RNaseH is considered the primary enzyme responsible for the direct degradation of various RNAs within cells and is also essential for maintaining cellular genome stability. There are two types of RNaseH in cells: RNase H1 and RNase H2. Both are capable of resolving the RNA strand in DNA-RNA hybrid structures by removing the RNA primers from Okazaki fragments, while RNase H2 also participates in ribonucleotide excision repair (RER) (Hyjek et al., 2019). Modulating the expression of cellular RNaseH to regulate the levels of R-loops has become a well-recognized fact (Camino et al., 2023). Reduced R-loop levels were observed through the detection of the S9.6 antibody in the knockdown assay of RNaseH (Drolet et al., 1995). The absence of RNaseH means that cells lose the ability to modify RNA, and the strong hydrogen bonds between nascent RNA and template DNA would tightly bind the two strands, which is a major contributor to R-loop formation. In fact, RNase H is a key enzyme that degrades the RNA moiety of R-loops, thereby preventing the accumulation of R-loops (Camino et al., 2023). Studies have shown that R-loop levels in RNase H1 knockout mice are elevated, leading to impaired mitochondrial function and liver degeneration (Lima et al., 2016). Furthermore, in cells where RNase H1 activity is absent, R-loop levels increase, resulting in DNA damage and replication fork stalling (Zhang W. et al., 2024). These findings highlight the crucial role of RNase H in maintaining genomic stability by resolving R-loops.

Although the mechanism of degrading R-loops in cells by RNaseH can achieve satisfactory results, it is a significant waste to hydrolyze the nascent RNA chain, especially the long-chain RNA. Many reports have revealed that numerous RNA-dependent ATPases are involved in the unwinding of RNA-DNA hybrid duplexes (He et al., 2022). This class of ATP-dependent helicases includes senataxin (SETX) (Suraweera et al., 2007), Aquarius (AQR) (Sollier et al., 2014), and DEAD-box helicase (DDX) family members such as DDX5, DDX41 (Shinriki and Matsui, 2022; Mersaoui et al., 2019). They are the second largest class of enzymes that remove R-loops after RNase H. SETX primarily removes the R-loop generated in the transcription termination region. AQR possesses 3′-5′ RNA helicase activity and single-stranded RNA binding activity, which can release the RNA strand through untwisting and facilitate mRNA splicing. DDX19 enables RNA binding activity and RNA helicase activity, and it participates in the mRNA export from the nucleus (Arul Nambi Rajan and Montpetit, 2021). In addition to above helicases, some novel DNA helicases have been continuously identified, such as DHX9, which has been found to be a driver for R-loop formation, as it releases the nascent RNA to produce a dissociative RNA terminal which can then combine with single-stranded DNA during the transcription process, leading to the formation of R-loops (Cristini et al., 2018). The phosphorylation of the DHX9 gene by ATR was required for the formation of R-loops (Liu et al., 2024). Evidence above indicates that DHX9 is an effective inducer of R-loops formation. In addition, DNA-RNA helicase AQR, SETX, and DHX9 could synergistically degrade the R-loops at the 3′-ends of a gene sequence. These three enzymes are important in transcriptional termination and the DNA damage resulting from the unstable abundance of R-loops (Figure 4B) (Cristini et al., 2018; Sakasai et al., 2017). The above three enzymes induce the untwisting of RNA-DNA hybrids, and the detached RNA strand is subsequently degraded by exonucleases such as XRN2. (Morales et al., 2016). The decrease in RNA abundance within the nucleus reduces the formation of DNA-RNA hybrid strands. Helicases degrade R-loops while retaining the original RNA, which is more conducive to the rational distribution of energy. These helicases are involved in the regulation of R-loops metabolism, and changes in their expression would also alter the abundance of R-loops.

In addition to acting directly on R-loops to affect their metabolism, cells also prevent production of R-loop by modulating chromatin structure. DNA topoisomerases regulate and alter the topological states of chromatin by catalyzing the transient breaking and reconnection of DNA strands, thereby ensuring the proper release of supercoiling and torsional tension on the DNA. Recent research on the HeLa cell has suggested that the knockdown of TopI increased the abundance of R-loops by inhibiting the supercoiling that occurs during the transcription process (Promonet et al., 2020). This is consistent with the previously described fact that the excessively supercoiled structure contributes to the formation of R-loops. The enzymes mentioned have different physiological characteristics, but they prevent the damage caused by R-loops by inhibiting their synthesis and promoting their degradation. Apart from working independently, these enzymes cooperate to maintain the R-loop content. For example, a sustained lack of TopI leads to replication stress and replication-transcription collision (RTC), but this could be eliminated by overexpression of RNase H (Manzo et al., 2018; Lang et al., 2017).

The function of R-loops varies greatly in different cell states, cell stages, and cell types. With the deepening of R-loops research, multiple functions of R-loops have been explored and fully elaborated. One typical example is the function of R-loops in class-switch recombination (CSR), which contributes to the diversity of antibodies. At the beginning of CSR in plasma cells, the R-loop is formed in the G-rich region of the IgH locus during the transcription process of antibody synthesis (Wiedemann et al., 2016). The DNA-RNA hybrid can free a long and stable ssDNA, enhancing the binding force of activation-induced deaminase (AID) to ssDNA (Yu and Lieber, 2003). The enzyme can deaminate cytosine residues on ssDNA to form uracil, which is then processed by base excision repair factors into DSB (Harris et al., 2002). After being repaired by non-homologous end-joining (NHEJ), intrachromosomal deletional recombination can occur to produce different kinds of immune proteins and improve immunity (Yu and Lieber, 2019; Stavnezer et al., 2008). In addition, there are indications that the R-loop is also implicated in immune signaling (Brickner et al., 2022). Because intracellular recognition of non-self and self nucleic acids can lead to the initiation of potent pro-inflammatory and anti-viral cytokine responses, the R-loops could be specifically recognized by cGAS and TLR9 thereby facilitating immune signaling, which has recently been found to promote the innate immune response in humans by activating IRF3 (Crossley et al., 2023; Mankan et al., 2014; Rigby et al., 2014).

The role of R-loops in DNA topology has also been gradually revealed (Mackay et al., 2020). As we know, in the process of DNA transcription and replication, in order to promote the efficiency of dsDNA unwinding, a negative superhelix is formed on the DNA strand and provides suitable conditions for the formation of R-loops (Roy et al., 2010; Massé and Drolet, 1999). In return, R-loops can be used as a superhelix stress eliminator to return the negative superhelix domain to a more energetically favorable (lower) state, relaxing the remaining DNA structural domains in part or completely (Stolz et al., 2019). In vitro experiments on plasmids have shown that the relaxation efficiency of R-loops is tens of times higher than that of nucleosomes of the same length (Stolz et al., 2019).

Gene expression is the cornerstone of biological activities, and studies have shown that the R-loop regulates gene expression in a variety of ways (Ginno et al., 2012; Colak et al., 2014). In epigenetics, the R loop can influence the methylation process of many genes. Since many R-loops are often found on unmethylated CGI, their presence shields the promoter from the action of DNA methyltransferases, thus facilitating the transcription of downstream genes (Chedin, 2016). For instance, DNA methyltransferase 1 (DNMT1), as a vital enzyme in the methylation process, has a strong affinity for double-strand DNA (Mohan, 2022). However, the RNA-DNA hybrid in R-loops makes it difficult for DNMT1 to bind to the promoter region of the gene (Galbraith and Snuderl, 2022; Grunseich et al., 2018). The nascent RNA has a space-occupying effect, impedes DNA methylation and regulates gene expression (Figure 5A). The R loop can also reduce methylation levels of CGI by attracting ten-eleven translocation (TET) DNA demethylases, and thousands of R-loop-dependent TET1 binding sites are present on CGI in human embryonic stem cells (Arab et al., 2019). To identify the existence of R-loops in CGI promoters, the DRIP method combined with deep sequencing was used and verified the universality of R-loops (Ginno et al., 2013). This was further investigated by Christopher et al. who, by employing DRIP and methyl C-seq techniques, revealed that in human cells the R-loop promotes transcription of more than 1,200 genes by preventing DNA methylation-directed gene silencing (Grunseich et al., 2018). However, there is a contrary view that R-loop binding to the DNA promoter region may occupy the binding sites of transcription factors in DNA (Boque-Sastre et al., 2015). This would disrupt the transcription process and accelerate the disorder of downstream RNA synthesis and gene expression, which has a huge impact on the cell cycle (Mackay et al., 2020). Besides the abundant binding in promoter regions, R-loops also gather in the G-rich terminator (Zhao et al., 2016). This assists RNA polymerase Ⅱ to stop on the template DNA strand and is conducive to transcriptional termination, resulting in precise regulation of gene expression, which has been verified in several gene transposons (Zhao et al., 2016; Skourti-Stathaki et al., 2011). One study has demonstrated that the recruitment of the helicase SETX by the R-loop at the 3′end of the gene promotes efficient transcription termination (Giannini and Porrua, 2024). The interaction between the R-loop and the G-quadruplex (G4) as secondary structures of the cellular nucleic acid chain is of equal interest. The R-loop formed in transcription promotes mRNA production by facilitating the formation of G4 in the non-template strand of DNA on the one hand (Lee et al., 2020), and on the other hand has been shown to recruit CCCTC-binding factors (CTCF) in mouse embryonic stem cells, where CTCF proteins would regulate gene expression by binding to target DNA sequences (Wulfridge et al., 2023; Dehingia et al., 2022).

Recently, R-loops critical role in DNA repair has been highlighted by increasing studies. Although excessive aggregation of R-loops in cells is a major cause of genomic instability, RNA-DNA hybrid was reported to promote homologous recombination (HR) or C-NHEJ between chromosomes, which is important for repairing DSBs in DNA (Refaat et al., 2023; Aguilera and Gómez-González, 2017). When R-loop levels are lower than normal, homologous recombination and NHEJ repair are less efficient (Ohle et al., 2016; Yasuhara et al., 2018). The occurrence of DSB is usually accompanied by the emergence of an R-loops structure. Given the loss of integrity of the DNA strand, the DNA repair requires NHEJ, and RNA polymerase II can be recruited through the MRE11-RAD50-NBS1 (MRN) complex for nascent RNA synthesis (Chang et al., 2019). This type of repair can cause DNA repair errors when the RNA strand in the R-loop can replace the damaged DNA strand and function as a template strand for DNA repair (Petermann et al., 2022). HR is the repair modality with the highest fidelity in DNA repair, especially in DSBs, and the R-loop acts mainly by facilitating HR (García-Muse and AguileraLoops, 2019; Xue and Greene, 2021). Moreover, the R-loop structure is also a robust recruitment platform for a variety of DNA damage repair factors (Figure 5B). The R-loop, as an essential molecule for efficient HR genesis, promotes the disappearance of DSBs through various pathways, although the general principle is similar: DSB induces the production of the R-loop, which in turn exerts its repair function by recruiting a series of DNA damage repair factors, such as RAD52 (Yasuhara et al., 2018), BRCA1 (D'Alessandro et al., 2018), RPA (Domingo-Prim et al., 2019), etc., and further enhances the loading of other repair factors, such as RAD51 (Wahba et al., 2013), BRCA2 (Teng et al., 2018), among which heterogeneous nuclear ribonucleoprotein U (HNRNPU) has been shown to interact with the R-loops to promote C-NHEJ-mediated DNA repair (Refaat et al., 2023). DNA repair factors and the corresponding hydrolases have been summarised (Table 1). The R-loops are subsequently hydrolysed by enzymes such as XPG (Yasuhara et al., 2018), SETX (Cohen et al., 2018a), RNase1 (Burger et al., 2019), RNase2 (Ohle et al., 2016) and dead-box helicase 1 (DDX1) (Li et al., 2016). Notably, some studies have found that Drosha enzymes are responsible for regulating the recruitment of repair factors to promote DNA repair. This suggests that Drosha enzymes may be related to the R-loops in some way, but further studies are needed (Camino et al., 2023; Liu et al., 2022).

Targeting the DSBs pathway in tumor cells has become an important tool for cancer treatment in precision medicine (Gillyard and Davis, 2021). Numerous studies have shown that DSBs and R-loops are often causally related. R-loops can be involved in the induction of DSBs or can be induced by DSBs (Promonet et al., 2020; Tan et al., 2020b; Arnould et al., 2023). R-loops primarily cause DSBs by stalling replication forks: the stable R-loop persists in the genome and ssDNA, the template strand, is encapsulated by the R-loop structure, which temporarily loses its ability to act as a template. During subsequent genome replication, DNA polymerase II is unable to bind to the template strand to attach free deoxy-ribonucleoside triphosphate (dNTP), and the DNA strand cannot be formed, resulting in a discontinuous ssDNA containing gaps, loss of the double helix structure, and easy breakage of the DNA under stress at the gaps (Hamperl et al., 2017). In addition, the R-loop has been found to disrupt DNA structure through non-replication dependence in some types of cells (Sollier and Cimprich, 2015; Cristini et al., 2019). DNA topoisomerase I (TOP1) enzymes can unwind DNA superhelix during transcription by forming topoisomerase 1-DNA cleavage complexes (TOP1cc) (Chowdhuri and Das, 2021). For instance, in fibroblasts, scientists first discovered that removal of TOP1cc produces single-strand breaks (SSBs) intermediate and that DSBs are ultimately generated by R-loops, which are then subjected to cleavage by the enzymes XPF, XPG and flap structure-specific endonuclease 1(FEN1). Similarly, depletion of TOP1 enzyme and RNase H enzyme in yeast cells was observed to enhance the results of R-loop-mediated non-dependent replication to increase DSBs (Stuckey et al., 2015). However, scientists observed diametrically opposite results in mammary epithelial cells, indicating that the induction of DSBs by the TOP1 enzyme and the R-loop is a very complex process that requires further study (Hidmi et al., 2024).

The pernicious effects of R-loops on genome stability in bacteria and humans by interfering with DNA replication and causing DSBs or chromosomal translocations have been demonstrated in both eukaryotic and prokaryotic cells (Gan et al., 2011). A widely accepted theory is that the dissociation of ssDNA in the R-loop makes the base which was originally inside the double helix structure free outside and is directly exposed to various harmful substances, such as multiple nucleases and genotoxic compounds, which are sensitive factors for DNA damage (Paulsen et al., 2009). Another fact is that the occupation of the R-loop in the genome hinders the progress of the replication fork during replication, which may lead to replication fork breakage or stalling (Stirling et al., 2012), and for a prolonged period this prevents replication from restarting, which is consistent with the instability of the genome (Gan et al., 2011). Meanwhile, the imbalance in R-loops leads to increased sensitivity of DNA to damaging factors, making it easier to induce DNA damage (Sarkar et al., 2018). An example of this is that in B cells overexpressing RNase H, the “collapsed R-loop” is susceptible to the AID enzyme, which deaminates dC to produce dU, more likely to cause DSB and recombination (Yu and Lieber, 2019; Yu et al., 2003). Another mechanism of DNA damage caused by R-loops may be mediated by base-excision repair (BER), such as the recruitment of nucleotide excision repair factors RAD1/XPG and RAD2/XPF by R-loop aggregation (Su and Freudenreich, 2017; Safari et al., 2021). In the process of removing R-loops, both of them will produce base deletions and DSBs in the replication process, leading to genome damage and cell genome instability (Mackay et al., 2020). All these are the main sources and strong evidence of genomic instability caused by the R-loop.

ALS is the most common neurodegenerative disease, characterized by the loss of motor neurons in the brain and spinal cord, resulting in muscle weakness or atrophy (Al-Chalabi and Hardiman, 2013). Several genes are involved in ALS pathogenesis, but R-loop involvement in ALS is not negligible (Cappella et al., 2019). The hexanucleotide GGGGCC (G4C2) amplification in the C9ORF72(C9) gene is the most prevalent genetic cause of ALS, and previous studies have shown that the C9 repeat amplification induces the DNA damage response (DDR) in ALS (Farg et al., 2017). The number of genomic CGI is significantly higher in some ALS patients than in normal subjects, and these duplicated CGI transcriptionally generate the R-loops (Perego et al., 2019; Reddy et al., 2014). Although the role of CGI in the formation of pathological R-loops is unclear, persistent R-loops certainly induce DNA damage in neurons affecting the stability of repetitive nucleotide sequences and even entire genomes (Reddy et al., 2014; Konopka and Atkin, 2018). This is supported by the finding of Manal et al. that in ALS, the expansion of the C9 repeat activates the DDR in neurons (Farg et al., 2017). Ultimately, this R-loop-dependent DNA damage leads to the degeneration and even the death of nerve cells in the brain.

ALS4 is a subgroup of ALS, which is characterized by the fact that it affects adolescents and progresses slowly (Grunseich et al., 2020). AOA2 is an autosomal recessive disorder in which patients often present with polyneuropathy and microcephaly (Anheim et al., 2009). The two diseases are caused by different mutations in the SETX gene (Moreira et al., 2004). SETX is a highly conserved RNA/DNA helicase that can function by mediating R-loop resolution (Mischo et al., 2011). When a dominant mutation of SETX occurs, the expression of helicase increases accompanied by an increase in the degradation of R-loops, resulting in ALS4 (172). The detailed mechanism is that the mutation of SETX decreases its expression, the abundance of R-loops in the promoter region of the BMP and activin membrane-bound inhibitor (BAMBI) gene decreases, the effect of preventing promoter methylation weakens and gene expression decreases (as mentioned above), while downregulation of BAMBI protein specifically activates the TGF-β pathway and finally contributes to the occurrence of ALS4 (Grunseich et al., 2018). TDP-43 is an RNA-binding protein whose depletion sensitizes neurons to DSBs (Mitra et al., 2019). According to clinical statistics, dominant mutations in the TDP-43 gene have a 4% and 1% incidence of familial and sporadic ALS, respectively (Sreedharan et al., 2008). In ALS4 patients, mutant TDP-43 was found to cause accumulation of R-loops through mislocalization or small interfering RNA (siRNA) depletion, with subsequent pathological R-loops exacerbating the disease phenotype by producing DSBs or promoting TRCs (Giannini et al., 2020). Notably, pathological changes in TDP-43 localization were caused by acquired mutations in SETX, indicating that SETX may contribute to the ALS disease pathway by indirectly modulating R-loop levels (Bennett et al., 2018). In contrast, when there is a recessive mutation in the SETX gene, helicase levels are reduced and a large number of body-damaging R-loops are produced in the cell, ultimately leading to AOA2 (Becherel et al., 2015). In addition, compared to ALS4, AOA2 cells also block SETX from interacting with the exosome subunit Rrp45, inhibiting exosomes from functioning (Richard and Manley, 2014; Richard et al., 2013). Normally, exosomes are responsible for breaking down the RNA produced by SETX de-localization (Houseley et al., 2006); therefore, the remaining RNA in AOA2 cells would still have the potential to exacerbate the disease phenotype.

SMA is a recessive genetic disorder caused by mutations in the motor neuron 1 (SMN1) gene on chromosome 5q13 and is characterized by muscle atrophy caused by degeneration of motor neurons in the spinal cord (Wirth, 2000). Mutations in the ZPR1 zinc finger (ZPR1) or SMN1 genes in the cellular genome were screened in SMA patient cells by deep sequencing. The proteins encoded by these two genes interact intracellularly to increase the efficiency of pre-mRNA splicing (Ito et al., 2018a). This can induce the differentiation of neurons and promote the growth of axons. Both gene defects contribute to the obstacle of pre-mRNA progression, produce R-loops, and damage the cell genome (Figure 6A) (Kannan et al., 2020; Kannan et al., 2022).

Viral infection can also cause malignant transformation of cells, which directly affects the integration of the viral genome into cell chromosomes and results in unregulated cell growth (D'Souza et al., 2020). Secondly, viral infection can result in the accumulation of R-loops in cells by inhibiting the expression of related proteins, leading to the malignant transformation of cells (Templeton and Laimins, 2023; Kumari et al., 2023). For example, Kaposi’s sarcoma-associated herpes virus (KSHV) is carcinogenic (Sandhu and Damania, 2022). The virus expresses the ORF57 protein, which is a regulatory factor of viral mRNA processing. ORF57 can tightly bind the export complex (hTREX) protein and inactivate it by masking the binding sequence, preventing it from participating in mRNA processing. This causes R-loops to accumulate in cells, leading to cancer (Figure 6B) (Jackson et al., 2014). In addition to KSHV, other viruses such as human papillomavirus (HPV), and Burkitt’s lymphoma also damage the genome directly or indirectly after infection (Templeton and Laimins, 2023; Kumari et al., 2023).

Hematopoietic stem and progenitor cells (HSPCs) have the capacity for long-term self-renewal and the potential to differentiate into various types of mature blood cells, the homeostasis of which is of great significance to the hematological system. The abundance of the R-loop has been found to affect the stability of HSPCs and has been implicated in acute myeloid leukemia (AML), myelodysplastic syndromes (MDS), adult T-cell leukemia and B-cell lymphoma (Frame and North, 2021; Weinreb et al., 2021; Lee et al., 2023; Chen et al., 2018; He et al., 2021). Among them, activation of DDX41-inflammation pathways is one of the important approaches to R-loop-induced disease progression. Dead-box helicase 41 (DDX41), a helicase, not only unwinds DNA-RNA hydrids but also activates the cGAMP synthase (cGAS)-Stimulator of Interferon Genes (STING) pathway in blood disorders (Winstone et al., 2024; Smith et al., 2023). Mutant DDX41 in myeloid cells results in the accumulation of pathological R-loops causing genomic instability and aberrant activation of the inflammatory response pathway (cGAS-STING), ultimately leading to AML or MDS (Mosler et al., 2021). In addition, aberrant expression of lncRNAs like HOTTIP (Luo et al., 2022), epigenetic like m6A methylation (Hwang et al., 2024), and transcriptional regulators like SRSF2 and U2 small nuclear RNA auxiliary factor 1 (U2AF1) can lead to abnormally elevated levels of R-loops and ultimately to blood disorders (Pellagatti and Boultwood, 2020).