Here we present a rare case of rapid tumor debulking response subsequent to a short course of combination ICI immunotherapy in a heavily pre-treated pMMR/MSS mCRC patient presenting with radiological progression after third-line experimental AlloStim® immunotherapy and a short course of low dose regorafenib. Here we consider the question whether the ICI immunotherapy alone or in combination with previous AlloStim®, or in combination with previous regorafenib alone, or with prior AlloStim® and regorafenib together was most likely responsible for eliciting this rare objective response in this cold tumor indication.
It seems unlikely that the combination ICI immunotherapy could be solely responsible for the observed response. ICI-based regimens both as monotherapy [40] [41] or as combination therapies [42] have not previously shown any meaningful positive outcomes in pMMR/MSS colorectal cancers.
For example, an initial phase II study assessed the efficacy of tremelimumab, a monoclonal antibody against CTLA4, in patients with treatment-refractory mCRC, which resulted in no improvement post-treatment [43]. Furthermore, two phase I studies of anti-PD-1 [44] and anti-PD-L1 [45] in previously-treated mCRC patients produced no responses. ICI regimens also failed as maintenance therapy after first line therapy in the MODUL study [46].
In general, ICI immunotherapy combining CTLA-4 and PD-L1 inhibitors have also shown very limited clinical benefit in patients with non-selected mCRC. A rare partial response (PR) (1/119) was reported in a randomized phase 2 clinical trial which evaluated the efficacy of combination durvalumab (anti-PD-L1) and tremelimumab (anti-CTLA-4) in patients with advanced refractory mCRC. In this study, 119 patients were assigned to the treatment group and 61 patients were assigned to best supportive care (BSC) alone. Patients in the treatment group received a median of 12 weeks of durvalumab and 12 weeks of tremelimumab [47], while in the present case only 5 weeks of ICI combination therapy was administered.
The phase II KEYNOTE-016 trial was performed to evaluate the clinical efficacy of single agent pembrolizumab in patients with pMMR/MSS mCRC, dMMR/MSI-H mCRC and or dMMR/MSI-H non-CRC. No responses were noted in 18 patients in the pMMR/MSS mCRC group [48]. In a clinical study which included 59 pMMR/MSS mCRC patients treated with ICI beyond radiological progression by RECIST 1.1, no patient demonstrated subsequent radiographical tumor shrinkage at a median of 42 days [49].
It has been reported that a small subset (~ 2%) of patients with pMMR/MSS colorectal cancer with a mutation in POLE and POLD1 enzymes and those without liver metastases have a higher chance of a response to ICI immunotherapy [50]. While the patient in the present case did not have POLE or POLD1 mutations, no liver metastases were present. Therefore, it is possible this patient was more susceptible to ICI immunotherapy, but seems unlikely that the short course of combination ICI alone was solely responsible for the extensive tumor debulking observed.
ICI strategies in combination with other drugs or procedures are under investigation, including evaluations of ICI in combinations with chemotherapy, radiotherapy, vascular endothelial growth factor (VEGF)/VEGF receptor (VEGFR) inhibitors, mitogen-activated protein kinase (MEK) inhibitors, and signal transducer and activation of transcription 3 (STAT3) inhibitors [51]. However, these combination approaches have yet to demonstrate any significant anti-tumor activity in the clinical setting [12, 52], [53].
Could the short course of regorafenib alone or in combination with ICI immunotherapy be responsible for the objective tumor response observed in this case?
Regorafenib is approved for third-line mCRC based on the results of the CORRECT trial which demonstrated only a 1.4 month increase in the median survival compared to a placebo control (6.4 months vs. 5.0 months) [54]. In the Phase II TEXCAN trial, no objective responses were reported in 35 mCRC patients after 2 months of treatment with regorafenib according to RECIST 1.1, Choi, and modified Choi [55]. Therefore, it seems unlikely that the prior short course of low dose regorafenib alone could be responsible for the rare objective response reported here.
Regorafenib is a multi-kinase inhibitor that targets several receptor tyrosine kinases involved in angiogenesis and metastases (VEGFR1, VEGFR2, VEGFR3, FGFR1, FGFR2, TIE2, PDGFRs), oncogenesis (KIT, RET, RAF1), and tumor immunity (CSF1R). While regorafenib does not directly convert cold tumors to hot tumors, regorafenib is believed to possibly contribute to shifting the tumor microenvironment toward a more immune-responsive state. This constellation of mechanisms suggests that regorafenib could potentially be a combination partner for ICIs [56].
There are mixed results on the combination of regorafenib with ICI in clinical trials. Regorafenib in combination with PD-1 antibody as a third‐line mCRC therapy has been evaluated in several studies. For example, 24 patients with MSS mCRC were included in the REGONIVO study. In this study, regorafenib was administered at 80–160 mg once daily for 21 days on and 7 days off together with nivolumab at 3 mg/kg every 2 weeks. A 33.3% objective response rate was reported with this regimen [57]. However, this promising activity has not been observed in other studies.
In a single site study, 18 mCRC patients treated with a combination of regorafenib and nivolumab, no objective responses were observed. In this study, 13 patients (69%) had progressive disease, and the median progression-free survival (PFS) was only 2 months. Four out of five patients in this study evaluated with stable disease (SD) occurred in patients without liver metastases, whereas a short disease stabilization was seen in 1 of 14 patients with history of liver metastases [58].
In another study in MSS mCRC patients, a combination of regorafenib and toripalimab, an anti-PD-1 ICI yielded an objective response rate of 15.2% (5 of 33 patients) with all (3 of 3) with lung-only metastasis responding [59]. In a retrospective study that involved 14 Chinese medical centers, a partial response rate of 5% (4 of 84 patients) was reported in MSS mCRC patients administered regorafenib combined with ICIs [60].
In a phase 2 study in patients from the USA with pMMR/MSS mCRC, regorafenib plus nivolumab yielded an objective response rate of 7%, with all responses observed in patients without liver metastases [61]. In this study, regorafenib was administered at 80 mg/day on a 3 weeks on/1 week off schedule and was increased to 120 mg/day if the 80 mg/day was well tolerated. Nivolumab was administered at 480 mg every 4 weeks.
Based on these data, we cannot rule out the possibility that the regorafenib pre-treatment may have primed for responsiveness to the ICI immunotherapy in this pMMR/MSS mCRC patient that presented without liver metastases.
However, in this case, the doses and frequencies of both regorafenib and of the combination ICI immunotherapy that were actually administered were significantly less that the doses administered in clinical trials where objective responses were observed.
In addition, in the present case, corticosteroids (CS) were administered 6 weeks after start of ICI administration. In a retrospective single institution study, patients were evaluated in two cohorts based on timing of initiation of CS after initiation of ICI immunotherapy (≥ 2 months vs < 2 months). The administration of CS < 2 months after initiation of ICI immunotherapy was found to significantly hinder ICI efficacy [62].
Since regorafenib does not directly convert cold tumors to hot tumors, which is necessary for priming ICI responsiveness, and the doses and frequencies of both regorafenib and the ICI immunotherapy used in the present case were at sub-optimal therapeutic levels, combined with the early use of CS, we believe, while possible, it is unlikely that the regorafenib priming was responsible for the rare objective tumor response observed here and it is more likely that a combination that converted the cold tumors to hot was responsible.
Therefore, we finally consider whether the experimental AlloStim® priming alone or in combination with regorafenib contributed to the ICI objective response.
We hypothesized that if the restaging CT scan after AlloStim® immunotherapy reported as progressive disease (PD) by RECIST 1.1, was actually pseudoprogression due to ‘hot’ inflammation of the tumor lesions which would make them appear to be larger than the actual tumor burden, that ICI immunotherapy would elicit a rapid tumor debulking response due to resident infiltrating effector immune cell release from suppression.
The present subject was negative for serum IL-12 at baseline. After three cycles of experimental AlloStim® immunotherapy the subject seroconverted to IL-12 positivity, supporting that the host immune system was modulated. We previously reported that IL-12 positivity correlated with long-term survival after AlloStim® immunotherapy [63].
IL-12 is an effector cytokine that promotes anti-tumor immunity by activating an effector Th1 response, which is required for the activation of cytotoxic T and NK cells [64]. IL-12 promotes production of IFN-γ which acts to upregulate PD-L1 in the tumor microenvironment (TME), which may make these tumors more susceptible to anti-PD-L1 ICI immunotherapy [65–67] .
The presence of IL-12 can have many beneficial anti-tumor effects, including: increasing production of IFN-γ from NK and T cells [68]; stimulation of growth and cytotoxicity of activated NK cells and CD8+ and CD4+ T cells [69], shifting the Th1/Th2 balance in favor of the Th1 phenotype [70]; induction of antiangiogenic cytokine and chemokine production [71]; remodeling of the peritumoral extracellular matrix and tumor stroma [72], reprogramming of myeloid-derived suppressor cells [73], and increasing expression of MHC class I molecules necessary of cytolytic T-lymphocyte (CTL) recognition of tumor cells [74].
Soluble heat shock protein (HSP)-70 was also detected in the serum after AlloStim® administration. HSP-70 is a stress-inducible chaperone that is overexpressed within tumor cells, including CRC [75]. The finding of HSP-70 in serum suggests that tumor cells have been killed in a manner where the cell membrane is disrupted (immunological cell death), releasing the HSP along with danger signals into the tumor microenvironment. Hsp70 extracellular function is believed to be immunogenic and extracellular Hsp70 can serve as an adjuvant to activate the innate immune system [76] and can eventually lead to tumor-specific adaptive immunity [77]. Endogenous HSP chaperone all tumor cell antigens, including self- and neo-antigens. Tumors accumulate mutations that can cause tumor-specific neo-antigen expression. Since these neo-antigens are intracellular, they may not have been previously exposed to the immune system, as the tumors sequester these neoantigens. Thus the presence of soluble HSP-70 supports that AlloStim® modified the TME in a manner that caused tumor lysis and release of chaperoned neoantigens. Exposure of tumor neoantigens to the immune system increases responsiveness to ICI [78].
The mechanism of action of AlloStim® is also consistent with the conclusion that an inflammatory cold to hot conversion occurred which caused the dramatic 73% increase in target lesion size by RECIST 1.1. The immune systems of patients with metastatic cancers are dysregulated resulting in a shift toward Th2 dominance [79–81]. AlloStim® experimental immunotherapy modulates the dysregulated immune systems of these patients to a Th1 dominance using a strategy of allo-priming [82]. The STIMVAX protocol incorporated a first series of intradermal injections of AlloStim®. The host rejection of the intentionally mis-matched AlloStim® cells shortly after administration results in increased titers of allo-specific Th1 and CTL cells, modulating the resident Th1/Th2 balance.
These allo-specific cells elicited after intradermal injections are non-specifically activated by cytokine release after intravenous infusion of AlloStim® through a bystander activation mechanism [83]. Activated T-cells can extravasate and local inflammation attracts these cells into tissues and sites of inflammation, including tumors [84]. Thus, the intravenous infusion of AlloStim® after allo-priming can convert “cold” tumors into “hot” tumors with extensive infiltration of Th1/CTL memory cells, which could possibly account for the 73% increase in target lesion size.
Inflamed tumor lesions can enlarge and appear as PD by RECIST 1.1. However, the putative anti-tumor mechanism leading to tumor debulking immunity may take several additional months before a radiological response can be detected and patients are often removed from the treatment protocols before a later assessment can be conducted.
We hypothesized that based on the mechanism of action of AlloStim® that radiological progression after 3 cycles likely represents a beneficial immune response that has primed the tumor lesions for an eventual debulking anti-tumor response. In order to support this hypothesis, we administered a short course of combination ICI immunotherapy. Since pMMR/MSS mCRC is known not to be responsive to ICI immunotherapy, we predicted that if a rapid tumor debulking response were observed, this would provide evidence supporting that the tumor lesions had been previously primed with infiltrating effector immune cells.