For Whom the Bell Tolls? A Toll-Like Receptor 9 Agonist's Journey from Vaccine Adjuvant to Promising Agent in Anti–PD-1–Resistant Melanoma

Since 2011, there have been 12 new drugs and/or regimens approved to treat patients with advanced, unresectable stage III or stage IV melanoma. Remarkably, other than the original approvals of the antagonist antibodies that inhibit PD-1 and that of the anti-CTLA4 mAb ipilimumab outside the United States, none of these approvals were for patients in the treatment-refractory population. Because the long-term data with first-line anti–PD-1–based therapy, either as a single agent or in combination with anti-CTLA4, show durable survival in 40% to 50% of patients compared with around 5% in historical series, there is a tendency toward celebrating this remarkable achievement and overlooking the fact that more than half of patients with advanced melanoma will still die of their disease (1, 2). Perhaps the greatest unmet in the field is to develop effective therapies for the anti–PD-1–resistant population.

To adequately overcome this major challenge, an understanding of the mechanisms of this resistance is necessary. The identification of numerous primary and secondary resistance mechanisms has helped to clarify the types of therapeutic approaches that should be tested in these patients, although to date there has been limited success in general, as well as in specifically identifying which patients may benefit from a specific approach (3). Unfortunately, a major limitation to the trials presently in this space has been the absence of identifying the mechanism of resistance to anti–PD-1 and populations of patients (either defined clinically or via biomarker) that identifies which patients are most likely to benefit from a therapy tested in those with anti–PD-1 resistance.

One of the approaches being explored in patients with anti–PD-1 resistance is intralesional therapy with oncolytic virus, Toll-like receptor (TLR) agonists, and other agents. TLR9 agonists have been of particular interest given the diverse effects that activating TLRs have on the immune microenvironment, and due to recent reports of their use in patients with anti–PD-1–resistant melanoma (refs. 4, 5; Ideara press release, March 18, 2021: https://ir.iderapharma.com/news-releases/news-release-details/idera-pharmaceuticals-announces-results-illuminate-301-trial). Specifically, TLR9 is activated by CpG dinucleotides, which are found regularly in various microbes and, once engaged, lead to a number of consequences, including enhanced T-cell effector function. CpG is a potent inducer of monocyte differentiation into plasmocytic dendritic cells and has been used as a vaccine adjuvant for the better part of two decades (6, 7). More recently, it has been hypothesized that intralesional CpG may reinvigorate a tumor microenvironment devoid of antitumor immune elements and lead to a reversal of therapeutic resistance to anti–PD-1, and it has led to the launch (and completion) of a number of clinical trials. Furthermore, it has been shown that CpG-A is more potent than CpG-B and CpG-C in leading to differentiation of IFNγ-producing plasmacytoid dendritic cells (8). Whether this will drive a more potent antitumor response than other CpGs or TLR agonists is not yet known, although there are now available clinical data with both CpG-C and CpG-A molecules to discuss (4, 5, 9).

The initial data from a phase Ib study of a synthetic CpG-C, SD-101, in combination with the anti–PD-1 pembrolizumab, in patients with advanced melanoma determined the combination to be safe and led to responses in patients with PD-1 inhibitor–naive (7 of 9) and PD-1 inhibitor–experienced patients (2 of 13), one of whom had PD-1 inhibitor–refractory disease (4). Furthermore, in the preliminary analysis of a phase Ib/II study of tilsotolimod, a CpG-C, in combination with ipilimumab in 25 patients with advanced melanoma, the combination was determined to be safe and the response rate was 36% (5). However, a follow-up phase III trial randomizing patients with ipilimumab with or without intralesional tilsotolimod recently reported, in a press release, that this trial did not meet its coprimary endpoint of improvement of objective response rate (the analysis of the other coprimary endpoint, overall survival, is immature), with 8.8% of patients treated with the combination versus 8.6% with ipilimumab alone having a response (Ideara press release, March 18, 2021: https://ir.iderapharma.com/news-releases/news-release-details/idera-pharmaceuticals-announces-results-illuminate-301-trial). It is not clear why there is such a discrepancy in the response rates of this combination in the phase Ib/II versus the phase III trials, but perhaps it relates to a stricter eligibility criterion regarding anti–PD-1 refractoriness in the definitive study than the earlier-phase trial. For certain, these data ensure that this combination will not be studied any further in patients with advanced melanoma unless longer-term analysis of overall survival is compelling.

In this issue of Cancer Discovery, Ribas and colleagues (9) summarize the clinical and limited correlative analyses of a phase Ib study of vidutolimod, a CpG-A TLR9 agonist, given in combination with the anti–PD-1 pembrolizumab. Vidutolimod is a CpG-A packaged within a viral-like particle formed by Qb, a bacteriophage coat protein. This is an important trial as it demonstrates the tolerability of this combination and preliminary efficacy in this previously described unmet-need population, namely, patients with advanced melanoma previously treated with anti–PD-1. Specifically, the patient population was represented by patients who rarely previously responded to prior anti–PD-1 (3 of 44; 7%) and nearly all whom had progressed on anti–PD-1 (40 of 44; 91%) before enrollment. The median time on prior anti–PD-1 was 5 months, and 27 (61%) patients had received a most recent dose of anti–PD-1 within 2 months. Pembrolizumab was given per the FDA label every 3 weeks, whereas vidutolimod was initially given weekly for the first 7 (schedule A) or 2 (schedule B) weeks and then every 3 weeks thereafter. In general, the combination was tolerable, with only 2 of 44 patients needing to stop therapy due to treatment-related adverse events. Still, 20 (45%) patients experienced a grade 3 and/or grade 4 treatment-related adverse event, and these types of events were almost exclusively seen in the first 7 weeks during the weekly administration of vidutolimod. From an efficacy standpoint, 11 (25%) patients had a partial (7 of 11) or complete (4 of 11) response to therapy, and responses were seen in injected and noninjected tumors as well as visceral, noninjected tumors.

Strengthening the clinical data is a well-designed correlative analysis plan that yielded a number of interesting findings. First, there appeared to be no correlation between baseline PD-L1 expression and/or CD8⁺ T-cell infiltration and/or inflammatory gene signature and response to therapy. Second, vidutolimod plus pembrolizumab led to an increase in tumor-infiltrating CD8⁺ T cells, PD-L1 score, and CXCL9 and CXCL10 levels in a patient with response to therapy, whereas this pattern was seen in only one of four nonresponding patients. Third, the combination of vidutolimod and pembrolizumab led to an increase in circulating CXCL10 levels with the third dose and beyond (but not with the first or second), and those patients who responded to therapy had a numerically increased (although not statistically significantly so) fold increase of CXCL10 and IL6, as well as lower tumor CD8⁺ T-cell levels. Finally, treatment was associated with an increase in anti-Qb antibody induction, an expected response to the viral-like particle packaging of the CpG-A, but the level of induction did not correlate with efficacy as was predicted in animal models.

In summary, Ribas and colleagues (9) have completed a compelling phase Ib/II trial of the combination of vidutolimod, a TLR9 agonist, with pembrolizumab. The response rate of 25% in this convincingly PD-1–resistant population and the pretreatment analysis of the tumor suggest that this treatment may be an important option for patients with PD-1–resistant melanoma whose tumors are immunologically cold, a finding that, if reproduced in a larger data set and ideally in biomarker-selected (e.g., low CD8, low PD-L1 expression) patients, would lead to an change in the way that this group of patients is managed.

R.J. Sullivan reports grants and personal fees from Merck; grants from Amgen; personal fees from Novartis, Eisai, Iovance, AstraZeneca, Pfizer, and Array Biopharma, and personal fees from OncoSec outside the submitted work. No other disclosures were reported.