Comparing VIP and PD‐L1 expression as cancer biomarkers

Immune checkpoint molecules are critical targets of cancer therapies due to their ability to modulate immune responses to cancer. Vasoactive intestinal peptide (VIP) has been proposed as an immune checkpoint molecule, but its predictive and prognostic values have not been established. We evaluated expression levels of VIP and programmed death‐ligand 1 (PD‐L1) across different cancer types and identified specific cancer histologies in which the expression of these markers is elevated. We conducted systematic analyses of the prognostic and predictive values of VIP and PD‐L1 in various cancers using publicly available patient databases and analysis tools including the Gene Expression Profiling Interactive Analysis, PrognoScan, Protein Atlas, cBioportal, and Timer2.0. We also assessed the relationship of PD‐L1 and VIP expression levels with survival and the frequencies of tumor‐infiltrating immune cells in various cancers. We observed a negative correlation between PD‐L1 and VIP expression across cancer types, suggesting the functional redundancy of VIP and PD‐L1 immunosuppressive pathways as mechanisms of immune escape. High expression levels of VIP and the association of VIP expression with immune cell infiltrates in the pancreatic adenocarcinoma tumor microenvironment suggest that VIP may be a predictive biomarker for treating pancreatic adenocarcinoma patients with drugs that inhibit the VIP signaling pathway.

antitumor immunity. More recently, the expression of PD-L1 by cancer cells and the cognate PD1 receptor on T cells have been identified as clinically relevant immune checkpoint molecules in cancers. Antibodies targeting this pathway have been approved for numerous malignancies including melanoma, non-small cell lung cancer, renal cell carcinoma, Hodgkin lymphoma, bladder cancer, head, and neck squamous cell carcinoma, and Merkel cell carcinoma, and are currently being evaluated for the treatment of other cancers. 2 Recent preclinical studies have identified vasoactive intestinal polypeptide (VIP), a 28-amino acid neuropeptide, as a potential ICI target. VIP has previously been shown to be a potent suppressor of T cell activation and proliferation. As such, VIP antagonists were found to enhance adaptive immunity to cytomegalovirus infection in murine models. 3 Past studies also show that VIP and its receptors were overexpressed in breast, prostate, and lung cancer, leading to the growth and metastasis of tumors. [4][5][6][7][8] T cells have been found to upregulate VIP receptors during T cell activation. In response to this VIP receptor signaling, T cell activation and proliferation become inhibited while Treg and Th2 cells were generated. [9][10][11] Concerning cancer immunotherapy VIP antagonists enhanced anti-leukemia T cell response and downregulated myeloid-derived suppressor cells in acute myeloid leukemia murine models. 12 Furthermore, VIP expression by pancreatic ductal adenocarcinoma cells was identified as a possible immunomodulator that protects cancer from immune surveillance by decreasing T cell proliferation, trafficking, activation, and function. Thus, VIP production by cancer cells has been proposed as an immune escape pathway used by some cancers. A multitude of clinical studies has proved the therapeutic value of anti-PD-L1/PD-1 antibody treatments in cancers with overexpressed PD-L1. 13 However, treatment targeting PD-L1 and PD-1 are traditionally believed to have low efficacy in cancers with low PD-L1 expression such as pancreatic adenocarcinoma. 14 The objective of this paper is to evaluate the relative expression levels of PD-L1 and VIP, compare the prognostic values of PD-L1 and VIP expression in cancers, and identify cancers that might be responsive to agents that target VIP signaling pathways. The over-arching hypothesis motivating this study is that drugs targeting VIP signaling might be effective in cancers that express high levels of VIP but low levels of PD-L1.

Differential expression of VIP and PD-L1 across multiple human cancers
To assess the aberrant expression of VIP and PD-L1, VIP and PD-L1 RNA-seq transcripts per million (TPM) data from the TCGA and GTEx were analyzed across various cancer histology. We used GEPIA analysis to compare PD-L1 and VIP expression in cancer tissues to adjacent normal tissues across various cancer types. Data on gene expression in cancer tissue was obtained from the TCGA. Additionally, data on gene expression in adjacent normal tissue in cancer patients was obtained from both the TCGA and GTEx databases, as GEPIA did not have a sufficient sample size of normal tissue data. As shown in Figure 1, lung adenocarcinoma, lung squamous cell carcinoma, breast invasive carcinoma, colon adenocarcinoma, skin cutaneous melanoma, and glioblastoma multiforme tissue had lower levels of VIP expression than adjacent normal tissues. In contrast, pancreatic adenocarcinoma had higher VIP expression compared to adjacent normal tissue. Acute myeloid leukemia and diffuse large B cell lymphoma had similar VIP expression between cancer and normal tissue. In contrast, PD-L1 expression was higher in pancreatic adenocarcinoma, colon adenocarcinoma, skin cutaneous melanoma, glioblastoma multiforme, acute myeloid leukemia, and diffuse large B cell lymphoma when compared to corresponding normal tissues, while lower in lung adenocarcinoma, lung squamous cell carcinoma, and breast carcinoma when compared to corresponding adjacent normal tissues. Additionally, compared with other cancer types, VIP is relatively overexpressed in pancreatic adenocarcinoma and colon adenocarcinoma, while PD-L1 is overexpressed in lung adenocarcinoma, lung squamous cell carcinoma, and diffuse large B cell lymphoma. Overall, both VIP and PD-L1 expression differ across different cancers, and some cancers had higher levels of VIP or PD-L1 compared to adjacent normal tissues. These results suggest that cancers may be selectively targeted by VIP or PD-L1 pathway inhibitors.

Prognostic values of VIP and PD-L1 expression
Protein Atlas analysis of TCGA data was used to evaluate the prog-

Association of VIP expression with PD-L1 expression
Next, we correlated PD-L1 and VIP expression in 17 cancers using the TCGA RNA-seq dataset and RNA-seq by expectation maximization (RSEM). Among the 17 cancers, lung adenocarcinoma, lung squamous cell carcinoma, colorectal adenocarcinoma, and pancreatic adenocarcinoma, showed interesting associations between VIP and PD-L1 expression ( Figure 3). Including adenocarcinoma and lung squamous cell carcinoma, PD-L1 expression tended to be high while VIP expression was low ( Figure 3A and B). In both colorectal adenocarcinoma and pancreatic adenocarcinoma, VIP expression was generally high while PD-L1 was low ( Figure 3C and D). These results suggest that VIP and PD-L1 expression might be negatively correlated, and even mutually exclusive as immune checkpoint pathways are overexpressed in cancers ( Figure 3F).

Association of VIP and PD-L1 expression with immune cell infiltrates in the tumor
High levels of immune cells infiltrating the tumor microenvironment have positive prognostic values in many cancers, including ovarian cancer. 23 It is hypothesized that local expression of PD-L1 and VIP in the tumor microenvironment mediates an immunosuppressive state and hampers anti-cancer immunity mediated by T cells. Therefore, we correlated the levels of immune cells in the TME with VIP and PD-L1 expression levels across lung adenocarcinoma, lung squamous cell cancer, pancreatic adenocarcinoma, and colorectal cancer histology.
Analysis of immune cell infiltration in the TME included the content of B cells, CD8+ T cells, CD4+ T cells, macrophages, neutrophils, and dendritic cells (Figure 4 and Table 1). In lung adenocarcinoma, PD-L1   To further investigate the potential interaction between the two molecules, we found that the expression of VIP and PD-L1 is mutually exclusive in colorectal adenocarcinoma, lung adenocarcinoma, lung squamous cell carcinoma, and pancreatic adenocarcinoma ( Figure 3F).
The results showed that VIP expression was high and PD-L1 expression was low in individual tumors from patients with pancreatic adenocarcinoma and colorectal adenocarcinoma. In contrast, PD-L1 expression was high and VIP expression was low in individual tumor specimens from lung adenocarcinoma and lung squamous cell carcinoma patients.
These results may indicate functional redundancy between the two immune-suppressive molecules. Thus, it may be possible that some cancers may use either VIP or PD-L1 as a primary means of immune escape. We also found higher levels of PD-L1, and VIP expression was associated with an increase of immune effector cells in the TME Thus, there is limited availability of data collection from tumors that commonly undergo neoadjuvant therapy. In turn, conclusions based upon analyses of these datasets may not translate to modern clinical practices for cancer. 26 Although the findings herein may suggest that VIP could function like PD-L1 as a key immune checkpoint molecule in cancer, further laboratory experiments and clinical trials are needed to explore the prognostic and predictive value of VIP expression in cancer.

CONCLUSION
In conclusion, PD-L1 expression levels are prognostic biomarkers for breast and colorectal cancer patients. PD-L1 is a proven predic-