Sex in cardiovascular disease: Why this biological variable should be considered in in vitro models

Abstract

Cardiovascular disease (CVD), the world’s leading cause of death, exhibits notable epidemiological, clinical, and pathophysiological differences between sexes. Many such differences can be linked back to cardiovascular sexual dimorphism, yet sex-specific in vitro models are still not the norm. A lack of sex reporting and apparent male bias raises the question of whether in vitro CVD models faithfully recapitulate the biology of intended treatment recipients. To ensure equitable treatment for the overlooked female patient population, sex as a biological variable (SABV) inclusion must become commonplace in CVD preclinical research. Here, we discuss the role of sex in CVD and underlying cardiovascular (patho)physiology. We review shortcomings in current SABV practices, describe the relevance of sex, and highlight emerging strategies for SABV inclusion in three major in vitro model types: primary cell, stem cell, and three-dimensional models. Last, we identify key barriers to inclusive design and suggest techniques for overcoming them.

Sex-specific preclinical CVD models should be the norm to ensure safe and effective treatments for marginalized patient groups.

INTRODUCTION

In the past few decades, the role of biological sex in understanding human health has gained increasing recognition. The notable 2001 publication by the National Institutes of Health (NIH) titled Exploring the Biological Contributions to Human Health: Does Sex Matter? emphasized the importance of evaluating sex in biological research as it affects the multimodal functionality of living beings on cellular, tissue, and organismal levels. Despite this, it took nearly 15 years for the NIH, one of the largest funding bodies for biomedical research in the United States, to implement a policy (NOT-OD-15-102) requiring the inclusion of sex as a biological variable (SABV) in the design, analysis, and reporting of all human and vertebrate animal studies (1). A shift in science culture to pursue sex- and gender-specific research in clinical and in vivo studies has progressed over the past decade. In 2019, the NIH Office of Research on Women’s Health (ORWH) announced their 2019–2023 Trans-NIH Strategic Plan for Women’s Health Research “Advancing Science for the Health of Women,” meant to drive sex and gender consideration in biomedical research for the improvement of women’s health (2). In the same year, the NIH ORWH also announced a RO1 funding opportunity (RFA-OD-19-029) for research on the “influence and intersection of sex and gender in health and disease” (3). These moves by the NIH have had sweeping implications for the general standard of scientific rigor in the field. However, the demand for SABV inclusion still does not extend to much of preclinical biomedical research, namely, research using primary and stem cell lines, despite the vulnerability of these models to sex-based variability.

The impact of sex and gender on human health is best observed in the heavily sexually dimorphic cardiovascular disease (CVD). This disease, or rather a group of diseases, has long been viewed as predominantly affecting men (4). This misconception has led to a widely overlooked disease demographic: women. Following decades of public health initiatives, women are now more aware of cardiovascular risks and educated on treatment and prevention. However, the tailoring of CVD prevention, diagnosis, and treatment strategies toward male patients still leaves women vulnerable to these diseases.

A more robust understanding of the driving factors of CVD sex differences may help improve health equity. A recent push to explore the male/female divide in CVD has uncovered divergence in everything from cardiovascular biology and pathophysiology to CVD clinical manifestation and epidemiology (5). Sex-biased CVD risk factors can generally be categorized as either inherent/biological, as in the case of hormonal and chromosomal imbalances, or environmental/situational, such as differences in lifestyle, treatment disparity, and clinical underrepresentation. While movements to improve the latter have been underway, there remains a need for a better understanding of the underlying biological differences causing sex and gender disparity in CVD. Furthermore, this new body of evidence demands more targeted treatments and therapeutics, particularly for women, who have not historically comprised a substantial percentage of the testing population (6).

In this review, we outline what is known about CVD sexual dimorphism and the resulting health inequity. Inclusive design, particularly the increased representation of biologically female samples, is highlighted as a potential solution to this ongoing problem. We highlight underlying biological and environmental factors that drive CVD sex differences and discuss their relevance to in vitro CVD models involving primary cells, stem cells, and three-dimensional (3D) cell cultures. We explore how researchers are implementing SABV into preclinical CVD models to improve understanding of CVD and provide more targeted diagnostics and therapeutics. Last, we identify current bottlenecks to SABV inclusion and argue that the clinical translatability of in vitro CVD work can be greatly improved by overhauling current research practices. By identifying these areas of improvement, we can shrink the existing CVD knowledge gap and improve health outcomes across the board.

In this review, the term “sex” is used to refer to the classification of “male” or “female” based on chromosomes, gonadal hormone levels, and reproductive organs. This binary classification is simplified from the accurate understanding of sex, which presents more like a spectrum with various intersex identities (7). In addition, the terms “men” and “women” refer to genders, which are personally and socially constructed and may or may not relate to biological sex. Research articles do not always clearly state whether sex or gender is being reported on. The terminology may at times be interchanged, despite carrying distinct definitions, to reflect this. While intersex and transgender/gender fluid individuals must be considered in all areas of human health research, this review is focused primarily on filling a gap in women’s health through SABV inclusion and thus does not explicitly mention these populations. However, this work stands to pave the way for inclusivity of other marginalized groups as well. The authors acknowledge the inability of this review to encompass all possible perspectives and identities relevant to CVD and encourage readers to further explore the relationship between sex, gender, and health through these suggested readings (8–10).

SEX AND GENDER IN CVD

Once assumed to be a mostly male-afflicting disease, CVD is garnering attention as the leading cause of mortality in women both in the United States (11) and globally (12). Research into this previously neglected cohort has implicated sex and gender in the clinical and epidemiological presence of CVD (5, 13). Summarizing the role of sex/gender in CVD is challenging because CVD encompasses several distinct diseases each with their own associated risk factors, pathophysiology, and clinical manifestation, among other metrics. Sex and gender can be an important factor in any one of these categories, and the trend of its effect is not necessarily consistent across diseases. Nevertheless, numerous studies have proven that CVD pathophysiology, clinical manifestation, management, treatment response, disease outcome, and epidemiology all exhibit sex/gender-based differences (4, 5, 14–16).

A long history of underrepresentation of women in clinical trials, gender bias in medicine, imbalanced treatment offering and diagnostic techniques, and lack of sex-specific therapeutics have caused a gap in CVD care standards between men and women (Fig. 1). Historically, the presence of female patients in clinical trials was limited because of misguided concerns over hormone fluctuation-induced variability (17). Trial results, collected primarily on male patients, were then extrapolated to female patients, leading to undesirable health outcomes. Female representation in cardiac clinical trials has notably improved over the past decade, with women comprising 38.2% of relevant clinical trials between 2010 and 2017 (18). Yet for stroke, arrhythmia, coronary heart disease, acute coronary syndrome, and heart failure trials, the rate of women’s participation is still much lower than the rate of prevalence. Participation to prevalence ratio (PPR) measures a group’s representation in trials with respect to their proportion in the disease population. Good representation for a group is considered PPR 0.8 to 1.2. After adjustment for prevalence, trials studying heart failure, acute coronary syndromes, and coronary artery disease maintain persistent underrepresentation of women with PPR <0.8 (18). A potential reason for women’s underrepresentation in modern clinical trials is the male-tailored recruitment and enrollment process (6, 18). Jin et al. (18) highlight multiple points in the clinical trial recruitment process that are subject to female patient fallout. Key takeaways include a need for better communication of trial opportunities, improved access to trial centers including childcare offerings and logistical support, and more informative and comforting leadership for women throughout the trial process. Gender-based psychological and social differences (e.g., decision-making processes and risk perception) drive disparity in recruitment success. Consideration of women’s perspectives may boost enrollment and improve therapeutic efficacy for this group.

Fig. 1. Major drivers of CVD sex and gender disparity.

Structural, societal, and/or behavioral changes are needed to lessen the impact of each contributing factor on CVD health disparity. This review focuses on the underrepresentation of women in clinical and preclinical studies, specifically preclinical and basic research, and strategies for remedying this. Created with BioRender.com.

Sex and gender bias is known to permeate society; the health care field is no exception. Cardiology and relevant subspecialties including CVD medicine suffer from a severe lack of women representation (19). The male dominance in this field may result in conscious or unconscious bias in disease recognition and treatment between sexes/genders, likely in favor of men. In a 2018 study conducted by Lichtman et al. (20), 53% of women interviewed who were hospitalized for acute myocardial infarction (AMI) reported that their provider did not recognize their symptoms as cardiac-related, compared to 37% of men, which likely delayed treatment. Underdiagnosis of CVD in women can be traced to sex/gender differences in CVD manifestation. Symptoms of CVD in men typically include heavy chest pain with radiated pain in the arm and upper abdomen and shortness of breath. While these are widely viewed as tell-tale signs of ischemia, for women typical symptoms include a stinging sensation in the chest, broader abdominal pain, and weakness (4). There are numerous other examples of sex-skewed CVD recognition and treatment by providers. Kim et al. (21) provide a comprehensive overview of common sex and gender biases observed in the treatment of several CVDs and draw direct links between the lack of women in cardiology leadership, provider bias, and undesirable health outcomes for women. Although 50% of medical school graduates in the United States are women, cardiology fellowship training includes only 21% of women. Only 13% of practicing cardiologists are women and among interventional cardiologists who perform coronary interventional procedures, only 4.5% are women (22, 23). It stands to reason that improving gender and sex diversity in the cardiac health care workforce would reduce bias and improve women’s health. To achieve this, Schnabel and Benjamin (24) describe an evidence-based plan (“Diversity 4.0”) for individuals and institutions to improve diversity, equity, and inclusion in the cardiovascular health care workforce. Their goal is to improve health equity across historically underrepresented patient populations.

CVD risk factors and comorbidities are skewed by patient sex. For men, age, hypertension, total and low-density lipoprotein (LDL) cholesterol levels greatly affect their risk of heart disease (25). Whereas for women, menopause, systolic arterial hypertension, smoking, diabetes, triglyceride, and high-density lipoprotein (HDL) cholesterol levels are more relevant (4). Providing sex-specific strategies for monitoring and addressing risk factors may greatly improve CVD prevention. While female-specific risk factors have been identified, there is still a need for better CVD prediction using this risk factor data. Cardiovascular risk prediction models, especially those trained on non-diverse cohorts, lack predictive accuracy for female patients. In their 2011 study, Cook et al. (26) found that the Framingham risk score, trained primarily using white patient data, greatly overestimated the 10-year risk of major CVD when tested on an independent, diverse cohort (C-statistic = 0.750). The newer Reynolds score, which factors in additional female-relevant risk factors such as the family history of MI and inflammatory markers, outperformed the Framingham score with a C-statistic of 0.765 (P < 0.0001). Given that predictive models such as these can often be used to inform primary prevention strategies, higher accuracy can promote earlier identification and treatment of individuals at high risk.

Age is strongly correlated with CVD risk and thus often obscures sex effects (27). A large analysis of 2264 patients aged 18 to 55 with AMI from the VIRGO (Variation in Recovery: Role of Gender on Outcomes of Young AMI Patients) study and 2264 population-based controls from the National Health and Nutrition Examination Survey matched by age, sex, race, and ethnicity demonstrated profound variation in risk factor profiles. Significant differences in multiple (often modifiable) risk factors were identified. For AMI in young women, diabetes, depression, hypertension, current smoking, and family history of diabetes had stronger associations, whereas for young men, hypercholesterolemia was more heavily associated (28). Further analysis of young women with MI shows this demographic fails to accurately assess their personal risk of heart disease, despite having a family history of CVD, and misattributes symptoms as non-cardiac. Expectedly then, women (aged 30 to 55) reported seeking minimal preventive cardiac care before onset of AMI (29). Though age-specific data typically show higher CVD risk in men than women, this obscures the steep increase in CVD risk and mortality observed in aging female populations. The average age of a woman presenting her first AMI is 72.0 years compared with 65.6 years for men (11). This older age at onset has traditionally been attributed to the protective role of circulating estrogen on the vascular endothelium, though it remains difficult to unravel the impact of age from that of menopause on cardiovascular risk (30). Identifying barriers to high-quality CVD prevention, identification, and management for both old and young populations is critical for improving health outcomes.

General awareness around CVD and its risk factors for women needs improvement. CVD was only recognized as a public health concern for women in the early 1990s. Around the turn of the century, major public health initiatives began spreading awareness of women’s CVD risks (31). The American Heart Association (AHA) launched the “Go Red for Women” campaign in 2004 to educate and empower women to take control of their cardiovascular health. Despite initial improvements (32), a recent downturn in women’s awareness of CVD risk and prevention is raising concerns (33). In 2022, the AHA released a “Call to Action” advocating for increased awareness of CVD in women and improved sex-specific cardiovascular health care (34). This is crucial, as the decrease in mortality from CVD in both men and women over the past few decades can be largely attributed to increased awareness of risk factors and improved preventative programs.

Therapeutic solutions still lack a female perspective, leading to reduced efficacy and even toxicity for this patient population. Statins, widely used for managing cholesterol levels in patients with CVD and at-risk populations, are prescribed differently between men and women, with men receiving more aggressive treatment on average (35). Women’s LDL-cholesterol levels are therefore less controlled as compared to men. Peters et al. confirmed this with their 2019 study showing less adequate control of dyslipidemia in women relative to men (51% versus 63%, respectively). Conversely, women had better control of hypertension (30% versus 22%) and diabetes mellitus (30% versus 20%) than men, showing the complicated landscape of gender in CVD risk factor treatment (25). For patients with type 1 diabetes, a major risk factor for CVD, common interventions including angiotensin-converting enzyme (ACE) inhibitors, angiotensin II receptor blockers, and statins are used much less frequently among women than men (36). While undergoing drug-based cardiac arrhythmia treatment, women have a higher chance of developing serious adverse effects yet are under-referred for alternative treatment such as ablation (37). Digoxin, a medication used to treat heart failure and arrhythmia, has been associated with a higher risk of death in female patients, but not males. This discrepancy in digoxin’s effect between sexes was missed during clinical trials, likely due to the heavily male-biased (four male:one female) patient population. Consequently, clinical guidelines for digoxin administration were never designed with women in mind, leading to adverse effects post-approval (38). The quality of care for patients hospitalized with ST-segment elevation myocardial infarction (STEMI) differs widely based on gender. A 2021 meta-analysis showed that women experience delayed medical treatment, including longer time to first medical contact (mean difference = 42.5 min) and door-to-balloon time (mean difference = 4.9 min). While in hospital, women are less likely than men to receive a range of STEMI treatments such as aspirin (89.5% versus 92.1%), coronary angioplasty (59.5% versus 68.2%), P2Y12 inhibitors (67.6% versus 75.4%), glycoprotein IIb/IIIa inhibitors (22.7% versus 29.3%), beta-blockers (75.1% versus 76.1%) and ACE inhibitors (55.6% versus 59.4%). Suboptimal and delayed care contributes, at least in part, to worsened health outcomes observed in women relative to men, such as higher in-hospital mortality rates (P < 0.00001), increased risk of repeat MI (P = 0.05), stroke (P < 0.001), and major bleeding (MB; P < 0.00001). Note that in this study pool, women were more likely to have several comorbidities to begin with (higher age, diabetes mellitus, hypertension, etc.), which increased mortality and repeat cardiac event risk (39). Long-term clinical outcomes, as well as in-hospital, are skewed by patient sex, with females having a higher likelihood of MB and major adverse cardiac events (MACE) including death, reinfarction, ischemia-induced target vessel revascularization, and stroke at 30 days (MB: P < 0.0001; MACE: P = 0.01) and 3 years (MB: P < 0.0001; MACE: P = 0.046). Participants in this study show significant baseline differences between males and females such as risk factor status and family medical history, which could confound trends in overall clinical outcomes. Still, the female sex remains an independent predictor of long-term MB after adjusting for patient baseline characteristics and treatment (40). These studies highlight the need for sex-specific treatments to combat the ongoing disparity in CVD management and outcomes.

Considering intersectional identities reveals additional concerning statistics. For example, Black women have a 2.5- to 3-fold higher pregnancy-related CVD mortality rate than non-Hispanic (NH) white and Hispanic females (25). Cardiovascular risk factors are also biased by racial and ethnic groups, with hypertension highly prevalent in NH Black women, LDL-cholesterol in NH white women, diabetes in Hispanic women, and overweight and obesity in NH Black and Hispanic women (34). For young women, low annual household income (<$10,000) is a much greater AMI risk factor than for young men (P < 0.001), indicating the pronounced role of socioeconomic status in women’s cardiovascular health (28). Further collection of risk factor data on women subgroups could be used to develop targeted risk prediction algorithms. Nakai et al. (41) developed a CVD risk prediction model using Japanese cohort data with significantly higher (P < 0.001) predictive accuracy over the Framingham risk score (C-statistic = 0.782 and 0.768, respectively). The net reclassification index between the models was 0.06, highlighting the importance of population-specific inputs on risk prediction model accuracy.

UNDERLYING DRIVERS OF SEX-BASED DIFFERENCES IN CVD

While technically distinct, sex and gender are inarguably linked, forming a complex relationship between biological and sociocultural factors that affect patient health (42). The extent of this impact is still under investigation, although it is widely acknowledged to be substantial. In this review, we will focus primarily on sex as a biological determinant of cardiovascular health. Biological differences between the sexes at the molecular, cellular, tissue, and organismal levels may drive the observed sexual dissimilarity in a clinical setting (Fig. 2). Hormonal and chromosomal imbalance between males and females, both major contributors to sex-based health differences, have been investigated for their role in CVD differences (43).

Fig. 2. Overview of cardiovascular sex differences from a molecular to population level.

Sexual dimorphism manifests in various ways, where seemingly minor differences at a small scale can lead to meaningful variations on both an individual and population level. Many of these phenomena are interconnected, such as sex hormone–dependent ion channel regulation and electrophysiological differences at the tissue/organ scale, and are driven by hormonal and/or chromosomal imbalance. Note that this figure does not serve as an exhaustive list. PPARα, peroxisome proliferator-activated receptor alpha; ESC, embryonic stem cell; ECG, electrocardiogram; E-C, electrochemical; AP, action potential; SMC, smooth muscle cell; EC, endothelial cell; CM, cardiomyocyte; CF, cardiac fibroblast. Created with BioRender.com.

Sex effects begin at development. The sex-determining gene Sry, found on the Y chromosome, dictates the formation of testes over ovaries and is the root of male/female gonadal hormone imbalance, a major factor in general sexual dimorphism. Gonadal hormones, such as testosterone, estrogen, and progesterone, then act directly on the cardiovascular system, affecting form and function differently between sexes (43). Sex hormones regulate ion channel function, a key aspect of cardiac electrophysiology, through intracellular signaling via receptor binding and transcriptional regulation (44). Estrogens exhibit a cardioprotective effect that has been linked with improved recovery following ischemic events in women (45), decreased CVD incidence (46) and mortality risk for premenopausal women (47) and accompanied increase in CVD risk post-menopause (48–50). Alongside driving hormonal imbalance, Sry and sex chromosome complement influence autosomal gene expression, furthering sexual dimorphism (51, 52). Ronen and Benvenisty’s (52) 2014 study found that out of 227 genes exhibiting differential expression between male and female pluripotent cells, 85 genes had Sry binding sites. This suggests that the Y-specific gene functions as a genetic modulator.

Sex chromosomes can yield sex-specific effects independent of hormones through gene expression modulation. Both Y chromosome presence and X chromosome dosage contribute to sex bias at the transcriptional level, driving downstream sexual dimorphism. This is especially pronounced during early cardiac development. Differential expression of homolog genes on the X and Y chromosomes drives sexual divergence during cardiogenesis, likely underpinning sex-specific CVD phenotypes (53). X dosage plays a major role in early development before X chromosome silencing. Distinct female embryonic stem cell (ESC) pluripotency phenotypes arise in an X dosage–dependent manner. A set of X-linked genes involved in differentiation pathway activity and pluripotency factor expression were identified as mechanistic determinants of these phenotypes (54).

Genomics is relevant in adulthood as well. Winham et al. (55) review the sexually divergent impact of genetics on CVD risk factors and anatomical pathologies. As an example, the high risk of hypertension in males can be linked to Sry regulation of tyrosine hydroxylase, whose activity is increased as a result. Genetics plays a role in inflammation, lipids, and “metabolic syndrome” (a term for several risk factors linked to metabolism) as well. Ethnicity and hormonal status are highlighted for their confounding effects on the role of genetics in sex differences. Sex-specific gene network regulation underscores some differences in CVD pathology. Regulation of cardiac hypertrophy, a sexually dimorphic pathological phenomenon, can be in part attributed to peroxisome proliferator-activated receptor alpha (PPARα) signaling. Harrington et al. (56) show that PPARα, amidst a wider gene regulatory network, drives the sexually divergent hypertrophic response. Deciphering the mechanistic pathways driving sex-based differences in CVD pathology can help identify therapeutic targets and inform better treatment strategies.

Sex divergence through development yields sexual dimorphism in adult cardiac physiology. Beginning in adolescence, the increased rate of hypertrophy in males results in higher cardiac muscle mass than in females throughout adulthood (57). Sex, particularly gonadal hormone imbalance, plays a role in non-myocyte cardiac cell populations as well. Male and female murine cardiac tissue show distinct cellular compositions, with differing relative amounts of endothelial cells (ECs), resident mesenchymal cells, and leukocytes. Four weeks after gonadectomy, cardiac cellularity is no longer distinct between male and female mice, revealing the role of gonadal hormones in driving sex-specific cardiac composition (58). Sexually divergent cellular composition is seen in human hearts as well. Female hearts tend to have more ventricular CMs and cardiac fibroblasts (CFs), while male hearts have more ECs, among other cell types (59). In their 2022 review, St. Pierre et al. (60) describe sex-based differences in cardiac size, geometry, function, and mechanical properties. Both female heart size and ventricular wall thickness are lower than their male counterparts. Twenty-one different metrics for cardiac geometry and functionality are shown to have sex differences, with male heart mass, volume, and output being greater and female ejection fraction and heart rate being slightly higher. Last, the healthy female heart has 10 to 14% higher contractility, as measured by myocardial strain, than the male heart. Electrophysiological metrics show sex bias from cell to organ level, including electrocardiograms (ECGs), contractility and excitation-contraction (E-C) coupling, action potential (AP), and ionic currents (44).

Cardiac form and function affect CVD pathophysiology in a sex-dependent manner. Walker et al. (59) review how cell sex drives CVD differences for various cardiac cell types. Smooth muscle cells (SMCs), a vital component of vascular walls, contribute to atherosclerosis via plaque formation and extracellular matrix (ECM) deposition. Phenotypic differences exist between male and female SMCs. They also exhibit the highest amount of differentially expressed genes of all cardiac cell types. It is no surprise then that SMC-related pathologies show sexual dimorphism: Females typically exhibit a higher risk of coronary artery stiffness, while males carry a higher risk of mortality from atherosclerosis and hypertension. Sex-biased CVD pathophysiological phenomena exist for ECs, CMs, macrophages, valve cells, and CFs as well (59).

These apparently benign sexual dimorphisms translate to meaningful clinical differences, warranting further exploration. As more light is shed on the deterministic relationship between biological sex and CVD, the biomedical community must determine how these findings might influence and improve research going forward.

CONSIDERATION OF SABV IN PRECLINICAL CVD MODELS

While CVD sexual dimorphism has been widely established, the incorporation of sex into basic and preclinical experimental design has been slow to non-existent. Including SABV earlier in CVD research can improve scientific rigor and ultimately human health. In the following sections, we will identify weaknesses in current standards for integrating SABV in vivo and in vitro, describe the impacts of avoiding SABV inclusion, and identify emerging strategies for sex-specific modeling.

In Vivo

Animal Studies

For experiments involving animals, sex-inclusive design begins with an even ratio of male to female subjects. The U.S. Food and Drug Administration’s (FDA) 1987 Guidance for Industry set the expectation that applicants use both sexes of animals in the preclinical studies where the investigational new drug (IND) is intended for use in males and females. There are no specific requirements for the male-to-female ratio, however. While much work has been done to improve women’s clinical representation, the U.S. FDA has made no further amendments to expectations for sex inclusivity at the preclinical stage, leaving the decision up to applicants.

This lack of solid regulation over sex representation and reporting is reflected in the literature. Studies using mouse models published from 1994 to 2014 reported using both sexes only 10.4% of the time (61). A shocking 34.3% of the 15,311 papers evaluated did not specify the sex of the mice used. For cardiovascular mouse studies, both sexes were reportedly used at a similar rate (approximately 10%) to that of all disciplines combined. However, male dominance is more pronounced in this discipline with nearly half (47%) of studies using only male mice. By comparison, 23% of all mouse studies analyzed reported using only males, and 32% used only females. The increased sex reporting in cardiovascular studies (78% reported sex) compared to the total (66% reported sex) may reveal an otherwise obscured bias for male animal use. Trends in sex reporting and diversity show some improvement over this period. By 2014, nearly 75% of publications made some mention of animal sex, up markedly from 0% in 1994. Yet, the reported use of both sexes in mouse studies shows far slower improvements, reaching only 10% of publications in 2014 from 0%. Male- and female-only studies increased at a greater rate, reaching 23 and 32%, respectively (61). The American Journal of Physiology-Heart and Circulatory Physiology (Am J Physiol-Heart Circ) reviewed sex inclusivity in articles published in the journal from 1980 to 2020 and found a modest improvement in female representation over time. Yet overall, male animals are consistently overrepresented. For studies using rats, the most common animal type reported, 8.2% of subjects were female and 81% were male over the period evaluated. The remaining 10.8% were unidentified. For all other animal types, 16.6% were female and 38.3% were male. As of 2020, approximately 35% more male animals are used than females in Am J Physiol-Heart Circ–published studies (62).

Guidelines set by funding agencies like the NIH have encouraged balancing male and female animal counts in basic and preclinical research. Though a critical first step, this metric should not be the end goal. Further recommendations for evaluating sex-specific phenomena can be applied depending on primary research goals. Sex hormone– and chromosome-specific phenomena can be observed through careful design of mouse models. Arnold et al. (63) provide a comprehensive review of strategies for hormone/chromosome dose effect isolation in mice, including gonadectomy for hormone level control, four core genotype modeling for isolating sex chromosome complement from gonadal phenotype (XX/XY with testes versus. XX/XY with ovaries), and the XY* model for investigating chromosome type and dosing effects (XO versus XX versus XY versus XXY). While each model has its own limitations, performing a combination of models may effectively minimize confounding effects. When not explicitly looking at sex-based differences, having a general working knowledge of sex differences relevant to one’s field is recommended. By using age-matched animals and disaggregating data by sex, researchers can observe potential trends without requiring larger group sizes or additional experiments. Last, considering variables that may confound sex effects, such as hormone levels and menstrual cycles, may further reveal important trends (64).

In Vitro

Cultured cells and tissues are used to support basic science research, identify therapeutic targets, and drive early drug discovery. Conclusions made in vitro inform the design of subsequent animal and human studies, a critical step for market approval. To minimize the substantial costs associated with programs failing at the clinical stage, an ever more likely outcome in today’s regulatory landscape, scientists have developed more representative in vitro models using primary and immortalized cell lines in both 2D and 3D culture systems (65). While recapitulation of biological systems continues to improve, the field has yet to adequately address the question of exactly whose biology is being modeled. Given the wide intraspecies diversity of humans, it is critical that biomedical research is built upon models that reflect that diversity. Here, we evaluate SABV inclusion in CVD-related studies using primary cell lines, stem cell lines, and 3D models. The findings are summarized in Table 1.

Table 1. Summary of sex effects and SABV strategies by model type.

Evidence of sex-driven variability in the four most common in vitro CVD model types provides a rationale for SABV inclusion. Emerging strategies can be model-specific (e.g., isogenic hiPSC lines) or generally applied (e.g., informed sex selection).

Model type	How sex impacts in vitro modeling	Current SABV strategies
Primary cells	• Increased Cx43 expression in female rat CMs under normal and pathological conditions (68)	• Use single donor/sex at a time
	• Sexually dimorphic cardiac fibrosis response in human and rodent CFs driven by estrogen signaling (71)	• Replicate sex-specific physiological conditions (e.g. sex steroid levels)
	• Increased elastin-to-collagen ratio in vascular SMC matrix deposition exposed to female sex steroids versus male (73)	• Increase sample size by choosing easy to acquire cell types
	• Differential expression of genes related to atherosclerosis and plaque instability in primary aortic ECs via distinct estrogen receptor phenotypes (75)	• Match M and F samples by other donor characteristics such as age, health condition, hormone status, ethnic background, etc.
	• Favorable energetic response to serum starvation and VEGF treatment in F-HUVECs (79)	• Pool multiple donors of the same sex into a single sample
	• Higher migratory capacity, cell proliferation, capillary formation, NOS3 mRNA, and protein expression in F-HUVECs over males (77)
	• Increased autophagy (marked by elevated beclin-1 and LC3-II/LC3-I ratio) in M-HUVECs (77)
Embryonic stem cells	• Differing transcriptomes in male and female mESCs throughout cardiac development (91)	• Initial and periodic karyotyping of all ESC lines
	• Three distinct classes of F-hESCs based on XCI reliability and related epigenetic stability (94)
	• Increased pluripotency and slowed differentiation initiation in F-hESCs due to X dosage effects and reduced MAPK signaling (54, 93)
	• Amplified sex chromosome-linked differences due to aneuploidy arising in long-term ESC culture (95, 96)
Induced pluripotent stem cells	• Epigenetic variability arising from unreliable XCR/XCI in F-hiPSCs (94)	• Provide a statement on how sex was considered in study design/evaluation and its potential role in the phenomena being studied
	• Lingering epigenetic signatures of somatic cells through incomplete XCI (101)	• Use single sex for all experiments
	• Sex-biased genomic imprinting aberrations in miPSCs (102) and hiPSCs (103, 104)	• Collect qualitative data rather than quantitative
	• Differential expression of Y-specific gene Sry in hiPSCs (52)	• Isolate sex effects using isogenic cell lines (106)
	• Naïve pluripotency state of F-iPSCs versus primed state of M-iPSCs (106)
	• Increased cardiac differentiation capacity of F-hiPSCs linked to higher X dosage (107)
	• Increased sensitivity of F-hiPSC-CMs to hERG channel blockers cisapride and dofetilide (111)
3D model systems	• Increased potential for sex-based variability in higher-order systems as complexity increases	• Select relevant cell sex
	• Sex differences in gene expression, cell composition and behavior, ECM, contractility, electrophysiology, calcium handling, and metabolism in 3D models (119)	• Build 3D models using cells from both sexes to evaluate sex effects on model performance
		• Add sex hormones to recapitulate native cardiac tissue development
		• Issue warning statements on sex-to-sex translatability of 3D models

Primary cells

Cardiovascular research using primary cell lines suffers from severe underreporting and male bias. Taylor et al. (66) showed that only 28% of research articles in this space reported sex and, of those, 68.9% used only male cells. No groups used exclusively female cells. In a 2021 study of sex reporting in the American Journal of Physiology-Cell Physiology (Am J Physiol-Cell Physiol), only 66% of publications using primary cells reported cell sex, with human primary cell studies exhibiting greater sex omission than cells of other species. Female cells were reportedly used in 38% of human primary cell studies (15% used only female cells) (67). Underreporting leads to poor translatability and reproducibility, ultimately stunting scientific innovation.

It is likely that in vitro primary cells faithfully recapitulate cellular sex differences observed in vivo, making cell sex a critical consideration. Stauffer et al. (68) showed that isolated female rat cardiomyocytes (CMs) have higher connexin43 (Cx43) expression than male rat CMs under both normal and pathological conditions. This supports the theory that human females’ reduced susceptibility to sudden cardiac death is a function of differential Cx43 expression and response to pathologic stimuli between sexes. Cardiac fibrosis, the scarring of heart tissue by CFs, is a major contributor to heart failure. A sexually divergent fibrotic response has been attributed in part to estrogen signaling, leading to reduced fibrosis in female patients (69). Given the conservation of CF subphenotypes across in vivo and in vitro primary cell models (70), cell-level sex-based differences are likely translated to primary rodent and human CF studies. Dworatzek et al. (71) determined the mechanistic role of estrogen signaling on sexually dimorphic fibrotic responses in both isolated human CFs and rat CFs. As both male and female CMs and CFs have functional estrogen receptors (72), the role of this sex hormone may be significant in other primary cell models as well. In vascular SMCs, sex hormones affect ECM protein deposition and metalloproteinase expression in vitro. Specifically, female sex steroids (17β-estradiol and progesterone) greatly increase the elastin-to-collagen ratio relative to testosterone (73). Given the importance of the structural protein composition of ECM on vascular biomechanics, human vasculature models using primary cells must consider cell sex in this regard. ECs, a critical component in vasculature lining and cardiovascular regulation, contribute to CVD and associated risk factors through endothelial dysfunction (74). Their role in both supporting healthy cardiac function and driving CV pathologies makes ECs ideal candidates for disease and drug modeling; however, they too are subject to sex-based variability. Varying estrogen receptor phenotypes among primary aortic ECs have been linked with differing expression of key genes in atherosclerosis and plaque instability, both major drivers of CVD mortality (75). Human umbilical vein ECs (HUVECs) are another major cell source for cardiovascular modeling and have been used to study hypertension, atherosclerosis, hypercholesterolemia, diabetes mellitus, angiogenesis, and other relevant phenomena (76). Sex affects HUVEC cell proliferation, migration, protein expression, and autophagy (77, 78). Sex-specific responses to metabolic perturbations have been observed, revealing a more favorable energetic profile in female HUVECs after starvation and VEGF treatment (79). This metabolic sexual dimorphism is theorized to drive increased cell motility and capillary formation in female HUVECS over their male counterparts (77).

Quantitative differences between major cardiac cell types in males and females bolster the need for consideration of sex in research models. Yet, due to high genetic variability between primary cell donors, direct comparison is often impractical. In this case, a single donor, whether male or female, is often used. Despite this limitation, groups are still able to evaluate sex effects by replicating sex-specific physiological conditions. Natoli et al. (73) investigated the role of sex steroids in human SMC (hSMC) protein deposition and arterial stiffening by exposing female hSMCs to physiological levels of 17β-estradiol, progesterone, 17β-estradiol plus progesterone, and testosterone. By using a single donor, the group was able to isolate matrix protein deposition as a variable and provide mechanistic evidence for increased arterial stiffening in post-pubescent males, relative to age-matched premenopausal females. When possible, it is generally preferential to use multiple donors from each sex. This raises the issue of how best to match donors of the opposite sex for comparison. Strategies for donor matching can vary widely depending on cell type and source. For smaller sample sizes, it is advised to match cells by as many non-sex characteristics (e.g., age and health condition) as possible, thus minimizing sources of variability. Adult arterial and venous ECs, though more biologically representative, are harder to acquire than HUVECS, making these the most used cell type (80). With higher feasible n values, HUVEC studies naturally provide more statistical power and reduce donor-to-donor variability concerns. Addis et al. (77) report using many male (n = 85) and female (n = 91) HUVECs to fully investigate the male versus female HUVEC phenotype. Yet, for most cell types, achieving such high donor numbers is neither feasible nor necessary from a statistical standpoint. A more common practice involves using pooled HUVEC samples, wherein cells from multiple donors of the same sex are pooled together (80, 81). Pre-pooled HUVECS are commercially available and advertised for their ability to overcome donor-to-donor variability.

Stem cells

Access to primary cells is often limited by donor availability and funds. CMs, one of the most common cell types used in CVD research, do not proliferate and thus are an unsustainable resource. Conversely, stem cells offer great potential for sex-specific modeling and therapeutic development due to their highly proliferative and pluripotent nature. Directed differentiation into CMs has been described for ESCs (82) and induced pluripotent stem cells (iPSCs) (83). Stem cells can theoretically provide an endless source of functional, biocompatible CMs for modeling or implantation.

Research articles using cell lines suffer from sex omission even more so than primary cell studies. After advising authors to include reporting, Am J Physiol-Cell Physiol saw an increase in sex reporting from 25 to 50% of articles (67). This still leaves half of their publications without adequate sex information, and they are deemed a leading publication in this effort.

Embryonic stem cells

Mouse ESCs (mESCs) have been a widely used resource in preclinical cardiovascular research since being proven capable of cardiac-specific differentiation in 1985 (84). Since then, mESC-derived CMs have been instrumental in investigating cardiac development (85) and developing cardiac tissue engineering strategies (86). However, these cells are limited in their ability to accurately model human cardiac (patho)physiology due to developmental and functional differences. Human ESCs (hESCs) helped overcome these limitations, though these cells present complications of their own, namely, ethical concerns over the use of human embryos, the sensitivity of differentiation capacity to culture conditions, genetic variability yielding wide-ranging cell line quality, and risk of arrhythmia upon implantation in adult hearts (87). Still, hESCs provide invaluable insight into human cardiac development, pathophysiology, drug response, and regenerative strategies/outcomes.

In a 2007 survey of more than 40 international hESC derivation groups, 55% of responders deemed sex “essential data” when reporting hESC line derivation information. While this majority is encouraging, there remains 42% who find sex only “desirable data” and 3% who deem it “not required.” Overall, sex (determined through karyotyping) was ranked 13th out of 18 surveyed data points in order of reporting need (88). Despite the call for the inclusion of sex/karyotype, strict standards for sex reporting and SABV inclusion have yet to be established. In their 2021 updated Guidelines for Stem Cell Research and Clinical Translation, the International Society for Stem Cell Research (ISSCR) recommends both male and female animals undergo safety and efficacy testing, barring the existence of “a scientifically valid reason not to do so.” Their recommendation for in vitro models amounts to a single sentence, stating cells of both sexes should be used “whenever possible” (89). These vague, lax guidelines perpetuate sex bias and underreporting.

Unlike many other cell types, hESCs may have an overrepresentation of female lines. Ben-Yosef et al. (90) showed that hESC line derivation protocols and culture conditions yield a disproportionately high F-hESC population (55%), likely through an inadvertently favorable growth environment for XX cells. This is in line with their finding that 66.9% of 397 similarly derived hESC lines across 35 stem cell centers were female.

Several sex-associated phenomena have been shown in ESCs, revealing the need for stricter SABV standards. At each stage of murine cardiac differentiation, cell transcriptomes reveal sex-biased gene expression independent of sex hormones (91). Variability in cell line quality can be linked to X chromosome activation status. X chromosome inactivation (XCI), a gene dosage compensation mechanism, is marked by a near-total silencing of one of the two X chromosomes present in female genomes. Early in development, hESCs have not yet undergone XCI and still contain two active X chromosomes. X dosage effects are thus more pronounced, with female embryonic cells receiving a double dose of X-encoded genes compared to males (92). Sex-specific effects linked to this include increased pluripotency and slowed initiation of differentiation in F-mESCs versus M-mESCs, mediated in part through reduced mitogen-activated protein kinase (MAPK) signaling (54, 93). XCI neutralizes X dosage effects in utero, removing the differentiation block on F-ESCs and allowing development. However, the questionable reliability of XCI in vitro raises concerns about the epigenetic comparability of F- and M-ESCs in modeling systems. While F-mESCs undergo XCI ex vivo reliably, F-hESC lines fall into one of three categories based on XCI likelihood (94). Class I lines most closely resemble mESCs by initially carrying two active Xs and undergoing XCI during differentiation. Class II lines carry one active and one inactive X chromosome, likely conferring a partially differentiated phenotype. Class II lines are considered the least epigenetically stable due to the loss of XIST expression after XCI (94). This epigenetic variability brings into question the quality of F-hESC lines and may require the time-consuming classification of each female cell line.

Through prolonged culture, ESCs, and SCs in general, are susceptible to chromosomal and genetic abnormalities (95, 96). Aneuploidy in sex chromosomes may add to sex-specific chromosomal effects. To monitor this, some groups recommend karyotyping ESCs periodically. Several methods for relatively easy karyotyping exist and can be selected according to personal preference (97, 98).

Induced pluripotent stem cells

The field of regenerative medicine, including tissue engineering, personalized medicine, and stem cell therapy, is growing rapidly with mouse (miPSCs) and human iPSCs (hiPSCs) playing a central role in its success. Despite noted gender- and sex-based variability in many of the diseases this field seeks to treat, research on iPSCs still suffers greatly from a lack of reporting on cell donor demographic information and bias for male cells (99). Vendors often do not disclose cell line sex information, and when they do, typically offer far more male stem cells than female. This drives insufficient reporting in academic papers and further obscures sex-related phenomena.

Donor sex can play a role in the reprogramming, pluripotency, and differentiation of hiPSC lines—all of which engender experimental variability. Reprogramming of somatic cells introduces sex-based variability likely due to X chromosome dosage/silencing (100) and differential epigenetic signatures (52). The two are interconnected, with X dosage determining global DNA methylation in female iPSCs (F-iPSCs) (100). Upon reprogramming, female cells undergo X chromosome reactivation (XCR), doubling X dosage and lowering DNA methylation as compared to male cells at the same stage. Typically, F-iPSCs undergo subsequent XCI during differentiation as in embryonic development, returning their epigenetic state to one similar to male iPSCs (M-iPSCs) (100). However, this XCR/XCI process does not always occur, leading to epigenetic variability among female cell lines (94) and between male and female lines. Tchieu et al. (101) showed that epigenetic signatures of the somatic cell type can persist through reprogramming as a result. Specifically, F-hiPSCs that do not undergo XCR during reprogramming retain one inactive X chromosome in their pluripotent state, resembling the somatic (fibroblast) XCI status more so than the targeted embryonic state. This is confirmed by the nonrandom silencing pattern observed in each hiPSC line, where the same X chromosome is inactivated in each cell. Had reprogramming induced XCR and subsequent XCI, X silencing would display a mosaic, random pattern as in embryonic development.

Sex-biased genome imprinting defects, phenomena occurring through reprogramming, also highlight the influence of sex on the miPSC (102) and hiPSC genomes (103, 104). Aberrant genome imprinting and DNA methylation can persist through culture and differentiation (103), rendering cell lines incapable of disease modeling and even dangerous for regenerative medicine applications (104). This ultimately calls for careful consideration of XCI status and epigenomic reprogramming fidelity, two sex-biased phenomena, as indicators for iPSC line quality.

It should be noted that sex is not indicated as a determinant of reprogramming efficiency itself (105), yet these genomic and epigenomic differences may have concerning downstream effects on cell line quality. Notably, expression of the Sry gene, described above as a major driver in sex-specific embryonic development, is induced through pluripotency reprogramming (52). Its discriminatory expression in M-hiPSCs, and not F-hiPSCs, may confer sex-specific phenotypes upon both the iPSC populations and their lineages.

Whatever the underlying cause, differences have been noted between male and female cell lines. For instance, F-iPSCs are described as being in a more naïve state, whereas M-iPSCs are more primed toward specific cell fates (106). D’Antonio-Chronowska et al. (107) showed that F-iPSCs have a higher propensity for cardiac differentiation than M-iPSCs. Furthermore, they showed that F-iPSCs with increased X dosage (through XCR) have higher cardiac fate determination, linking differentiation capacity with chromosomal imbalance.

Sex differences between M-iPSCs and F-iPSCs persist through differentiation as well, though this research area remains to be further explored. HiPSCs can be successfully directed toward cardiovascular cell types, including CMs (83), CFs (108), vascular SMCs (109), and ECs (110). According to Zeng et al. (111), hiPSC-derived CMs (hiPSC-CMs) respond in a sex- and ethnicity-specific manner to several classes of pharmaceuticals. Namely, female hiPSC-CMs (F-hiPSC-CMs) exhibit higher sensitivity to torsadogenic drugs such as cisapride and dofetilide than male hiPSC-CMs (M-hiPSC-CMs). Preliminary CVD drug toxicity and/or efficacy studies must be sex-specific considering this.

Knowing that donor sex contributes to much of iPSC interline variability, how is this accounted for in iPSC-based research? In their 2023 article on the use of iPSC-derived SMCs (iPSC-SMCs) in cardiac modeling, Liu et al. (112) acknowledge under “limitations” that in only using three cell lines, they were unable to evaluate any influence of sex on model variability. Similarly, Kim et al. (113) address the limitations of their endothelial dysfunction model using both healthy and diseased donor-derived hiPSC-ECs. Two of the three hiPSC lines used for both the healthy and diseased phenotype were female by origin. They highlight known sex differences in epigenetic marks of hiPSCs and how this may obfuscate the results of their study. Their suggestion is to eliminate sex effects by isolating one sex at a time. Explicitly addressing how sex is considered in the study design, even if not at all, can be helpful to other researchers hoping to learn from the published work.

When sex is being evaluated, research goals may need to be adjusted based on feasibility and available resources. When comparing hiPSC-CMs from donors of different sexes, Zeng et al. (111) avoided seeking quantitative data, instead opting for simple “yes/no” classifications. With only three male and three female donors, achieving statistical power for more quantitative data would have been unlikely. Yet, despite inter- and intraline variability, they were able to show increased susceptibility of F-hiPSC-CMs to early afterdepolarization events following exposure to several different hERG channel blockers.

Waldhorn et al. (106) have pioneered new mechanisms for isolating sex-related effects from confounding genetic variability via isogenic cell lines. They took advantage of abnormal chromosome pairing found in mosaic Klinefelter syndrome to generate four iPSC lines with differing sex chromosome complements (XXY/XX/XY/X0) from a single patient. Using this platform, the effect of X chromosome dosage (X0 and XY versus XX and XXY) and Y chromosome presence (XX and X0 versus XY and XXY) on various cellular processes can be evaluated without the need for large sample sizes. While this model can successfully evaluate chromosomal imbalance, it does not account for hormonal imbalance, which is also known to drive sex-based differences. This demands consideration when drawing conclusions from experiments using this method.

3D models

With the advancement of technology, traditional 2D in vitro models are increasing in complexity to form 3D systems that more accurately represent the tissues and organs they are used to study. 3D modeling techniques, such as 3D cell culturing, engineered tissues, organoids, and organ-on-a-chip technologies, are being developed for various tissue and organ types, including cardiovascular. These 3D techniques can be harnessed for disease, tissue, and organ modeling (114, 115) and as regenerative therapeutic tools (116, 117). In combination with patient-derived iPSCs, cardiac tissue constructs could provide personalized CVD prevention, diagnostic, and treatment platforms. The expansion of precision and regenerative medicine demands robust development of these 3D techniques.

Despite considerable sexual dimorphisms in native tissues and organs, in vitro models of these systems still suffer from a lack of sex specificity and male bias. A 2022 study found more than half (53%) of 100 recent papers on organ-on-a-chip and organoid technologies did not specify the sex of the cells used. While 20% used both male and female cells, none of these groups performed a statistical evaluation of sex effects. Of the remaining 27%, only male cells were used in 22% while female-only studies made up just 5% of the total (118).

Increasing complexity amplifies interline sex-based variability. From genetics to tissue functionality, sex-based variability could be introduced at every level of a 3D model. Sex differences in gene expression, cellular composition and behavior, ECM, contractility, electrophysiology, calcium handling, and energy metabolism in native tissue must be considered when building a 3D cardiac model (119). The multiplex nature of these models makes them more susceptible to the unintended introduction of bias. This is especially important as they are often intended for disease or drug response modeling, two areas known to be sexually divergent. Understanding and designing for sex-based variability becomes crucial to the success of higher-order biological platforms.

Because 3D models aim to mimic native organs, additional parameters should be considered to avoid sex bias in modeling. Selecting relevant cell sex remains an important tool for ensuring clinical translatability, as in 2D studies. At a minimum, the use of both male and female cells could be considered standard when developing microphysiological model systems. Sex hormone addition to culture media can be useful in both 2D and 3D systems as well, depending on the phenomenon being studied. This may prove additionally useful in developmental studies, where sex hormones are known to play a role in dictating cardiac cellularity (58). The structural integrity of the native tissue/organ being modeled is a critical additional component in 3D. As noted above, cardiac physiology differs by size, geometry, and cellularity between the sexes. Accurate recapitulation of these sexual dimorphisms is critical to yielding physiologically relevant data.

Some researchers are now listing warning statements in their publications on the potential experimental variability introduced by differing cell line sex. Drakhlis et al. (120) provide a cautionary note that experimental results using their described cardiac organoid platform may vary according to cell line sex. Thus, it is “important to use cells from both sexes.” They go on to say that their protocols have been successfully tested using both male and female cell lines. Transparent communication of cell-based platforms’ robustness across sexes, whether it was tested, could be easily implemented as a publication standard. Furthermore, indicating the degree to which sex-based differences have been evaluated can aid future researchers in cell line selection. This requires no added time or energy from the research group but may increase awareness of sex-based variability within a field and improve translatability, a serious concern for the scientific community right now.

CURRENT BARRIERS TO INCLUSIVE DESIGN

In the following section, we describe key barriers and potential solutions to SABV inclusion in CVD preclinical research. Key findings are summarized in Fig. 3.

Fig. 3. SABV inclusion strategies and major players.

Five key groups are identified as central to the SABV practice overhaul. Inclusion strategies are divided into research- and reporting-focused initiatives. SABV, sex as a biological variable. Created with BioRender.com.

The first step toward more inclusive design is assessing the field’s current standards. The lack of donor demographic reporting in the literature presents a sizeable hurdle to this. For cell line sex, insufficient reporting has been noted in biomaterials research (121), commercial cell vendors (99), and basic science research (66, 67). Sex reporting in academic papers is likely hindered in part by cell sex omission by commercial vendors. Kim et al. (67) theorize that reduced sex reporting observed in purchased primary cell studies (24% reported) relative to self-collected cell studies (72% reported) is likely a result of vendors not providing cell sex. A 2013 analysis of three major cell banks (American Type Culture Collection, European Collection of Cell Cultures, and Japanese Collection of Research Bioresources) revealed that, out of 1495 human primary cells offered, 96.6% lacked cell sex information. More than 91% of mouse, 82% of rat, and nearly 16% of total available human cell lines also did not provide sex information. For both primary and stem cell lines (those with sex data), male cells were offered more frequently than female cells (99). In addition, voluntary self-reporting by cell donors, the most common method for acquiring sex information for human cells, is vulnerable to error. Donors may confuse sex and gender distinctions, leading to inconsistencies. One solution for this is comparing donor-provided sex with standard sex-specific genetic markers and excluding or relabeling samples that do not match (105). Furthermore, the lack of cell sex reporting can be overcome through gene expression analysis. Using PCR, Liu et al. (112) evaluated Sry expression in two iPSC lines, which were both determined to be female. This simple technique can allow researchers to continue using any cell lines they have on hand, despite missing donor information.

Major publishers could establish standards for reporting donor sex and other demographic information. Am J Physiol-Heart Circ published an editorial evaluating sex and gender reporting in cardiovascular research in response to the reader and contributor outcries for better practices (122). The following year, Am J Physiol-Heart Circ provided explicit instructions for increasing SABV inclusion in experimental design and reporting. Clear guidance and standards from major publishers could greatly incentivize sex reporting among researchers. As of 2020, the following journals regularly publishing cardiovascular research have some editorial policy regarding sex reporting: Am J Physiol-Heart Circ; Arteriosclerosis, Thrombosis, and Vascular Biology; Circulation; Circulation Research; and Journal of the American College of Cardiology. Other major journals including Cardiovascular Research, European Heart Journal, Journal of Cardiovascular Pharmacology, and Journal of Molecular and Cellular Cardiology do not have explicit policies on this (123).

Isolating demographic variables presents a huge barrier to inclusive design. It requires substantial time and resources and thus is unattractive to include, especially in conjunction with primary research work. Kilpinen et al. (124) found that 5 to 46% of hiPSC line phenotypic heterogeneity can be attributed to donor genetic variability. Detangling sex effects from genetic variation demands large sample sizes for strong statistical significance. Increasing sample size may require more resources, such as cell lines, animals, labor, and consumables. For primary cell studies, sample sizes would need to be effectively doubled to include both male and female cells and prevent sex bias. Furthermore, if the aim of the study is to analyze sex effects specifically, then more than three sources of male and female cells are required, increasing sample size and experimental conditions sixfold (118). Understanding that acquiring these resources is often impractical, Am J Physiol-Heart Circ does not require groups seeking publication to investigate the mechanisms causing observed sex-based differences where the goal of the manuscript is not to do so (64). In other words, if the research is not explicitly investigating sex-related phenomena, yet sex-specific trends are observed, then authors need only report, not explain, these findings. This is meant to encourage reporting of significant sex-related patterns but minimize undue labor. As cell lines represent an unlimited resource, they may more easily be shared in support of sex-inclusive design. One solution might be encouraging cell line sharing among groups within a particular research area. By alleviating the financial burden of smaller, particularly academic, groups, the field may collectively benefit from increased scientific rigor and excellence.

Comprehensive best practices can offer a helping hand to investigators wanting to incorporate biological variables into their research. Rich-Edwards et al. (125) provide straightforward and thorough advice on SABV inclusion from bench to population-level studies. Potential pitfalls and opportunities for improvement are provided for each stage of research design, from motivation and subject selection through data collection, analysis, and reporting. With broad contributions from researchers, these best practices can be further optimized and tailored for certain disciplines. Integrating this knowledge into science education at earlier stages can help establish SABV inclusion as the “status quo.” The NIH ORWH’s “Sex as a Biological Variable Primer” course provides four modules (SABV and the Health of Women and Men; SABV and Experimental Design; SABV and Analyses; SABV and Research Reporting) which describe the importance of SABV and how to begin incorporating it into the experimental process (126). The basic outline of this course could be adapted and expanded for more in-depth learning at the collegiate level. Requiring some base understanding of sex as it relates to health equity could improve standards for inclusive design.

Simply identifying these strategies is not enough to move the needle. Major players will need to make changes, both immediate and gradual, toward achieving this collective goal. Some changes, such as increasing cell sex reporting and cell line sharing, can be implemented effectively overnight, ignoring some short-term logistics. Meanwhile, other efforts like crowdsourcing best practices, integrating SABV into education programs, and grant fund allocation will likely require extensive coordination and lengthy roll-out periods. Note that each player will have individual circumstances which may alter their respective implementation timeline. For example, a researcher may not have access to cell sex information for reporting and thus are dependent on vendor disclosure or sex genotyping to do so. In this example, the respective timelines for acquiring cell sex information via either vendors or in-lab testing could vary widely. Thus, there is some built-in ambiguity as to which steps are the quickest to implement, and each major player is expected to decide for themselves a realistic timeline. Strong leadership will likely be needed to ensure timely changes. If powerful players like funding bodies and higher education institutions set clear goals within a specified timeframe, then change can be enacted on both individual and systemic levels.

SUMMARY AND OUTLOOK

The state of sex-specific CVD research and drug development remains stunted despite numerous calls to action. While efforts to improve women’s clinical participation have seen progress, the same cannot be said for earlier, preclinical research. The lack of female representation in vitro has wide negative consequences for the field, as well as broadly across disciplines. Sex differences are highly prevalent in human health and disease, with phenomena reported for all leading disease-related causes of death [i.e., cancer (127), coronavirus disease 2019 (128), stroke (129), chronic lower respiratory diseases (130), Alzheimer’s disease (131), diabetes (132) and more]. With the highest mortality rates worldwide, CVD presents a critical opportunity for improving human health through better sex representation. As the biomedical sciences have learned, conclusions drawn from non-diverse research groups cannot successfully be extrapolated to other groups. Parallels can be drawn to the historical adaptation of all-male clinical trials to female patients, which yielded many negative outcomes. To avoid falling into the same trap, SABV must be integrated into CVD research not only at the clinical stage but also during preclinical and basic research timelines. Achieving this goal will take cooperation between all major stakeholders, from scientists and funding bodies to patients and medical providers. With concerted effort and strategic changes, we might approach a more equitable future for global cardiovascular health.