The integration of immune checkpoint inhibitors with VEGF targeted agents in advanced gastric and gastroesophageal adenocarcinoma: a review on the rationale and results of early phase trials

Bevacizumab was tested in the first line setting in treatment of metastatic GC patients in combination with chemotherapy compared with chemotherapy alone. The combination showed advances in response rates (46% vs 37%) and PFS (6.7 vs 5.3 months). Unfortunately, the benefit of OS was not statistically significant (12.1 vs 10.1 months) [18].

On the other hand, the survival benefit was demonstrated with ramucirumab, the anti-VEGFR2 antibody, either as monotherapy versus best supportive care (5.2 vs 3.8 months) or in combination with paclitaxel versus paclitaxel alone (9.6 vs 7.4 months) in the second-line setting, which validated the role of VEGF/VEGFR-2 signaling as an important therapeutic target in advanced GC/GEJC and esophageal adenocarcinomas [19, 20].

In addition, VEGFR TKIs have also been explored in patients with GC/GEJC. Regorafenib, was tested in the phase II randomized placebo-controlled trial (INTEGRATE). The trial enrolled 147 evaluable patients. A significant improvement was seen in median PFS, 2.6 versus 0.9 months favoring regorafenib with a hazard ratio of 0.4 [21].

Cabozantinib, another multi-TKI inhibitor targeting VEGFR2, MET, RET, and AXL showed broad single agent activity in multiple solid tumors including gastric cancer, and is currently FDA approved for patients with RCC and HCC after progression on sorafenib. A Phase 2 randomized discontinuation trial evaluated the efficacy and safety of cabozantinib in 526 patients with 9 tumor types, including a cohort of 21 patients with GC. Patients received cabozantinib 100 mg orally once daily over a 12-week lead-in phase. ORR, the primary endpoint of the lead-in phase at week 12 for the GC cohort was 10% and the disease control rate was 33%. The most common grade 3 or higher adverse events included fatigue, diarrhea, hypertension, and palmar-plantar erythrodysesthesia [22]. (Table 1).

There is extensive evidence that angiogenesis pathways are able to foster an immune suppressive tumor microenvironment (TME) in various ways including through direct suppression of antigen presenting cells and immune effector cells or through augmentation of immune suppressive cells such as Treg, MDSCs, and TAMs.

VEGF-driven angiogenesis involving its cognate receptors VEGFR1-3 is among the most important angiogenic factors associated with multi-pronged immune suppressive effects in the TME [34‐36]. For example, binding of VEGF to its receptors on DCs inhibits their maturation and antigen presentation and induces PD-L1 expression on the cell surface [37]. Furthermore, increased VEGF levels leads to inhibition of cytotoxic T cell trafficking, proliferation, and effector function [9, 38, 39]. In addition, VEGF drives the expansion of Tregs and MDSCs as well as TAMs in the TME [40]. Moreover, angiogenic pathways other than VEGF have also been shown to contribute to immune suppression in the TME. For example, Ang-2 is a cytokine ligand of the receptor Ang1/Tie2, which modulates VEGF-mediated angiogenesis [41, 42]. Ang-2 expression leads to increased endothelial adhesion of TIE-2 expressing monocytes/macrophages and stimulates their production of immunosuppressive cytokines such as IL-10 [43, 44].

In turn, targeting VEGFR using various approaches including VEGF multi-TKIs have been found to promote an immune permissive TME by normalizing vascularization and reducing MDSCs (CD11b + , Gr +). In addition, VEGF TKIs have been described to have immuno-modulatory effects in vitro and in murine models for several cancers. For instance, cabozantinib appears to drive its impact on Tregs via the HGF/c-Met pathway, where this receptor signaling cascade moderates several immune cell functions [45]. HGF was shown to induce Tregs (CD4 + CD25 + FoxP3) via hindering of DC function in a murine central nervous system autoimmunity model [46]. HGF cultured monocytes differentiate into monocytic cells that generate soluble factors (e.g., IL-10) which are known to esteem immune suppressive conditions ideal for tumor development [47].

The immunomodulatory effects of cabozantinib have been studied mostly in in vitro and in vivo tumor models other than GC/GEJC. For example, in murine colorectal cancer tumor cells, cabozantinib led to an increase in major histocompatibility complex (MHC) class 1 and cell surface molecules FAS expression, as well as promoting intercellular adhesion molecule 1, and calreticulin expression, which together heighten immune cell recognition and enhance sensitivity to T-cell mediated lysis [48]. Furthermore, treatment with cabozantinib boosted peripheral CD8 + T cells and reduced Treg and MDSCs in the spleen. In murine tumor models, marked tumor infiltration of CD8 + T cells and Tregs along with a reduction in tumor infiltration of MDSCs and TAMs was seen. These results highlight that cabozantinib alleviates the immune suppressive environment in mouse tumors. In a chimeric murine model of metastatic castrate resistance prostate cancer, a modest effect on tumor mass was seen when cabozantinib was used as single agent, but in combination with anti-CTLA-4 or anti-PD-1 a robust synergistic response was noted, mediated by neutralization of MDSCs (CD11b + , Gr1 +) [49]. PI3K signaling was also found to be hampered by cabozantinib, which impaired the release of cytokines by prostate cancer cells; which in turn prompted expression of MDSC genes responsible for tumor suppressive activity. The negative modulatory impact on GR1 + MDSCs was associated with an increase in CD8 + T cell tumor infiltration in prostate tumors, which endorses the antagonistic activity of Gr1 + MDSCs on the CD8 + population of T cells.

In addition to cabozantinib, other VEGF TKIs have shown immunomodulatory effects in both in vitro and in vivo studies, further supporting the potential role of VEGF TKIs in combination IO therapy. For instance, lenvatinib has been associated with reduction of TAMs in multiple tumor models and increased tumor infiltration of interferon-γ producing CD8 + T cells as well as memory T cells [50‐52]. Furthermore, regorafenib has also been associated with decreased infiltration of TAMs and increased M1 macrophages in colorectal tumor models [53] (Fig. 1).

Although both VEGFR inhibitors and PD-1/PD-L1 inhibitors have demonstrated some activity as single agents in the treatment of GC and GEJC, the mechanisms by which the VEGFR plus PD-1/PD-L1 targeting combinations lead to benefit and predictors of response to therapy need further investigations. To date, PD-L1 expression, high TMB and MSI-High status correlated with treatment response and improved survival from monotherapy with PD-1/PD-L1 blockade. Both PD-L1 expression using the CPS score and MSI status have also been validated as predictors of response to PD-1 plus chemotherapy combinations in this population as per the CheckMate-649 trial, KEYNOTE-590, and others. Ongoing and future trials testing the VEGFR plus PD-1/PD-L1 inhibitors should test the predictive value of those similar markers as well as explore additional predictive & prognostic markers which will add essential data to the field supporting personalized medicine in immuno-oncology. There remains a challenge in the interpretation of biomarkers and evaluation of their predictive values in the setting of combination immune and angiogenesis blockade. Future studies focusing on biomarker development aimed at predicting synergy between IO and VEGF targeting agents are warranted. Considering the toxicity profile of this novel combination, which is less favorable then either drug alone, tailoring this strategy to the population that would benefit most would be ideal.

Furthermore, results from early phase trials testing VEGFR plus PD-1/PD-L1 inhibitors in advanced stage GC/GEJC suggest that using this therapeutic approach in earlier vs later lines setting is associated with more favorable tumor responses and survival outcomes. This could theoretically be related to more immune-permissive tumor microenvironment in early line settings and encourages future direction to validate those results in larger pivotal phase III trials. Such trials could test this combinatorial strategy alone in sequential approach post chemotherapy or in combination with chemotherapy. If the survival advantage of such combination is confirmed in phase III trials, a next step would be the incorporation of this therapy with multimodality approaches in early stage GC/ GEJC.