Wrong Unit of Survival

Stress, metabolism and the interests of the host

The Wrong Unit of Survival

Keeping a cell alive and keeping a tissue working are different achievements. Hypoxia and disturbed nitrogen metabolism show how they can come apart—and why ammonia deserves a place in the explanation.

A cell can survive a condition that makes its tissue progressively less functional. That is a simple possibility, but it changes the question we should ask about adaptation. When a response is described as protective, we need to know what it protects.

The body has many mechanisms that reconcile cellular behavior with the needs of the whole: limits on division, tissue architecture, differentiation, cell death and immune surveillance. Their success is the ordinary condition of being alive. But a response that preserves one population under stress can carry a cost elsewhere, and a response that is useful briefly can become harmful when the stress persists.

Three outcomes need to be assessed separately: the persistence of a cell population, the function of its tissue, and the health of the organism. Improvement in the first does not establish improvement in the other two.

This is an argument about how those interests can diverge. Pulmonary vascular disease offers a concrete starting point. Cancer research adds evidence about environmental stress and clonal behavior. Ammonia supplies further connections: it can participate in metabolic adaptation while changing the fitness and function of surrounding cells. The connections are experimentally supported in particular systems. Their assembly into one route from persistent stress to cancer remains a hypothesis.

Cell populationCan it persist?

Survival, proliferation and resistance to stress

TissueCan it still work?

Architecture, blood flow, repair and immune function

OrganismDoes the host benefit?

Organ function, disease burden and survival

Three questions, not one scale of success. This is a conceptual distinction, not a claim that every cellular survival response harms its host.

Part I / A response with consequences

The Alkaline Cell in a Narrowing Vessel

When resistance rises in the lung's circulation, the right side of the heart has to work harder. Several diseases can produce this problem. Pulmonary arterial hypertension, or PAH, is one category; pulmonary hypertension associated with lung disease and hypoxia is another. Low oxygen can contribute to vascular remodeling, but it does not make these diagnoses interchangeable.^[1]

Rios and colleagues examined one specific preparation in 2005: mice exposed to 10% oxygen for three weeks. They isolated smooth-muscle cells from pulmonary arteries and measured their internal pH. In bicarbonate-containing solution, basal pH rose from 7.34 ± 0.03 to 7.59 ± 0.04. Na⁺/H⁺ exchange activity and expression of the exchanger NHE1 also increased.^[2]

That result places pH regulation within the response to chronic hypoxia. It does not, by itself, show that the alkaline shift was necessary to keep the cells alive, or that the same shift caused the animal's death. The study's contribution is narrower and still important: a measurable change in cellular acid handling that could contribute to proliferation and remodeling. Its pH-recovery assay followed an imposed acid load; it was not a measurement of continuous acid export into living lung tissue.

It is tempting to make the story cleaner than that. A stressed cell defends its interior; its growth narrows a vessel; the heart pays. But each arrow requires evidence. Vascular remodeling also involves cell death, migration, extracellular matrix, inflammation and hemodynamics. A measured pH change is a mechanism to investigate within that process, not the whole explanation.

There is a second complication that helps the larger argument. In pulmonary hypertension associated with chronic lung disease, senescent vascular cells have been implicated in stimulating neighboring smooth-muscle proliferation and migration through their secretions. A cell can contribute to harmful remodeling after it has stopped dividing. Persistence, proliferation and tissue damage are related possibilities, not synonyms.^[3]

Part II / Resemblance and restraint

Clonal Growth Is Not Yet Cancer

A different set of observations comes from severe human PAH. In 1998, Lee and colleagues found a monoclonal X-inactivation pattern in 17 of 22 endothelial plexiform lesions from four patients with what was then called primary pulmonary hypertension. Nineteen lesions in the secondary-disease comparison were polyclonal. The comparison concerned disease associated with congenital heart malformations, not the hypoxic mouse smooth-muscle preparation. Four patients also remain four patients, however many lesions were sampled.^[4]

Yeager and colleagues subsequently reported microsatellite abnormalities involving genes associated with growth and apoptosis, including TGFBR2 and BAX. Those findings make the cancer comparison biologically interesting. They do not establish that ordinary hypoxic adaptation progressed through these changes in the same cells over time.^[5]

Clonality tells us about ancestry. Genetic abnormalities tell us that a population has changed. Neither observation alone establishes invasive malignancy. The relevant question is which constraints remain, which have weakened, and what behavior follows in that particular tissue.

There is no demonstrated single p53 or p16 checkpoint that divides pulmonary vascular disease from cancer. Experimental activation of p53-dependent senescence can restrain vascular remodeling, but it does not reveal the genotype or fate of every human lesion. Nor does p16 or p21 staining in part of a tissue certify that all of its proliferating cells remain constrained.^[6]

Telomerase also resists a simple answer. Work in pulmonary hypertension identified growth effects independent of detectable telomere-length changes over the experimental interval. That does not mean telomere maintenance is absent. Human idiopathic-PAH smooth-muscle cells have also been reported to have longer telomeres and greater proliferative persistence than controls. These findings justify studying replicative capacity; they do not establish immortality.^[7][8]

The comparison is useful because the diseases share some mechanisms. It becomes misleading if separate observations are treated as a single disease marching toward cancer and stopping at one missing step.

Part III / What the environment favors

Stress Can Change Which Cells Prosper

Stress can generate damage. It can also change the relative advantage of cells that already differ. Those processes can operate together. There is no need to choose between mutation and selection as if one explanation excluded the other.

In a mouse-esophagus experiment, a defined p53-mutant population existed before low-dose irradiation. Radiation changed its competitive position. A particular combination of antioxidant treatment and irradiation could reverse that advantage. The experiment concerned clonal competition; it did not show that simply removing stress prevents cancer.^[9]

Nor does this diminish the evidence for direct mutagenesis. Tobacco-associated mutational signatures demonstrate that exposure can leave a substantial record in DNA. The relative contribution of damage and selection depends on the stressor, the tissue and the stage of disease.^[10]

The same discipline applies to acid resistance. Gatenby and colleagues combined mathematical modeling, breast-cancer spheroids and human specimens to investigate how hypoxia and acidosis could favor invasive phenotypes. Their work makes pH regulation part of an evolutionary explanation. But an increase in a transporter can be a reversible response to a region's conditions; it need not be a fixed trait selected over generations.^[11]

And successful competition does not always harm the host. In mouse esophagus, expanded mutant populations in otherwise normal tissue could eliminate incipient tumors. The effect of a clone depends on what it does, where it sits and whom it competes with. “More competitive” is not a synonym for “more malignant.”^[12]

The defensible proposition is more specific: persistent disturbance can favor cell states or clones whose behavior conflicts with tissue function. Establishing selection requires following the populations. Establishing host harm requires measuring more than their number.

Part IV / Oxygen has a pattern

Reoxygenation Matters, but It Is Not a Cancer Switch

Duration is only one dimension of oxygen stress. The nadir, the frequency of fluctuations and the return of oxygen can change the response. An experiment by Verduzco and colleagues shows why that possibility deserves attention—and why its limits matter.^[13]

They repeatedly exposed an immortalized, nonmalignant breast epithelial line to 16 hours at 0.2% oxygen followed by eight hours of normoxia. After 50 cycles, characterized clones showed persistent changes, including loss of one copy of TP53 and CDH1, and greater resistance to several stresses. A chronic-hypoxia comparison did not produce the same pattern.

The altered cells nevertheless did not form tumors in the reported animal test. They also remained dependent on growth factors. This is evidence that a particular cycling regimen can produce stable, consequential changes. It is not evidence that oxygen return supplies the missing step between remodeling and cancer.

The human comparisons are harder to interpret. Altitude and obstructive sleep apnea (OSA) differ in far more than oxygen waveform. Ecological altitude findings do not measure matched tissue oxygen exposure, and sleep apnea brings differences in sleep fragmentation, ventilation, obesity and other factors.^[14]

Selected newer sleep-apnea studies: a mixed picture

Study	Population	Relevant result
Kendzerska et al., 2021[15]	33,711 participants	Severe OSA and prolonged nocturnal hypoxemia were associated with incident cancer after adjustment.
Marriott et al., 2023[16]	20,289 participants	OSA measures did not independently predict incident cancer after adjustment.
Claesson et al., 2025[17]	82,059 participants	Self-reported witnessed apnea was associated with particular cancers; oxygen exposure was not measured.

A metric needs the same care as a diagnosis. T90 means time below 90% oxygen saturation. It does not incorporate the depth of each desaturation: ten minutes at 89% and ten minutes at 70% both add ten minutes. Event-related hypoxic burden, which integrates desaturation area, is a different measure. These observations justify better exposure measurement, not a universal division between harmless sustained hypoxia and carcinogenic intermittent hypoxia.^[18]

Part V / Nitrogen changes the question

Ammonia Is More Than a Waste Product

Oxygen and acidity are only part of the environment. Cells also have to manage nitrogen. When amino acids are used and broken down, ammonia can be produced; the liver, other tissues and excretory pathways normally keep its movement and disposal coordinated. Disease can disturb that coordination.

Here the distinction between cell and host becomes unusually concrete. Nitrogen that one population can recycle into growth may burden another. A pH change tolerated in one compartment may disable a function in another. The question is not whether ammonia is universally beneficial or toxic. It is who is exposed, in which compartment, at what concentration, and with what capacity to use or remove it.

“Ammonia” also needs a definition. NH₃ is the uncharged form; NH₄⁺ is ammonium. They interconvert according to pH, and many biological assays report their combined pool. At physiological pH, most of the free pool is ammonium; a relatively small pH change can substantially alter the fraction present as membrane-permeant NH₃. Incorporating this nitrogen into glutamine or glutamate is different from retaining free NH₃/NH₄⁺. A larger intracellular pool does not, by itself, tell us what a neighboring immune cell encounters.^[19]

Low-oxygen signaling can weaken nitrogen disposal

A 2026 study provides a direct upstream connection. Cantwell and colleagues exposed primary mouse hepatocytes to 2% oxygen. HNF4α, a regulator of hepatic function, fell, followed by expression of the urea-cycle genes CPS1 and OTC; deleting HIF2α prevented these changes. In a separate in-vivo arm, genetic activation of hypoxia signaling through hepatocyte Vhl deletion increased liver and serum ammonia, while additional Hif2α deletion normalized it.^[20]

These experiments connect hypoxia signaling to impaired ammonia-disposal capacity. Actual low oxygen was tested in the cultured hepatocytes; the measured whole-animal ammonia rise came from the genetic model. The distinction matters, but the connection is substantial: nitrogen clearance can become part of the response to oxygen-related signaling.

Related work in established hepatocellular carcinoma (HCC), the most common primary liver cancer, identified another route. Hypoxia-associated reduction of SLC25A15, a mitochondrial transporter needed for the urea cycle, was linked to ammonia accumulation and altered glutamine metabolism supporting tumor progression. Here oxygen stress and nitrogen reprogramming were studied within a connected cancer model.^[21]

A metabolic burden can become a resource

In breast cancer cells, Wang and colleagues linked CtBP/SIRT4-regulated glutaminolysis to ammonia production and intracellular pH. Interfering with the pathway reduced ammonia output and increased intracellular acidity. Bicarbonate restored pH and rescued survival after CtBP suppression. This supports a contribution of ammonia-producing metabolism to acid tolerance, although the intervention changes more than ammonia alone. It does not establish a retained free-ammonia reservoir as the mechanism.^[22]

Isotope tracing supplies a different kind of evidence. Spinelli and colleagues showed breast cancer cells incorporating ammonia nitrogen into amino acids. Added ammonium supported three-dimensional growth through glutamate dehydrogenase. In their mouse model, tumor interstitial ammonia exceeded plasma concentrations: the environment around the cells could differ materially from the blood used to assess the animal. Assimilation here uses up free ammonia while making biomass.^[23]

Ammonia can also trigger a survival program. In ovarian cancer stem-like cells, ammonia exposure activated HIF1α, which slowed proliferation while protecting against death. A glutamine-synthetase-dependent response favored disposal and growth. Much of the mechanistic work used 10 mM ammonium chloride, so this should not be generalized to modest circulating elevations. But it makes the central point sharply: a cell may persist by growing less.^[24]

Another study connected ammonia to SCAP/SREBP-dependent lipid synthesis. A targeted mutation in SCAP disrupted the ammonia-dependent response and suppressed experimental tumor growth. Ammonia was functioning as a regulatory input as well as a potential nitrogen source. These experiments began with cancer cells; they explain how malignancy can be supported, not how every malignancy begins.^[25]

The immune cell can stay alive and stop doing its job

The cost to the neighborhood is not merely hypothetical. Bell and colleagues found that ammonia detoxification improved T-cell function in experimental colorectal cancer. Ornithine reduced syngeneic tumor growth in immunocompetent mice, but not in nude mice or after depletion of CD8 T cells, supporting T-cell involvement. That intervention provides a stronger causal link than finding ammonia beside exhausted-cell markers.^[26]

Domagala and colleagues identified a more specific failure. Ammonia raised the pH of secretory lysosomes and reduced mature perforin, a component of the machinery used by NK cells and cytotoxic T cells to kill targets. At millimolar concentrations, killing could be impaired while NK-cell viability remained intact. The cellular mechanism was persuasive; in-vivo ammonia-addition experiments were limited by rapid clearance and did not establish an equally clear tumor-outcome effect.^[27]

Survival is an inadequate measure of immune competence. A cytotoxic cell can remain present and alive while losing the function for which the tissue needs it.

The immune cell's own metabolism can also create a burden. A 2024 study linked ammonia generated by activated T-cell glutaminolysis to lysosomal accumulation, alkalinization, subsequent mitochondrial injury and cell death. Interfering with production or compartmental handling protected the cells and improved experimental adoptive therapy. This is an intracellular failure of nitrogen management, not necessarily poisoning by a neighboring tumor cell.^[28]

Nor do all lymphocytes respond alike. Gu and colleagues found regulatory T cells adapted to ammonia-rich liver-cancer regions through nitrogen metabolism and a spermine-associated program supporting suppressive function. Together with the effector-cell studies, this suggests an unequal consequence: the environment can impair populations that attack the tumor while favoring populations that restrain them. The observations concern established HCC; they do not date that immune imbalance to before cancer.^[29]

Acidity belongs in this account too. Extracellular acidification can impair effector T cells; ammonia has no monopoly on metabolic immune suppression. The meaningful comparison is between mechanisms and compartments, not between an immunotoxic nitrogen product and harmless protons.^[30]

The disturbed environment does not require a tumor

Evidence from established cancers leaves a temporal question: did the metabolic environment help the tumor arise, or did the tumor create it? Noncancerous liver disease supplies part of the answer. Human steatosis and steatohepatitis biopsies, together with experimental models, show reduced urea-cycle function and ammonia accumulation in noncancerous disease. These observations establish that cancer is not required for the disturbance.^[31]

Ammonia can then feed back on the diseased organ. Primary human hepatic stellate cells exposed to 50–300 μM ammonia showed activation, contraction and oxidative and endoplasmic-reticulum stress. In a rat liver-disease model, ammonia lowering reduced portal pressure. Related fatty-liver experiments linked ammonia to hepatocyte injury and fibrosis, with improvement after ammonia-lowering treatment. These are experimental contributions to tissue dysfunction; the drugs also affect other parts of metabolism.^[32][33]

Immune disturbance in this setting need not mean uniformly less activity. A 2026 study implicated ammonia produced by Clostridium perfringens in gut-barrier damage and harmful CD8-cell-mediated liver inflammation. Bacterial gene deletion, direct ammonium administration and immune interventions helped establish the pathway. There was no cancer endpoint, but there was a nitrogen-dependent alteration of the tissue environment without a tumor being present.^[34]

Small human experiments make the same distinction between activity and useful function. In cirrhosis, an ammonia-generating challenge impaired neutrophil phagocytosis compared with placebo. Related experiments found increased spontaneous oxidative activity alongside poorer engulfment of targets. Inflammation and inadequate defense can coexist. Neither observation alone establishes failure to eliminate abnormal clones.^[35]

Promotion has been tested

There are experiments in which ammonia exposure precedes a worse cancer outcome. In a gastric-carcinogenesis program, rats first received the carcinogen MNNG and then ammonia in drinking water. Ammonia increased tumor incidence and multiplicity. A separate colonic study found greater high-grade dysplasia and adenocarcinoma after ammonium acetate treatment in carcinogen-initiated rats. These are direct promotion experiments, although the exposures were local or substantial and cannot be mapped casually onto blood ammonia.^[36][37]

A 2026 liver study tested impaired nitrogen disposal alongside defined cancer drivers. Disrupting individual urea-cycle enzymes increased ammonia, accelerated HCC development and shortened mouse survival. The liver cells had also received oncogenic changes. The experiment shows cooperation between impaired nitrogen handling and those drivers; because the enzyme changes affect multiple metabolites, ammonia cannot be assigned the entire effect.^[38]

Human temporal evidence is consistent with a contribution, but less decisive. In a retrospective cirrhosis cohort containing 48,476 people with ammonia measurements, higher ammonia was associated with subsequent HCC in an adjusted two-year landmark analysis: hazard ratio 1.97, with a 95% confidence interval of 1.36–2.85. Residual confounding remains possible. The same paper's cell-transplantation experiments support tumor establishment by already malignant cells, not conversion of ordinary hepatocytes into cancer.^[39]

Ammonia has stronger support as a contributor to a susceptible environment than as a universal initiating cause. It can participate in injury, alter immune function and promote cancer after other changes have occurred. Those are meaningful roles even if it is neither necessary nor sufficient in every tumor.

The direction depends on the setting

There are important counterexamples to a one-directional story. At 1% oxygen, some cancer cells diverted glutamine nitrogen into secreted dihydroorotate, limiting ammonia accumulation. Low oxygen therefore does not invariably enlarge a free-ammonia pool.^[40]

Extracellular acidity can also protect cancer cells against ammonia toxicity. In one study, reducing medium pH from 7.5 to 6.8 mitigated effects of 4–10 mM ammonium chloride on growth and lysosomal degradation. Lower availability of membrane-permeant NH₃ was a proposed explanation. The result makes compartment and pH indispensable to the hypothesis.^[19]

Even a beneficial ammonia-lowering result is not a universal therapeutic rule. In experimental liver metastasis, treatment with L-ornithine L-aspartate (LOLA) lowered ammonia, reduced metastatic burden and changed stromal and myeloid states. Yet combining LOLA with anti-PD-L1 weakened the benefit. These were mice with introduced malignant cells, not a test of preventing cancer in healthy tissue.^[41]

Part VI / The links

A Causal Proposal, Not a Completed Sequence

The evidence now supports several connections. Hypoxia signaling can alter nitrogen disposal. Ammonia-related metabolism can support survival or biosynthesis. Ammonia exposure can impair immune killing, change stromal behavior and aggravate pre-existing disease. Experimental nitrogen perturbations can also promote tumors. What remains unestablished is that these events occur in this order, through these mediators, in the same initially nonmalignant tissue.

Measured in particular experimental systemsOxygen stress and nitrogen handling

Hypoxia / HIF signaling→Ammonia-disposal capacity

Cellular responsesSurvival and biosynthesis

pH regulation, nitrogen assimilation and signaling

Effects on other cellsImmune and stromal function

Cytotoxic impairment, regulatory-cell adaptation and tissue remodeling

Proposed connection requiring longitudinal testingDo these changes permit abnormal clones to persist and progress before cancer is established?

The solid connection and response boxes summarize findings from different models. The dashed box marks the unjoined pre-cancer hypothesis. This is not a pathway demonstrated end to end in one tissue.

Part VII / What this means

Does the Tissue Still Work?

Keeping cells alive is not enough to show that a response is good for the body. We also need to ask whether the tissue still works: can immune cells kill dangerous targets, can the liver remove metabolic waste, and can damaged tissue heal without excessive scarring? A change that improves one cell population's survival may make one of those functions worse.

Ammonia makes this problem concrete. In some experimental settings, cancer cells use its nitrogen or activate responses that help them survive. Other studies show ammonia impairing immune-cell killing or contributing to liver injury and fibrosis. These effects depend on the concentration, acidity and cells involved. But they support ammonia as a contributor to cancer.

References

Sources are listed in order of first citation. The notes identify the kind of evidence each contributes.

1. Humbert et al. (2022). 2022 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension. European Respiratory Journal.Clinical classification guideline ↩
2. Rios et al. (2005). Chronic hypoxia elevates intracellular pH and activates Na+/H+ exchange in pulmonary arterial smooth muscle cells. American Journal of Physiology—Lung Cellular and Molecular Physiology.Hypoxic mice; isolated smooth-muscle-cell measurements ↩
3. Noureddine et al. (2011). Pulmonary artery smooth muscle cell senescence is a pathogenic mechanism for pulmonary hypertension in chronic lung disease. Circulation Research.Human disease observations and cellular experiments ↩
4. Lee et al. (1998). Monoclonal endothelial cell proliferation is present in primary but not secondary pulmonary hypertension. Journal of Clinical Investigation.Small cross-sectional human lesion study ↩
5. Yeager et al. (2001). Microsatellite instability of endothelial cell growth and apoptosis genes within plexiform lesions in primary pulmonary hypertension. Circulation Research.Human lesion study; primary abstract verified ↩
6. Mouraret et al. (2013). Activation of lung p53 by Nutlin-3a prevents and reverses experimental pulmonary hypertension. Circulation.Experimental intervention ↩
7. Mouraret et al. (2015). Role for telomerase in pulmonary hypertension. Circulation.Human samples and experimental models ↩
8. Izikki et al. (2015). Telomere Maintenance Is a Critical Determinant in the Physiopathology of Pulmonary Hypertension. Journal of the American College of Cardiology.Human-cell research letter ↩
9. Fernandez-Antoran et al. (2019). Outcompeting p53-mutant cells in the normal esophagus by redox manipulation. Cell Stem Cell.Mouse clonal-competition experiment ↩
10. Alexandrov et al. (2016). Mutational signatures associated with tobacco smoking in human cancer. Science.Human tumor genomics ↩
11. Gatenby et al. (2007). Cellular adaptations to hypoxia and acidosis during somatic evolution of breast cancer. British Journal of Cancer.Modeling, spheroids and human specimens ↩
12. Colom et al. (2021). Mutant clones in normal epithelium outcompete and eliminate emerging tumours. Nature.Mouse tissue competition ↩
13. Verduzco et al. (2015). Intermittent hypoxia selects for genotypes and phenotypes that increase survival, invasion, and therapy resistance. PLOS ONE.Cell models; negative tumorigenicity result in Supplementary Figure S10 ↩
14. Simeonov and Himmelstein (2015). Lung cancer incidence decreases with elevation: evidence for oxygen as an inhaled carcinogen. PeerJ.Ecological county-level analysis; title states authors' interpretation ↩
15. Kendzerska et al. (2021). Obstructive sleep apnea and incident cancer: a large retrospective multicenter clinical cohort study. Cancer Epidemiology, Biomarkers & Prevention.Observational cohort ↩
16. Marriott et al. (2023). Does OSA Increase Risk for Cancer?: A Large Historical Sleep Clinic Cohort Study. Chest.Observational cohort ↩
17. Claesson et al. (2025). Self-reported witnessed episodes of apnea during sleep is associated with incident lung and breast cancer. Sleep Medicine.Observational cohort; self-reported exposure ↩
18. Azarbarzin et al. (2019). The hypoxic burden of sleep apnoea predicts cardiovascular disease-related mortality: the Osteoporotic Fractures in Men Study and the Sleep Heart Health Study. European Heart Journal.Definition and evaluation of a desaturation-area metric ↩
19. Dravecka et al. (2025). Low extracellular pH protects cancer cells from ammonia toxicity. Cell Death Discovery.Cancer-cell lines; pH and ammonia interventions ↩
20. Cantwell et al. (2026). Hepatocyte HIF2α downregulates the urea cycle through suppression of HNF4α. Gastro Hep Advances.Primary hepatocytes under hypoxia; genetic mouse models ↩
21. Zhang et al. (2024). Deficiency in SLC25A15, a hypoxia-responsive gene, promotes hepatocellular carcinoma by reprogramming glutamine metabolism. Journal of Hepatology.Established HCC models; primary abstract and publisher material verified ↩
22. Wang et al. (2015). CtBP maintains cancer cell growth and metabolic homeostasis via regulating SIRT4. Cell Death & Disease.Cancer-cell metabolism and pH-rescue experiments ↩
23. Spinelli et al. (2017). Metabolic recycling of ammonia via glutamate dehydrogenase supports breast cancer biomass. Science.Isotope tracing and experimental breast cancer ↩
24. Kitajima et al. (2017). Hypoxia-inducible factor-1α promotes cell survival during ammonia stress response in ovarian cancer stem-like cells. Oncotarget.Cancer-cell survival and nitrogen-metabolism experiments ↩
25. Cheng et al. (2022). Ammonia stimulates SCAP/Insig dissociation and SREBP-1 activation to promote lipogenesis and tumor growth. Nature Metabolism.Biochemistry, mutagenesis and tumor models ↩
26. Bell et al. (2023). Microenvironmental ammonia enhances T cell exhaustion in colorectal cancer. Cell Metabolism.Tumor models; ammonia detoxification and CD8-dependence experiments ↩
27. Domagala et al. (2025). Ammonia suppresses the antitumor activity of natural killer cells and T cells by decreasing mature perforin. Cancer Research.Strong cellular mechanism; limited in-vivo exposure experiment ↩
28. Zhang et al. (2024). Ammonia-induced lysosomal and mitochondrial damage causes cell death of effector CD8+ T cells. Nature Cell Biology.T-cell-intrinsic mechanism and experimental adoptive therapy ↩
29. Gu et al. (2026). Tumor-produced ammonia is metabolized by regulatory T cells to further impede anti-tumor immunity. Cell.Human HCC and mechanistic experiments; online December 2025 ↩
30. Calcinotto et al. (2012). Modulation of microenvironment acidity reverses anergy in human and murine tumor-infiltrating T lymphocytes. Cancer Research.Extracellular-pH and T-cell-function experiments ↩
31. De Chiara et al. (2018). Urea cycle dysregulation in non-alcoholic fatty liver disease. Journal of Hepatology.Noncancer human disease and experimental models ↩
32. Jalan et al. (2016). Ammonia produces pathological changes in human hepatic stellate cells and is a target for therapy of portal hypertension. Journal of Hepatology.Primary human stellate cells and rat intervention ↩
33. De Chiara et al. (2020). Ammonia scavenging prevents progression of fibrosis in experimental nonalcoholic fatty liver disease. Hepatology.Noncancer liver injury and fibrosis models ↩
34. Qu et al. (2026). Metabolic dysfunction–associated steatohepatitis exacerbated by Clostridium perfringens–derived ammonia is attenuated by tripeptide DT-109. Journal of Clinical Investigation.Microbial, metabolic and immune interventions; no cancer endpoint ↩
35. Shawcross et al. (2008). Ammonia impairs neutrophil phagocytic function in liver disease. Hepatology.Human-cell experiments and small cirrhosis challenge study ↩
36. Tsujii et al. (1995). Mechanism for ammonia-induced promotion of gastric carcinogenesis in rats. Carcinogenesis.Ammonia exposure after chemical carcinogenic initiation ↩
37. Clinton et al. (1988). Effects of ammonium acetate and sodium cholate on N-methyl-N′-nitro-N-nitrosoguanidine-induced colon carcinogenesis of rats. Cancer Research.Ammonium-salt exposure after chemical initiation ↩
38. Han et al. (2026). Impaired nitrogenous waste clearance promotes hepatocellular carcinoma. Science Advances.Urea-cycle disruption alongside oncogenic drivers in mice ↩
39. Elaimy et al. (2024). SLC4A11 mediates ammonia import and promotes cancer stemness in hepatocellular carcinoma. JCI Insight.Retrospective cirrhosis cohort plus cancer-cell and transplantation experiments ↩
40. Wang et al. (2019). Coordinative metabolism of glutamine carbon and nitrogen in proliferating cancer cells under hypoxia. Nature Communications.Hypoxic nitrogen-disposal experiments ↩
41. Sun et al. (2026). Ammonia detoxification inhibits liver metastasis by reshaping hepatic microenvironment. Advanced Science.Mouse metastatic-colonization models; combination-treatment caveat ↩

Evidence note. This essay draws on targeted primary-literature review through 9 October 2026. Experimental models, observational human findings and the proposed synthesis are distinguished in the text. It is not a systematic review; several recent mechanisms await independent replication.