← Writing

Field notes

Migration Patterns Behind HCMV Diversity

Human betaherpesvirus 5 (HHV-5), better known as human cytomegalovirus (HCMV), is one of the most widespread and host-specific viruses we know. Herpesviruses are thought to have co-evolved with their hosts across hundreds of millions of years1, and as a result they tend to be tightly adapted to a single host species, rarely infecting anything outside their normal range. HCMV is also extremely common: seroprevalence in adults runs as high as 80% worldwide2. Most infected adults never notice, the virus simply establishes life-long residency in their cells and stays quiet.

That long, quiet coexistence is the interesting part. Most of the pathogens that make headlines, SARS-CoV-2, Ebola, HIV, are zoonotic in origin: they jumped into humans relatively recently, and spillover events like that tend to be the source of the deadliest outbreaks, the host hasn't had time to adapt. HCMV is the opposite case. It has been shaped by, and shaping, human hosts for a very long time, and that history shows up directly in its genome. Recent work on HCMV strain collections has found that strains cluster geographically, sequences sampled from southeast Africa, for instance, sit apart from the rest of the dataset as one of its most distinct groups. That's the pattern this post digs into.

What 330 HCMV genomes look like in strain-space

Below is a multidimensional scaling (MDS) projection of 330 publicly available HCMV genomes, built from pairwise genomic distance and reduced to two components so the overall structure is visible at a glance. Each point is one strain, colored by the continent it was sampled from, and you can hover any point for its strain name and country of origin.

n = 330 strains · Europe (266) and Africa (36, mostly Zambia and Kenya) make up the bulk of publicly available sequence, which is itself a sampling bias worth keeping in mind, but the African cluster's separation from the European core holds up regardless.

The African strains, Zambian sequences especially, sit toward the outer edge of the projection rather than blending into the larger Eurasian cluster. That's consistent with deeper genomic divergence, which is exactly what you'd expect if HCMV has been differentiating alongside human populations since long before recent global travel started mixing everything back together.

A viral map that mirrors a human one

If HCMV diversified as it traveled with us, its genomic map should echo our own migration map, at least roughly. Human genetic diversity carries a well-documented signature of the out-of-Africa expansion: populations that migrated furthest from the origin generally show reduced diversity relative to those that stayed closer to it, a serial founder-effect pattern repeated at every step of the journey. Immune-related genes are no exception, historic migration and the pathogens encountered along the way have measurably shaped how human immune variation is distributed across populations today4.

Map of early human migration out of Africa, showing approximate routes and timing in thousands of years ago, including introgression events with Neanderthals and Denisovans
Approximate routes and timing of early human migration out of Africa, with Neanderthal and Denisovan introgression events marked. From Domínguez-Andrés & Netea, 20194.
World map of Y-DNA haplogroups showing dominant paternal lineages in native populations and their proposed migration routes
Y-DNA haplogroup distribution and proposed migration routes, a complementary view of the same population movements a host-adapted virus would have traveled alongside.

Line the two maps up against the HCMV projection above and the parallel is hard to miss: the same broad strokes, an African origin, a long Eurasian spread, more isolated founder populations at the far ends of the migration routes, show up in both the human genetic record and, more faintly so far, in HCMV's.

An open question worth testing

Primary HCMV infection is usually silent in healthy adults, but not always, some people develop mononucleosis3, and we still don't have a good explanation for why. One hypothesis this pattern suggests: HCMV strains that co-evolved with a given human population over a long stretch of shared history may simply be better tolerated by descendants of that population than a strain introduced more recently from elsewhere, the same logic that makes a pathogen most dangerous right after it jumps into a new host applies, in a milder form, to mismatches between a well-adapted virus and an unfamiliar host population. That's a hypothesis, not a finding, the strain-clustering pattern above is suggestive, not proof, but it's a concrete, testable direction: pairing strain phylogeography with host ancestry and clinical outcome data is the natural next step.

References

  1. Davison AJ. Evolution of the herpesviruses. Vet Microbiol. 2002 Apr 22;86(1-2):69-88. doi: 10.1016/s0378-1135(01)00492-8. PMID: 11888691.
  2. Zuhair M, Smit GSA, Wallis G, et al. Estimation of the worldwide seroprevalence of cytomegalovirus: A systematic review and meta-analysis. Rev Med Virol. 2019;29:e2034. doi: 10.1002/rmv.2034.
  3. Mayo Clinic Staff. Mononucleosis: Symptoms and causes. Mayo Clinic. mayoclinic.org.
  4. Domínguez-Andrés J, Netea MG. Impact of Historic Migrations and Evolutionary Processes on Human Immunity. Trends Immunol. 2019 Dec;40(12):1105-1119. PMID: 31786023.