The choice of an experimental system will depend on the question being asked and the hypothesis to be tested. In contemporary publishing, methodological details including the sex of the research material or animals are often omitted from research papers or relegated to supplementary material. However, it is just these methodological details that allow investigators to reproduce experiments of others and to understand how important variables such as sex or hormonal status may influence outcomes. The Institute for Laboratory Animal Research and their official publication (ILAR Journal) is a rich resource for details regarding reproductive, hormonal and developmental information of various animal species used in cardiovascular disease research. Several general considerations are provided below for some of the more common experimental systems employed to address mechanisms of cardiovascular function.
Cells in culture and isolated tissues
Cultured vascular endothelial cells, smooth muscle cells and cardiac myocytes (including neonatal cells), as well as isolated blood vessels and hearts, are used in experiments that explore intracellular signaling mechanisms in the development and treatment of cardiovascular disease. While most signaling pathways may be common in cells/tissues derived from female and male animals, it is important to understand which pathways may or may not show a sex difference as gene and protein expressions are influenced by both sex and hormones [55–62]. The same can be said for stem and progenitor cells being cultured for cell-based therapies [63–66]. For example in mice, cells derived from female animals appear to be more effective at both reversing disease and restoring a pre-disease level of cells to bone marrow than male-derived cells . Differences in the efficacy of male and female cells could possibly be attributed to differences in paracrine factors (for example, cytokines) . Comparison of these preclinical vascular data to early human clinical data suggest that sex-based differences in progenitor cell numbers and function seen in mouse models of atherosclerosis and acute myocardial infarction reflect human clinical scenarios .
The following points should be addressed in design of experiments using cultured cells, isolated stem and progenitor cells, and isolated tissues: (1) is expression of the receptor, pathway, enzyme, and so on, or the number of isolated progenitor cells affected by the sex or hormonal status of the donor animal? For example, would gene expression/signaling differ depending on whether the donor animal was studied before sexual maturity, was sexually mature (null or multiparous), or was reproductively senescent? (2) Would number of cells, gene expression or phenotype be affected by gonadectomy of the donor either before or after sexual maturity? (3) How does the expression of the pathway of interest or phenotype of progenitor cells change with passage of cells exposed to hormones in culture media using fetal bovine serum? (4) Is the choice of cell/tissue donor appropriate for the mechanism of interest related to human disease in regard to sex, age and hormonal status? (5) For cell-based therapies, given potential sex differences in cell number, phenotype, and potency, should sex mismatched allogeneic cells be considered as a therapeutic option? If so, is the donor appropriate for the mechanism of interest related to human disease in regard to sex, age and hormonal status?
The age/hormonal status of the donor for cell/tissue may not be available depending on the source of the material (that is, abattoir, human material for which medical information/records are not available, some immortalized cell lines). If the sex of the cell/tissue donor is not known, it can be determined by a PCR-based assay to identify specific fragments of the X and Y chromosomes. For example, a multiplex PCR-based method was developed to measure a 475-bp fragment from the ABCD1 gene (X chromosome) and a 231-bp fragment from the SRY gene (Y chromosome) . This method has been refined and extended to a high-throughput automated setting .
Sex of neonatal animals (including mice and rats) can be determined by examining the anogenital distance . In larger animals the presence of gonads should be confirmed as some male animals obtained from commercial suppliers (in particular, pigs) may be castrated at birth, and, therefore, studies in these animals would yield cells developed or differentiated under a sex hormone-depleted environment.
As sex differences are influenced both by the sex chromosomes and sex hormones, care must be taken to control the hormonal environment of cultured cells. Some media, in particular, fetal calf serum, contain sex steroid hormones that could influence the pathway/signal of interest. Media can be stripped of hormones by charcoal treatment. When hormones are added to the media, care should be taken to control for the solvent used for lipophilic compounds and to consider that both testosterone and 17β-estradiol may be metabolized by the tissue. Thus, a given concentration of sex steroid added at day 1 of a culture cycle may not be sustained over a long period of time. Alternatively, exposure of cells to hormones may 'imprint' the cell phenotype over several passages [70, 71]. As sex steroid hormones initiate rapid actions that do not require gene transcription (non-genomic actions) as well as effects on gene transcription (genomic actions) that represent different temporal sequences, duration of exposure to the hormone of interest is a critical consideration in study design.
For example, prolonged exposure of certain cells to 17β estradiol increases synthesis of endothelial nitric oxide synthase (eNOS) via a genomic mechanism caused by binding of the ligand-bound estrogen receptor α (ERα) to an estrogen response element (ERE) on the eNOS gene . However, estradiol also acutely increases eNOS activity and release of nitric oxide via a non-genomic effect that is due to an increase in intracellular calcium . Data also suggest that there is a plasma membrane-associated, G-protein coupled estrogen receptor (GPR30 or GPER) . However, the specificity of this receptor is controversial since GPR30 also mediates rapid aldosterone-mediated effects on the vasculature [32, 33, 72]. Androgens also produce genomic effects via androgen response elements (AREs) in genes, but cause acute vasodilation by a non-genomic mechanism that involves activation of calcium activated potassium channels [73, 74]. Future studies are necessary to completely understand the genomic and non-genomic effects of sex steroids and the mechanisms responsible for modulating cardiovascular function including whether the chromosomal constitution of the cell modulates responses to steroids via transcription factors, differential expression of protein chaperones, methylation of DNA, and so on [32, 33, 72].
Several mammals are used in studies of cardiovascular disease. General considerations for the selection of the appropriate species include cost, size, housing requirements, diurnal activity cycle, diet requirements (soy or plant based influencing consumption of phytoestrogen) , lifespan, frequency of estrous cycle in females, ability to perform genetic manipulations, ability to investigate developmental and/or environmental components of disease modulation . Stress imposed by disruption of sleep/wake cycles, patterns of social interaction (or degree of social isolation), or handling may influence some parameters (for example, corticosteroids and catecholamines) that interact with pathways activated by the sex-steroid hormones [77, 78].
Of the small mammals, rats and mice are used most often to model cardiovascular disease because of their short life span, short estrous cycle and gestation, modest cost of housing, and the ability to perform genetic manipulation in them. Both the strain and sex of the animals are known to influence expression of cardiovascular pathologies [79, 80].
Transgenic mice (and some transgenic rats) that contain estrogen and androgen receptor knockouts are commercially available for study. Although preliminary, use of these animals to study the onset and progression of vascular disease has shown that the number of vascular progenitor cells found in bone marrow and blood is decreased in estrogen receptor knockout (ERKO) mice compared to wild type controls. These transgenic animals are especially valuable to study the role of sex steroids and their receptors in cardiovascular disease. While the global ERKO mice are viable and fairly healthy, global androgen receptor knockouts (ARKO) have serious developmental problems . For example, the global ARKO exhibit neutropenia, osteopenia, low levels of serum testosterone, arrested spermatogenesis, and females have a reduced number of pups compared to wild type mice . Therefore, especially for androgen receptor, specific tissue-directed androgen receptor knockouts, using Cre-loxP methods, are preferred [82–85]. Cre-loxP technology allows for knockout or expression of a gene of interest in a specific tissue . Briefly, the site-specific recombinase Cre (cyclization recombination) from the bacteriophage P1 is used to induce recombination between two 34 bp recognition sites (loxP = 'locus of crossing over in P1') inserted into the genome. These loxP sites contain two 13 bp inverted repeats surrounding an 8 bp core sequence that provides directionality. Thus, two transgenic mouse strains are necessary to develop a cell-specific knockout: one strain that has the Cre recombinase expressed under the control of a cell specific promoter, and one that contains the gene of interest (or a critical exon of it) that is flanked by two loxP sites (termed 'floxed'). Following one round of crossbreeding, double transgenic pups containing both the floxed transgene and the Cre transgene are selected by genotyping. In a second round of breeding the double heterozygotes are either inbred or bred to the floxed mouse line. In the F2 generation, heterozygotes for the floxed allele and the Cre transgene are selected since the floxed gene/exon will be excised selectively in the cell types that express Cre recombinase (taken from review in ). For the androgen receptor, there have been several Cre/loxP mice developed for reproductive tissues, as well as adipose tissues, skeletal muscle, bone, T cells and B cells, liver and skin, to name a few . The use of cell-specific knockouts of AR thus allows the investigator to specifically study the role of AR in their tissue/cell of choice while minimizing the effects of global changes that would complicate the study and perhaps make interpretation ambiguous. It should be kept in mind that the gene for the androgen receptor resides on the X chromosome. Thus, studies androgen actions in females are complicated by X inactivation which results in mosaic expression of any polymorphisms that reside in this gene or others which have the potential to affect cardiovascular function in females.
One caveat about the use of mice, transgenic or otherwise, is that uncertainties still exist regarding the degree to which cardiovascular phenomena, especially the pathobiology of atherosclerosis in these rodents, can be translated to human beings (see review in ). However, Taylor and colleagues [65, 66, 88] have shown in apoE-/- mice and preliminarily in ERKO mice: (1) the onset and progression of atherosclerosis differs in male and female animals with the onset of disease occurring earlier in males and catching up with aging in females, similar to that in humans; (2) that the composition of the bone marrow in male and female animals differs and changes over time with aging consisting with the onset and progression of disease; (3) that the use of bone marrow mononuclear cells from females is sufficient to decrease atherosclerotic plaque in males but that the converse is not true; and (4) that the number of circulating cytokines differs based on sex and degree of disease with proinflammatory cytokines being higher in the circulations of males than in females.
Based on a spontaneous deletion of the SRY gene from the Y chromosome, Arnold and colleagues have developed transgenic animals for evaluating whether a phenotype follows gonadal sex or sex chromosome complement . For example, these investigators developed transgenic animals to 'knock-in' the SRY gene on an autosome of XX animals to provide mice that have testes but XX chromosomes; alternatively, with the spontaneous deletion of the SRY gene from the Y chromosome, mice have ovaries but XY chromosomes. This allows the investigator to determine if a sex difference in phenotype correlates with the type of sex chromosomes (XX or XY) or with gonadal sex of the animal. Using these intact and gonadectomized animals, investigators can separate the contribution of sex steroids from sex chromosomes (the interaction of sex and sex hormones, that is, sex specific hormone action) in mediating sex differences in cardiovascular function and dysfunction.
In addition to transgenic animals, inbred and congenic strains of rats have been used extensively to study mechanisms responsible for sex differences in cardiovascular disease etiology. For example, Dahl salt sensitive rats exhibit hypertension with aging or more rapidly with a high salt diet . High salt diet-mediated hypertension is exacerbated in males and ovariectomized females and attenuated in castrated males and intact females [91, 92]. Spontaneously hypertensive rats (SHR) also exhibit sex differences in blood pressure control, with males exhibiting higher blood pressures than females . Sex differences in the pressor response to angiotensin II have also been documented in mice and rats [94, 95].
With regard to rat models that mimic menopause, both SHR and Dahl salt sensitive rats have been used. Both strains of rats exhibit naturally increasing blood pressure when they stop estrous cycling, and have been used to study the mechanisms responsible for postmenopausal hypertension. However, in female rats after cessation of estrous cycling, estradiol levels do not fall as low as in women following menopause which is a common criticism of using rats as a model of menopause. To address this problem, investigators have used 4-vinylcyclohexene diepoxide (VCD), a chemical toxin that causes ovarian failure by targeting pre-antral follicles [96–99]. VCD can be used in rats and mice and allows for studies to be performed during a perimenopausal period [96, 100]. Treatment of animals with VCD causes a follicular depleted state and a cessation in the production of female ovarian hormones within approximately 50 to 75 days after injection, with full cessation of estrogen production and obsolescence of the follicles by 129 days . The advantages of this model are twofold. First, the rodents undergo a sustained period with no estrogen production. Second, VCD can be given in a young adult animal to simulate early ovarian failure . Since the VCD ovary produces androgens just as do ovaries of postmenopausal women, perhaps this best models ovarian failure in women [102–106]. The investigator should keep in mind that studies in women with early ovarian failure (19 to 39 years of age) showed that they responded to transdermal estradiol with reductions in blood pressure, angiotensin II and creatinine , whereas in women who go through natural menopause (average age 51 years), the effects of estradiol are not as clear which was evident by the results of the Women's Health Initiative . Thus, chemical menopause may be a better option to examine interactions of aging with estrogen depletion rather than the loss of all ovarian hormones as would result from ovariectomy.
There is a growing literature that considers sex differences in mouse strains and cardiovascular disease, sex-specific differences in the impact of particular bone marrow, blood, cardiac and vascular progenitors cells [65, 66] and sex-specific influence of particular gene effects [87, 109, 110]. In this light, many of the strategies used by geneticists to identify and characterize specific gene effects on cardiovascular disease-related phenotypes can easily accommodate tests of hypotheses surrounding the sex specificity of those gene effects, as will be discussed in the section on 'Identifying genes that exhibit sex-specific or sex-interaction effects' below.
Rabbits historically have been used in studies of autonomic and pharmacological regulation of vascular tone and vascular remodeling associated with atherosclerosis, ischemia reperfusion and stroke [111–122]. The Watanabe breed develops spontaneous atherosclerosis [123, 124], but other breeds develop atheromatous lesions in conduit arteries after feeding with high cholesterol diets [112, 113, 125, 126]. Sex differences for some vascular functions and castration with hormone replacement indicate that both sex and hormones modulate these effects [127–129]. Rabbits also express the same myocardial contractile protein isoforms as humans (for example, β versus α myosin heavy chain in the adult ventricle) and have a myocardial blood reserve that more closely resembles humans  more so than that of rats or dogs . Based on these similarities to humans, rabbits have also been used to evaluate acute myocardial infarction therapies [132–134].
Rabbits and rodents are coprophagic thus influencing dietary sources of protein . It is unclear how this activity influences immunological responses in these species related to nitric oxide synthase Il (iNOS) or other isoforms of NOS that might be regulated by estrogen and are associated with inflammatory responses proposed as a stimulus for endothelial dysfunction and vascular disease [136–139].
Larger mammals, such as cattle, sheep, dogs, pigs and primates, offer the advantage of scale for testing devices and procedures to be developed in humans (for example, ) but are disadvantaged by availability, cost, and special requirements for handling and husbandry. While some large animals are available from commercial breeders, others are not, and it may be economical for some tissues to be obtained from abattoirs (cattle, pigs). However, limitations on sources may make it difficult to determine the sex and or hormonal status of the tissue donor. Commercially available animals may be castrated at birth, which affects hormonally mediated developmental processes, or may be sexually immature at the time of study making extrapolation of data to adult animals problematic. Retired breeder females have been used for studies of aging, but it is unclear if cardiovascular function(s) of multiparous animals are the same or different from age-matched nulliparous females or aged matched males. Reckelhoff and colleagues reported that renal function was decreased in Sprague Dawley female rats who had had six to seven pregnancies and lactations compared to virgin females , whereas Baylis and colleagues reported that renal function and blood pressure were not affected by three pregnancies and lactations in spontaneously hypertensive rats . Retired breeder rabbits have been used for models of cell-based treatment for acute myocardial infarction and seem to be a reasonable model for human disease [132–134]. Thus, studies in retired breeders may more closely mimic the cardiovascular systems of most women than do virgin female animals.
Historically, dogs were the experimental animal of choice for investigating mechanisms of cardiac regulation and autonomic control of the vasculature, renal function, models of disease, development of imaging modalities, testing novel therapeutics, basic pharmacology of endothelial and smooth muscle function and aging [143–153]. Depending on the source of the animals, however, it may not always be possible to obtain information on age or reproductive history. While studies of ovariectomized and hormone replaced female dogs have been performed [152, 154], studies in castrated male dogs are not reported in cardiovascular literature.
Swine smaller than 100 kg used in research are usually sexually immature, with the exception of Yucatan mini-swine. Depending on the supplier, males may be castrated at birth, so designation of 'male' with a weight of less than 100 kg may not represent results comparable to sexually mature animals [155–157]. Alternatively, sexually immature females (about 3 months of age) are used to maintain manageable sizes. The level of maturity or hormonal status for these animals is often not reported in methods sections of scientific papers. Ossabaw miniature swine (Sus scrofa) have a 'thrifty genotype' that when fed a high caloric diet enables them to store fat in order to survive seasonal food shortages. The phenotype of these female animals including central obesity, insulin resistance, impaired glucose tolerance, dyslipidemia and hypertension, are characteristics comparable to those used to define metabolic syndrome in humans. The atherosclerotic lesions in coronary arteries of the diabetic swine are similar to those found in humans and the size of the arteries are acceptable for testing coronary interventions such as stenting [158–160], thus providing an appropriate experimental animal to evaluate basic mechanisms of type II diabetes and treatment strategies which might be more easily be translated to human disease/treatment .
Non-human primates have been used extensively to study the natural history of atherosclerosis in relation to sex differences (See [77, 161]). Rhesus and cynomolgus monkeys (Macaca mulatta, Macaca fascicularis, respectively) are particularly useful when fed diets that elevate blood lipids and the animals develop lesions consistent in morphological characteristics and location with those in humans with hyperlipidemia . For example, atherosclerosis develops first in the aorta and proximal portions of the main branch coronary arteries and later in the common and internal carotid arteries. At risk animals (that is, those consuming an atherogenic diet) experience myocardial infarction at a rate similar to that of their human counterparts . Finally, the macaques and other Old World anthropoid primates uniquely resemble women in reproductive function, as exemplified by similarities in ovarian hormone profiles, the presence of a menstrual cycle, and the occurrence of menopause .
An extensive series of studies, most conducted using socially housed female and male cynomolgus monkeys fed a diet relatively high in fat and cholesterol (designed to mimic typical consumption in industrialized countries) has resulted in seminal observations affecting design of future investigations with translation to human studies and evaluation of cardiovascular risk. The points below delineate these key findings.
Females typically develop less atherosclerosis than males (see [77, 164]). However, among socially housed monkeys this female 'protection' extends only to animals dominant within their social group; subordinate females exhibit a 'precocious acceleration' of atherosclerosis and are equivalent to males in coronary artery atherosclerosis extent.
The precocious atherosclerosis that characterizes subordinate females likely results from a subclinical stress-induced, reversible ovarian impairment that resembles functional hypothalamic amenorrhea/anovulation (FHA) observed in women [165, 166]. This hypothesis is supported by the observation that ovariectomy eliminates the protection of dominant females, rendering them equivalent to subordinate females and males in atherosclerosis extent [165, 167]. Treatment of subordinate animals with exogenous estrogen (oral contraceptives) inhibits the development of atherosclerosis [168, 169].
The trajectory of premenopausal atherosclerosis predicts postmenopausal atherosclerosis extent. In a two-part study reproductively intact monkeys consumed an atherogenic diet for 2 years, half were also treated with an oral contraceptive ; all animals were ovariectomized and continued to consume an atherogenic diet for 3 more years. Atherosclerosis extent at ovariectomy (as determined in an iliac artery biopsy) predicted the amount of atherosclerosis present at the end of the study [169, 170], irrespective of postmenopausal interventions.
The extensive atherosclerosis that accompanies ovariectomy can be substantially inhibited by treatment with exogenous estrogens (for example, conjugated equine estrogens or 17β estradiol), only if treatment begins immediately following ovariectomy in animals initially free of atherosclerosis ; treatment of animals with pre-existing atherosclerosis is ineffective in proportion to the amount of lesion present [172, 173].
The primary lesson from studies of female monkeys is that stress-induced premenopausal ovarian dysfunction, a common and subclinical condition, puts affected individuals on accelerated trajectory for atherosclerosis. Furthermore, the trajectory of atherosclerosis established in the premenopausal years appears to determine postmenopausal lesion extent, underscoring the importance of early events in the development of the postmenopausal disease burden.
Although studies in non-human primates have provided data for several decades of research and form the basis of studies in humans, the cost and difficulty of working with these animals make them prohibitive for many investigators in the future. In this regard, the US National Institutes of Health (NIH)-sponsored National Primate Research Center program and specialized colony resources (the Animal and Biological Materials Resources program) may provide investigators with access to animals and expertise to conduct studies with monkeys. Access to these resources is described at the website of the Division of Comparative Medicine (currently part of the National Center for Research Resources but soon to be moved into the Office of the NIH Director).