Genetic variation in reproductive life-history traits is not correlated with estrogen-receptor positive breast cancer risk

This study finds that, contrary to life-history trade-off theories suggesting a link between reproductive traits and estrogen-receptor positive breast cancer risk, there is no significant genome-wide genetic correlation between the disease and key reproductive life-history traits such as age at menarche, age at first birth, or number of children.

The Gist

The Big Question: Is "Having a Big Family" in Your DNA Linked to Breast Cancer?

Imagine your body is a complex factory. For decades, scientists have noticed a strange pattern in how this factory runs:

• Women who start their periods early, have their first baby young, or have many children tend to have a higher risk of developing a specific type of breast cancer (called ER+ breast cancer) later in life.
• Evolutionary biologists have a theory for this: They call it "Life-History Trade-offs."

The Analogy: The "Fast Car" vs. The "Long-Lasting Car"
Think of life-history traits like driving a car.

• The "Fast Car" Strategy: Some cars are built to speed up quickly, accelerate early, and carry a heavy load (many children) right away. But because they are pushed hard early on, the engine might wear out faster or break down later (cancer).
• The Theory: The idea is that our DNA might have kept these "Fast Car" genes because they helped our ancestors survive and have lots of babies when they were young. The cost? A higher risk of the engine failing (cancer) when we get older.

What This Study Did

The authors of this paper wanted to know: Is this "Fast Car" theory actually written in our genetic code?

They didn't just look at a few specific genes; they looked at the entire instruction manual (the whole genome) of over 97,000 women. They asked:

1. Do the genes that make you start your period early also make you more likely to get breast cancer?
2. Do the genes that make you have many children also make you more likely to get breast cancer?

They used two different high-tech methods to check this:

1. The "Family Tree" Check (GREML): They looked at a massive Dutch database (Lifelines) to see how closely related women were and compared their DNA to their health records.
2. The "Global Search" Check (LD Score Regression): They took the biggest genetic studies from around the world and ran a statistical super-comparison to see if the same genes showed up for both traits.

The Surprise Result: The Connection is Missing!

The Verdict: The study found no significant link between the genes for reproductive traits and the genes for breast cancer risk.

The Analogy: The "Ghost in the Machine"
Imagine you have a library of 20,000 books (your genes).

• The Old Theory: Scientists thought that if you opened the "Reproduction" chapter, you would find a warning label saying, "Warning: This chapter also causes Cancer."
• The New Finding: The authors opened the books and looked at every single page. They found that the "Reproduction" chapter and the "Cancer" chapter are written by completely different authors. They don't seem to share any sentences.

Even though women who have more children phenotypically (in real life) have different cancer risks, it doesn't seem to be because of their genes.

Why Does This Matter?

If the genes aren't the link, what is? The paper suggests the link is environmental, not genetic.

The Analogy: The "Mismatched Wardrobe"
Think of our bodies as wearing a wardrobe designed for the Stone Age.

• The Past: In the past, women started periods later, had fewer children, and breastfed for years. This gave their bodies long breaks from estrogen (the hormone that fuels this type of cancer).
• The Present: Today, thanks to better nutrition, women start periods earlier. Cultural shifts mean women have fewer children and breastfeed for shorter times.
• The Result: Our bodies are still wearing that "Stone Age wardrobe," but we are living in a "Modern World." We are exposed to estrogen for much longer than our bodies evolved to handle. This extra exposure causes the cancer, but it's not because our DNA changed to be "bad." It's because our lifestyle changed faster than our genes could adapt.

The Takeaway

1. Evolutionary Trade-offs might not be the main culprit: While a few specific genes (like BRCA1/2) might follow the "Fast Car" rule, the average genetic risk for breast cancer isn't shaped by a trade-off with having children.
2. It's not "Bad Genes": You don't need to worry that having a large family is "genetically programmed" to give you cancer.
3. It's about Lifestyle: The real driver is likely the modern environment (diet, timing of birth, breastfeeding habits) creating a mismatch with our ancient biology.

In short: The paper tells us that the link between reproduction and breast cancer is likely a story of how we live today, not a story of what we inherited from our ancestors.

Technical Summary

1. Problem Statement

The study addresses a central question in evolutionary medicine: Why do genetic variants predisposing individuals to life-threatening diseases like Estrogen-Receptor Positive Breast Cancer (ER+ BC) persist in human populations?

• The Hypothesis: Evolutionary life-history theory and the antagonistic pleiotropy theory suggest that these risk variants may be maintained because they confer reproductive advantages earlier in life (e.g., faster maturation, earlier reproduction, higher fertility), creating a trade-off where early-life fitness benefits outweigh late-life disease costs.
• The Conflict: While specific genetic loci (e.g., BRCA1/2, p53) and some genome-wide significant variants show evidence of such trade-offs, epidemiological data presents a paradox. Phenotypically, earlier menarche increases ER+ BC risk, but earlier first birth and higher parity (number of children) decrease risk.
• The Gap: Previous studies have largely focused on individual genome-wide significant loci or failed to account for confounding factors like population stratification and assortative mating. It remains unclear whether genome-wide genetic correlations (the average shared genetic architecture) support the existence of life-history trade-offs shaping ER+ BC risk.

2. Methodology

The authors employed a multi-faceted approach using two distinct datasets and analytical methods to estimate genetic correlations ($r_g$) between ER+ BC risk and three reproductive life-history traits: age at menarche, age at first birth, and number of children born.

A. Dataset 1: Lifelines Biobank (Individual-Level Data)

• Cohort: 97,131 women of European ancestry from the Dutch Lifelines biobank.
• Phenotypes:
  • ER+ BC: Diagnoses linked to the Dutch Nationwide Pathology Databank (Palga), resulting in 182–198 cases depending on the genomic array used.
  • Life-History Traits: Self-reported age at menarche, age at first birth, and total number of children.
• Genomics: Three different genomic arrays (Global Screening Array, Affymetrix, CytoSNP) were harmonized.
• Analysis: Genomic Restricted Maximum Likelihood (GREML) using GCTA.
  • Used a Genomic Relatedness Matrix (GRM) to estimate heritability and genetic correlations.
  • Controlled for age, birth year, and population stratification (top 20 principal components).
  • Analyzed separately by array type to ensure consistency in SNP coverage.

B. Dataset 2: GWAS Summary Statistics (Population-Level Data)

• Cohort: Largest available European GWAS data.
  • ER+ BC: Michailidou et al. (2017) meta-analysis (69,501 cases, 95,042 controls).
  • Life-History Traits: Tan et al. (2024) family-based GWAS (5,671,230–6,551,047 SNPs).
• Analysis: Linkage Disequilibrium (LD) Score Regression using ldsc.
  • Family-Based GWAS: Utilized data that conditions on parental genotypes to isolate direct genetic effects, removing confounding from assortative mating and population stratification.
  • Population GWAS: Repeated analysis using standard population data to check for power differences and potential underestimation of effects.

3. Key Results

Heritability Estimates

• Lifelines (GREML): Significant heritability was found for age at menarche ($h^2 \approx 43\%$) and age at first birth ($h^2 \approx 24\%$). Heritability for ER+ BC risk ($11.8\%$) and number of children ($12.8\%$) was estimated but not always statistically significant in the primary array, likely due to low case numbers.
• GWAS (LD Score): Confirmed significant heritability for ER+ BC ($10.7\%$) and number of children ($8.6\%$). Age at menarche showed lower heritability ($14.1\%$) in family-based data compared to Lifelines, while age at first birth showed negligible heritability ($1.6\%$) in family-based data, suggesting previous estimates may have been inflated by environmental confounding.

Genetic Correlations ($r_g$)

Across all analyses (Lifelines GREML, Family-based LDSC, and Population LDSC), the study found no statistically significant genetic correlations between ER+ BC risk and any of the three life-history traits.

• Lifelines GREML:
  • Age at menarche: $r_g = 0.05 \pm 0.17$ ($p > 0.016$)
  • Age at first birth: $r_g = 0.10 \pm 0.24$ ($p > 0.016$)
  • Number of children: $r_g = 0.10 \pm 0.35$ ($p > 0.016$)
• Family-Based LDSC (Most precise):
  • Age at menarche: $r_g = 0.079 \pm 0.15$ ($p = 0.602$)
  • Age at first birth: $r_g = -0.052 \pm 0.178$ ($p = 0.764$)
  • Number of children: $r_g = -0.010 \pm 0.037$ ($p = 0.795$)
• Population LDSC: Results were consistent with family-based data, showing correlations near zero (e.g., number of children $r_g = -0.0063$).

The 95% confidence intervals allowed the authors to rule out moderate-to-strong genetic correlations (e.g., ruling out $r_g > 0.17$ or $<-0.39$ for age at menarche).

4. Key Contributions

1. Resolution of the Phenotypic-Genetic Paradox: The study demonstrates that while phenotypic associations exist between reproductive traits and ER+ BC risk (e.g., more children = lower risk), these associations are not driven by shared genetic architecture across the genome.
2. Methodological Rigor: By combining individual-level biobank data with family-based GWAS summary statistics, the study effectively controls for population stratification and assortative mating, providing a more robust test of evolutionary hypotheses than previous studies.
3. Refutation of Genome-Wide Antagonistic Pleiotropy: The findings suggest that while specific high-effect loci (like BRCA1/2) may exhibit antagonistic pleiotropy, this pattern does not hold on average across the polygenic architecture of ER+ BC risk.
4. Evolutionary Implications: The lack of genetic correlation implies that ER+ BC risk variants are likely not under strong directional selection driven by reproductive trade-offs in contemporary European populations. Instead, their persistence may be better explained by mutation accumulation (weak selection at older ages) rather than antagonistic pleiotropy.

5. Significance and Future Directions

• Evolutionary Medicine: The results challenge the broad application of life-history trade-off theories to ER+ BC. They suggest that the observed epidemiological links between reproduction and cancer are largely environmental (e.g., changes in diet, breastfeeding duration, and cultural shifts in family size) rather than genetic. This supports the "evolutionary mismatch" hypothesis, where modern environments expose women to estrogen levels or reproductive patterns their genomes did not evolve to handle.
• Clinical and Public Health: Understanding that these risks are not genetically linked to reproductive success implies that genetic counseling regarding reproductive choices based on cancer risk may need to focus on environmental modifiers rather than inherent genetic trade-offs.
• Future Research: The authors recommend:
  • Investigating local genetic correlations (e.g., using $\rho$-HESS) to detect trade-offs at specific genomic loci that are masked in genome-wide averages.
  • Examining gene-environment interactions and gene networks related to estrogen metabolism.
  • Extending these analyses to non-European populations and other breast cancer subtypes (e.g., ER- BC).

In conclusion, the paper provides robust evidence that genetic variation in reproductive life-history traits does not overall shape the risk of ER+ breast cancer, indicating that the persistence of ER+ BC risk variants is likely due to mutation accumulation rather than a genome-wide trade-off with reproductive fitness.