Increased hyaluronan by naked mole-rat Has2 improves healthspan in mice

Zhang Z, Tian X, Lu JY, Boit K, Ablaeva J, Zakusilo FT, Emmrich S, Firsanov D, Rydkina E, Biashad SA, Lu Q, Tyshkovskiy A, Gladyshev VN, Horvath S, Seluanov A, Gorbunova V · Nature 621(7977):196–205 · 2023 · DOI: 10.1038/s41586-023-06463-0

TL;DR

This is the first whole-animal demonstration that a naked mole-rat longevity adaptation — accumulation of high-molecular-mass hyaluronic acid (HMM-HA) via the HAS2 synthase — can be transferred to a second mammalian species and produce measurable lifespan extension, cancer resistance, and multi-organ healthspan improvements. Transgenic mice carrying the naked mole-rat Has2 ORF showed +4.4% median lifespan, +12.2% maximum lifespan, reduced spontaneous cancer mortality, attenuated inflammaging, preserved gut barrier integrity, a longevity-associated microbiome shift, and a modest deceleration of epigenetic aging. This constitutes a methodological precedent for xenogeneic gene transfer as a longevity intervention modality: isolating a single molecular mechanism from a long-lived species and re-instantiating it in a shorter-lived mammal produces a quantifiable aging benefit.

Study design

Transgenic construct

The naked mole-rat (Heterocephalus glaber) Has2 open reading frame was cloned into the pCALNL-GFP vector under a ubiquitous CAG promoter. A loxP-flanked STOP cassette keeps the transgene silent until Cre-mediated excision. Mice were crossed with R26-CreERT2 mice on a C57BL/6 background to produce the inducible nmrHas2 line. Both nmrHas2 transgenic mice and creER-only controls received tamoxifen (80 mg/kg, 5 consecutive days; route not specified in paper methods) at 3 months of age to induce recombination. This design ensures tamoxifen exposure is identical between groups, isolating the effect of nmrHas2 expression.

Lifespan cohort

175 mice total: nmrHas2 n=84 (50 female, 34 male); creER controls n=91 (54 female, 37 male). Cages were inspected every night; dead animals were examined macroscopically. Animals were humanely euthanized if showing significant signs of morbidity. No mice were censored from survival analysis.

Companion cohorts

Separate cohorts for tissue timepoints at 6 months and 24 months (transcriptomics, cytokines, histology, gut physiology), young/old pairings for healthspan assays (frailty, rotarod, grip strength, bone micro-CT), and 5-month males for in vitro/LPS challenge experiments. Three biological replicates per sex, age group, and genotype in most molecular assays.

Primary outcomes

Lifespan

nmrHas2 mice showed significantly extended lifespan relative to creER controls (two-tailed log-rank test, p<0.05):

• Overall median lifespan: +4.4% (n=84 vs n=91)
• Overall maximum lifespan: +12.2%
• Female nmrHas2: +9% median lifespan (n=50 vs 54 controls)
• Male nmrHas2: similar median to controls but +16% maximum lifespan (n=34 vs 37 controls)

The sex-stratified split — median improvement driven by females, maximum improvement by males — is not mechanistically explained in the paper and warrants caution in interpretation. contradictory-evidence

Cancer incidence

• Spontaneous cancer mortality across all ages: 57% in nmrHas2 vs 70% in creER controls (full lifespan cohort: n=84 nmrHas2, n=91 creER; pooled sexes)
• In animals surviving past 27 months: 49% nmrHas2 vs 83% creER (subanalysis cohort n=74 nmrHas2, n=81 creER; p<0.05, two-tailed χ² test)
• DMBA/TPA-induced skin papilloma protocol (20-week chemical carcinogenesis): nmrHas2 mice formed significantly fewer papillomas vs creER (pooled sexes: n=7 acetone, n=13 creER, n=11 nmrHas2; p<0.05 two-tailed unpaired t-test; sex-stratified from Extended Data: females n=4 nmrHas2 vs n=5 creER; males n=7 nmrHas2 vs n=8 creER)

The cancer-mortality reduction at >27 months is striking (49% vs 83%). The DMBA/TPA cohort is small and should be regarded as directionally confirmatory, not independently conclusive. needs-replication

Healthspan measures

Assay	Groups	Result
Frailty index	Old: nmrHas2 n=14, creER n=13; young: nmrHas2 n=16, creER n=17	Old nmrHas2 scored substantially lower (healthier)
Rotarod latency-to-fall	6F + 6M per group	Old creER: significant decline; old nmrHas2: maintained youthful performance
Grip strength	Same young/old cohorts as frailty	Old nmrHas2 outperformed old creER
Bone connectivity density (femur, tibia, tibia subchondral)	Old female nmrHas2 n=6, creER n=5	Higher connectivity in nmrHas2 females; males NS
Body weight	n=9-11 per sex per genotype	Similar across life — benefit not confounded by weight differences

Epigenetic clock

24-month liver samples analyzed with the HorvathMammalMethylChip40 mammalian clock (not the original 2013 Horvath human-tissue clock — see horvath-clock-2013 §Note for the distinction). nmrHas2 mice showed approximately −0.2 years of age acceleration relative to creER controls (n=9 per group, p<0.05). Of 6,553 CpG sites analyzed: 165 were differentially methylated between genotypes. Among CpGs that normally gain methylation with aging: 145 showed lower methylation in nmrHas2. Among CpGs that normally lose methylation with aging: 20 showed higher methylation in nmrHas2. Both directions indicate that nmrHas2 partially resists the canonical aging methylation drift. needs-replication (n=9)

Mechanistic findings

HA accumulation

Pulse-field gel electrophoresis of muscle, heart, kidneys, and small intestine confirmed higher molecular mass and greater abundance of HA in nmrHas2 mice. HABP (hyaluronic acid binding protein) fluorescence quantification (n=3 biological replicates) confirmed tissue-level accumulation. The authors note that accumulation was “mild” despite very high nmrHas2 mRNA levels across tissues (Log2 expression vs endogenous mouse Has2: liver ≈7.7, muscle ≈9.5, kidney ≈10.6, intestine ≈5.7 in females; comparable male values reported). The authors attribute the modest accumulation to high endogenous mouse hyaluronidase activity degrading the newly synthesized HMM-HA. This is an important mechanistic caveat: the gain-of-function may be substantially attenuated in mice relative to what occurs in naked mole-rats, whose hyaluronidase activity is reportedly lower ¹. dose-response-unclear

Inflammation suppression

Plasma cytokines in aged females (n=7 nmrHas2, n=8 creER): IL-12p40, MIP1α, MIP1β, and CCL7 were significantly lower in old nmrHas2; the majority of pro-inflammatory cytokines and chemokines trended lower. Young cohorts (n=11 each) showed no significant differences, consistent with an age-dependent phenotype.

In vivo LPS challenge (young males, n=3–4): At 4 hours post-injection, nmrHas2 males had reduced plasma TNF; both sexes showed reduced IL-6. At 24 hours, female nmrHas2 showed lower plasma TNF. Liver, spleen, and kidney Il1b and Tnf mRNA were significantly lower in nmrHas2 24h post-LPS (n=4 females); Il6 was similar between groups.

Bone-marrow-derived macrophage (BMDM) analysis (n=3 males): nmrHas2 BMDMs expressed ~6× more Has2 mRNA. After LPS challenge: Il1b and Il6 were significantly lower, Tnf trended lower, Il10 and Arg1 (arginase-1, M2-polarization marker) were higher. This macrophage polarization shift — less M1-like, more M2-like — is proposed as a key effector of the systemic inflammation suppression.

Oxidative stress resistance

Skin fibroblasts from 5-month male nmrHas2 mice showed higher survival after 24-hour H₂O₂ exposure (100, 200, and 400 µM; n=3). Critically, mouse cells overexpressing mouse Has2 (not the NMR transgene) showed the same protective effect. This is the key control demonstrating that HMM-HA itself — not any species-specific property of the NMR Has2 protein — confers the oxidative stress resistance. This reframes the field’s interpretation of the Tian 2013 NMR-specific-HA narrative: the beneficial effects may be size/abundance-dependent and portable to mouse Has2 as well ¹. contradictory-evidence

Gut barrier integrity

FITC-dextran gut permeability assay: Young creER and nmrHas2 (n=10 each) were similarly permeable. In old mice (n=12 each), creER showed age-associated increase in gut permeability while nmrHas2 maintained low, youthful permeability (p<0.05).

Intestinal epithelial structure (n=10 pooled per timepoint): Both young and old nmrHas2 had more goblet cells in small intestine and distal colon vs age-matched controls. Young nmrHas2 had more Paneth cells; the age-related increase in Paneth cell number was attenuated in nmrHas2 (lower relative increase). Lgr5+ intestinal stem cell (ISC) counts were similar across genotype and age.

Organoid formation (n=3): Old creER crypts formed far fewer organoids than young controls. Old nmrHas2 crypts maintained young-like organoid-forming capacity. Supplementing old creER crypt medium with exogenous HMM-HA (20 µg/mL) was sufficient to rescue organoid formation; LMW-HA (low-molecular-mass) had no effect. This molecular weight selectivity confirms HMM-HA (not any HA) as the active species and supports a receptor-mediated or physical-scaffold mechanism.

Transcriptome and longevity signature

Transcriptome analysis across liver, muscle, white adipose tissue, kidneys, and spleen at 6 and 24 months (3 biological replicates per sex/age/genotype): nmrHas2 mice accumulated fewer age-associated transcriptome changes and maintained a younger transcriptome signature. Pathway analysis showed:

• Downregulated in nmrHas2 with age (relative to creER): interleukin and interferon signaling; genes correlated with negative mouse longevity signatures (neg-MLS)
• Upregulated: oxidative phosphorylation, respiratory electron transport, mitochondrial translation; genes correlated with positive longevity signatures (pos-MLS)

The authors report that the nmrHas2 transcriptome signature showed correlation with rapamycin and caloric restriction longevity signatures but with stronger downregulation of inflammation and senescence pathways specifically. This positions HMM-HA as an intervention with a distinct mechanistic emphasis relative to mTOR inhibition or CR.

Microbiome

16S rRNA sequencing of 7- and 24-month mice (n=9 nmrHas2, n=10 creER controls, sexes pooled): Old nmrHas2 showed increased Bacteroidetes and decreased Firmicutes relative to old creER (p<0.05 for old groups; young groups did not differ). At family level: Muribaculaceae (associated with murine longevity) higher in old nmrHas2; Deferribacteraceae, Streptococcaceae, and Lachnospiraceae (pro-inflammatory families) higher in old creER. The direction of change — higher Bacteroidetes, lower Firmicutes, higher Muribaculaceae — matches the firmicutes-bacteroidetes-ratio pattern associated with healthier aging and longer-lived mice. Whether the microbiome shift is a cause or consequence of the gut barrier preservation cannot be determined from this design. no-mechanism

Limitations and gaps

• Modest median lifespan extension (~4.4%) relative to the robust healthspan and cancer-resistance signals. Maximum lifespan improvement (+12.2%) is larger but calculated from a smaller effective sample (oldest animals) and is more sensitive to individual outliers. needs-replication
• Sex-dependent result: median lifespan improves in females (+9%), while males gain primarily in maximum lifespan (+16% max). No mechanistic explanation is provided for this split. contradictory-evidence
• Mild HA tissue accumulation despite high transgene mRNA expression. The mouse hyaluronidase activity ceiling may substantially limit the achievable HMM-HA benefit. Whether co-inhibition of mouse hyaluronidase would amplify lifespan benefits is untested. dose-response-unclear
• HA molecular mass not directly measured in vivo across tissues — pulse-field gel confirms higher mass, but quantitative size-distribution in circulation and individual tissues at multiple ages is not provided. no-mechanism
• Small n in some key assays: LPS challenge n=3–4, methylation clock n=9, DMBA/TPA carcinogenesis n=7–13. Results from these cohorts should be regarded as preliminary pending replication.
• No human translation pathway yet established. HAS2 gene therapy in humans raises regulatory and safety questions not addressed here. needs-human-replication
• Mouse Has2 overexpression replication: the finding that mouse Has2 alone reproduces the oxidative stress resistance in fibroblasts (noted above as a key control) is from a single cell-type experiment and requires in vivo confirmation. If mouse Has2 overexpression alone is sufficient for healthspan/lifespan benefits in vivo, it would be a simpler translational target than xenogeneic NMR Has2 expression. needs-replication
• Reframes Tian 2013 species-specificity narrative: Tian 2013 attributed NMR cancer resistance partly to species-specific properties of NMR HAS2 (lower turnover, different regulation). The present mouse Has2 fibroblast result suggests HA mass/abundance may be more important than the species of origin of HAS2. This is a substantive reframe of existing claims on heterocephalus-glaber; that page should be reviewed to reflect this nuance. contradictory-evidence

Broader implications: xenogeneic gene transfer as a longevity modality

This paper establishes a proof-of-concept framework that extends well beyond HMM-HA: longevity adaptations evolved in extreme-lifespan species can be functionally transferred to shorter-lived mammals via single-gene xenogeneic overexpression. Prior work demonstrated species-correlation between HAS2 biology and longevity ¹, and characterized HA size and abundance differences between NMR and other species ² (with del Marmol 2021 partially challenging Tian 2013’s ultra-HMW claim), but these remained correlational. Zhang 2023 is the first whole-animal test.

The obvious next-target candidates in the same framework:

• Elephant TP53 retrogenes — Loxodonta africana carries ~20 extra copies of a TP53 retrogene (LIF6/others); elephant-derived TP53 introduced into mouse cells shows enhanced DNA-damage-induced apoptosis. See loxodonta-africana for current status. (Related precedent: Tyner 2002 super-p53 mouse extends cancer resistance but not lifespan in isolation; Tomás-Loba 2008 combined super-p53/Ink4a/Arf + telomerase did achieve lifespan extension — composite gene-dosage strategy.)
• Bowhead whale CIRBP (cold-inducible RNA binding protein variant) — bowhead whales carry a distinct CIRBP variant associated with enhanced proteostasis; candidate xenogeneic transfer target.
• NMR p15/p16 hybrid — naked mole-rats express an unusual p15/p16 hybrid ARF-INK4a-like protein that provides uniquely robust cell cycle arrest under hyperplastic stimuli. Whether this can be stably expressed in mice without toxicity is unknown.

The Tomás-Loba 2008 design is an important conceptual precursor: combining enhanced tumor suppression (super-p53, Ink4a/Arf overexpression) with telomerase activation (to avoid accelerated cellular senescence from over-active tumor suppression) achieved lifespan extension and healthspan improvement in mice — a multi-gene strategy for balancing safety vs. longevity mechanisms. Zhang 2023 is simpler (single gene), suggesting that some longevity adaptations have sufficiently clean action-separation to be portable without a counter-balancing gene.

See engineered-negligible-senescence for the broader theoretical framework.

Follow-up work (Gorbunova/Seluanov lab, 2023–2024)

Three closely related papers from the same lab published concurrent with or shortly after Zhang 2023:

• Zhao Y et al. (2023), Nature Communications (doi:10.1038/s41467-023-43623-2) — showed that evolution of HMM-HA is broadly associated with subterranean lifestyle across multiple species, suggesting convergent evolution under similar selective pressures (confinement, hypoxia, CO₂ accumulation, cancer risk). Locally available in archive. ³
• Takasugi M et al. (2023), Cell Reports (doi:10.1016/j.celrep.2023.113130) — demonstrated that CD44 (the primary HMM-HA receptor) correlates with longevity across species and confers enhanced ATF6 ER-stress resistance. This identifies the downstream receptor axis through which HMM-HA likely exerts cellular protection. Links directly to cd44. ⁴
• McGuire J et al. (2024), Scientific Reports (doi:10.1038/s41598-024-64924-6) — delphinidin (an anthocyanidin) inhibits hyaluronidase activity and suppresses cancer metastasis, suggesting a small-molecule approach to boosting endogenous HMM-HA. If confirmed, this could be a pharmacological alternative to gene transfer for reproducing a subset of the HMM-HA phenotype. ⁵

Extrapolation to humans

Dimension	Status	Notes
Pathway conserved in humans?	yes	HAS2 and HA biology are fully conserved; humans express HAS1/HAS2/HAS3
Phenotype conserved in humans?	unknown	HMM-HA manipulation in humans has not been tested for aging endpoints
Replicated in humans?	no	No human data; gene therapy approach raises unresolved safety and delivery questions

needs-human-replication

See also

• has2 — the transgene; NMR-derived hyaluronic acid synthase 2
• hyaluronic-acid — HMM-HA vs LMW-HA biology; tissue accumulation
• cd44 — primary HMM-HA receptor; Takasugi 2023 longevity-correlation data
• heterocephalus-glaber — NMR model organism page; NMR-specific HA biology; review Tian 2013 species-specificity claim in light of Zhang 2023 mouse-Has2 fibroblast result
• chronic-inflammation — inflammaging hallmark; IL-12p40/MIP1α/MIP1β/CCL7 reduction mechanism
• cellular-senescence — transcriptome shows stronger senescence downregulation than rapamycin/CR
• dysbiosis — Bacteroidetes↑/Firmicutes↓ microbiome shift
• firmicutes-bacteroidetes-ratio — directional alignment with longevity-associated microbiome patterns
• gut-barrier — FITC-dextran permeability, goblet/Paneth cells, ISC organoid rescue
• frailty-index — frailty scoring in aged nmrHas2 cohort
• horvath-clock-2013 — note: Zhang 2023 used HorvathMammalMethylChip40 mammalian multi-species clock, not the original 2013 353-CpG human-tissue clock; these are distinct instruments
• cancer-aging-tradeoffs — cancer incidence reduction as the dominant lifespan benefit
• engineered-negligible-senescence — xenogeneic gene transfer as a modality; broader framework
• loxodonta-africana — elephant TP53 retrogenes; next-target candidate in same transfer framework

---

Footnotes

1. doi:10.1038/nature12234 · Tian X et al. (Gorbunova V, Seluanov A) · Nature 2013 · 779 citations · local PDF available in a local paper archive · in-vivo + in-vitro; established that NMR-derived HMM-HA mediates cancer resistance; identified the NMR Has2 promoter as conferring higher Has2 expression than mouse ortholog. contradictory-evidence — Zhang 2023 mouse-Has2 fibroblast result challenges the species-specificity interpretation. ↩ ↩² ↩³

2. doi:10.1038/s41598-021-86967-9 · del Marmol D et al. · Scientific Reports 2021 · local PDF available in a local paper archive · quantified HA abundance and size distribution in NMR, guinea pig, and mouse tissues/serum using HABP-based histochemistry and SEC. Partially challenges Tian 2013: found NMR HA peak MW ~2.5 MDa (not 6–12 MDa), found no ultra-HMW HA (≥3–4 MDa) in NMR skin by gel electrophoresis, and found that HA localization pattern is broadly similar across the three species. Did confirm that HA amount is higher in NMR than guinea pig in most tissues (serum: >tenfold higher in NMR vs guinea pig), and that the HMW-to-LMW ratio is shifted toward HMW in NMR vs guinea pig. Compared NMR to guinea pig (Cavia porcellus), not to rats. contradictory-evidence (vs Tian 2013 ultra-HMW claims) ↩

3. doi:10.1038/s41467-023-43623-2 · Zhao Y et al. (Seluanov A, Gorbunova V) · Nature Communications 2023 · in-vivo/comparative · comparative genomics + biochemistry across subterranean species; HMM-HA evolution convergent with subterranean niche. ↩

4. doi:10.1016/j.celrep.2023.113130 · Takasugi M et al. (Seluanov A, Gorbunova V) · Cell Reports 2023 · in-vitro + cross-species · CD44 longevity correlation; ATF6/ER-stress mechanism downstream of HMM-HA receptor engagement. ↩

5. doi:10.1038/s41598-024-64924-6 · McGuire J et al. (Seluanov A, Gorbunova V) · Scientific Reports 2024 · in-vivo + in-vitro · delphinidin as hyaluronidase inhibitor; cancer metastasis suppression; potential pharmacological route to sustained HMM-HA. needs-replication ↩