1. Immune Signature of Retrotransposons in Cancer
Most of the (above-mentioned) genome-wide studies were focused on identifying new insertions and characterizing their effect on tumor-modulating genes. There is also a growing interest in identifying the factors controlling retrotransposon RNA expression or the factors triggered by its activation, such as the emerging data demonstrating that retrotransposon activation can be immunogenic and may instigate IFN and apoptosis signaling
[1][2][3][4][5].
Tumors with high immune activity, such as those associated with the Epstein–Barr virus (EBV) infection, demonstrated a low number of L1 insertions
[1]. Reports also indicate high retrotransposon activity in head and neck squamous cell carcinoma (HNSCC) patients. The overexpression of retrotransposons in HNSCC was shown to be associated with robust DNA CpG demethylation of tumor tissue
[6]. A high expression of the long terminal repeat (LTR) retrotransposon HERVs in HNSCC cases was accompanied by high cytolytic effectors, which correlated positively with cytolytic immune activity
[7]. This activity could be related to the oncogenic human papillomavirus (HPV), whose infection is among the etiological factors contributing to a subset of HNSCC tumors. HPV-positive cases often present with better outcomes
[8] and are less likely to have
TP53 mutation
[9]. These tumors have also demonstrated less retrotransposon somatic insertions (i.e., activity)
[9]. The examples above suggest the involvement of a defense mechanism against retrotransposons resembling antiviral actions.
Many retrotransposon regulation mechanisms are similarly used to protect cells from exogenous viral infections. When nucleic acids of foreign origin are detected by endosomal or pattern recognition receptors (PRR), an IFN-driven immune response is initiated to eliminate the affected cell populations
[10]. The cell is equipped with a heterogeneous group of PRRs that includes but is not limited to Toll-like receptors (TLR3, TLR7, TLR8, and TLR9); the RNA sensors RIG-I (retinoic acid-inducible gene I), MDA5 (melanoma differentiation-associated protein 5), and LGP2 (RIG-I-like receptor LGP2); and the DNA sensors cGAS (cyclic GMP-AMP synthase) and AIM2 (absent in melanoma 2)
[11].
Specific criteria, including location, nucleic acid sequence pattern, and threshold quantity, determine which nucleic acid each PRR senses
[11]. TLR nucleic acid binding domains face the lumen of endosomal compartments, and the other PRRs are present in the cytoplasm
[11]. TLR3 binds dsRNA of >40 bp size; TLR7/8 bind fragmented RNA with unmodified nucleosides; and TLR9 binds ssDNA of >11 nt size with a high affinity to the unmethylated cytosine CpG motif
[11]. RIG-I binds >20 bp dsRNA with blunt end conformation; MDA5 binds >1–2 Kb dsRNA; cGAS binds dsDNA of >20–40 bp size; and AIM2 binds dsDNA of >50–80 bp size
[11]. The quantity of detected nucleic acid can be affected by the increased supply that causes the accumulation of nucleic acids and the defective mechanisms of their clearance.
The failure of one or more of the (above-described) retrotransposon regulatory mechanisms (due to aging, tumorigenesis, or autoimmune disease) can result in retrotransposon activation. This activity promotes dsRNA or dsDNA (sequences of different sizes and motifs) release into the cytoplasm and their detection by cGAS or MDA5, respectively
[4][5][12]. Most ADAR-mediated A-to-I RNA editing sites are found in close proximity to retrotransposons. Upon the depletion of ADAR1 in conditions such as Aicardi–Goutières syndrome and some cancers, unedited endogenous RNAs trigger a chronic type I IFN response via MDA5 facilitated by the LGP2 RNA sensor
[13]. The activation of L1 during cellular senescence triggered the release of L1 dsDNA in the cytoplasm and promoted type I IFN responses and sterile inflammation
[3].
In addition to the evidence summarized in
Table 1 below, many examples suggest the retrotransposon activation of innate immune response in cancer. By analyzing TCGA RNA sequencing data, specific HERV elements were highly enriched in tumor samples compared to their normal counterparts, and this enrichment was associated with an increased immune response
[7]. Another piece of evidence showed that cytosolic ssDNA and dsDNA in several tumor cell lines were mainly retrotransposon-derived and associated with the cGAS-activated STING and type I IFN response
[12]. Activating HERV expression using DNMT inhibitors (DNMTi) in cancer cells triggered cytosolic dsRNA release, and MDA5 stimulated immune response
[5]. In addition, expressing ERV sequences in TLR3, TLR7, and TLR9 triple-deficient mice failed to induce a sufficient immune response, resulting in their development of T-cell acute lymphoblastic leukemia and their early death
[14]. Blood samples from individuals with the autoimmune disease SLE, systemic lupus erythematosus, were enriched in
Alu RNA associated with high levels of type I IFN response
[15]. Although the triggers of retrotransposon activation in the disorders mentioned above may differ, their induction of TEs is likely to be the cause of the IFN responses as a means of protection. A feedback loop may be generated to inhibit L1 activity, as suggested by specific interferon-stimulated proteins directly interacting with its encoded ORF1p
[16].
Table 1. Retrotransposon activity and associated immune response in cancer.
Tumor-specific characteristics may alter the tumor microenvironment and play a role in retrotransposon expression and its associated immune response. TP53, for example, has immunomodulatory roles, and its dysfunction associates with immunosuppression
[7][28], which is consistent with the evidence of gastrointestinal tumors with
TP53 mutations showing low immune activity and higher loads of L1 insertions than tumors with wild-type
TP53 [1]. Also, evidence from colon cancer shows that in response to viral infection in cells, TP53 induces an IFN-dependent antiviral response by activating IFN-stimulated genes
[29]. Another piece of evidence showed that
TP53 cooperates with DNA methylation to maintain the silencing of SINEs and other non-coding RNAs
[30]. The
TP53-deficient cells in this study exhibited high SINE element expression accompanied by a high type I IFN response
[30]. However, not all tumors exhibit the same type of
TP53 mutation, and not all mutations result in TP53 protein deficiency
[31].
TP53 mutation can contribute to tumorigenesis by losing TP53 function and gaining mutant functions
[31]. Whereas frequent
TP53 loss of function mutations in basal-like breast cancer could increase retrotransposon expression and the associated IFN response, the
TP53 gain of function mutations in high-grade serous ovarian tumors could reduce retrotransposon expression and its associated IFN response
[32].
Apart from
TP53, gastrointestinal tumors had strong associations between retrotransposons and TLRs or IFN-induced mRNAs, which was not the case in breast and ovarian cancers
[1][32]. Also, IFNε, which is hormonally regulated and expressed in the cells of reproductive organs
[33], presented high associations with retrotransposon expression in breast and ovarian cancers
[32]. It could be that because retrotransposons contain several binding motifs for estrogen response elements (ERE)
[34], they may play a role in IFNε expression in the tumors of reproductive organs. Therefore, the effect of IFN on retrotransposons could be related to the hormone-regulated microenvironment and might be tumor type-specific. The context-dependent IFN signaling associated with the ER+ and ER-negative breast cancer subtypes, which impacts their response to therapy and overall outcomes, reinforces the above notion
[35]. It could be interesting to extend these experiments to identify the levels of retrotransposon expression among ER+ and ER- breast tumors. The examples above support the assertion that the tumor type and specific characteristics could affect the retrotransposon’s expression and linked immune response. These variabilities should be considered when studying the retrotransposon’s activity in different types of cancer.
2. Therapeutic Opportunities for Retrotransposon Activity in Cancer
Throughout their evolutionary timeline, significant retrotransposon-related activities at the genomic and cellular levels have been attributed to their RT
[36]. However, retrotransposon genomic insertions in cancer have drawn considerable attention beyond the attention given to retrotransposon RT activity
[37]. RT activity was shown to increase during tumorigenesis. Anti-retroviral non-nucleoside reverse transcriptase inhibitors (NNRTIs), such as efavirenz and nevirapine, reduced RT activity significantly by inducing conformational changes in the enzyme
[38][39]. The NNRTIs reduced tumor growth by decreasing cellular proliferation and promoting differentiation
[40][41]. The effect of inhibiting RT using NNRTIs was similar to that of the L1 siRNA suppressing effect; therefore, they were assumed to target L1 activity
[42]. Other lines of evidence suggest that another class of RT inhibitors, nucleoside reverse transcriptase inhibitors (NRTIs), are capable of inhibiting L1 activity and having anticancer effects in cells
[43][44]. This evidence suggests that L1-encoded RT is a potential marker for diagnostic purposes and a potential target for therapeutic intervention. However, further work is still required to understand the exact mechanism of the observed effect of RT inhibitors on cancer
[45]. Although both NRTIs and NNRTIs could inhibit cancer cell growth, only NRTIs inhibited telomerase RT in vitro
[46], which may suggest a mechanism related to L1 RT particularly to affect cancer growth.
Among the mechanisms that activate retrotransposons, demethylating agents such as DNMTi act by releasing the epigenetic restriction placed on retrotransposons
[4][6][47]. Activating various TE classes in glioblastoma cells triggered type I and II IFN responses
[6]. TE-derived peptides were processed and presented on MHC class I molecules that activated adaptive immunity
[6]. Activation of HERVs resulted in a viral mimicry response of dsRNAs, inducing the MDA5/MAVS RNA recognition pathway and the downstream activation of interferon response factor 7 (IRF7)
[4]. Recent evidence (based on TCGA data analysis and in vitro DNMTi treatment of ovarian cancer cells) suggested that high HERV expression in patients was associated with better survival and correlated with the infiltration of cytotoxic T cells
[48]. The use of DNA-hypomethylating agent 5-azacitidine (AZA) in colon and ovarian cancer cell models was associated with the increased expression of HERV and L1 RNA
[5][49]. HERV expression was linked to regulatory T cell tumor infiltrates and predicted cytolytic activity in AZA-treated cells
[49].
In contrast, L1 expression correlated with TP53 status and predicted AZA drug sensitivity
[49]. A dinitroazetidine derivative (RRx-001), another hypomethylating drug less toxic than AZA, is currently in phase II clinical trials
[50]. RRx-001 induced antitumorigenic effects by activating the expression of HERV and IFN-responsive genes
[50]. Similarly, treating colon cancer cells and tumor organoids with another derivative of a hypomethylating agent (5-aza-2′-deoxycytidine) was sufficient to induce a growth-inhibiting immune response by triggering retrotransposon expression
[4][47]. Interestingly, the combination of DNMTi and HDACi selectively induced LTR retrotransposons more efficiently than using each drug individually
[51]. The treatment-activated TSS of LTR elements induced them de novo from non-annotated TSS
[51]. This activation resulted in chimeric products with predicted immunogenic functions
[51].
In addition, some targeted cancer therapeutics and chemotherapeutic agents were shown to activate retrotransposon expression in cancer cells
[2][52]. Cyclin-dependent kinases 4 and 6 (CDK4/6) inhibitors repressed DNMT1 and caused activation of repeat elements, including retrotransposons in breast cancer
[52]. This activation promoted cytotoxic T-cell-mediated clearance of tumor cells and increased tumor immunogenicity
[52]. However, some cells within a heterogeneous cancer population may develop adaptation mechanisms to survive the challenging tumor microenvironment conditions
[2]. These cells could modulate retrotransposon expression with lethal drug exposures by maintaining their epigenetic repression
[2]. This evidence suggests combining HDACi with other targeted therapeutics may enhance their efficacy in treating cancer
[10].
The examples mentioned above support the notion that retrotransposon activation in tumors may contribute to their turning into ‘hot tumors’, which are inflamed and T-cell- infiltrated tumors
[53]. In such a microenvironment, the antitumor immune response will reduce the tumor burden and sensitize it to other targeted therapies and immunotherapy
[53]. Retrotransposon activity in cancer probably occurs more in specific tumor types than in others
[54][55]. It is unclear whether this is related to a more vigorous immune defense or a higher level of cellular adaptation by implementing changes in their epigenome or transcriptome
[56].
Tumor-derived extracellular vesicles (EVs) are enriched in retrotransposon RNA and involved in the horizontal transfer of retrotransposons to normal cells. They may broadly influence the tumor microenvironment and immune response
[57][58]. This evidence suggests that EVs facilitate the release and transfer of retrotransposons to other cells, contributing to tumor evolution or metastasis (if derived from tumor cells). Also, retrotransposon RNA transfer can influence recipient cells’ transcriptional and post-transcriptional regulation. For example, the increased L1-derived RNA transcripts in recipient cells after the EVs transfer activate members of the APOBEC3
[58]. EVs are currently subject to multiple clinical trials at different phases and are to be used as non-invasive tools for diagnosis and therapeutics. They can serve as cargo for drug delivery in cancer and other conditions (as referred to
https://clinicaltrials.gov/, accessed on 18 June 2023). The increased expression of retrotransposons in EVs derived from tumor cells compared to those derived from normal cells
[57] could potentially serve as a valuable biomarker for diagnostic purposes. Studies to characterize the origin, biogenesis, and destination of EVs containing retrotransposon RNA and protein in cancer patients are currently needed to understand their potential fully.