DSBs are primarily repaired via three major pathways in cancer cells: NHEJ, HR and alt-EJ, with single-strand annealing (SSA) also used in certain circumstances (Fig. 1). These pathways differ in mechanism, repair fidelity and cell cycle regulation, collectively maintaining genome stability while balancing repair efficiency and accuracy (21,35-38).

NHEJ is the predominant DSB repair pathway throughout the cell cycle and is especially active during G₁ phase. It involves recognition of broken DNA ends by the Ku70/80 heterodimer, which recruits and activates the kinase of DNA-dependent protein kinase catalytic subunit (DNA-PKcs) and Artemis to process DSB ends, followed by ligation with XRCC4-LIG4-XLF complex (39-41). NHEJ typically re-ligates breaks with minimal processing, which occasionally causes small insertions or deletions.

HR offers high-fidelity repair by using a homologous template, primarily in S/G₂ phases. The MRN complex (MRE11-RAD50-NBS1) recognizes DSBs, which activates ATM kinase, initiates MRE11-mediated resection and produces 3'single-stranded DNA (ssDNA) overhangs coated by RPA, facilitating replacement by RAD51 through BRCA1/2 mediation (13). This mediates strand invasion and template-directed repair via gene conversion or break-induced replication. HR deficiency (HRD) tumors usually show hypersensitivity poly (ADP-ribose) polymerase (PARP) inhibitors (PARPi) (42-44).

SSA, although less common, uses extensive homology between repeat sequences flanking a DSB, which contributes to genome instability when dysregulated (38). Initiation of SSA requires extensive 5'-3' end resection producing long 3' (ssDNA) tails bound by RPA, followed by RAD52-mediated annealing of complementary repeats (45,46).

Alt-EJ was first discovered in NHEJ-deficient cells and considered an alternate repair route for rejoining DSBs in the presence of compromised NHEJ and HR (47,48). At present, alt-EJ has gained prominence as a distinct, regulated pathway that contributes to both physiological repair processes and pathological genome instability, particularly in cancer. Alt-EJ utilizes short microhomologous sequences, typically 2-20 base pairs, near DSB ends to facilitate alignment and repair. Unlike NHEJ, which ligates DNA ends with minimal processing, or HR, which uses templates for error-free repair, alt-EJ requires DNA end resection and often results in deletions, insertions and chromosomal rearrangements, thereby contributing markedly to genomic instability in cancer (49,50).

The Alt-EJ process is initiated by rapid detection of DSBs and stabilization of broken DNA ends, predominantly mediated by PARP1, which recruits repair factors and modulates chromatin at break sites (51,52). The MRN complex, in concert with CtIP, catalyzes 5'-3' resection of the DNA ends, generating 3' ssDNA overhangs required for microhomology exposure (1,53). When microhomology sites are spaced farther apart, exonuclease 1 (EXO1) and Bloom syndrome helicase (BLM) extend this resection, producing longer ssDNA tracts (21,53,54). This extensive resection sharply distinguishes alt-EJ from NHEJ, which repairs minimally processed ends.

Alt-EJ exploits short stretches of microhomology exposed on the complementary 3' ssDNA overhangs for annealing. This step is inherently mutagenic, frequently generating nucleotide deletions or insertions at repair junctions (21,49). DNA Pol θ, encoded by POLQ, is the key mediator of this process, performing DNA synthesis that extends and stabilizes annealed microhomologies, thereby bridging the broken ends (55,56).

POLQ is unique in its domain architecture and consists of an N-terminal helicase-like domain and a C-terminal A-family polymerase domain linked by a central region. The domains both contribute to the end-joining activity of POLQ by stabilizing DNA synapses and catalyzing synthesis across partially paired or mismatched templates (18). POLQ's low-fidelity polymerase lacks 3'-5' proofreading and tolerates base mispairing, enabling it to efficiently extend from short microhomologies but often introducing mutations or templated insertions that constitute the signatures of alt-EJ repair (57,58). The final ligation step in alt-EJ is mediated by LIG1 or LIG3, in complex with XRCC1 (59-61).

Beyond POLQ, DNA polymerase λ (Polλ), an X-family polymerase typically involved in NHEJ, has been implicated in an alternative alt-EJ mechanism. Polλ promotes alt-EJ independently of canonical NHEJ factors by stabilizing DNA end synapses with minimal base pairing and inserting nucleotides at gaps flanked by microhomologies of 4-6 base pairs. This polymerase acts on a range of substrates, including 5-12 nucleotide single-stranded overhangs and small gaps, expanding the diversity of polymerases supporting alt-EJ (62).

Following gap-filling synthesis by POLQ or Polλ, displaced 3' ssDNA flaps must be removed to generate proper DNA ends for ligation. The apurinic/apyrimidinic endonuclease APE2 has recently emerged as a critical nuclease in this process. APE2 exhibits 3'-5' exonuclease and flap endonuclease activities, enabling it to cleave 3' flaps generated during alt-EJ and facilitate end processing necessary for subsequent repair (63-65). Intriguingly, APE2 depletion sensitizes homologous recombination-deficient cells, highlighting its complementary role and synthetic lethality relationship to POLQ (63). This synergy underscores the importance of APE2 in maintaining genome stability by enabling POLQ-driven alt-EJ in contexts where NHEJ or HR is defective.

PARP1 recruitment is essential, particularly when NHEJ and HR are compromised, as it promotes assembly of alt-EJ factors at damage sites (23). POLQ is a central determinant of alt-EJ, not only for carrying out annealing and extension at microhomologies but also for antagonizing HR by preventing RAD51 filament formation (66-68). The 9-1-1 (RAD9A-RAD1-HUS1)/Rad9-Hus1-Rad1-interacting nuclear orphan (RHINO) complex also directs POLQ to DSBs during mitosis, highlighting the cell cycle-specific regulation of alt-EJ (69).

Although microhomology usage is often considered a defining feature of alt-EJ, numerous repair junctions formed by this pathway lack clear microhomologous sequences. For example, studies quantifying chromosomal aberrations suggest that up to 50% of events attributed to alt-EJ occur without detectable microhomology, even when they require POLQ or other essential alt-EJ components (18,70). This may reflect variability in DNA end resection, the influence of local sequence context, or the capacity of alt-EJ polymerases such as POLQ to facilitate end joining through template-independent synthesis (56,71). These findings indicate that alt-EJ is a mechanistically flexible and heterogeneous pathway and that reliance solely on microhomology as a marker may underestimate its role in genome instability and cancer.

The efficiency and reliance on alt-EJ in cells, especially cancer cells, are shaped by a complex interplay between intrinsic genomic factors, regulatory molecules, cell cycle status and external signals (Fig. 2). Alt-EJ activity is governed by the degree of 5'-3' resection at DSB ends, initiated by the MRN complex with CtIP and further extended by nucleases/helicases such as EXO1 and BLM (53,72). This resection exposes microhomologies, with the length of ssDNA determining whether short (2-6 bp) or longer (>10 bp) microhomologies are used (73). Factors such as 53BP1 restrain resection, favoring classical NHEJ and loss or suppression of 53BP1 skews repair towards alt-EJ (74). Deficiencies in other factors, such as the ssDNA-binding protein RPA, can also potentiate alt-EJ by increasing the use of longer microhomologies and promoting chromosomal rearrangements (73,75). Moreover, components mutated in Fanconi Anemia (FA), a disorder characterized by defective interstrand crosslink (ICL) repair, have been implicated in modulating alt-EJ activity, suggesting cross-talk between replication stress responses and alt-EJ (23,76,77).

Alt-EJ is also subject to regulation by extracellular signals, notably TGFβ. Upon activation in the tissue microenvironment, TGFβ ligands bind to heteromeric complexes of serine/threonine kinase receptors, TGFβ receptor type I and type II, initiating downstream phosphorylation cascades primarily mediated by receptor-regulated suppressor of mother against decapentaplegic (Smad) proteins (Smad2, Smad3 and Smad4) that translocate to the nucleus, or by non-canonical TGFβ pathways (Fig. 2). These TGFβ signaling transduction pathways orchestrate a broad array of biological functions that govern cell cycle, differentiation, apoptosis and importantly, DDR pathways (25,78,79). Active TGFβ signaling supports ATM-dependent DDR and suppresses alt-EJ gene expression (80). Conversely, TGFβ pathway inhibition, frequently observed in the tumor microenvironment or certain viral infections, upregulates alt-EJ components (POLQ, PARP1 and LIG1) and increases reliance on this mutagenic pathway, leading to more frequent chromosomal aberrations (25,81). Mechanistically, this effect is partly mediated by the microRNA miR-182, which is upregulated upon TGFβ inhibition and acts to repress key HR and NHEJ repair effectors, thereby shifting repair pathway choice towards alt-EJ (24).

Oncogenic viruses such as human papillomavirus (HPV) further hijack the alt-EJ machinery. A critical step in HPV-mediated oncogenesis is the integration of viral DNA into the host genome, which not only ensures persistent viral replication but also promotes genomic instability and malignant transformation. Parfenov et al (82) performed comprehensive genomic profiling of 279 head and neck squamous cell carcinoma (HNSCC) specimens using next-generation sequencing techniques and found that ~60% of HPV integration events occurred within regions characterized by microhomology at the junction sites. Since microhomology usage is a hallmark of alt-EJ, the authors' study suggests that HPV exploits this error-prone repair pathway to facilitate viral insertion and tumorigenesis.

This hypothesis is supported by several mechanistic studies demonstrating elevated alt-EJ activity in HPV-positive cancers (Fig. 2). For example, Liu et al (24) reported markedly higher expression of key alt-EJ factors in HPV-positive HNSCC compared with HPV-negative tumors. Functional assays in that study showed that HPV-positive tumor cells preferentially rely on PARP1-dependent alt-EJ for repairing DSBs, highlighting a shift in DSB repair pathway choice induced by HPV infection. Furthermore, Liu et al (25) analyzed transcriptomic data from The Cancer Genome Atlas (TCGA) that comprised 243 HPV-negative and 36 HPV-positive HNSCC samples, where the authors found unsupervised clustering based on alt-EJ gene signatures distinctly segregated HPV-positive tumors into a cluster characterized by high expression of alt-EJ-associated genes. Similarly, Guix et al (79) validated this signature in patient-derived xenografts and primary tumor tissues via NanoString assays, finding that HPV-positive samples exhibited an upregulated alt-EJ signature. Independent validation by Leeman et al (83) also demonstrated that HPV infection increases the frequency of DNA deletions flanked by microhomology by alt-EJ. The authors' mechanistic dissection revealed that the HPV oncoprotein E7 promotes a shift in DSB repair pathway choice toward alt-EJ. The aforementioned evidence suggests the existence of a strategy wherein a virus hijacks alt-EJ in its host for its oncogenic agenda.

The interplay between alt-EJ and the canonical DSB repair pathways, HR and NHEJ, reflects the remarkable plasticity of the cellular DNA damage response. Disruption or inhibition of HR or NHEJ frequently redirects DSB repair toward alt-EJ, suggesting that alt-EJ is a flexible and adaptable mechanism, which may be exploited by cancer cells to maintain survival under genotoxic stress (24,84). Studies have revealed that the regulation of alt-EJ is a dynamic and multifaceted process involving complex coordination between numerous signaling pathways and repair proteins, including TGFβ, PARP1 and POLQ and factors involved in DNA end resection (22,62,85). Such regulation is critical because dysregulated repair pathway choice can profoundly affect genome stability, with significant implications for tumorigenesis and therapeutic resistance.

Although the alt-EJ mechanism was originally described as a backup pathway to NHEJ and HR, emerging evidence now positions alt-EJ as an independent and potentially competitive repair mechanism that operates even in the presence of intact HR and NHEJ pathways (1,86). Alt-EJ is preferentially engaged during the S and G₂ phases of the cell cycle when limited DNA end resection has occurred, contributing 10-20% of DSB repair activity in various mammalian cell contexts (48). This is particularly salient in conditions of replication stress, where alt-EJ facilitates rapid repair of collapsed replication forks, balancing between the cytotoxic risk of error-prone repair and the cellular imperative to maintain genomic integrity.

A key nexus of crosstalk exists between HR and alt-EJ centers on POLQ, a multifunctional enzyme with both DNA polymerase and N-terminal helicase domains endowed with ATPase and DNA unwinding activities (50,57,87,88). POLQ's helicase domain mediates critical molecular antagonism of HR by binding RAD51 and thereby preventing RAD51 nucleoprotein filament assembly on RPA-coated ssDNA, a prerequisite for HR-mediated strand invasion. This inhibitory interaction facilitates the suppression of HR and promotes alt-EJ, effectively channeling repair towards a more error-prone pathway (68,88). The interaction is structurally mediated by a disordered central domain within POLQ (residues 847-894) that is essential for RAD51 binding and displacement, thus highlighting a direct molecular mechanism for pathway choice regulation (68,89). However, the question of whether POLQ's ATPase activity regulates the stability and dynamics of RAD51 filaments remains to be elucidated.

The synthetic lethal relationship between HRD and alt-EJ dependency further exemplifies their functional interplay. HR-deficient tumors rely heavily on alt-EJ for repairing complex DSBs, especially those that arise from replication fork collapse, events characterized by single-ended DSBs that cannot be repaired by classical NHEJ due to the absence of a second DSB end (90). Therefore, POLQ-mediated alt-EJ is potentially vital in bridging ssDNA gaps formed during replication stress and fork collapse, particularly important in HRD contexts or following treatment with PARPi (85,91,92).

Single-ended DSBs typically arise when an unrepaired single-strand break is converted into a DSB, leading to replication fork collapse. It is also important to appreciate that single-ended DSBs, the predominant form of DSB in unperturbed cells due to unrepaired single-strand breaks converted into DSBs during replication, are primarily repaired by HR (93,94). Alt-EJ repair of single-ended DSBs tends to be detrimental, often causing chromosomal abnormalities and increased genome instability and is thus suppressed under normal conditions to preserve genomic integrity (94). However, alt-EJ can act as a salvage pathway in HR-compromised scenarios, albeit at the cost of increased mutagenesis and chromosomal rearrangements (90). This delicate balance illustrates an intricate regulatory network where repair pathway choice is tightly controlled to limit deleterious outcomes while facilitating survival.

Alt-EJ often introduces insertions and deletions (indels) at DSB repair sites. This pathway not only fosters genetic alterations within coding regions but also leaves distinctive 'genomic scars'. These scars frequently involve microhomology regions that facilitate end alignment prior to repair, leading to characteristic deletions, rearrangements and templated or non-templated insertions, which can be detected as a genomic signature for alt-EJ activity.

In the context of cancer, the mutagenic potential of alt-EJ is especially significant. For example, loss of TGFβ signaling increases alt-EJ in pan-cancer, which lead to increased genomic alterations (25). Notably, a subset of types of cancer with upregulated alt-EJ has shown a specific indel mutation signature, characterized by >5 base pair deletions and overlapping microhomology at deletion boundaries (25). Furthermore, in BRCA-mutant tumors, increased reliance on alt-EJ not only drives tumor progression but also engenders specific mutational signatures that can be exploited for diagnosis or therapeutic targeting. These mutational signatures include recurrent deletions featuring microhomology at breakpoint junctions and complex rearrangements that impact genome stability. Such patterns serve as genomic scars indicative of defective HR repair and are associated with increased sensitivity to PARPi (95-98).

POLQ facilitates annealing of microhomologous regions and contributes to the characteristic junctional mutations. In a C. elegans study, the results analyzing ~7,000 deletion breakpoints demonstrated that POLQ-dependent alt-EJ accounts for a significant portion of mutations induced by alkylating agents such as ethyl methane sulfonate and UV/TMP treatments (95). These breakpoints often include small deletions due to microhomology and show inserted DNA sequences, either copied from close (templated) or newly added (non-templated) region.

The frequent insertion of short sequences at breakpoints can be attributed to POLQ's terminal transferase activity and its propensity for template switching processes that generate diverse repair signatures, including partial ssDNA intermediates (99). This accumulation of ssDNA renders cells more susceptible to further damage or mutations, especially under conditions of external genotoxic stress or endogenous replicative stress. Alt-EJ leads to exposure of ssDNA regions and increased opportunities for nucleotide misincorporation or further damage. Notably, hypermutagenesis tends to extend over a defined physical distance (~7-9 kb) from the breakpoints, indicating that error-prone repair is not confined solely to immediate junctions but can propagate along adjacent DNA regions (100). Notably, this hypermutability appears to be an intrinsic feature of alt-EJ and not solely dependent on POLQ activity, suggesting that the pathway's inherent mechanics predispose to extensive mutagenic outcomes.

Alt-EJ often results in chromosomal aberrations and translocations. In Saccharomyces cerevisiae (budding yeast), the presence of multiple simultaneous DSBs increases the likelihood of promiscuous end joining via alt-EJ, generating chromosomal translocations and complex rearrangements (101). This finding highlights the potential for alt-EJ to misrepair disparate DNA ends. Consistently, a high frequency of microhomology at translocation junctions has been detected in human tumor cells, supporting a mechanistic link between alt-EJ and chromosomal rearrangements in cancer (102,103). This microhomology enrichment at breakpoints contrasts with the more blunt-end ligation characteristic of NHEJ, underscoring the distinct repair signature of alt-EJ. Furthermore, loss of TGFβ signaling increases IR-induced chromosome aberrations in HNSCC cells, but knock-down of POLQ suppresses the effects, suggesting a critical role for alt-EJ (24). Cisplatin treatment also induces chromosomal aberrations, especially in human papillomavirus (HPV)-positive HNSCC with preference for alt-EJ (76).

To investigate the direct role of alt-EJ in chromosomal translocation formation, studies induced targeted DSBs on distinct chromosomes using site-specific nucleases such as CRISPR-Cas9 or I-SceI endonuclease (104,105). Experiments performed in both murine and human cells deficient in NHEJ factors demonstrated that loss of key NHEJ components, notably XRCC4, substantially elevated the frequency of reciprocal translocations up to fivefold compared with wild-type controls (104,105). Furthermore, microhomologies were observed at ~60% of these breakpoint junctions, strongly implicating alt-EJ in the formation of the chromosomal aberrations. Beyond XRCC4, other proteins critical to alt-EJ function, such as CtIP, which initiates DNA end resection, and LIG3, which catalyzes the final ligation, have been shown to influence chromosomal rearrangement frequencies. For example, mouse cells depleted of CtIP or LIG3 exhibit reduced alt-EJ activity and corresponding decreases in chromosomal aberrations, indicating that a functional alt-EJ machinery is required for these instability phenotypes (70,106).

Telomeres, the specialized nucleoprotein structures capping chromosome ends, are essential for preserving genomic integrity by preventing chromosome ends from being recognized as DSBs. However, critically short or dysfunctional telomeres that result from replicative attrition or loss of protective factors become 'uncapped', exposing chromosome ends as DNA breaks vulnerable to erroneous repair pathways. In these contexts, alt-EJ plays a significant role in driving telomere fusions, which contribute to profound genome instability. Indeed, telomere fusion events mediated by alt-EJ have been observed in aggressive cancers including glioblastomas, where they correlate with poor prognosis and increased genomic chaos (107-109).

Recent investigations have provided critical mechanistic insights into alt-EJ mediators in telomere fusion. For example, one study that used mouse embryonic fibroblasts lacking TRF2 and Ku demonstrated robust NHEJ-independent telomere fusions that are highly dependent on POLQ and this POLQ-dependent fusion process is sensitive to inhibition by PARPi, highlighting a potential therapeutic vulnerability in tumors reliant on alt-EJ for telomere maintenance (69). Importantly, the same study revealed that APE2 acts as an epistatic partner of POLQ in executing these fusion events, as simultaneous depletion of POLQ and APE2 did not further reduce telomere fusions compared with single depletions. These findings not only strengthen evidence that POLQ is a central player in alt-EJ-mediated telomere fusion but also identify APE2 as a novel cooperating factor in this pathway.

Consistent with the pronounced effects of alt-EJ in genome instability, recent discoveries have also increasingly linked aberrant activation of alt-EJ to cancer development, progression and therapeutic resistance. In HRD tumors, such as those harboring BRCA1/2 mutations, reliance on alt-EJ fosters the accumulation of microhomology-associated deletions and templated insertions, generating genomic scars that drive tumor evolution and heterogeneity (25,55,95,98). These genomic features are associated with poor clinical prognosis and have been proposed as predictive biomarkers for response to genotoxic therapies and PARPi. Large-scale pan-cancer transcriptomic analyses have further revealed that tumors with elevated alt-EJ activity frequently exhibit high expression of POLQ and PARP1, correlating with increased chromosomal instability and resistance to standard treatments (22,25,68,98). Together, these findings support the emerging view that alt-EJ is an active contributor to malignant transformation. Understanding these oncogenic consequences of alt-EJ and factors that affect alt-EJ activity are critical for harnessing its components as potential diagnostic markers and novel therapeutic targets in cancer.