CETSA Overview: Protein Denaturation and Drug Target Validation

The cellular thermal shift assay, known as CETSA, is a method that detects whether a drug or small molecule physically binds to its protein target inside living cells. It works by exploiting a straightforward biophysical reality: when a small molecule binds to a protein, it makes that protein harder to unfold with heat.¹ First described in 2013, CETSA gave drug researchers something they had long lacked: a way to confirm that a compound actually reaches and grips its intended target in a cellular environment, rather than just in a purified test-tube setup.

Why Drug Researchers Needed CETSA

Before CETSA, proving that a drug candidate physically engaged its protein target inside a cell was surprisingly difficult. Most binding assays required either purified proteins in a dish or chemical labels attached to the drug molecule, both of which introduce artifacts. A compound might bind beautifully to a purified protein but never reach that protein inside a living cell because of membrane barriers, metabolism, or competition with other molecules. Conversely, attaching a fluorescent tag or radioactive label to a drug can change the way it behaves. CETSA sidesteps both problems. It works with unmodified drugs and can be run on intact cells or even tissue samples, giving a readout of what is actually happening in a biologically relevant context.²

The Core Principle

Every protein has a temperature at which it unfolds, loses its three-dimensional shape, and clumps together. Think of it like cooking an egg: the clear albumin turns white and solid because the proteins have denatured. In a cell, denatured proteins aggregate and become insoluble. The key insight behind CETSA is that when a small molecule binds to a protein, it acts like a brace, holding the protein’s structure together and raising the temperature at which it unfolds. The tighter the binding, the greater the thermal stabilization.

So if you heat two batches of cells to the same temperature and one batch was pre-treated with a drug, the target protein in the treated batch will survive the heat better. You can measure how much soluble protein remains after heating, and the difference between treated and untreated samples tells you whether the drug is engaging the target.²

How a CETSA Experiment Works in Practice

A typical CETSA experiment follows a handful of clear steps. First, living cells are treated with the compound of interest (or a vehicle control). After allowing time for the compound to enter cells and find its target, the cells are heated to a series of temperatures, usually spanning from around body temperature up to roughly 70–80 °C. At each temperature step, cells are lysed open, and the mixture is separated so that soluble (still-folded) proteins are collected while aggregated (denatured) proteins are removed.³

The amount of target protein remaining soluble at each temperature is then measured, typically by western blot for individual targets or by mass spectrometry for broader surveys. Plotting soluble protein against temperature produces a melting curve. If the drug is engaging the target, the melting curve shifts to the right, meaning the protein survives to higher temperatures compared to the untreated control. The size of that shift reflects how strongly the compound stabilizes the protein.

A complementary format called the isothermal dose-response fingerprint holds temperature constant at a single point near the protein’s melting temperature and varies the drug concentration instead. This approach generates dose-response curves and helps determine how much drug is needed for meaningful engagement in cells.⁴

Scaling Up to the Whole Proteome

The original CETSA experiment looks at one protein at a time using antibody-based detection. That is useful when you already know the intended target, but drug molecules rarely touch only one protein. Off-target binding can cause side effects, explain unexpected toxicity, or reveal new therapeutic opportunities. This is where the proteome-wide version of CETSA, often called MS-CETSA or thermal proteome profiling, becomes valuable.

By combining CETSA with mass spectrometry, researchers can simultaneously monitor the thermal stability of thousands of proteins in a single experiment. If a compound shifts the melting curve of a protein nobody expected, that protein is a candidate off-target. This approach has been used to discover previously unknown targets for existing clinical drugs, to identify why certain compounds are toxic, and to explain why some drugs affect multiple pathways at once.⁵

That same proteome-wide strategy has been used to map how natural cellular molecules like nucleotides interact with the full range of proteins in a cell. By treating cell extracts with specific nucleotides and running CETSA, researchers recaptured known interactions and uncovered novel ones across diverse protein families, from enzymes to regulatory proteins.⁶

Making CETSA Fast Enough for Drug Screening

A classic CETSA experiment using western blots is informative but slow. Screening thousands or millions of compounds to find drug candidates requires a format that works in microtiter plates with automated readouts. Several groups have developed high-throughput adaptations to meet this need.

One widely used approach fuses a small peptide tag to the target protein, then adds a complementary protein fragment after heating. Only the soluble, still-folded target protein can reassemble the tag into an active enzyme that produces a light signal, making the readout as simple as measuring luminescence in a plate reader. This split Nano Luciferase CETSA platform has been validated in 384-well and 1536-well plate formats, making it compatible with large-scale screening campaigns.⁷ Another approach uses acoustic liquid handling and reverse-phase protein arrays to run CETSA across 96- and 384-well plates from start to finish.⁸

Comparative analyses between high-throughput CETSA and conventional screening technologies show that CETSA correlates well with other methods while adding the distinct advantage of confirming target engagement in cells rather than in purified biochemical assays.⁹ High-throughput CETSA has been applied at multiple stages of a drug campaign: from early screening to hit confirmation to optimizing lead compounds through dose-response experiments.¹⁰

Pushing CETSA to Ultra-High Throughput

Even 1536-well plates were not fast enough for some campaigns. Recent work has pushed CETSA into what researchers call ultra-high throughput screening (uHTS) territory. A key innovation involves an isothermal CETSA platform designed specifically for 1536-well primary screens, paired with a gradient Peltier device for retesting hits with full melting curves. Rather than the traditional method of heating cells with a sudden thermal shock, the optimized platform uses a controlled temperature ramp, which works especially well for adherent cells that remain in their normal growth state during the experiment.¹¹

A separate line of development has explored real-time CETSA, where the signal is monitored continuously as temperature rises rather than measuring individual endpoints at discrete temperatures. All earlier CETSA methods require a separate sample for every temperature point being tested. A real-time readout would dramatically reduce material requirements and speed up data collection.¹²

Membrane Proteins and Other Difficult Targets

One early concern was that CETSA might only work for proteins floating freely in the cell’s interior. Membrane-embedded proteins, which are anchored in lipid bilayers and often have multiple stretches crossing the membrane, behave differently during heating. They do not simply fall out of solution the way soluble proteins do, because they are already tangled up in membranes.

Researchers addressed this by modifying the protocol. After heating intact cells, they used detergent extraction to pull membrane proteins into solution, then separated soluble from aggregated forms. Using this approach, three different classes of multipass transmembrane proteins were successfully tested. One case, involving the outer mitochondrial protein TSPO, showed a large and robust stabilization upon compound binding. A second case with SERCA2, a calcium pump, showed a more modest shift. A third involved the receptor PAR2 and produced a stabilization pattern different from the textbook expectation.¹³ Solute carrier proteins, another notoriously tricky class of membrane targets, have also been measured. The monocarboxylate transporter SLC16A1 showed clear ligand-dependent stabilization at temperatures between 65 and 85 °C when treated with known inhibitors, while unrelated proteins in the same samples showed no change.¹⁴

Uses Beyond Drug Binding

CETSA was invented to measure drug-target engagement, but the underlying principle is broader: anything that changes a protein’s stability can produce a thermal shift. Researchers have taken advantage of this to study phenomena that have nothing to do with drug binding.

One creative application involved tracking how a particular chemical modification on proteins, called O-GlcNAcylation, affects their stability. By increasing the level of this modification on an immune receptor protein called Nod2 and then running CETSA, researchers could detect the resulting change in thermal stability. This turned CETSA into a tool for studying post-translational modifications rather than small-molecule drugs.¹⁵

Similarly, proteins that exist in complexes with other proteins, nucleic acids, or metabolites may show different melting behavior depending on whether those complexes are intact. A sensitive detection method coupling CETSA with proximity extension assays revealed that several well-known proteins, including kinases and signaling molecules, showed unusual thermal profiles that appeared to reflect whether they were bound in complexes with other cellular components. The detection method registered different protein forms (such as splice variants or modified versions) with different sensitivities, highlighting how much context matters when interpreting CETSA data.¹⁶

From Cells to Living Animals

Confirming that a drug engages its target in cultured cells is a major step, but drug developers ultimately need to know whether the same engagement occurs in living organisms. CETSA has been extended to tissue samples and even to in vivo settings in animal models.

In one study, the target engagement of taxane chemotherapy drugs was measured using CETSA in mouse xenograft tumors. Mice bearing human-derived tumors were injected with drug or vehicle, and tumors were collected shortly afterward for CETSA analysis. At a dose equivalent to roughly 150 mg/m² for humans, a significant stabilization of β-tubulin (the known taxane target) was observed at 60 °C, demonstrating that the same principle works in actual tumor tissue from a treated animal.¹⁷

More recently, CETSA was combined with advanced mass spectrometry techniques to measure target engagement of the enzyme DPP9 directly in rat tissue. The experiment demonstrated both a temperature-dependent melting curve from ex vivo tissue incubation and a dose-dependent response after dosing live animals, effectively confirming that the compound reached its target in vivo.¹⁸ These tissue-level and in vivo applications bring CETSA closer to having clinical relevance as a biomarker for whether a drug is actually doing its job in a patient’s body.

Analyzing the Data Is Harder Than It Looks

Generating CETSA data, especially in the proteome-wide mass spectrometry format, produces enormous datasets. Deciding which proteins show a “real” shift versus random noise is not trivial. The melting curve for every protein has to be fitted, a melting temperature calculated, and the shift between treated and control conditions assessed. Each computational choice along this pipeline can dramatically change the final list of proteins flagged as hits.

One group developed a publicly available analysis tool called Inflect, which systematically evaluated each step of the melt-shift calculation and showed that seemingly minor decisions, like how curves are fitted or how outliers are handled, have outsized effects on the final results.¹⁹ Another approach called NPARC introduced a statistical framework based on comparing the shapes of entire melting curves rather than just comparing single melting-temperature values, improving the ability to detect real shifts while controlling for false positives across thousands of proteins.²⁰

For anyone using CETSA data, the practical takeaway is that the analysis pipeline matters as much as the experiment itself. Two labs running the same raw data through different software with different parameter choices can get meaningfully different hit lists. Standardization and transparency in the computational steps are active areas of development in the field.

Guiding Medicinal Chemistry

One of the most practical ways CETSA has influenced drug development is in guiding the optimization of drug candidates. Traditionally, medicinal chemists tweak a molecule’s structure and then test the result in biochemical assays with purified proteins. The problem is that improvements in test-tube potency do not always translate into better cellular activity, because factors like cell permeability, protein binding, and metabolism intervene. CETSA lets chemists check whether a structural change to a molecule actually improves target engagement inside cells, providing what amounts to a reality check at every design cycle.

This approach has been used for generating structure-activity relationships that reflect genuine cellular engagement rather than artificial test-tube conditions, for validating screening hits, and for ranking lead compounds based on their ability to reach and stabilize the target in a cellular context.²¹ When the gap between biochemical potency and cellular activity is large and mysterious, CETSA often provides the explanation: the compound might bind the target in a dish but fail to get inside the cell in sufficient quantities.

Applications in Anti-Infective and Agricultural Research

CETSA’s principle is target-agnostic, and researchers have applied it beyond human cancer and metabolic disease programs. In antibacterial research, a luminescence-based CETSA was developed to evaluate whether new antibiotic candidates engage their target enzyme inside bacterial cells. One recent study used this approach to assess a series of compounds aimed at dihydropteroate synthase, an enzyme in the folate biosynthesis pathway of gram-negative bacteria. The CETSA readout, combined with accumulation assays, revealed that modifying the compound’s chemical structure to remove a negatively charged group improved both drug accumulation inside the bacteria and whole-cell target engagement.²²

In agricultural biotechnology, CETSA coupled with mass spectrometry has been used alongside other methods to identify the protein targets of natural herbicidal compounds. One study used this combination to identify binding proteins for the plant-derived compound drupacine in a weed species, pinpointing shikimate dehydrogenase as a novel herbicidal target. The proteome-wide CETSA identified 68 candidate binding proteins, while a complementary technique captured 51, with nine proteins appearing in both datasets.²³ These cross-disciplinary applications highlight that the thermal shift principle is not limited to human therapeutics: anywhere a small molecule is supposed to bind a protein, CETSA can ask whether it actually does.

What CETSA Cannot Tell You

CETSA is powerful for confirming that binding occurs, but it has real limitations that are worth understanding. First, not every binding event produces a detectable thermal shift. If a ligand binds at a surface site without substantially rigidifying the protein’s core, the melting temperature may not change enough to measure. A negative CETSA result does not prove that no binding occurs; it may just mean the stabilization is too subtle to detect.

Second, CETSA measures thermal stability, not function. A compound might stabilize a protein without inhibiting or activating it. Confirming that engagement leads to a biological effect still requires functional assays. Third, the assay depends on having a way to detect the target protein, whether by antibody, mass spectrometry, or a reporter tag. For proteins expressed at very low levels or lacking good antibodies, the readout can be noisy or unattainable.

Fourth, interpretation gets complicated when a compound causes protein destabilization rather than stabilization. Some binding events or cellular effects actually lower a protein’s melting temperature. Distinguishing genuine destabilization from protein degradation or other indirect effects requires careful controls. Finally, in the proteome-wide format, the sheer volume of data makes false discoveries a real concern. As noted with the Inflect and NPARC tools, the statistical analysis requires as much rigor as the bench work.

Despite these caveats, CETSA fills a gap that no other single technique covers as cleanly: confirming, in cells or tissues, that a molecule physically interacts with its intended target using the native, unmodified versions of both the drug and the protein. That fundamental capability is why the method has spread from a handful of academic labs to standard practice across pharmaceutical research in just over a decade.