Immune Checkpoint Inhibitors in pMMR/MSS Colorectal Cancer

Abstract

Background

Immune checkpoint inhibitors have recently replaced over chemotherapy as the first-line treatment for microsatellite instability-high or mismatch repair deficient (dMMR/MSI-H) stage 4 colorectal cancers. Considering this success, many studies have tried to replicate the use of immune checkpoint inhibitors, either as a single agent or in combination with other therapeutic agents, in the treatment of proficient mismatch repair (pMMR/MSS) stage 4 colorectal cancers. This review summarizes the seminal clinical data about the immune checkpoint inhibitors used in pMMR/MSS colorectal cancers and some future directions.

Results

Studies concerning the use of immune checkpoint inhibitors as a single agent or in combination with other immune checkpoint inhibitors, targeted therapy, chemotherapy, or radiotherapy have proven inefficient in the treatment of pMMR/MSS colorectal cancer. However, a small subset of patients with pMMR/MSS colorectal cancer who has a mutation in POLE and POLD1 enzymes may respond to immunotherapy. Moreover, patients without liver metastasis appear to have a better chance of response. New immune checkpoint targets are being identified, such as VISTA, TIGIT, LAG3, STING signal pathway, and BTLA, and studies are ongoing to determine their efficiency in this disease type.

Conclusion

Immune checkpoint inhibitor-based regimens have not yet shown any meaningful positive outcomes for most pMMR/MSS colorectal cancers. A beneficial effect among a minority of these patients has been observed, but concrete biomarkers of response are lacking. Understanding the underlying mechanisms of immune resistance should guide further research for overcoming these obstacles.

Background

Colorectal cancer is the third leading cause of cancer-related deaths globally [1], accounting for more than one million deaths in 2019 [2]. Traditionally, chemotherapy and targeted therapies are the foundation in the management of metastatic colorectal cancers. However, immune checkpoint inhibitors have recently triumphed over chemotherapy as the first-line treatment for a small subset (~ 15%) of stage 4 colorectal cancers, characterized by microsatellite instability (MSI-H) or mismatch repair deficiency (dMMR) [3]. Tumors of this molecular subtype carry abnormalities in mismatch repair (MMR) genes, either as germline mutations [4] or as acquired mutations that silence the gene expression by promoter region hypermethylation [5]. The MMR genes encode proteins that correct nucleotide base mispairings [6]. In their absence, errors accumulate in the DNA microsatellite regions [7]. These regions are short, repetitive DNA base pairs sequences that occur throughout the genome and are prone to nucleotide base mispairing, especially in the setting of MMR deficiency. The remaining colorectal cancers have alterations at a chromosomal level [8] such as loss of heterozygosity, aneuploidy, and chromosomal translocations [9, 10]. The microsatellite regions in such cancers remain relatively unaffected, and these tumors are described as proficient in mismatch repair (pMMR) or microsatellite stable (MSS). Most of the national and international consensus guidelines recommend early testing for dMMR or MSI-H status in colorectal cancer patients [11, 12]. Testing uses either immunohistochemistry (IHC) to detect the nuclear expression of one of the four mismatch repair proteins (MLH1, MSH2, MSH6, PMS2) [13–15] or microsatellite instability (MSI-H) testing by a PCR-based assay [15–18]. Tumors are defined as dMMR if the IHC test shows loss of expression of one of the MMR proteins or PCR shows microsatellite instability at ≥ 2 loci (out of 5) compared to normal tissue [18, 19]. Commercially available next-generation sequencing (NGS) testing platforms can also reliably identify MSI-H disease by measuring tumor mutation burden [20–23]. On NGS tests, dMMR/MSI-H tumors generally have a higher tumor mutation burden score (TMB-high) (≥ 20 mutations/MB and in some cases > 1000 mutations/MB) due to the inability of cancer cells to repair DNA mutations, compared to pMMR/MSS tumors that show a low tumor mutation burden score (TMB-low). TMB-high score and dMMR/MSI-H are highly correlated [24], while tumors with TMB-low correlate with pMMR/MSS [25, 26]. Analysis of human tumor samples shows that pMMR/MSS tumors have a median of 6 mutations/Mb, compared with a median of 52–54 mutations/megabase in dMMR/MSI-H tumors [27, 28].

Depending upon the genetic mutation profile of the pMMR/MSS colorectal cancers, we use a combination of fluoropyrimidine-based cytotoxic chemotherapy regimens and biological therapies. These biological therapies include vascular endothelial growth factor (VEGF) inhibitors or epithelial growth factor receptor (EGFR) inhibitors [29]. In contrast, for the dMMR/MSI-H colorectal cancers, we use pembrolizumab in the first-line metastatic setting based on the KEYNOTE-177 trial. This trial showed that the median progression-free survival (PFS) was 16.5 months (95% confidence interval [CI], 5.4 to 32.4) with pembrolizumab and 8.2 months (95% CI, 6.1 to 10.2) with chemotherapy [30]. Due to the success of the immune checkpoint inhibitor therapy in dMMR/MSI-H colorectal cancers [3, 31–33], many studies have tried to replicate its benefit in pMMR/MSS tumors, either as a single-agent or in combination with other therapeutic agents active in colorectal cancers. This review summarizes the seminal clinical data about the immune checkpoint inhibitors used in pMMR/MSS colorectal cancers and some future directions.

Early Studies Using Immune Checkpoint Inhibitors in Colorectal Cancers

Earlier studies of immune checkpoint inhibitors in refractory colorectal cancer aimed to emulate the success of these therapies in melanoma [34]. These studies did not distinguish between the dMMR/MSI-H or pMMR/MSS groups [35, 36]. In one of the earliest studies using single-agent PD-1 immune checkpoint inhibitor, nivolumab, in refractory metastatic colorectal cancer patients, 13 of 14 patients progressed on their first tumor response assessment. Only one patient with dMMR/MSI-H phenotype achieved a complete and durable response lasting more than 3 years [37, 38]. The lack of response to nivolumab in pMMR/MSS colorectal cancer was also demonstrated when escalating doses of nivolumab (1.0 mg/kg, 3.0 mg/kg, or 10.0 mg/kg every 2 weeks) showed no objective response in the 19 colorectal cancer patients enrolled in the study [39]. Another phase II trial investigated pembrolizumab at the dose of 10 mg/kg administered every 2 weeks in 41 patients groups to three groups; dMMR/MSI-colorectal cancers, pMRR/MSS colorectal cancers, and non-colorectal dMMR/MSI-H patients [40]. The objective response rate in the pMMR/MSS group was 0%, with a stable disease rate of 11% (2 of 18 patients: 95% CI, 1–35). Out of 18 patients with pMMR/MSS tumors, 11 had disease progression during the study period. The PFS at 20 weeks was 0%, and the median overall survival (OS) was 6 months [40]. In contrast, among patients with dMMR/MSI-H colorectal cancers, the objective response rate was 40% (95% CI, 12–74), and PFS and OS at 24 months were 61% and 66%, respectively [40]. Hence, these early studies demonstrated a lack of efficacy of the single-agent immune checkpoint inhibitors in pMMR/MSS colorectal cancers and excellent responses in dMMR/MSI-H cancers (Table 1).

Table 1.

Early studies about immune checkpoint inhibitors in colorectal cancers

Study name	Agent	Target	Study population	Primary endpoint	Phase	Seminal results
NCT00730639 [37, 38]	Nivolumab	PD-1	Treatment refractory metastatic solid tumors, 14 CRC	Safety, objective response	I	Response in 1 dMMR/MSI-H patient No response in pMMR/MSS patient
NCT00730639 [39]	Nivolumab	PD-1	Treatment refractory advanced solid tumors, 19 CRC	Safety	I	No objective response in patients with CRC
NCT01876511 [40]	Pembrolizumab	PD-1	dMMR/MSI-H CRC, pMMR/MSS CRC, non-CRC dMMR/MSI-H	PFS, Objective response rate	II	Objective response rate was 50% (95% CI 31–69%) in dMMR/MSH-I group and 0% in pMMR/MSS group Median PFS was not reached in dMMR/MSH-I and was 2.4 months for pMMR/MSS (HR = 0.137, p = < 0.0001)

Anti-PD-1 and Anti-CTLA-4 Antibody Combination in pMMR/MSS Colorectal Cancers

The combination studies with anti-PD-1 and anti-CTLA-4 immune checkpoint inhibitors performed in refractory metastatic colorectal cancers showed a similar trend as the single-agent in pMMR/MSS disease (Table 2). CheckMate-142 trial that studied nivolumab with or without ipilimumab in metastatic colorectal cancer patients regardless of MSI status showed a median PFS of 1.4 months (pooled across nivolumab vs. nivolumab and ipilimumab treatment groups) [41]. The Cancer Trial Group CO.26 phase II study compared the combination of a PD-L1 inhibitor durvalumab, and CTLA-4 antibody tremelimumab, with the best supportive care in all patients with metastatic colorectal cancers [42]. In the durvalumab (1500 mg every 4 weeks) plus tremelimumab (75 mg every 4 weeks) arm, 98% of patients had pMMR/MSS status compared to 88% in the best supportive care arm. The objective response rate was 1% and 0%, respectively. The median PFS was 1.8 months compared to 1.9 months, respectively (HR = 1.01, p-value = 0.97), and the median OS was 6.6 months compared to 4.1 months, respectively (HR = 0.72, p-value = 0.07). Recently, a phase I study of the combination of a novel FC-enhanced IgG1 CTLA-4 inhibitor, botensilimab, with balstilimab, an anti-PD-1 antibody, was reported in pMMR/MSS colorectal cancer [43]. The overall response rate was 22%, and the disease control rate was 73%. The median OS was not reached at the time of the report. The durability of these responses was seen with 9 out of 13 responses, and the 12-month landmark OS was 61% for the overall population (NCT03860272) [43].

Table 2.

Summary of the combinations of immune checkpoint inhibitors with chemotherapy, targeted therapy, and radiotherapy in pMMR/MSS colorectal cancers

Study name	Agent	Target	Study population	Primary endpoint	Phase	Seminal results
CheckMate-142 NCT02060188 [41]	Nivolumab only vs Nivolumab with Ipilimumab	PD-1, CTLA-4	Recurrent and metastatic MSI-H and non-MSI-H in colon cancer	Overall response rate (ORR)	II	Overall response rate 27% in the nivolumab arm vs 15% in the nivolumab + ipilimumab across all patient groups In non-MSI-H, PFS was 1.4 months (95% CI, 1.2–1.9) across the two arms, and PFS for MSI-H was not reached
Cancer Trial Group CO.26 NCT02870920 [42]	Durvalumab with Tremelimumab with best supportive care (BSC) vs BSC only	PD-1, CTLA-4	Refractory metastatic colorectal cancer	Overall survival (OS)	II	Median OS was 6.6 months for durvalumab + tremelimumab regimen vs 4.1 months for BSC (HR = 0.72, p = 0.07) PFS was 1.8 months and 1.9 months respectively (HR = 1.01, 90% CI, 0.76–1.34)
NCT03860272 [43]	Botensilimab + Balstilimab	CTLA-4, anti-PD-1	Metastatic pretreated pMMR/MSS CRC	Safety	Ia/Ib	Overall response rate was 22% Median PFS was 4.1 months Median OS was not reached
AtezoTRIBE NCT03721653 [45]	FOLFOXIRI + Bevacizumab + Atezolizumab vs FOLFOXIRI + Bevacizumab	PD-L1	Chemotherapy naïve unresectable metastatic colorectal cancer	Progression free survival (PFS)	II	73 patients enrolled in the control group vs 145 in the atezolizumab arm Median PFS was 13.1 months (80% CI 12.5–13.8) in the atezolizumab arm vs 11.5 months (10.0–12.6) in the control group (HR = 0.69, p = 0.012)
BACCI NCT02873195 [46]	Capecitabine + Bevacizumab vs Capecitabine + Bevacizumab + Atezolizumab	PD-L1	Refractory metastatic colorectal cancer	PFS	II	Median PFS was 4.4 months (95% CI, 4.1–6.4 months) in the atezolizumab arm vs 3.6 months (95% CI, 2.2–6.2 months) in the control group Among patients in pMMR/MSS CRC group, the HR for PFS was 0.66 (95% CI, 0.44–0.99) No significant change in OS between the atezolizumab arm and the control group
CheckMate-9X8 NCT03414983 [47]	FOLFOX + Bevacizumab + Nivolumab vs FOLFOX + Bevacizumab	PD-1	Chemotherapy naïve metastatic colorectal cancer	PFS	II/III	Similar PFS pf 11.9 months in the two arms (HR = 0.81, p-value = 0.30) Objective response rate was 60% in the nivolumab arm vs 46% in the control arm OR = 1.72, 95% CI 0.96–3.10)
AVETUX NCT03174405 [48]	FOLFOX + Cetuximab + Avelumab	PD-L1	RAS/BRAF wild type chemotherapy naïve metastatic colorectal cancer (2 patients were dMMR/MSI-H, 1 MSI-Low, and 40 pMMR/MSS)	PFS	II	12-month PFS of 40% Median PFS of 11.1 moths Objective response rate of 79.5%
REGONIVO NCT03406871 [57]	Nivolumab + Regorafenib	PD-1	Advanced and metastatic solid tumors	Safety	Ib	Objective response rate was 33% in pMMR/MSS group Median PFS across all groups was 7.9 months 52% of patients had liver metastasis vs 64% with lung metastasis Objective response in patients with liver vs lung metastasis was 8.3% compared to 63.6% respectively
NCT03712943 [58]	Regorafenib + Nivolumab	PD-1	Refractory colorectal cancer	Safety	I/Ib	Disease control rate was 63% Median OFS was 4.3 months Median OS was 11.1 months
NCT03657641 [59]	Regorafenib + Pembrolizumab	PD-1	Advanced or metastatic colorectal cancer	Safety, PFS, OS	I/II	Median PFS was 2 months (1.8–3.5) Median OS was 10.9 months (5.3-NR) Better outcomes in patients with lung metastasis vs liver metastasis, patients with non-liver metastatic disease had a PFS of 4.4 months (1.9–8.4)
REGOMUNE NCT03475973 [60]	Regorafenib + Avelumab	PD-L1	Advanced or metastatic solid tumors	Safety, objective response, progression-free rate	I/II	Median PFS was 2.5 months (95% CI, 1.9–5.5) Median OS was 11.9 months (95% CI, 6.2-NA) Partial response for 4 patients (13.8%), stable disease for 11 patients (37.9%), progressive disease for 14 patients (48.3%)
NCT02713373 [61]	Cetuximab + Pembrolizumab	PD-1	Advanced RAS wild-type colorectal cancer	Safety, PFS	Ib/II	Median PFS was 4.1 months (95% CI, 3.9–5.5)
CAVE-2 GOIM NCT05291156 [62]	Cetuximab + Avelumab vs Cetuximab only	PD-L1	Pre-treated RAS/BRAF Wild-type metastatic colorectal cancer	OS	II	Trial is active and awaiting results
NCT04017650 [64]	Encorafenib + Cetuximab + Nivolumab	PD-1	pMMR/MSS, BRAF V600E mutated unresectable or metastatic colorectal cancer	Overall response rate, safety and tolerability	I/II	Objective response rate of 45% (95% CI, 23–68) Median PFS was 7.3 months (95% CI, 5.5-NA) Median OS was 11.4 months (95% CI, 7.6-NA)
BEACON CRC NCT02928224 [63]	Encorafenib + Binimetinib + Cetuximab vs Encorafenib + Cetuximab vs Cetuximab + irinotecan or FOLFIRI (investigators’ choice)	PD-1	BRAF V600E mutated metastatic colorectal cancer with disease progression after one or two previous regimens	Safety, OS, overall response rate	III	In the doublet group: Objective response rate of 20% (95% CI, 13–29%) PFS of 4.2 months (95% CI, 3.7–5.4) Median OS was 8.4 months (HR for death vs control, 0.60, 95% CI, 0.45–0.79, p < 0.001)
NCT05308446	Nivolumab + Encorafenib + Cetuximab vs Cetuximab + Encorafenib	PD-1	Previously treated, pMMR/MSS BRAF V600E metastatic and unresectable colorectal cancer	PFS	II	Trial is active and awaiting results
NCT05019534	Vemurafenib + Cetuximab + Camrelizumab	PD-1	BRAF V600E mutated pMMR/MSS metastatic colorectal cancer	Safety	I	Trial is active and awaiting results
NCT02437071 [67]	Radiation + Pembrolizumab	PD-1	Metastatic colorectal cancer	Response rate	II	1 response in pMMR/MSS CRC group (out of 22 patients)
NCT03104439 [65]	Radiation + Ipilimumab + Nivolumab	CTLA-4, PD-1	pMMR/MSS and dMMR/MSI-H metastatic colorectal and pancreatic cancers	Disease control rate	II	Out of the 40 patients in the pMMR/MSS CRC group: Disease control rate was 25% (95% CI, 13–41%) Objective response rate was 10% (95% CI, 3–24%) In the CRC cohort: Median PFS of 2.4 months (95% CI, 1.8–2.5) Median OS of 7.1 months (95% CI, 4.3–10.9)
411 INNATE NCT04130854 [68]	APX005M + radiotherapy	CD40	Previously untreated rectal adenocarcinomas	Pathological complete response rate	II	Trial is active and awaiting results

Immune Checkpoint Inhibitor and Cytotoxic Chemotherapy Combinations in pMMR/MSS Colorectal Cancers

Multiple studies tested the combination of immune checkpoint inhibitors with active therapies in pMMR/MSS metastatic colorectal cancers. The premise is that the chemotherapeutic agents known to work in colorectal cancer may have immunomodulatory effects in tumors by increasing the expression of the PD-L1 receptors or altering the tumor microenvironment, thus making the tumors more responsive to immune checkpoint inhibitors. This was demonstrated in other solid tumors such as non-small cell lung cancers [44].

The AtezoTRIBE phase II trial was a randomized controlled trial of FOLFOXIRI with bevacizumab with or without atezolizumab in all previously untreated colorectal cancer patients, regardless of MMR status (Table 2) [45]. In the atezolizumab arm, the objective response rate was 59% compared to 64% in the standard-of-care arm, and this difference was not statistically significant (p-value = 0.412). However, the median PFS was significantly improved in the atezolizumab arm, with 13.1 months compared to 11.5 in the standard-of-care arm (HR = 0.69, p-value = 0.012). In the subgroup of pMMR/MSS patients, there was no statistical difference between the two arms in terms of median PFS with 11.4 months compared to 12.9 months (HR = 0.78, p-value = 0.71). A phase II randomized clinical trial BACCI evaluated bevacizumab and capecitabine with or without atezolizumab in metastatic colorectal cancer patients after induction chemotherapy [46]. In the pMMR/MSS subgroup, the objective response rate was 10% in the capecitabine, bevacizumab, and atezolizumab arm compared to 5% in the doublet arm without a statistically significant difference (p-value = 0.33). However, this study met its primary endpoint of improvement in median PFS among all subgroups, which was slightly better in the triplet arm at 4.4 months compared to 3.6 months in the standard-of-care arm (HR = 0.75, p-value = 0.07). Moreover, in the pMMR/MSS subgroup, the median PFS was better in the triplet arm compared to the doublet arm (5.3 vs. 3.3 months, HR = 0.66, 95% CI: 0.44–0.99). The median OS was the same in both arms, among all colorectal cancer subgroups, at 10.3 vs. 10.2 months (HR = 0.96, p-value = 0.42) [46]. Moreover, the CheckMate-9X8 trial investigated the combination of nivolumab with modified FOLFOX and bevacizumab compared to the standard of care (modified FOLFOX with bevacizumab) in chemotherapy naïve colorectal cancer patients, regardless of the microsatellite status [47]. Both arms had a similar median PFS of 11.9 months (HR 0.81; 95% CI 0.53–1.23; p = 0.30), which was the study’s primary endpoint. The objective response rate was 60% in the nivolumab-containing arm compared to 46% in the standard-of-care arm (odds ratio 1.72; 95% CI 0.96–3.10] [47].

Furthermore, the single-arm phase II AVETUX trial studied FOLFOX, cetuximab, and avelumab (a PD-L1 inhibitor) in the first-line setting in RAS/BRAF wild-type metastatic colorectal cancer, regardless of MSI status [48]. This combination was active and showed a 12-month PFS of 40%, a median PFS of 11.1 months, and an objective response rate of 79.5%. However, as compared to historical controls, these numbers are similar to the combination of FOLFOX and cetuximab in this patient subgroup.

These trials showed that the use of immune checkpoint inhibitors in combination with chemotherapy had a very limited benefit in the pMMR/MSS colorectal cancer patients, whether they were chemotherapy naïve or previously treated.

Immune Checkpoint Inhibitor and Targeted Chemotherapies Combination in pMMR/MSS Colorectal Cancers

Targeted therapies play a critical role in the management of metastatic colorectal cancer. These treatments include VEGF inhibitors such as bevacizumab [49, 50], ramucirumab [51, 52], and regorafenib (an oral multi-kinase inhibitor with predominantly anti-VEGF inhibition) [53], anti-EGFR antibodies such as cetuximab and panitumumab in RAS wild-type disease [54, 55], and anti-BRAF V600E targeting agents such as vemurafenib and encorafenib in BRAF mutated cancers. In preclinical murine models, regorafenib demonstrated higher T-cell activation and synergy with immune checkpoint inhibitors [56]. Based on these data, a phase Ib study, the REGONIVO trial, investigated the maximum tolerable dose of the combination of nivolumab (3 mg/kg every 2 weeks) with regorafenib (80 mg daily, 21 days on/7 days off) in 25 colorectal cancer patients refractory to at least two previous lines of chemotherapy agents (Table 2). The objective response rate was 33% in the pMMR/MSS subgroup while the median PFS of all enrolled patients was 7.9 months [57]. In a US study, 40 patients were treated with nivolumab 240 mg, given every 2 weeks, and 80 mg of regorafenib on days 1–21 of a 28-day cycle [58]. The 80 mg dose of regorafenib is lower than the FDA-approved dose of 160 mg likely due to additional toxicities with the combination. Among the 40 evaluable patients, 4 patients (10%) achieved an objective response including one unconfirmed response, 21 (53%) achieved stable disease, and the disease control rate was 63%. The median PFS and OS were 4.3 and 11.1 months, respectively [58]. The difference in the objective response rates of these two studies may be due to the number of patients enrolled (24 patients with MSS in the REGONIVO study compared to 52 in the US-based study) and the number of prior lines of therapy. Regorafenib 80 mg daily (days 1–14 of a 21-day cycle) has also been studied in combination with pembrolizumab 200 mg every 3 weeks in another phase I/II study [59]. Like the previous studies, the response rate was low (0%), and stable disease was observed in 49% of patients, with a median duration of stable disease of 2 months (0.2–18.8). The median PFS was 2 months, and the median overall survival was 10.9 months. The authors noticed that patients with lung-only metastasis performed better than liver metastasis [59]. The combination of regorafenib with an anti-PD-L1 antibody, avelumab (REGOMUNE) also proved ineffective with an objective response rate of 0%, a median PFS of 3.6 months, and a median OS of 10.8 months [60].

As mentioned above, all metastatic colorectal cancer patients usually undergo tumor mutation assessment, especially of the RAS and BRAF genes. In RAS wild-type metastatic colorectal cancers, patients benefit from the use of EGFR antagonists such as cetuximab or panitumumab as a single agent or in combination with chemotherapy [54, 55]. A phase I/II study looking at the antitumor efficacy of cetuximab in combination with pembrolizumab in patients with RAS wild-type metastatic colorectal cancers showed that among the 44 evaluable patients for efficacy, the objective response rate was 2.6%, 6-month PFS was 31% and median PFS was 4.1 months (95% CI 3.9–5.5 months) [61]. Another combination trial using cetuximab with avelumab is awaiting results [62].

For BRAF V600E mutated colon cancers, the BEACON study established the current standard of care treatment, which is the combination of BRAF V600E inhibitor encorafenib with cetuximab, after progression on the first-line therapy [63]. Recently, a single-arm single institution phase I/II trial investigated the use of encorafenib, cetuximab, with nivolumab in patients with pMMR/MSS metastatic colorectal cancers harboring BRAF V600E mutation [64]. The early results demonstrated that the triplet regimen had an objective response rate of 45% (95% CI, 23–68) and a disease control rate of 95% (95% CI, 75–100). The median PFS was 7.3 months (95% CI, 5.5-NA) while the median OS was 11.4 months (95% CI, 7.6-NA) [64]. In comparison, the doublet encorafenib and cetuximab regimen from the BEACON trial showed an objective response rate of 20% (95% CI 13–29%) and a PFS of 4.2 months (95% CI, 3.7 to 5.4) [63]. Based on these results, a randomized clinical trial is ongoing (NCT05308446). A phase I study is ongoing, investigating the maximum tolerable dose of vemurafenib, a BRAF inhibitor in combination with cetuximab and camrelizumab (an anti-PD-1) at fixed doses, and is awaiting results (NCT05019534). The combinations of immune checkpoint inhibitors to anti-EGFR or anti-BRAF targeted therapy show promise, but we will have to wait for results.

Immune Checkpoint Inhibitors and Radiotherapy Combination in pMMR/MSS Colorectal Cancer

The abscopal effect is a phenomenon in which local treatment of a tumor with radiation therapy leads to an antitumor response at a distant site by increasing the likelihood of systemic response to immunotherapy (Table 2) [65]. In 2020, McLaughlin et al. published a study in a preclinical setting that showed that radiotherapy could cause an increase of antigen presentation by antigen-presenting cells enhancing the antitumor effects of immune checkpoint inhibitors [66]. Based on these results, several early-phase studies tested the combination of immune checkpoint inhibitors with radiotherapy. A single-arm phase II study reported only one patient with pMMR/MSS colorectal cancer who responded in non-radiotherapy targeted lesions to the combination of pembrolizumab and external beam radiation (out of a total of 22 patients) [67]. Another phase II single-arm study combined radiation, ipilimumab, and nivolumab in patients with metastatic pMMR/MSS colorectal cancer and pancreatic cancer [65]. Among 40 pMMR/MSS colorectal cancer patients, the disease control rate was 25% (95% CI, 13–41%) and an objective response rate of 10% (95% CI, 3–24%) in the intention to treat analysis. The colorectal cancer cohort had a median PFS of 2.4 months (95% CI, 1.8–2.5) and a median OS of 7.1 months (95% CI, 4.3–10.9) [65]. In the ongoing phase II INNATE trial [68], a novel tumor microenvironment immune modulating agent, targeting CD40, is being studied in combination with radiotherapy in patients with rectal cancer after this combination showed promising results in animal models [69].

pMMR/MSS Tumors Harboring DNA Polymerase Mutations

Colorectal cancers with mutation of polymerases epsilon (POLE) and delta (POLD1) are rare subsets (< 1%) within pMMR/MSS cancers that can benefit from immunotherapy treatment [70, 71]. POLE and POLD1 enzymes carry a unique proofreading ability through their exonuclease domains to ensure the precision of DNA replication [72, 73]. Cancers with deleterious POLE or POLD1 mutation demonstrate a hypermutation profile on NGS testing characterized by a TMB-H score on these tests. The accumulation of mutations due to replicative errors increases the neoantigen production and leads to potential sensitivity to immune checkpoint inhibitors [28]. In a small case series of 16 such patients with POLE mutated cancers, 6 patients achieved a response, and 5 out of these 6 patients had colorectal cancer [74]. Therefore, among the pMMR/MSS group, POLE and POLD1 mutated cancers are highly likely to respond to immune checkpoint inhibitors.

Site of Metastasis and Its Implication in Immune Checkpoint Inhibitors Treatment

The REGONIVO study and its US-based phase Ib study of regorafenib and nivolumab [57, 58] demonstrated low objective responses but a trend towards better response according to metastatic sites was observed. In the REGONIVO study, 64% of the patients had lung metastasis, 60% had lymph node metastasis, 52% had liver metastasis, and 16% had peritoneal carcinomatosis [57]. The objective response among patients with liver-only metastasis was 8.3% compared to 63.6% for patients with lung-only metastasis. Lung-only metastasis is a prognostic marker in colon cancer and has a better prognosis on regorafenib monotherapy compared to liver metastasis patients [75]. In a separate phase II study conducted by Fakih M et al. the combination of nivolumab with regorafenib in patients with pMMR/MSS colorectal cancer, patients were stratified according to the status of liver metastatic involvement [76]. The objective response rate in the subgroup of patients with liver metastasis was 0% compared to 21.7% in the subgroup of patients without liver metastasis. Metastasis in the liver was found to be associated with lower ratios of CD8/Foxp3 + regulatory T cells and a decreased percentage of activated PD-1 + / CTLA4 + CD8 + cells [77]. However, current data are lacking due to the lack of a control arm and the exploratory nature of these studies. It is also unknown if the slow response has any relationship to tumor volume in such patients. At this point, there is insufficient data to recommend this strategy in routine clinical practice.

Possible Mechanisms of Immunotherapy Resistance in pMMR/MSS Colorectal Cancer

There have been several proposed mechanisms to explain the resistance of pMMR/MSS colorectal cancers to immunotherapy. These include decreased mutational burden and distinct qualities of the lymphocytic infiltrate [78, 79]. The tumor mutational burden is a measure of the total number of non-inherited mutations present per megabase pair in a tumor. It correlates with response to immunotherapy across nearly all cancer types [80]. Cancer cells with a TMB-H are, by extension, more likely to present mutant peptides or “neo-antigens” via their major histocompatibility (MHC) class I surface molecules [81, 82]. These interact with T-cell receptors to mark the cancer cells as non-self, marking the first step in cancer cell destruction. The importance of TMB in colorectal cancer was highlighted by Chen et al. in 2020 in a study that evaluated dual immune checkpoint inhibitor therapy with tremelimumab (CTLA-4 inhibitor) plus durvalumab (PD-L1 inhibitor), versus best supportive care in a sample of 180 patients with refractory CRC [42]. Nearly all patients had pMMR-MSS disease. While dual immunotherapy was associated with a decreased HR of death (HR 0.66; 90% CI 0.48–0.89; p = 0.02), this effect was markedly more pronounced for the subgroup of pMMR-MSS stable patients in the treatment arm with a TMB ≥ 28, compared to those with a TMB < 28 (HR for death 0.34; 90% CI, 0.18–0.63; P = 0.004). In colorectal cancer, microsatellite instability and TMB are highly correlated [24], and tumors with lower mutational burden express lower levels of neoantigens [25, 26]. The low neoantigen expression may be a mechanism of resistance to immunotherapy. In mouse models of colorectal cancer, the number of neoantigens remains stable over time in pMMR/MSS cancer cells but increases predictably in dMMR/MSI-H cells [83]. Analysis of human tumor samples shows that pMMR/MSS tumors have a median of 6 mutations/megabase, compared with a median of 52–54 mutations/megabase in dMMR/MSI-H tumors [27, 28]. Interestingly, pMMR/MSS colorectal tumors also have fewer clonal populations of neoantigen [25]. Clonal populations are more likely to elicit effective responses to immunotherapy [84]. Another possible mechanism of immunotherapy resistance in pMMR/MSS colorectal cancer is decreased lymphocyte infiltration in the tumor microenvironment. A study of 499 resected stage I–III colorectal tumors found that 68% of MSI-H tumors and 4.5% of MSS tumors had a high level of tumor-infiltrating lymphocytes defined as > 4/high power field [28]. Narayan et al. studied tumors from 204 patients with pMMR/MSS and 205 patients with dMMR/MSI-H colorectal cancer [26]. MSS tumors had significantly fewer T cells (p = 0.0013), fewer M1-type macrophages (p < 0.0001), and fewer resting natural killer cells (p = 0.0001). In another analysis that directly compared stage-matched pMMR/MSS and dMMR/MSI-H tumors, pMMR/MSS tumors had significantly less intratumoral and peritumoral lymphocytic infiltrate [85]. The difference in T-cell infiltrate is distinguished by the T-cell subtype. For example, pMMR/MSS colorectal tumors have consistently been found to have fewer CD8 + infiltrating lymphocytes than dMMR/MSI-H tumors [79]. There are no differences in helper and regulatory T cells between these tumors types [25]. Despite these findings, there is evidence to suggest that it is the quality of tumor-infiltrating T cells, and less in part the quantity, that influence resistance to immunotherapy in pMMR/MSS colorectal cancers. A recent study of colorectal cancer tumor-infiltrating CD8 + lymphocytes found that less than 1% show neo-antigen specificity in colorectal cancer tumor samples [86]. The sample was largely pMMR/MSS (13 of 14 tumors) with a low TMB < 10 mutations/megabase. The proportion of neoantigen-specific lymphocytes may vary with TMB or with exposure to immunotherapy. Aggressive treatment with dual checkpoint inhibitor therapy in pMMR/MSS does not increase tumor lymphocytic infiltration but decreases tumor-infiltrating CD8 + T-cell PD-1 expression ⁸⁷. In mouse models, infiltrating T-cells from pMMR/MSS, low TMB colorectal malignancies are less likely to lose TCF1 transcription factor expression and gain cytoplasmic Granzyme B [25]. These are markers of effector differentiation, an early step in T cell maturation [88, 89]. Their absence in pMMR/MSS tumors is evidence of impaired T cell function in response to low TMB [25].

New Possible Immunotherapy Targets

V-Domain Ig Suppressor of T-Cell Activation (VISTA)

Human VISTA is a 279-amino acid protein that acts as an immune checkpoint molecule for human T cells (Table 3) [90]. In a preclinical study targeting colorectal cancer, it has been shown that VISTA expression was significantly lower in pMMR/MSS colorectal cancers compared to dMMR/MSI-H tumors [90]. Anti-VISTA immune checkpoint inhibitors trials are in early-phase clinical trials in solid tumors (results are pending, NCT04475523).

Table 3.

New possible immunotherapy targets

Study name	Agent	Target	Study population	Primary endpoint	Phase	Seminal results
NCT04475523	CI-8993	VISTA	Advanced solid tumor malignancies	Safety	I	Trial is active and awaiting results
NCT05061628	JS006 only vs JS006 + Toripalimab	TIGIT, PD-1	Advanced refractory tumors	Safety	I	Trial is active and awaiting results
RELATIVITY-047 NCT03470922 [100]	Relatlimab + Nivolumab vs Nivolumab only	LAG3, PD-1	Untreated advanced melanomas	PFS	II/III	Median PFS was 10.1 months (95% CI, 6.4–15.7) in the doublet group compared to 4.6 months (95% CI, 3.4–5.6) in the nivolumab-only group (HR for progression or death = 0,775, p = 0.006) Percentage pf patients with PFS at 12 months was 47.7% (95% CI, 41.8–53.2) in the doublet group and 36% (95% CI, 30.5–41.6) in the nivolumab-only arm

T-Cell Immunoglobulin and ITIM Domain (TIGIT)

TIGIT is a receptor of the Ig superfamily, and multiple studies have demonstrated that it plays a role in limiting adaptive and innate immunity against tumors [91, 92]. There is a suggested role of TIGIT in the regulation of the innate immune response and the intestinal microbiome [93]. Anti-TIGIT antibodies are in early-phase clinical trials in solid tumors (results are pending, NCT05061628).

cGAS-STING Signal Pathway

The cyclic guanosine monophosphate-adenosine monophosphate synthase (cGAS protein) is one of the major pathogens’ recognition receptors in innate immunity [94]. STING is an adaptor protein whose expression is seen on dendritic cells, lymphocytes, macrophages, and endothelial and epithelial cells [95]. It has been studied that the cGAS-STING pathway could regulate CD8 + T-cells differentiation to initiate a response against multiple solid tumors, including colon cancer [96, 97]. The cGAS-STING pathway could represent a new target for immunotherapy-based treatments by regulating macrophage activation.

Lymphocyte-Activation Gene 3 (LAG3)

LAG3 is a transmembrane receptor-function protein and belongs to the immunoglobulin family [98]. It binds with MHC II, suppressing the T-cell receptor signal [99]. A recent study by Zhou et al. found that CD8 + lymphocytes at liver metastasis showed a high frequency of LAG3 compared to the primary tumor. Moreover, it demonstrated that pMMR/MSS colorectal cancer patients with liver metastasis had an increased expression of LAG3 compared with primary tumors [100]. Knowing that liver metastasis in colorectal cancers is the major cause of death from this disease [101], developing new checkpoint inhibitors and targeting these new pathways is being explored. A LAG3 inhibitor co-formulated with PD-1 inhibitor is approved for treatment in melanoma [102].

B- and T-Lymphocyte Attenuator (BTLA)

BTLA is a co-signaling protein expressed by T cells, B cells, natural killer cells, and antigen-presenting cells and belongs to the CD28 immunoglobulin superfamily [103, 104]. When bound to its ligand, HVEM (Herpes virus entry mediator), it recruits SHP-1 and SHP-2 protein tyrosine phosphatases, suppressing T-cell receptor activation [104, 105]. In colorectal cancer patients, it has been shown that the gene expression of BTLA was significantly upregulated compared to the control group [106]. However, the role of BTLA in the prognosis and immunotherapy of colorectal cancer remains unclear and requires further investigation.

Conclusion

In conclusion, current immune checkpoint inhibitors, although essential in dMMR/MSI-H colorectal cancer management, have failed in the treatment of pMMR/MSS colorectal cancers despite many attempts in recent years. There are no FDA-approved indications for the pMMR/MSS group, but in subsets of POLE/POLD1 mutation subsets the benefit is clear, while in other subsets, e.g., with BRAF V600E mutation and immunotherapies combinations with targeted agents is showing promise. Understanding the underlying mechanisms of immune resistance and evasiveness should guide further research in this field to develop novel strategies for overcoming these obstacles.