Research Article

Adaptive mutability of colorectal cancers in response to targeted therapies

See allHide authors and affiliations

Science  20 Dec 2019:
Vol. 366, Issue 6472, pp. 1473-1480
DOI: 10.1126/science.aav4474

A cross-kingdom tale of drug resistance

Physicians who treat bacterial infections and those who treat cancer often face a common challenge: the development of drug resistance. It is well known that when bacteria are exposed to antibiotics, they temporarily increase their mutation rate, thus increasing the chance that a descendant antibiotic-resistant cell will arise. Russo et al. now provide evidence that cancer cells exploit a similar mechanism to ensure their survival after drug exposure (see the Perspective by Gerlinger). They found that human colorectal cancer cells treated with certain targeted therapies display a transient up-regulation of errorprone DNA polymerases and a reduction in their ability to repair DNA damage. Thus, like bacteria, cancer cells can adapt to therapeutic pressure by enhancing their mutability.

Science, this issue p. 1473; see also p. 1458


The emergence of drug resistance limits the efficacy of targeted therapies in human tumors. The prevalent view is that resistance is a fait accompli: when treatment is initiated, cancers already contain drug-resistant mutant cells. Bacteria exposed to antibiotics transiently increase their mutation rates (adaptive mutability), thus improving the likelihood of survival. We investigated whether human colorectal cancer (CRC) cells likewise exploit adaptive mutability to evade therapeutic pressure. We found that epidermal growth factor receptor (EGFR)/BRAF inhibition down-regulates mismatch repair (MMR) and homologous recombination DNA-repair genes and concomitantly up-regulates error-prone polymerases in drug-tolerant (persister) cells. MMR proteins were also down-regulated in patient-derived xenografts and tumor specimens during therapy. EGFR/BRAF inhibition induced DNA damage, increased mutability, and triggered microsatellite instability. Thus, like unicellular organisms, tumor cells evade therapeutic pressures by enhancing mutability.

More than 75 years ago, Luria and Delbrück demonstrated that bacterial resistance to phage viruses was due to random mutations that spontaneously occurred in the absence of selection (1). Resistance to targeted therapies in human tumors is also widely thought to be due to mutations that exist before treatment (2). The conventional view is that relapses occur because drug-resistant mutant subclones are present in any detectable metastatic lesion before the initiation of therapy. According to this view, resistance is a fait accompli, and the time to recurrence is merely the interval required for preexisting drug-resistant (mutant) cells to repopulate the lesion (3).

Here, we explore the hypothesis that resistance to targeted therapies can also be fostered by a transient increase in genomic instability during treatment, leading to de novo mutagenesis. A similar process has been shown to increase the emergence of microbial strains resistant to antibiotics (4, 5). In a stable microenvironment, the mutation rate of microorganisms is usually low, which precludes the accumulation of deleterious mutations. However, several mechanisms of stress-induced genetic instability and increased mutability, known as stress-induced mutagenesis (SIM), have been described in bacteria and yeast (612).

Bacterial persister cells can survive lethal stress conditions imposed by antibiotics through a reduction in growth rate. A subsequent reduction in the efficiency of DNA mismatch repair (MMR) (4, 9, 13) and a shift to error-prone DNA polymerases increases the rate at which adaptive mutations occur in the surviving population (4, 9, 14, 15). Selection then allows the growth of mutant subpopulations capable of replicating under stressful conditions. Once the stressed population has adapted to the new conditions, the hypermutator status is counterselected to avoid the accumulation of deleterious mutations and to prevent the continuous increase of mutational load (9, 1620). Together, these processes boost genetic diversity, foster adaptability to new microenvironments, and contribute to the development of resistance (9, 12, 18, 19).

In the setting of cancer, the emergence of a drug-tolerant persister population is often observed when oncogene-dependent tumor cells are challenged with targeted agents (21). Persister cancer cells survive exposure to targeted therapies through poorly understood mechanisms (21) and represent a reservoir from which genetically divergent, drug-resistant derivatives eventually emerge (22, 23). Recent work showed that drug-resistant mutant cancer cells can originate not only from rare, preexisting mutant clones, but also from drug-tolerant subpopulations (24). The probability that the latter resistance mechanism occurs would be greatly increased if the genetic diversity of tumor cells were enhanced during treatment. Accordingly, we hypothesized that during the persister state, tumor cells, like unicellular organisms, alter DNA-repair and DNA-replication mechanisms to enhance adaptive mutability.

Targeted therapy–induced down-regulation of MMR and HR proficiency of CRC cells

To test our hypothesis, we studied the response of microsatellite-stable (MSS) human colorectal cancer (CRC) cell lines to the anti-EGFR (epidermal growth factor receptor) antibody cetuximab, which is approved, together with panitumumab, for the treatment of patients with metastatic CRC whose tumors lack RAS and BRAF mutations (25), or with the BRAF inhibitor dabrafenib (DAB) as combinatorial treatment, which has shown promising activity in patients with CRC harboring BRAF mutations (26). We selected human CRC cell lines that are RAS and BRAF wild-type and sensitive to EGFR blockade (DiFi cells, fig. S1A) or that carry the oncogenic BRAF p.V600E mutation and are sensitive to concomitant EGFR and BRAF inhibition (WiDr cells, fig. S1A). Treatment with targeted agents led to G1 cell-cycle arrest (fig. S1B). However, a small number of drug-tolerant persister cells survived several weeks after treatment initiation (fig. S1, C and D). Indeed, when drug pressure was removed, these cells rapidly resumed growth and again showed sensitivity to targeted therapy, thus demonstrating that persisters are only transiently and reversibly resistant to the treatment (fig. S1, E and F). By contrast, prolonged treatment led to the generation of permanently resistant cells, which did not reacquire sensitivity after the removal of drug pressure (fig. S1, E and F).

We next assessed whether CRC cells modulate the expression of DNA-repair genes upon drug treatment. Transcriptional profiles revealed decreased expression of the MMR genes MLH1, MSH2, and MSH6, as well as of homologous recombination (HR) effectors such as BRCA2 and RAD51 (Fig. 1A and fig. S1, G and H). Expression of EXO1, a gene coding for an exonuclease that participates in mismatch and double-strand break (DSB) repair, was also affected (Fig. 1A and fig. S1, G and H). A time-dependent down-regulation of MMR and HR proteins was also observed (Fig. 1B and fig. S2, A and B). Comparable results were obtained in another cetuximab-sensitive human CRC cell line, NCIH508 (fig. S3, A to C), and in BRAF-mutant HT29 cells that were derived from the same patient from whom the WiDr cell line originated (fig. S3, D and E). Furthermore, we confirmed that down-regulation or loss of DNA-repair components is maintained in persister cells (fig. S4, A to D). Therapy-induced modulation of DNA-repair gene expression was transient and expression levels returned to normal upon removal of treatment (fig. S5A). Cancer cells that had previously developed permanent resistance to targeted agents did not modulate the expression of DNA-repair genes in response to drugs (fig. S5, B and C).

Fig. 1 CRC cells modulate DNA-repair effectors in response to targeted agents.

(A) CRC cells were treated with cetuximab alone (DiFi) or in combination with the BRAF inhibitor DAB (WiDr) for 96 hours and RNA-sequencing analysis was performed. MMR (yellow), HR (green), and DNA polymerase (blue) genes are reported. Results represent means of two independent experiments. (B) CRC cells were treated and analyzed at the indicated time points by Western blot. CTX, cetuximab; pERK, phosphorylated extracellular signal–regulated kinase. (C) CRC cells were transfected with G:C undamaged (UNDAMAG) plasmid or with G:G mismatch-damaged (DAMAG) plasmid. Where indicated (DRUG), cells were treated with targeted therapies for 50-60 hours and analyzed by flow cytometry. A mock transfection was used as a control. Quantification of MMR capacity of each cell line relative to control is reported in the bar graph. LIM1215, MMR-deficient CRC cells, were used as a positive control for MMR loss. Results represent means of two independent experiments. *p < 0.05 (Student’s t test). (D) pDRGFP-stably expressing CRC cells were transfected with the pCBASce-I plasmid and then either left in the absence of drug or treated with targeted therapies for 50-60 hours and analyzed by flow cytometry. A mock transfection was used as a control. Quantification of HR capacity of each cell line relative to mock is reported in the bar graph. Results represent means ± SD (n = 3). **p < 0.01 (Student’s t test).

To ascertain whether targeted therapies affect DNA-repair competence in CRC cells, we used fluorescence-based multiplex host-cell reactivation (FM-HCR) assays (27). CRC cells were transfected with a G:G mismatch-containing plasmid to determine the impact of drug treatment on MMR capacity. An MMR-deficient (MMRd) human CRC cell line (LIM1215) was used as a positive control for MMR loss. We found that in CRC cells treated with targeted agents, MMR proficiency (MMRp) was significantly reduced (Fig. 1C and fig. S6A).

We next evaluated cellular HR capability by using the two-step, plasmid-based pDRGFP/pCBASce-I assay (28). Upon stable expression of the pDRGFP plasmid, we measured the generation of a green fluorescent signal upon DSBs induced by Sce-I expression. This assay showed that both DiFi and WiDr cells had a marked reduction in HR proficiency upon treatment with targeted therapies (Fig. 1D and fig. S6B).

MMR proteins are down-regulated in samples of CRC residual disease after targeted treatment

To determine whether the cell-based findings extend to patient-derived tumor samples, we exploited our CRC biobank of molecularly and therapeutically annotated patient-derived xenograft (PDX) models (29, 30). We selected six MSS PDX models with wild-type KRAS, NRAS, and BRAF in which EGFR inhibition by cetuximab led to tumor regression to a variable extent, paralleling the clinical scenario (Fig. 2A). Immunohistochemistry analysis unveiled areas with down-regulation of MLH1 and/or MSH2 in all neoplastic samples obtained when tumors were at the point of maximum response to cetuximab but still contained residual persisters (Fig. 2, B and C, and fig. S7, A to D), as compared with placebo-treated controls.

Fig. 2 MMR down-regulation in CRC PDXs and patients treated with targeted therapies.

(A) Extent of tumor regression in PDX models after treatment with cetuximab (20 mg/kg twice weekly) for 6 weeks. Each bar is the average of tumor volumes from six mice. (B) Growth-curve kinetics in two out of six PDXs. Shown are mean tumor volumes ± SEM (n = 6). Gray arrows indicate treatment initiation. (C) Immunohistochemical staining with anti-MLH1 and anti-MSH2 antibodies of histologic tumor sections derived from indicated PDXs treated with cetuximab for 6 weeks. Tumor section derived from the placebo arm was used as a control. Scale bar, 0.1 mm. Magnifications are 40× (scale bar, 0.05 mm). (D) Immunohistochemical staining with anti-MLH1 and anti-MSH2 antibodies of tumor sections derived from two CRC patients treated with FOLFOX + the anti-EGFR monoclonal antibody panitumumab. Tumor sections were derived from the primary lesion at diagnosis (pretreatment) and at the time of partial response (PR) when the lesions shrank. Scale bar, 0.1 mm. Magnifications are 40× (scale bar, 0.05 mm).

We next investigated whether down-regulation of DNA-repair proteins also occurs in clinical specimens from two CRC patients who achieved an objective partial response upon treatment with FOLFOX (folinic acid, 5-fluorouracil, and oxaliplatin) plus panitumumab. In both instances, tumor specimens were longitudinally collected at diagnosis and at maximal therapeutic response, when a limited number of tumor cells persist despite treatment. MLH1 and MSH2 were down-regulated in tumor samples obtained at response compared with pretreatment specimens, confirming the clinical relevance of our findings (Fig. 2D).

Induction of DNA damage and error-prone DNA polymerases in CRC cells treated with targeted therapies

In addition to reduced DNA-repair ability, we found that targeted therapies triggered a switch from high-fidelity to low-fidelity DNA polymerases. DNA polymerases usually involved in accurate DNA replication, such as POLδ and POLε, were down-regulated, whereas DNA polymerases characterized by poor accuracy, low processivity, and absence of proofreading capacity (i.e., error-prone polymerases) were induced (Fig. 1A and fig. S4A). These included Polι, Polκ, and Rev1 (which belong to the Y family of polymerases, orthologous to the bacterial stress–induced polymerases Pol IV and Pol V), as well as Polλ and Polμ (31) (Fig. 1, A and B, and figs. S1, G and H; S2B; S3, B to C and E; and S4, A and D). Error-prone polymerases replace canonical high-fidelity polymerases that stall when encountering a DNA lesion and facilitate DNA replication across DNA damage sites in a manner that introduces errors into the genome (15, 16, 20); this may lead to base mispairings, incorporation of aberrant DNA primer ends, and increased mutagenesis rate (32, 33).

We therefore investigated whether treatment with targeted therapies leads to genomic damage in cancer cells and if error-prone–mediated repair of DNA damage is favored when CRC cells encounter the hostile environment imposed by targeted therapies. Indeed, quantification of phosphorylation of H2AX at Ser139 (γH2AX), a common marker of DNA damage (34), revealed a dose- and time-dependent increase in the number of foci-positive nuclei upon drug treatment (Fig. 3, A and B, and fig. S8, A and B), whereas no further increase was observed in permanently resistant cells upon drug treatment (fig. S8, C and D). In addition, we observed a dose- and time-dependent increase in the number of 53BP1-positive nuclei upon EGFR and BRAF blockade (fig. S9, A and B). In direct opposition to BRCA1, 53BP1 promotes nonhomologous end joining–mediated DSB repair while preventing HR through restriction of end resection (35). These data suggest that targeted therapies trigger a switch from high-fidelity to error-prone–mediated repair of DNA damage, thereby potentially increasing the occurrence of mutations conferring drug resistance.

Fig. 3 Targeted therapies trigger a stress response, increase ROS levels, and induce DNA damage in CRC cells.

(A) CRC cells were treated as reported and fixed and stained with anti-γH2AX antibody at the indicated time points. Vehicle-treated cells (NT) were used as controls. Nuclei are stained with DAPI (blue) and anti-γH2AX antibody (red). Scale bar, 50 μm. Representative images for each condition are shown. (B) Quantification of nuclear γH2AX foci in DiFi (left panel) and WiDr (right panel) cells. Results represent means ± SD (n = 3 for 48 and 72 hours; n = 2 for 96 hours). *p < 0.05; **p < 0.01; ***p < 0.001 (two-way ANOVA). (C) CRC cells were treated as indicated and ROS levels were measured. NAC was used as a control to rescue ROS production. Results represent means of two independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001 (Student’s t test). (D) CRC cells were treated with targeted therapies and analyzed by Western blot at the indicated time points. pAMPK, phosphorylated adenosine monophosphate kinase. (E) Wild-type DiFi (left panel) and BRAF-mutated WiDr (right panel) cells were transfected with the indicated siRNA or combination of siRNAs for 72 hours and analyzed by Western blot. ALL STAR, nontargeting siRNA.

We next explored the possible causes of the DNA damage observed upon the administration of targeted therapies. Although several chemotherapeutic agents directly generate DNA damage, drugs interfering with oncogenic signaling (such as EGFR or BRAF inhibitors) are not directly genotoxic. However, it has been shown that certain targeted therapies, such as ABL and BRAF inhibitors, increase the levels of reactive oxygen species (ROS) in cancer cells (36, 37), potentially contributing to DNA damage during treatment. ROS levels significantly increased when CRC cells were exposed to EGFR and BRAF inhibitors (Fig. 3C). By contrast, ROS levels were not increased in permanently drug-resistant (adapted) cells upon drug treatment (fig. S9C).

The drug-induced increase in ROS levels was abrogated when targeted therapies were administered in the presence of the antioxidant N-acetyl-L-cysteine (NAC) (Fig. 3C). NAC administration partially reduced the number of γH2AX foci-positive nuclei upon EGFR and BRAF blockade (fig. S10, A and B). However, cotreatment with NAC did not prevent or rescue down-regulation of DNA-repair genes (fig. S10C). The addition of NAC delayed onset of relapse to targeted therapies when administered together with mitogen-activated protein kinase (MAPK) pathway inhibitors (fig. S10, D and E) (38, 39).

Interfering with oncogenic dependencies initiates a stress response in CRCs

To elucidate the mechanistic basis of therapy-induced mutagenesis in cancer cells, we tested whether the adaptive mutability that we observed in response to targeted therapies was simply a secondary response to G1 cell-cycle arrest or DNA damage or if it represented an active stress response. We found that thymidine-mediated cell-cycle stress (fig. S11, A to C) or direct DNA damage with the alkylating agent oxaliplatin (fig. S11, D to F) instead promoted the up-regulation of the MMR- and HR-repair systems (fig. S11, C and F), and G1 cell-cycle arrest by nutrient starvation did not lead to modulation of DNA-repair gene expression (fig. S11, G to I). In bacterial cells, both the DNA damage–activated SOS response and the general stress response appear to be required to induce adaptive mutagenesis (14). We therefore examined the modulation of the kinase mammalian target of rapamycin (mTOR), which is a master regulator of mammalian cellular stress response (40). Indeed, the mTOR effectors pS6K–p70K were down-regulated with kinetics comparable to that of MMR and HR regulation upon EGFR and BRAF pharmacological blockade (Fig. 3D). However, silencing of mTOR did not affect the expression of DNA-repair proteins or γH2AX (Fig. 3E). It is therefore plausible that down-regulation of mTOR contributes to stress-induced mutagenesis of cancer cells but is not sufficient to activate this phenotype.

The exquisite sensitivity of DiFi and WiDr cells to EGFR and BRAF blockade reflects cell-specific oncogenic alterations. The EGFR locus is amplified in DiFi cells (2); the WiDr cells carry the BRAF p.V600E oncogenic mutation, but they also become dependent on feedback activation of EGFR when treated with BRAF inhibitors (41). We therefore investigated whether interfering with the oncogenic dependency of cancer cells could directly initiate the drug-induced stress phenotype. Indeed, small interfering RNA (siRNA)–mediated knockdown of EGFR or KRAS in DiFi cells and of BRAFEGFR) in WiDr cells led to reduced expression of DNA-repair proteins, triggered DNA damage and mTOR down-regulation (Fig. 3E), and increased ROS levels (fig. S12). These results exclude the possibility that drug-induced down-regulation of DNA-repair pathways could be due to a nonspecific (off-target) effect of the anti-EGFR antibody cetuximab or the BRAF inhibitor dabrafenib.

Targeted therapies induce adaptive mutability in CRC cells

Next, we tested whether the stress response induced by targeted therapies translated into increased mutagenesis in CRC cells. We used a reporter assay in which a dinucleotide CA-repeat microsatellite drives the NanoLuc enzyme coding sequence out of frame (Fig. 4A). Random mutations that introduce frameshifts in this region, in the absence of a functional MMR, would restore the NanoLuc open reading frame, leading to bioluminescence. Analogous approaches have previously been used to measure MMR defects in cancer cells (4244). To validate the assay, we first introduced the CA-NanoLuc vector into a MMRd human CRC cell line (HCT116) and three MMRp human CRC cell lines (DiFi, WiDr, and NCIH508). The NanoLuc signal was significantly higher in MMRd cells after 48 hours of standard growth conditions (Fig. 4B). This difference was further increased when HCT116 cells were kept in culture for several days, whereas the signal in the MMRp lines remained low (Fig. 4B), indicating that the CA-NanoLuc assay effectively detects MMR deficiency in cancer cells.

Fig. 4 Treatment with targeted therapies promotes mutagenesis in CRC cells.

(A) Schematic representation of the CA-NanoLuc reporter assay. (B) MMRd HCT116 and MMRp DiFi, WiDr, and NCIH508 CRC cells were transduced with the NanoLuc lentivirus. At the indicated time points, NanoLuc signal was evaluated and normalized to cell viability. Results represent means ± SD (n = 3). **p < 0.01; ***p < 0.001 (Student’s t test). NS, not a statistically significant difference. (C) NanoLuc signal in HT29 MLH1-KO clones (cl. 1 and cl. 2). NanoLuc signal was evaluated after 72 and 96 hours of growth in standard conditions and normalized to cell viability. NanoLuc signal from MLH1 KO clones was then compared with signal detected in MLH1 wild-type cells (CTR). Results represent means ± SD (n = 4). *p < 0.05; **p < 0.01; ***p < 0.001 (Student’s t test). (D) DiFi, WiDr, and NCIH508 CRC cells were treated as indicated. NanoLuc signal was normalized to cell viability. NanoLuc signal from treated cells was then compared with signal detected in untreated (NT) cells. Results represent means ± SD (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001 (Student’s t test).

We next used the CA-NanoLuc system to measure the impact of ectopic inactivation of MMR in CRC cells. To this end, we used CRISPR-CAS9 to inactivate the MLH1 gene in the HT29 human CRC cell line. After the isolation of two independent MLH1 knockout (KO) clones (fig. S13, A and B), they were transduced with the CA-NanoLuc vector. MLH1 KO clones exhibited higher levels of NanoLuc signal as expected, confirming that the assay can detect inactivation of DNA MMR (Fig. 4C). Next, drug-dependent (transient) MMR down-regulation was evaluated. EGFR and BRAF inhibition led to time-dependent increases of bioluminescence (Fig. 4D), paralleling the down-regulation of DNA-repair effectors and the up-regulation of low-fidelity polymerases. We further found that permanently resistant derivatives no longer exhibited adaptive mutability in response to targeted therapies (fig. S14).

Genomic alterations in CRC cells upon treatment with targeted therapies

To determine whether molecular evidence of adaptive mutability was present in the genome of CRC cells treated with EGFR and BRAF inhibitors, we analyzed whole-exome sequencing (WES) data from DiFi and WiDr parental, persister, and drug-resistant derivative cells. The overall mutational burden (i.e., the number of mutations per megabase) of persisters and the drug-resistant cell population was only marginally affected (fig. S15, A and B). As a control, we assessed whether MMR permanent inactivation induced by MLH1 KO affected the mutational burden of HT29 CRC cells and found that it was only marginally affected (fig. S16A).

Given these results, we changed our approach. Because treatment with targeted therapies led to a transient MMR-deficient phenotype, we reasoned that MMR status could be more easily detected by examining microsatellite regions, where DNA replication slippage errors occur frequently and are ineffectively repaired in the absence of MMR. Indeed, WES analysis unveiled alterations in microsatellite regions of HT29 in which the MLH1 gene was genetically knocked out (fig. S16, B and C). We also detected increased genetic instability in the microsatellite regions of CRC cells made resistant to targeted agents (Fig. 5, A and B), as shown by a shift in the length of microsatellite regions, highlighting the impact of targeted therapies on the DNA-repair process and mutagenicity. To detect the occurrence of microsatellite alterations in nonclonal cell populations, we utilized a high-depth capture panel that detects hotspot somatic variants and shifts in the length of microsatellite regions. Indeed, such high-sensitivity analysis unveiled a significant shift in the length of microsatellite regions in both persister and drug-resistant cells (Fig. 5C and fig. S17).

Fig. 5 Adaptive mutability leads to genetic instability in CRC cells in response to therapy-induced stress.

(A) Percentage of unstable microsatellite regions in DiFi and WiDr persister and resistant cells compared with their parental counterpart (CTRL). (B) Length distribution of one representative microsatellite region for drug-resistant DiFi and WiDr cell lines. ***p < 0.001 (χ2 test). (C) Number of unstable microsatellite sites detected by NGS-based high-depth capture panel in WiDr cells (parental) treated with cetuximab + DAB for 14 days (persisters) and at resistance. (D) DNA was collected from one vehicle-treated and one cetuximab-resistant PDX. Percentage of unstable microsatellite regions of the tumor collected from the cetuximab-resistant mouse (PDX CTX-R) compared with the vehicle-treated (CTRL) mouse is reported. (E) Length distribution of one representative microsatellite region. ***p < 0.001 (χ2 test).

We next assessed the impact of targeted therapies on the genomic landscape of PDXs by studying a PDX model (CRC0078) (Fig. 2A and fig. S7D) that was continuously treated with cetuximab until it developed resistance (fig. S18). WES analysis of the cetuximab-resistant tumor tissue revealed alterations in microsatellite genomic regions that were not present in the PDX tumor collected from the corresponding untreated mouse (Fig. 5, D and E). Overall, these results indicate that CRC cells and a CRC PDX model exposed to targeted therapies experience loss of replication fidelity in regions of nucleotide repeats.


The development of resistance has emerged as a major limitation of targeted therapies directed against oncoproteins such as EGFR, BRAF, and ABL (25).

In this study, we tested the hypothesis that cancer cells treated with targeted therapies activate stress-induced mutagenic mechanisms. We found that persister (drug-tolerant) cancer cells that survive EGFR and/or BRAF inhibition exhibit DNA damage, down-regulate mismatch and HR repair proteins, switch from high-fidelity to error-prone–mediated repair of DNA damage, and transiently increase their mutagenic ability.

Stress-induced mutagenesis is a characteristic trait of unicellular organisms to transiently accelerate genetic diversity in a fraction of the population when encountering a hostile environment. (16). Indeed, we found that therapy-induced modulation of DNA repair in cancer cells is also transitory and reverts back once a mutational landscape able to restore the ability to grow in the presence of the drug is achieved. We postulate that in cells of multicellular organisms, stress-induced mutagenesis is not operational. However, in cancer cells that have lost tissue-imposed homeostasis—and in many ways operate like unicellular organisms—this ancestral program is still available and is unleashed by oncoprotein-targeted drugs. A similar process has also been observed in cancer cells undergoing hypoxia-driven stress (7, 45, 46).

The analysis of mutational signatures has emerged as a valuable tool with which to document the mutational processes operative in cells (47). In future studies, it will be interesting to establish whether specific mutational signatures emerge under targeted therapies. Resolving such processes, which we postulate occur transiently in small cell subpopulations, is likely to require extensive genomic comparisons of multiple clones and independent data points.

These results may have clinical implications. The knowledge that cancer cells under therapeutic stress down-regulate key effectors of the DNA-repair machinery, such as MMR and HR, exposes a vulnerability that could be clinically exploited. For example, it will be important to assess whether down-regulation of HR proteins confers sensitivity to poly-ADP-ribose polymerases (PARP) inhibitors as observed in HR-deficient cancers (4850). Moreover, pharmacological or genetic interference could be deployed to curb the cellular mechanisms that initiate drug-driven adaptive mutagenesis with the goal of reducing the generation of new variants during treatment. This strategy could potentially increase and prolong the clinical efficacy of targeted therapies.

Supplementary Materials

Materials and Methods

Figs. S1 to S18

Tables S1 and S2

References (5161)

References and Notes

Acknowledgments: We thank members of Molecular Oncology Laboratory at Candiolo Cancer Institute for critically reading the manuscript, A. Cassingena and F. Sassi for their help with experiments, and G. McKenzie for the NanoLuc plasmid design. Funding: This research was supported by Fondazione AIRC under the 5 per Mille 2018 ID 21091 Program (P.I. A. Bardelli, G.L. S. Siena, G.L. A. Bertotti, G.L. L. Trusolino, G.L. F. Di Nicolantonio); AIRC 2010 Special Program Molecular Clinical Oncology 5 per Mille, Project No. 9970 Extension Program (A. Bardelli); AIRC IG no. 16788 (A. Bardelli); AIRC IG 2018–ID 21923 project (P.I. A. Bardelli); H2020 grant agreement no. 635342-2 MoTriColor (A. Bardelli); Ministero Salute, Ricerca Corrente 2019 RC2019 (A. Bardelli); H2020 grant agreement no. 754923 COLOSSUS (L.T.). A. Bardelli has also received research funding from PhoreMost and NeoPhore. Author contributions: M.R. and A. Bardelli conceived the study. M.R., A.S., N.M.R., S.A., S.L., V.A., A.M. M.C., M.G., and M.C.L. conducted the experiments and analyzed data. G.C. performed NGS and bioinformatics analysis. A. Bartolini performed NGS panel-based experiments. L.N. performed RNAseq bioinformatics analysis. Z.D.N. and C.G.P. provided plasmids harboring DNA damage. A. Bertotti and L.T. analyzed and provided PDX material. I.S. performed and analyzed IHC data; A.A., A.S.-B., and S.S. provided patient samples. F.D.N. analyzed data. M.R. and A. Bardelli wrote the manuscript. A. Bardelli supervised the study. Competing interests: L.T. is a paid consultant for Eli Lilly, AstraZeneca, and Merck KGaA. Z.D.N. is an inventor on a patent (U.S. 9,938,587) covering methods and kits for determining DNA-repair capacity. A. Bardelli is a member of the scientific advisory board of NeoPhore and a shareholder of NeoPhore and PhoreMost. The other authors declare no competing interests. Data and materials availability: The FM-HCR reporter plasmids are available from Z.D.N. under a material transfer agreement. The pDRGFP and the pCBASce-I plasmids are available from AddGene under a material transfer agreement. The NanoLuc-expressing plasmid is available from PhoreMost, Ltd. (Cambridge, UK) under a material transfer agreement. The HT29 empty and MLH1 KO cells are available from A. Bardelli (UNITO) under a material transfer agreement. RNA-sequencing and DNA-sequencing data have been deposited in ENA (from EBI; no. PRJEB28674).

Stay Connected to Science

Navigate This Article