Scientists study complete androgen insensitivity syndrome (CAIS) through a surprisingly wide range of research models, from mice carrying a naturally occurring mutation discovered decades ago to zebrafish engineered with CRISPR and human stem cells reprogrammed in the lab. Because CAIS is rare and the androgen receptor touches nearly every organ system, no single model captures the full picture. Each one illuminates a different piece of the puzzle, and together they have reshaped our understanding of how androgens shape bodies, brains, bones, and metabolism.
Why CAIS Needs So Many Research Approaches
CAIS occurs when the androgen receptor, a protein that lets cells respond to testosterone and related hormones, is completely nonfunctional. People with CAIS have a 46,XY karyotype and internal testes that produce testosterone, but their bodies cannot use it. The result is female external anatomy, absent or sparse body hair, and a host of downstream effects on bone density, fat distribution, and reproductive development. Studying all of these consequences in living humans is limited by the condition’s rarity and by obvious ethical constraints, so researchers have built an ecosystem of laboratory models, each designed to answer different questions about what goes wrong when androgen signaling fails.
The Tfm Mouse and Its Legacy
The oldest and most influential animal model is the testicular feminized mouse, or Tfm mouse. These animals carry a naturally occurring mutation in the androgen receptor gene on the X chromosome. Like humans with CAIS, Tfm males have testes and produce testosterone but develop female-like external features because their cells cannot respond to the hormone. This model has been used since the 1970s to explore everything from genital development to brain structure.
One area where Tfm mice proved especially revealing is bone biology. When researchers removed the testes of Tfm mice and then gave them testosterone replacement, the cortical bone (the dense outer shell) did not respond, while the same treatment fully maintained bone thickness and density in normal control mice. This demonstrated that the androgen receptor itself, not just the presence of testosterone, is essential for building and maintaining the outer layer of bone.
Tfm mice have also been valuable for studying how androgen receptor loss scrambles metabolism across different tissues. One study found that Tfm mice had reduced expression of a key glucose transporter in muscle and lower activity of sugar-burning enzymes in muscle, liver, and subcutaneous fat. Meanwhile, enzymes involved in making new fat were increased in the liver and subcutaneous fat tissue. The pattern suggests that without functional androgen receptors, the body’s ability to process glucose and fatty acids in peripheral tissues breaks down, potentially shunting excess fuel toward visceral fat, the liver, and artery walls.
Engineered Knockout Mice and Tissue-Specific Models
The Tfm mouse has a global receptor defect, meaning every cell in the body is affected at once. That makes it hard to tell which tissue is actually responsible for a given symptom. If a Tfm mouse develops weak bones, is it because the bone cells themselves need androgen signaling, or because some distant organ stopped sending the right hormonal or metabolic signals? To untangle this, researchers developed conditional knockout mice using a genetic tool called the Cre-loxP system, which lets them delete the androgen receptor in one specific cell type while leaving it intact everywhere else.
This approach has generated a whole library of tissue-targeted models. Scientists have knocked out the androgen receptor selectively in the testes, ovaries, muscle, bone, fat, liver, skin, brain, and immune cells, among others. Each model isolates the receptor’s role in that one tissue. The result is a much finer-grained map of androgen action than any global model could provide.
Bone offers a clear example of why this matters. When researchers created mice lacking the androgen receptor only in bone-forming cells (osteoblasts), the animals developed high-turnover osteoporosis: their bones were being broken down faster than they were being rebuilt. Crucially, the problem was traced to the osteoblasts themselves ramping up production of a molecule called RANKL, which stimulates bone-destroying cells. Without the androgen receptor reining them in, osteoblasts were essentially calling in their own demolition crew. A separate study confirmed that deleting the receptor either before or after puberty led to reduced bone density throughout the skeleton, reinforcing the idea that androgen signaling in bone is not just needed for initial development but for ongoing maintenance in adulthood.
Fat tissue tells an equally nuanced story. Mice with the androgen receptor knocked out specifically in fat cells developed smaller fat pads on a normal diet but were hyperinsulinemic, meaning their pancreas was pumping out extra insulin to keep blood sugar under control. On a high-fat diet, these mice became more susceptible to visceral obesity and eventually showed failing insulin secretion and high blood sugar. A protein called retinol binding protein 4 (RBP4), linked to insulin resistance, was selectively elevated in their blood and fat tissue. The researchers concluded that androgen receptor signaling in fat cells independently protects against both visceral obesity and metabolic dysfunction, even apart from its effects in other tissues.
A separate global knockout study found that male mice lacking the androgen receptor entirely became obese over time despite eating no more than normal controls. They were simply less active and burned less energy. Their fat tissue secreted more of the insulin-sensitizing hormone adiponectin, which may explain why they did not immediately become diabetic despite their expanding waistlines. But the underlying metabolic trajectory was clearly unfavorable.
In Vitro Models Using Human Cells
Animal models can only approximate the human condition. To study CAIS at the molecular level in human tissue, researchers have long relied on genital skin fibroblasts, cells grown from small skin biopsies taken from the genital area of affected individuals. These cells retain the androgen receptor (or its mutant form) in culture, making them a living test tube for studying how specific mutations alter receptor behavior.
One of the earliest and most informative uses of this approach involved DNA microarrays, a technology that measures the activity of thousands of genes at once. When researchers compared fibroblasts from normal males with those from individuals with CAIS caused by known androgen receptor mutations, they found over 400 genes with significantly different activity levels. That number was striking because it suggested the androgen receptor is not just a simple on-off switch for a handful of targets but a broad regulator of gene programs involved in genital differentiation.
Fibroblast studies have also revealed that CAIS is not a single molecular story. In some patients, the receptor protein is entirely absent from the cell’s nucleus. In others, the receptor is present and even binds its hormone normally but still fails to activate genes. One study described two siblings with a classical CAIS phenotype whose fibroblasts contained androgen receptor concentrations seven standard deviations above the normal mean, with binding properties that looked completely normal in every assay the researchers could run. The receptor was there in abundance, it grabbed onto its hormone, it moved into the nucleus, but it still could not do its job. That finding pointed to a category of CAIS caused not by a broken receptor but by a failure somewhere downstream in the signaling chain, perhaps in how the receptor interacts with other proteins to switch genes on.
These fibroblast studies remain essential for confirming the functional consequences of newly discovered mutations. When a family is identified with a novel variant in the androgen receptor gene, researchers often grow fibroblasts from affected individuals to see whether the receptor protein is made, whether it binds hormone, and whether it reaches the nucleus. This step bridges the gap between finding a DNA change and proving it actually causes the disease.
Induced Pluripotent Stem Cells
Fibroblasts have one major limitation: they are skin cells. They cannot become neurons, or bone cells, or gonadal tissue. To study CAIS in cell types that are harder to biopsy, researchers have started generating induced pluripotent stem cells (iPSCs) from affected individuals. These are adult cells that have been reprogrammed back to a stem-cell-like state, capable of being coaxed into becoming virtually any cell type in the body.
The first published CAIS-derived iPSC lines came from a patient carrying a specific point mutation in the androgen receptor gene. The resulting cell lines expressed stem cell markers, could differentiate into cells from all three embryonic germ layers, maintained a normal 46,XY karyotype, and carried the expected mutation. In principle, these lines can now be directed to become neurons, bone cells, gonadal tissue, or any other androgen-responsive cell type, giving researchers a human-specific platform to study the downstream effects of complete androgen insensitivity in tissues that were previously out of reach.
iPSC technology also opens the door to gene correction experiments. Using CRISPR to repair the androgen receptor mutation in an iPSC line and then comparing the corrected cells to their uncorrected siblings would let researchers isolate the receptor’s contribution in a genetically matched background, removing the confounding differences between cells from unrelated donors.
Organoids and Bioengineered Tissue
A step beyond flat cell cultures, organoids are three-dimensional mini-organs grown from stem cells. They self-organize into structures that roughly mimic real tissue architecture, complete with multiple interacting cell types. Testicular organoids derived from human pluripotent stem cells have already been used to study hormone production and drug effects in a system that is closer to the real organ than any flat culture dish can be. While testicular organoids have not yet been widely applied to CAIS specifically, the platform is a natural fit. Engineering organoids from CAIS-patient iPSCs could allow researchers to watch how the absence of androgen receptor signaling distorts testicular development in real time, in three dimensions, and with human cells.
Zebrafish Models
Zebrafish may seem like an odd choice for studying a human condition involving sex hormones, but they offer something mammals cannot: rapid generation times, transparent embryos, and easy genetic manipulation. Using CRISPR, two independent groups knocked out the androgen receptor gene in zebrafish and found results that echoed the mammalian condition in some ways but diverged in others.
Male zebrafish without a functional androgen receptor developed female secondary sex characteristics, were infertile due to defective sperm production, and were heavier than their wild-type siblings. The proportion of female offspring in crosses also increased. Unexpectedly, female zebrafish lacking the receptor experienced premature ovarian failure during growth, a finding with no obvious parallel in mammalian CAIS but one that hints at an underappreciated role for androgen signaling in maintaining ovarian function. Because zebrafish gonads can, under some circumstances, switch between ovary-like and testis-like states, the fish model offers a unique window into how androgen receptor loss alters the tug-of-war between male and female developmental programs.
Non-Human Primate Studies
Rodents differ from humans in important ways when it comes to androgen biology. Fat distribution, metabolic set points, and brain organization are not identical, and some findings in mice have failed to translate to people. Non-human primates offer a closer approximation. Research in rhesus monkeys over the past half-century has confirmed that prenatal androgen exposure permanently shapes brain circuits governing sexually dimorphic behaviors, including play styles, mounting behavior, and vocalizations. In primates, both testosterone and its non-aromatizable cousin dihydrotestosterone are equally effective at masculinizing behavior, indicating that the androgen receptor pathway (rather than conversion to estrogen) is the key driver.
Primate studies have also addressed the metabolic effects of androgen manipulation. Researchers have noted that rodents and humans sometimes respond differently to androgen imbalance, and nonhuman primates have been used to investigate the sex-specific effects of gonadectomy and hormone replacement on fat tissue architecture and function. These studies help bridge the gap between what a mouse model predicts and what a human clinician observes.
Brain and Behavior
Some of the most provocative research model work has focused on the brain. Tfm rodent models have shown that androgen receptors are normally involved in masculinizing many brain regions and a range of behaviors, including sexual behavior, aggression, stress responses, and certain forms of learning. In one landmark study, male mice carrying a completely null androgen receptor mutation showed a near-total loss of male-typical sexual and aggressive behaviors. Treating these mice with dihydrotestosterone, a potent androgen that cannot be converted to estrogen, failed to restore sexual behavior but partially rescued aggression. This dissociation revealed that different male behaviors depend on the androgen receptor to different degrees, and some aspects of brain masculinization involve both receptor-dependent and receptor-independent androgen pathways.
These findings have direct relevance to understanding the psychological dimensions of CAIS in humans. People with CAIS generally identify as women and report female-typical gender identity, consistent with what the animal models predict: without a functioning androgen receptor, the brain does not undergo masculinization regardless of how much testosterone circulates in the blood.
Immune Function
Androgens influence the immune system in ways that are still being mapped, and research models for CAIS have contributed to this area. Tfm mice have larger thymuses and spleens than normal male controls, and their T-cell populations look different, suggesting that androgen receptor signaling normally shapes how immune cells mature. Tissue-specific knockout models have further shown that the androgen receptor in various cell types plays roles in both the production and maturation of immune cells.
However, interpreting immune data from Tfm mice requires caution. One careful study found that while some immune features of Tfm mice tracked with their androgen insensitivity, others, such as high production of the cytokine interleukin-4, appeared to be related to other genetic factors on the same mouse strain background rather than to the androgen receptor mutation itself. This is a reminder that no animal model is a clean single-variable experiment; background genetics always matter.
Cardiovascular Insights
Androgen receptor signaling also appears relevant to blood vessel function. Using mice with the androgen receptor knocked out in smooth muscle cells alone, or in both smooth muscle and endothelial (blood vessel lining) cells, researchers found impaired recovery of blood flow after an induced blockage in the hind limb. The mice with the double knockout showed the most persistent deficit, still lagging behind controls three weeks after surgery. This finding suggests the androgen receptor in vascular cells plays a direct role in the body’s ability to restore circulation after an ischemic event, with potential implications for understanding cardiovascular risk in individuals with impaired androgen signaling.
Clinical Registries as a Different Kind of Model
Not all research models live in a petri dish or a cage. International registries that track people with differences of sex development (DSD) function as observational research platforms. The I-DSD Registry, for instance, collects clinical data from centers worldwide and has been shown to have high data completeness for conditions like CAIS. Registry-based studies have documented how clinical practices such as the timing and approach to gonadectomy (surgical removal of testes) vary across regions and have shifted over time. These databases allow researchers to study the natural history of CAIS in large cohorts, something impossible in any single clinic given the condition’s rarity.
One registry-based analysis examined gonadal tumor risk in people with CAIS and found cases of germ cell neoplasia in situ, a precancerous change, occurring in a small number of individuals, with most cases arising after puberty. This kind of data directly informs clinical decisions about whether and when to remove retained testes, a question that has been debated for decades and remains without a firm consensus.
Molecular Genetics and the Growing Mutation Catalog
Underpinning all of these models is a continuously expanding catalog of androgen receptor mutations found in people with CAIS. The mutations are remarkably diverse. Some families carry whole exon deletions; others have single-letter changes in the DNA that swap one amino acid for another and cripple the receptor’s ability to bind its hormone. One particularly unusual case involved a mutation deep within an intron, a stretch of DNA between coding regions that is normally spliced out. This intronic change created a new splice site that tricked the cell’s machinery into inserting extra sequence into the messenger RNA, leading to a premature stop signal and degradation of the receptor message before it could ever be translated into protein. No receptor protein was detectable in the patient’s cells.
Newer studies have combined traditional sequencing with computational molecular dynamics simulations. In one Iranian family with CAIS, whole-exome sequencing identified a novel missense variant that had never been reported before. The researchers then modeled the mutant protein computationally and showed that the amino acid swap prevented the receptor from properly gripping its hormone ligand, explaining the complete loss of function. Each new mutation enriches the global database and helps researchers understand which parts of the receptor are most critical for its function.
Developmental Biology and Wolffian Duct Fate
One of the oldest questions in CAIS biology is why the internal male reproductive tract, derived from embryonic structures called Wolffian ducts, fails to develop. Animal models have provided a nuanced answer. In normal female rodent embryos, the Wolffian ducts regress through a well-defined sequence of events: cells stop dividing, programmed cell death ramps up, and the basement membrane that holds the duct together falls apart. When researchers genetically or chemically blocked androgen action in male embryos, the ducts regressed through an essentially identical process, just two to three days later than in females.
That delay is the interesting part. It suggests that androgens are not the only factor keeping the Wolffian duct alive in males. Some other, as yet unidentified, signals may also promote duct survival in males or actively drive duct regression in females, implying the developmental program is more complex than a simple “androgens on, duct stays; androgens off, duct goes.” This finding has pushed researchers to look beyond the androgen receptor for additional molecular players in sex-specific duct fate decisions.