Combination Therapy Approaches to Enhance the Efficacy of ERV-Targeting Vaccines in Cancer Open Access

Retroviruses are unique among RNA viruses as their replication requires integration into the genome of their host. Although infection and integration usually occur in somatic cells, on rare occasions they may occur in germline cells. In these instances, the provirus can be transmitted to subsequent generations through a process called vertical transmission. Endogenous retrovirus (ERV) sequences are thought to comprise 8% of the modern human genome (1). ERV proviruses retain the retrovirus structure and are composed of gag, pro, pol, and env genes flanked at either side by long terminal repeats (LTR) and can be classified into three classes (γ, β, and spuma-like) based on their relationship to other retroviruses (Fig. 1; ref. 5). The vast majority of ERV sequences integrated into the genome thousands of years ago and have accumulated mutations over the course of human evolution. These mutations result in the fragmentation and/or inactivation of the provirus, and despite their abundance, it is estimated that approximately 100 ERVs conserve an intact or nearly intact provirus that could theoretically produce intact viral particles. However, multiple studies have not been able to detect ERV-derived viruses in human samples (6, 7). Although mutated proviruses may not be capable of producing viral particles, several encode full-length proteins.

Expression of ERV-encoded proteins is tightly regulated in healthy tissues epigenetically (8). One of the hallmarks of cancer is epigenetic dysregulation (9), which may result in the aberrant expression of ERV-encoded proteins in multiple tumor types, with human ERV-K (HERV-K) being the most common (10). The immune system is not ignorant to this aberrant expression of ERVs in carcinomas. Several studies have demonstrated ERV-reactive T cells and ERV-specific antibodies in patients with breast, ovarian, and kidney carcinomas (10–15). This demonstration of immunogenicity as well as the high expression of many ERVs in carcinoma cells, as compared with adjacent healthy tissues (16), makes ERVs an attractive therapeutic target for cancer immunotherapy.

In this review, we summarize the current literature reporting promising preclinical studies targeting ERV sequences using various therapeutic modalities, such as antibodies, cellular therapies, and therapeutic cancer vaccines. Furthermore, we provide potential combination strategies of ERV-targeting agents with other immuno-oncology therapies such as epigenetic modifiers, immune checkpoint inhibitors (ICI), and cytokines. In addition to describing characteristics that make them targetable, we also highlight potential challenges to their clinical development.

The majority of ERV-encoded open reading frames (ORF) are not known to play any role in healthy tissues, and their expression is actively silenced (17–19). However, a small subset of ERV ORFs has been reported to play critical roles in normal cellular functions and processes, such as embryogenesis, immune regulation, and aging (19–23). The required function of ERVs in healthy tissues suggests the presence of a dynamic, selective regulation of expression. For example, human ERV sequences such as HERV-L, HERV-H, and HERV-K, among others, are dynamically expressed and regulated (activated or silenced) during embryogenesis, placentation, placental evolution (24, 25), and somatic development (Fig. 2). Aberrations in their regulation can lead to pathologic changes (28, 29). In addition, it was recently shown that ERV-derived ORF expression is thought to protect trophoblasts from infection by exogenous retroviruses (26, 30, 31). Overall, the majority of ERV sequences are not transcriptionally active in healthy tissues; however, some have been reported to play essential cellular functions.

In most cases, the silencing of ERV sequences in healthy adult tissues is thought to be achieved via promoter methylation. As global DNA hypomethylation is a common occurrence in most cancers, it is not surprising that many ERV-encoded proteins are aberrantly expressed in tumor tissues (32, 33). In support of this theory, Gimenez and colleagues (16) performed an HERV-targeted microarray assay using testicular cancer samples and found that the LTRs of six HERV-W loci that were upregulated in tumors had methylation levels ranging from 0% to 30% compared with 82% to 100% in healthy tumor-adjacent tissues. This reduction in methylation associates with enhanced expression of ERV sequences. In addition to promoter methylation, recent publications have highlighted the role of regulators of chromatin structure in the silencing of ERV sequences, such as SETDB1 and TRIM28. SETDB1 is a histone 3 lysine 9 (H3K9) methyltransferase that has been shown to silence ERV sequences by maintaining trimethylation of H3K9 and heterochromatin status (34). TRIM28 is a transcriptional corepressor that is recruited to ERV sequences in the genome and forms a complex with SETDB1 and other proteins to actively silence ERV transcript expression (35, 36). In mouse tumor models, loss of SETDB1 results in expression of ERV sequences, which induce ERV-derived type I IFN responses, which associates with enhanced lymphocyte infiltration, reduced tumor growth, and enhanced responsiveness to radiotherapy (37, 38). Conversely, the SETDB1–TRIM28 complex has been reported to suppress antitumor immunity (39). In a recent publication, Alcazer and colleagues (40) were not only able to distinguish between healthy hematopoietic and acute myeloid leukemia (AML) cells but also to classify AML into distinct subtypes with differing prognosis based upon their chromatin structure and ERV expression. These observations demonstrate the interdependence of chromatin architecture and ERV expression in cancers.

It has been reported that an average of 2,736 ERV sequences are expressed in each tumor cohort evaluated (pancreatic, liver, clear-cell kidney, ovarian, bladder, lung, and breast carcinomas), with kidney cancers having the highest number of expressed ERVs and liver cancers having the lowest number (bioRxiv.2024.02.07.579350). The authors found that the most expressed ERV families across these cohorts were HERVH, HERVL, ERVLE, ERV316A3, MER4, MER41, and ERVLB4.

In general, ERV sequences have been found to be highly expressed in hematologic malignancies (i.e., leukemias and lymphomas) as well as in prostate, ovarian, bladder, lung, colon, and liver cancers (41). In a recent review, Müller and colleagues (10) surveyed the literature and summarized ERV-derived proteins detected in samples of patients with cancer and/or cancer cell lines (Table 1). Among all ERV sequences, HERV-K is the most frequently expressed ERV across different cancer types, and it is expressed in germ cell tumors, neurologic cancers, melanoma, and prostate cancer, among others. After HERV-K, the other most common cancer-associated ERV types are HERV-H, HER-W, and HERV-R (42).

The CancerHERVdb database provides an overview of the literature reporting the expression of specific ERV sequences across different types of cancer, and it serves as a tool to identify ERV sequences that could potentially be targeted in cancer. To illustrate, this database shows that a high percentage of samples from patients with cancer express different ERVs, such as HER-V, HERV-K, and HERV-H LTR–associating protein 2 (HHLA-2), among others (summarized in Table 2; ref. 43).

It has become increasingly apparent that ERV-encoded proteins may play active roles in tumor progression. ERV-derived Env proteins have immunosuppressive properties and play important roles in cancer development, progression, and resistance to therapies (56). Expression of the ERV-encoded Env protein ERVH48-1 (suppressyn) has been shown to promote the proliferation of prostate cancer cells and resistance to doxorubicin (57). Furthermore, additional Env proteins such as those of ERV3-1, ERVW-1, and ERVFRD-1 have tumor suppressor roles, whereas ERVV-1, ERVK13-1, and ERVMER34-1 may promote oncogenesis in breast cancer (58).

Additionally, ERVs actively dampen the generation of effective antitumor immunity. For example, the ERV-derived immune checkpoint HHLA-2 is expressed in many types of human cancers such as melanoma, breast, colon, pancreatic, among others, and high HHLA-2 expression is associated with poor prognosis (59). HHLA-2 has immunosuppressive properties as it interacts with KIR3DL3 to inhibit CD8⁺ T-cell and NK-cell function and induce resistance to tumor-cell killing (60). Additionally, overexpression of ERVE-4 and HERV 4700 was observed in patients with clear cell renal cell carcinoma (ccRCC) who did not respond to anti–PD-1 therapy (61). Ng and colleagues (62) found that HERV-targeting B-cell responses can be amplified by ICIs in mouse models and patients with lung adenocarcinoma, and ERV expression can predict the outcome of ICIs in these patients. In contrast, it has also been observed that expression of other ERV sequences may associate with responsiveness to immunotherapy. In non–small cell lung cancer, patients who have high MER4 expression associate with better progression-free survival and overall survival as these patients also display inflammatory gene signatures and demonstrate improved responsiveness to ICIs (63).

Currently, researchers are evaluating the relationship between ERV sequence expression signatures (comprising multiple ERVs) and resistance or response to immunotherapy. Panda and colleagues (64) investigated the association between ERV expression in tumors and local immune checkpoint activation and responses to ICIs. They reported that tumors expressing high levels of ERV sequences had increased immune cell infiltration, upregulated checkpoint pathways, and higher CD8⁺ T-cell fractions compared with ccRCC tumors with low ERV expression. They observed similar trends in colon, neck squamous cell, and breast tumors. Furthermore, the authors discovered that ERV3-2 expression was significantly higher in anti–PD-1/PD-L1 responders compared with nonresponders in patients with metastatic ccRCC (mccRCC). Likewise, Zhou and colleagues (65) evaluated three clinical trials in which patients with ccRCC were treated with anti–PD-1 and found an ERV signature comprising nine ERVs that could stratify patients into high- or low-risk categories. Patients in the low ERV risk category had higher CD8⁺ T-cell infiltration, higher probability of survival, and better prognosis than patients in the high ERV risk category. Therefore, this ERV signature may serve as a predictive and prognostic biomarker for patients with advanced ccRCC who receive anti–PD-1 therapy. Lastly, another group identified a relationship between ERV abundance, immunogenicity, and epigenetic dysregulation in metastatic breast, colorectal, and pancreatic ductal adenocarcinoma tumors (66). The authors reported a positive correlation between tumors that express high levels of ERV sequences with gene signatures of immunogenicity and autonomous antiviral responses in the three tumor types evaluated. Specifically, they found significantly strong positive correlations between ERV sequence expression and genes associated with neutrophils, T cells, and IFN signaling. Interestingly, the authors also found that colorectal and pancreatic tumors that had a viral mimicry phenotype had increased expression of TET2, a DNA demethylation gene (67), thus suggesting that aberrations in DNA methylation may play an important role in the expression of ERV sequences and in gene signatures of immunogenicity and viral mimicry.

In summary, the high expression patterns of ERV sequences in numerous cancer types compared with healthy adult tissues, in addition to the critical roles that ERVs play in cancer progression and resistance to conventional therapies and immunotherapies, make ERVs potential targets for cancer immunotherapy.

Although most ERV proviral sequences have accumulated mutations preventing the production of viral particles, some encode full-length ORFs that encode retroviral proteins (68). In cancers, ERV-derived antigens can be generated via multiple mechanisms. ERV-derived immunogenic epitopes can be derived via canonical ERV ORFs, chimeric endogenous retroelement–exon tumor antigens (69), or noncanonical splicing junctions between exons and transposable elements (70). The latter two mechanisms generate aberrant ERV-derived proteins that are uniquely expressed in tumor cells and absent from healthy tissues. ERV-specific immunity, detected against ERV proteins expressed by each of the three above processes, has been identified in patients with cancer (summarized in Table 3; refs. 78, 79).

Multiple research groups have demonstrated that ERV-derived peptides are processed and presented by the immunopeptidome, and these presented ERV epitopes can expand antigen-specific T cells with cytotoxic activity and high avidity against tumor cells or organoids (12, 71, 72). In addition, several reports have shown that ERV-specific CD8⁺ T-cell responses directed against multiple ERV-derived proteins (i.e., gag, rec, env, and reverse transcriptase) have been detected in the blood of a large proportion of patients diagnosed with different types of cancer (78). In one such study, Saini and colleagues (80) found that half of the patients with hematologic malignancies who were included in the study had specific T cells against ERV peptides derived from HERVE-3, HERVH-5, and HERVW-1. In addition, patients with melanoma have HERV-K–specific T cells in circulation, and these are absent in healthy donors as HERV-K is a tumor-associated antigen that is expressed by melanoma cells but not normal cells (74, 81). Besides inducing a T-cell response, ERV-derived proteins may also induce humoral responses. As such, HERV-K immunoreactive antibodies have been found in the serum of patients with ovarian cancer (12), seminoma (76), germinal cell cancer (77), and breast cancer (15).

Although there are several studies evaluating the presence of ERV-specific T cells and demonstrating their ability to kill cancer cells, relatively few studies have identified specific ERV-derived T-cell epitopes and characterized the T-cell clones generated (11, 13, 73–75, 82). One such study, performed by Bonaventura and colleagues (71), identified nominal CD8⁺ T-cell epitopes from conserved Gag and Pol HERV-K motifs in patients with triple-negative breast cancer and found that 7 of 11 patients had T-cell receptors (TCR) with specificities for these HERV-K epitopes.

It has been well established that one mechanism by which retroviruses promote their propagation is by actively dampening the immune system within infected hosts. Multiple studies have described the presence of a highly conserved immunosuppressive domain (ISD) within the retroviral Env protein. This ISD can suppress proliferation and activation of macrophages, lymphocytes, and monocytes (83–86). As such, tumor cells expressing envelope proteins containing the ISD may also utilize a similar mechanism to evade immune surveillance (Fig. 3). A study by Mangeney and colleagues (87) demonstrated that expression of the full-length HERV-H Env protein in a murine carcinoma cell line enabled tumor growth when cells were implanted in either allogeneic or syngeneic immunocompetent mice. Other examples of ERV Env proteins containing immunosuppressive regions include HERV-P(b) and HERV-V Env proteins (41). Thus, the initial challenge of ERV-targeting therapies resides in that they must generate an immune response strong enough to surmount the immunosuppression of the ISD of ERV Env proteins.

It is well known that several ERVs, such as HERV-K, are expressed in the pluripotent human stem cells in the placenta and may also be present in other healthy adult tissues (88–92). Burn and colleagues (93) characterized the expression of HERV-K (HML-2) provirus in healthy tissues sampled at autopsy using RNA sequencing datasets from the Genotype Tissue and Expression project. Following assessment of 54 different tissues corresponding to 948 donors, HERV-K provirus transcripts were found in every tissue evaluated, with higher levels being detected in the cerebellum, pituitary, testis, and thyroid (93). HERV-K–encoded rec and np9 transcripts have also been detected in normal human tissues, such as the heart, brain, pancreas, and spleen, among others (94). However, these reports are balanced by negative protein staining in healthy adult tissues (95). Sacha and colleagues (95) also assessed this concern directly, by vaccinating rhesus macaques with consensus sequences of simian ERV-K Gag and Env and demonstrated that vaccination induced T-cell responses without vaccine-related pathologies. In humans, HERV-K‒specific T cells can be detected in the peripheral blood of patients with cancer without adverse effects (11, 12). Other studies have reported the presence of autoantibodies to HERV-K in patients with autoimmune and degenerative diseases such as systemic lupus erythematosus, rheumatoid arthritis, and Sjörgen syndrome, among other diseases (96, 97). However, there is a lack of consistency in autoimmune manifestations of autoantibodies targeting HERV-K in different diseases. Furthermore, autoantibodies targeting HERV-K have also been observed occasionally in healthy individuals (96).

Recently, a phase I clinical trial in patients with mccRCC assessed off-target toxicities following treatment with an HERV-E–directed therapy. In the study, patients were treated with adoptively transferred HERV-E TCR-transduced T cells without dose-limiting toxicities, off-target toxicities, or treatment-related deaths being reported at the end of the study (98). This lack of toxicity may depend on the ERV being targeted; thus, careful selection of the ERV target is key to ensure low or no expression in normal tissues and elevated expression in tumors to reduce off-target effects during treatment.

ERV Env proteins are expressed on the cell surface, so one could potentially target them using specific antibodies. ERV-specific antibodies could neutralize ERV proteins and block their interaction with other proteins and mediate antibody-dependent cell cytotoxicity by NK cells, complement activation, or antibody-dependent cell phagocytosis by macrophages. Wang-Johanning and colleagues (99) developed mAbs against the HERV-K Env protein, which they found to be expressed in breast cancer cell lines and patient samples. They showed that treatment of breast cancer cells with these mAbs led to growth inhibition and apoptosis in vitro. Importantly, treatment with these mAbs significantly reduced tumor growth in xenograft models in vivo. Other groups have also shown that targeting the murine ERV envelope protein with mAbs reduces tumor growth, with some antibody clones curing mice and significantly enhancing survival of myeloid leukemia–bearing mice (100).

Adoptive transfer of T cells engineered to express either chimeric antigen receptor (CAR) or TCR is an active area of investigation within the field of immuno-oncology. In preclinical models, Krishnamurthy and colleagues (81) engineered T cells to express a CAR that is derived from a mAb that targets HERV-K Env. These HERV-K Env–specific CAR-T cells were shown to have cytotoxic activity, recognize shed HERV-K Env, and reduce tumor growth of an HERV-K Env⁺ metastatic melanoma model. In addition, the same group tested the ability of HERV-K Env–specific CAR-T cells to inhibit breast cancer growth and metastasis in xenograft models (101). Only a handful of clinical studies targeting ERV⁺ cancers through cell-based therapies exist. Results from the first clinical trial evaluating TCR-engineered T cells targeting an ERV in patients with mccRCC (NCT03354390) showed that it is feasible to manufacture CD8⁺ HERV-E TCR-transduced T cells. These cells did not show any dose-limiting toxicities or off-target toxicities, proliferated in vivo, trafficked to metastatic sites, and resulted in partial responses in 7% of patients and stable disease in 29% of patients (98, 102).

Several preclinical studies have tested ERV-targeted vaccines in various cancer models (Table 4). Neukirch and colleagues (104) targeted the envelope protein of melanoma-associated retrovirus, composed of p15E and gp70 subunits, by using vector-encoding virus-like particles. Vaccination increased secretion of IFNγ and TNFα by CD8⁺ T cells and reduced tumor progression in mice bearing CT26 murine colorectal tumors. Mice were also protected against rechallenge with 4T1 mammary tumors, which also express the target antigen gp70 (108). In a separate study, melanoma-associated retrovirus was targeted using virus-like particles encoding the envelope with a mutated ISD (Env ISDmut) as a surface target to improve the immunogenicity and efficacy of the vaccine (105). Vaccination resulted in increased multifunctional CD8⁺ T-cell responses and greater tumor control when compared with the vaccine targeting the nonmutated Env in the syngeneic CT26 murine colorectal tumor model. Additionally, combination with anti–PD-1 also showed higher curative efficacy in Env ISDmut‒vaccinated mice.

Another study developed a customized oncolytic vaccine, ERV-PeptiCRAd, targeting ERV peptides previously identified through an MHC-I ligandome analysis in the 4T1 murine mammary tumor model (106). The PeptiCRAd vaccine platform consists of tumor peptides that are adsorbed onto an oncolytic viral capsid. Following intratumoral administration, the ERV-PeptiCRAd showed antitumor efficacy by reducing growth of 4T1 mammary tumors in mice. Interestingly, combination with anti–PD-1 did not increase the level of protection in this tumor model (106).

Other studies have focused on developing vaccines targeting human ERVs such as HERV-K and HERV-W instead of murine retroviruses (84, 103, 107, 109). Kraus and colleagues (84, 103) tested vaccines targeting HERV-K Env and HERV-K Gag proteins in murine renal carcinoma cells genetically modified to express these targets. Both vaccines used the recombinant modified vaccinia virus Ankara as vaccine platforms. The HERV-K Env vaccine was able to reduce pulmonary metastasis in vivo and increase cytotoxicity against tumor cells. Moreover, prophylactic vaccination was able to protect mice against tumor development. Similarly, vaccination against HERV-K Gag also decreased tumor growth and pulmonary metastasis in vivo (103).

Recently, Skandorff and colleagues (107) targeted HERV-W (syncytin-1) by evaluating vaccines encoding either the nonmutated or mutated envelope ISD in a murine renal carcinoma model genetically modified to express HERV-W Env. The ISD of HERV-W is atypical due to its lack of immunosuppressive function that can be reversed to the immunosuppressive role following mutation (110). In this study, the antitumor effects of the wild-type HERV-W vaccine outperformed those observed with the ISD-mutated counterpart in this scenario. Vaccination with the wild-type HERV-W vaccine resulted in increased specific T-cell responses, enhanced activation of antigen-presenting cells, and augmented survival.

Together, these preclinical studies using ERV-targeting vaccines in tumor models have demonstrated immunogenicity of the targets and increased tumor control. Nonetheless, additional studies are needed to translate these preclinical findings into the clinical setting as these models are genetically engineered to express the target HERV and further improve antitumor efficacy.

One approach that could be used to further increase the efficacy of ERV-targeting vaccines in cancer models is to develop a combination strategy that adds other immuno-oncology agents as part of the treatment regimen. These combinations could potentially contribute to the antitumor immune response by increasing, expanding, and activating vaccine-reactive T cells in the tumor microenvironment (TME) and by opposing immunosuppressive mechanisms such as the presence of regulatory T cells and upregulation of immune checkpoints. Here, we discuss immuno-oncology agents that could potentially enhance the immune responses generated by ERV-targeting vaccines against ERV-expressing tumors (Fig. 4).

One of the mechanisms by which tumors avoid immune surveillance is by upregulating inhibitory pathways such as the PD-1/PD-L1 axis. PD-1 and CTLA-4 are immune checkpoints expressed on T cells that, following binding to their cognate ligands, hinder the cytotoxic function of T cells (111–113). Previous studies have demonstrated that ICIs, such as those targeting PD-1/PD-L1 or CTL4, can expand the amount of neoepitope-specific CD8⁺ T cells infiltrating the TME and decrease the negative regulation of T-cell activation. Additional studies have demonstrated enhanced tumor regression, prolonged survival, increased cytotoxic T-cell responses, and protection from tumor rechallenge following combination with ICIs (114, 115).

As shown in Table 4, so far, most of the ERV-targeting vaccines studied preclinically have been tested in combination with anti–PD-1 (104–106). The addition of PD-1 blockade has generally enhanced the antitumor immunity of ERV vaccines, with some reports of tumor eradication following treatment (105, 109). Therefore, the combination of ERV-targeting vaccines plus ICIs continues to be a promising strategy, the efficacy of which should further be evaluated in clinical studies.

Therapeutic ERV cancer vaccines may offer a promising avenue to accomplish tumor regression in tumors with a high ERV expression; however, they may offer limited clinical benefit when other mechanisms of immune failure are also at play. Cytokines and chemokines within the TME regulate immune responses and cellular processes such as proliferation, apoptosis, and differentiation, all of which impact the efficacy of cancer therapies including immunotherapy (116). Some of the cytokines with reported protumor activity include TGF-β, IL-6, CXCL8, VEGF, and colony stimulating factor-1. Strategies for neutralizing these cytokines within the TME utilizing either antibodies or small-molecule inhibitors are currently being evaluated clinically. On the other hand, cytokines such as IFNγ, GM-CSF, IL-2, IL-12, and IL-15 have demonstrated antitumor effects in preclinical studies (117). These cytokines slow growth by inhibiting tumor cell proliferation and stimulating an antitumor immune response. Hence, a therapeutic strategy that could potentiate the immune responses of ERV-targeting vaccines in tumors could either block the production/function of immunosuppressive cytokines/chemokines or administer immunostimulatory cytokines, such as NHS–IL-12 or N803, as combination therapy targeting ERV proteins (118, 119).

NHS–IL-12 is a fusion protein consisting of IL-12 bound to an antibody that targets DNA in necrotic areas present in solid tumors (119). Necrosis supports the targeted delivery of IL-12 to the tumor and exposes intracellular antigens within the TME that could further be targeted with cancer vaccines. Mice treated with NHS–IL-2 have been found to have long-lasting lytic CD8⁺ T-cell responses specific to the endogenous retroviral protein p15E (120). It has been reported that the combination with NHS–IL-12 allows for vaccine-induced T cells to better infiltrate the tumor and remain active (120–122).

On the other hand, N803 (Anktiva) is an IL-15 superagonist recently approved by the FDA that binds to circulating immune cells through the IL-15 receptor, leading to the activation and expansion of NK cells and central memory T cells (118, 120, 123–125). Treatment of mice with a combination of N803 and therapeutic neoepitope cancer vaccine also promoted the expansion of p15E-specific T cells in the MC38 colorectal tumor model (120). The addition of NHS–IL-12 and anti–PD-L1 to the regimen further enhanced tumor regression and increased CD8⁺ T-cell infiltration and clonality of T cells in the TME (120).

Altogether, combining these immuno-oncology agents with therapeutic ERV cancer vaccines could potentially increase the immunogenicity of the vaccine to levels capable of rejecting the tumor. Although the combination of cytokines with ERV vaccines has yet to be evaluated preclinically, there is the potential to obtain enhanced antitumor immune responses based on previous studies evaluating the combined regimen of therapeutic cancer vaccines and cytokines at the preclinical and clinical levels.

It is well known that ERV expression is regulated epigenetically. Among the epigenetic processes that are relevant for controlling ERV expression are CpG methylation, histone deacetylation, and histone methylation. CpG-rich promoters tend to indicate the silencing of corresponding genes through methylation. There is evidence that hypomethylation of the genome leads to expression of ERVs in cancerous tissues (41). It has been reported that the use of DNA methyl transferase inhibitors (DNMTi), such as azacitidine (Aza), enhances the expression of specific transcripts of ERVs in several cancer types, such as ovarian cancer, endometrial cancer, melanoma, and neuroblastoma, by removing methylation from promoter regions of silenced ERVs (41, 79).

Prior studies have also demonstrated that DNMTis can induce “viral mimicry” by producing ERV transcripts that could activate innate immunity and stimulate viral defense response mechanisms (126). Chiappinelli and colleagues (79) show that treatment with Aza upregulates sense and antisense ERV transcripts, and that this bidirectional transcription results in double-stranded RNA (dsRNA) transcripts that trigger type I IFN responses and apoptosis in ovarian cancer cell lines. Furthermore, DNMTis can also promote the expression of antitumor cytokines such as IL-2, IFNγ, and Th1 chemokines like CXCL9, which could then enhance the infiltration of effector T cells to the TME (127). Hence, combination treatment with DNMTis may sensitize the response to ERV cancer vaccines by further increasing tumor cell visibility to immune surveillance, intensifying the antigen presentation process, enhancing the trafficking of effector T cells, and promoting cytolytic T-cell responses in the TME (128).

Another epigenetic mechanism that could regulate the expression of ERVs is histone deacetylation. Acetylation of lysine residues in histones is catalyzed by histone acetyltransferases and counteracted by histone deacetylases (HDAC). Nonetheless, previous studies have reported that the use of HDAC inhibitors (HDACi) as a monotherapy have failed to significantly increase ERV expression in humans (129). However, there is a possibility that histone deacetylation may act in combination with CpG methylation to regulate ERV expression; therefore, additional studies are needed to further understand this mechanism.

In addition to modulating ERV expression, immunogenic neoantigens derived from ERV elements can also be induced by DNMTis and HDACis (130). Targeting neoantigens represents a way to increase antitumor immune responses (131). Goyal and colleagues (130) recently identified 45 validated ERV-derived neoantigens that were presented by HLAs upon treatment with DNMTis and HDACis. This phenomenon was conserved across lung cancer, colon cancer, glioblastoma, and AML cell lines in vitro and in patients with cancer as the authors identified ERV-derived neoepitopes in patients with AML who were treated with the hypomethylating agent decitabine. More importantly, these neoantigens were able to activate T cells and elicit cytotoxic responses against tumor cells. This immune recognition could potentially be enhanced by addition of autophagy inhibitors as a combination strategy, which could inhibit the degradation of endogenous retroelement transcripts induced by Aza, increasing their capacity to generate MHC-I–associated peptides, as shown in a recent study in AML (132).

The aberrant expression of endogenous retroviral proteins within human carcinomas represents a class of tumor-associated antigens that can potentially be immunologically targeted in cancer. There are currently several diverse treatment modalities being evaluated in preclinical models to target ERVs, including the use of ERV-specific antibodies, ERV-specific T-cell transfer, and ERV-targeting therapeutic cancer vaccines. Combinations with other immuno-oncology agents such as ICIs, cytokines, and epigenetic regulators could further enhance antitumor immune responses by reducing immunosuppression in the TME, augmenting the antigen presentation process, promoting cytolytic T-cell responses, and increasing tumor cell visibility to immune surveillance. So far, most studies conducted are still in an early stage, with preclinical studies demonstrating preliminary antitumor efficacy in various tumor models. Nonetheless, careful selection of the target ERV is key to ensure minimal or no expression in normal tissues and reduce off-target effects during treatment. Altogether, mounting preclinical evidence strongly supports the continued development of ERV-targeting strategies as a potential treatment for tumors with high ERV expression.

No disclosures were reported.

The authors thank Debra Weingarten for her editorial assistance. This research was supported by the Intramural Research Program of the Center for Cancer Research, NCI, NIH.