Cancer metabolism

Cancer metabolism · from mechanism to treatment

What Makes a Tumour Acidic?

From acid production and poor clearance to combinations that could restore immune killing.

In a 1998 mouse experiment, tumours grown from cells deficient in lactate dehydrogenase contained about 76% less lactate than their parental counterparts. Yet both groups had the same mean extracellular pH: 7.03. Normal muscle measured approximately 7.43. Much of the lactate was gone; the acidic environment remained.^[1]

That result matters for treatment. A drug may suppress a metabolic pathway without removing the condition that makes a tumour hostile to immune cells. To change that environment, we need to understand what sustains it.

The answer involves more than acid production. Cells move acid across their membranes. Carbon dioxide exchanges with bicarbonate. Buffers bind protons, and blood flow helps carry metabolic products away. Tumour acidity emerges from the balance among these processes. Each offers a possible point of intervention; each requires its own evidence.

Find what sustains the hostile environment, change it measurably, and pair that change with a treatment that can exploit it.

From mechanism to intervention

One tissue. Three places to intervene.

Change the extracellular environment, disrupt cancer-cell pH control, or interrupt the immune response to acidity.

2

Cancer-cell interior

Metabolism and ATP turnover generate an acid-handling demand.

Substrate oxidation → CO₂CO₂ diffuses across the membrane ↓

Lactate + H⁺MCT transporters couple their export ↓

Extracellular space

CA IX acts at the outer cell surface

CO₂ + H₂O⇅ CA IX accelerates exchangeHCO₃⁻ + H⁺

↩ Bicarbonate can return to the cancer cell through transporters, supporting intracellular buffering.

Local buffers bind acid equivalents. Diffusion and vascular exchange influence what accumulates.

3

Nearby immune cell

Extracellular H⁺ → GPR65

Acid sensing can change immune function. Blocking that signal is a candidate strategy.

1

Nearby capillary

↑ Buffer deliveryFrom blood to tissue↓ Metabolite clearanceFrom tissue to blood

Blood also delivers oxygen and substrates, so changing perfusion can change production and removal together.

1

Relieve extracellular acidity

Buffering, transport interventions and functional perfusion can alter the external environment.

2

Disrupt cancer-cell pH control

CA IX and MCT inhibition can interfere with the mechanisms that protect the cell interior.

3

Interrupt immune acid sensing

GPR65 is a candidate target for reducing acidity-driven immune suppression.

Conceptual tissue map showing candidate intervention sites. CA IX accelerates reversible chemistry; MCTs couple lactate and proton movement, shown here during export. Buffering binds acid equivalents, while vascular exchange helps remove them. Arrows show relationships, not measured fluxes or proven treatment effects.

The tumours that stayed acidic

Five years before the LDH experiment, Newell and colleagues had studied a different glycolysis-deficient cell line. In culture, its lactate-production rate was less than 4% of the parental rate. Tumours grown from those cells still had a mean extracellular pH of 6.78, compared with 6.65 in parental tumours and 7.29 in muscle. The difference between the tumour groups was not statistically significant (P = 0.13).^[2]

These experiments established a useful starting point: severely impairing tumour-cell lactate production did not eliminate extracellular acidity. They did not assign percentages of the acid load to different sources. A pH measurement describes hydrogen-ion activity in a buffered environment; identifying the processes that maintain it requires additional measurements.

Helmlinger and colleagues took that next step in 2002. Alongside measurements in a tumour-window model, they sampled tumour interstitial fluid and examined lactate, CO₂, bicarbonate and pH. Glycolysis-impaired tumours again remained acidic. Their interstitial lactate was lower, while CO₂-related measurements were similar to those in parental tumours. The study provided evidence that CO₂ also contributes to tumour acidity.^[3]

The important shift is from following one metabolite to measuring the environment as a whole. Lactate concentration, CO₂ production and extracellular pH describe different features of that environment. Their relationships can change when metabolism adapts.

Two routes converge outside the cell

Glycolytic metabolism has a clear route to extracellular acidification. Monocarboxylate transporters, including MCT1 and MCT4, couple lactate movement to proton movement. During export, they carry lactate and H⁺ into the surrounding fluid. The pyruvate-to-lactate reaction itself consumes a proton; the net acid balance belongs to the broader metabolic system, including ATP turnover and membrane transport.

Oxidation provides another route. Carbon dioxide crosses cell membranes and participates in a reversible reaction:

CO₂ + H₂O ⇌ HCO₃⁻ + H⁺

Carbonic anhydrase IX, or CA IX, accelerates this chemistry at the outer face of the cancer-cell membrane. Often induced through hypoxia signalling, it works with bicarbonate transporters and other pH regulators. Bicarbonate imported into the cell can neutralize intracellular acid; the resulting CO₂ can diffuse outward again. This coupling can support intracellular pH while contributing to extracellular acidification.^[4]

There is direct evidence that changing CA IX expression changes tumour pH. In Lee and colleagues’ colorectal-cancer xenograft model, CA IX-expressing tumours had a median extracellular pH of 6.71, compared with 6.86 in controls. Intracellular pH did not differ detectably. A membrane-associated regulator had altered the external environment without a corresponding measured shift inside the cells.^[5]

The enzyme accelerates a reversible reaction rather than changing its equilibrium. Its tissue-level effect depends on ongoing metabolism, transport and exchange with blood. This makes CA IX a component of a coupled system—and a target whose effect needs to be measured in that system.

What accumulates depends on what can leave

A tumour does not have to produce more CO₂ than normal tissue for CO₂ to accumulate locally. Production and removal both matter. If clearance falls sufficiently, even a lower production rate can coexist with a higher local concentration.

Blood flow helps deliver bicarbonate and remove metabolic products. It also brings oxygen and nutrients that support further metabolism. Improving perfusion therefore changes several terms at once. The resulting pH depends on which changes dominate, where they occur and how long they persist.

This helps explain why oxygenation and acidity need separate maps. Helmlinger’s high-resolution measurements found discordant local oxygen and pH profiles. Molavian’s subsequent modelling examined how heterogeneous delivery of oxygen and buffers could generate such patterns. Together, they support treating acidity as a tissue-transport problem as well as a metabolic one.^[6][7]

The same principle applies to experiments in culture. Gas exchange, buffer composition and measurement time affect how much respiratory CO₂ contributes to acidification. Mookerjee and colleagues measured substantial but variable respiratory contributions in cell assays. A falling medium pH is therefore not, by itself, a measurement of lactate production.^[8]

The glucose arithmetic, in brief

Complete oxidation yields six CO₂ per glucose; conversion to lactate directly yields none. If a hypothetical cell ferments 90% of its glucose and fully oxidizes 10%, its yield is 0.6 CO₂ per glucose. At 5×, 10× or 20× the glucose-consumption rate of a fully oxidizing comparator, its glucose-derived CO₂ output becomes 0.50×, 1.00× or 2.00×. A 95%-oxidizing comparator instead gives 0.53×, 1.05× or 2.11×.

This illustrates how high flux can offset low yield. It is not a measured tumour average, and other fuel contributions belong on both sides of the comparison. Human lung-tumour tracing has demonstrated enhanced glycolysis alongside enhanced glucose oxidation.^[9] FDG uptake alone establishes neither the oxidation fraction nor a universal glucose-consumption multiplier.^[10]

Changing the environment is only the first step

An acidic tumour can contain immune cells that recognize cancer yet struggle to attack it. Calcinotto and colleagues showed that acidic conditions impaired human and mouse tumour-specific T-cell functions, with recovery when pH was restored in the experimental system. This makes acidity a potentially reversible barrier to immune killing.^[11]

Removing that barrier does not guarantee tumour eradication. The immune response may still lack sufficient recognition, activation or access. A cancer cell whose pH regulation has been disrupted may stop growing without dying. These are reasons to combine treatments—but cancer’s complexity alone does not tell us which treatments belong together.

A useful combination should remove distinct barriers, or block an adaptation caused by its first component. For an acidity intervention, the clearest partner is often an immune treatment capable of exploiting the changed environment. Additional components should earn their place by addressing a demonstrated remaining problem: ammonia, poor delivery, antioxidant defence or suppressive immune signalling.

The clearest starting point: acidity intervention plus immune killing

Bicarbonate has direct preclinical support. Robey and colleagues found that oral bicarbonate raised tumour extracellular pH and reduced metastases in selected mouse models, with effects that varied across models.^[12] Pilon-Thomas and colleagues subsequently showed improved responses to selected checkpoint inhibitors and adoptive T-cell therapy when these treatments were combined with bicarbonate. Their buffering began before tumour inoculation, an important limitation when considering treatment of an established cancer.^[13]

The logic is complementary: buffering aims to make the environment more permissive, while checkpoint blockade releases an inhibitory signal restraining an existing immune response. Adoptive therapy supplies tumour-reactive cells. Neither immune partner should be expected to overcome every barrier merely because average pH rises.

Bicarbonate’s human record also needs precise wording. Small studies encountered palatability, gastrointestinal and adherence problems and did not establish anticancer efficacy. This leaves both delivery and biological effectiveness unresolved; it does not prove that every method of correcting tumour acidity will fail.^[14] A second drug cannot compensate for a buffer that never changes the relevant tumour regions.

Another route is to reduce acid export. In a mouse colorectal-cancer study, MCT4 inhibition increased pH measured in dissected tumours after treatment and enhanced checkpoint-blockade effects.^[15] The attractive setting would be a tumour with a demonstrable MCT4-dependent acidic environment and an immune response that recovers when that environment changes. This is a proposed selection rule, not a validated clinical biomarker.

Transporter selectivity matters. Deleting MCT1 specifically in regulatory T cells weakened their intratumoural function and improved immunotherapy responses in mice.^[16] A systemic MCT1 inhibitor would act on additional cell types. Lactate can also be used as a fuel, while lactic acidosis can impair effector T cells; those conditions should be separated experimentally.^[17] The aim is to improve the balance between cancer and immune cells, rather than simply minimize a metabolite everywhere.

CA IX inhibition: decide which compartment should change

Carbonic anhydrase inhibition offers another immune combination, but its objective needs care. Chafe and colleagues found that SLC-0111, an inhibitor of tumour-associated carbonic anhydrases including CA IX and CA XII, enhanced checkpoint-blockade effects in mouse melanoma and breast-cancer models. Their extracellular-pH experiment was performed in culture; the animal results did not establish extracellular alkalinization as the mechanism of benefit.^[18]

This distinction is consequential. In pancreatic-cancer xenografts, Nakaoka and colleagues observed slower tumour growth with U-104, also known as SLC-0111, while extracellular pH tended to become more acidic.^[19] A treatment can damage cancer-cell pH control without producing a more alkaline extracellular environment.

There are therefore two different combination proposals. One seeks extracellular-pH correction to support immunity. The other attacks the cancer cell’s ability to maintain its interior, then adds immune killing or a second metabolic stress. Both are plausible. They require different measurements, and evidence for one should not be used as proof of the other.

Bicarbonate plus phenylbutyrate: address acidity and ammonia together

This pairing asks a useful question: what if correcting acidity leaves a second metabolic obstacle untouched? A tumour may contain both an acidic extracellular environment and excess ammonia. Improving pH alone would not necessarily restore immune cells exposed to damaging nitrogen metabolism.

Bell and colleagues linked microenvironmental ammonia to T-cell exhaustion in colorectal-cancer models. Glycerol phenylbutyrate reduced MC38 tumour size in a supplementary experiment. The study’s direct checkpoint-combination evidence, however, concerned ornithine plus anti-PD-L1, not phenylbutyrate. These are separate interventions, and their results cannot be interchanged.^[20]

Human hepatocellular-carcinoma tissue provides a further mechanistic reason to investigate ammonia. Gu and colleagues identified glutaminolytic, ammonia-rich regions in which effector T cells were disadvantaged while regulatory T cells used nitrogen-handling and metabolic adaptations to persist. Their experimental work also implicated ammonia released after anti-PD-1 treatment in reinforcing regulatory-cell suppression. This supports ammonia as a potential resistance mechanism; it does not establish phenylbutyrate as its solution.^[21]

Phenylbutyrate supplies an alternative route for nitrogen disposal. It is converted to phenylacetate, which conjugates with glutamine; the resulting phenylacetylglutamine is excreted in urine. It therefore removes nitrogen through metabolism and excretion rather than directly trapping CO₂ or neutralizing extracellular acid. Glycerol phenylbutyrate and sodium phenylbutyrate deliver the same precursor through different formulations, with different delivery characteristics and sodium burdens.^[22]

The proposed roles are distinct: bicarbonate changes extracellular acid–base conditions; a phenylbutyrate formulation lowers harmful ammonia exposure, if effective in that tumour; an immune treatment supplies the killing response. The targeted literature search did not identify a cancer study directly testing bicarbonate plus phenylbutyrate. The pair—and especially its extension to checkpoint blockade—is a research hypothesis assembled from separate lines of evidence.

What could prevent ammonia toxicity as pH rises?

The chemistry explains why nitrogen handling belongs in the experiment. At a fixed temperature, a rise of 0.3 pH units approximately doubles the NH₃:NH₄⁺ ratio. Uncharged NH₃ crosses membranes readily. In cancer-cell culture, Dravecka and colleagues found that low extracellular pH protected cells against ammonia-associated lysosomal dysfunction and growth inhibition.^[23]

A nitrogen scavenger is therefore a logical candidate to reduce unwanted ammonia exposure during pH correction. Reducing tumour ammonia production is another possibility. For adoptive therapy, protecting the transferred T cells themselves offers a more selective direction. None of these additions has been shown to solve this specific buffering problem in patients.

There is also a potential trade-off: lowering ammonia could protect immune cells while rescuing cancer cells from toxicity exposed by the pH increase. That is an inference from the separate findings, not demonstrated antagonism between bicarbonate and phenylbutyrate. The decisive outcome is better immune-mediated cancer killing, assessed alongside the survival of each cell population.

Phenylbutyrate is also an imperfect mechanistic probe because it has histone-deacetylase and chemical-chaperone activities. In one lung-cancer cell study, 4-phenylbutyrate attenuated endoplasmic-reticulum stress and partially protected cells from an experimental cisplatin-containing combination.^[24] Its partner should therefore be chosen with attention to how that partner kills. Lower ammonia alone would be insufficient evidence of a useful combination.

Attack the adaptation created by the first treatment

An effective second component can block a defence induced by the first. Venkateswaran and colleagues found that CA IX inhibition increased glutamine uptake and glutathione-dependent antioxidant protection in hypoxic tumour cells.^[25]

Adding CB-839, a glutaminase inhibitor, or interventions against glutathione synthesis or GPX4 enhanced ferroptotic killing in culture; the CB-839 pair was also tested in spheroids. Animal evidence used SLC-0111 plus inducible GCLC knockdown, not two-drug CB-839 treatment.^[25]

The proposed selection criterion is the induced defence: measure it, then test a matched second intervention. These findings concern redox adaptation; they establish neither extracellular-pH correction nor reduced CO₂ production.

Glutamine-directed treatment can also have an immune partner. In KEAP1-mutant lung-cancer mouse models, the glutamine-antagonist prodrug DRP-104 enhanced responses to anti-PD-1 treatment.^[26] That study supports a genotype-defined metabolic combination. It does not establish ammonia reduction or pH correction as the explanation, and a broad glutamine antagonist should not be treated as equivalent to a selective glutaminase inhibitor.

This is a candidate alternative to the scavenger approach: reduce a tumour’s dependence on—and possibly its ammonia production from—glutamine metabolism. The ammonia effect remains a hypothesis requiring measurement. Because immune cells also need metabolic resources, preserving their function belongs in the same experiment.

Improve delivery and clearance where the tissue is the bottleneck

If poorly functioning vessels prevent a buffer or immune treatment from reaching viable tumour regions, adding another metabolic drug may leave the central obstacle intact. A vascular intervention could improve access and exchange. Its effect on pH remains uncertain because improved blood flow can increase metabolic supply as well as removal.

Park and colleagues combined the vascular-normalizing compound CU06-1004 with anti-PD-1 in mouse tumours. They reported improved delivery and cytotoxic T-cell activity, reduced hypoxia and better tumour outcomes. These results support a vascular–immune combination, but did not demonstrate that pH rose or CO₂ clearance increased.^[27]

Adding buffering to such a combination is a logical extension for a tumour with documented poor perfusion and acidity. It is still a proposed three-component experiment. The vascular intervention would first need to improve functional exchange; buffering would then need to change extracellular pH; immune treatment would need to exploit those changes. Dosing sequence should follow the measured delivery window.

Local buffer delivery offers a different solution. Abumanhal-Masarweh and colleagues used liposomal sodium bicarbonate in a 4T1 breast-cancer mouse model, measured increased tumour pH and improved doxorubicin uptake, and found greater tumour control with the combination.^[28] Here the partner had a specific physicochemical rationale: extracellular acidity can hinder entry of a weak-base drug. This result should not be generalized to every chemotherapy, and it does not establish that the same formulation would improve immunotherapy.

The broader lesson is that formulation can be part of the combination. Getting a physiological intervention into the right tissue may matter as much as adding another target.

Prevent the immune system from interpreting acidity as a suppressive signal

Another strategy is to change the immune response to acidity while the tissue remains acidic. The proton-sensing receptor GPR65 offers an example. Truong-Andrievici and colleagues found that its genetic disruption reduced growth of strongly acidic B16F10 melanoma but not the less acidic MC38 model. Pharmacological inhibition reduced melanoma growth and suppressive myeloid-cell recruitment.^[29]

This suggests combining acid-sensing inhibition with immune killing in a suitable tumour. The direct checkpoint combination in that study, however, used a CXCR2 antagonist plus anti-PD-1 to address myeloid recruitment; it was not a demonstration of GPR65 inhibitor plus anti-PD-1. That distinction identifies what remains to be tested.

Context may determine the direction of benefit. A separate study found that loss of GPR65 worsened inflammation and tumour progression in mouse colon-cancer models.^[30] Selecting a treatment merely because a tumour is acidic would therefore be insufficient. The relevant receptor-bearing cells, their suppressive behaviour and the consequences of blocking the signal need to be established.

Build immune cells that can work in the environment

It may be possible to improve an adoptive immune treatment without correcting every region of a tumour. Navarro and colleagues engineered T cells to overexpress the proton channel HVCN1. The cells better tolerated acidic conditions, and engineered antigen-specific T cells and CAR T cells improved tumour outcomes in mouse experiments.^[31]

Ammonia resistance offers a complementary direction. Zhang and colleagues showed ammonia-associated lysosomal and mitochondrial injury in effector CD8 T cells. In a B16-OVA adoptive-transfer experiment, ex-vivo pretreatment with either DON or the lysosome-acidifying compound C381 improved the therapeutic performance of the transferred cells.^[32] Those interventions and their schedules are experimental; the result does not establish systemic administration as an equivalent strategy.

A more exploratory proposal would combine acid tolerance with independently validated ammonia protection in the same therapeutic cells. The dual strategy has not been established by those separate studies. Its attraction is selectivity: improve the cells intended to kill the tumour while minimizing protection of the cancer. Tests would need to confirm antigen-specific killing, persistence and normal activation control, rather than assuming that greater stress tolerance is always beneficial.

Combination options: what has actually been tested?

Combination	Why pair them?	Evidence boundary
Bicarbonate + checkpoint blockade or adoptive T cells	Relieve an environmental barrier while strengthening immune killing.	Direct mouse combinations. Buffering began before tumour inoculation in the cited immunotherapy study; human efficacy is unestablished.
MCT4 inhibition + checkpoint blockade	Reduce a transport-associated barrier and improve immune response.	Direct mouse combination. Post-treatment tissue pH was measured; clinical benefit remains unestablished.
SLC-0111 + checkpoint blockade	Disrupt tumour carbonic-anhydrase biology while releasing immune inhibition.	Direct mouse combinations. Animal benefit does not establish extracellular alkalinization as its cause.
Bicarbonate + phenylbutyrate, then an immune partner	Address extracellular acidity and harmful ammonia exposure together.	Proposed combination. Separate preclinical support; no direct test of the exact buffer–scavenger pair identified.
CA IX inhibition + glutamine or glutathione intervention	Block induced antioxidant defence.	Direct cell combinations. The animal combination used genetic GCLC suppression.
Vascular normalization + immune treatment, with buffering if needed	Improve access, exchange and immune killing in poorly perfused tissue.	Direct mouse vascular–immune pair. Adding a buffer and attributing benefit to acid clearance remain hypotheses.
Liposomal bicarbonate + doxorubicin	Change local pH to improve delivery into cancer cells.	Direct mouse combination with pH measurement. Evidence is specific to the formulation, drug and model.
Acid-tolerant + ammonia-protected adoptive cells	Preserve the killing cells despite two environmental stresses.	Proposed dual protection. The individual engineering or pretreatment approaches have separate mouse support.

These are research options with different levels of support. A favourable combination result does not by itself prove synergy or identify the mechanism responsible.

What has reached patients?

AZD3965: the MCT1 inhibitor showed target engagement in a 40-patient phase I dose-escalation study. Dose-limiting toxicities included ocular changes, cardiac troponin elevation and acidosis. It does not inhibit MCT4.^[33]

SLC-0111: a 17-patient phase I study reported no objective responses; two patients had stable disease beyond 24 weeks.^[34] A subsequent CA IX-positive pancreatic-cancer trial with gemcitabine enrolled six participants before termination in 2024 as the treatment landscape changed; no results are posted.^[35]

Phenylbutyrate: its established nitrogen-disposal use in urea-cycle disorders does not establish cancer efficacy or the safety and effectiveness of the proposed combinations.^[22]

Early human target engagement, practical tolerability and anticancer efficacy answer different questions. The combinations proposed here require evidence for all three before they could become treatment options.

A study that could resolve the bicarbonate–phenylbutyrate question

The most informative starting model would have an established tumour, an intact immune system, measurable extracellular acidity and elevated local ammonia. A model with only one of those barriers would be a poor test of the two-barrier hypothesis. Include a contrasting model to examine where the combination should fail.

1. Map both exposures. Measure extracellular pH and total ammonia in viable tumour regions, alongside perfusion and oxygenation. Estimate the uncharged NH₃ fraction from local conditions, with appropriate uncertainty. Plasma ammonia alone cannot establish what immune cells encounter inside a tumour.
2. Test the pair before building the triple. Compare control, bicarbonate, a defined phenylbutyrate formulation and both, using matched established tumours. Measure early changes before tumour-size differences confound interpretation. Check nitrogen disposal through urinary phenylacetylglutamine, together with tumour and plasma glutamine and ammonia.
3. Add immune killing in a factorial experiment. Once both intended physiological effects are demonstrated, test those four conditions with and without checkpoint blockade: eight groups. Prespecify the interaction analysis, cancer control, survival and toxicity outcomes. Improvement over one component alone is not sufficient to call the result synergistic.
4. Find out which cells were protected. Combine controlled-pH and ammonia experiments in cancer-cell cultures, isolated immune cells and tumour–immune cocultures. Measure cancer-cell survival, effector-cell survival and antigen-specific killing. Use appropriate immune-depletion experiments to assess whether tumour control depends on the proposed immune mechanism.

The central comparison is whether the scavenger adds useful immune recovery to buffering, and whether that gain exceeds any protection it gives the cancer. A measured reduction in ammonia with worse tumour control would argue against that combination, even if an immune marker improved.

Timing is a testable variable. If ammonia lowering is demonstrated, establishing it before pH correction could reduce an early exposure increase; this proposed sequence has not been validated. Likewise, a vascular delivery window or an induced antioxidant response may determine when another partner should be given.

Measurement must fit the mechanism. Hyperpolarized ¹³C-bicarbonate can report tumour pH, but interpretation depends on bicarbonate–CO₂ label exchange. Carbonic-anhydrase inhibition changes that exchange, potentially altering the readout independently of pH. An independent method or validated kinetic correction is needed when the enzyme is being perturbed.^[36][37]

Let the result determine the next combination

pH corrected; ammonia remains high; immune cells remain impaired

Test a nitrogen-disposal or production-reducing partner, after confirming that ammonia is contributing to the impairment.

pH and ammonia improve; cancer killing does not

Look for inadequate recognition, additional immune suppression or cancer-cell protection. More environmental correction may not address the remaining barrier.

Target engaged; the intended physiological change is absent

Investigate delivery, exchange or compensation before adding treatments on the assumption that the first component worked.

Cancer control improves; extracellular pH is unchanged

Pursue the demonstrated intracellular, metabolic or signalling effect. The combination may be useful without being an alkalinizing treatment.

More targets are useful only when they close a real escape route

Even two apparently complementary pH targets can disappoint. Parks and colleagues found that separate genetic loss of NHE1 or CA IX markedly reduced colon-cancer xenograft growth. Yet tumours lacking both grew faster than either single-knockout group, alongside increased CA XII expression. Compensation was a proposed explanation, not a proven complete mechanism.^[38]

This is why combinations should follow the tumour’s response. If buffering restores immune function, pair it with effective immune killing. If ammonia remains a barrier, test whether nitrogen disposal helps. If disrupted pH control induces antioxidant defence, target that defence. If delivery fails, solve the tissue-access problem. These are different biological situations, each with a different next experiment.

The therapeutic opportunity lies in making a tumour less able to survive and more vulnerable to attack. A strong combination connects each component to a measured obstacle—and demonstrates that removing those obstacles produces better cancer killing.

References

Primary studies and their evidence settings. Numbered citations link to the corresponding source.

1. Yamagata M et al. (1998). The contribution of lactic acid to acidification of tumours: studies of variant cells lacking lactate dehydrogenase. British Journal of Cancer 77:1726–1731.LDH-deficient variant cells and mouse tumours; carbonic-acid contribution proposed. ↩
2. Newell K et al. (1993). Studies with glycolysis-deficient cells suggest that production of lactic acid is not the only cause of tumor acidity. PNAS 90:1127–1131.Cell experiments and mouse tumours; microelectrode pH measurements. ↩
3. Helmlinger G et al. (2002). Acid production in glycolysis-impaired tumors provides new insights into tumor metabolism. Clinical Cancer Research 8:1284–1291.Cell and animal models; acidification beyond lactate production. ↩
4. Swietach P et al. (2009). The role of carbonic anhydrase 9 in regulating extracellular and intracellular pH in three-dimensional tumor cell growths. Journal of Biological Chemistry 284:20299–20310.HCT116 tumour spheroids; pH regulation and coupled transport. ↩
5. Lee SH et al. (2018). Carbonic anhydrase IX is a pH-stat that sets an acidic tumour extracellular pH in vivo. British Journal of Cancer 119:622–630.Engineered CA IX expression in HCT116 mouse xenografts; direct pH measurements. ↩
6. Helmlinger G et al. (1997). Interstitial pH and pO2 gradients in solid tumors in vivo: high-resolution measurements reveal a lack of correlation. Nature Medicine 3:177–182.Animal tumours; local extracellular pH and oxygen microelectrode measurements. ↩
7. Molavian HR et al. (2011). Deriving mechanisms responsible for the lack of correlation between hypoxia and acidity in solid tumors. PLOS ONE 6:e28101.Computational model of vascular delivery, buffering and metabolism. ↩
8. Mookerjee SA et al. (2015). The contributions of respiration and glycolysis to extracellular acid production. Biochimica et Biophysica Acta 1847:171–181.Cell and subcellular preparations; respiratory contributions to assay acidification. ↩
9. Hensley CT et al. (2016). Metabolic heterogeneity in human lung tumors. Cell 164:681–694.Intraoperative glucose isotope tracing in nine patients; no treatment-efficacy claim. ↩
10. Williams SP et al. (2012). Quantitation of glucose uptake in tumors by dynamic FDG-PET has less glucose bias and lower variability when adjusted for partial saturation of glucose transport. EJNMMI Research 2:6.Quantitative tracer modelling and preclinical imaging. ↩
11. Calcinotto A et al. (2012). Modulation of microenvironment acidity reverses anergy in human and murine tumor-infiltrating T lymphocytes. Cancer Research 72:2746–2756.Human and mouse T-cell experiments; mouse immunotherapy studies. ↩
12. Robey IF et al. (2009). Bicarbonate increases tumor pH and inhibits spontaneous metastases. Cancer Research 69:2260–2268.Mouse pH and metastasis experiments; model-dependent effects. ↩
13. Pilon-Thomas S et al. (2016). Neutralization of Tumor Acidity Improves Antitumor Responses to Immunotherapy. Cancer Research 76:1381–1390.Mouse pH and immunotherapy experiments; bicarbonate started before tumour inoculation. ↩
14. Gillies RJ et al. (2022). Back to basic: Trials and tribulations of alkalizing agents in cancer. Frontiers in Oncology 12:981718.Report of the authors’ small human trials; recruitment and tolerability limitations. ↩
15. Babl N et al. (2023). MCT4 blockade increases the efficacy of immune checkpoint blockade. Journal for ImmunoTherapy of Cancer 11:e007349.Mouse tumours; pH measured in dissected tissue after in-vivo treatment. ↩
16. Watson MJ et al. (2021). Metabolic support of tumour-infiltrating regulatory T cells by lactic acid. Nature 591:645–651.Treg-specific MCT1 deletion and immunotherapy in experimental models. ↩
17. Fischer K et al. (2007). Inhibitory effect of tumor cell-derived lactic acid on human T cells. Blood 109:3812–3819.Human T-cell experiments; importance of distinguishing lactate from lactic acidosis. ↩
18. Chafe SC et al. (2019). Targeting Hypoxia-Induced Carbonic Anhydrase IX Enhances Immune-Checkpoint Blockade Locally and Systemically. Cancer Immunology Research 7:1064–1078.Culture pH experiments and mouse combination therapy; human prognostic association. ↩
19. Nakaoka R et al. (2023). Electron Paramagnetic Resonance Implemented with Multiple Harmonic Detections Successfully Maps Extracellular pH In Vivo. Analytical Chemistry 95:3940–3950.EPR pH imaging in xenografts, including U-104 treatment; acidification described as a tendency. ↩
20. Bell HN et al. (2023). Microenvironmental Ammonia Enhances T Cell Exhaustion in Colorectal Cancer. Cell Metabolism 35:134–149.e6.Cell, human-tissue and mouse studies. Glycerol phenylbutyrate reduced MC38 tumour size; the checkpoint pair used ornithine, not phenylbutyrate. ↩
21. Gu J et al. (2026). Tumor-produced ammonia is metabolized by regulatory T cells to further impede anti-tumor immunity. Cell 189:418–434.e24.Human HCC spatial profiling and experimental mechanisms; ammonia-associated Treg adaptation. Abstract-verified; no phenylbutyrate combination efficacy claimed. ↩
22. DailyMed · RAVICTI prescribing information (2025). Glycerol phenylbutyrate: mechanism of action and clinical pharmacology. Official prescribing information; reviewed 11 October 2026.Nitrogen-disposal mechanism and approved urea-cycle-disorder setting; not evidence of cancer-treatment efficacy. ↩
23. Dravecka M et al. (2025). Low extracellular pH protects cancer cells from ammonia toxicity. Cell Death Discovery 11:137.Cancer-cell culture; extracellular pH altered ammonia-associated lysosomal dysfunction and growth inhibition. No buffer–scavenger combination tested. ↩
24. Tang Z et al. (2022). Icariside II enhances cisplatin-induced apoptosis by promoting endoplasmic reticulum stress signalling in non-small cell lung cancer cells. International Journal of Biological Sciences 18:2060–2074.4-Phenylbutyrate attenuated ER-stress-associated killing in cell experiments; not a clinical interaction study. ↩
25. Venkateswaran N et al. (2023). A Carbonic Anhydrase IX/SLC1A5 Axis Regulates Glutamine Metabolism Dependent Ferroptosis in Hypoxic Tumor Cells. Molecular Cancer Therapeutics; doi:10.1158/1535-7163.MCT-23-0041.Cell and spheroid drug combinations; mouse combination used SLC-0111 plus inducible GCLC knockdown, not two-drug CB-839 treatment. ↩
26. Pillai R et al. (2024). Glutamine antagonist DRP-104 suppresses tumor growth and enhances response to checkpoint blockade in KEAP1 mutant lung cancer. Science Advances 10:eadm9859.KEAP1-mutant lung-cancer mouse models; DRP-104 plus anti-PD-1. Ammonia or pH mediation was not established. ↩
27. Park S et al. (2021). CU06-1004-Induced Vascular Normalization Improves Immunotherapy by Modulating Tumor Microenvironment via Cytotoxic T Cells. Frontiers in Immunology 11:620166.Mouse vascular–immune combination; delivery, hypoxia and immune outcomes. No demonstrated pH correction or CO2 clearance effect. ↩
28. Abumanhal-Masarweh H et al. (2019). Sodium bicarbonate nanoparticles modulate the tumor pH and enhance the cellular uptake of doxorubicin. Journal of Controlled Release 296:1–13.Liposomal bicarbonate plus doxorubicin in a 4T1 mouse model, with tumour-pH measurement and drug-uptake experiments. ↩
29. Truong-Andrievici N et al. (2026). Solid tumors exploit proton-sensing GPR65 for orchestration of an immunosuppressive tumor microenvironment. Signal Transduction and Targeted Therapy 11:180.Cell and mouse experiments; context-dependent effects of GPR65 disruption and early pharmacological treatment. ↩
30. Malagola E et al. (2026). Loss of proton-sensing TDAG8 increases tumor progression in mouse models of colon cancer. Molecular Oncology; doi:10.1002/1878-0261.70283.Mouse colon-cancer models; receptor loss increased tumour burden and inflammatory-cell recruitment. ↩
31. Navarro F et al. (2022). Overcoming T cell dysfunction in acidic pH to enhance adoptive T cell transfer immunotherapy. OncoImmunology 11:2070337.HVCN1-engineered T-cell and CAR-T experiments, including mouse tumour outcomes; no combined ammonia-resistance strategy established. ↩
32. Zhang H et al. (2024). Ammonia-induced lysosomal and mitochondrial damage causes cell death of effector CD8+ T cells. Nature Cell Biology 26:1892–1902.Cell and animal work; ex-vivo DON or C381 pretreatment improved adoptive-transfer outcomes in B16-OVA mice. ↩
33. Halford S et al. (2023). A Phase I Dose-Escalation Study of AZD3965, an Oral Monocarboxylate Transporter 1 Inhibitor, in Patients with Advanced Cancer. Clinical Cancer Research 29:1429–1439.First-in-human dose escalation; 40 treated patients; safety and pharmacodynamics. ↩
34. McDonald PC et al. (2020). A Phase 1 Study of SLC-0111, a Novel Inhibitor of Carbonic Anhydrase IX, in Patients With Advanced Solid Tumors. American Journal of Clinical Oncology 43:484–490.Seventeen-patient phase I trial; safety, pharmacokinetics and preliminary activity. ↩
35. ClinicalTrials.gov · NCT03450018. SLC-0111 and gemcitabine in CA IX-positive pancreatic cancer. Trial registry; reviewed 11 October 2026.Terminated; actual enrolment six; completion 16 May 2024; no results posted. ↩
36. Gallagher FA et al. (2008). Magnetic resonance imaging of pH in vivo using hyperpolarized 13C-labelled bicarbonate. Nature 453:940–943.Method development and mouse tumour pH imaging. ↩
37. Gallagher FA et al. (2015). Carbonic Anhydrase Activity Monitored In Vivo by Hyperpolarized 13C-Magnetic Resonance Spectroscopy Demonstrates Its Importance for pH Regulation in Tumors. Cancer Research 75:4109–4118.Carbonic-anhydrase activity and exchange-kinetics validation; imaging-method limitation. ↩
38. Parks SK et al. (2017). Genetic disruption of the pHi-regulating proteins Na+/H+ exchanger 1 (SLC9A1) and carbonic anhydrase 9 severely reduces growth of colon cancer cells. Oncotarget 8:10225–10237.Colon-cancer xenografts; double-knockout tumours grew faster than either single-knockout group, with increased CA XII expression. ↩

Evidence note. Targeted primary-literature review through 11 October 2026. Cell experiments, animal studies, modelling and early human trials are distinguished in the text. Combinations assembled from separate studies are explicitly labelled as hypotheses. The staged study is an author-proposed research programme, not a clinical treatment protocol.