For those of you who have “been around the block” in the online autistic community, you’ve probably heard of the “Neanderthal Theory of Autism,” proposed by rdos (Leif Ekblad) back in 2001, along with the subsequent Aspie Quiz he designed, which was used, in part, to further explore these ideas. Back in my early days on Wrong Planet (WP) circa 2004-ish when I went by the handle, Sophist, I remember the numerous discussions on the topic. From my perspective, many auties found the idea appealing– in the same way that the alien logo of WP plucked a primal string amongst Aspergians who had frequently taken the brunt of bullying or, at the very least, ostricism and “otherness” throughout their lives. The idea of being more Neanderthal was perhaps appealing to some on the spectrum.
In those days, I will admit, I thought the Neanderthal Theory unfounded and a bit ludicrous (sorry, Leif!). And while I still don’t prefer to venture into hypotheticals about Neanderthal-like behaviors in modern people, my own recent research has led me down an unexpected path regarding the roles of ancient Neanderthal DNA in autism susceptibility. After reading The Pattern Seekers by Simon Baron-Cohen, in which he draws attention to something called the Upper Paleolithic Revolution, which began roughly 50,000 years ago and is purportedly an early intellectual renaissance in our species (at least within Europe), I found it interesting that this “renaissance” roughly coincided with the time periods in which early Eurasian people would have been hybridizing with Neanderthals. It made me wonder: Has Neanderthal DNA played an important role in our own intellectual evolution? Given the links between autism, savantism, creativity, and even genius, could it still be playing a role in autism today??? All of a sudden, the Neanderthal Theory maybe didn’t sound so ludicrous.
After gathering my thoughts into a testable hypothesis, I brought the idea to my collaborator, Dr. Alex Feltus at Clemson University. Thankfully, Alex is usually game for slightly wacky ideas provided they’re testable and the research man power is available, so we started planning our approach. It took awhile for us to get our scientific ducks in a row, but fast forward to summer 2023 when we had our first results. It turns out, not that autistic folks have more Neanderthal DNA overall, but that they have more Neanderthal variants that are otherwise rare in the general population. We’ve just published our results in the journal of Molecular Psychiatry, and have a series of future studies in the works, also expanding our research to include Denisovan DNA (another of our ancient cousins with whom early Asian people intermixed).
People on the spectrum also seem to have more of a subset of common SNPs (SNPs that occur in at least 1% or more of the general population). In particular, we studied SNPs that are known as “quantitative trait loci” or “QTLs.” A QTL is any SNP in the genome that, when it varies or changes, it results in a measurable effect on a gene’s expression. For this study, we specifically only looked at QTLs that influence gene expression in the brain. A notable example includes a SNP in the SLC37A1 gene, which we found is associated with epilepsy and multiplex standing (i.e., more than one person in the family has autism). In particular, sixty-seven percent of white non-Hispanic autistics who were multiplex and had epilepsy carried this variant compared to 22% of ethnically-matched controls. Interestingly, SNP associations typically differed by ethnic group, so the same SNPs don’t appear to be susceptibility factors across all ethnic backgrounds. That means that the ethnic background of the person is also playing an important role, which has a lot of implications for genetics research in general.
What does this all mean? First off, I want to stress that autistic people are not “more Neanderthal” than non-autistic people. True, our original hypothesis predicted that autistic people, on average, would have more Neanderthal-derived SNPs than non-autistic people based on earlier work by Gregory and colleagues (2017). But this hypothesis wasn’t supported by the data. Instead, the reality is far more nuanced than that.
How Do Mutations Actually Work?
Most of us have grown up with the (false) idea that there’s a gene for everything: a gene for blue eyes, multiple genes for height, and even “autism genes.” But biology is never so simple. These are human generalizations that are akin to slapping a Campbell’s soup label on what are very complex processes. Individual genes are used as templates to create a variety of gene products (RNA, proteins) and variations in these genes and their products can definitely have a measurable influence on phenotype (observable characteristics). But the behavior of most genes is complicated and influences the behavior of numerous– if not all– cells in your body at one time or another. It’s important to understand this distinction in order to better appreciate what some Neanderthal SNPs may actually be doing in autism and how those effects can vary across different human populations.
Let’s imagine an extremely simplified example using two imaginary genes. We’ll call them Genes A and B. Genes A and B are next door neighbors to each other on Chromosome C. These A and B genes are often expressed at the same time, largely because the A and B proteins that ultimately get produced from these gene templates bind together to form the AAB complex, which contains two parts A to every B. Now, let’s pretend that the AAB complex is important in the development of neurons. You can start to imagine that the ratios at which Genes A and B are expressed are really important because if we produce too little or too much of either gene, that’s going to mess up that delicate balance and you won’t be able to make as much of the AAB complex, which might affect how your neurons develop.
Let’s say we reduce the expression of Gene A by 1/6th (figure below). This will lead to a 1/3rd reduction in the amount of the AAB complex we can produce in order to support neuron growth and development. In essence, our AAB complex is unbalanced. Because of this tendency to easily get off-balanced, Genes A and B are really sensitive to mutations and they don’t tend to tolerate them well. As a result, evolution of Genes A and B are relatively slow compared to other genes that are more tolerant of change. A lot of neurodevelopmental genes share this mutation sensitivity.
Now, a lot of genetic variants occur and are retained– not because they have some obvious positive adaptive benefit, like creating a new feature the lineage never had before. Instead, they tend to compensate for previous mutations that were not so beneficial (see Ohto, 1992). Basically, it’s a Red Queen genetic race where mutations occur and are retained because they help maintain (or re-attain) the status quo. However, when genetic variants are split apart that are normally used to compensating for each other, as may occur with sexual recombination in sex cells where we receive a mishmosh of our grandparents’ DNA, this can uncover individual mutations that lead to imbalance. This often happens in the case of two species when they hybridize, because unfamiliar variants can suddenly get thrown together and unbalance delicate networks like our imaginary AAB complex. In the illustration below, you can see how this plays out with the AAB complex when Homo sapiens and Neanderthals intermix, creating a hybrid offspring whose A and B gene expressions are out of balance with each other. This may have unanticipated effects on neuron development in the hybrid, even though we don’t see similar issues in either of the parents. Hopefully, you can appreciate that this scenario is more nuanced than simply saying, “Gene A is the gene for neuron development.” Otherwise, we lose context and, ultimately, understanding of the biological process that regulates the maturation of neurons.
In the early days following hybridization, we find that a subset of genetic variants that are spliced together thanks to sexual recombination reeeeaaaallllyy don’t work well together and these combinations are pretty quickly deleted from the gene pool, usually because they’re either lethal or they seriously affect fertility somehow. This most likely happened very early on when we began intermixing with Neanderthals. Eventually, however, Neanderthals went extinct (whether it was through competition, infection, or they got absorbed into our own larger species), such that we were left only with our new hybrid mostly-Homo-sapiens-plus-a-little-bit-Neanderthal species. Once the nasty SNPs got weeded out in the early days, that left a subset of variants that weren’t necessarily deadly or that impaired fertility, but they were just a little bit bad. Those SNPs, which are probably mildly mismatched with the H. sapiens background, are still being pushed out of the population genome, slowly and little by little.
These are probably some of the rare SNPs we’re seeing associated with autism. And while they’re likely being selected against because they’re mismatched with our ancestral genome and have some negative effects (e.g., health issues like immune disorders, connective tissue disorders, mild fertility issues such as PCOS), there’s also the possibility they’re simultaneously maintained in small numbers because they provide some benefit as well (e.g., human intelligence, creativity, etc.). This last bit is of course all conjecture and is not particularly testable, but these queries are nevertheless what led me to ask the question in the first place: Is Neanderthal DNA playing a role in autism susceptibility? Regardless of the reason some of these SNPs are sticking around in a subset of people, the answer appears to be “yes.” Even though this hybridization took place tends of thousands of years ago, it’s still playing a role in autism today.
Wow. Just… wow.