iiinova — where your digital self keeps evolving

Preprint: www.biorxiv.org/content/10.6... Thank you to @ishahjain.bsky.social, @physiologyiscool.bsky.social and everyone in the Jain Lab for their support and advice through the twists and turns of this project. Stay tuned for our related review about electron acceptors, coming out tomorrow!

BCAA fermentation is an ancient metabolic pathway and is latent in humans as a marker of reductive stress. Meanwhile, rodent sperm evolved to engage this pathway with higher efficiency to support anaerobic metabolism and hypermotility.

In humans and in mice, reductive stress caused by hypoxia or alcohol intake increases BCAA fermentation. Therefore, BCHAs serve as a sensitive biomarker of reductive stress.

Meanwhile, in rodent sperm, LDHC drives BCAA fermentation with higher efficiency, supporting anaerobic metabolism and hypermotility.

100

Why sperm? Fermentation is a high-flux electron sink localized to the flagellum. Under ETC blockade, BCKAs relieved reductive stress, and maximizing fermentation boosted hyperactive motility, which is required for fertilization.

ETC inhibition with cyanide increased the NADH/NAD+ ratio in mouse sperm significantly, but supplementing with alternate electron acceptors (pyruvate or BCKAs) alleviated this reductive stress.

Forcing sperm to perform fermentation by inhibiting respiration and supplying exogenous electron acceptors (like pyruvate or BCKAs) consistently increased hyperactive motility.

100

Expression of mouse LDHC dramatically increased BCAA fermentation, but human LDHC did not. Across 12 mammalian species, every rodent LDHC sequence increased BCAA fermentation while the non-rodents failed, suggesting a specific evolution in rodent sperm.

Over-expression of mouse LDHC increases BCAA fermentation by up to 400-fold, but over-expression of human LDHC does not have this effect.

Over-expression of LDHCs from many rodents, but not from non-rodent mammals, significantly increases fermentation of BCAAs.

100

In mice, we found a surprising result! Female mice had far lower circulating levels of BCHAs, and castration of male mice also decreased the levels of these metabolites. The testes exhibited the highest absolute BCHA levels of any tissue, pointing to the sperm-specific enzyme LDHC.

Female mice had lower levels of BCHAs at baseline and in hypoxia than male mice.

100

Because BCAA fermentation increases during reductive stress, the resulting BCHAs are sensitive biomarkers of reductive stress in humans and are enriched in response to resistance exercise, severe COVID-19, and alcohol intake, all conditions involving accumulation of electrons.

Severe COVID causes reductive stress by decreasing oxygen in the blood. Resistance exercise causes reductive stress because the demand for oxygen exceeds the supply of oxygen in the blood. And alcohol consumption causes hepatic reductive stress because ethanol oxidation produces NADH faster than it can be turned over by the mitochondria.

The trigger for activating this pathway is cytosolic reductive stress, or the accumulation of electrons in the cytosol. Inhibition of the ETC or the malate-aspartate shuttle has the same effect as hypoxia, and NAD+ regeneration by LbNOX decreases the effect of hypoxia.

Expression of LbNOX, a bacterial enzyme that directly oxidizes NADH and alleviates reductive stress, decreases the fermentation of valine and isoleucine.

100

BCHAs are much less abundant than lactate, but they can still be produced by mammalian cells. So which enzyme catalyzes their production? Human metabolite GWAS pointed to LDHA, and knocking out Ldha suppressed valine and isoleucine fermentation in cells.

In a large human dataset, BCHA abundance in plasma was associated with genetic variants upstream of the LDHA locus.

Knocking out Ldha in cells decreased the fermentation of valine and isoleucine.

100

We identified another set of electron carriers: branched-chain hydroxyacids (BCHAs), which come from branched-chain amino acids (BCAAs). BCAA derivatives are fermented instead of oxidized, regenerating NAD+. This pathway is known in micro-organisms but not well-studied in mammals.

The metabolism of branched-chain amino acids starts with their conversion to branched-chain keto acids, which can subsequently be oxidized (generating NADH) or fermented (consuming NADH and regenerating NAD+)

100

Aerobic metabolism relies on the flow of electrons from fuels to NADH to oxygen. In hypoxia, mammals use anaerobic electron carriers, like lactate. We built a time- and organ-resolved hypoxia metabolome dataset to identify other electron carriers: jain-lab-ucsf.github.io/hypoxia-meta...

Mice were housed in 3 different oxygen levels for up to 3 weeks. Multiple tissues were collected at different timepoints, and untargeted metabolomics was performed. Then, we looked for the enrichment of anaerobic electron carriers.

100

Ayush Midha @ayushmidha.bsky.social · 19/03/2026

New preprint! www.biorxiv.org/content/10.6... We uncovered surprising roles for a little-known metabolic pathway, branched-chain amino acid fermentation, with implications for hypoxia, mitochondrial dysfunction, alcohol toxicity, and fertility.

122

This is super cool, congrats Andrei and Ton!

010

Reposted by Ayush Midha

Fred Hutch Basic Sciences Division @basicsci.fredhutch.org · 26/02/2026

“These awardees stood out for their scientific originality, rigor and dedication to asking important scientific questions. They reflect the strength and promise of the next generation of scientific leaders.” - Dr. Bai, director of the Weintraub Award. www.fredhutch.org/en/news/rele...

Top row, left to right: Keene Abbot, Gabriella Chua, Lifei Jiang, Won Jun Kim, Ruchita Kothari and Ayush Midha. Bottom row, left to right: Rohith Rajasekaran, Yusha Sun, Andrea Terceros, Wendy Valencia Montoya, Zachary Walsh and Peter Yoon.

153

This project has been an incredibly fun collaboration, and I’m grateful to have had the opportunity to work on this alongside Brandon Chew, Benedict Choi, Timmy Suh, and Chris Carpenter! Huge thank you to many others @gladstoneinst.bsky.social, @arcinstitute.bsky.social, and @ucsfhealth.bsky.social

000

But treating real patients with hypoxia would pose practical challenges. Recently, we showed that HypoxyStat, a small molecule that increases hemoglobin’s binding affinity for oxygen, can mimic the effects of hypoxia. And this drug also slowed tumor growth!

The small molecule HypoxyStat increases hemoglobin's binding affinity for oxygen, decreasing oxygen offloading at downstream tissues.

Daily HypoxyStat dosing suppressed tumor growth.

100

So it looks like hypoxia slows tumor growth, but could it be used alongside existing cancer therapies? We found that hypoxia could augment the tumor suppressive effects of the chemotherapy drug gemcitabine and of anti-CTLA4 immunotherapy.

Combination therapy of hypoxia and the chemotherapy gemcitabine suppressed tumor growth more than either treatment alone.

Combination therapy of hypoxia and anti-CTLA4 immunotherapy suppressed tumor growth more than either treatment alone.

100

Nucleotides can come from two sources: de novo synthesis, which is energetically demanding, and salvage, which requires abundant nucleobase substrates. We measured the contribution of each pathway; hypoxic tumors suppressed de novo synthesis, resulting in smaller purine pools.

Schematic of stable isotope-labeled tracer experiment for measuring purine synthesis pathways. De novo purine synthesis was measured using a 15N-glutamine tracer, and salvage was measured using a 13C-adenine tracer.

The results of the tracer experiment show that hypoxic tumors exhibited decreased de novo purine synthesis and relied more heavily on purine salvage.

120

Coinciding with the transcriptomics data, we observed that the smaller tumors in hypoxic mice had lower levels of purine nucleotides, which could contribute to decreased cancer cell proliferation.

Volcano plot showing enrichment or depletion of metabolites in hypoxic tumors. Nucleotide metabolites were less abundant in hypoxic tumors.

100

Using the GENEVA data, we found that a few cell lines out-competed others in hypoxia, and those cell lines tended to activate genes involved in the synthesis of new purine nucleotides.

Most cell lines were less fit in hypoxia. A few cell lines out-competed the others in hypoxia, including some renal cell carcinoma lines.

Correlation between fitness in hypoxia and expression of different genes. Genes involved in de novo purine synthesis were associated with greater fitness in hypoxia.

100

To see how hypoxia would affect different cancer cell lines, we used a platform called GENEVA, developed by @thejohnnyyu.bsky.social and @genophoria.bsky.social . We implanted 20 pooled cancer cell lines into normoxic and hypoxic mice, and performed single-cell RNA-sequencing after 8 days.

We implanted pooled tumors containing 20 different human cancer cell lines into normoxic and hypoxic mice. After 8 days, we performed single-cell RNA-sequencing to determine the fitness and transcriptomic profiles of the tumors in hypoxia and normoxia.

But little is known about how systemic hypoxia affects tumor progression, so we exposed mice with solid tumors to different oxygen levels: 21% (normoxia), 11% (moderate hypoxia) and 8% O2 (hypoxia). The tumors grew much less in the hypoxic mice!

Subcutaneous xenograft tumors grew slower in hypoxic mice compared to normoxic mice.

100

We started investigating this question based on two seemingly conflicting observations: 1. Tumor-localized hypoxia is associated with worse prognosis 2. People who live at altitude (where the air is thinner) have lower cancer mortality

Map of age-adjusted cancer mortality across the US. High altitude counties tend to have lower cancer mortality.