PD-1/PD-L1 co-inhibition shapes anticancer T cell immunodominance: facing the consequences of an immunological ménage à trois

Abstract

PD-1- and PD-L1-blocking monoclonal antibodies have shown significant promise in clinical settings and rekindled the hope for successful cancer immunotherapy. We recently demonstrated that interfering with PD-1/PD-L1 signaling selectively augments CD8⁺ T cell (T_CD8) responses to subdominant determinants (SDDs) of a model tumor antigen. This was likely due to decreased lysis of SDD-specific T_CD8 by neighboring immunodominant clones co-engaging the same antigen-presenting cells (APCs). We therefore proposed that PD-1-based checkpoint inhibitors widen the range of tumor determinants that can be effectively targeted by T_CD8. Subsequently and using different tumor models, Chen et al. reported, in Proceedings of the National Academy of Sciences of the United States of America, that PD-L1 protects APCs from the lytic function of immunodominant T_CD8 and that PD-L1 blockade narrows, rather than broadens, the overall anticancer T cell response. Here, we briefly compare and contrast the experimental systems employed by the two groups, which may account, at least partially, for the opposing conclusions drawn. We argue that the pathway(s) of tumor antigen presentation, direct presentation versus cross-presentation, and the intensity of PD-1 expression by immunodominant and subdominant T_CD8 must be taken into consideration in rational design of anti-PD-1/PD-L1-adjuvanted tumor vaccines and therapies.

Immunodominance is a fascinating feature of both anticancer and anti-pathogen CD8⁺ T cell (T_CD8) responses, and is responsible for their hierarchical pattern [1]. Accordingly, only a very limited number of peptides from complex antigens (Ags) induce immune responses of sufficient magnitude to give rise to detectable effector T_CD8 clones. Even among the ‘chosen’ peptides, only one or few stand out as immunodominant determinants (IDDs), while T_CD8 that recognize subdominant determinants (SDDs) elicit modest-to-moderate responses. Immunodominance is deemed an impediment to effective therapeutic vaccination against cancer and infectious diseases [2]. An IDD is not necessarily the most protective epitope. Moreover, narrowly focused T_CD8 responses may favor the escape of mutating bugs and the outgrowth of neoplastic cells that do not display bona fide Ags within heterogeneic tumors.

Immunodominance hierarchies are shaped by multiple factors operating at multiple levels, including central T cell education [3, 4], Ag processing and presentation [5–7], and peripheral regulatory mechanisms [8]. Up until recently, it was unclear if signaling through PD-1 (CD279) influences tumor-specific T_CD8 immunodominance. This is a particularly important question in light of the promising clinical responses to PD-1-based ‘checkpoint inhibitors’ in several malignancies [9].

Sustained T cell receptor (TCR) triggering, due for instance to a high tumor load, results in continued expression of PD-1 whose engagement with PD-L1 (aka. B7-H1 or CD274) hampers anticancer T_CD8 functions [10]. This has been ascribed primarily to T_CD8 exhaustion, which can be reversed by anti-PD-1 and/or anti-PD-L1 monoclonal antibodies [11, 12]. We recently established a link between the PD-1-PD-L1 co-inhibitory axis and tumor-specific T_CD8 immunodominance, which prompted us to propose ‘epitope spreading’ as an additional new mechanism underlying the beneficial effects of anti-PD-1-/PD-L1 agents [13].

Using a powerful mouse model in which T_CD8 responses can be monitored against well-defined epitopes of a clinically relevant oncoprotein, namely simian virus 40 (SV40) large tumor Ag (T Ag) [14], we found SDD-specific T_CD8 to express higher PD-1 levels in comparison with their IDD-specific counterparts [13]. Consequently, blocking PD-1 or PD-L1 selectively expanded the clonal size of effector and memory T_CD8 recognizing SDDs without incapacitating IDD-specific T_CD8. Furthermore, we utilized fibrosarcoma and renal epithelial tumor cells expressing T Ag variants, recombinant vaccinia viruses (rVVs) encoding full-length T Ag or T Ag-derived epitope minigenes, and peptide-pulsed dendritic cells (DCs) to demonstrate that PD-1 blockade prevents the fratricidal lysis of SDD-specific T_CD8. In our peptide-pulsing experiments modeling many DC-based vaccination protocols, anti-PD-1 could boost the SDD-specific response only when subdominant and immunodominant T_CD8 engaged the same DCs, thus provoking their participation in an ‘immunological threesome’. Accordingly, interfering with PD-1 signaling should rescue SDD-specific T_CD8 (Fig. 1).

Fig. 1.

Curbing PD-1–PD-L1 interactions may spare PD-1^high subdominant T_CD8 clones. A Many, if not most, tumor cell types are unable to directly prime T_CD8. However, they may provide DCs with Ag processing substrates (a), resulting in cross-presentation of tumor-derived peptides (b) along with a costimulatory signal(s) (c). This, in turn, induces the cross-priming of immunodominant (ID) and subdominant (SD) antitumor T_CD8. On the other hand, peptide-pulsed DCs can be used as therapeutic tumor vaccines to directly activate ID and SD T_CD8 (d). Sustained TCR triggering maintains the expression of PD-1 on T_CD8, which binds PD-L1 on APCs and/or tumor cells and mediates T_CD8 exhaustion. In addition, when/if SD T_CD8 expresses high levels of PD-1 (e), strong signaling through this co-inhibitory molecule may induce or upregulate death receptors (f) and allow SD T_CD8 to receive the fratricidal kiss of death from adjacent ID clones (g). Therefore, PD-1-PD-L1 interactions contribute to T_CD8 immunodominance. B We suggest that Ab-mediated blockade of PD-1 or PD-L1 (h) should prevent death receptor expression by SD T_CD8 and their lysis by ID T_CD8 (i). This enables SD T_CD8 to mount a more vigorous response at a clonal level (j)

More recently, Chen et al. [15] proposed that the presence of PD-L1 on DCs moderates immunodominance disparities in subsequent antitumor T_CD8 responses. While citing our work, these investigators surprisingly stated that “it is unknown whether the B7-H1/PD-1 pathways affects immunodominance”. Regardless, we applaud them for extending the literature on this important subject.

Utilizing DC immunization and elegant tumor models, Chen et al. [15] reported preferential invigoration of IDD-specific T_CD8 in the absence of PD-L1. This was attributed to elimination of co-pulsed DCs by immunodominant T_CD8, thus making Ag-presenting cells (APCs) unavailable for subdominant T_CD8 activation. Of note, the authors did not rule out the possibility that IDD-specific T_CD8 in their models may express higher PD-1 levels, potentially rendering them susceptible to lysis by SDD-specific T_CD8 while co-engaging PD-L1⁺ DCs. We are of the opinion that the presence/strength of PD-1 signaling in antitumor T_CD8 clones may alter their death receptor expression levels and/or cytolytic machinery, which could, in turn, confer a survival advantage upon select clones regardless of their rank in immunodominance hierarchies.

It is also noteworthy that Chen et al. [15] used peptide-pulsed/Ag-presenting DCs to activate T_CD8 and also employed pre-activated T_CD8 against peptide minigene-expressing tumors. These approaches only enable the examination of ‘direct priming’, which is relevant for certain tumor vaccination strategies. We too used costimulation-sufficient DCs and minigene-expressing rVVs to study direct priming. In addition, however, we examined T_CD8 ‘cross-priming’ by costimulation-deficient, MHC-incompatible, and beta-2 microglobulin (β2M)^−/− tumor cells. Cross-priming is instrumental and more relevant to immune surveillance against many malignant cell types that fail to activate naive T_CD8 on their own, due, for instance, to their inability to generate peptide:MHC I complexes and/or suboptimal expression of costimulatory molecules [16, 17] (Fig. 1).

Collectively, we propose that the nature and presentation mode of tumor Ags, and the expression levels of PD-1 in immunodominant and subdominant T_CD8 must be considered in optimizing PD-1/PD-L1-based therapies.

It is now well appreciated that checkpoint inhibitors typically work only in a fraction of cancer patients. Interestingly, in a previous study, serial biopsies from patients with metastatic melanoma revealed many more expanded tumor-infiltrating T cell clones among responders to pembrolizumab than in progressors [18]. Predicting whether and to what extent cancer patients may benefit from treatment with PD-1- or PD-L1-blocking agents remains an area of intense investigation. Future studies will need to address the potential relationship between the ‘breadth’ of antitumor T_CD8 responses and the clinical outcomes of treatment with various immune checkpoint inhibitors.

Abbreviations

Ag(s)

Antigen(s)

APC(s)

Antigen-presenting cell(s)

β2M

Beta-2 microglobulin

DC(s)

Dendritic cell(s)

IDD(s)

Immunodominant determinant(s)

rVV(s)

Recombinant vaccinia virus(es)

SDD(s)

Subdominant determinant(s)

SV40

Simian virus 40

T Ag

[Large] Tumor antigen

T_CD8

CD8⁺ T cell(s)

TCR

T cell receptor

Author contributions

S. M. Mansour Haeryfar conceived the theme of this commentary, collected the relevant literature, and wrote the manuscript. Todd D. Schell provided intellectual input into the paper.

Funding

The authors’ work highlighted in this commentary was funded by a Canadian Institutes of Health Research Operating Grant (MOP-130465) to S. M. Mansour Haeryfar and by a National Cancer Institute/NIH Grant (CA-025000) to Todd D. Schell.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.