How the DSS Colitis Model Works in IBD Research

The DSS colitis model is the most widely used animal model for studying inflammatory bowel disease, and for good reason: it is fast, reproducible, and strikingly simple. Researchers add dextran sodium sulfate, a synthetic sulfated sugar, to the drinking water of mice, and within days the animals develop colon inflammation that shares many features with human ulcerative colitis. Since Ohkusa’s group first demonstrated the approach in hamsters in 1985 and refined it in mice by 1990, the model has become a workhorse of IBD research, enabling studies of everything from gut barrier biology to drug screening to cancer development. But the apparent simplicity hides real complexity, and getting reliable results from DSS colitis requires paying close attention to variables that many newcomers underestimate.

How DSS Damages the Gut

DSS is not absorbed into the bloodstream in meaningful amounts. Instead, it works locally, directly disrupting the epithelial barrier that lines the colon. The chemical strips away the protective mucus layer and damages the tight junctions between epithelial cells, the molecular glue that keeps the gut lining sealed. One key protein in those junctions, ZO-1, begins disappearing within a single day of DSS exposure, and by day three the colon becomes measurably more permeable to molecules that would normally be kept out.¹ By day seven of treatment, ZO-1 is essentially absent from the tissue, while other junction proteins like claudin-1 are upregulated, likely as a compensatory but insufficient response. This permeability increase is further driven by an enzyme called MMP-9, which rises in epithelial cells during DSS exposure and contributes to junction breakdown.²

Once the barrier is compromised, luminal contents, including bacteria and their products, flood into the underlying tissue. This is what triggers the inflammatory cascade. The damage is not subtle: under the microscope, DSS-treated colons show loss of crypt architecture, massive infiltration of immune cells into the mucosa and submucosa, swelling of the tissue layers, and outright ulceration.³ Goblet cells, the specialized cells responsible for producing protective mucus, are destroyed. Deeper in the colon wall, collagen fibers become thinner and disorganized, and smooth muscle cells swell with visible nuclear deformation. In short, DSS creates a cascading structural collapse from the surface inward.

The Immune Response in Acute Versus Chronic Disease

The DSS model can produce either acute or chronic colitis depending on how the chemical is administered. A single course of DSS in the drinking water for five to seven days produces acute inflammation. Repeating that cycle, with recovery periods in between, drives chronic disease. The immune profiles in these two phases look quite different, which is one of the model’s strengths for studying different aspects of IBD.

In acute DSS colitis, innate immune cells respond first. Changes in neutrophils, macrophages, and other innate populations appear as early as one day after DSS exposure, and these changes spread through both the gut and systemic immune tissues like the spleen over the first five to twelve days.⁴ The acute inflammation is driven by a mix of pro-inflammatory signals, including TNF-alpha, IL-6, and IL-17, consistent with what immunologists call a Th1/Th17 response.⁵ The NLRP3 inflammasome, a molecular alarm system inside cells, plays a critical role in amplifying this early inflammation. Mice lacking the NLRP3 gene develop less severe colitis, and pharmacologically blocking the downstream enzyme caspase-1 provides a similar degree of protection.⁶

As the disease becomes chronic, the picture shifts. T and B cells accumulate in the colon, while in the spleen, their numbers drop and neutrophils and macrophages take over.⁷ The cytokine balance flips as well: the Th1/Th17 signature of acute disease gives way to a Th2-dominated profile, with increases in IL-4 and IL-10 and declining levels of TNF-alpha, IL-6, and IL-17.⁸ This transition is relevant for researchers choosing the model, because the immune mechanisms they want to study may only be active in one phase.

Why Mouse Strain Matters More Than Most People Realize

One of the biggest pitfalls in using the DSS model is assuming that results from one mouse strain will apply to another. They often do not. A systematic comparison of inbred strains found major differences in susceptibility: C3H/HeJ mice are highly sensitive to DSS, while NON/LtJ mice are quite resistant, despite being closely related genetically to the highly susceptible NOD strain.⁹ C57BL/6 and DBA/2J mice fall in between, with their susceptibility depending on which part of the colon you examine.

Even the two most commonly used strains, C57BL/6 and BALB/c, behave quite differently under DSS. C57BL/6 mice tend to show early, severe weight loss and higher mortality at higher DSS doses, along with elevated systemic inflammatory markers, greater intestinal permeability, and even signs of brain damage at high concentrations. BALB/c mice, by contrast, lose weight more gradually but show more pronounced goblet cell loss and more significant damage to the liver and pancreas.¹⁰ These are not minor differences. A study designed around one strain’s response could reach entirely different conclusions if repeated in the other.

The molecular weight of the DSS itself adds another layer of variability. Different commercial lots come in different sizes, and this turns out to matter enormously. When researchers tested DSS at 5 kilodaltons, 40 kilodaltons, and 500 kilodaltons, the 500 kD form did not produce colitis at all. The 5 kD version caused inflammation mainly in the cecum and upper colon, while the 40 kD form produced the most severe disease and concentrated it in the lower colon.¹¹ Most modern protocols use DSS in the 36–50 kD range, but the lesson is clear: switching suppliers or lots without checking the molecular weight can sabotage an experiment.

Sex Differences in DSS Colitis

Male and female mice respond to DSS differently, and this is not a subtle effect. Male mice develop significantly worse disease activity scores than females, driven primarily by greater weight loss and worse physical deterioration, though interestingly, stool consistency and fecal blood scores are comparable between sexes.¹² The underlying immune landscape also differs by sex. At baseline, female mice carry a higher proportion of most immune cell types in the colon, while males have higher levels of Th17-associated cytokines in the middle and lower colon. During active colitis, the pattern of immune cell redistribution and cytokine shifts varies between sexes and across different segments of the colon.¹³

Estrogen appears to play a protective role. Female mice develop longer colons (a marker of less severe disease), lose less body weight, and show less inflammatory infiltration and crypt damage than males. Critically, when researchers remove the ovaries of female mice and then supplement them with estradiol, the protection returns, strongly suggesting estrogen mediates the sex difference rather than some other chromosomal factor.¹⁴ For any lab running DSS experiments, this means that mixing sexes without accounting for the difference, or using only one sex without reporting which, introduces noise that can obscure genuine treatment effects.

The Microbiome as a Hidden Variable

Perhaps the most underappreciated variable in the DSS model is the gut microbiome. Genetically identical wild-type mice that differ only in their microbiota composition can show dramatically different susceptibility to DSS-induced colitis.¹⁵ This is not a theoretical concern. In cohousing experiments, where mice share bedding and therefore share gut bacteria, microbiota from more susceptible mice can transfer their vulnerability to resistant animals. One study demonstrated that caspase-deficient mice, which are normally resistant to DSS, developed more severe colitis after cohousing with susceptible wild-type mice, and the transfer correlated with an increase in Prevotella species in their guts.¹⁶

The method of microbiota transfer matters too. In one experiment, mice that were cohoused with animals carrying a low-diversity microbiome fared dramatically worse than any other group: over ninety percent required euthanasia due to disease severity, with the most severe cases showing extensive epithelial ulceration and erosion.¹⁷ By contrast, mice receiving high-diversity microbiomes fared far better regardless of how the transfer was performed. The practical implication is stark: two labs running what they believe is the same DSS protocol can get wildly different results if their mouse colonies harbor different microbial communities. This is one reason why many papers on DSS colitis are difficult to reproduce across institutions.

Probiotic and bacterial consortium approaches have had mixed results in the model. A consortium of four rare short-chain fatty acid-producing bacteria protected germ-free mice from DSS colitis by thickening the mucus layer and boosting microbial diversity.¹⁸ A well-studied probiotic mixture called VSL#3 completely prevented the spike in intestinal permeability seen in acute colitis and maintained tight junction protein expression.¹⁹ However, simply supplementing with butyrate or mixed short-chain fatty acids in the drinking water did not alleviate gut inflammation in at least one DSS study, illustrating that the relationship between microbial metabolites and colitis is not as straightforward as “more short-chain fatty acids equals less inflammation.”²⁰

Measuring Disease Severity

Researchers track DSS colitis using a composite score called the Disease Activity Index, which combines three observations: body weight change, stool consistency, and the presence of blood in the stool. Each is scored on a numerical scale, and the total provides a daily readout of disease progression that can be followed without sacrificing the animal.²¹ Noninvasive approaches like endoscopy and fecal biomarker testing (lipocalin-2 is a common one) can also track disease over time without killing the mice.²²

At endpoint, tissue histology provides the most detailed assessment. Scoring systems typically evaluate the extent of immune cell infiltration, the degree of tissue damage, and how disrupted the normal mucosal architecture has become.²³ Importantly, the descending colon is a more informative site for histological evaluation than the ascending colon, because DSS-induced damage concentrates distally in the most commonly used molecular weight range.²⁴ Colon length at necropsy is another standard measure; shorter colons indicate more severe disease.

Drug Testing and Translational Value

The DSS model’s primary practical purpose is evaluating candidate therapies for IBD. It holds up reasonably well in this role. When researchers tested four drugs already approved for human IBD in DSS-treated mice, three of them showed clear efficacy, supporting the model’s relevance for predicting clinical utility.²⁵ A broader comparison of established IBD drugs found that the thiopurine 6-thioguanine and cyclosporine A performed well as reference compounds, and that certain drug combinations were more effective than either agent alone, mirroring what clinicians sometimes see in patients.²⁶

That said, no mouse model perfectly reproduces human IBD, and the DSS model has well-known limitations. It primarily damages the epithelial barrier from the outside in, whereas human ulcerative colitis involves a complex interplay of genetic susceptibility, immune dysregulation, and environmental triggers that often begins with more subtle immune activation. Responses to drug testing can also vary by sex, which not all studies account for.²⁷ The model’s strength is not perfect fidelity to human disease but rather the ability to study specific aspects of colitis biology, particularly barrier disruption, innate immune activation, and mucosal healing, in a controlled and reproducible way.

The AOM/DSS Model for Colitis-Associated Cancer

One of the most important extensions of the DSS model combines it with a carcinogen called azoxymethane (AOM) to study the progression from chronic inflammation to colorectal cancer. The standard protocol involves a single injection of AOM followed one week later by a course of DSS in the drinking water. In one well-characterized version, all animals developed colon adenocarcinomas by twelve weeks.²⁸ The AOM provides a genotoxic hit to initiate tumor formation, while DSS supplies the chronic inflammatory environment that promotes tumor growth, mimicking how chronic IBD raises colorectal cancer risk in humans. The model is inexpensive, highly reproducible, and widely used in chemoprevention research and studies of the inflammation-to-cancer transition.

Mucosal Healing and Stem Cell Biology

Because DSS causes such dramatic tissue destruction followed by a natural recovery phase once the chemical is removed, the model has become valuable for studying how the gut heals. This recovery depends on intestinal stem cells, and DSS experiments have revealed key molecular players in the process. One signaling molecule, prostaglandin E2, activates a protein called YAP that promotes colon regeneration after injury. Mice lacking YAP cannot regenerate their colon tissue after DSS treatment and die, while mice engineered to have higher prostaglandin E2 levels recover faster, regaining body weight, colon length, and normal tissue architecture more quickly.²⁹ The catch is that the same prostaglandin/YAP pathway also promotes carcinogenesis, highlighting the delicate balance between healing and cancer.

More recently, researchers identified a gene called CMG2 as essential for replenishing the intestinal stem cell pool after DSS injury. Mice missing this gene develop colitis of similar severity to normal mice but fail to regenerate afterward because they cannot transition from an injury-activated fetal-like stem cell state back to normal adult stem cells.³⁰ Findings like this, enabled by the DSS model’s clean injury-then-recovery design, have reshaped understanding of how the gut rebuilds itself.

Effects Beyond the Gut

Although DSS is administered orally and acts primarily in the colon, the inflammation it produces does not stay there. DSS-treated mice develop measurable changes in distant organs, a phenomenon that mirrors the extraintestinal manifestations often seen in human IBD patients. The spleen, for example, shows altered immune cell ratios during chronic DSS colitis, with increased neutrophils and macrophages and decreased T and B cells, and probiotic treatment can partially normalize splenic inflammatory signaling pathways.³¹

The lungs are also affected. DSS-treated mice develop lung inflammation characterized by immune cell infiltration, elevated IL-17 and interferon-gamma, and increased levels of bacterial DNA in lung tissue, suggesting that gut barrier breakdown allows bacterial products to reach the lungs through the bloodstream. Strikingly, this lung inflammation showed reduced responsiveness to steroid treatment.³² The strain comparison work mentioned earlier found that C57BL/6 mice developed signs of brain damage at higher DSS concentrations, while BALB/c mice showed more liver and pancreatic injury.³³ These systemic effects make the DSS model useful for studying gut-lung, gut-brain, and gut-liver interactions, research questions that are increasingly important in understanding how IBD affects the whole body.

Emerging Alternatives and Complementary Platforms

For all its strengths, the DSS model has gaps that have pushed researchers toward complementary approaches. Human-derived colon organoids, miniature 3D tissue structures grown from induced pluripotent stem cells, can now be treated with inflammatory cytokines to model ulcerative colitis in a dish. When treated with a cocktail of TNF-alpha, interferon-gamma, and IL-1-beta, these organoids upregulate inflammatory genes in a pattern that correlates reasonably well with gene expression profiles from UC patients. They allow researchers to test drugs on human tissue without animal experiments and to study patient-specific biology, though they obviously cannot capture the complexity of a whole immune system, a microbiome, or systemic organ crosstalk. They are best understood as a complement to animal models rather than a replacement, filling in the human-biology gap that no mouse model, DSS or otherwise, can fully bridge.