Human cartilage has less water-holding sugar. Genes offer a clue why

The central finding is striking. Across eight joint types, human cartilage samples contained substantially less sulfated glycosaminoglycan, or GAG, than samples from other great apes. These long sugar chains help cartilage retain water and withstand compression. After adjustment for DNA content, human levels were 2.6 to four times lower, with an overall difference described as roughly threefold. That is a difference in tissue chemistry, not a measured increase in the likelihood of osteoarthritis.⁠3

The study, published in Nature on September 23, 2026, received renewed attention through a November 14 Jerusalem Post interview with co-leader David Gokhman. It was a collaboration between his laboratory at Israel’s Weizmann Institute of Science and Fumitaka Inoue’s laboratory at Kyoto University, with Nadav Mishol and Yizhi Yan leading the research work. The November coverage brought an interview, not a newly published experiment.⁠3⁠12⁠1

GAGs matter because cartilage’s cushioning depends on more than its cells. Those cells make an extracellular matrix—the material surrounding and supporting them. GAG chains carry a strong negative electrical charge, helping draw water into a hydrated network that contributes to elasticity and resistance to pressure. They account for roughly 15–30 percent of cartilage’s dry weight; that figure excludes the water itself.⁠13

The researchers’ proposal is that lower production leaves humans with a smaller biochemical reserve, from which further losses could contribute to degeneration. It is a useful hypothesis precisely because it identifies something specific to investigate. It does not explain every painful joint, establish an inevitable decline or show that restoring an ape-like level would prevent disease.⁠3

The tissue work covered hips, knees, shoulders, elbows, wrists, knuckles and two categories of finger joints. Researchers measured sulfated GAG using a dye-binding assay, then normalized the measurements to DNA content. They did not directly measure total joint cushioning, cartilage volume or the mechanical strength of each joint.⁠3

There were 111 samples from a reported 42 human donors and 28 samples from seven great apes. Those distinctions matter: several pieces of cartilage from one individual do not become several independent donors. For the overall comparison, repeated measurements were averaged to one value per individual. Some finger-joint comparisons had tissue from only one ape, sharply limiting how well those particular results can represent a wider population.⁠3

Nevertheless, every joint category pointed in the same direction. The difference persisted after adjustments for age and sex, and after two approaches to matching ages across species. It also crossed joints experiencing different mechanical loading in humans. That makes an explanation based solely on how much a particular joint is used less persuasive, without ruling out every environmental influence.⁠3

All specimens were collected after death; no animals were killed for the study. Human tissue came through the Technion’s consenting body-donor program, while zoological collaborations supplied ape material. Collection conditions were not identical: human specimens were chemically preserved, while specified chimpanzee hand material had been frozen. This was a small, cross-sectional comparison—not a record of how people’s and apes’ cartilage changed throughout their lives, or a direct measurement of when the difference begins in human development.⁠3

The genetic work gives the chemistry a possible mechanism. Rather than starting with a suspected arthritis gene, the team compared genomes from 15,708 modern humans and 139 great apes. Three Neanderthal genomes and one Denisovan genome helped identify changes shared across modern and ancient human lineages. From that search, researchers selected 561,410 single-letter substitutions in candidate regulatory regions: DNA sequences that help control when, where and how strongly genes operate.⁠3

They built paired human and ancestral versions of those regions and tested their activity in human cartilage cells. Of 65,576 sequence pairs showing activity, 15,077—23 percent—behaved significantly differently between versions. Most human and ancestral versions behaved similarly. And a sequence quiet in this assay could still matter in another tissue or developmental stage.⁠5

A complementary experiment put human and ape chromosomes inside the same laboratory cells. Researchers fused cultured human cells with chimpanzee or gorilla cells, then guided the hybrids toward an early stage capable of forming bone and cartilage. These were hybrid cells, not hybrid animals. The shared cellular environment helped distinguish effects of nearby regulatory DNA from differences that arise when species are grown in separate cultures.⁠3⁠12

Both approaches implicated the machinery that builds and maintains the extracellular matrix. In the hybrid-cell analysis, 17 of 21 GAG-related genes whose activity differed were less active on the human side. The result suggests coordinated changes across a production pathway, rather than a single mutation switching cartilage production off.⁠3

There was also an important complication. The human ACAN gene, which encodes the aggrecan scaffold carrying many GAG chains, had expanded. Human versions averaged six extra repeats, creating 12 additional potential attachment sites. Those extra sites might compensate for reduced GAG production, the researchers suggest, but that remains unproved. More places to attach chains do not necessarily mean more chains are made—which is why the direct tissue measurements were essential.⁠3

Reduced cartilage GAG is already associated with osteoarthritis and other degenerative conditions. The researchers propose that the species-level reduction contributes to human vulnerability, with further losses occurring through life. But this study did not follow disease incidence or calculate an individual’s additional risk. Great apes can develop joint disease too; the comparison is not human frailty against ape immunity.⁠3

Why the change persisted is less settled still. One proposal concerns the nervous system, where GAGs help stabilize connections between neurons: lower levels might have prolonged brain plasticity and opportunities for learning. Another links reduced production to a lighter, more slender skeleton compatible with endurance running. The study tested neither learning nor running performance. Neither advantage has been demonstrated here.⁠13

Nor must every persistent evolutionary change represent a successful bargain. The paper considers both adaptation and a relaxation of negative selection—the possibility that constraints against lower production weakened. It would therefore be premature to declare that natural selection traded healthy joints for intelligence or stamina.⁠3

For patients, the boundary is equally clear. Injections involving GAGs are already used in osteoarthritis care, but this was not a treatment trial and did not establish that earlier injections prevent disability. It produced no validated new screening test, preventive regimen or gene therapy. In his Jerusalem Post interview, Gokhman cautioned: “But we are far from using it for gene therapy.”⁠1⁠3

The work gives researchers a more concrete biological mechanism and potential targets to investigate. Whether changing that machinery could safely preserve mobility remains unanswered. The discovery is not a cure for painful joints; it is a better-grounded explanation to test.