Targeting MHC-I inhibitory pathways for cancer immunotherapy

MHC-I downregulation and immunotherapy resistance

The TME is a multifaceted ecosystem comprising heterogeneous cancer cells, stromal cells, and immune cells. Within the TME, CD8⁺ cytotoxic T lymphocytes (CTLs) play a crucial role in eliminating tumor cells. The generation of tumor-specific CD8⁺ T cell responses relies on the effective sensing of tumor antigens presented by MHC class I (MHC-I), or human leukocyte antigen (HLA), which is a heterodimer comprising a heavy chain (HC) and a β2 microglobulin (B2M) light chain, from the antigen-presenting cells (APCs) through antigen-specific T cell receptors (TCRs) on T cells. Cancer cells, in an effort to combat immune elimination, often shut down T cell activities for immune evasion. This process involves an array of inhibitory receptor–ligand interactions [1], the so-called immune checkpoints such as PD-L1/PD-1, in the interface between tumor cells and T cells. The emergence of immune checkpoint blockade (ICB) has significantly advanced the treatment of many cancer types; it is led by anti-PD-L1/PD-1 antibody therapies targeting these immune inhibitory pathways to restore or repair tumor-specific T cell immunity [2]. The success of this therapy is generally thought to arise from ‘inflamed or hot’ tumors with abundant T cell responses, and a large subset of patients do not respond to ICB therapy (primary resistance) or develop acquired resistance, especially in ‘cold’ tumor types with weak T cell infiltration and activity. With the antigenic or ‘first’ signal for CD8⁺ T cell activation, the downregulation of MHC-I by tumor cells or tumor-associated APCs has been closely linked to cold tumors and resistance to immunotherapy, as reviewed elsewhere [3–6]. Nevertheless, a significant challenge persists in developing therapeutic strategies that can precisely and efficiently restore MHC-I AP (Box 1) in the TME with a favorable response:immunotoxicity ratio [7]. Many existing approaches related to MHC-I modulation are intertwined with various other biological mechanisms. Here, we discuss known and possible mechanisms of MHC-I regulation in cancers, with a focus on inhibitory pathways of MHC-I expression and/or AP, as well as their presumed roles in cancer immune evasion. We think that these approaches might offer more viable therapeutic options than other strategies. We also propose, as potential future cancer immunotherapies, strategies targeting these MHC-I inhibitors beyond the modulation of positive regulators related to the complex MHC-I AP pathway (Figure 1, Key figure).

MHC-I modulation by genetic alterations of positive regulators

A substantial body of research on MHC-I regulation in cancers has focused on genetic alterations or reduced expression of ‘positive regulators’ related to key MHC-I AP machinery (APM) components (Figure 1). For instance, homozygous loss of the MHC-I B2M light chain was reported in melanoma patients (three/69) and non-small cell lung cancer patients (one/14) who are resistant to ICB treatment [8,9]. Mutations in MHC-I HC have been also identified in several cancers, such as stomach, cervical, and head/neck cancers, as well as glioma, although at a very low frequency (<4% in pan-cancer) [3,5,10]. In addition, genetic alterations involved in peptide generation (immunoproteasome subunits PSMB8/9/10), transportation [transporter associated with antigen processing (TAP)1/2], trimming (ERAP1/2), and loading [calreticulin (CALR) and tapasin (TAPBP)] have been documented, as well as transcription factors or signaling proteins related to MHC-I induction such as JAK1/2 [downstream of interferon (IFN) cytokine signaling], CITA/NLRC5, and IRF1/2. Such alterations have been observed across multiple types of cancer, such as skin cutaneous melanoma, lung adenocarcinoma, diffuse large B cell lymphoma, and stomach adenocarcinoma (data obtained from cBioportal.org and reviewed in [3,5]). Considering the relatively infrequent occurrence of genetic alterations in these positive regulators among most cancer patients (and even in such patients), only a small fraction of tumor cells seem to harbor such genetic changes [3]. Thus, an ongoing debate persists on whether these mechanisms universally constitute the primary reasons behind the low immunogenicity observed in cancers. Additionally, the therapeutic targeting of these MHC-I positive regulators can be challenging, particularly when attempting to (re)introduce or overexpress these factors specifically within the TME. In this context, type-I IFN therapy may represent one of the earliest therapeutic approaches associated with the positive regulation of MHC-I [11]. However, it is challenging to discern its specific impact on MHC-I compared with its various other effects, due to its pleiotropic nature. Furthermore, its effectiveness has thus far been predominantly noted in the treatment of a select group of cancers including melanoma, hairy cell leukemia, and Kaposi’s sarcoma, while being accompanied by notable immune toxicities [11]. Besides type-I IFN, numerous other MHC-I positive regulators, including STING that induces type-I IFN [12], are intracellular proteins with relatively widespread expression patterns. Introducing these dysregulated factors in vivo through gene therapy or small molecules to enhance their expression or activate those pathways may provide potential solutions to this problem. However, challenges such as specific tumor delivery, efficiency, potential pleiotropic effects, and immunotoxicities could pose significant obstacles [11,13,14].

MHC-I modulation by inhibitory mechanisms as an emerging field of study

In contrast to our more comprehensive understanding of immune checkpoints as co-signaling modulators for T cells, our knowledge of inhibitory mechanisms related to the MHC-I antigenic signaling in cancers remains somewhat limited. We hypothesize that, similar to PD-L1 or other checkpoint ligands induced by cancer cells to suppress T cell activities, tumors might actively upregulate inhibitory molecules to downregulate MHC-I expression and/or AP. Targeting these pathways could potentially restore endogenous MHC-I responses and facilitate tumor control. Notably, negative regulators of MHC-I have emerged as common immune evasion mechanisms produced by many pathogens, particularly viruses, yet their study in the context of cancer remains less explored. Viruses can produce a spectrum of viral factors that actively inhibit almost every step of the MHC-I AP pathway (Table 1), as reviewed elsewhere [15,16]. Notably, during a long course of coevolution with a host, many pathogens might ‘engineer’ new proteins as advantageous MHC-I inhibitory strategies. Hypothetically, for cancer cells, it might be harder to escape immunity compared with viruses because the former can only use the tools encoded in a cell and may attempt immune escape in a limited time frame. In addition, most viruses may not induce genetic alterations at the MHC-I locus (HLA genes) in host cells, which may differ for cancer cells. However, relative to viruses, it remains unclear whether and how tumors employ similar MHC-I inhibitory tactics and whether these mechanisms represent more general and effective immune evasion strategies for cancer cells.

A few recent cancer studies suggested a potential role for signaling proteins and epigenetic or transcriptional regulators in MHC-I downregulation. For instance, high expression of TNF receptor-associated factor 3 (TRAF3), Nedd4-binding protein 1 (N4BP1), and TNFAIP3-interacting protein 1 (TNIP1), suppressors of the NF-κB and IFN signaling pathways, was negatively correlated with MHC-I expression as well as with survival in cancer patients (several cancer types, including neuroblastoma) [17,18]. Knocking out of the genes encoding these proteins also promoted MHC-I expression in melanoma and neuroblastoma cells and sensitized cancer cells to CTL cytotoxicity and ICB responses [17,18]. PTPN2 and DUX4 suppressed IFN-γ-mediated MHC-I induction through distinct mechanisms related to JAK1 and/or STAT1 [19,20]. Moreover, epigenetic regulators such as Polycomb repressive complex 2 (PRC2), DNA methyltransferases (DNMTs), and histone deacetylases (HDACs) impeded access to promoter/enhancer regions of transcription factors related to the expression of MHC-I and the APM components [3,5,6]. Treatment using demethylating drugs or inhibitors targeting EZH2 (the catalytic core of PRC2) or HDACs increased MHC-I expression and enhanced CTL activity [3,5,6,21]. Moreover, thymidylate synthetase, involved in DNA replication and repair, has been associated with MHC-I regulation in human lymphoma cells [21]. Some miRNAs were also reported to impede the translation of APM components such as NLRC5, HLA-C, TAP, Erp57, calnexin, and PSMB8/LMP7 [5,6]. In addition, the RNA-binding protein MEX3B and RNA-binding E3 ubiquitin-protein ligase MEX3C can bind to the 3′-UTR of HLA-A mRNA, leading to its destabilization and reduced surface expression [22,23] (Figure 1).

Furthermore, recent findings suggest very specific mechanisms related to MHC-I inhibition. The oncoprotein staphylococcal nuclease and Tudor domain containing 1 (SND1) directly interacted with the MHC-I HC in the endoplasmic reticulum (ER) lumen and hijacked the nascent HC to ER-associated degradation (ERAD) by the cytosolic proteasome. Consequently, knocking out of SND1 enhanced MHC-I expression and reduced tumor progression in melanoma mouse models [24]. Also, the intracellular protein autophagy cargo receptor NBR1 and receptor-interacting protein kinase 2 (RIPK2) synergistically directed MHC-I for autophagy-mediated lysosomal degradation in pancreatic cancer mouse models [25,26]. Knocking down of NBR1 or RIPK2 by shRNA or inhibition of autophagy restored MHC-I expression and AP in pancreatic cancer cells and sensitized pancreatic ductal adenocarcinoma (PDAC) to ICB therapy in orthotopic mouse models [25,26]. In addition, the secreted protein PCSK9, a regulator of low-density lipoprotein (LDL) receptors, interacted with MHC-I on the cell surface, leading to its internalization and subsequent lysosomal degradation [27]. Consequently, PCSK9 neutralizing antibodies potentiated ICB therapy in melanoma and mammary carcinoma murine models [27]. Beyond these pathways, in various tumor models, a surface MHC-I masking mechanism was shown to hinder the accessibility of MHC-I to CTL TCR recognition and anti-MHC-I antibodies via B3GNT5-mediated promotion of cell-surface glycosphingolipid (GSL) synthesis; this in turn was overcome by overexpression of SPPL3, an enzyme that proteolytically inactivates B3GNT5 [28]. Accordingly, the application of clinically approved GSL synthesis inhibitors enhanced MHC-I accessibility and CTL responses in vitro [28], suggesting an important role for the GSL pathway in MHC-I regulation (Figure 1).

In general, it is exciting to witness the growing body of work evaluating MHC-I inhibitors in the context of cancer. Future studies are eagerly anticipated to delve into their detailed mechanisms, their cancer relevance, and the development of targeted approaches and to examine their therapeutic potential in the treatment of cancer. However, given the wide range of functions of many of these molecules – which extend beyond MHC-I regulation – it also becomes essential to meticulously dissect their specific influence on MHC-I in contrast to their diverse effects, especially in cancers. For instance, it is worth noting that targeting NBR1 and PCSK9 might also impact postnatal osteoblastic bone formation and lipid metabolism, respectively [29,30]. Furthermore, although specific types of GSLs have been proposed as tumor-associated antigens, GSL synthesis inhibitors are used to treat lysosomal storage diseases such as Gaucher disease [31]. Additionally, most of these molecules, except PCSK9, are intracellular proteins with broad expression, which pose challenges when attempting to target them, similar to the potential obstacles in targeting MHC-I positive regulators.

In our perspective, there is an increasing need to uncover additional cancer-specific MHC-I inhibitory mechanisms, which include but are not limited to: (i) other extracellular proteins that are specifically produced by cancer cells or the TME and are directly associated with the downregulation of MHC-I expression, and which might serve as putative targets for antibody or small molecule-based therapeutics; (ii) intracellular proteins with specific expression in cancer cells or the TME and a predominant role in regulating MHC-I, and which might serve as candidate therapeutic targets for siRNA or small-molecule drug interventions; and (iii) other factors that go beyond the general regulation of MHC-I expression and instead specifically modulate the process of MHC-I AP. Studies to uncover other key MHC-I evasion mechanisms in the TME will also be pivotal. For instance, acidification, which is a cancer hallmark and can regulate MHC-II on APCs, should be investigated [32,33]. We posit that these areas of research may hold significant promise in advancing our understanding of cancer immune evasion of MHC-I AP and the potential development of novel therapeutic strategies to increase the quality and quantity of tumor-specific CD8⁺ T cells.

A membrane-associated MHC-I inhibitory axis

When reviewing the MHC-I inhibitory mechanisms employed by viruses, it becomes apparent that, in addition to hijacking various intracellular steps of MHC-I AP (as outlined in Table 1), viral factors possess the ability to directly engage with MHC-I through cis interactions on the plasma or organelle membrane, mediating MHC-I downregulation. For instance, the murine cytomegalovirus (MCMV) m4/gp34 protein is associated with peptide-loaded MHC-I (pMHC-I) on the cell surface to prevent specific MHC-I allele-restricted CTL recognition [16,34]. Increased H2-K^b-restricted CTL-mediated lysis of B6 mouse embryo fibroblasts was observed in cells infected with a MCMV m4/gp34 deletion mutant [34]. Moreover, overexpression of Kaposi’s sarcoma-associated herpesvirus (KSHV) kK3/kK5 in BJAB cells or HeLa cells and Epstein–Barr virus (EBV) BILF1 in 293 or MJS cells induced robust MHC-I internalization and lysosomal degradation through direct association with MHC-I [35–37]. Surprisingly, this type of mechanism has rarely been studied in the context of cancer. Nevertheless, recent findings suggest the importance of immune cis interactions in regulating T cell responses, such as cis PD-L1/B7–1 to disrupt PD-1 interaction and facilitate T cell activation through CD28 [38], as well as cis PD-L1/CMTM6 interactions to stabilize PD-L1 [39]. Therefore, a question arises: could tumors employ similar cis membrane interactions to mediate surface MHC-I inhibition?

To answer this question, our group recently performed genome-wide pMHC-I-guided CRISPR-Cas9 screens in both human and mouse acute myeloid leukemia (AML) cell lines and constructed a comprehensive network of both positive and negative regulators of membrane MHC-I expression [40]. Many of those negative AP regulators are associated with protein ubiquitination, vesicular transport, DNA metabolism processes, and the cell cycle. In particular, sushi domain containing 6 (SUSD6), transmembrane protein 127 (TMEM127), and the E3 ubiquitin ligase WWP2 emerged as the top candidates affecting both MHC-I expression and MHC-I AP [40]. Genetic ablation of SUSD6 and TMEM127 enhanced tumor MHC-I expression and MHC-I AP as well as tumor-specific CD8⁺ T cell responses in vitro. Furthermore, SUSD6 and TMEM127 depletion by shRNA or sgRNA significantly controlled tumor growth and extended animal survival in our well-established C1498 leukemia and B16F10-OVA melanoma mouse syngeneic tumor models; moreover, survival was associated with improved CD8⁺ T cell infiltration and effector function in the TME. Mechanistically, SUSD6 and TMEM127 directly interacted with MHC-I molecules in cis and recruited the E3 ligase WWP2 to the proximity of the MHC-I cytosolic tail for its ubiquitination. The ubiquitinated MHC-I was subsequently targeted for lysosomal degradation, leading to reduced MHC-I expression and T cell activation [40]. Our findings highlight the SUSD6, TMEM127, and WWP2 (STW) axis as a tumor-associated MHC-I inhibitory axis in blood and solid cancers. Importantly, we also found that these molecules are highly expressed in a number of cancers (most of which are cold tumors), including AML, pancreatic cancers, and others; the expression of these molecules was negatively associated with patient survival or response to immunotherapy [40]. Of note, TMEM127 and WWP2 can directly associate with MHC-II degradation in the presence of the Salmonella effector protein SteD [41]. This is interesting as it suggests a convergent, albeit different, mechanism that can be employed by both tumors and certain pathogens.

Several intriguing questions remain for further investigation regarding this membrane-associated MHC-I inhibitory axis. These include: (i) the underlying mechanisms regulating the expression of these molecules in various cancer or immune cells; (ii) the specific interacting interfaces among these molecules; (iii) the intracellular pathways employed by the STW complex to transport surface MHC-I into the lysosome; (iv) the potential functions of these molecules beyond their role in MHC-I modulation; and (v) the development of therapeutic strategies to target these molecules for immunotherapy in cancers and/or other human diseases. Furthermore, it is also worthwhile to study other putative inhibitors that were identified in our laboratory screens as well as those of others. The investigation of such molecules in MHC-I modulation and tumor immunity may facilitate our understanding of immune regulation in the TME and foster potential MHC-I inhibitor-based cancer immunotherapies.

Clinical studies related to MHC-I inhibitory mechanisms

Given the central roles of MHC-I regulation in antitumor T cell immunity, efforts have been made to uncover therapeutic strategies to restore aberrant MHC-I expression and/or boost MHC-I AP in cancers. As mentioned above, several approaches related to the reintroduction of positive regulators, such as type-I IFN or STING agonists, have been used in a number of cancer trials [11,14]. For instance, type-I IFN can be effective in some cancer patients despite the potential to elicit severe side effects and broader effects beyond MHC-I modulation [11]. Unfortunately, STING agonists have been often linked to high immune toxicities and limited efficacy when administered systemically [14]. Tumor-specific delivery of these agents, along with gene therapy targeting other positive regulators, might constitute possible options to improve the therapeutic window; however, delivering treatments to tumors has remained a longstanding challenge [13]. Although radiotherapy and chemotherapy have been shown to upregulate cancer MHC-I expression [42], this was not our focus here.

On the flip side, we are still in the early stages of exploring the therapeutic potential of MHC-I inhibitory pathways, and several programs targeting these aforementioned mechanisms are still being evaluated in clinical trials. For instance, the EZH2 inhibitors tazemetostat (NCT05467748)ⁱ and XNW5004 (NCT06022757)ⁱⁱ, in combination with PD-1 blockage, are now under Phase 1/2 evaluation (Table 2). Given the potential involvement of autophagy in MHC-I regulation [25], several autophagy inhibitors, such as chloroquine and ezurpimtrostat, are now being assessed for cancer immunotherapy in different clinical settings (NCT04214418ⁱⁱⁱ, NCT05448677^iv). Moreover, two PCSK9 monoclonal antibodies have entered clinical trials in combination with anti-PD-1 ICB in metastatic cancers (NCT05553834^v and NCT05128539^vi), following promising preclinical data [27]. Inhibitors of GSL synthetase ([28]; NCT04944888^vii), SMAC ([17]; NCT05082259^viii, NCT04122625^ix), thymidylate synthase ([21]; NCT04396457), DNMTs ([43]; NCT02260440^x), HDACs (NCT02437136^xi), MEK (NCT04216953^xii), BET (NCT05422794^xiii), and the androgen receptor ([44]; NCT04946370^xiv), among others, are currently also under clinical investigation to explore their potential in cancer immunotherapy (Table 2).

Strictly speaking, many of these pathways are associated with other critical cellular functions, making it challenging to determine their specific contribution to MHC-I modulation. While we argue that targeting MHC-I inhibitory pathways may represent a highly viable option in cancer therapeutics, it is essential that we critically evaluate these molecules and further identify those pathways with strong relevance to tumors and specific functions in MHC-I modulation, especially for different types of cancers. Given the specific role of STW complexes in MHC-I inhibition and their high expression in blood and solid cancers, it will be interesting to further explore this inhibitory axis and develop therapeutic strategies for cancer immunotherapy that rely on antibodies or small molecules.

Concluding remarks

While MHC-I AP represents a fundamental pathway for T cell immunity, there is limited knowledge of how tumor cells actively evade this critical immunological process. MHC-I evasion can occur via the induction of MHC-I inhibitors that extend beyond genetic alterations of MHC-I positive regulators. As discussed in this opinion, many known MHC-I inhibitors execute broad functions beyond MHC-I modulation. This highlights the need to delve deeper into the elucidation of specific mechanisms of regulated MHC-I expression that are robustly relevant for certain tumors. This is a rising field of study with many key questions that remain to be addressed (see Outstanding questions).

One important limitation to our viewpoint is that immunotherapy resistance can be caused by other mechanisms beyond MHC-I AP, such as MHC-II regulation, and other putative mechanisms of immune suppression of T cells, myeloid cells, or stromal cells. The relative contributions of diverse mechanisms in the control of tumor immunity require thorough examination, especially for different types of cancers. Other than cis regulation of MHC-I, trans regulation via cell–cell interactions is also possible. Moreover, MHC-I inhibitory mechanisms might fluctuate over time during cancer development; therefore, it will be important to design diagnosis tools to monitor these changes. Conceivably, certain cancer types might not exhibit a dominant MHC-I inhibitory mechanism, which highlights the importance of patient stratification. Nevertheless, we posit that the discovery and rigorous characterization of tumor-associated MHC-I inhibitory pathways will be pivotal to furthering our understanding of MHC-I regulation in cancers and inform novel MHC-I inhibitor-based candidate therapies to overcome immunotherapy resistance.