Ethnicity of crucial medical cell lines misclassified

When it comes to health, race can be a double-edged sword.

On one hand, it’s a dismal fact that being in certain ethnic groups is a sentence to poorer health. For example, a 2018 study of nearly 900,000 cases of lung, breast, bowel and prostate cancer found black patients had the lowest survival rates.

Some of that miserable outlook may be socio-economic, including less social support and access to healthcare, and more risk factors such as smoking. But part of the problem is biology.

Race can influence how invasive a cancer is and how well it does with treatment. A February 2019 study, for example, found African American men were more likely to have genes that predict aggressive prostate cancer.

But, and here’s the upside, armed with a person’s ethnicity doctors can get a jump on the problem with targeted screening and treatment. It’s all part of the push for “precision medicine” that, in theory, works better because it is tailored to the individual.

A new study, however, led by Rick Kittles at the City of Hope Comprehensive Cancer Centre in Duarte, California, US, puts something of a spanner in those precision works.

The team decided to take a fresh look at a key research tool for targeted medicine: so-called “immortal” cell lines. These are cells, often sourced from human cancers, that replicate indefinitely and are crucial for modelling disease. They can be used to test, for instance, the effectiveness and toxicity of cancer drugs.

But there’s a catch.

Because ethnicity is a risk factor for some cancers, it’s essential the cell lines come from racially diverse groups to guide treatment across the divide. Kittles’ team decided to test 15 commercially available cell lines to see if their ethnicity, as claimed by suppliers, squared with their genetic ancestry.

To do that, they used methods similar to those of the genealogy companies 23andMe and Ancestry.com. For a fee, those outfits take a spit sample and tell you whether your forebears were Manchurian, Iberian or perhaps Native American.

The technique scours your DNA for so-called “ancestry informative markers” or AIMs. The markers are a bunch of genetic changes called “single nucleotide polymorphisms” (SNPs) that are heritable and, when linked to the owner’s ethnic traits, act as a genetic signature of race.

The results were, to the say the least, a mixed bag.

Cell lines labelled as White/Caucasian were mostly that; on average 97% of their ancestry was found to be European. But when it came to cell lines that purported to be Black/African American the researchers were in for a shock.

Of two lines of prostate cancer cells previously classified as Black/African American, one was found to be 99% European and the other had 92% European ancestry.

That’s a mix up with very troubling implications.

First, as reported in a commentary on the study in the publication Chemical & Engineering News (C&EN), one of those cell lines has been used in multiple studies, with one published as recently as last September, potentially impacting on their findings.

Perhaps worse, there are only a limited number of African American cell lines as it stands. The authors write that a search for healthy and cancerous breast cell lines from one US tissue bank yielded 59 Caucasian/White specimens but only 11 Black ones.

Kittles told C&EN that it was upsetting to have so few African American cell lines to start with, only to find those resources misclassified. Sean Kimbro, a co-author on the study, from North Carolina Central University, said the lack of diversity biases the science against non-European ancestry groups, although “not necessarily intentionally but just by default because the cell lines don’t include those individuals”.

How could such a staggering case of mistaken identity even happen?

The authors explain that “most commonly used human cell lines were developed at a time when self-reported race was considered a sufficient demographic detail.”

In a 2018 op-ed in the journal JAMA Surgery, cancer researchers Lisa Newman and John Carpten give more clues to why self-reported race can diverge from what the genes say.

“Despite varying degrees of admixture, individuals self-identify (or are categorised by others) as being white, black/African American, Hispanic/Latino, Asian, or Native American based on familial ties, physical attributes, culture, language, and/or geography,” they write.

But if racial self-report is wonky, the path to rectification would seem to be clear. Kittles and colleagues conclude that genetic ancestry estimates should be provided for all cell lines to avoid these kinds of errors.

But in an age where Joe Public can find his DNA relatives with a kit from Amazon why haven’t science databases screened all cell lines in the same way already? Newman and Carpten hazard a guess.

“It is possible that fears of being perceived as promoting something akin to ‘race medicine’ (the rationale for the horrific Tuskegee experiments, which left syphilis untreated in African American men from 1932 to 1972) has hindered integration of AIMs genotyping into mainstream cancer research,” they write.

In the US, at least, the medical system’s struggle with racial biases is well documented. A 2016 study, for example, found some hospital doctors (falsely) believe black people have higher pain tolerance.

All of which seems to lead to a clear take-home message: if race is to play any role in medicine at all, it must be as an objective scientific tool that crafts better, not worse health outcomes for the groups concerned.

The study appears in the journal Cancer Epidemiology, Biomarkers & Prevention.