Kyoto-Led Team Maps the Gene Switches Behind Human Skeletal Evolution

Why did the human skeleton become so different from those of other great apes? Fossils document the emergence of upright walking, a reshaped pelvis, longer lower limbs, a lighter skeleton and a skull capable of housing an unusually large brain. What fossils cannot easily show is which changes in DNA altered the developmental programs that produced those forms. A Kyoto University-led international team has now built what it describes as the first functional atlas of human-specific gene-regulatory changes involved in skeletal evolution.^[1][2]

The study, published in Nature on September 23, was led by Associate Professor Fumitaka Inoue and researcher Yizhi Yan at Kyoto University’s Institute for the Advanced Study of Human Biology, together with David Gokhman and Nadav Mishol at the Weizmann Institute of Science and collaborators across several institutions. Rather than looking primarily for protein-coding genes that are unique to humans, the team focused on DNA sequences that control when, where and how strongly genes are expressed.^[1]

Looking beyond genes to the instructions that control them

Protein-coding genes are only part of the biological program that builds an organism. The same gene can produce very different developmental outcomes depending on when it is activated, in which tissue and at what level. Promoters and enhancers—cis-regulatory elements near or linked to genes—act more like switches and dials than like protein blueprints.

The idea that regulatory evolution may explain major human–ape differences has deep roots. In 1975, Mary-Claire King and Allan Wilson drew attention to the striking similarity of human and chimpanzee proteins compared with the large differences in anatomy and behavior, arguing that changes in systems controlling gene expression could be central to human evolution. Decades of comparative genomics have since strengthened the case that non-coding regulatory DNA can contribute substantially to phenotypic divergence.^[3][4]

Finding a sequence that differs between species, however, is not the same as proving that the sequence changes gene activity. The new study’s significance lies in testing an exceptionally large collection of human-derived substitutions in cells directly relevant to skeletal biology.

Testing more than half a million substitutions

The researchers assembled 561,410 fixed or nearly fixed human-derived substitutions located in candidate promoters and enhancers. They synthesized matched human and great-ape versions of the surrounding sequences and tested them using a massively parallel reporter assay, or MPRA, in human chondrocytes—cells central to skeletal development and the maintenance of cartilage.^[2]

In an MPRA, candidate regulatory sequences are coupled to DNA barcodes. By measuring the amount of RNA produced from each barcode, researchers can test hundreds of thousands of sequences in parallel and quantify how strongly each version drives expression. The study identified 15,077 candidate cis-regulatory elements with human-specific regulatory activity.^[1][2]

Putting human and ape chromosomes in the same cellular environment

MPRA captures only part of regulatory evolution. Insertions, deletions and structural variants can also affect expression, and differences between separately cultured human and ape cells may reflect environmental or experimental conditions rather than DNA acting locally. The team therefore used another unusual system: experimentally fused human–chimpanzee and human–gorilla cells that were differentiated into osteochondral progenitors.^[2]

In these interspecies hybrid cells, human and ape chromosomes occupy the same nucleus and are exposed to the same pool of transcription factors and other trans-acting molecules. If the human copy of a gene is expressed differently from the ape copy under those shared conditions, the difference is much more likely to come from cis-regulatory information attached to the chromosomes themselves.

The hybrid-cell analysis identified 4,463 human-specific changes in cis-regulatory expression. The two methods therefore work at complementary levels: MPRA identifies sequence variants capable of changing regulatory activity, while hybrid cells identify genes whose expression diverged on the human lineage. Integrating the two creates a route from DNA variant to gene to biological pathway.^[2]

A broad signal emerges in the extracellular matrix

When the datasets were integrated, one of the strongest patterns involved the extracellular matrix, or ECM—the network of molecules surrounding cells that gives tissues structure and mechanical properties. Within that network, the researchers found widespread downregulation of pathways that synthesize glycosaminoglycans, or GAGs.^[1][2]

GAGs are long sugar molecules that are major components of cartilage and other extracellular matrices. They bind water, help tissues resist compression and participate in cell signaling. In joints, they are integral to the ability of cartilage to remain hydrated and absorb mechanical loads.

The researchers compared joint tissue and found that GAG content in humans was approximately three- to four-fold lower than in non-human apes. They also found signatures consistent with lineage-specific selection affecting this pathway, along with two human-specific expansions of GAG-anchor repeats in aggrecan, a major extracellular-matrix protein.^[2]

What lower GAG levels might mean for human form

The team then compared known phenotypes associated with reduced GAG biosynthesis with anatomical traits that distinguish humans. Traits linked to lower GAG levels—including shorter stature, facial flattening, a higher forehead, shorter fingers and changes in thumb orientation—showed significant enrichment among human-specific phenotypes. The paper reports a 2.9-fold enrichment.^[2]

That result is suggestive, not proof that GAG reduction directly produced those features. Human skeletal anatomy reflects many genes, developmental processes, mechanical forces and life-history changes. The paper instead argues that coordinated remodeling of extracellular-matrix biology was likely one contributor to the evolution of human skeletal morphology.

The fossil record supplies the anatomical timeline

The skeletal transition itself is visible across millions of years of fossils. The Smithsonian’s Human Origins Program notes evidence for upright posture and bipedal locomotion among early hominins by roughly six million years ago. A 4.1-million-year-old tibia from Australopithecus anamensis shows weight-bearing adaptations consistent with frequent upright walking. The 3.2-million-year-old Australopithecus afarensis skeleton known as Lucy combined a short, broad pelvis and inward-angled femora suited to bipedal walking with upper-body features that still supported climbing. By about 1.9 million years ago, Homo erectus had a pelvis and long lower limbs much closer to the pattern associated with efficient long-distance terrestrial walking.^[5][6]

Those fossils tell researchers when anatomical changes appeared. Genomics asks a different question: which molecular changes made altered development possible? It is rarely possible to draw a straight line from one nucleotide substitution to one fossil feature. The value of the new atlas is that it provides thousands of experimentally supported candidates that can now be tested against developmental and anatomical hypotheses.

Could the biology that shaped us also leave us vulnerable?

The study also raises a medically provocative possibility. Because GAGs help maintain cartilage mechanics, a reduction in GAG abundance could conceivably trade some skeletal properties for reduced long-term tissue resilience. The Nature paper notes that several degenerative skeletal disorders linked to low GAG levels are more common in humans than in non-human apes even after attempts to control for age, sex and environmental variables. The authors therefore propose that human-specific reduction in GAG content may contribute to our unusual susceptibility to degenerative skeletal disease.^[2]

That should not be read as a claim that evolution “caused arthritis.” Lifespan, body mass, activity, environment and diagnosis all influence disease prevalence. The finding is better understood as a possible evolutionary trade-off: some regulatory changes that contributed to a distinctly human skeleton may also have altered the long-term properties of cartilage.

An experimental candidate list for the next generation of skeletal-evolution research

The long-term value of the work may be the resource itself. The 15,077 candidate regulatory elements and 4,463 human-specific expression changes create a set of hypotheses that future studies can interrogate one by one. Genome editing could, in principle, replace selected human regulatory sequences with ancestral versions—or introduce human versions into experimental systems—to test effects on cartilage, extracellular matrix and developmental programs.

There are important limitations. MPRAs measure regulatory activity in an experimental context and cannot fully reproduce three-dimensional chromosome architecture or every stage of embryonic development. Hybrid cells are powerful controls for cis regulation, but they are not a developing human skeleton. The authors therefore describe the work as a framework for dissecting the genetic basis of human skeletal biology, not as a completed causal map of human evolution.^[2]

From the shape of bones to the logic of gene control

Human-origins research began with bones, teeth, tools and geology. Molecular biology later added protein comparisons, and genome sequencing exposed millions of DNA differences between humans and our closest relatives. The frontier is now shifting again—from cataloguing which sequences differ to determining what those differences actually do inside cells.

The new work offers one concrete answer to a question that has persisted for half a century: how can species with highly similar genomes acquire dramatically different bodies? Part of the answer may lie not in inventing entirely new genes, but in rewiring when and how existing genes are used.

After testing 561,410 substitutions, the researchers did not find a single “human skeleton gene.” They found a network of regulatory changes touching extracellular matrix, cartilage biology and multiple developmental pathways. Human skeletal evolution increasingly looks less like the flip of one genetic switch than the gradual rewiring of thousands of controls. This atlas gives researchers a new map for finding out which of those controls mattered most.