Combinational blockade of MET and PD-L1 improves pancreatic cancer immunotherapeutic efficacy

Dysregulated expression and activation of receptor tyrosine kinases (RTKs) are associated with a range of human cancers. However, current RTK-targeting strategies exert little effect on pancreatic cancer, a highly malignant tumor with complex immune microenvironment. Given that immunotherapy for pancreatic cancer still remains challenging, this study aimed to elucidate the prognostic role of RTKs in pancreatic tumors with different immunological backgrounds and investigate their targeting potential in pancreatic cancer immunotherapy. Kaplan–Meier plotter was used to analyze the prognostic significance of each of the all-known RTKs to date in immune “hot” and “cold” pancreatic cancers. Gene Expression Profiling Interactive Analysis-2 was applied to assess the differential expression of RTKs between pancreatic tumors and normal pancreatic tissues, as well as its correlation with immune checkpoints (ICPs). One hundred and fifty in-house clinical tissue specimens of pancreatic cancer were collected for expression and correlation validation via immunohistochemical analysis. Two pancreatic cancer cell lines were used to demonstrate the regulatory effects of RTKs on ICPs by biochemistry and flow cytometry. Two in vivo models bearing pancreatic tumors were jointly applied to investigate the combinational regimen of RTK inhibition and immune checkpoint blockade for pancreatic cancer immunotherapy. MET was identified as a pancreatic cancer-specific RTK, which is significantly associated with prognosis in both immune “hot” and “cold” pancreatic cancers. MET was observed to be highly upregulated in pancreatic cancer tissues, and positively correlated with PD-L1 levels. Elevated MET and PD-L1 expressions were closely associated with lymph node metastasis, tumor TNM stage, and overall survival in pancreatic cancer. Mechanistically, MET could interact with PD-L1, and maintain its expression level in multiple ways. MET deficiency was found to facilitate lymphocyte infiltration into pancreatic tumors. Finally, significant benefits of combining MET inhibition with PD-1/PD-L1 blockage were verified in both orthotopic and subcutaneous mouse models of pancreatic cancer. This study systematically investigated the potential effectiveness of a novel pancreatic cancer immunotherapy targeting RTKs, and revealed the function of MET in PD-L1 regulation as well as the combined therapeutic efficacy of MET and PD-L1 in pancreatic cancer.


Background
As an aggressive malignant cancer driven by multiple molecular mechanisms, pancreatic cancer is associated with an extremely high mortality rate [1][2][3]. Pancreatic ductal adenocarcinoma (PDAC), which is the most common type of pancreatic cancer, has become the fourth deadliest malignancy in Western countries [4,5]. Further, given the late diagnosis, rapid metastatic progression, resistance to conventional therapeutics, and increased incidence rate of pancreatic cancer, it is expected to become the second leading cause of cancerrelated deaths in the United States by 2030 [6]. This poses a grave threat to human health and survival. Thus, it is imperative to identify more efficacious therapies against this dreadful disease, especially because of its intrinsic resistance to the immune system [7][8][9].
Receptor tyrosine kinases (RTKs) are high-affinity cell surface receptors that mediate environmental signal transmission and cellular responses [10][11][12][13]. Dysregulated expression and aberrant activation of RTKs are associated with a range of human disease-related processes, especially tumorigenesis and cancer progression [14,15]. Accordingly, multiple RTK-targeting drugs have been developed and applied, benefiting numerous cancer patients [16]. However, current RTK-targeting therapies unfortunately exert little effect on pancreatic cancer. Epidermal growth factor receptor (EGFR) blockage in PDAC previously reported with disappointing outcomes and dampened the enthusiasm in the examination of RTKs for pancreatic cancer therapy [17]. Moreover, the association between RTKs and the immune status in pancreatic cancer still remains unclear. Accordingly, to enhance our understanding of the efficacy of RTK-targeting therapy, further investigation with regard to the selection of targets and therapeutic strategies is warranted.
MET is a classical RTK and the receptor of hepatocyte growth factor (HGF) [18][19][20]. MET was observed to be upregulated in pancreatic cancer, and that was associated with tumor grade [21]. Further, the HGF-MET axis is involved in the metastatic progression of pancreatic cancer and functions as a bridge linking the tumor with stroma, resulting in a positive loop for pancreatic cancer progression [21]. In view of these findings, there is continuous research on targeting HGF-MET signaling in pancreatic cancer with MET overexpression. However, few therapeutic outcomes have been obtained from these studies because of HGF-MET-targeting drug resistance in pancreatic cancer [18,22]. Recent reports fortunately provide novel directions for pancreatic cancer treatment. Noguchi et al. reported that the upregulation of MET induces gemcitabine resistance in mice with pancreatic cancer [23]. Tomihara et al. showed that radiation therapy upregulates MET and activates its downstream signaling in PDAC [24]. Rucki et al. suggested that treatment strategies targeting HGF/MET and Hedgehog pathways alleviate monodrug resistance; they further reported that HGF/MET-targeted drug therapy in combination with chemotherapy shows more significant antitumor effects compared with monotherapy [25]. These data indicate that combination therapy enhances the therapeutic efficacy of MET-targeted drugs and may serve as a new treatment strategy for pancreatic cancer.
The property of the immune system to discriminate tumor-derived antigens as non-self and to be activated to eliminate cancer cells indicates a connection of tumor immunotherapy with MET-targeted therapy [19,26]. Immune checkpoint (ICP) inhibitors targeting programmed cell death protein 1 (PD-1)/PD-1 ligand 1 (PD-L1) are currently the most popular weapon in cancer immunotherapy. PD-L1 binds to PD-1 to inhibit T cell activation and cytokine production, resulting in cancer cell immune tolerance and survival [27,28]. PD-L1targeted drugs block this interaction and reactivate the immune system to attack and eradicate tumor cells in cancer patients with PD-L1 overexpression; this has benefited a proportion of patients in multiple cancers [29,30]. However, the therapeutic efficacy of PD-L1 blockage remains limited in pancreatic cancer [31]. In clinical trials, the administration of ipilimumab and BMS-936,559 (an anti-PD-L1 monoclonal antibody) showed poor anti-tumor effects in advanced pancreatic cancer cases [32][33][34]. However, an increasing number of preclinical studies have shown that combination therapy effectively improves the efficacy of PD-L1 blockage. For example, Pan et al. reported that JQ1 and anti-PD-L1 antibody synergistically inhibited PDAC [35]. Mace et al. showed that joint blockage of IL-6 and PD-L1 suppresses pancreatic cancer growth in mice and elevates the levels of intra-tumoral effector T cells [36]. Therefore, exploring the regulatory mechanisms associated with PD-L1 expression could help identify biomarkers for predicting the therapeutic response after ICP blockage (ICB) in PDAC.
In this study, we systematically investigated the potential effectiveness of targeting RTKs as a novel strategy for pancreatic cancer treatment. We first focused on elucidating the prognostic role of all known RTKs in pancreatic tumors with different immunological backgrounds. Furthermore, we identified pancreatic cancer-specific RTKs, which are significantly upregulated in pancreatic tumors and associated with ICPs. Most importantly, we determined the endogenous correlation of MET with PD-L1 and the function of MET in the regulation of PD-L1 expression as well as the efficacy of their combination therapy in pancreatic cancer treatment. Thus, our study provides novel insights for the selection of targets and treatment strategies for pancreatic cancer patients.

Data collection
To analyze prognoses and expression levels of RTKs as well as to determine their correlation with ICPs in pancreatic cancer, the Pancreatic Adenocarcinoma datasets from TCGA database (http://cancergenome.nih.gov) were collected by two individual web servers.

Kaplan-Meier survival curve analysis
The Kaplan-Meier plotter (KM-plotter, http://kmplot. com, updated on 2020.07.22) [37] was used to analyze the effects of each RTK on the prognosis of immune "hot" and "cold" pancreatic cancers. The KM-plotter can assess the effect of 54k genes (mRNA, miRNA, and proteins) on survival in 21 cancer types. The database sources include GEO, EGA, and TCGA. The primary purpose of this tool is meta-analysis-based discovery and validation of survival biomarkers. In this study, KM-plotter was used to analyze the prognostic roles of each RTK in both immune "hot" (CD8 + T-cell enriched) and "cold" (CD8 + T-cell decreased) pancreatic cancers. Overall survival analysis of each RTK was performed using the Kaplan-Meier method with the best cutoff to divide patients into lower and higher expression groups. The log-rank test (Mantel-Cox test) was used for hypothesis testing. The cox proportional hazards regression model was applied to calculate the hazard ratio, and a p-value < 0.05 was considered to be statistically significant.

Gene expression profiling interactive analysis
Gene Expression Profiling Interactive Analysis (GEPIA, http://gepia2.cancer-pku.cn, version 2) [38] is an openaccess online tool for the interactive exploration of RNA sequencing expression data of 9,736 tumors and 8,587 normal samples from the TCGA and GTEx projects. GEPIA uses the UCSC Xena project-based datasets (http://xena.ucsc.edu), following a standard processing pipeline to avoid data imbalance [39,40]. In this study, GEPIA was used to analyze the differential expressions of RTKs between pancreatic cancer tissues and normal pancreatic tissues, as well as the correlations of RTKs with ICPs in pancreatic cancer. One-way ANOVA was used to analyze the differential expression of RTKs, and genes with |log2FC| values > 1 and q values < 0.05 were considered to be differentially expressed. For correlation analyses, we focused on the RTKs that were upregulated in pancreatic cancer tissues compared with normal pancreatic tissues, and six most well-established inhibitory ICPs on the tumor side were selected and integrated as the ICP signature in our study; these included PD-L1 (also known as CD274 and B7H1), CD276 (also known as B7H3), CD155 (also known as PVR), CD112 (also known as NECTIN2 and PVRL2), LGALS9 (also known as Galectin 9), and HVEM (also known as CD258 and LIGHT). Spearman correlation analysis was used to analyze the pair-wise gene expression correlations between RTKs and ICP signature, and a p-value < 0.05 was considered to indicate a statistically significant difference.  B Expression summary of RTKs. All 26 RTKs significantly associated with the prognosis of immune "hot" pancreatic cancer were subjected to differential expression analysis. Up, upregulated; Ns, not significant. C Correlation analysis between the expression of MET and ICPs in pancreatic cancer. D Correlation summary between RTKs and ICPs. All 17 RTKs significantly upregulated in immune "hot" pancreatic cancer were subjected to correlation analysis  qPCR was conducted based on a previous report [41]. Total RNA was isolated using TRIzol reagent (Invitrogen, USA). Following this, cDNA was synthesized using PrimeScript RT reagent Kit (Takara). SYBR Green Real-Time PCR Master Mix (Takara) was used for qPCR on an Applied QuantStudio™ 5 System, as directed by the manufacturer. The 2 −ΔΔCt method was followed for the analysis. Primers for human c-MET (Cat HP100082) and PD-L1 (Cat HP100170) were provided by Sino Biological (Beijing, China).

Western blotting
Immunoblotting was performed based on a previous report [41] and has been briefly described below. Total protein for Western blotting was obtained using radioimmunoprecipitation assay (RIPA) lysis buffer supplemented with protease inhibitor and phosphorylase inhibitors, separated via sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and electrotransferred onto PVDF membranes (Millipore, USA). overnight at 4°C, and corresponding biotinylated secondary antibodies for 2 h at ambient temperature. DAB kits were used for detection, and counterstaining was performed using hematoxylin and eosin before analysis under a microscope (Leica, Germany). The immunostained tissues were reviewed and graded by two independent pathologists. The samples were scored using the histochemistry score (H-score) method combining the values of immunoreaction intensity and percentage of tumor-cell staining.

Mouse models
Male C57BL/6 mice (6 weeks) provided by Shanghai Experimental Animal Center (Shanghai, China) were housed under Specific-pathogen-free (SPF) conditions. All animal protocols in this study were approved by the Animal Care and Use Committee of the Zhejiang University School of Medicine. In xenograft studies, 5 × 10 5 KPC cells were subcutaneously or orthotopically injected. Following this, xenograft tumors were assessed at three-day intervals. When the tumor volume reached approximately 30 mm 3 , the mice were intraperitoneally injected with PD-L1 antibody (200 µg, Bio X Cell) or PD-1 antibody (100 µg, Bio X Cell) once every three days or orally administered capmatinib daily (with 0.5 % methylcellulose and 5 % dimethyl acetamide) at a dose of 10 mg/kg. The tumor volume was determined as follows: (length × width 2 )/2. After two weeks of treatment, the mice underwent CO 2 inhalation for euthanasia, and tumor specimens were collected.

Statistical analyses
SPSS v22 (SPSS, USA) and GraphPad Prism 7 (Graph-Pad Software, USA) were used for data analysis. Data are presented as mean ± standard deviation (SD). Continuous and categorical variables were compared by unpaired Student's t-test and the Chi-squared or Fisher exact test, respectively. Overall survival was compared by the Kaplan-Meier method, and significance was determined by log-rank test. P < 0.05 was considered statistically significant.

RTKs are potential immunotherapeutic targets in pancreatic cancer
To further investigate whether any of the 26 RTKs associated with immune "hot" pancreatic cancer are suitable immunotherapeutic targets, we compared the expression level of each RTK between pancreatic tumor tissues and normal pancreatic tissues. We found that 17 RTKs, including MET, were significantly upregulated in pancreatic cancer tissues compared with normal pancreatic tissue, suggesting that they are potentially viable therapeutic targets ( Fig. 2A and B). Given that several previous studies have documented the direct and indirect regulatory effects of RTKs on ICPs, the most critical mediators for tumor immune resistance, we further analyzed the correlation between upregulated RTKs and representative ICPs (PD-L1, CD276, CD155, CD112, LGALS9, and HVEM) in pancreatic cancer and found that 16 RTKs significantly positively correlated with ICPs in terms of expression level (Fig. 2C and D). Additionally, we also investigated the correlation between each RTK and each ICP, and the results showed significantly positive associations between most of the RTKs and ICPs in pancreatic cancer. This suggests that RTKs and ICPs may synergistically contribute to immune resistance and thus should be co-targeted during pancreatic cancer treatment.

MET is significantly associated with PD-L1 expression in pancreatic cancer
Based on abovementioned in silico data, we selected MET and PD-L1 a representative ICP for further investigation. We first evaluated the expression levels of MET and PD-L1 in paired PDAC and para-cancerous tissues and found that both were markedly upregulated in PDAC tissue (Fig. 3A-C and Figures S1A-S1B). Further analysis showed a positive correlation between the expression levels of MET and PD-L1 (p < 0.0001, R = 0.8601) (Fig. 3D). We then examined MET and PD-L1 levels in tumor tissue microarray from 140 PDAC patients (Fig. 3E) and found that approximately 68 % specimens with high MET expression levels exhibited intense PD-L1 staining, confirming the positive correlation between them in pancreatic cancer (Fig. 3F). Subsequently, we assessed the associations between MET and PD-L1 expressions in terms of clinical factors and found that elevated MET and PD-L1 levels were significantly associated with lymph node invasion (p = 0.034), and an advanced TNM stage (p = 0.006) ( Table 2). Moreover, patients with low MET and PD-L1 expression levels showed significantly prolonged survival compared with those with high MET and PD-L1 levels (Fig. 3G). Taken together, these data suggest that the MET-PD-L1 axis indeed functions as a prognostic marker in PDAC.

MET maintains PD-L1 expression in pancreatic cancer cells
To determine the association of MET with PD-L1 in pancreatic cancer, we first investigated the influence of MET deficiency on PD-L1 expression. As excepted, PD-L1 (Fig. 4A-B)  expression were starkly down-regulated in the shMET group. Moreover, capmatinib, a small-molecule tyrosine kinase inhibitor (TKI) with high specificity for MET [44,45], also induced PD-L1 downregulation in a dosedependent manner (Fig. 4E, F). During the investigation for a potential regulatory mechanism, we found that the mRNA level of PD-L1 was remarkably reduced in the shMET group ( Figure S2A). In addition, we observed decreased phosphorylation of signal transducer and activator of transcription 1 (STAT1), a critical transcriptional factor for PD-L1, and PD-L1 downregulation when MET was knocked down ( Figure S2B). In support of this, IFNγ-induced phosphorylation of STAT1 and PD-L1 upregulation were inhibited in shMET group ( Figure  S2B, Figures S3A-S3B). Further, the administration of capmatinib significantly alleviated IFNγ-associated PD-L1 upregulation on the cell membrane ( Figure S3C). Thus, MET-STAT1 axis promoted the transcription of PD-L1. In addition to the mRNA level, the half-life of PD-L1 was shortened in the shMET group, suggesting that MET can maintain the stability of PD-L1 (Figures S4A-S4B). Further exploration showed an endogenous interaction between MET and PD-L1 (Fig. 4G, H) as well as between MET and CMTM6 ( Figure S4C), a principal regulator of PD-L1. CMTM6 was found to be required for preventing PD-L1 from endo-lysosomal pathwaymediated degradation, as well as for interacting with PD-L1 to reduce its ubiquitination [46,47]. The expression level of CMTM6 was decreased when MET was knocked down ( Figure S4D). Moreover, CMTM6 overexpression abrogated the down-regulation of PD-L1 caused by MET deficiency ( Figure S4E). Thus, MET maintained the stability of PD-L1 in a CMTM6dependent manner. Taken together, MET maintained the expression of PD-L1 at least at transcriptional and post-transcriptional stages in pancreatic cancer cells.

MET deficiency inhibits tumor growth and enhances immune cell infiltration in pancreatic cancer
To determine whether MET is indeed associated with pancreatic tumor immunity, mouse PDAC KPC cells stably transfected with shMET and CTRL shRNA were individually implanted into immunocompetent C57BL/6 mice. The growth of tumors was significantly decelerated in the shMET group compared with that in the CTRL group (Fig. 5A-B). We further found that CD3 + T-cell infiltration was significantly increased in the shMET group (Fig. 5C). Similarly, the subpopulation of GZMB+, IFNγ+, and TNF-α + in infiltrated CD8 + T lymphocytes also showed significant upregulation in the shMET group (Fig. 5D, F). Taken together, these results indicate that MET silencing indeed suppressed PDAC growth in an immune-related manner.

Dual-targeting MET and PD-L1 cooperatively inhibits pancreatic cancer
Based on the regulatory role of MET in PD-L1 expression, we hypothesized that MET inhibition improves the therapeutic efficacy of PD-L1/PD-1 blockage. To prove this, we applied three therapeutic strategies in KPC tumor-bearing orthotopic mice, including monotherapy with capmatinib, anti-PD-L1 antibody treatment, or a combination of both (Fig. 6A). First, no significant weight loss was observed in all experimental mice (Fig. 6B). Further, as expected, the combination therapy showed more significant inhibitory effects on tumor size and burden compared with the control and both monotherapies ( Fig. 6C-E). Moreover, we observed significant elevation of tumor-infiltrating CD3+, CD4+, and CD8 + T cells but not CD11B + myloid cells in mice administered combination therapy, suggesting that strategies coblocking MET and PD-L1 synergistically potentiate antitumor immunity in mice ( Fig. 6F-G, Figure S5A-5B). We also assessed the efficacy of combination treatment and monotherapy in a subcutaneous KPC PDAC model (Fig. 7A). Consistent with KPC-bearing orthotopic model, no significant weight change was observed after combination treatment (Fig. 7B), and significant inhibitory effects were observed on tumor size and tumor weight even in this experimental model (Fig. 7C-E). Moreover, the numbers of tumor-infiltrating CD3 + T cells and CD8 + T cells but not CD4 + T cells and CD11B + myloid cells, were also significantly elevated in mice administered with a combination of capmatinib and anti-PD-L1 antibody (Fig. 7F-G, Figures S5C-5D).
We also administered capmatinib or PD-1 antibody alone or in combination to mice bearing subcutaneous KPC tumors ( Figure S7A). We found that no significant weight loss was observed in all experimental mice (Figure S6B). The combination of capmatinib and anti-PD-1 impaired tumor growth and decreased tumor burden more effectively than control treatment and capmatinib or anti-PD-1 monotreatment (Figures S6C-S6E). The  (Figures S6F-S6G). Taken together, these findings indicate that MET inhibition and PD-L1/PD-1 pathway blockage cooperatively inhibited tumor growth.

Discussion
To the best of our knowledge, this is the first study presenting a comprehensive analysis investigating an association between RTKs and pancreatic cancer prognosis, with the consideration of the immune status of patients. Moreover, in this study, we characterize the relationships between the expression levels of RTKs and ICPs in pancreatic cancer. RTKs are potential immunotherapeutic targets for immune "hot" pancreatic cancers. As shown above, the expression of different RTKs is associated with the differential prognosis of pancreatic cancer. Based on this knowledge, we believe that the treatment strategy for each patient should be individualized via precision medicine. Specifically, we identified three RTKs, MET, EPHB6, and RYK, which are associated with a poor prognosis regardless of the immune status in patients with pancreatic cancer. We suggest that further preclinical validation of the inhibitory potential of these RTKs be conducted using blockage antibodies or small molecule inhibitors. In contrast, we identified two RTKs, EPHA10 and ALK, which are associated with a good prognosis in patients with pancreatic cancer regardless of the immune status. For these RTKs, we suggest that strategies to activate them, rather than inhibit them, using natural ligands may be therapeutically beneficial. Of note, we also identified RTKs (e.g., FLT1, also known as VEGFR1) showing contrasting effects on prognoses between immune "hot" and "cold" cancers. For these RTKs, we suggest that the infiltration of CD8 + T cells be detected before deciding how to manage these patients.
Recently, Li et al. have reported that MET suppression upregulates PD-L1 in liver cancer, and MET knockout (KO) induces PD-L1 expression by preventing GSK3Bassociated PD-L1 degradation rather than by affecting its transcription [48]. In contrast, our report first revealed that high MET expression has a positive correlation with PD-L1 upregulation in PDAC. This is consistent with another previous study reporting that elevated PD-L1 expressions are correlated with high MET levels in metastatic clear cell renal cell carcinoma [49]. Interestingly, we showed that MET deficiency resulted in reduced protein stability of PD-L1, thereby lowering PD-L1 expression in pancreatic cancer cells. Burr et al. [47] identified CMTM6 as a critical regulator of PD-L1 in a broad range of cancer cells, and reported that CMTM6 is required for efficient endocytic recycling of PD-L1, thus preventing PD-L1 degradation in lysosomes. Mezzadra et al. [46] demonstrated that CMTM6 associates with the PD-L1 protein, reducing its ubiquitination and increasing PD-L1 protein half-life. In this study, MET deficiency-induced PD-L1 downregulation could be restored by overexpressing CMTM6. In addition to decreased stability, MET deficiency could induce mRNA level downregulation of PD-L1, which might be associated with the inactivation of STAT1. These results imply that MET regulates PD-L1 expression probably at the transcriptional, translational, and post-translational levels. Additionally, it is well-established that IFNγ upregulates PD-L1 expression to promote cancer immune resistance [50,51]. Previous studies have mainly focused on IFNγ-induced activation of IRF1, which directly interacts with the PD-L1 promoter to enhance its transcription via the JAK-STAT signaling pathway [52]. A recent study reported that IFNγ-associated PD-L1 upregulation in medulloblastoma is mediated by CDK5, the depletion of which downregulates PD-L1 in malignant cells [53]. In this study, we revealed that MET may be a crucial mediator of IFNγ-associated PD-L1 upregulation in PDAC. Taken together, these results suggest that MET-mediated PD-L1 regulation likely involves several proteins and has different activation mechanisms in different cancers.
To date, anti-PD-1/PD-L1 administration has been regarded as one of the most successful immunotherapeutic options in various tumors [54]. However, its overall response rate usually does not exceed 40 % across multiple cancer types because relevant regulatory mechanisms of PD-L1 expression potentially influence the efficacy of ani-PD-L1 therapy [55]. Increasing evidence reveals that cotargeting these regulatory pathways with PD-L1 can improve the efficacy of immunotherapy in patients [56,57]. Capmatinib (INC280), an efficient and specific MET suppressor, is a promising anti-tumor drug [58]. This was proved in a phase II study in individuals with advanced liver cancer who expressed high levels of MET [59]. In the current study, tumor growth was inhibited in the monocapmatinib group compared with the controls. Moreover, capmatinib enhanced the efficacy of anti-PD-L1 in mouse models, indicating the potential benefits of a combined therapeutic strategy for treating pancreatic cancer. However, the vast majority of PDAC patients are diagnosed at advanced or metastasized stages. Compared with this, the treatment for mice starts early, which is discrepant with the real-word patient situation. Therefore, further studies targeting RTKs are warranted in the future for the clinical application of our findings.

Conclusions
In conclusion, this study showed that dual blocking of MET and PD-L1 enhances pancreatic cancer immunotherapy. Despite the conflicting correlation between RTKs and ICPs in other cancer types [26,48,60,61], we have taken critical first steps to improve therapeutic options in pancreatic cancer. Anomalous expression of RTKs may be used as a prognostic factor to design tailored therapeutic strategies for individual cancer patient. However, to provide immediate clinical significance, this work will need to be extended to clinical studies to validate the strategy of RTK targeting in pancreatic cancer in combination with ICP therapy, as appropriate.