Why Canagliflozin and Estradiol Extend Male Lifespan but Shorten Female Lifespan in Mice
ITP pathology study finds no clear histological explanation for why two lifespan-extending drugs harm female mice, suggesting cryptic sex-specific toxicity.
Summary
The NIA Interventions Testing Program tested two agents — 16α-hydroxyestradiol (OH_Est) started at 12 months and canagliflozin started at 16 months — in genetically heterogeneous UM-HET3 mice. Both drugs extended male median lifespan by roughly 15% but reduced female lifespan by 6–7%. To understand why, researchers performed blinded end-of-life histopathology on 164 mice across three sites. Neoplasia was the dominant cause of death in all groups (72% overall). Neither drug significantly shifted cause-of-death patterns in either sex. Incidental lesion analysis revealed complex, sex-specific changes — including effects on liver, reproductive organs, and adrenal glands — but no coherent pattern explaining female harm. The authors conclude that whatever is killing female mice faster is likely not detectable by standard histopathology, pointing toward functional or molecular toxicity rather than gross tissue damage.
Detailed Summary
Two drugs tested by the NIA-funded Interventions Testing Program (ITP) produced a striking and puzzling result: both 16α-hydroxyestradiol (OH_Est, begun at 12 months) and canagliflozin begun at 16 months (Cana_16) extended median lifespan in male UM-HET3 mice by approximately 15%, yet each simultaneously reduced median lifespan in female mice — OH_Est by 7% and Cana_16 by 6%. This study was designed to determine whether end-of-life histopathology could explain this sex-specific divergence. The investigators examined 164 mice (30 female controls, 15 male controls, 29 female and 29 male OH_Est-treated, 30 female and 31 male Cana_16-treated) from three geographically distributed test sites: The Jackson Laboratory, University of Michigan, and UT Health San Antonio.
Mice were maintained until natural death or euthanasia at moribund state. Fixed tissues were processed for hematoxylin and eosin histology and evaluated by a blinded board-certified veterinary pathologist. A comprehensive panel of organs was examined, including brain, lungs, liver, kidneys, heart, pancreas, spleen, reproductive organs, adrenal glands, and thyroid. Lesions were scored as present/absent or on graded severity scales. Special stains (Congo red, Masson's trichrome) were applied to a subset of female reproductive tracts from OH_Est and control mice. Cause of death was assigned and incidental lesions tabulated. Statistical analysis used two-factor ANOVA with sex, treatment, and interaction terms.
Neoplasia was overwhelmingly the leading cause of death, accounting for 72% of all mice — 77% of females and 65% of males. The most common fatal neoplasms in females were hematopoietic neoplasia, mammary carcinoma, hemangiosarcoma, and pituitary neoplasia; in males, lung carcinoma and hepatocellular carcinoma were prominent. Critically, neither OH_Est nor Cana_16 produced statistically significant shifts in cause-of-death distributions in either sex. The authors note that statistical power for cause-of-death comparisons was low given sample sizes, which limits confidence in null findings.
Incidental lesion analysis — combining lethal and non-lethal pathological changes — revealed a complex mosaic of drug effects. Both agents produced significant or near-significant changes in multiple lesion categories, and some of these effects were sex-specific in direction or magnitude. For example, drug-associated changes were observed in hepatic vacuolation, reproductive organ pathology (including uterine cystic endometrial hyperplasia and ovarian degeneration in females, testicular changes in males), and adrenal findings. However, no single lesion or pattern of lesions emerged that could coherently account for the shortened female lifespan seen with either drug. Importantly, some effects were absent or reversed between sexes, reinforcing the idea that the two drugs act through sex-dependent biological pathways.
The investigators had originally selected OH_Est based on evidence that 17α-estradiol — which preferentially extends male lifespan — undergoes sex-specific metabolism in male mice to estriol-related sulfated compounds. The hypothesis was that providing OH_Est directly might confer benefit in both sexes. Instead, female mice were harmed. Similarly, earlier-start canagliflozin (at 6–7 months) had not significantly affected female lifespan, whereas the late-start protocol (16 months) produced measurable female mortality acceleration. Why late-start canagliflozin harms females but earlier-start does not remains unexplained by this pathology dataset.
The authors conclude that the female-specific harm induced by both agents likely reflects functional or biochemical toxicity that is not readily captured by conventional histopathological endpoints. Possible explanations include metabolic dysregulation, hormonal interference, or organ-level dysfunction that does not produce gross or microscopic structural lesions detectable at end of life. This study underscores an important methodological limitation in preclinical longevity research: histopathology, while invaluable, may be insufficient to explain sex-divergent lifespan effects. Future work integrating molecular, metabolomic, and functional endpoints at intermediate timepoints will likely be needed to resolve this puzzle — and to inform whether SGLT2 inhibitors and estrogenic compounds carry analogous sex-specific risks in aging humans.
Key Findings
- OH_Est (16α-hydroxyestradiol, started at 12 months) extended male median lifespan by ~15% but reduced female median lifespan by 7% in UM-HET3 mice
- Canagliflozin started at 16 months (Cana_16) extended male median lifespan by ~15% but reduced female median lifespan by 6%
- Neoplasia was the cause of death in 72% of all 164 mice examined — 77% of females and 65% of males — with no significant drug-induced shift in cause-of-death distribution in either sex
- Hematopoietic neoplasia was the most common lethal tumor overall (43 of 164 mice), especially prominent in females; lung carcinoma and hepatocellular carcinoma dominated in males
- Incidental lesion analysis revealed significant or near-significant sex-specific changes in hepatic, reproductive, and adrenal pathology for both drugs, but no coherent pattern explaining female-specific harm
- Earlier-start canagliflozin (6–7 months) had not significantly affected female lifespan, indicating that the timing of intervention — not just the drug itself — determines sex-specific outcomes
- Special staining (Congo red, Masson's trichrome) of female reproductive tracts from OH_Est-treated mice did not reveal a clear structural explanation for accelerated female mortality
Methodology
Blinded end-of-life histopathological study of 164 UM-HET3 genetically heterogeneous mice from three ITP sites (TJL, UM, UT San Antonio), encompassing control, OH_Est (5 ppm diet from 12 months), and Cana_16 (180 ppm diet from 16 months) groups, with sample sizes of 29–31 per treated group and 15–30 controls. A board-certified veterinary pathologist scored a comprehensive organ panel for graded lesion severity (0–4 scales) plus special stains on a subset of female reproductive tissues; a second pathologist reviewed vascular and adrenal neoplasms for consensus. Two-factor ANOVA with sex, treatment, and interaction terms was used; no correction for multiple comparisons was applied, and statistical power for cause-of-death comparisons was explicitly acknowledged as limited.
Study Limitations
Statistical power was low for cause-of-death comparisons given the sample sizes (~15–31 per group), making it difficult to detect modest shifts in tumor spectrum. No correction for multiple comparisons was applied to the incidental lesion analyses, increasing the risk of false-positive findings among the many lesion categories tested. The study was conducted entirely in mice (UM-HET3 F1 hybrids), and direct translation to human sex-specific risks of canagliflozin or estrogenic compounds requires independent clinical investigation; no conflicts of interest were declared.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: