HLA Class-I Antigen Processing Machinery Defects in Antitumor Immunity and Immunotherapy

Abstract

HLA class-I antigen processing machinery (APM) plays a crucial role in the synthesis and expression of HLA class-I-tumor antigen-derived peptide complexes; the latter mediate the recognition and elimination of malignant cells by cognate T cells. Defects in HLA class-I APM component expression and/or function are frequently found in cancer cells, providing them with an immune escape mechanism which has relevance in the clinical course of the disease and in the response to T cell-based immunotherapy. The majority of HLA class-I APM defects (>75%) are caused by epigenetic mechanisms or dysregulated signaling and therefore can be corrected by strategies which counteract the underlying mechanisms. Their application in oncology is likely to improve responses to T cell-based immunotherapies, including checkpoint inhibition.

HLA class-I APM component defects in cancer

The resistance to checkpoint inhibitor-based immunotherapy experienced by a high percentage of patients withsolid tumors has rekindled interest in the characterization of the defects in HLA class-I APM component expression and/or function frequently found in cancer cells. Recognition of cancer cells by cognate T cells is mediated by the transmembrane HLA class-I trimolecular complex, which consists of a three-domain heavy (α) chain, a β₂-microglobulin (β₂M) light chain and a loaded tumor antigen (TA)-derived peptide. These complexes are synthesized by HLA class-I APM, which generates 8–12 amino acid long peptides from TAs and loads them on β₂M-associated HLA class-I heavy chain heterodimers. The resulting trimolecular complexes are transported to the cell membrane for presentation to cognate T cells (Figure 1).

Figure 1: Function of HLA Class-I Antigen Processing Machinery.

Ubiquitinated proteins are degraded by proteasome catalytic subunits: β1 (Delta), β2 (Zeta) and β5 (MB1), or cytokine-inducible immunoproteasome catalytic subunits: LMP2, LMP10, and LMP7. The generated peptides enter the ER via the heterodimeric transporter associated with antigen processing complex (TAP1/TAP2). Often following length refinement by ER aminopeptidases (ERAP1, ERAP2), peptides are loaded onto β2M-associated HLA class-I heavy chain with assistance of chaperones tapasin, ERp57, and calreticulin. The resulting β2M-HLA class-I heavy chain-peptide trimers travel to the cell surface; their expression is regulated through clathrin-independent endocytosis. Mutations/deletions of any HLA class-I APM component may impair the synthesis/expression of β2M-HLA class-I heavy chain-peptide trimolecular complexes on the cell surface.

Typically, endogenous proteins are ubiquitinated and subsequently degraded by multi-subunit proteasome complexes in the cytosol. The constitutive proteasome catalytic subunits include β1, β2 and β5. Alternative catalytic subunits comprise the immunoproteasome, namely, low molecular weight protein 2 (LMP2), LMP10 and LMP7 (also known as β1i, β2i/MECL1, and β5i, respectively) to produce peptides with a hydrophobic C terminus (1). The immunoproteasome, which is induced under cell stress and inflammation, primarily through interferon-γ (IFNγ) stimulation, is thought to perform its proteolytic function more efficiently than the proteasome; whether the immunoproteasome plays a more significant role than the proteasome in the cleavage of TAs is unclear at present (1).

Proteasomal degradation products are shuttled by the heterodimeric transporter associated with antigen processing (TAP) complex into the lumen of the endoplasmic reticulum (ER). Peptide lengths are refined by ER aminopeptidase 1 (ERAP1) and ERAP2 (2). Furthermore, newly synthesized HLA class-I heavy chains are stabilized by calnexin in the ER and then associate with β₂M. Peptides with the correct length and sequence for HLA class-I allele binding are loaded onto β₂M-associated HLA class-I heavy chain dimers with the help of the peptide loading complex (PLC). The latter consists of TAP, oxidoreductase ER resident protein 57 (ERp57) and chaperone molecules calreticulin and tapasin (3). The resulting trimers then travel to the plasma membrane via the Golgi apparatus and are presented to cognate T cells.

HLA class-I trimolecular complexes are removed or recycled from the cell membrane via clathrin-independent endocytosis. These complexes are then internalized by Arf6-GTP-enriched vesicles and reach EEA1+/Rab5+ sorting endosomes (4). Here, they are recycled either quickly to the membrane through the Rab35 regulated early endosomal recycling pathway or moved to the Rab11+ endocytic recycling compartment for slow recycling. Rab22 mediates the formation of tubular recycling endosomes from the endocytic recycling compartment, which undergo Arf6-dependent fusion with the plasma membrane (5). Sub-optimally loaded HLA class-I molecules usually undergo degradation after reaching the sorting endosomes, but may also be recycled via the late endosomal recycling pathway (4).

Defects in HLA class-I APM component expression have been detected in most, if not all of the cancer types analyzed (6,7). HLA class-I heavy chain defects range between a minimum of 36% in urinary bladder cancer and a maximum of 80% in penile cancer, whilst β₂M defects range between a minimum of 17% in bone and soft tissue cancer and a maximum of 73% in uveal melanoma (6). Brain cancers often are associated with defects in TAP1 (40%) and TAP2 (88%) (7). Defects in other APM components have been observed in a range of cancer types, however, their extent has been poorly characterized. These defects play a role in disease progression and have a negative impact on the clinical response to T cell-based immunotherapy (6,7).

Analysis of the mechanism(s) underlying these defects has mainly focused on structural mutations in HLA class-I APM genes, although they constitute a small proportion of total HLA class-I APM defects. Indeed, comparison between HLA class-I APM component mutation frequency and the frequency of protein-level downregulation revealed that mutations in HLA class-I APM components constitute at most 25% of defects (8). Nevertheless, HLA class-I APM downregulation by epigenetic mechanisms or dysregulated signaling, which appear to be the most frequent cause of HLA class-I APM defects, have been neglected.

Here, we review HLA class-I APM defects found in malignant cells. Then we describe their underlying mechanisms, emphasizing the major role played by epigenetic and signaling abnormalities. Finally, we discuss the potential impact of these defects on the clinical response to T cell-based immunotherapy as well as the available approaches to restore HLA class-I APM expression and/or function in malignant cells.

Structural mechanisms underlying HLA class-I APM component defects

Defects in HLA class-I alleles found in cancer pertain to loss of the gene products of one or two HLA class-I loci or of the HLA class-I allo-specificities encoded by the genes present in the MHC region of parental chromosome 6. In the absence of one or more HLA class-I allele(s), the diversity of peptides presented on cancer cells generally decreases. Mutations in β₂M, as well as TAP1 and tapasin can result in total loss of HLA class-I expression, whilst structural defects in other HLA class-I APM genes generally cause HLA class-I downregulation (8). In view of the co-dominance of the two genes, one of paternal and the other one of maternal origin, encoding APM components as well as β₂M, two events are required to cause lack of their expression. Typically, in the case of β₂M, the total or partial loss of chromosome 15, which carries the β₂M encoding gene, is followed by a structural mutation in the other β₂M allele (9–11). In the case of APM, the loss of one component typically results from the combination of a cancer-unrelated germ-line mutation, which targets one encoding allele, with a cancer-related repression of the other allele. HLA class-I APM component mutation frequency is significantly higher in patients with high microsatellite instability (MSI-H) (12). At variance with tumor mutation burden, which is one of the determinants of immunotherapy response (13–16), structural defects in HLA class-I subunits are implicated in primary and acquired resistance to T cell-based immunotherapy. Indeed, homozygous deletions and mutations in B2M are associated with resistance to ICI-based therapy in advanced melanoma patients (17,18). Moreover, loss of heterozygosity of germline HLA class-I heavy chain at one or more loci in advanced cancer patients receiving ICIs is associated with poor overall survival (19). Likewise, knockout of B2m in mouse lung cancer cells confers resistance to anti-PD-1 mAb-based therapy in syngeneic, immunocompetent mouse tumor models (20).

Non-structural mechanisms underlying HLA class-I APM component defects

Non-structural defects are the most frequent cause of HLA class-I APM component downregulation in malignant cells. At least three-fourths of defects in HLA class-I heavy chain, β₂M, and APM components are non-structural (8). Most defects occur transcriptionally or post-translationally, and often result in HLA class-I downregulation, which may lead to undetectable HLA class-I expression. These defects can often be corrected by cytokines and/or by demethylating agents, highlighting the role of epigenetic mechanisms in the regulation of HLA class-I APM component expression (Figure 2).

Figure 2: Mechanisms Regulating HLA Class-I Antigen Processing Machinery Component Expression in Cancer Cells.

HLA class-I APM components are involved in the synthesis and expression of HLA class-I trimolecular complexes. Their expression is regulated by transcription factors IRF1 and NF-κB, as well as enhanceosome transactivator NLRC5. Overexpression and/or activation of histone deacetylases, DNA methyltransferases, and polycomb repressive complex 2 (PRC2) enzymatic component EZH2 may reduce DNA accessibility of HLA class-I APM component promoter/enhancer regions, preventing transcription factors from binding. Activating mutations or amplifications in the receptor tyrosine kinase/MAPK signaling pathway may inactivate STAT1, downregulating IRF1, NLRC5, and HLA class-I APM component expression. MAPK activation may also inactivate STAT3, leading to HLA class-I APM component downregulation. Type I and type II interferons upregulate HLA class-I APM component expression through JAK/STAT1 activation. cGAS, a cytoplasmic DNA sensor that activates stimulator of interferon genes (STING) and downstream TBK1/IKKε, may also induce HLA class-I APM component transcription. IRF3/IRF7-dependent type I interferon secretion represents one of the underlying mechanisms. The other mechanism is canonical NF-κB activation. The latter is also induced by tumor necrosis factor α (TNFα) and Toll-like receptor (TLR)/interleukin-1 receptor (IL-1R) agonists. Non-canonical NF-κB activation mediated by binding of ligands to TNF receptor superfamily members can also induce HLA class-I APM component transcription. Post-translationally, autophagy receptor NBR1 mediates HLA class-I trimolecular complex degradation via the autophagy-lysosomal pathway, while secreted enzyme PCSK9 mediates degradation of HLA class-I trimolecular complexes via the endosomal-lysosomal pathway.

HLA class-I heavy chain expression is regulated at three major transcription factor binding sites: Enhancer A region, IFN-stimulated response element (ISRE) and SXY module. They are recognized by nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB), interferon regulatory factor 1 (IRF1) and HLA class-I enhanceosome, respectively. The latter consists of regulatory factor X (RFX) family transcription factors (RFX5, RFXAP, RFXANK), activating transcription factor (ATF)/cAMP responsive element binding protein (CREB) family transcription factors (ATF1, CREB1), nuclear factor Y (NFY) transcription factor complex (NFYA, NFYB, NFYC), and enhanceosome transactivator NLR family CARD domain containing 5 (NLRC5) (21,22). The distal promoter displays additional upstream stimulatory factor 1 (USF1), USF2, and specificity protein 1 (Sp1) binding sites in a locus and allele-specific manner (23). Notably, NF-κB enhancer A binding is also locus-specific, with two functional κB sites for HLA-A, one for HLA-B, and none for HLA-C (24). Many of these transcriptional regulators also promote β₂M and APM component gene expression (25), along with several others, such as E2F1 in the case of tapasin (26) (Figure 3). In non-malignant cells, these proteins maintain HLA class-I heavy chain, β₂M, and APM component expression at a basal level. However, malignant cells may display alterations that i) impair binding of these proteins to the promoter/enhancer region of HLA class-I APM genes, ii) downregulate the expression of these transcriptional regulators, or iii) accelerate post-translational HLA class-I trimolecular complex degradation. The relative frequency of these alterations varies by cancer type.

Figure 3: Transcriptional Regulation of HLA Class I APM Components.

Processed ChIP-sequencing and ATAC-sequencing data in the HepG2 hepatocellular carcinoma cell line published as part of the ENCODE project were used to generate tracks on the UCSC Genome Browser (accession ID annotated at the top of each track). GENCODE V36 was used to annotate transcripts and CpG islands were generated using the default method on the UCSC Genome Browser. ATAC-seq, H3K27 acetylation ChIP-seq, H3K4 trimethylation ChIP-seq, H3K27 trimethylation ChIP-seq, SP1 ChIP-seq, E2F1 ChIP-seq, RELA ChIP-seq, IRF1 ChIP-seq, RFX5 ChIP-seq, RFXAP ChIP-seq, RFXANK ChIP-seq, CREB1 ChIP-seq, ATF1 ChIP-seq, NFYA ChIP-seq, NFYB ChIP-seq, and NFYC ChIP-seq tracks are shown. Analysis was performed on HLA-A (A), HLA-B (B), HLA-C (C), B2M (D), PSMB5 (MB1; E), PSMB6 (Delta; F), PSMB7 (Zeta; G), PSMB8 (LMP7; H), PSMB9 (LM P2)/TAP1 bidirectional promoter (H), TAP2 (H), ERAP1 (I), ERAP2 (I), PSMB10 (LMP10; J), CANX (calnexin; K), CALR (calreticulin; L), TAPBP (tapasin; M), and PDIA3 (ERp57; N). HLA-B appears to be dysregulated in this cell line, displaying low levels of H3K27 acetylation and an abnormal distribution/intensity of transcription factor ChIP peaks compared to HLA-B in other cell lines.

HLA class-I APM component downregulation by epigenetic mechanisms

Repressive hypermethylation of DNA at the 5-position of cytosine in HLA class-I heavy chain, β2m and APM component encoding gene promoter regions has been described in several cancer cell lines and surgically removed patient tumors (27–29). Treatment with DNA methyltransferase (DNMT) inhibitors reverses DNA methylation, and enhances cancer cell susceptibility to T cell-based immunotherapy through restoration of HLA class-I APM component expression (30,31). Similar results have been obtained with inhibitors of enhancer of zeste homolog 2 (EZH2), a histone methyltransferase enzyme (32), which is overexpressed in many cancer types (33). EZH2 inhibition reverses repressive trimethylation of histone H3 at lysine 27 (H3K27me3) and upregulates HLA class-I expression on head and neck squamous cell carcinoma (HNSCC) cells. The resulting improvement in TA presentation to cognate cytotoxic T cells restores the efficacy of anti-programmed cell death protein-1 (anti-PD-1)-based immunotherapy in HNSCC models (34). Furthermore, histone deacetylase (HDAC) inhibitors upregulate HLA class-I APM component expression in breast and Merkel cell carcinoma cell lines in vitro (35,36). Enhanced histone acetylation at lysine(s) represents the underlying mechanism, as this process increases accessibility of HLA class-I APM chromatin to transcription factors and possibly reverses the repressive chromatin state induced by lack of histone acetylation (37). The induced HLA class-I upregulation enhanced the clinical efficacy of immune checkpoint inhibition (ICI) in a Merkel cell carcinoma patient (38). Identifying other cancer types where DNMTs, EZH2, and HDACs regulate HLA class-I APM expression is essential in the rational design of clinical trials combining epigenetic therapy with T cell-based immunotherapy, a strategy which has become increasingly popular. Epigenetic agents display a wide range of immunostimulatory activities beyond HLA class-I APM component upregulation. They may increase TA expression including cancer/testis antigens (CTAs), induce type I IFN signaling by activating endogenous retroviruses (ERVs), and inhibit MYC-mediated immunosuppression (39). Moreover, when combined with ICIs they may enhance exhausted T cell rejuvenation (40). Early-stage clinical trials suggest that the combination of ICIs and epigenetic agents is well-tolerated; however, further testing is required to prove greater efficacy of the combinatorial therapy as compared to ICIs alone (39,41).

HLA class-I APM component downregulation by interferon signaling defects

IFN signaling plays an important role in HLA class-I trimolecular complex synthesis by upregulating genes such as TAP and ERAP (42). IFN stimulation activates Janus kinases (JAKs) and signal transducer and activator of transcription 1 (STAT1), inducing the expression of IRF1 and NLRC5, which promote transcription of HLA class-I APM genes (43). Defective IFN signaling, which affects HLA class-I APM component expression and consequently HLA class-I trimolecular complex expression, has been described in many cancer types (44,45). Embryonic transcription factor double homeobox 4 (DUX4) upregulation decreases JAK1/JAK2 and STAT1 signaling, reducing constitutive and IFNγ-inducible HLA class-I APM gene expression in DUX4+ cancer cells (46). Additionally, Src homology region 2 domain-containing phosphatase 2 (SHP2) depletion in HNSCC cells activates JAK1/JAK2 and STAT1, and upregulates HLA class-I expression (47). Lastly, hypoxia inhibits STAT1 and IFN-inducible gene expression in renal cell carcinoma cells. Indeed, von Hippel-Lindau (VHL) mutations, which are known to stabilize hypoxia-inducible factor 1 (HIF-1), upregulate the transcriptional repressor stimulated by retinoic acid 13 (STRA13), which inhibits STAT1 transcription. Thus, VHL mutations are associated with HLA class-I APM component downregulation (48). Defects in HLA class-I APM expression caused by this mechanism can be potentially corrected with chemotherapy or radiotherapy, which exert their antitumor activity not only through cytotoxicity, but also through induction or enhancement of a TA-specific immune response (49,50). Both chemotherapy and radiotherapy increase IFN signaling in several cancer types. The resulting HLA class-I upregulation can improve the recognition and elimination of cancer cells by cognate T cells in pre-clinical models (51,52). Several ongoing clinical trials are combining chemo- and/or radiotherapy with ICI or adoptive T cell therapy (49,53). More recently, stimulator of interferon gene (STING) agonists, promoting type I IFN secretion, have emerged as another potential strategy to correct defects in IFN-inducible genes (54).

HLA class-I APM component downregulation by NF-κB signaling defects

NF-κB signaling can be activated by the binding of ligands to the tumor-necrosis factor receptor superfamily (TNFRSF) and to Toll-like receptors (TLRs) via canonical and non-canonical pathways (44,55). Members of the NF-κB family of transcription factors (Canonical: NF-κB1 p50, RELA, and c-REL; Non-canonical: NF-κB2 p52, and RELB), can migrate from the cytosol to the nucleus following NF-κB activation and bind to the Enhancer A region of HLA class-I APM component encoding genes. Abnormalities in downstream signaling of the NF-κB pathway cause HLA class-I downregulation in neuroblastoma (56,57). NEDD4 Binding Protein 1 (N4BP1) and TNFα-induced protein 3 interacting protein 1 (TNIP1) inhibit NF-κB-mediated HLA class-I upregulation (57). Their depletion upregulates HLA class-I expression and increases neuroblastoma cell susceptibility to recognition and elimination by cognate TA-specific CD8+ T cells. Furthermore, transforming growth factor β (TGFβ)-mediated HLA class-I downregulation has been described in prostate cancer (58). This phenotype is likely to be mediated by the upregulation of the transcription factor Snail (SNAI1), which downregulates HLA class-I by targeting NF-κB/p65.

HLA class-I APM component downregulation by oncogenes

Various oncogenes may contribute to MHC class-I downregulation including receptor tyrosine kinase (RTK)/mitogen-activated protein kinase (MAPK) pathway components, cyclin-dependent kinase 4/6 (CDK4/6), and c-Myc. The effect of the former is most prominent in cases of activating mutations or amplifications in RTK/MAPK pathway genes (epidermal growth factor receptor [EGFR], KRAS, BRAF, SHP2, and MEK) (59). KRAS mutations are associated with HLA class-I downregulation in NSCLC tumors (60). MAPK inhibition enhances HLA class-I expression on several types of cancer cell lines including breast, esophagus, lung, stomach, liver and pancreas (61,62). This effect is mediated by STAT1 and STAT3 activation by the MAPK inhibitor (59,63). Anaplastic lymphoma kinase (ALK) and RET, two RTKs that are associated with MAPK and phosphoinositide 3-kinase (PI3K) pathway activation, are mutated in various cancer types (64).Inhibition of these molecules can suppress MAPK signaling and upregulate HLA class-I expression (64). Human epidermal growth factor receptor 2 (HER2/neu) overexpression is also associated with HLA class-I downregulation due to impaired LMP2, TAP1, TAP2, and tapasin transcription (26). These defects and the resulting decrease in cancer cell susceptibility to cytotoxic T cell-mediated lysis can be corrected in HER2/neu+ cells by IFNγ treatment (65). Similarly, HER2 downregulation induced by a small interfering RNA, or MAPK pathway inhibition can upregulate HLA class-I expression (66,67). Targeting RTKs with mAbs can also inhibit MAPK signaling. The anti-EGFR mAb cetuximab upregulates HLA class-I expression by head and neck cancer cells in patients’ tumors displaying a post-treatment volume reduction (68). Moreover, CDK4/6 are involved in MHC class-I downregulation in breast cancer cells. Their inhibition stimulates the production of type III interferons and induces STAT1 phosphorylation by downregulating DNMT1. Consequently, MHC class I APM components are upregulated, and a synergy is observed with anti-PD-L1 mAbs in transgenic mouse tumor models (69). Lastly, c-Myc transfection downregulates HLA class-I subunits at the mRNA level in melanoma cells (70) likely by binding to the initiator (Inr) element of their promoters (71).

HLA class-I downregulation by lysosomal degradation

HLA class-I expression is downregulated in a large percentage of pancreatic ductal adenocarcinoma (PDAC) tumors (72). This abnormality is caused by the autophagy-lysosomal degradation pathway (73), a result which corroborates the previously described MHC class-I antigen upregulation by autophagy inhibition in a murine melanoma cell line (74). The autophagy cargo receptor NBR1 associates with HLA class-I trimolecular complexes mediating their transport to lysosomes where they are degraded (73). Inhibition of lysosomal degradation by chloroquine restored MHC class-I antigen expression on murine PDAC models and enhanced the efficacy of T cell-based immunity triggered by ICIs. The same restoration is likely to occur in humans; hydroxychloroquine, administered with the rationale of inhibiting autophagy-mediated cancer cell survival, in combination with chemotherapy, resulted in increased immune cell infiltration compared to chemotherapy alone in a Phase II trial in PDAC patients (75). Whether this effect was mediated through HLA class-I upregulation remains to be proven.

The enzyme proprotein convertase subtilisin/kexin type 9 (PCSK9), which is involved in regulation of cholesterol levels by low-density lipoprotein (LDL) receptor lysosomal trafficking, has also been reported to regulate HLA class-I levels (76,77). Specifically, PCSK9 is secreted by malignant and non-malignant cells, binds to the extracellular alpha-1 region of HLA class-I heavy chain and mediates trimolecular complex degradation via the endosomal-lysosomal pathway. PCSK9 knockdown or inhibition of its function with blocking mAbs restored MHC class-I expression and improved responses to ICI-based therapy in multiple syngeneic, immunocompetent mouse tumor models (76,77).

HLA class-I APM component regulation by microRNAs

Several microRNAs (miRNAs) binding to the 3’-untranslated regions (3’UTRs) of HLA class-I APM components regulate their expression. Despite limited investigation of miRNA-mediated HLA class-I heavy chain regulation due to low DNA sequence homology of these genes with their orthologs, HLA-C is negatively regulated by miR-148a. However, the binding affinity is highly dependent on the germline HLA-C 3’UTR variant (78). Additionally, TAP1 is negatively regulated by miR-26b-5p and miR-21–3p in melanoma cells. Overexpression of these miRNAs downregulates both TAP1 and HLA class-I trimolecular complex expression (79). TAP2, instead, is negatively regulated by miRNA-125a in esophageal adenocarcinoma cells. These cells display concomitant TAP2 and HLA class-I trimolecular complex downregulation upon transfection of a miRNA-125a mimic (80). Other APM components are also negatively regulated by miRNAs, including PSMB8 (LMP7) targeted by miR-451a (81), PDIA3 (ERp57) targeted by miR-148a and miR-330–5p (82,83), (CANX) (calnexin) targeted by miR-711 (84), and (CALR) calreticulin targeted by miR-455 and miR-27a (85,86). Furthermore, NLRC5 is negatively regulated by miR-34a in HPV16-positive cervical cancer cells (87). This activity is likely to influence HLA class-I expression as NLRC5 exclusively transactivates HLA class-I APM component genes (88) and miR-34a inhibition enhances MHC class-I antigen expression in mouse neuroblastoma cells (89).

HLA class-I APM component modulation by cytokines

In addition to IFNs and TGFβ, other cytokines modulate HLA class-I expression. For instance, cytokines inducing NF-κB activation such as tumor necrosis factor α (TNFα) enhance HLA class-I expression on cancer cells in an allele specific manner due to the differential ability of NF-κB to bind to κB sites among HLA class-I heavy chain loci promoter regions (90,91). Furthermore, IL-2, IL-12, and IL-27 enhance HLA class-I expression by papillary thyroid cancer, melanoma, and small cell lung cancer cells, respectively (92–94). While the underlying mechanism is not characterized for IL-2 or IL-12, STAT1 or STAT3 phosphorylation may regulate IL-27-mediated HLA class-I upregulation, which is accompanied by enhanced TAP1 and TAP2 expression. The role of IL-10 in HLA class-I modulation remains controversial; downregulation in monocytes and upregulation in papillary thyroid cancer cells have been reported (95,96).

HLA class-I APM component modulation by the tumor microenvironment

Abnormal conditions within the tumor microenvironment (TME) are also believed to influence HLA class-I APM component expression by cancer cells (97). Inadequate angiogenesis results in low blood supply to rapidly proliferating malignant cells in the TME, decreasing oxygen levels and causing hypoxia (98). In this environment, hypoxia-inducible factors (HIFs) are not degraded and can relocate to the nucleus to drive transcriptional modulation of HLA class-I APM components (99). Similarly, adenosine triphosphate (ATP) levels are significantly increased at tumor sites due to the glycolytic metabolism of cancer cells (100,101). As the TAP complex is regulated by ATP hydrolysis in the cytoplasm (102), its function is likely to be influenced.

Concluding Remarks

Experimental evidence convincingly indicates that non-structural mechanisms are the major cause of HLA class-I APM downregulation, which is frequently present in most cancer types. These abnormalities have clinical relevance, as shown by their association with the clinical course of the disease and with response to T cell-based immunotherapies, including ICI therapy, which are at the forefront of cancer treatment (103–106). Nevertheless, the role of HLA class-I APM defects in resistance to T cell-based immunotherapies (107) and in disease recurrence following a clinical response (108,109) has often been overlooked, likely because these defects are assumed to be rare, given the low frequency of structural mutations in HLA class-I APM component encoding genes (8). As a result, the value of HLA class-I APM defects as biomarkers to identify patients who may benefit from T cell-based immunotherapy has been explored to a limited extent, despite the reported association of HLA class-I APM component expression level with clinical response to ICI therapy in some cancer types (110,111) and its increased efficacy when combined with strategies that correct HLA class-I APM abnormalities in targeted cancer cells (30,31,34,38,51,52,73,76,77,98). Similarly, HLA class-I alleles expressed by targeted cancer cells, with few exceptions (112), have not been taken into account when assessing the immunogenicity and clinical relevance of neoantigens present in malignant cells, although they are crucial for their presentation to cognate T cells. Therefore, systematic studies to explore the value of HLA class-I APM component expression level as a biomarker on malignant cells, to identify potential responders to ICI-based therapy, and those who may require HLA class-I APM restoration to achieve a response, are warranted (see Outstanding Questions’ box).

Outstanding Questions.

• What is the frequency and impact of defects in HLA class-I APM components in various cancer types on HLA class-I trimolecular complex expression?

• How do defects in HLA class-I alleles impact response to immune checkpoint inhibitors?

• Does the evaluation of mutation load in malignant tumors in the context of HLA class-I alleles expressed by cancer cells and APM component expression by cancer cells improve its correlation with clinical response to ICI therapy?

• Can more quantitative diagnostic assays be developed to assess the extent of HLA class-I APM component downregulation in patients?

• What is the level of downregulation or loss of function of HLA class-I APM that has clinical relevance?

• What is the role of HLA class-I recycling components in the expression and functionality of HLA class-I molecules in malignant cells?

Several examples have been described in which tumors displaying HLA class-I APM component defects have recurred following a clinical response to T cell-based immunotherapy (108,109). Furthermore, the available evidence, although limited, suggests that the frequency of HLA class-I APM defects in metastases is increased in patients treated with T cell-based immunotherapy (27). It is likely that a cancer cell subpopulation can develop novel escape mechanism(s) under the selective pressure imposed by an immunotherapy. When the targeted restricting HLA class-I allele is lost, the host’s immune system can change the TA and the restricting HLA class-I allele (113). These clinical findings emphasize the need to develop strategies which maintain the expression and function of HLA class-I APM components to enhance recognition and elimination of cancer cells by cognate T cells. However, strategies which may have beneficial effects by upregulating HLA class-I expression, could have broad effects and therefore also upregulate molecules which have a negative impact upon immune effector cell-cancer cell interactions, such as radiotherapy-induced upregulation of programmed death ligand-1 (114,115).

From a practical viewpoint, analysis of HLA class-I APM expression and function will greatly benefit from standardization of the methodology used to quantitate expression levels in malignant and normal cells in formalin-fixed, paraffin-embedded tissues. The latter represent the standard to score the HLA class-I APM component expression level in malignant cells. It is also important to consider that HLA class-I APM component expression and source of defects will differ among patients even with the same cancer type, as well as between primary and metastatic tumors (116). Therefore, mechanisms and potential interventions may have to be individualized. Furthermore, mAbs that detect shared HLA class-I epitopes are not practical for determining selective allele loss (117); therefore, selection of a panel of mAbs to identify specific HLA class-I alleles is critical. Finally, HLA class-I APM component expression levels may not necessarily correspond to functionality of HLA class-I trimolecular complexes. Variance in the repertoire of TA-derived peptides presented by HLA class-I alleles, due to differential expression of immunoproteasome subunits, is associated with patient outcomes in various cancer types (118,119). Assessment of the efficiency of peptide presentation by malignant cells before and/or after HLA class-I APM expression restoration will help determine a patient’s likelihood of clinical response to T cell-based immunotherapy.

Highlights.

• HLA class-I APM component defects are present in most, if not all, cancer types at high frequency.

• The mechanisms underlying these defects are predominantly non-structural and therefore, are potentially correctable with rationally designed therapeutic strategies.

• Restoration or upregulation of HLA class-I APM component expression in malignant cells enhances the efficacy of T cell-mediated immunity, including that triggered by immune checkpoint inhibitors.

• HLA class-I APM component expression is a potential biomarker for the selection of patients who may benefit from immune checkpoint inhibitor-based therapy.