How a Single Glycolytic Enzyme Decides Whether Cancer Cells Grow or Senesce

Deep inside every dividing cancer cell, an enzyme called triosephosphate isomerase 1, or TPI1, performs a deceptively simple job: it shuffles a molecule known as dihydroxyacetone phosphate, or DHAP, into its mirror image, glyceraldehyde-3-phosphate, keeping the glycolytic assembly line moving smoothly. For decades, that conversion was viewed as little more than metabolic plumbing. A new study published in Advanced Science now reveals that when this plumbing fails, the accumulating DHAP is not inert waste but an active signal—one that triggers a chain reaction running from a stress enzyme called SARM1 through calcium release and all the way to the mitochondria, where it produces a burst of reactive oxygen species strong enough to force tumor cells into a permanently arrested, senescent state.

The research team, led by investigators at the Chinese PLA General Hospital, began with an unbiased hunt rather than a hypothesis. Using a customized CRISPR-Cas9 library targeting 1,649 metabolic enzymes and regulators—6,596 guide RNAs plus controls—they systematically disabled each gene in A498 clear cell renal cell carcinoma cells and tracked which losses crippled proliferation over 21 days. TPI1 emerged as one of the most robustly depleted hits, sitting at a hub position among the top ten negatively selected genes. The finding was more than a screening artifact: tissue microarrays built from 234 renal tumors and 172 adjacent normal specimens showed markedly elevated TPI1 protein in cancer tissue, a result corroborated by data from the TCGA and an independent cohort. TPI1 expression even climbed with tumor grade, alongside rising frequencies of gain-of-function alterations.

To confirm that TPI1 truly drives tumor growth, the researchers shut it down with short hairpin RNAs in renal cancer cells and watched proliferation collapse in growth curves, EdU incorporation assays, and colony formation tests. In mice bearing xenografts of TPI1-deficient OS-RC-2 cells, tumors were significantly smaller and lighter, with reduced Ki67 staining marking the slowdown in cell division. Crucially, deleting TPI1 caused DHAP to pile up inside cells to roughly ten times normal levels. When the team replaced endogenous TPI1 with a wild-type copy, growth recovered; when they substituted a catalytically dead K14M mutant that cannot handle DHAP, it did not. The lesson was unambiguous: it is TPI1’s enzymatic activity, not some scaffold function, that cancer cells depend on.

Because DHAP cannot cross cell membranes and no eukaryotic transporter for it is known, the team exploited a clever chemical workaround. They fed cells dihydroxyacetone, a membrane-permeable precursor that is converted intracellularly into DHAP by the enzyme TKFC. Supplementing with dihydroxyacetone suppressed cell viability, and silencing TKFC blunted that toxicity—direct evidence that the accumulating upstream metabolite, rather than the loss of the downstream product GAP, was doing the damage. Indeed, adding GAP itself to TPI1-deficient cells offered no rescue at all. The experiment repositioned DHAP in the scientific imagination: no longer a passive glycolytic intermediate, but a molecule with signaling consequences, echoing recent work showing that DHAP can inform the nutrient-sensing complex mTORC1 about glucose availability.

What actually happens inside a cell drowning in DHAP turned out to be the study’s most striking revelation. Transcriptome sequencing of TPI1-deficient cells showed positive enrichment of inflammatory signaling, the p53 pathway, and a senescence-associated secretory phenotype gene set, alongside suppression of E2F targets and G2/M checkpoint programs. The cells stained positive for senescence-associated beta-galactosidase, accumulated the DNA damage marker gamma-H2AX, activated the p53-p21 axis, lost Lamin B1 from their nuclear envelopes, and arrested in the G0/G1 phase. Luminex multiplex analysis and ELISA confirmed a flood of inflammatory secretions—IL-6, IL-8, TNF-alpha, and GM-CSF—the classic SASP signature. In xenografted tumors, the same markers rose in tissue sections. This was not cell death; it was cellular senescence, a durable proliferative arrest that oncologists have long hoped to impose on tumors deliberately.

The engine behind that arrest was mitochondrial. TPI1 knockdown sharply increased overall reactive oxygen species, and probes specific for mitochondrial superoxide and hydrogen peroxide showed the origin clearly: mitochondrial ROS surged. A mitochondria-targeted antioxidant, Mito-TEMPO, brought intracellular ROS back down, reduced the fraction of beta-galactosidase-positive cells, and lowered gamma-H2AX, p53, and p21—mimicking the effects of the general antioxidant NAC. Mitochondrial oxidants were therefore acting as the critical effector, converting a metabolic glitch into DNA double-strand damage and, ultimately, senescence. The remaining question was what had switched the mitochondria into this overdrive state in the first place.

Integrated metabolomic and transcriptomic analysis pointed to calcium. TPI1-deficient cells showed elevated intracellular calcium by both colorimetric assay and fluorescent imaging with Fluo-4 AM. When the researchers compared two chelators, the result was decisive: BAPTA-AM, which binds calcium inside the cell, rescued viability and dampened senescence markers, SASP secretion, and DNA damage, while EGTA, which only chelates extracellular calcium, did essentially nothing. In living tumors, peritumoral BAPTA-AM injections let TPI1-deficient xenografts grow significantly faster than vehicle-treated controls. The signal itself proved to be cyclic ADP-ribose, or cADPR, a second messenger that opens ryanodine receptors on the endoplasmic reticulum and liberates calcium into the cytosol. Metabolomics revealed cADPR accumulating after TPI1 loss, and providing cADPR directly to healthy cells was sufficient on its own to induce calcium overload, beta-galactosidase activity, DNA damage, p53-p21 activation, and SASP release—all of which BAPTA-AM reversed. Conversely, blocking cADPR signaling with the antagonist 8-Br-cADPR suppressed the calcium rise and the senescence program triggered by TPI1 depletion.

Upstream of cADPR, the team identified an unexpected culprit: SARM1, a sterile alpha and TIR motif containing protein long known for its role in axonal degeneration and increasingly appreciated as an NAD-consuming enzyme with catalytic activity rivaling the classical ADP-ribosyl cyclase CD38. In TPI1-deficient cells, CD38 was undetectable, but SARM1 rose markedly—dihydroxyacetone treatment produced the same effect. Overexpressing SARM1 alone reproduced the entire phenotype, from gamma-H2AX and p53-p21 induction to Lamin B1 loss, ROS generation, and SASP secretion, while suppressing proliferation and clonogenic capacity. Removing SARM1 from TPI1-deficient cells restored viability, normalized calcium and cADPR levels, reversed the DNA damage and redox stress, and let xenografts grow faster in mice. Notably, mTORC1 appeared to feed into this circuit: TPI1 loss activated the pathway, read out through phospho-S6K1, and the mTORC1 inhibitor rapamycin attenuated SARM1 induction—yet forced SARM1 expression still drove senescence signaling even under rapamycin, placing SARM1 centrally in the cascade with mTORC1 as a modulator.

The mechanism appears to travel well beyond kidney cancer. Knocking down TPI1 in prostate cancer C4-2, lung cancer A549, and bladder cancer T24 cells likewise curbed viability, raised beta-galactosidase activity, elevated SARM1 and gamma-H2AX, depleted Lamin B1, and increased IL-6, IL-8, and TNF secretion, with TPI1-deficient bladder cancer xenografts growing more slowly in mice. In each line, cADPR and calcium climbed after TPI1 loss, and both BAPTA-AM and 8-Br-cADPR blunted the senescence phenotype, as did SARM1 knockdown—a conserved SARM1-cADPR-calcium axis operating across tumor types.

The implications cut in two directions. For patients with TPI1 deficiency, the rare autosomal recessive disorder caused most often by the E104D mutation, the study offers a mechanistic clue: neither patients nor fly models show overt ATP depletion, and the new work suggests their hemolytic anemia and neurodegeneration may instead stem from a redox-calcium pathway in which DHAP overload drives mitochondrial oxidant stress. For oncology, the findings sketch a metabolite-driven vulnerability: suppressing TPI1, or pharmacologically reinforcing the SARM1-cADPR-calcium-mtROS cascade it controls, might push tumors into senescence deliberately, potentially priming them for immune clearance, while calcium chelation or cADPR antagonism could protect vulnerable tissues. The authors caution that the sensor connecting DHAP to SARM1 remains unidentified, that rapamycin only partially suppressed the phenotype, and that whether the cascade is fully reversible is unknown. Still, the conceptual shift is hard to miss: a century after glycolysis was first mapped in tumors, one of its smallest intermediates has turned out to be a messenger—and cancer cells, it seems, listen closely.

Subject of Research: Metabolite-driven mitochondrial redox signaling linking TPI1 loss to cellular senescence via the SARM1–cADPR–Ca2+ axis in cancer

Article Title: TPI1 Loss Triggers a Metabolite‐Driven Mitochondrial Redox Vulnerability via the SARM1–cADPR–Ca2+ Axis

Article References: Liu, C., Wu, S., Wang, C., Zhou, Z., Cai, T., Tao, W., Zuo, S., Zhang, C., Dong, Y., Feng, Y., Huang, Q., Wang, B., Ma, X., Ma, H., Zhang, X., & Huang, Y. (2026). TPI1 Loss Triggers a Metabolite‐Driven Mitochondrial Redox Vulnerability via the SARM1–cADPR–Ca 2+ Axis. Advanced Science, 13(56), Article e76614. https://doi.org/10.1002/advs.76614

Image Credits: AI Generated

DOI: 10.1002/advs.76614

Keywords: TPI1, DHAP, SARM1, cADPR, calcium signaling, mitochondrial ROS, cellular senescence, clear cell renal cell carcinoma, CRISPR screen, glycolysis, p53-p21 pathway, mTORC1