to an increase in apoptotic cells, while cell cycle progression was unimpaired. Previously, our group showed that IL-13-induced growth acceleration of PC cells was associated with an increased S-phase cell fraction and reduced percentage of cells in G0/G1 [26]. Combined with our current finding with increased apoptosis after IL-13Rα1 KD, the vital role of IL-13Rα1 for PC cell survival is stressed.
IL-4 and IL-13 are believed to carry out abundant functions in tumour cells through several pro-oncogenic pathways involving signal factors such as STAT3 [25,27], STAT6 [38], PI3K/Akt [39], and ERK1/2 [39,40]. In our present study, we can confirm the involvement of STAT3, STAT6, PI3K, Akt and ERK1/2 in the response to exogenous IL-4 and IL-13. Furthermore, we demonstrated that downregulation of IL-13Rα1 in Capan-1 cells leads to a decrease in baseline expression of Akt. Furthermore, responsiveness towards ligand binding is reduced, displayed as reduced phosphorylation of mainly STAT6, Akt. As those are key mediators in regulating cell survival and growth [41][42][43], we propose that the reduced cell survival through enhanced apoptosis mechanistically is due to the suppression of STAT6 and Akt. This is in line with previous results demonstrating the activation of the Type II IL-4 receptor through IL-4 and IL-13 stimulation with phosphorylation of downstream JAK1 and STAT6 [35].
Thus, one could argue that the reduced apoptosis resistance after IL-13Rα1 KD may render the cells less cancerous. However, besides sustained proliferation and evasion of cell death, the ability to invade tissues and form metastasis is considered one of the "hallmarks on cancer" [44].
Wound healing and migration were significantly increased after IL-13Rα KD. This is paralleled by an increase in EMT markers with increased expression of Vimentin and reduced E-cadherin expression, although cellular morphology was unaltered.
Convincing evidence suggests that EMT is involved in promoting invasion and metastasis in pancreatic cancer [28]. In other visceral malignancies, IL-13 stimulation was associated with increased EMT [45,46]. Similarly, in our results, the exogenous stimulation resulted in increased cell migration. This was especially obvious in the difference between the non-directed migration in the wound healing assay and the directed migration in the Boyden chamber assay. IL-13 stimulation consistently increased migration in the Boyden chamber assay in both WT and control cells, while this effect was far less pronounced in the wound healing assay.
However, we are the first to report that the loss of IL-13Rα1 also promotes an EMT phenotype. This is a highly interesting new finding as the results of exogenous cytokine stimulation needs to be re-evaluated. Possibly, the promoted EMT phenotype through IL-4/-13 stimulation is due to cellular signalling mainly through the IL-4Rα receptor and the relative loss of IL-13Rα1 further promotes this phenotype. However, the precise mechanisms underlaying these observations require further experiments.
Up to date, IL-13-focussed treatments were studied in clinical trials, which utilized the Pseudomonas Exotoxin coupled IL-13. These trials, such as the PRECISE Trial (randomized controlled Phase III clinical trial), were conducted in Glioblastoma multiforme patients and achieved prolonged time to progression in treated patients [47]. In pancreatic cancer, IL-13 or IL-4 is not studied in clinical trials currently. However, the RECAP Trial (NCT01423604) utilizes ruxolitinib, a JAK1/JAK2-Inhibitor, targeting the downstream pathways of IL-13 and IL-4. In metastatic PDAC, ruxolitinib was able to prolong overall and progression free survival [48].
Overall, we can conclude that IL-13Rα1 is vital for cell survival and apoptosis resistance. However, its loss induces an EMT phenotype and consistently promotes cell migration. The more detailed knowledge in IL-4/-13 signalling we have received from this study helps in designing more promising clinical studies, as the multiple functions and cross-play of ligands and receptor chains need to be taken into account.

Cell Lines and Cell Culture
Human pancreatic cancer cell lines A818-6, AsPC-1, Capan-1 and were cultured in RPMI (Roswell Park Memorial Institute medium). MIA PaCa-2 and PANC-1 were cultured in DMEM (Dulbecco's Modified Eagle's Medium). All media were supplemented with 10% foetal calf serum (FCS), 1% Penicillin (10,000 U/mL)/Streptomycin (10,000 µg/mL). The maintenance media for transfected clones containing a neomycin resistance gene were supplemented with 550 and 1100 µg/mL geneticin (G418) sulphate, respectively. Cells were cultured in 100 mm cell culture dishes and maintained in monolayer culture at 37 • C in humidified air with 5% CO 2 . All cells were tested as mycoplasma-free.

Immunoblotting
Western blotting was performed as previously described [27]. Cultured cells at around 80% confluence were washed twice with ice-cold DPBS (Dulbecco's Phosphate Buffered Saline) and were incubated with lysis buffer for 30 min on ice. Protein concentration was measured using the Pierce ® BCA Protein Assay kit (Thermo Fisher Scientific

Cell Growth Assay
The basal anchorage-dependent growth of cultured cells was determined by the MTT colorimetric assay as described before [27]. Briefly, 10,000 cells/well were seeded and viable cells were detected after incubation with the MTT reagent at the absorbance of 570 nm after 24, 48, 72, and 96 h.
Colony formation assay was performed to assess the basal anchorage-independent growth of cancer cells. For the base layer, 2 mL of 0.9% agar solution was gently added into each well of a prewarmed 6-well plate. When the base agar solution was solidified, 4000 vital cells were gently resuspended in 0.35% agar solution and added onto the base layer. After the top agar solution solidified, plates were maintained at 37 • C in 5% CO 2 atmosphere. After 21 days, 9 photos were taken per well, as shown in Figure S4. Afterwards, the mean number of colonies and mean colony size were measured by ImageJ 1.52a.

Cell Migration Assay
Cell movement was studied in the scratch assay. Confluent cells in a monolayer were scratched to make equidistant wounds by yellow tips, as shown in Figure S5. Distances of the wounds were recorded in quadruplicate by taking pictures at defined positions. Gap distances were quantitatively evaluated by ImageJ 1.52a. The wound healing rate was determined as (A − B)/A × 100%, where A was the primary wound width and B was the wound width after 24 or 48 h.
The modified Boyden Chamber assay was performed to investigate cell migration as mentioned before [27]. Then, 5 × 10 4 cells suspended in 100 µL of medium containing 1% FCS were seeded into each insert, which was placed in the 24-well plate, as shown in Figure S6. Non-migratory cells were scraped off with wet cotton swabs after 24 h, while migratory cells on the underside of the membrane were rinsed by dH 2 O, fixed with 4% paraformaldehyde and stained with DAPI for 5 min. Afterwards, fluorescence photographs were taken at 6 random positions at 10x magnification. Migratory cells were counted using ImageJ 1.52a.

Giemsa Staining Assay
Exponentially growing cells in 100 mm dishes were rinsed by 10 mL of DPBS, fixed in 5 mL of methanol for 15 min, and then stained in Giemsa staining solution (Giemsa Stain, Sigma-Aldrich, Taufkirchen, Bavaria, Germany, diluted with dH2O in the ratio of 1:20) for 15 min. Cell morphology was observed and recorded by taking photos under an inverted light microscope.
Annexin V-FITC Kit (Miltenyi Biotec, Bergisch Gladbach, North Rhine-Westphalia, Germany) was used to detect apoptotic cells. Experiments were performed according to the protocol supplied by the manufacturer. Apoptotic cells are stained positively for Annexin V-FITC but are negative for staining with PI.

Statistics
Statistical analysis was performed using GraphPad Prism 8.0.1 (GraphPad Software, San Diego, California, USA). Paired t test, ratio paired t test, Tukey's multiple comparisons test and uncorrected Fisher's LSD were used for evaluating group differences. p values <0.05 were taken as level of significance. p values are shown as follows: ns p > 0.05, * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001.

Conclusions
Overall, IL-13Rα1 plays a critical and diverse role in the survival and migration of cultured pancreatic cancer cells. The findings of this study may help to better understand the different functions and mechanisms involving IL-13Rα1 in pancreatic cancer progression. As cytokines such as IL-4 and IL-13 play a vital role in the interaction of tumour cells and components of the TME, their understanding is crucial in order to design better therapies for pancreatic cancer. Acknowledgments: Thanks for the technical support and kind help of Pengfei Xu, Nadine Suessner, Fabian Gärtner, Adrian Gihring and Aileen Roth.

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