Cellular Metabolism and Metabolic Reprogramming: A Complete Map from Pathways to Validated Cell Models

Metabolomics measures pools. A change in pool size is not, by itself, evidence that metabolic flux has changed.

Why a Map Rather Than a Pathway List

Metabolism is taught as a set of linear pathways and practised as a set of measurements that rarely connect to them. A metabolomics experiment returns a table of hundreds of named compounds; a pathway diagram explains none of them, because the interesting information is not in which metabolites are present but in which rates changed and which node caused the change.

This article is organised around that gap. The pathways are summarised compactly, because they are well known. The substance is in three things that determine whether a metabolic claim holds: the currency being traded, the compartment in which it is traded, and the level of evidence the claim requires.

Boundary

This domain owns metabolic pathway logic and the validation of metabolic models: which substrate feeds which route, how a rate is measured, and what a cell model must satisfy for a metabolic claim to be interpretable. Redox chemistry, stress-response biology and transcriptional regulation are owned elsewhere in this series; where they intersect metabolism — as they do at every node — the intersection is named and handed off rather than duplicated.

The scope note at the end states what this map does not claim.

Part 1 — Six Interacting Metabolic Modules

Cellular metabolism can be organised into six interacting modules that exchange substrates, intermediates, energy and reducing equivalents. Most observed metabolic phenotypes involve changes across more than one module.

ModuleCore functionsRepresentative control pointsTypical readouts
Glucose uptake and glycolysisATP production and biosynthetic precursorsGlucose transporters; HK; PFK1/PFKFB3; PKGlucose uptake and lactate release; glycolytic proton efflux; ¹³C tracer analysis
Hexosamine and sugar interconversionUDP-GlcNAc production, glycosylation and mannose/fructose interconversionGFPT1/2; GNPNAT1; PGM3; MPI; PMM2UDP-GlcNAc; sugar-phosphate profiling; glycan analysis
Pentose phosphate pathwayNADPH generation and ribose-5-phosphate supplyG6PD; 6PGD; transketolase; transaldolaseNADPH/NADP⁺; ribose labelling; ¹³C positional/isotopologue analysis
Mitochondrial oxidationTCA-cycle intermediates, oxidative phosphorylation and biosynthetic precursor supplyPyruvate transport/PDH; TCA-cycle dehydrogenases; respiratory chain; CPT1-dependent fatty-acid entryOCR with inhibitor controls; substrate-dependency assays; ¹³C tracing and flux modelling
Amino acid metabolismNitrogen transfer, anaplerosis, one-carbon metabolism and nucleotide synthesisGLS; GOT1/2; PHGDH; IDO1/TDO2 (tryptophan branch)Amino-acid uptake/release; ¹⁵N/¹³C tracing; serine/glycine and kynurenine measurements
Lipid synthesis and storageMembrane lipids, signalling lipids and neutral-lipid storageACLY; ACC; FASN; SCD1; DGAT enzymes¹³C acetate/glucose tracing; lipidomics; neutral-lipid staining with orthogonal quantification

Two features of this table matter more than the entries.

Readout note: ECAR is an extracellular acidification measurement, not a glycolysis-specific measurement. Respiratory CO₂ and other acid-producing processes can contribute. Pair it with lactate release, appropriate inhibitor controls or glycolytic proton-efflux analysis when making a glycolysis claim.

The modules are coupled through shared intermediates, so a perturbation in one can propagate into others. For example, inhibiting glutaminase (GLS) can reduce glutamine-derived anaplerotic input to the TCA cycle and alter downstream citrate availability and lipid synthesis. The magnitude and direction of these effects depend on cell type, nutrient availability and alternative substrate use. A growth phenotype after GLS inhibition therefore does not, on its own, establish a specific glutamine-to-lipogenesis mechanism.

The enzyme with the largest expression change is not necessarily the controlling step. Metabolic flux depends on substrate supply, transport, allosteric regulation, enzyme activity, cofactor availability, compartmentation and post-translational modification. The control of a pathway is distributed and condition-dependent; avoid labelling a fixed enzyme as rate-limiting without experimental evidence.

Part 2 — Four Analytical Currencies

For practical analysis, a metabolic phenotype can be organised around four analytical currencies: carbon, nitrogen, reducing equivalents and energy, and signalling metabolites. This is a useful framework for choosing measurements, not a universally accepted or exhaustive classification of cellular metabolism.

CurrencyCarriers / examplesWhat a change meansRepresentative measurements
CarbonGlucose, glutamine, serine, acetate, fatty acidsWhere carbon skeletons are routed¹³C tracer analysis; isotopologue patterns; formal ¹³C-MFA when absolute fluxes are required
NitrogenGlutamine, arginine, branched-chain amino acidsHow nitrogen is transferred, incorporated or released¹⁵N tracing; amino-acid uptake/release; ammonia measurement
Reducing equivalentsNADH/NAD⁺; NADPH/NADP⁺; FAD/FADH₂Redox state and electron-transfer capacityCompartment-aware redox sensors; targeted cofactor assays; NAD(P)H lifetime imaging
EnergyATP, ADP, AMP; phosphocreatine in relevant tissuesEnergy charge and ATP supply/demandATP/ADP/AMP or energy-charge assays; ATP sensors; OCR interpreted with coupling controls
Signalling metabolitesLactate, succinate, itaconate, 2-hydroxyglutarate, acetyl-CoAMetabolites acting through receptors, enzymes or as modification substratesTarget-specific receptor/enzyme assays; site-specific PTM assays; orthogonal metabolite quantification

Signalling metabolites are particularly easy to over-interpret. A metabolite may act through a receptor, regulate an enzyme, or serve as a substrate for a protein modification. For example, acetyl-CoA supplies acetyl groups; succinate and 2-hydroxyglutarate can inhibit α-ketoglutarate-dependent dioxygenases and thereby influence DNA or histone modification; and lactate-associated protein lactylation requires direct evidence for the modified residue, metabolite dependence and functional consequence. A correlation between intracellular lactate and a histone mark is not sufficient to establish this mechanism.

Part 3 — Three Levels of Evidence

Important distinction: ¹³C labelling patterns can reveal nutrient contribution and relative pathway routing, but isotopologue data alone do not automatically provide absolute fluxes. Formal ¹³C metabolic flux analysis integrates tracer data with extracellular rates, a defined reaction network and model-fit assessment. Choose the level of analysis to match the strength of the claim.

This is the discipline the rest of the article depends on.

LevelClaimMinimum evidenceWhat does not count
1. PoolThe amount of a metabolite changedTargeted, calibrated quantification with suitable internal standards and normalisation to cell number, protein or DNAAn uncalibrated relative fold-change presented as an absolute concentration
2. FluxThe rate through a step changedTracer design plus isotopologue data interpreted with an appropriate network model; for quantitative flux, include uptake/secretion rates, mass balance, fit diagnostics and flux confidence intervalsPool size; enzyme expression; a single labelling timepoint interpreted as an absolute flux
3. DependenceThe change is required for the phenotypeNutrient restriction, pathway inhibition, or genetic perturbation, each with rescueCorrelation between flux and growth

A larger metabolite pool can arise from increased production, decreased consumption, reduced export, increased import or changes in compartment-specific exchange. These possibilities have different mechanistic implications and may require different perturbations.

A dependency claim needs more than an inhibitor-associated growth effect. Remove or perturb the proposed substrate, transporter or enzyme; measure the relevant metabolic change; demonstrate a phenotype; and test specificity with an appropriate rescue. Genetic and pharmacological approaches are complementary, and off-target effects must be considered.

Part 4 — Compartmentation

The same metabolite in two compartments is two different measurements, and most bulk methods cannot distinguish them.

MetaboliteCompartmentsWhy the distinction matters
Acetyl-CoACytosolic, mitochondrial, nuclearCytosolic pool feeds lipogenesis and acetylation; mitochondrial pool feeds the TCA cycle
LactateCytosol and extracellular space; mitochondrial transport/oxidation is context-dependent and should not be assumedProduction, export, uptake and oxidation jointly determine intracellular and extracellular lactate; measure both sides of the exchange when relevant
NAD⁺/NADHCytosol, mitochondria and nucleusCompartment-specific redox ratios differ; bulk extraction averages pools and may obscure local redox state
ATPCytosol and mitochondria; extracellular glutathione is a distinct extracellular poolCompartment-specific redox buffering and the GSH–GPX4 system influence ferroptosis susceptibility; no single pool alone defines sensitivity
GlutathioneCytosol and mitochondria; extracellular ATP is a separate signalling poolBulk ATP can mask local supply-demand differences and does not report extracellular purinergic signalling

Three practical consequences follow. First, bulk measurements of compartmentalised metabolites have limited mechanistic resolution, even when quantitatively precise. Second, transporters are part of the pathway: for example, MCT-mediated lactate transport and the mitochondrial pyruvate carrier can change the interpretation of intracellular pools. Third, isotope labelling can reveal nutrient contribution and pathway routing, but it does not automatically identify the compartment in which a reaction occurred. Compartment-specific sensors, validated fractionation or compartment-aware modelling may be needed.

Part 5 — Three Claim Patterns and Their Evidence Requirements

The three published studies used here illustrate distinct claim types. They are examples for experimental reasoning, not universal templates; each conclusion must be interpreted within the model, design and evidence reported by the original investigators.

Pattern A — Substrate rerouting

A 2026 study, subsequently published as “Renal glycosuria triggers a coordinated mannose and glutamine metabolic program to maintain glucose homeostasis,” examined substrate adaptation during glycosuria. It is a substrate-rerouting claim: when glucose is lost in the urine, mannose and glutamine metabolism contribute to compensatory metabolic responses. The study used metabolite profiling, isotope tracing and perturbations in mouse models. The example is useful because it illustrates why substrate contribution and physiological dependence should be tested separately, and why results from an in vivo systemic model should not be assumed to transfer directly to an isolated cell line.

What this claim requires:

  • • A measured loss of the primary substrate, not an inferred one.
  • • Evidence that the alternative substrate is used, ideally by tracing its carbon into the pathway it is claimed to feed.
  • • A setpoint measurement that is independent of the pathway being studied — otherwise the claim is circular.
  • • A perturbation that removes the compensation, showing that homeostasis fails without it.

The key design question is whether the measured homeostatic output is independent of the pathway used to define compensation. Use orthogonal physiological endpoints and perturbations that distinguish substrate contribution from the final maintained state.

Pattern B — Metabolite as a post-translational signal

A 2026 study in hepatocellular carcinoma reported that AARS1 promotes tumour progression and immune evasion through ATF6 K424 lactylation and downstream tryptophan–kynurenine metabolism. This is a metabolite-linked post-translational modification claim: the proposed modification stabilises ATF6 and changes transcriptional output, with downstream effects on immune cells. The paper provides a concrete example of the evidence chain required to connect a metabolic signal to a phenotype; it should not be generalised to all lactylation systems.

What this claim requires:

  • • Demonstrate the modified residue directly, for example by site-resolved mass spectrometry and/or a rigorously validated site-specific antibody. Include appropriate negative controls.
  • • Test dependence on the proposed lactylation machinery and lactate-related metabolic state, while controlling for changes in cell viability, pH and global metabolism.
  • • Test functional consequence with a non-modifiable residue mutant and a suitable rescue. Confirm that the mutation does not simply disrupt protein abundance, folding, localisation or basal function.
  • • Separation of the two arms: does the immune phenotype require the tryptophan route, the modification, or both?

A residue mutant strengthens causal inference but is not sufficient by itself: expression, localisation, protein stability and relevant basal activity should also be checked. The goal is to show that the specific modification, rather than loss of the whole protein, explains the phenotype.

Pattern C — Metabolic memory

A 2026 study reported that exercise-associated metabolic memory protects against pathological cardiac hypertrophy through sustained regulation of PDK4, involving METTL3-dependent m6A RNA methylation. This is a persistence claim: a transient physiological stimulus is associated with a metabolic state that remains after the initial adaptation period and influences a later challenge. The example also shows that metabolic memory can involve RNA modification, not only DNA methylation or chromatin marks.

What this claim requires:

  • • A defined stimulus and a defined washout, with the metabolic state measured after washout rather than during the stimulus.
  • • A persistence timescale, which is the actual finding.
  • • A mechanism for persistence — for example, chromatin or DNA methylation, RNA modification, stable protein or organelle turnover, or altered cell composition — supported by direct measurements.
  • • Reversal, showing that removing the persistent factor abolishes the protection.

A persistent state must be distinguished from a slowly resolving response. Define the washout, sampling window and duration of persistence in advance; the appropriate timescale depends on the biological system and endpoint.

Part 6 — What a Metabolic Model Must Satisfy

RequirementWhyCommon violation
Defined medium compositionGlucose, glutamine, serine and pyruvate set every flux"Standard medium" reported without formulation
Medium ageGlutamine degrades in culture, ammonia accumulatesMedium prepared weekly and used throughout
Serum documentationLipids, hormones and metabolites vary by lotLot not recorded
Documented confluency and densityProliferation rate sets biosynthetic demandHarvest at unspecified density
Oxygen tensionOxidation and glycolysis shift with O₂Incubator setpoint assumed
A flux measurement, not only poolsPools are steady states of ratesMetabolomics only
A dependency testFlux without dependence is a descriptionGrowth assay only at one condition

Glutamine can be depleted by cellular uptake and can also undergo spontaneous degradation in culture medium at 37 °C, generating ammonia. The rate depends on medium composition, storage and handling. Medium age is therefore an experimental variable: record preparation and storage conditions, and measure glutamine and ammonia when they are likely to affect interpretation.

Part 7 — Failures That Recur

  • • Pool reported as flux. The single most common error in the field.
  • • Seahorse OCR and ECAR should be normalised to an appropriate measure such as viable cell number, DNA or protein, with plate layout and cell distribution controlled. OCR is not automatically equivalent to mitochondrial ATP production, and ECAR is not specific to glycolysis because respiratory CO₂ can contribute to extracellular acidification. Use assay-specific controls and, where needed, glycolytic proton-efflux calculations or orthogonal lactate measurements.
  • • Enzyme expression used as a proxy for flux. Abundance and rate diverge whenever substrate or allostery changes.
  • • Glutamine dependence concluded from one glutamine-free condition without matched nutrient rescue and appropriate pH, ammonia and osmolality controls where relevant.
  • • Compartment-blind conclusions from bulk metabolite measurements.
  • • A modification site reported without a site-specific functional test and controls for mutant protein integrity, leaving the causal role uncertain.
  • • “Memory” claimed without a defined washout, a sufficiently long follow-up or evidence that the persistent state is distinct from a slow response.
  • • Proliferation unmeasured, although biosynthetic demand changes everything.

Part 8 — Where This Map Connects

Metabolism is the currency layer beneath the rest of this series. Stress responses consume ATP and reducing equivalents; the redox pillar measures them directly. Cell fate decisions are gated by metabolic state, and the fate pillar owns the programmes that follow. Signalling hubs such as AMPK and mTORC1 read the currencies described here, and the signalling pillar owns their regulation.

What this domain owns is the accounting: which substrate, at which rate, in which compartment, and whether the phenotype depends on it. The three claim patterns above are the templates for the domain's support articles.

If you need metabolic models with traced flux, defined medium and dependency testing built in, we can scope the build with you:

  • • Custom ¹³C tracing design for a specific node
  • • Isogenic panels for nutrient-dependency testing
  • • Validated medium and sampling protocols for metabolomics

Contact EDITGENE to request a project consultation or quote.

Scientific scope and validation note

Pathway use, flux distribution and nutrient dependence vary with cell type, medium composition, oxygen tension, growth rate and experimental timing. The maps here describe common principles and representative readouts, not the behaviour of every model.

A change in metabolite abundance does not establish a change in flux, and a change in flux does not establish that the pathway is required for a phenotype.

The three studies are discussed as examples of claim types based on their published reports. Their findings are model- and context-specific; this article does not make therapeutic recommendations.

Scientific interpretation and experimental design

  • • Metabolite pools are concentrations at a sampling time, not rates. Flux inference requires appropriate tracer design and/or uptake–secretion measurements with mass balance and a suitable metabolic model.
  • • Name the currency. Carbon routing, nitrogen transfer, redox state, energy charge and signalling-metabolite activity require different measurements.
  • • Compartment first. Bulk acetyl-CoA, NAD(H), ATP or lactate measurements can average pools with different functions.
  • • Medium is part of the experiment. Composition, age, serum lot, pH, oxygen and cell density can alter metabolic measurements.
  • • Dependency is the third level. Combine perturbation with a relevant metabolic readout, functional phenotype and specificity/rescue controls.
  • • Modification needs site-specific evidence and functional controls. A metabolite–mark correlation is not a mechanism.
  • • Memory needs a defined washout and persistence window. Without these, a slow response can be mistaken for memory.

Selected references for further reading

  • • Vander Heiden MG, DeBerardinis RJ. Understanding the intersections between metabolism and cancer biology. Cell. 2017;168:657–669.
  • • DeBerardinis RJ, Chandel NS. Fundamentals of cancer metabolism. Science Advances. 2016;2:e1600200.
  • • Jang C, et al. Metabolomics and isotope tracing. Cell. 2018;173:822–837.
  • • Metallo CM, et al. Evaluation of ¹³C isotopic tracers for metabolic flux analysis in mammalian cells. Molecular BioSystems. 2009;5:1667–1677.

Antoniewicz MR. A guide to 13C metabolic flux analysis for the cancer biologist. Experimental & Molecular Medicine. 2018;50:1–13. doi:10.1038/s12276-018-0060-y.

Publishing 13C metabolic flux analysis studies: A review and future perspectives. Metabolic Engineering. 2018. PMCID: PMC5866723.

Rashid N, et al. Renal glycosuria triggers a coordinated mannose and glutamine metabolic program to maintain glucose homeostasis. Cell Reports. 2026;45(8):117804. doi:10.1016/j.celrep.2026.117804.

Wang Y, et al. AARS1 promotes tumor progression and immune evasion via ATF6 lactylation-mediated tryptophan metabolism in hepatocellular carcinoma. Cell Metabolism. 2026;38(8):1702–1721.e11. doi:10.1016/j.cmet.2026.06.003.

Zheng C, et al. Exercise Metabolic Memory Halts Pathological Cardiac Hypertrophy via PDK4. Circulation Research. 2026;138(11):e326889. doi:10.1161/CIRCRESAHA.125.326889.

Tritsch GL, Moore GE. Spontaneous decomposition of glutamine in cell culture media. Experimental Cell Research. 1962;28(2):360–364. doi:10.1016/0014-4827(62)90290-2.

Related Products

Product name Cat.No. Species Gene ID
PBRM1 Knockout A-549 Cell Line EDJ0001-K01 Human 55193 Details Get a Quote
YTHDC1 Knockout A-549 Cell Line EDC07652 Human 91746 Details Get a Quote
TP53 Knockout HCT 116 Cell Line EDC07854 Human 7157 Details Get a Quote
Sirt2 Knockout MH-S Cell Line EDJ-KQ78173 Mouse 64383 Details Get a Quote
TFE3 Knockout KGN Cell Line EDJ-KQ48 Human 7030 Details Get a Quote
Stub1 Knockout MB49 Cell Line EDJ-KQ53 Mouse 56424 Details Get a Quote
ARID1A Knockout SNK-6 Cell Line EDJ-KQ64 Human 8289 Details Get a Quote
Tet1 Knockout Kupffer Cell Line EDJ-KQ88 Mouse 52463 Details Get a Quote
NR1H3 Knockout HEK293T Cell Line EDJ-KQ109 Human 10062 Details Get a Quote
NR1H2 Knockout HEK293T Cell Line EDJ-KQ110 Human 7376 Details Get a Quote
RPS6KB2 Knockout HEK293 Cell Line EDJ-KQ125 Human 6199 Details Get a Quote
SMAD6 Knockout HEK293 Cell Line EDJ-KQ126 Human 4091 Details Get a Quote
JUN Knockout HEK293 Cell Line EDJ-KQ176 Human 3725 Details Get a Quote
JUN Knockout HEK293T Cell Line EDJ-KQ184 Human 3725 Details Get a Quote
ALKBH5 Knockout HEK293 Cell Line EDJ-KQ186 Human 54890 Details Get a Quote
Contact Us
*
*
*
How did you hear about us: