The stakes could hardly be higher. Brain tumors remain among the most lethal malignancies in medicine, characterized by dismal prognoses and stubborn resistance to conventional therapies, while neuroblastoma—a cancer of the sympathetic nervous system that strikes mainly in early childhood—poses its own formidable treatment challenges. Patients undergoing intensive chemotherapy for these diseases are profoundly immunosuppressed, and bacterial infections during treatment are common, feared, and frequently fatal. Antibiotic prophylaxis and empiric treatment are therefore standard practice, not optional extras. Clinical studies have documented significant rates of bacterial and fungal infection during induction chemotherapy for high-risk neuroblastoma, and febrile neutropenia in solid tumor patients is routinely managed with broad-spectrum cephalosporins like cefazolin’s relatives. If those protective drugs simultaneously erode the efficacy of the anticancer regimen, the trade-off at the bedside becomes far more complicated than anyone assumed.

The Turkish team, led by İlayda Şişli together with Melek Yüce and Esra Albayrak, set out to probe precisely this dual role. Their question was deceptively simple: what happens to cancer cells when a widely used antibiotic and a chemotherapy drug share the same culture medium? Cefazolin, marketed in some countries as Cefamezin, is a beta-lactam antibiotic that works by disrupting bacterial cell wall synthesis—utterly irrelevant machinery for human cells, which have no cell walls. Yet a accumulating literature suggests cephalosporins are not pharmacologically inert toward tumor cells. Some studies have reported that certain cephalosporins can suppress cancer growth, for example by triggering ferroptosis, an iron-dependent form of cell death, in nasopharyngeal carcinoma cells, or by targeting Aurora B kinase in lung cancer models. Others have found the opposite: ceftazidime and cefepime were shown to antagonize the effect of 5-fluorouracil in colon cancer cells. The picture is inconsistent, drug-specific, and cell-type-specific—which is exactly why direct experimental testing matters.

To interrogate the interaction, the researchers used two complementary experimental readouts. The first was the MTT assay, a colorimetric test in which metabolically active cells convert a yellow tetrazolium compound into a purple formazan product; the amount of purple dye serves as a proxy for the number of viable, metabolically competent cells. The second was flow cytometry, a technique that fires cells one by one through a laser and quantifies their DNA content and fluorescent markers, allowing researchers to determine what fraction of the cell population sits in each phase of the cell cycle—G0/G1, S, G2/M—and what fraction has fallen into the sub-G0/G1 debris characteristic of dying cells. Together, these methods capture both the net effect on cell survival and the mechanistic fingerprint of how the drugs act on the cell division machinery.

The results were clear on both fronts. Irinotecan, a topoisomerase I inhibitor that works by trapping the enzyme responsible for relieving torsional stress in DNA during replication, thereby converting an essential enzyme into a DNA-damaging poison, induced significant cytotoxicity in both LN-18 glioblastoma and SH-SY5Y neuroblastoma cells. Flow cytometry showed the expected signature: an increase in the sub-G0/G1 fraction, indicating cells with fragmented DNA heading toward death, alongside arrest in the S phase, where DNA replication stalls, and the G2/M phase, where cells with damaged DNA halt before mitosis. This is textbook topoisomerase poison behavior—the drug prevents cells from copying and dividing their genome cleanly, and the damaged cells accumulate and die.

Then came the twist. When cefazolin was added alongside irinotecan, the chemotherapy’s grip on the cells loosened. In the SH-SY5Y neuroblastoma cells, co-treatment with the antibiotic attenuated both the cytotoxicity and the cell cycle disruption that irinotecan alone had produced; the drug’s ability to push cells into the lethal sub-G0/G1 compartment and to arrest them in S and G2/M was measurably blunted. In the LN-18 glioblastoma cells, the pattern was similar but not identical: cefazolin partially weakened irinotecan’s inhibition of cell viability, while necrosis rates—death by a messier, non-programmed route—were not significantly altered. In both cell lines, the direction of the interaction was the same, and it was the wrong direction for a cancer patient: the antibiotic softened the chemotherapy’s blow.

What might explain this antagonism at the molecular level? The study does not pin down a mechanism, but the authors and the surrounding literature point to several plausible candidates. Antibiotics and chemotherapeutics can compete for cellular transport systems, altering how much drug actually reaches its intracellular target—a phenomenon first documented more than fifty years ago, when researchers showed that certain antibiotics interfered with the cellular transport and antitumor activity of methotrexate in leukemia cells. Antibiotics can also modulate drug efflux pumps such as BCRP/ABCG2, the molecular bouncers that expel chemotherapy agents from cells; indeed, one antibiotic, novobiocin, is famous precisely because it reverses such resistance. More speculatively, cefazolin could be altering mitochondrial function, reactive oxygen species generation, or the metabolic state of the tumor cells in ways that buffer them against topoisomerase-induced stress. Disentangling these possibilities will require the kind of mechanistic follow-up this in vitro study deliberately leaves open.

The broader context makes the finding more than a laboratory curiosity. A substantial clinical literature has already linked antibiotic exposure to worse cancer outcomes, most prominently with immune checkpoint inhibitors, where antibiotic use around the time of treatment has been associated with reduced clinical activity in advanced renal cell and non-small-cell lung cancer—largely attributed to disruption of the gut microbiome. A retrospective cohort study in glioblastoma patients has examined whether antibiotic drug use affects outcome and therapy-related toxicity, and reviews have framed antibiotics as a double-edged sword in oncology: indispensable against infection, potentially harmful against the tumor. Drug–drug interactions in cancer patients treated with oral anticancer agents are common and often overlooked, and fatal adverse drug events remain a sobering reminder that polypharmacy in oncology is a minefield. The new study extends this concern from the microbiome and the immune system down to the level of direct pharmacological interference between a cephalosporin and a topoisomerase inhibitor inside the tumor cell itself.

The researchers are careful about what their data can and cannot say. This was an in vitro study: two immortalized cell lines, culture dishes, and defined drug concentrations. Real glioblastomas and neuroblastomas live behind the blood–brain barrier, embedded in a microenvironment that no monolayer culture reproduces, and the concentrations and timing of cefazolin exposure in a treated patient differ from those in a flask. The authors explicitly note that their findings suggest cefazolin may antagonize the cytotoxic effects of irinotecan in vitro and underscore the need for further validation in in vivo and clinically relevant models. Animal studies and, ultimately, carefully designed clinical pharmacology studies would be needed before anyone could say whether a patient receiving cefazolin during irinotecan-based therapy actually experiences a meaningful reduction in tumor control. It is also worth remembering that irinotecan is not a mainstay of standard glioblastoma care, so the immediate clinical implications for current treatment protocols are limited; the finding speaks more to the principle than to a specific prescription today.

Even so, the message lands with force. Antibiotics are among the most frequently co-administered drugs in oncology, handed out for prophylaxis, for fever during neutropenia, and for intercurrent infections across every tumor type and every age group, from children with neuroblastoma to adults with glioblastoma. If a drug as ordinary as cefazolin can measurably weaken a chemotherapy agent in the dish, then the possibility deserves systematic scrutiny rather than assumption of safety. The study, funded by the Ondokuz Mayıs University Scientific Research Project Office and conducted with cell lines provided by Yeditepe University in Istanbul, is a small piece of evidence—but it points at a large blind spot. As cancer therapy grows ever more complex, the drugs given to protect patients from infection must be evaluated as potential players in the treatment itself, not as pharmacological bystanders. The next experiments, in animal models and eventually in patients, will determine whether this laboratory warning becomes a clinical rule.

Subject of Research: Interaction between the antibiotic cefazolin and the chemotherapeutic drug irinotecan in glioblastoma and neuroblastoma cells

Article Title: Antibiotic-chemotherapy interaction in brain tumors: cefazolin reduces chemotherapeutic efficacy of irinotecan in glioblastoma and neuroblastoma cells

Article References: Şişli, İ., Yüce, M., & Albayrak, E. (2026). Antibiotic-chemotherapy interaction in brain tumors: cefazolin reduces chemotherapeutic efficacy of irinotecan in glioblastoma and neuroblastoma cells. Medical Oncology, 43(11), Article 302. https://doi.org/10.1007/s12032-026-03403-6

Image Credits: AI Generated

DOI: 10.1007/s12032-026-03403-6

Keywords: cefazolin, irinotecan, glioblastoma, neuroblastoma, antibiotics, chemotherapy, drug interactions, cell cycle, flow cytometry, MTT assay, topoisomerase inhibitor, cancer pharmacology

The concept of BRCAness has dominated precision oncology discussions of DNA damage response defects for more than a decade. Tumors with loss of homologous recombination repair, the high-fidelity pathway that accurately fixes double-stranded DNA breaks, become exquisitely vulnerable to agents that create or exploit such breaks, including platinum drugs like oxaliplatin and cisplatin and PARP inhibitors such as olaparib. This logic works well for tumors with BRCA1 or BRCA2 mutations. However, ATM and CHEK2 mutations tell a different story. ATM, the gene encoding ataxia-telangiectasia mutated kinase, sits at the apex of a signaling cascade that coordinates the cellular response to DNA damage, orchestrating cell cycle checkpoints and the recruitment of repair machinery. CHEK2, encoding the checkpoint kinase 2 that acts downstream of ATM, is part of the same pathway. Crucially, defects in this pathway do not necessarily abolish homologous recombination repair itself, which means that tumors carrying ATM or CHEK2 mutations may retain enough repair capacity to withstand platinum and PARP-inhibitor therapy.

The research team, led by investigators at the Medical College of Wisconsin’s LaBahn Pancreatic Cancer Program together with collaborators at Wayne State University’s Karmanos Cancer Institute, approached the problem from two complementary directions. First, they mined an institutional real-world database to identify patients with advanced or metastatic pancreatic ductal adenocarcinoma whose tumors harbored ATM or CHEK2 mutations, assembling a cohort of 16 patients whose treatment histories spanned 48 separate lines of chemotherapy. Second, they turned to the laboratory, where they used CRISPR gene editing to create a panel of isogenic PANC-1 pancreatic cancer cell lines carrying either homozygous or heterozygous knockout of ATM. The cell lines were allele-verified by whole-exome sequencing, ensuring that any differences in drug response could be attributed specifically to the degree of ATM loss rather than to background genetic variation between different cell models.

The clinical results were striking. When the researchers compared the outcomes of second-line treatments for these ATM or CHEK2-mutant patients, irinotecan-containing regimens outperformed every other therapeutic category, including the platinum-based and PARP-inhibitor-containing combinations that the BRCAness framework would have predicted to be most effective. The median progression-free survival on irinotecan-based second-line therapy was 11.5 months, compared with just 3 months for the other treatment types, a difference that reached statistical significance with a p value below 0.001. In a disease where median survival is typically measured in months and treatment options after first-line failure are limited, an eight-and-a-half-month difference in disease control represents a clinically meaningful advance for this molecularly defined subgroup.

The laboratory experiments provided a mechanistic explanation for this clinical observation. The researchers exposed their isogenic cell line panel to SN-38, the active metabolite of irinotecan, which the liver generates from the prodrug in patients. SN-38 works by poisoning topoisomerase I, an enzyme that relieves torsional stress in DNA by creating transient single-strand breaks during replication and transcription. When topoisomerase I is inhibited, the enzyme becomes trapped on DNA as covalent topoisomerase I-DNA adducts, and the collision of these adducts with advancing replication forks generates double-stranded breaks that require an intact DNA damage response to resolve. Preliminary work had suggested that certain forms of ATM pathway dysfunction might prevent the removal of these topoisomerase I-DNA adducts, allowing the lesions to persist and accumulate into lethal genomic instability.

The CRISPR-edited panel confirmed this dose-dependent relationship with remarkable precision. In colony formation assays, the homozygous ATM knockout cells showed a half-maximal inhibitory concentration, or IC50, of just 0.3 nanomolar for SN-38, while the heterozygous knockout cells showed an IC50 of 0.8 nanomolar and the wild-type controls an IC50 of 7 nanomolar. In other words, complete loss of ATM rendered the cells more than twenty-fold more sensitive to the drug than unedited cells, and even the loss of a single ATM copy, which better models the heterozygous mutations frequently seen in patient tumors, conferred nearly nine-fold sensitization. Proliferation and viability assays told the same story, with all comparisons reaching statistical significance at p values below 0.01. The graded response across the isogenic panel demonstrates that the degree of ATM loss directly determines the degree of topoisomerase I inhibitor sensitivity.

These findings carry substantial implications for the clinical management of pancreatic cancer. Currently, comprehensive genomic profiling of pancreatic tumors routinely identifies ATM and CHEK2 mutations, and many oncologists extrapolate from the BRCA literature to offer platinum chemotherapy or PARP inhibitors to these patients. The new data suggest that this extrapolation may be misguided for a meaningful proportion of the ATM pathway-mutant population, since these tumors may not exhibit homologous recombination repair deficiency and may therefore not respond consistently to those agents. Instead, the study proposes that topoisomerase I inhibition exploits a distinct vulnerability: the inability of ATM-defective cells to process the stalled topoisomerase I-DNA complexes that SN-38 generates. Irinotecan is already an established component of the liposomal irinotecan combination used in second-line metastatic pancreatic cancer, which makes the prospect of biomarker-guided deployment particularly practical.

The study’s design also highlights the growing value of pairing real-world evidence with engineered laboratory models. The clinical cohort, while modest in size at 16 patients and 48 chemotherapy lines, was drawn from institutional protocols approved by the Medical College of Wisconsin Institutional Review Board, including the MCW Master Predict observational program registered as NCT05802069, and reflects the actual treatment decisions and outcomes recorded in routine care. Because each patient served in part as their own comparison across different treatment lines, the analysis captures the heterogeneity of real clinical practice while still permitting a statistically robust comparison. The isogenic CRISPR-edited cell lines then provided the controlled experimental system that retrospective clinical data can never offer, isolating the effect of ATM dosage on drug sensitivity with allele-level verification.

Several caveats temper the enthusiasm. The clinical cohort was small, and retrospective real-world analyses are inherently susceptible to selection biases in how patients were assigned to different treatments. The laboratory work relied on a single pancreatic cancer cell line background, PANC-1, albeit engineered with rigor, and in vitro sensitivity to SN-38 does not guarantee that the pharmacokinetics, toxicity profile, and combination effects of irinotecan in patients will mirror the dish. The authors themselves are careful to frame their conclusions as hypothesis-generating, calling for prospective investigation of topoisomerase I inhibitors in ATM pathway-mutated pancreatic ductal adenocarcinoma rather than claiming an immediate change to standards of care. Nonetheless, the convergence of a large clinical signal, a statistically significant survival difference, and a clean dose-dependent molecular mechanism is exactly the kind of evidence that justifies launching biomarker-stratified clinical trials.

Looking forward, the study opens several avenues. Antibody-drug conjugates that deliver topoisomerase I inhibitors selectively to tumor cells are an increasingly prominent class of anticancer agents, and the mechanistic logic established here suggests that ATM pathway status could serve as a predictive biomarker for such drugs across multiple tumor types, not only pancreatic cancer. More broadly, the work challenges the field to move beyond the binary of homologous recombination proficiency and deficiency, recognizing that different DNA damage response defects create different, drug-specific vulnerabilities. For the five to ten percent of pancreatic cancer patients whose tumors carry ATM or CHEK2 mutations, the message is one of cautious optimism: a therapy already in the clinic may work far better in their specific molecular context than anyone had previously demonstrated, and prospective trials to confirm that promise may not be far behind.

Subject of Research: Sensitivity of ATM/CHEK2-mutant pancreatic ductal adenocarcinoma to topoisomerase I inhibition

Article Title: Beyond BRCAness: ATM pathway defects confer sensitivity to topoisomerase I inhibition

Article References: Kamgar, M., McFall, T., Mehdi, M., Thapa, B., Szabo, A., Ahmed, G., Davidson, R., Scheuber, G., Shreenivas, A., Thomas, J. P., Sriram, D., Evans, D. B., Tsai, S., Christians, K. K., Erickson, B., Hall, W. A., Chen, H.-Z., Lytle, N., Sarkar, N. D., … Kurzrock, R. (2026). Beyond BRCAness: ATM pathway defects confer sensitivity to topoisomerase I inhibition. Genome Medicine. https://doi.org/10.1186/s13073-026-01774-z

Image Credits: AI Generated

DOI: 10.1186/s13073-026-01774-z

Keywords: ATM, CHEK2, pancreatic cancer, irinotecan, SN-38, topoisomerase I inhibitor, DNA damage response, BRCAness, CRISPR, precision oncology, progression-free survival, homologous recombination repair