Sex Matters for Mechanism

See allHide authors and affiliations

Science Translational Medicine  15 Oct 2014:
Vol. 6, Issue 258, pp. 258fs40
DOI: 10.1126/scitranslmed.3009859


Some funding agencies now require consideration of sex and gender in preclinical research, a policy that heralds opportunities and challenges for researchers.

Personalized medicine is predicated on the notion that the identification of functional variants in human genomes will improve medical care by defining specific targets for new therapies. For example, identification of polymorphisms that affect drug metabolism and clearance can define diagnostic tests and improve dosing practices, thus enhancing drug safety and efficacy. The identification of somatic mutations that drive cancer growth or drug resistance is critical for the development of targeted therapeutics based on molecular mechanisms.

In contrast to this keen focus on genetic variation among individuals, only a minority of preclinical research studies include sex as a variable. Most studies report using animals of a single sex, usually male, or do not mention the sex of their animal models. Fundamental research fuels the mechanistic discoveries that create the pipeline for new diagnostics and therapeutics; thus the paucity of sex-based analysis in preclinical studies has a direct impact on medical advances. The U.S. National Institutes of Health’s (NIH’s) Office of Women’s Health has sought input from the scientific community in a request for information on areas of research in which inclusion of sex as a biological variable may have the greatest impact as well as comments on how best to develop such policies ( The launching point for these considerations is precise definitions for the terms “sex” and “gender.”

Sex and gender exert wide-ranging effects on human health. Sex is the constellation of biological attributes of sexually reproducing organisms, including physical characteristics. Gender refers to cultural and social attitudes that influence a continuum of traits considered to be feminine or masculine, social interactions, and issues of gender identity. In cell and animal models, biological sex reflects the combined influences of sex-chromosome composition, gene expression patterns, hormones, and responses to environmental stimuli. Not all traits, genes, or biological pathways display sex differences, and in some settings, intra-sex variation may equal or exceed intersex variation. The crucial issue is to design studies that can capture variation within and between sexes. How can the research community foster the recognition that sex should be considered in preclinical study design and analysis?


Sex equity.

Men and women display differences in presentation and risk for certain human diseases, but the underlying mechanisms remain elusive.


Medical research agencies have begun to launch policies to promote considerations of sex and gender in the projects they support. Since 1993, NIH has required the inclusion of female subjects in clinical studies, which previously had been strongly male biased. Two decades later, this policy has not permeated NIH-funded preclinical biomedical research. In response to this situation, NIH recently announced a new policy to compel applicants either to consider both sexes or to provide convincing evidence that sex is not relevant in their preclinical models (1). Nine European funding agencies have partnered to increase the proportion of studies that include consideration of sex and gender. These efforts mirrored that of the Canadian Institutes of Health Research (CIHR; Canada’s largest federal medical research funder), which in 2010 implemented a sex- and gender-analysis policy that requires applicants to declare whether sex or gender will be considered or why they are not relevant to their research proposals. A recent analysis of funded CIHR research grants reviewed since enacting the policy reveals variable inclusion of sex or gender considerations across research disciplines. Clinical and population-based projects are more likely to report analysis of sex or gender in their studies than are basic biomedical proposals. In this latter group—which includes molecular biology, neuroscience, genetics, and immunology—~80% of applicants indicated that biological sex was not considered in their studies (2). These data imply that many members of the community devoted to the discovery of new biological mechanisms do not consider sex to be an important variable.

Multiple practices and beliefs underlie the dominance of this unisex view. Historically, many animal models of human syndromes are primarily studied in only one sex, and the sex of cell lines is rarely acknowledged. Perhaps more influential are prevailing beliefs on this topic: (i) Biological mechanisms unrelated to sex organs and reproduction do not differ appreciably between sexes. As highlighted below, many examples contradict this view. (ii) Female hormonal fluctuations confound biological measurements and so studies of males are favored. However, a recent meta-analysis of 293 publications reported that trait variability was no greater in female than in male mice despite not considering estrous cycle stage in the primary studies (3). (iii) Considering sex as a variable requires more animals and thus demands greater resources. This is a serious issue for most research labs in the current challenging funding environments in the United States, Canada, and other nations.

Although these are valid funding concerns, there are substantial health risks and societal costs associated with the failure to recognize sex and gender differences. For example, 8 of 10 drugs withdrawn from the U.S. market from 1997 to 2000 resulted from serious adverse reactions that posed greater risks for women than for men ( The human-health and financial costs of pulling drugs from the market are enormous; it is far better to consider sex and gender during their preclinical and clinical development.

For biomedical researchers, a central motivation is the opportunity to make impactful discoveries. Authors of the recent policies on inclusion of sex and gender in biomedical research might engage the scientific community by highlighting the fascinating, foundational, and clinically relevant discoveries made when considering these variables. Striking examples abound in genetics, neuroscience, immunology, and beyond.


Humans and model organisms display a sex-specific genetic architecture in which interactions between genotype variation and sex contribute to phenotypes, including the frequencies and severity of some diseases (4). Sex-chromosome effects on phenotype underlie X chromosome–linked recessive diseases in males (hemophilia, X-linked agammaglobulinemia, and Duchenne’s muscular dystrophy). Although males and females share the sequence of the autosomal genome, extensive sexually variable gene regulation has been identified in flies, worms, and mice (5). Hundreds of genes are differentially expressed between the ovary and testis of inbred mice. More remarkable are the extensive sex-specific differences in liver and kidney gene expression in age- and strain-matched mice and the hormone-dependent and -independent transcriptional networks that mediate these sex effects. Among these are Cyp450 family members and other enzymes that regulate the metabolism of steroid hormones and many drugs. Sex differences in human CY450 may contribute to differential clearance rates of commonly used cancer chemotherapeutics, antidepressants, and antihypertension drugs (6). Thus sex differences in the regulatory genome underlie physiological differences that must be addressed to fill knowledge gaps in clinical pharmacology and thus improve patient care.

The frequency and age at onset of some common neuropsychiatric disorders differ between females and males. Females have a higher prevalence of diseases that pre­sent from puberty into adulthood, including affective and anxiety disorders, and a higher risk of dementia compared to males. In contrast, neurodevelopmental disorders that present in childhood, such as attention-deficit hyperactivity and autism spectrum, are two- to fivefold more common in males. Exposure of female mice to moderate stress during early pregnancy results in elevated stress hormones and impaired cognition in their male but not female offspring (7). Molecular analyses of the maternal-fetal interface in these models have identified hormonal, metabolic, and proinflammatory mediators that contribute to sex-specific epigenetic programming of the offspring. These findings suggest that maternal stress exerts on their offspring sex-dependent effects underlying some neurodevelopmental disorders. This avenue of research may identify evolutionarily conserved biomarkers of risk for these neurodevelopmental disorders during pregnancy and provide a new window for early therapeutic intervention.

Abundant evidence supports sex differences in susceptibility to infections and tissue-specific and systemic autoimmune diseases. More than 80% of patients with Sjogren’s syndrome, autoimmune thyroiditis, scleroderma, and systemic lupus erythematosus (SLE) are females. Rheumatoid arthritis and multiple sclerosis (MS) are more common in women; ankylosing spondylitis is more common in men. Autoimmune disease severity also differs by sex. Male MS patients generally have a more debilitating course than do women, and compared to females, male SLE patients have greater prevalence of nephropathy, vascular thrombosis, and neuropathy (8). However, we have limited insight into the mechanistic origins of these sex differences. We need to understand the impact of genetic and environmental risk factors on these diseases in a sex-specific framework.

The experimentally induced autoimmune encephalomyelitis (EAE) mouse models of MS and the nonobese diabetic (NOD) mouse model of spontaneous type 1 diabetes (T1D) both display a higher incidence in females. These models have been instrumental in leveraging genetic manipulation and experimental interventions to define causal events in these diseases. In EAE, the nuclear receptor peroxisome proliferator–activated receptor α (PPARα) controls helper T cell proliferation and cytokine production uniquely in male mice, and administration of PPARα ligand protects males but not females from demyelinating disease. Female mice instead display a helper T cell cytokine pattern enriched in interferon-γ expression. A similar sex bias identified in human T cell responses depends on sex hormone–regulated gene expression (9). Because cytokine modulators are a major focus of clinical development for autoimmune diseases, identification of sex effects on the control of T cell cytokine expression in preclinical mouse models may portend sex-dependent differences in patient responses to these interventions.

The incidence of immune-mediated diseases has increased dramatically in the last 50 years while improvements in public health such as vaccination and antibiotic treatment have contributed to a decline in infectious diseases. These trends suggest that a modern environment has changed patterns of microbial exposure in humans and animals—changes that likely have contributed to the increased risk of autoimmune and inflammatory diseases. The composition and function of the human intestinal microbial community have been associated with inflammatory diseases, but the mechanisms that mediate these associations are poorly understood. NOD mice display a high spontaneous incidence of T1D, which shares genetic risk factors and immunopathogenic features with the human disease. In an environment free of pathogens, microbe-colonized male NOD mice display a lower frequency of T1D than do females, but the disease incidence equalizes between the sexes when NOD mice are reared in completely germ-free conditions (10). Furthermore, germ-free male NOD mice have lower testosterone levels than do their specific pathogen–free male counterparts. Oral transfer of male gut microbiota into young female NOD mice elevates their testosterone levels, alters their gut microbiome and serum metabolite profiles, and confers strong protection from T1D. Pharmacological inhibition of androgen signaling in female recipients of male microbiota restores the female-typic metabolome and high T1D incidence (10). The focus on sex as a variable in these studies revealed an unforeseen causal relationship between sex hormones, the gut microbiota, and control of autoimmunity that may have relevance for multiple human diseases.

Responses to the new funding policies on inclusion of sex and gender considerations in preclinical research will likely be mixed. As discussed at a recent workshop at Stanford University, gaining broad support for these new policies within the biomedical community will depend on several factors. First, it must be made clear that these policies are not intended to convert researchers to the study of sex effects but rather to promote the design of studies capable of discovering sex effects in their selected biological systems, if they exist. Second, NIH, CIHR, and other funders must educate grant applicants on study designs that consider the possibility of sex differences, including specifications for appropriate statistical power. Third, the requirement for including sex as a variable must be balanced against the increase in study cost by providing incremental additional funding needed to include subjects of both sexes. Fourth, editors and publishers can contribute to changing the research culture by establishing reporting standards for animal model studies that include documenting strain, sex, age, and housing conditions. Enacting these changes in biomedical publications will also be essential to monitor the impact of new policies on scientific output.

Changing the biomedical research culture so as to enshrine sex equity in preclinical research will incur new costs, but the human and economic costs of failing to implement these policies are far higher. The NIH-mandated changes in clinical trials to include equal numbers of men and women, begun in the 1990s, have enhanced our understanding of the impact of sex and gender on disease presentations and therapeutic responses. A near-term reward for expanding preclinical studies to include sex as a variable will be the discovery of new biological mechanisms that may translate to improved human health. For most researchers, there is no greater incentive.


Stay Connected to Science Translational Medicine

Navigate This Article