Part IV · Molecular biology and therapeutic vulnerabilities · Chapter 18

DNA repair, replication stress, and genomic instability

A deficiency leaves a scar, and the scar outlives the deficiency.

1 · Homologous recombination and its failure modes

Homologous recombination repairs double-strand breaks using the sister chromatid as a template. It is therefore available only in S and G2 phases, when a sister chromatid exists. That restriction is why the pathway matters most at the replication fork, and why its failure produces a replication-linked phenotype rather than a general repair defect.

The sequence has distinct steps with distinct owners. Break ends are resected. Single-stranded DNA is coated by replication protein A. RAD51 replaces it to form the nucleoprotein filament. The filament invades the homologous duplex, and repair is completed and resolved.

BRCA1 and BRCA2 act at different points, and calling them interchangeable obscures the clinically useful part. BRCA1 acts early, at resection and at the choice between homologous recombination and non-homologous end joining, in opposition to 53BP1. BRCA2 acts late, loading RAD51 onto single-stranded DNA. A BRCA1-deficient cell can be restored by removing 53BP1, because resection becomes possible again. A BRCA2-deficient cell cannot be rescued that way, because the missing step is downstream.

The gene list is wider than BRCA1 and BRCA2, and the phenotype is not uniform across it. PALB2 partners BRCA2 in RAD51 loading, and RAD51C and RAD51D are paralogues of the recombinase. ATM and CHEK2 are checkpoint kinases rather than core recombination proteins, and that distinction has a clinical readout. TBCRC 048 enrolled 54 patients with metastatic breast cancer and mutations in homologous recombination related genes. Confirmed responses to olaparib occurred only with germline PALB2 mutations, at an objective response rate of 82%, and with somatic BRCA1 or BRCA2 mutations, at 50%. No responses occurred in patients whose only alteration was in ATM or CHEK2¹.

A DNA repair gene is not a homologous recombination gene. The trial separates them, and a panel report that groups them does not.

2 · Homologous recombination deficiency as a measurable phenotype

Homologous recombination deficiency is a functional state, and four families of assay try to capture it. They measure different things, and the differences are the content of this section.

The first measures cause. Sequencing BRCA1, BRCA2 and related genes reports a lesion. It does not report function. It misses BRCA1 promoter methylation entirely, and it returns variants of uncertain significance that no threshold resolves.

The second measures scar. Genomic instability scores count the marks that failed repair leaves on chromosomes: loss of heterozygosity, telomeric allelic imbalance and large-scale state transitions. The combined homologous recombination deficiency score is an unweighted sum of the three. Deficiency was defined as a score of 42 or above, or a BRCA1 or BRCA2 mutation. On that definition it predicted residual cancer burden of 0 or I and pathologic complete response across three neoadjuvant triple-negative trials of platinum-containing therapy. The association held in tumours carrying no BRCA1 or BRCA2 mutation².

The third measures signature. Mutational signatures record the processes that generated a genome rather than the breakpoints they produced. HRDetect is a weighted model built from six signatures, with 98.7% sensitivity for BRCA1 or BRCA2 deficient samples. Applied to 560 breast cancers, of which 22 carried a known germline BRCA1 or BRCA2 mutation, it identified a further 22 with somatic loss and 47 with functional deficiency and no detectable mutation. The proportion of breast cancers with BRCA1 or BRCA2 deficiency rose to as much as 22% on that classification³.

The fourth measures function now. RAD51 nuclear foci report whether the filament forms. Low RAD51 was associated with objective response to PARP inhibition, and the assay is evaluable in untreated tumours because endogenous DNA damage supplies the stimulus⁴.

3 · PARP trapping and synthetic lethality

The synthetic lethal relationship was established directly. BRCA-deficient cells are acutely sensitive to PARP inhibition while BRCA-proficient cells are not, and the difference is large enough to be a therapeutic strategy rather than a therapeutic index⁵.

The original explanation was that PARP inhibition blocks single-strand break repair, that unrepaired single-strand breaks become double-strand breaks when a fork encounters them, and that a homologous-recombination-deficient cell cannot resolve those. The model is incomplete. Clinical PARP inhibitors trap PARP1 and PARP2 on DNA, and trapping potency does not track catalytic inhibition potency across agents⁶. The trapped protein-DNA complex is more cytotoxic than the unrepaired lesion it replaces.

That reframes the drug class. These agents are not only enzyme inhibitors. They are poisons that convert an enzyme into a lesion. Two consequences follow. Agents matched for catalytic potency are not matched for cytotoxicity or for myelosuppression. And the relevant measure of exposure is occupancy on chromatin rather than plasma concentration, which no clinical assay reports.

The randomised evidence in metastatic disease is consistent across two agents. Olaparib against single-agent chemotherapy of the physician's choice, in 302 patients with germline BRCA mutations and HER2-negative disease, gave median progression-free survival of 7.0 months against 4.2 months, hazard ratio 0.58, with response in 59.9% against 28.8%⁷. Talazoparib in the same design, in 431 patients, gave 8.6 months against 5.6 months, hazard ratio 0.54, with response in 62.6% against 27.2%⁸.

The adjuvant result is the one that changes outcome rather than interval. In OlympiA, 1,836 patients with high-risk HER2-negative early breast cancer and germline BRCA1 or BRCA2 pathogenic variants received one year of olaparib or placebo. Three-year invasive disease-free survival was 85.9% against 77.1%, hazard ratio 0.58⁹. At that interim analysis there were 59 deaths on olaparib and 86 on placebo, a difference that did not cross the prespecified significance boundary⁹.

4 · Reversion mutations and other routes to PARP inhibitor resistance

A reversion mutation is a second genetic event that restores the reading frame of a gene already broken by the first. It is the least ambiguous resistance mechanism in oncology, because the restored sequence is directly readable.

Two reports published together in 2008 established it in BRCA2. Selecting a BRCA2-mutated cell line with cisplatin produced five different secondary mutations that all restored the wild-type reading frame. Every resistant clone was resistant to both cisplatin and a PARP inhibitor, and a recurrent ovarian tumour that acquired cisplatin resistance had reverted its BRCA2 mutation¹⁰. In parallel, PARP-inhibitor-resistant clones from a BRCA2-mutant line carried intragenic deletions that removed the frameshift and restored the open reading frame. Those cells regained RAD51 foci and once again limited genotoxin-induced instability¹¹.

One detail in that second report deserves to be pulled out. Most of the deletions occurred at short tracts of homology, consistent with error-prone repair caused by the BRCA2 deficiency itself¹¹. The defect generates the events that cure it. That is not a metaphor. It is the mechanistic link to Polymerase theta and alternative end joining.

Reversion is not the only route, and treating it as the default is the common error. In PARP-inhibitor-resistant germline BRCA1 patient-derived xenografts, exome sequencing found no in-frame secondary mutations at all. Hypomorphic BRCA1 protein was present in 60% of the models, TP53BP1 loss in 20%, and RAD51 amplification in one sample. None of these were mutually exclusive⁴. One of three resistant germline BRCA2 tumours in the same series did show BRCA2 restoration by exome sequencing.

The measurement consequence is unambiguous. A genomic scar assay detects none of these mechanisms, because none of them erases the scar. A sequencing assay detects reversion and misses the rest. RAD51 foci were the only feature shared across every resistant model and sample regardless of underlying mechanism⁴. Reversions are detectable in circulating tumour DNA, which is the application developed in Liquid biopsy and longitudinal monitoring.

5 · Replication stress, ATR, WEE1, and checkpoint targeting

Replication stress is the slowing or stalling of replication forks. In breast cancer its usual source is oncogenic: MYC amplification and cyclin E1 amplification both drive excessive origin firing and deplete the nucleotide pool. The genes in question are the ones already met in MYC, FGFR, and other recurrent amplification-driven dependencies and Rb loss, cyclin E amplification, and CDK6 upregulation, which is worth noticing. The alteration that produces cell cycle escape also produces the vulnerability.

The checkpoint that manages this stress is a small network with two druggable nodes. ATR senses replication protein A coated single-stranded DNA at stalled forks and signals through CHK1. The signal suppresses new origin firing and stabilises existing forks. WEE1 restrains CDK1 and CDK2, which does much the same job from the other direction.

Inhibiting either node removes a brake rather than adding a block. A cell already firing too many origins fires more, exhausts its replication protein A, and leaves single-stranded DNA unprotected. Forks then collapse into double-strand breaks. The therapeutic logic is the inverse of PARP inhibition. PARP inhibition exploits a repair defect. Checkpoint inhibition exploits an excess of stress.

Two obstacles keep this class out of practice. The first is a selection problem. There is no validated clinical assay for replication stress, and the candidate markers are surrogates whose predictive value has not been established. The second is that the normal proliferating compartment also depends on these checkpoints, so the therapeutic index is narrow and combination schedules dominate development.

One combination signal is worth recording. PARP inhibitor resistance in patient-derived xenografts was reverted by adding an ATM inhibitor to the PARP inhibitor⁴. That result is preclinical. It is nonetheless the clearest statement of what the class is for, which is to be given with something rather than instead of it.

6 · Polymerase theta and alternative end joining

When homologous recombination and classical non-homologous end joining both fail, a backup operates. Polymerase theta, encoded by POLQ, drives microhomology-mediated end joining. It aligns broken ends at short stretches of sequence identity and joins them. The repair is mutagenic by construction, because the sequence between the microhomologies is lost.

The dependency is synthetic lethal with homologous recombination deficiency, and the demonstration is symmetrical. Polymerase theta expression is inversely correlated with recombination activity. Knockdown in recombination-proficient cells raises that activity and RAD51 filament assembly, while knockdown in deficient cells enhances cell death. Combined genetic inactivation of a Fanconi anaemia gene and Polq is embryonically lethal in mice. Polymerase theta carries RAD51 binding motifs and blocks RAD51-mediated recombination directly¹².

The clinical argument is a second synthetic lethality in the same population, and it is not simply a substitute for the first. A tumour that has escaped PARP inhibition by restoring homologous recombination has also escaped polymerase theta dependence, because the dependence was a consequence of the deficiency. A tumour that has escaped by drug efflux, by hypomorphic protein or by altered trapping retains the deficiency and should retain the dependence. Which of those applies cannot be determined from a scar assay, and this is where the functional assays in Homologous recombination deficiency as a measurable phenotype stop being an academic preference.

The loop closes on the signature. Deletions flanked by microhomology are a polymerase theta footprint, and they are among the patterns that homologous recombination deficiency classifiers detect³. The backup pathway writes the scar that the assay reads. No agent in this class has yet produced randomised evidence in breast cancer, and nothing here should be read as a clinical recommendation.

7 · Genomic instability, cGAS-STING, and the bridge to immune recognition

Chromosome mis-segregation produces micronuclei, each a fragment of chromatin inside its own defective nuclear envelope. That envelope ruptures. Genomic DNA then reaches the cytosol, where cyclic GMP-AMP synthase binds it and generates the second messenger that activates STING. The synthase accumulates at micronuclei after envelope breakdown, the response is cell cycle dependent, and interferon-stimulated gene expression is induced specifically in micronucleated cells¹³. Genomic instability is therefore not only a source of mutations. It is a source of an innate immune signal.

Breast cancer supplies the clinical version. Among 184 breast cancer samples, the 65 belonging to a DNA damage response deficient subtype showed CD4-positive and CD8-positive lymphocytic infiltration. Cells of that subtype expressed the chemokines CXCL10 and CCL5 at 3.5 to 11.9 times the level in proficient cells, and their conditioned medium attracted peripheral blood mononuclear cells through those two chemokines. The cGAS, STING, TBK1 and IRF3 pathway was constitutively active in a cell cycle specific manner¹⁴.

The same study found the turn in the argument. S-phase DNA damage induced PD-L1 expression in a STING-dependent manner¹⁴. One pathway recruits the lymphocytes and installs the ligand that disables them. An infiltrated tumour that is not controlled is therefore the expected output rather than a contradiction. The immunology is developed in Breast cancer immunology.

Three things follow for Part X. Platinum and PARP inhibition increase cytosolic DNA, which is the mechanistic rationale for combining them with checkpoint blockade rather than sequencing them. Resistance to PARP inhibition splits into mechanisms that restore repair and mechanisms that do not, and only the second group keeps the vulnerabilities in Replication stress, ATR, WEE1, and checkpoint targeting and Polymerase theta and alternative end joining. And the assays defining this population detect a historical state, so the population they define at progression is not the one they defined at diagnosis. Treatment is taken up in Metastatic triple-negative disease and Targeted therapy, and the inherited context in Inherited susceptibility and familial breast cancer.

References

1. Tung NM, Robson ME, Ventz S, et al. TBCRC 048: phase II study of olaparib for metastatic breast cancer and mutations in homologous recombination-related genes. J Clin Oncol 2020 38:4274-4282. PMID 33119476
2. Telli ML, Timms KM, Reid J, et al. Homologous recombination deficiency (HRD) score predicts response to platinum-containing neoadjuvant chemotherapy in patients with triple-negative breast cancer. Clin Cancer Res 2016 22:3764-3773. PMID 26957554
3. Davies H, Glodzik D, Morganella S, et al. HRDetect is a predictor of BRCA1 and BRCA2 deficiency based on mutational signatures. Nat Med 2017 23:517-525. PMID 28288110
4. Cruz C, Castroviejo-Bermejo M, Gutierrez-Enriquez S, et al. RAD51 foci as a functional biomarker of homologous recombination repair and PARP inhibitor resistance in germline BRCA-mutated breast cancer. Ann Oncol 2018 29:1203-1210. PMID 29635390
5. Farmer H, McCabe N, Lord CJ, et al. Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy. Nature 2005 434:917-921. PMID 15829967
6. Murai J, Huang SY, Das BB, et al. Trapping of PARP1 and PARP2 by clinical PARP inhibitors. Cancer Res 2012 72:5588-5599. PMID 23118055
7. Robson M, Im SA, Senkus E, et al. Olaparib for metastatic breast cancer in patients with a germline BRCA mutation. N Engl J Med 2017 377:523-533. PMID 28578601
8. Litton JK, Rugo HS, Ettl J, et al. Talazoparib in patients with advanced breast cancer and a germline BRCA mutation. N Engl J Med 2018 379:753-763. PMID 30110579
9. Tutt ANJ, Garber JE, Kaufman B, et al. Adjuvant olaparib for patients with BRCA1- or BRCA2-mutated breast cancer. N Engl J Med 2021 384:2394-2405. PMID 34081848
10. Sakai W, Swisher EM, Karlan BY, et al. Secondary mutations as a mechanism of cisplatin resistance in BRCA2-mutated cancers. Nature 2008 451:1116-1120. PMID 18264087
11. Edwards SL, Brough R, Lord CJ, et al. Resistance to therapy caused by intragenic deletion in BRCA2. Nature 2008 451:1111-1115. PMID 18264088
12. Ceccaldi R, Liu JC, Amunugama R, et al. Homologous-recombination-deficient tumours are dependent on Poltheta-mediated repair. Nature 2015 518:258-262. PMID 25642963
13. Mackenzie KJ, Carroll P, Martin CA, et al. cGAS surveillance of micronuclei links genome instability to innate immunity. Nature 2017 548:461-465. PMID 28738408
14. Parkes EE, Walker SM, Taggart LE, et al. Activation of STING-dependent innate immune signaling by S-phase-specific DNA damage in breast cancer. J Natl Cancer Inst 2017 109:djw199. PMID 27707838
15. Bakhoum SF, Ngo B, Laughney AM, et al. Chromosomal instability drives metastasis through a cytosolic DNA response. Nature 2018 553:467-472. PMID 29342134