Petljak et al. (2022) — Mechanisms of APOBEC3 Mutagenesis in Human Cancer Cells

Bibliographic Reference

Petljak, M., Dananberg, A., Chu, K., Bergstrom, E. N., Striepen, J., von Morgen, P., Chen, Y., Shah, H., Sale, J. E., Alexandrov, L. B., Stratton, M. R., & Maciejowski, J. (2022). Mechanisms of APOBEC3 mutagenesis in human cancer cells. Nature, 607(7919), 799–807. https://doi.org/10.1038/s41586-022-04972-y

Core Argument

Two decades after APOBEC3 enzymes were first proposed as endogenous mutators in cancer, this paper provides the first direct causal evidence. By deleting APOBEC3A and APOBEC3B from human cancer cell lines that naturally acquire APOBEC3-associated mutations and tracking mutation acquisition via whole-genome sequencing of 251 clonal lines, the authors demonstrate that APOBEC3A — not APOBEC3B — is the main driver of the prevalent SBS2 and SBS13 mutational signatures. APOBEC3B can restrain APOBEC3A-dependent mutagenesis while contributing its own smaller mutation burdens. The uracil glycosylase UNG is required for APOBEC3-mediated transversions, and the translesion polymerase REV1 is a critical mediator of overall APOBEC3 mutational burdens. This study resolves a longstanding debate in the field by establishing causal mechanisms rather than relying on correlative expression and in vitro deamination data.

Methods

CRISPR–Cas9 deletion of APOBEC3A, APOBEC3B, both, UNG, SMUG1, and REV1 from a panel of human cancer cell lines (breast: BT-474, MDA-MB-453; B cell lymphoma: BC-1, JSC-1; bladder: HT-1376) that naturally acquire APOBEC3-associated mutations over time. Single-cell-derived parent clones were subjected to long-term cultivation (60–143 days), followed by subcloning into daughter clones. Whole-genome sequencing of 251 parent and daughter clones (plus 5 bulk cell lines). Mutational signature extraction and decomposition into COSMIC reference signatures. Surrogate assays: cytosine deaminase activity on linear and hairpin DNA probes, RNA editing at DDOST hotspot (ddPCR), protein levels (western blot), and cell proliferation/clonal survival assays.

Key Findings

• First causal evidence linking endogenous APOBEC3s to cancer mutational signatures. Deletion of APOBEC3A from breast cancer and B cell lymphoma cell lines significantly diminished SBS2 and SBS13a/b mutations. This is the first direct experimental demonstration — not inference from expression or in vitro deamination — that endogenous APOBEC3 enzymes generate these signatures in human cancer cells.

• APOBEC3A is the main driver, not APOBEC3B. Despite APOBEC3B having substantially higher mRNA and protein expression and being the dominant source of cytosine deaminase activity in cell extracts — the very evidence that led the field to designate APOBEC3B as the primary mutator — deletion of APOBEC3B did not significantly reduce SBS2 and SBS13a/b burdens. APOBEC3B contributes only small mutation burdens, revealed only after APOBEC3A deletion unmasks them. This directly challenges the prevailing APOBEC3B-centric model (Burns et al., 2013, 2015).

• APOBEC3B can RESTRAIN APOBEC3A. In MDA-MB-453 breast cancer cells, APOBEC3B knockout significantly increased SBS2 and SBS13a/b mutations — APOBEC3A-mediated mutagenesis was higher in the absence of APOBEC3B. This was accompanied by stabilized APOBEC3A protein levels. shRNA depletion of APOBEC3B confirmed increased APOBEC3A protein and RNA-editing activity. This provides a mechanistic explanation for the paradoxical observation that the APOBEC3B germline deletion polymorphism is associated with increased cancer risk and higher APOBEC3 mutation burdens.

• APOBEC3A prefers YTCA motifs; APOBEC3B prefers RTCA motifs. Endogenous APOBEC3A exhibits a preference for YTCA sequence contexts (Y = pyrimidine), while APOBEC3B prefers RTCA (R = purine). The enrichment of YTCA mutations in wild-type clones and their diminishment after APOBEC3A deletion confirms APOBEC3A as the major contributor.

• Residual APOBEC3 mutagenesis after APOBEC3A/B deletion. SBS2 and SBS13a/b were further diminished but not eliminated in APOBEC3A/APOBEC3B double knockouts. APOBEC3H (haplotype I, present in both BRCA cell lines) or another APOBEC enzyme may contribute small amounts.

• UNG is required for APOBEC3-mediated transversions. UNG deletion reduced the relative proportions of C>A and C>G mutations in TCN contexts and decreased SBS13a/b burdens. UNG–GFP expression in BC-1 cells increased transversion proportions. This directly links uracil excision to the generation of APOBEC3 transversion signatures — UNG converts APOBEC3-mediated C→U deamination events into C>G and C>A mutations.

• REV1 is a critical mediator of APOBEC3 mutagenesis. REV1 deletion led to decreased SBS2 and SBS13a/b mutations and reduced C>G proportions in TCN contexts, consistent with REV1’s deoxycytidyl transferase activity opposite abasic sites. REV1 also contributed to SBS5 — the ubiquitous clock-like signature — suggesting that SBS5 may in part represent a footprint of lower-fidelity REV1-dependent translesion synthesis.

• SMUG1 can partially substitute for UNG. SMUG1 deletion resulted in altered C>A/C>G ratios in TCN contexts and increased SBS13b (characterized by higher C>A relative to C>G), suggesting SMUG1 can occasionally substitute for UNG in excising APOBEC3-mediated uracil.

• Episodic APOBEC3 activity confirmed. SBS2 and SBS13a/b burdens varied enormously across daughter clones derived from the same parent (e.g., BC-1: 12,504 vs 954 SBS2 mutations in 108 days), confirming episodic APOBEC3 mutagenesis not explained by multiclonality, growth differences, or expression levels of candidate mutators.

• APOBEC3A drives both kataegis and omikli. APOBEC3A deletion significantly reduced clustered APOBEC3-associated mutations including kataegis and omikli.

Concepts Introduced or Used

APOBEC-mutagenesis, APOBEC3A, APOBEC3B, APOBEC3H, UNG, REV1, SMUG1, SBS2, SBS13, mutational-signature, kataegis, omikli, uracil-excision, translesion-synthesis, YTCA-motif, RTCA-motif, episodic-mutagenesis, causal-evidence

Entities Referenced

• Cell lines: MDA-MB-453, BT-474 (breast cancer), BC-1, JSC-1 (B cell lymphoma), HT-1376 (bladder cancer)
• Genes: APOBEC3A, APOBEC3B, APOBEC3H, UNG, SMUG1, REV1, DDOST
• Mutational signatures: SBS2, SBS13 (SBS13a, SBS13b), SBS5, SBS288A-E
• Cancer types: breast, bladder, B cell lymphoma, head and neck, cervical
• Methods: CRISPR–Cas9, whole-genome sequencing, mutational signature analysis, ddPCR
• Previous APOBEC3B-centric model: Burns et al. (2013, 2015); Harris et al. studies

Limitations (as stated by authors)

• The study is limited to five cancer cell lines representing three cancer types; the relative contributions of APOBEC3A vs APOBEC3B may differ in other cancer types, particularly those with enrichment of RTCA (APOBEC3B-preferred) motifs.
• Residual APOBEC3 mutagenesis after APOBEC3A/B deletion could not be definitively attributed — APOBEC3H or another enzyme may contribute, but small mutation burdens challenge signature quantification.
• The mechanisms by which APOBEC3B restrains APOBEC3A protein levels require further investigation — CRISPR edits do not resemble the APOBEC3B deletion polymorphism structure.
• The study was performed in vitro; the relative contributions of APOBEC3 enzymes in vivo may differ due to tumour microenvironment factors.
• APOBEC3 expression is dynamic and varies across individual clones; the study captures snapshot measurements.

Relevance to Clonal Evolution

This paper resolves a two-decade debate about which APOBEC3 enzyme drives cancer mutagenesis. The finding that APOBEC3A — not APOBEC3B — is the main mutator, and that APOBEC3B can restrain APOBEC3A, has major implications for therapeutic strategies that previously targeted APOBEC3B. The identification of UNG and REV1 as critical downstream mediators of APOBEC3 mutagenesis opens new therapeutic avenues: targeting the mutational consequences of APOBEC3 activity (uracil processing and translesion synthesis) rather than the deaminases themselves. The episodic nature of APOBEC3 mutagenesis — with 13-fold variation between daughter clones — suggests that mutagenic bursts rather than constant activity generate the subclonal diversity that fuels clonal-evolution and therapy-resistance.