A misleading case of “trauma inherited across generations”

Here we have a new paper in Nature Scientific Reports, accompanied by a news piece in Science, that sends a misleading message to the public, both about “inheritance of trauma” and the effects of epigenetic changes.  Both pieces are free to access; click on the first headline below to go to the news piece, and the second to go to the scientific report (its pdf is here). 

I must add that most of the “misleading” appears not in the paper but in the News piece by Andrew Curry, who suggests that trauma is inherited when in fact there’s not a scintilla of evidence for that. But the authors of the real paper don’t go to any great lengths to dispel that notion, either, and this suggestion is undoubtedly why Nature Scientific Reports found the piece clickbaity and publishable.

Note that the news piece suggests that what is inherited across three generations is trauma. That is false. What the researchers shows is that Syrian women exposed to trauma during their country’s wars have offspring and grand-offspring that inherited certain epigenetic markers in the DNA: methyl groups affixed to consistent positions in the offspring DNA.  This “epigenetic inheritance” may indeed be caused by maternal trauma, for trauma messes up the fetal environment, and since female fetuses already carry their own eggs after a few months, it could affect grandchildren at all.

But inheritance of trauma itself? NO EVIDENCE. They have no idea what the DNA positions that are methylated even do, much less that they’re in genes that affect trauma.

The situation described in both the news puffery and the paper resembles the “epigenetic” inheritance associated with the Dutch “Hunger Winter” of 1944-1945, during which a German blockade of food killed around 20,000 people in the Netherlands.  It turns out that the children of survivors who were pregnant during the famine had a higher frequency obesity, higher cholesterol, as well as higher incidences of diabetes and schizophrenia, than did children of survivors who were not pregnant. The former also lived less long, but what they inherited as not “famine”, but a panoply of diseases and conditions that may well have been the result of biochemical changes in a pregnant mother experiencing famine. These changes were certainly not adaptive, either!  However, the inheritance lasted only one generation (grandchildren of pregnant survivors were normal). PLUS, what was inherited in the famous Dutch case were conditions and behaviors, while in the present case the “trauma” appears to have caused only slight changes in the DNA sequence that had an unknown effect. There was no inheritance of trauma described at all. But look at the headline below!

The news piece:

It summarizes the scientific report this way:

Rana Dajani, a biologist at Hashemite University in Amman, Jordan, wondered whether the recent conflicts in neighboring Syria might have left traces in the epigenomes of people in the country—with implications for the health of future generations. “I wanted to ask if environmental exposure was impacting different genes,” Dajani says. “Can those changes be transferred across three generations, or more?”

To answer that, Dajani, a Jordanian researcher of Palestinian and Syrian descent, teamed up with researchers in the United States and Jordan, leveraging her family contacts to assemble a cohort of Syrian women living in Jordan. In one group were women and girls who were either pregnant or in utero themselves during the Syrian civil war that began in 2011 and had fled to Jordan. Another group included someone who was pregnant during a government-orchestrated massacre in the city of Hama in the early 1980s, her daughter and grandchildren, and other unrelated female descendants of survivors. As a control group, Dajani included Syrian families who emigrated to Jordan almost a century ago, sharing a culture with the rest of the participants but with no direct experience of violent conflict.

Biologist Dima Hamadmad, a co-author and a descendant of survivors of the Hama violence, spent hundreds of hours over the course of 5 years contacting potential participants and listening to their stories. Many of them had experienced trauma such as being severely beaten, witnessing wounded or dead bodies, or seeing someone being shot or killed. “It’s a lot of work, and the victims also deserve a lot of credit,” says Isabelle Mansuy, an epigeneticist at ETH Zürich who was not part of the research. “What they’ve done is remarkable.”

After using cheek swabs to collect DNA from more than 130 women, the team looked for patterns in DNA methylation, a process in which responses to environmental circumstances—such as trauma—add or subtract to genes chemical tags known as methyl groups that alter the gene’s function. DNA methylation is among the most studied examples of epigenetic change.

The team found that women who experienced wartime trauma directly shared such changes in 21 different spots in their genome; grandchildren in the study showed alterations in a different set of 14 sites. “We discovered a number of genes with signatures of trauma transferred across generations compared to the control group,” Dajani says. The function of the genes and proteins associated with the sites isn’t known.

Comparing those results with the surveys and interviews revealed the more wartime horrors someone experienced, the more methylation changes they seemed to have. “It doesn’t look random,” says Mulligan, who co-led the study with Dajani.

I’m prepared to believe all that, though I’m disturbed by the important control group, which is described as “Syrian families who emigrated to Jordan almost a century ago, sharing a culture with the rest of the participants but with no direct experience of violent conflict.” Well, one can debate whether a group that has been in non-warring Jordan for a century has experienced the same “culture” as Syrians who emigrated in 1980 and 2011. But others who know more about epigenetics than I have weighed in with other criticisms (see below). What was affected may not have been trauma, but just gum disease!

Click the article to read. I can’t find any description of the control group in the paper except for this—”In the control group, Syrian grandmothers and mothers lived in Jordan prior to 1980″, and it adds they were “unexposed to war,” but it doesn’t say that not that the ancestors of the control individuals been in Jordan for a century. Oh well, we’ll let that slide.

The paper:

Here’s a diagram of the experimental setup from the paper; the caption is also from the paper. Click to enlarge.

There are three groups: the control (right), consisting of pregnant women unexposed to war; the 1980 group, which included women who experienced violence when the fetuses had eggs (about 12 weeks into pregnancy); and the 2011 group, which included women who experienced violence in the early stages of pregnancy, before the (female) fetus developed eggs. Click diagram to enlarge:

Note the very small sample size of both women exposed to trauma and their children and grandchildren. Here is the violence the authors describe what was experienced by pregnant women:

“. . . . violent traumatic experiences that included being severely beaten, being persecuted (by the authorities/militia), seeing a wounded or dead body, and seeing someone else severely beaten, shot or killed.”

They then did DNA sequencing of all individuals using a sampling system that identified 850,000 nucleotide bases (SNPs). Out of these, they found 21 sites that were methylated in a pretty consistent way among those who experienced violence; these were in the pregnant women’s non-germline DNA, so could not be passed on. However, they found another 14 sites  methylated in the germline (mother’s or fetus’s eggs), and were inherited across not just one generation, but across two (this might be expected since fetal eggs can also be exposed to grandmother’s physiological conditions).  But in no case did they know which genes were involved in the changes, though they speculate that some regions could be involved in “gene regulation”.

The authors conclude this:

There is strong scientific evidence indicating that impacts of stress and trauma can reverberate far into the future, possibly through epigenetic mechanisms.

Well, that’s true if “far into the future” means “three generations,” but epigenetic marks are usually wiped clean from the DNA when gametes (sperm and eggs) are made, and four generations is about as far as any environmental alterations of mammalian DNA have persisted. What we do not have here is either inheritance of trauma or any kind of permanent evolutionary change produced by the environment. This is manifestly not “Lamarckian inheritance“!

The news piece does proffer some mild criticism:

These results are consistent with research in mice and other organisms that shows trauma can be passed down across generations. But other researchers note that the sample size isn’t big enough to confidently conclude that trauma passes from generation to generation through the germline—in this case via egg cells. “It’s important to do studies like this, and we need more of them, and with larger samples,” says Michael Pluess, a psychologist at the University of Surrey who was not involved in this study but whose own work with Syrian refugee children has found similar violence-related methylation changes in different places of the genome. “We also need to replicate the findings to know if they’re real or just chance.”

If you click on the first link in the preceding paragraph, you’ll find changes in biomarkers that may be associated with trauma in humans and mice, but not evidence for the inheritance of trauma itself.

But there is even stronger criticism of the methods and conclusions posted on Bluesky by John Greally, a professor of genetics at Albert Einstein College of Medicine, and he has  the chops to criticize.  Here’s his post thread in its entirety. One of his important criticisms appears to be that they got the Syrian DNA by using buccal (cheek) swabs, and, as Greally notes, “This could be a very expensive study of gingivitis.” Also, note the penultimate post in which Greally says that there’s not any convincing evidence (including this paper) for transmission of acquired characteristics in mammals.”  Just remember that when you hear about this study or the famous but misleading Dutch famine study.