‹ Back

When the Medicine Crosses Over: What a New Study in Zebrafish Tells Us About Antidepressants and Pregnancy

I'll be honest. I don't always read every paper I contribute to with the same emotional weight. This one (The sertraline metabolite, desmethylsertraline, may be implicated in adverse outcomes reported after gestational sertraline use: insights from a study in zebrafish) was different.

My wife was recently pregnant, which meant that a study about what antidepressants do to a developing baby hit considerably closer to home than usual.

And then there's the other irony: my wife is a passionate aquarist who keeps a thriving tank of GloFish danios, the very species at the centre of this research. When you've spent time staring down a mass spectrometer at zebrafish proteins, and then come home to a glowing tank of their cousins on the living room shelf, science has a way of becoming very personal, very quickly.

 

The Difficult Reality of Depression in Pregnancy

More than one in ten pregnant women worldwide experiences clinical depression: a serious medical condition that, left untreated, can lead to reduced self-care, difficulty following prenatal guidelines, and in severe cases, risk to both mother and baby.

Among the most commonly prescribed antidepressants in pregnancy is sertraline, an SSRI with a relatively well-established safety profile. Yet studies have suggested a possible link between gestational SSRI use and certain heart defects and neurodevelopmental differences in exposed children. The question researchers have been struggling to answer is: what exactly is causing the risk?

Having recently watched my wife navigate pregnancy, I found myself engaging with this paper not just as a scientist, but as a father.

 

A Metabolite in the Spotlight

When you swallow sertraline, your liver breaks it down into other compounds called metabolites. One of the most important is desmethylsertraline (DES). Research has shown that in umbilical cord blood, DES is typically present at concentrations roughly twice as high as the original sertraline, meaning the developing baby is exposed to far more of the metabolite than the drug the mother actually took. This has been measured directly in umbilical cord blood. By the time sertraline has travelled through a pregnant woman's body and reached her baby, it is DES, not sertraline, that dominates the picture.

And yet almost all research into sertraline's safety in pregnancy has focused on the parent drug. A new study published in Pharmacological Reports, led by Professor Carine Smith and Dr Tracy Kellermann at Stellenbosch University's Faculty of Medicine and Health Sciences, set out to change that. My contribution to the study was the proteomics analysis: the mass spectrometry work that forms the heart of its most novel findings, carried out during my time at Stellenbosch University's Central Analytical Facility before I established Trace Labs as an independent proteomics and protein chemistry service in October 2024.

 

Enter the Zebrafish: Including the Ones in Our Living Room

My wife keeps GloFish danios: genetically engineered zebrafish available in six fluorescent neon colours, namely Starfire Red, Sunburst Orange, Electric Green, Cosmic Blue, Galactic Purple, and Moonrise Pink. Under aquarium lighting, a school of them looks less like a fish tank and more like a living aurora.

Their origins are a fascinating story. The wild-type zebrafish they're based on are native to rivers in India and Bangladesh. In 1999, Singaporean scientists extracted a green fluorescent protein from jellyfish and inserted it into zebrafish embryos, not for aesthetics, but to create fish that would glow visibly in the presence of environmental pollutants. The team later developed red fluorescent fish using a gene from sea coral, and orange-yellow fish from a variant jellyfish gene. The aquarium trade eventually adopted them, and the GloFish line was born.

There is something quietly poetic about this: fish originally engineered to detect environmental toxins, glowing in our living room, while their wild-type cousins yielded the protein data I was analysing in the laboratory. My wife has always loved her danios for their beauty. I now look at them and think about what their species has contributed to human medicine.

Zebrafish share around 70% of their genetic material with humans, their embryos develop transparently, and they absorb substances directly through the water they swim in, making precise, developmentally-timed dosing straightforward. The team exposed larvae to concentrations of sertraline and DES calculated to match levels actually measured in human umbilical cord blood, from four hours after fertilisation until about five days old: one of the most critical windows of early development.

 

What We Found

The study measured behaviour, serotonin transporter (SERT) protein levels, and the full proteome of exposed larvae. The proteomics portion was my contribution, using label-free quantitative liquid chromatography-mass spectrometry. Proteomics essentially takes a molecular snapshot of an entire organism, revealing which biological processes are being turned up or down. It's painstaking, data-intensive work, but when it lands, it gives you a remarkably complete picture of what a substance is actually doing inside a living system.

Both sertraline and DES caused reduced movement and significantly lower SERT levels in the larvae, confirming that the doses used were achieving a genuine, physiologically relevant SSRI-like effect.

But the proteome told a striking story. Sertraline caused no measurable changes in protein expression whatsoever. DES, however, produced disruptions across hundreds of proteins, and both the zebrafish and human protein databases I interrogated pointed to the same organ systems: the heart and the developing brain.

Among the most concerning findings were significant reductions in MYL2, a protein critical to heart muscle contraction that has been linked to heart failure in humans; disruption to the VEGF pathway governing blood vessel development; and patterns associated with cardiomyopathy. Mitochondrial stress pathways, tryptophan metabolism (the pathway used to produce serotonin), and proteins involved in bone and cartilage development were also affected, potentially explaining musculoskeletal malformations previously reported in some SSRI-exposed children. Changes in estrogen biosynthesis proteins suggest female offspring may face disproportionate risk.

 

What This Means, and What It Doesn't

The headline finding is this: at concentrations found in umbilical cord blood, it is DES, not sertraline itself, that disrupts the developing proteome. This is a significant reframing. Future research into gestational SSRI safety may need to train its focus firmly on the metabolite, not the parent drug.

This does not mean pregnant women taking sertraline should stop their medication. Untreated gestational depression carries serious, well-documented risks of its own. Every medication decision in pregnancy involves a careful balance of risks and benefits, made with a healthcare provider. What this research does is add important nuance to that conversation, pointing toward specific molecular pathways that could, in future, become targets for protective intervention or clinical monitoring.

The close alignment between what we observed in zebrafish larvae and what has been reported in human cohorts also validates this model as a powerful tool for investigating gestational drug exposure, far faster and more detailed than waiting decades for human studies to accumulate.

My wife now looks at her tank of glowing danios with a slightly different expression than she used to. Those Cosmic Blues and Galactic Purples darting through the water have always been more than decoration. It turns out they've been, in a very real sense, standing in for us.

 

This research was published in Pharmacological Reports (2025) and conducted at Stellenbosch University's Faculty of Medicine and Health Sciences. The proteomics analysis was performed by Maré Vlok, founder and owner of Trace Labs, an independent proteomics and protein chemistry service established in October 2024.