Sex as a Biological Variable in Atherosclerosis

Abstract

Atherosclerosis is a chronic inflammatory vascular disease and the predominant cause of heart attack and ischemic stroke. Despite the well-known sexual dimorphism in the incidence and complications of atherosclerosis, there are relatively limited data in the clinical and pre-clinical literature to rigorously address mechanisms underlying sex as a biological variable in atherosclerosis. In multiple histological and imaging studies, overall plaque burden and markers of inflammation appear to be greater in men than women and are predictive of cardiovascular events. However, while younger women are relatively protected from cardiovascular disease, by the seventh decade the incidence of myocardial infarction in women ultimately surpasses that of men, suggesting an interaction between sex and age. Most pre-clinical studies in animal atherosclerosis models do not examine both sexes, and even in those that do, well-powered direct statistical comparisons for sex as an independent variable remain rare. This manuscript reviews the available data. Overall, male animals appear to have more inflamed yet smaller plaques compared to female animals. Plaque inflammation is often used as a surrogate endpoint for plaque vulnerability in animals. The available data supports the notion that rather than plaque size, plaque inflammation may be more relevant in assessing sex-specific mechanisms since the findings correlate with the sex difference in ischemic events and mortality and thus may be more reflective of the human condition. Overall, the number of pre-clinical studies directly comparing plaque inflammation between the sexes is extremely limited relative to the vast literature exploring atherosclerosis mechanisms. Failure to include both sexes and to address age in mechanistic atherosclerosis studies are missed opportunities to uncover underlying sex-specific mechanisms. Understanding the mechanisms driving sex as a biological variable in atherosclerotic disease is critical to future precision medicine strategies to mitigate what is still the leading cause of death of men and women worldwide.

A. Introduction: Sex as a Biological Variable (SABV) in Cardiovascular Disease

Despite recent advances in our understanding of the mechanisms driving ischemic heart disease and decades of innovation in medical and interventional cardiovascular care, heart disease remains the primary cause of death in men and women of most racial and ethnic groups. Ischemic cardiovascular diseases (CVD), including myocardial infarction (MI) and stroke, account for approximately 650,000 deaths annually. Each year, more than 600,000 persons have a first MI, and 200,000 persons have a recurrent event. The observation that the pattern of CVD prevalence varies by sex has been known for centuries. In his 1772 description of angina pectoris, William Heberden wrote: “I have seen nearly a hundred people under this disorder [of the breast], of which there have been three women…all the rest were men.”¹ Similarly, in a lecture at Johns Hopkins in 1896, Sir William Osler cited this quote from Heberden, adding: “In my own series of forty cases of true angina, there was only one woman.”² Modern epidemiological data continue to show that younger women have a decreased tendency to develop cardiovascular disease and experience lower rates of MI relative to men but that women catch up in the seventh decade of life and surpass men by the ninth decade (Figure 1).³ In contrast, this pattern of disease onset is not recapitulated in stroke. Women have a greater prevalence of stroke until the seventh decade of life when men surpass women, but women again have the highest prevalence in the growing elderly population. Despite having a higher prevalence of stroke, men have similar age-adjusted death rates among Whites and higher death rates among Blacks, Hispanics, and Asian/Pacific Islanders.³ Interestingly, both Black and White women have a greater probability of recurrent stroke within the first 5 years after stroke compared to men. These demographic statistics suggest underlying biological variation between the sexes and highlight the need to include sex as an important component of investigation from animal models through clinical trials. Despite these longstanding observations, the reason for the sexual dimorphism in cardiovascular ischemic events, recurrence, and outcomes remains incompletely understood.

Figure 1. Changes in the difference in prevalence of cardiovascular disease between men and women in the US with age.

Adapted from Virani and colleagues³ from unpublished NHLBI tabulation using NHANES, 2013 to 2016.

Ischemic CVD is largely caused by atherosclerosis, a process in which lipid-laden plaques form in the vasculature. From decades of investigation into mechanisms of atherogenesis and plaque progression in multiple pre-clinical animal models, it has become clear that atherosclerosis is a diffuse, chronic inflammatory condition of the vasculature. In response to damage induced by cardiovascular risk factors (i.e. hypertension, diabetes, dyslipidemia, and smoking), the endothelial cells (EC) lining the vasculature become dysfunctional and lose their normal anti-inflammatory and anti-thrombotic properties. As a result, inflammatory cells infiltrate the vasculature and take up oxidized lipids, becoming foam cells that comprise the plaque core and are covered by a fibrous cap that stabilizes the plaque. Plaque inflammatory cells release factors that recruit additional immune cells and inhibit their clearance, thereby inducing a chronic inflammatory state and plaque progression. Larger plaques (>70% stenosis) that occlude blood flow chronically cause symptoms such as angina or claudication. However, plaque inflammatory cells also secrete matrix metalloproteinases that can degrade the fibrous cap and eventually result in plaque rupture, thrombosis, and acute ischemia – the cause of most MIs and strokes.⁴ The crucial role of inflammation in atherosclerosis revealed by pre-clinical research has led to recent trials of anti-inflammatory therapy for prevention of cardiovascular events.^5,6 The purpose of this review is to summarize recent knowledge about sex as a biological variable in atherosclerosis in humans and in animal models and to identify gaps in our understanding of how sex impacts the development and complications of atherosclerosis.

B. Sex Differences in Cardiovascular Risk Factors: Important, but insufficient to completely explain SABV in atherosclerosis.

Sex differences in atherosclerosis development or complications could be due to sexual dimorphism in major modifiable risk factors for MI and stroke identified by the Framingham study including cigarette smoking, dyslipidemia, hypertension, and diabetes. Large data sets have shown that the impact of these risk factors on the associated risk of MI is more potent in women compared to men with odds ratios of 1.5, 1.6, and 1.3 for hypertension, diabetes mellitus, and smoking, respectively,^7,8 suggesting sexual dimorphism in the vascular response to each risk factor and its impact on atherogenesis. Sex differences in the incidence of each risk factor are summarized below.

Diseases attributable to cigarette smoking account for about 5.7% of total US medical costs⁹ and 20% of all deaths. In the US, about 16% of men and 12% of women smoke.¹⁰ Globally, more than 80% of smokers are men.¹¹ Recently, the use of traditional cigarettes has been declining in the US, but e-cigarettes use has increased and has undefined long-term effects.

Current estimates suggest that approximately 12% of the adults in the US have high total cholesterol.¹⁰ The prevalence of high total cholesterol is significantly greater in women among Whites, similar in both sexes among Blacks and Asians, and lower in women among Hispanics.¹² In addition, the composition of the lipid profile varies between men and women. Men are more than twice as likely to have low high-density lipoprotein (HDL) than women in all age groups. Men have modestly higher levels of LDL compared to women.³ Response to treatment has been shown to vary by sex as well. In the SATURN trial, female sex was independently associated with greater atheroma regression with statin therapy when LDL levels were ≤70 mg/dL.¹³

For the past 15 years, the prevalence of hypertension in adults in the US has remained around 29%. Men have a modestly higher prevalence of hypertension than women (30.2% and 27.7%, respectively).¹⁴ The rate at which the incidence of hypertension increases with age also varies by sex. Women under 55 years have a substantially lower incidence of hypertension compared to age-matched men, but women over 75 years have a greater incidence of hypertension compared to age-matched men. Women have better rates of controlled hypertension than men at ages 60 years and under, whereas the control rate is similar and generally only about 50% among women and men at ages 60 years and above.¹⁴ This sex difference in hypertension prevalence may underlie the increased prevalence of stroke in older women compared with men.

Among adults, it is estimated that 9.7% of adults (26 million) have diagnosed diabetes mellitus and 4.3% remain undiagnosed.¹⁵ Men are estimated to have higher rates of both diagnosed and undiagnosed cases with a 30% excess compared to women. The prevalence of pre-diabetes (fasting glucose between 100–126 mg/dL or impaired glucose tolerance) is skewed similarly, affecting 44% of men and 31% of women.³

Thus, in general, men tend to have a worse risk factor profile than women, and this difference likely participates in the increased likelihood of earlier adverse events in men. However, the impact of most risk factors on enhanced cardiovascular risk is greater in women. Moreover, age substantially modifies the sexual dimorphism in risk factors and in CVD risk, supporting the importance of understanding the interaction between age and sex as biological variables in atherosclerosis.

C. SABV in Plaque Size and Characteristics: Does size really matter?

1. SABV in Plaque Size and Morphology in Humans

a. The impact of plaque size and morphology on outcomes

Atherosclerotic disease is initially asymptomatic as it develops. When an individual plaque becomes sufficiently large to limit the ability to increase blood flow to meet the demand of downstream tissues (usually >70% stenosis), symptoms of exertional ischemia (i.e. angina, claudication) may occur. Plaque rupture is a distinct phenomenon that causes most acute ischemic events (i.e. MI and stroke) and hence mediates most of the morbidity and mortality from atherosclerosis. Post-mortem pathological studies reveal that smaller plaques (30–40% stenosis) with more inflammatory cells, a larger lipid core, and a thinner fibrous cap are more likely to rupture leading to vascular occlusion and death.¹⁶ Thus, sex differences in the degree of stenosis of an individual plaque may contribute to exertional symptoms, whereas overall plaque burden (the number of small plaques), plaque inflammatory state, and unstable plaque morphology may contribute to acute MI and stroke risk.

b. Sex differences in plaque burden and degree of stenosis in humans

Non-invasive, ultrasound-based methods have allowed for evaluation of carotid artery intima-media thickness (IMT), a quantifiable parameter of early atherogenesis prior to the development of frank plaque. In children and adolescents, IMT thickness was higher in boys compared to age-matched girls.¹⁷ Similar findings were noted in a cohort of middle-aged adults.¹⁸ Notably, in the Young Finns study, the greater IMT in males was attenuated after adjustment for conventional cardiovascular risk factors.¹⁹ However, other studies suggest the sex difference may be independent of risk factors. In the ELSA-Brasil study, men had significantly higher IMT than women even when stratified by race and when restricted to low-risk individuals who lack conventional cardiovascular risk factors.²⁰ Similarly in the Tromsø study of males and females matched for age, BMI, blood pressure, diabetes and smoking, men had a greater proportion of individuals with plaques, mean number of plaques, and plaque area compared to women.²¹ Another factor to consider is the age distribution in these studies. In the Gutenberg-Heart Study, IMT was greater in men at 35 years but not at 75 years, suggesting a late catch-up phenomenon in women.²² In two studies, the sex difference in IMT remained statistically significant after adjusting for vessel lumen diameter.^19,20 This is important since artery size varies in association with body size, which is also linked with age and sex. Overall, the data suggests that subclinical atherogenesis may occur earlier in men than women. However, the measurements may be impacted by vascular size and the sex difference is modified by age and perhaps other risk factors.

Computed tomography (CT) angiography can also provide assessments of plaque burden and intra-plaque characteristics. Multiple population studies have demonstrated that men tend to have more atherosclerotic plaques than women by this method. In a Canadian referral population, men were more than twice as likely to have any plaque, at least one segment with ≥50% stenosis, and greater overall plaque burden.²³ Plank and colleagues used propensity matching to limit variation in baseline demographics but still demonstrated significantly more atherosclerotic plaques, more calcified plaques, higher rates of coronary artery disease, and a higher rate of major adverse cardiac events over 5.6 years of follow up in men compared with women.²³ In 21,132 patients enrolled in the CONFIRM registry, 50% of women had normal CT scans compared with 31% of men.²⁴ Interestingly, once atherosclerosis was present and normalized for segment involvement, the risk of major adverse cardiovascular events, including MI and stroke, was higher for men under 60 years but was greater for women over 60 years of age.

Due to its lack of movement, larger size than coronary arteries, and proximity to the body surface, the carotid artery provides additional opportunities for quantifying the flow velocity to determine the degree of stenosis. In the REFINE-Reykjavik study, men were 50% more likely to have carotid plaque after controlling for age, sex, risk factors, education level, and medications.²⁴ In an ultrasonographic evaluation of more than 1,600 patients from an atherosclerotic prevention clinic, women had a greater degree of stenosis than men, but men had a greater plaque area than women.²⁵ In this cohort, adverse cardiovascular events were well predicted by plaque area but not by degree of stenosis, suggesting that the extent of the vessel affected by plaque may be more predictive than its worst single location.

Thus overall, non-invasive imaging reveals that men develop plaques earlier and have a greater plaque burden than women, even after accounting for differences in risk factors. Plaque area, rather that degree of stenosis, predicts adverse ischemic events. This is consistent with the greater incidence of ischemic events in males, although this relationship changes with advancing age.

c. Sex differences in plaque morphology in humans

Recent advances in imaging have permitted the assessment of individual atherosclerotic plaque features associated with adverse events. In patients undergoing percutaneous coronary intervention, intravascular ultrasound (IVUS) can be used to evaluate not only the degree of stenosis but also the extent of necrotic core in the plaque based on its echolucency.²⁶ In patients with acute coronary syndromes, IVUS demonstrated that women had a similar number of culprit but fewer non-culprit lesions, fewer involved coronary arteries with lesions, lower frequency of plaque rupture, and smaller total necrotic core volume despite an older average age and greater comorbidities.²⁵ This sex difference in necrotic core volume was also observed in the Tromsø study.²¹

The advent of frequency-domain optical coherence tomography (OCT) has allowed for higher resolution imaging of plaque morphological features.²⁷ In a study of patients who underwent percutaneous coronary intervention,²⁸ non-culprit plaques in women with stable coronary artery disease had a smaller lipid arc, were less likely to contain cholesterol crystals, and had less lesion calcification – features associated with plaque stability. However, plaques in these women were also more likely to exhibit plaque erosion, a feature that can progress to atherothrombotic events. Fibrous cap thickness and incidence of thin-cap fibroatheroma did not vary by sex. However, when stratified by age in a different study, men less than 70 years of age had a higher incidence of thin-cap fibroatheroma, whereas women older than 70 years had a higher incidence.²⁶

These sex differences were also observed in plaques evaluated by magnetic resonance imaging. In patients who underwent MRI of a carotid artery with >50% stenosis, men were more likely to have a plaque with an intraplaque hemorrhage, thin or ruptured fibrous cap, or thrombus in the contralateral artery with <50% stenosis.²⁹ In histological analyses of endarterectomy or post-mortem tissue, men also exhibit a higher prevalence of morphological features associated with instability or actual rupture in carotid^30–34 and coronary^35,36 artery plaques compared to women. Interestingly, in one study, plaque hemorrhage was associated with a greater frequency of recurrent events in men but not women.³³

Thus, human studies have demonstrated that women tend to have decreased atherosclerotic plaque burden and fewer high-risk plaque features compared to men. This tends to remain true even in the setting of clinical events, both in active lesions and non-culprit lesions. As many of these studies have been performed in young and middle-aged subjects, whether plaque burden or morphology contribute to the mechanism of catch up with increased events in older women remains less thoroughly explored.

2. SABV in Plaque Size, Burden and Morphology in Animal Models of Atherosclerosis

a. Limited reporting of sex as a biological variable in pre-clinical research on atherosclerosis

Despite a vast literature exploring molecular mechanisms and novel drug targets for atherosclerosis, data examining sex differences are relatively limited. Ramirez and Hibbert analyzed 771 pre-clinical articles on atherosclerosis and other vascular diseases published between 2006 and 2016 in leading American Heart Association journals.³⁷ Of those, 18.8% did not report the sex of animals studied. Of articles that specified the sex of the animals, 75.8% studied exclusively one sex; 55.4% and 20.4% studied only males or females, respectively. Thus, less than 25% of studies during this 10-year period studied both males and females. This proportion is similar to recent reports by Atherosclerosis, Thrombosis, and Vascular Biology editors reviewing manuscripts published in the years 2017 and 2018 from their journal, stating that only 21% and 28% of studies included both sexes, respectively.^38,39 Moreover, of the studies that include both sexes, fewer than half directly compare males and females with the appropriate statistical analysis to consider sex as an independent variable or the interaction of sex with a treatment or genotype. Thus, the pre-clinical data investigating sex as a biological variable in atherosclerosis lags substantially behind the rest of the field. In this section, we review the available data from pre-clinical animal atherosclerosis models (see Tables 1–4) and summarize what is known about sex as a biological variable in atherosclerosis. Due to the aforementioned limitations in the literature, for studies that reported data for both sexes but for which a direct statistical comparison was not performed or stated in the text, a “sex difference” is noted in the table if there was a difference in mean plaque size (or other relevant measurement) with non-overlapping error bars. Studies that directly compare males and females are indicated in Tables 1–4 with an asterisk. Overall, among the approximately 25% of studies that include both males and females, we found that only 44.8% rigorously tested sex as a biological variable in animal atherosclerosis models by using an appropriate statistical test.

Table 1.

Sex difference in atherosclerosis in non-murine animal models

Duration of diet intervention	Age at sacrifice	Vascular Region	Analysis	Sex Difference	Ref.
Non-human primates
69 wk	N/A	Coronary, Iliac	H&E	M > F*	40
Pig
Aortic Arch
26 wk	35–39 wk	Aortic arch	ORO	F > M	41
26 wk	56–78 wk	Aortic arch	EVG	F > M*	42
Aorta
13 wk	26 wk	Thoracic aorta	H&E/Stary	F > M	43
26 wk	35–39 wk	Thoracic aorta	ORO	F > M	41
26 wk	39 wk	Thoracic aorta	H&E	M = F	44
26 wk	48 wk	Abdominal aorta	VVG/H&E	M = F	45
26 wk	48–52 wk	Abdominal aorta	Sudan IV	M = F	46
43–48 wk	49–54 wk	Aorta	H&E	M = F	47
17 wk	104 wk	Aorta	Trichrome	M = F	48
78 wk	N/A	Abdominal aorta	Visual grade	M = F*	49
Coronary
26 wk	35–39 wk	LAD, RCA	H&E	F > M	41
26 wk	48 wk	Coronary	VVG/H&E	M = F	45
26 wk	48–52 wk	LAD, LCX	Movat’s	M = F	46
43–48 wk	49–54 wk	LAD	H&E	M = F*	47
78 wk	N/A	Coronary	Visual grade	F > M*	49
95 wk	165 wk	Coronary	Sudan IV	F > M	50
Miscellaneous
13 wk	26 wk	Iliac, cerebral	H&E/Stary	M > F	43
43–48 wk	49–54 wk	Iliac	H&E	M = F*	47
13 wk	26 wk	Common carotid	H&E/Stary	F > M	43
Rabbit
12 wk	24 wk	Aortic arch	H&E/EVG	M = F	51
14 wk	26–32 wk	Ascending aorta	ORO	M = F	52
12 wk	27 wk	Aortic arch	EVG	M = F*	53
15 wk	25–27 wk	Aortic arch	Sudan IV	M = F*	54
16 wk	N/A	Aortic arch	Sudan IV	M = F*	55
18 wk	N/A	Aortic arch	Sudan IV	M = F	56
78 wk	90 wk	Ascending aorta	ORO	M = F	57
14 wk	26–32 wk	Descending aorta	ORO	M > F	52
15 wk	25–27 wk	Abdominal aorta	Sudan IV	M = F*	54
16 wk	N/A	Thoracic aorta	Sudan IV	M > F*	55
10 wk	28 wk	Thoracic aorta	EVG	M > F*	58
15 wk	33 wk	Thoracic aorta	EVG	M = F*	58
Wean	52 wk	Abdominal aorta	EVG	F > M*	59
48 wk	56 wk	Thoracic aorta	EVG	F > M*	60
48 wk	56 wk	Abdominal aorta	EVG	F > M*	60
78 wk	90 wk	Thoracic aorta	ORO	F > M*	57
78 wk	90 wk	Abdominal aorta	ORO	F > M*	57
NZW Rabbits (HFD)
18 wk	N/A	Coronary	Sudan IV	M = F	61
48 wk	56 wk	Coronary	EVG	F > M*	60
WHHL Rabbits (NC)
--	26 wk	Coronary	EVG	F > M*	62
--	52 wk	Coronary	EVG	M = F*	62
--	104 wk	Coronary	EVG	M = F*	62
--	43–73 wk	Coronary	EVG	F > M*	63
--	74–99 wk	Coronary	EVG	M = F*	63
--	100–121 wk	Coronary	EVG	M = F*	63
--	48–150 wk	Coronary	EVG	M = F*	64
--	156 wk	Coronary	N/A	M = F*	65

Some ages were estimated based on starting weight reported in the articles and the growth chart from Lone Star Swine (pigs) and Charles River (rabbits). N/A = age at sacrifice not explicitly mentioned; ORO = Oil Red O; EVG = Elastin Verhoeff-Van Gieson; H&E = Hematoxylin and eosin; wk = weeks; F=female; M=male.

^*=direct statistical comparison between the sexes.

Table 4.

Sex difference in plaque area of LDLR^−/− mice by age and by site

Duration of diet intervention	Age at sacrifice	Vascular Region	Analysis	Sex Difference	Ref.
Normal chow
--	22 wk	Whole aorta	Cholesterol ester	F > M	106
--	28 wk	Whole aorta	Sudan IV	F > M	115
--	33 wk F / 52 wk M	Aorta cross-section	Sudan IV	F > M	115
Atherogenic diet
8 wk	12 wk	Aortic root	ORO	M = F*	116
10 wk	16 wk	Aortic root	H&E	F > M	117
Weaning	16 wk	Aortic root	ORO	F > M	90
14 wk	18 wk	Aortic root	ORO	F > M*	116
12 wk	20 wk	Aortic root	ORO	F > M*†	118
12 wk	20–22 wk	Aortic root	ORO	M = F*	119
12 wk	22 wk	Aortic root	ORO	M = F	120
16 wk	22 wk	Aortic root	ORO	F > M	121
18 wk	22 wk	Aortic root	ORO	F > M*	116
13 wk	23 wk	Aortic root	Movat	M = F	122
16 wk	24 wk	Aortic root	H&E	M = F*	123
9 wk	17 wk	Aortic root	Movat	M = F	124
16 wk	26 wk	Aortic root	H&E	M = F	125
17 wk	26 wk	Aortic root	Movat	F > M	124
22 wk	26 wk	Aortic root	ORO	F > M*	116
16 wk	26–28 wk	Aortic root	ORO	M = F*	126
24 wk	32 wk	Aortic root	Sudan IV	M = F	127
12 wk	15 wk	Whole aorta	Sudan IV	M = F	128
12 wk	16 wk	Whole aorta	ORO	M = F	129
12 wk	20–22 wk	Whole aorta	ORO	F > M*	130
12 wk	20–22 wk	Whole aorta	ORO	M = F	131
14 wk	22 wk	Whole aorta	Sudan IV	M = F*	132
13 wk	23 wk	Whole aorta	ORO	M = F	122
12 wk	20–24 wk	Whole aorta	ORO	M = F*	133
24 wk	20–24 wk	Whole aorta	ORO	M = F*	133
26 wk	52 wk	Whole aorta	Sudan IV	M > F*	114
Weaning	16 wk	Brachiocephalic	ORO	F > M	90
12 wk	20–22 wk	Aortic arch	ORO	M = F*	119
20 wk	28 wk	Descending aorta	Sudan IV	M > F	134

ORO = Oil Red O; H&E = Hematoxylin and eosin; wk = weeks; F=female; M=male.

^*=direct statistical comparison between the sexes.

^†=Utilized AAV-PCSK9 to create functional LDLR null mouse.

b. Sex as a biological variable in plaque size and the impact of anatomic location and age

i. SABV in atherosclerosis in non-human primates

Cynomolgus macaque monkeys have been the non-human primate of choice to study sex differences and the effects of stress in atherosclerosis. When treated with an atherogenic diet, male monkeys exhibit a greater extent of coronary and iliac artery atherosclerosis compared to females.⁴⁰ Atherosclerotic lesion size in the coronary, carotid, and ilio-femoral arteries was increased 2–10 fold in female monkeys that were ovariectomized compared to sham surgery controls,⁶⁶ and this was attenuated by estrogen replacement,⁶⁷ supporting a potential role for estrogen in the sex difference. Hormone contraceptive treatment⁶⁸ and pregnancy⁶⁹ have also been shown to attenuate atherosclerotic lesion size in monkeys. Interestingly, psychosocial stress by means of repeated social reorganization or subordination has been shown to exacerbate coronary artery atherosclerosis in both male and female monkeys.^70–72 Subordinate female monkeys exhibit hypercortisolemia and ovarian dysfunction, resulting in low circulating estrogen levels. These hormonal changes may contribute to the exacerbation of lesion size to levels comparable to that of males.

ii. SABV in pig models of atherosclerosis

Several breeds of minipigs (i.e. Gottingen, Hormel, Yucatan) are commonly used to study atherosclerosis because of their relatively smaller size and similarity to human coronary anatomy and plaque morphology when fed an atherogenic diet. Transgenic minipig models of familial hypercholesterolemia have also been recently generated.^45,47 In contrast to humans, female minipigs are consistently reported to have larger or equally sized plaques in the coronary artery compared to males. In the aortic arch, female minipigs have larger plaques at both 9 and 18 months of age. In the whole aorta, female minipigs have greater atherosclerotic burden compared to males until about 9 months old, when plaque burden appears to be equal between sexes. Multiple studies have posited that a potential explanation for this sex difference is the fact that female minipigs have worse diet-induced hyperlipidemia compared to males.^42,50,73 One study also noted that male pigs voluntarily exercise more than females.⁵⁰

iii. SABV in rabbit models of atherosclerosis

The two most common rabbit models of atherosclerosis are the New Zealand White (NZW) rabbit²⁰⁹ and the Watanabe Heritable Hyperlipidemic (WHHL) rabbit.⁷⁴ The NZW rabbit develops lesions within 8 to 12 weeks when fed a high cholesterol diet. The WHHL inbred strain spontaneously develops lesions (without an atherogenic diet) due to LDL receptor deficiency. In the aortic arch or ascending aorta of NZW rabbits, there is no reported difference in plaque size between male and female rabbits treated with a high cholesterol diet for 12–18 weeks (Table 1). In the abdominal or descending aorta, male NZW rabbits have greater lesion burden in earlier atherosclerosis (<18 weeks atherogenic diet), whereas female NZW rabbits have greater lesion burden in the descending aorta in later atherosclerosis. In the coronary vasculature, young female WWHL rabbits appear to have larger coronary artery lesions compared to young males, but this sex difference is abolished with age. The human relevance of sex differences in atherosclerosis in rabbit models is complicated by their dissimilar coronary anatomy as well as the fact that ovulation in female rabbits is coitally induced in the absence of a regular menstrual cycle.⁷⁵

iv. SABV in mouse models of atherosclerosis

The mouse is the most commonly used animal model to study atherosclerosis due to its small size, rapid breeding, lower cost of maintenance, and ease of genetic manipulation. Although mice are relatively resistant to spontaneous atherogenesis, there are variations in susceptibility to atherosclerosis among different strains.²⁰⁹ Thus, the more susceptible C57BL/6 background is the most commonly used in atherosclerosis research. The two most common mouse models of atherosclerosis are the apolipoprotein E-null (ApoE^−/−)^76,77 and low-density lipoprotein receptor-null (LDLR^−/−) mice,⁷⁸ both of which develop atherosclerotic plaques when fed an atherogenic diet with a morphology that is similar to human pathologic specimens. ApoE^−/− mice spontaneously develop hyperlipidemia and atherosclerosis when fed with normal chow but exhibit faster plaque progression when treated with high fat diet. Over 10,000 manuscripts have been published using these murine atherosclerosis models in the 25 years since they were developed, and plaque size or burden are the most commonly reported outcomes in mouse models. As mice do not develop spontaneous plaque rupture, surrogate endpoints of plaque instability are often reported, including lipid content, necrotic core size, inflammation, and fibrous cap thickness.

Atherosclerosis burden is typically measured in two ways: 1) plaque size is measured histologically in aortic root cross sections at the level where all three leaflets of the aortic valve can be visualized; or 2) plaque burden is measured in the whole aorta by en face staining (i.e. Oil Red O, Sudan IV) to quantify the percent of the aortic area covered with lipid-containing plaque. Tables 2 and 3 list studies that report data from both male and female ApoE^−/− mice and summarizes sex differences in atherosclerotic plaque size in the aortic root and burden in the whole aorta, respectively. The data are listed in order of increasing duration of diet exposure and hence also increasing age at the time of sacrifice in order to observe whether sex as a biological variable is modified by the stage of plaque or age of the mice. Overall, in the aortic root of ApoE^−/− mice, females are consistently reported to have larger plaques than their male littermates through 6 months of age on either normal chow or atherogenic diet. In 8+ month old ApoE^−/− mice on normal chow, males were reported to have equal or larger plaques in three studies (Table 2). When en face staining of the whole aorta is assessed in ApoE^−/− mice, data from normal chow fed mice reveals that in more studies, female mice have greater burden compared to male mice. However, in some studies, there is no sex difference and several papers report that males have greater plaque burden along the whole aorta at ages ranging from 3–8 months, providing no consistency to the changes in the sex difference with age. With exposure to an atherogenic diet, male and female ApoE^−/− mice are reported to have equal plaque burden along the whole aorta at less than 22 weeks old, but the few studies looking at late atherosclerosis in older mice on a longer duration of atherogenic diet show greater plaque burden in males (Table 3).

Table 2.

Sex difference in plaque area in the aortic root of ApoE^−/− mice by age

Duration of diet intervention	Age at sacrifice	Analysis	Sex Difference	Reference
Normal Chow
--	16 wk	ORO	F > M*	79
--	16 wk	ORO	F > M	80
--	16 wk	EVG	F > M	81
--	20 wk	ORO	F > M	82
--	22 wk	ORO	F > M	83
--	28 wk	ORO	F > M*	84
--	32–34 wk	EVG	M = F*	85
--	48 wk	ORO	M > F*	79
--	66 wk	ORO	M = F	86
Atherogenic diet
8 wk	14 wk	Sudan IV	F > M	87
6 wk	16 wk	Masson	F > M	88
10 wk	16 wk	EVG	F > M	81
12 wk	16 wk	ORO	F > M*	89
12 wk	16 wk	ORO	F > M	90
12 wk	17 wk	EVG	F > M*	91
10 wk	18 wk	H&E	F > M	92
10 wk	16–18 wk	Sudan IV	F > M	93
16 wk	20 wk	ORO	F > M	90
12 wk	20 wk	H&E	F > M	94
12 wk	20 wk	ORO	F > M*	95
14 wk	20 wk	EVG	M = F	96
10 wk	21 wk	Toluidine	F > M	97
13 wk	22 wk	H&E	F > M*	98
16 wk	22 wk	Sudan IV	F > M	87
12 wk	24 wk	ORO	F > M	99
16 wk	24 wk	ORO	F > M*	100
16 wk	24 wk	ORO	F > M*	101
18 wk	26 wk	Orcein	F > M	102

ORO = Oil Red O; EVG = Elastin Verhoeff-Van Gieson; H&E = Hematoxylin and eosin; wk = weeks; F=female; M=male;

^*=direct statistical comparison between the sexes.

Table 3.

Sex difference in plaque burden along the whole aorta of ApoE^−/− mice by age

Duration of diet intervention	Age at sacrifice	Analysis	Sex Difference	Reference
Normal chow
--	4, 9 wk	ORO	M = F*	103
--	13, 17 wk	ORO	M > F*	103
--	16 wk	ORO	F > M	81
--	16 wk	ORO	M = F	80
--	17 wk	Sudan IV	M = F*	104
--	22 wk	ORO	M > F*	103
--	22 wk	ORO	F > M	105
--	22 wk	Cholesterol ester	F > M	106
--	22 wk	Sudan IV	M = F	83
--	26 wk	ORO	M = F*	103
--	28 wk	ORO	F > M	107
--	30 wk	Sudan IV	M = F*	108
--	30 wk	ORO	F > M	105
--	32–34 wk	ORO	M > F*	85
--	35 wk	ORO	M = F*	103
--	42 wk	Sudan IV	F > M*	108
--	42 wk	ORO	F > M	109
--	52 wk	Sudan IV	F > M*	108
--	66 wk	ORO	M = F	86
Atherogenic diet
8 wk	14 wk	Sudan IV	M = F	87
10 wk	16 wk	ORO	M = F	81
12 wk	16 wk	ORO	M = F	89
12 wk	20 wk	ORO	M = F	110
12 wk	20 wk	Sudan IV	M = F	111
13 wk	22 wk	Sudan IV	F > M*	98
16 wk	22 wk	Sudan IV	M = F	87
16 wk	26 wk	Light microscopy	M > F*	112
17 wk	26 wk	ORO	M = F*	113
26 wk	52 wk	Sudan IV	M > F*	114

ORO = Oil Red O; wk = weeks; F=female; M=male.

^*=direct statistical comparison between the sexes.

Since LDLR^−/− mice develop plaques predominantly when exposed to atherogenic diet, most studies of this model use such a diet. On either normal chow or atherogenic diet, female LDLR^−/− mice are consistently reported to have greater or equal plaque size in the aortic root as well as plaque burden in the whole aorta compared to their male counterparts through 6 months of age (Table 4). One study found that male LDLR^−/− mice have more plaque in the descending aorta after 5 months of an atherogenic diet, and one long-term study reported that 1-year old male LDLR^−/− mice had larger lesions along the whole aorta compared to female LDLR^−/− mice treated with an atherogenic diet for 6 months.

Thus, in both ApoE^−/− and LDLR^−/− mice, females have larger plaques, a factor that may have contributed to the limited use of these models to compare sex differences. The exceptions to this general pattern include a few reports of greater plaque burden in the whole aorta of older male mice compared to age-matched female mice treated with a longer duration of an atherogenic diet. These mouse models share other limitations in their relevance to human atherosclerotic disease. On an atherogenic diet, both models develop serum cholesterol concentrations in excess of 1,000 mg/dL, levels which are not observed in humans with the exception of rare cases of familial hypercholesterolemia. Mice primarily develop plaques in the aortic root, aortic arch, and brachiocephalic (or innominate) artery, but even aged mice develop minimal lesions in the coronary and carotid arteries, the most clinically significant sites of atherosclerotic disease in humans.¹³⁵ Atherosclerotic plaques in mice are also resistant to rupture, even in old age. Genetic and mechanical plaque rupture models have recently been developed (reviewed in¹³⁶), but their results are highly variable, and the relevance of mechanical rupture remains to be clarified.

In summary, across these species, with the exception of non-human primates, younger female animals tend to have a greater atherosclerotic lesion size and overall burden compared to their male counterparts. With aging, male ApoE^−/− and LDLR^−/− mice, WHHL rabbits, and minipigs have larger or equally sized plaques, most notably along the thoracic and abdominal aorta in mice. Only in the lower aorta of NZW rabbits fed a high cholesterol diet was plaque size larger in males at a younger age and then larger in females at an older age.

c. Sex as a biological variable in plaque morphology in animal models

Plaque fibrosis is typically quantified by Masson’s trichrome, Movat’s, or picrosirius red staining for collagen and is viewed as a positive indicator of plaque stability. The necrotic core is typically quantified by the acellular, non-fibrotic plaque area void of hematoxylin or equivalent staining and is associated with plaque vulnerability. Three studies that directly compared male and female LDLR-deficient mice demonstrated no difference in the percent of aortic root plaque with collagen staining or composed of necrotic core after 16 weeks of an atherogenic diet.^121,123 Data in the ApoE^−/− model also suggests no sex difference in collagen staining in the plaques of the aortic root⁸⁸ and brachiocephalic artery¹³⁷ of ApoE^−/− mice, but males and females were not directly compared.

3. Conclusions from comparison of plaque size between animal models and humans

In addition to having comparable physiology and anatomy, whether the sex differences in a given animal model of atherosclerosis parallel that of humans is an important determinant of whether the animal is an appropriate model for studying sexually dimorphic mechanisms of atherogenesis and responses to treatment. In humans, plaque size and burden tend to be greater in males, particularly at a younger age, but females tend to catch up at a later age. Here we summarized the data on plaque size and burden as a measure of atherosclerosis in animal models, as this is the most commonly studied parameter and because ischemic events cannot be studied in most animal models. The primary observation is that relative to the total number of preclinical atherosclerosis studies, very few include both male and female animals and even fewer analyze the data by direct statistical comparison of the two sexes. Indeed, only about 25% of studies in the past decade include both sexes, and of all the studies listed in the tables in which males and females were analyzed separately, less than half (44.8%) statistically compared the two sexes.

From the data available, we conclude that sex differences in plaque size depends on a number of variables in the experimental design – diet (normal chow vs. atherogenic diet); timing (age when diet was started, duration of diet, age at endpoint); and analytic method (vascular bed or location tested, en face surface area vs. cross-sectional area, proportion of total area vs. absolute area). From this review of the literature, age may play an important role in sex differences in atherosclerosis in animals and in humans but with a different pattern depending on the species. For example, both female ApoE^−/− and LDLR^−/− mice have larger plaques in the aortic root regardless of diet until 6 months of age, at which point male ApoE^−/− and LDLR^−/− mice have larger or equally sized aortic root plaques. However, in humans, males have greater plaque burden, but females catch up later in life. More studies on sex differences in advanced atherosclerosis in older animal models are needed to confirm whether this age-dependent change in plaque size by sex is consistently reproducible. If so, perhaps older mice or more advanced atherosclerosis models would be more relevant to studying mechanisms of sex differences in plaque burden that may be applicable in humans.

While studies in mice focus on aortic atherosclerosis, development of coronary and carotid artery atherosclerosis in larger animal models allows for the study of sex differences in these clinically relevant vascular regions. Regardless of age, female NZW rabbits and minipigs have larger coronary plaques. Interestingly, younger female WHHL rabbits have larger coronary plaques, but older male and female WHHL rabbits have equally sized plaques. Male macaques seem to have larger coronary plaques, but there are too few studies to infer how this may vary with age in non-human primates.

In humans, there are sex differences in cardiovascular risk factors that impact atherosclerosis although the data suggest that this is insufficient to completely explain the sex difference. Similarly, in animal models, a number of studies have demonstrated that lipoprotein profiles do not fully explain sex differences in plaque size. In some cases, young female minipigs, rabbits, and mice developed larger diet-induced atherosclerotic plaques than their male counterparts despite having similar^{81,95,108,118,124,130} or even lower¹²¹ cholesterol levels. Despite older male ApoE^−/− and LDLR^−/− mice having larger plaques than age-matched females, their total cholesterol levels were not different.¹¹⁴ Even in young NZW rabbits in which plaque size in the thoracic and abdominal aorta is larger in males, age-matched females had higher total cholesterol levels^52,55 and no difference in LDL levels.⁵⁸

The finding that females are protected from acute MI relative to males only until older age (after the hormonal changes that take place at the time of menopause) has prompted substantial research on the impact of sex steroid hormones on atherosclerosis. Pre-clinical studies in this area generally study only a single sex and compare the burden of atherosclerosis in gonad-intact compared to gonadectomized animals in the presence or absence of hormone replacement or hormone receptor manipulation. In general, estrogen has been found to be atheroprotective, while the impact of androgens is more controversial. This literature has been reviewed elsewhere^138–140 and is beyond the scope of this review as the studies exploring the role of sex hormones rarely compare both sexes. Of note, neither ovariectomy nor subsequent estrogen supplementation modulated total cholesterol levels in female NZW rabbits compared to sham surgery controls or to ovariectomized rabbits treated with placebo, respectively,⁵¹ suggesting that the atheroprotective effect of estrogen is not mediated by simply inducing favorable lipoprotein profiles. A meta-analysis of Mendelian randomization studies found that higher endogenous testosterone levels are associated with elevated LDL and lower HDL.¹⁴¹ Testosterone treatment is associated with inconsistent results ranging from increased, decreased, or no change in HDL levels,^141,142 although there is evidence of an increased risk of cardiovascular-related events in men on testosterone.¹⁴¹

D. Sex as a Biological Variable in Vascular Inflammation in Atherosclerosis:

1. Sex Differences in Inflammation in Atherosclerosis in Humans

a. The role of inflammation in atherosclerosis

Although larger plaque size may be associated with symptoms such as angina and claudication in humans, plaque rupture and thrombosis mediate the most significant complications of atherosclerosis by inducing MI, stroke, and associated mortality. In humans, such vulnerable plaques have been found to have a greater burden of inflammation.^143,144 Specifically, there are more macrophages¹⁴⁵ and T cells¹⁴⁶ in the lipid-rich core and at the site of rupture of vulnerable versus stable plaques in humans. Since plaques do not rupture in most pre-clinical atherosclerosis models, much research has focused on plaque inflammation as a surrogate endpoint of plaque instability.

Plaque inflammation develops as circulating monocytes transmigrate through the damaged endothelium of the vessel into the subintima where they differentiate into macrophages, scavenge oxidized LDL, and become lipid-laden foam cells.¹⁴⁷ Foam cells in the plaque secrete matrix metalloproteinases that degrade the fibrous cap overlying the plaque.¹⁴⁸ Efferocytosis is also inhibited, leading to accumulation of a necrotic core. Furthermore, activated macrophages secrete chemokines that recruit other immune cells (i.e. T cells) that secrete their own cytokines, thereby maintaining a pro-inflammatory macrophage phenotype and propagating chronic vascular inflammation. Altogether, this leads to a vulnerable morphology in plaques that are more prone to rupture and cause ischemic events.^35,143,149

There are known sex differences in immune function with aging (reviewed in¹⁵⁰). Therefore, sexual dimorphism in the immune response to hyperlipidemia and other cardiovascular risk factors may contribute to sex differences in the plaque phenotype and incidence of ischemic events. In this section we review what is known about sex differences in inflammation in atherosclerosis in humans, sex as a biological variable in vascular inflammation in mouse atherosclerosis models, and potential mechanisms for this dimorphism in vascular inflammation.

b. Sex differences in systemic inflammation in atherosclerosis in humans

In humans, systemic inflammation can be interrogated through the measurement of blood-based biomarkers and plaque inflammation can be assessed through imaging of the arterial wall. C-reactive protein (CRP) is a marker of systemic inflammation and predictor of cardiovascular disease events.^151–153 In an evaluation of the NHANES survey, the median CRP level (2.1 mg/L) was the same for both men and women, although the mean for women (0.55±0.91) was higher than that for men (0.41±0.64).²⁰⁹ Despite having higher CRP levels in this study, women had a lower 10-year probability of coronary heart disease in all CRP categories. These results have been supported by the Dallas Heart Study, a cross-sectional analysis of over 2,700 subjects aged 30–65 years in which women had higher CRP levels than men (median 3.3 vs. 1.8 mg/L, respectively).¹⁵⁴ In the Great Smoky Mountains Study, children aged 9–13 years were followed until age 21. Changes in CRP began to vary between boys and girls around age 15, when the rate of CRP elevation accelerated in girls as opposed to a smaller, linear increase in boys.¹⁵⁵ Interestingly, the use of oral hormone replacement therapy (HRT) has also been associated with higher CRP levels.^156–160 In fact, post-menopausal women have similar levels of CRP as men, but any kind of HRT (progesterone with or without estrogen) increases CRP.¹⁶¹ While one potential explanation is that estrogen increases CRP by effecting hepatic protein synthesis, other markers of systemic inflammation (including IL-6, IL-1β, and TNFα) are also increased by estrogen.^156–160 Alternatively, female sex hormones could drive systemic inflammation, while other factors related to female sex may attenuate the vascular impact of enhanced systemic inflammation in women.

c. Sex differences in plaque inflammation in humans

In a large study of 763 carotid endarterectomy specimens, male sex was significantly associated with increased plaque inflammatory infiltrates by logistic regression.³² Indeed, multiple well-powered studies of endarterectomy specimens have shown that plaques from men have more CD68 staining (an inflammatory marker associated with macrophages) compared to plaques from age-matched women.^30,31,162 Interestingly, this was still true for both asymptomatic patients and patients with more than 90% stenosis of the carotid artery when each group was analyzed separately.³¹ Plaque macrophages in men also express higher levels of CD86, a co-stimulatory molecule that facilitates communication with T cells and propagating inflammation.¹⁶² There does not appear to be a sex difference in the overall abundance of T cells in plaques based on CD3 staining of carotid endarterectomy specimens.³¹

¹⁸F-fluorodeoxyglucose (FDG) uptake is a biomarker of plaque inflammation-related metabolic activity measured by positive emission tomography^163,164 and has been shown to be associated with high-risk morphological features such as CD45 and CD68 staining (total leukocytes and macrophages, respectively) as well as symptom severity.^165,166,167 In a study of 61 patients with transient ischemic attack or mild stroke, FDG uptake in carotid artery lesions was significantly greater in men compared to women,¹⁶⁸ although this sex difference was not seen in all FDG studies in patients with symptomatic carotid atherosclerosis.^164,167 In 521 asymptomatic patients who underwent carotid artery stenosis screening, men were more likely to have carotid and femoral artery FDG uptake compared with women.¹⁶⁹ Of the subjects with both carotid and femoral artery activity, men were more than twice as likely to manifest carotid artery stenosis. Interestingly, some of the variation may be related to the frequency of metabolic syndrome in men compared with women. Two studies have linked FDG uptake to metabolic syndrome, but Lee and colleagues have shown that carotid artery FDG uptake is higher in men compared to women both with and without the metabolic syndrome.^170,171

2. Sex Differences in Plaque Inflammation in Atherosclerosis in Animal Models

a. Sex differences in plaque inflammatory cells in mouse atherosclerosis

Even fewer studies have directly compared plaque inflammation in pre-clinical atherosclerosis models (see Tables 5–6). Indeed, only one study¹¹⁸ directly compared plaque inflammatory cell number by flow cytometry in male and female littermates in a model in which hyperlipidemia was induced by an adeno-associated virus expressing a gain-of-function mutant form of human PCSK9 (AAV-PCSK9), which degrades the LDL receptor.^172,173 In this study, Moss et al. demonstrated that 20-week old male C57BL/6 mice that had been fed an atherogenic diet for 12 weeks had significantly more CD45⁺ leukocytes in their atherosclerotic aortic arches compared to their female counterparts despite having similar levels of hyperlipidemia and smaller plaques in the aortic root. Males had increased number of both myeloid cells and T cells in the aortic arch, leukocytes from both innate and adaptive immune systems, respectively.

Table 5.

Sex differences in macrophage infiltration in atherosclerotic plaques

Diet	Duration of diet intervention	Age at sacrifice	Analysis	Sex Difference	Ref.
Rabbit
Atherogenic	16 wk	N/A	TA (IHC): RAM11	M = F*	55
Atherogenic	18 wk	20 wk	AA (IHC): RAM11	F > M	56
ApoE−/− mice
Atherogenic	6 wk	16 wk	AR (IHC): Gal-3	M = F	88
Atherogenic	12 wk	16 wk	AR (IHC): CD68	F > M	89
Atherogenic	12 wk	18 wk	BCA (IHC): Mac-3	M = F	137
Atherogenic	12 wk	20 wk	BCA (IHC): F4/80	M = F*	175
Atherogenic	13 wk	22 wk	AR (IHC): Mac-3	M = F	98
Atherogenic	16 wk	24 wk	AR (IHC): MOMA-2	F > M	100
Atherogenic	17 wk	26 wk	AR (IHC): MOMA-2	M > F*	113
Normal chow	--	22 wk	AR (IHC): CD68	F > M	83
Normal chow	--	26 wk	AA (IHC): Mac-3	M = F*	174
Normal chow	--	32–34 wk	AR (IHC): Gal-3	M > F*	85
LDLR−/− mice or AAV-hPCSK9
Atherogenic	9 wk	17 mo	AR (IHC): Mac-3	M = F	124
Atherogenic	12 wk	20 wk	AA (FC): CD11b	M > F*	118
Atherogenic	10 wk	23 wk	AR (IHC): MOMA-2	M = F	122

AA = aortic arch; AR = aortic root; BCA = brachiocephalic artery; TA = thoracic aorta; FC = flow cytometry; IHC = immunohistochemistry; wk = weeks; F=female; M=male.

^*=direct statistical comparison between the sexes.

Table 6.

Sex differences in T cell infiltration in atherosclerotic plaques

Diet	Duration of diet intervention	Age at sacrifice	Analysis	Sex Difference	Ref.
ApoE−/− mice
Atherogenic	13 wk	22 wk	AR (IHC): CD3	M = F	98
Atherogenic	16 wk	24 wk	AR (IHC): CD4 & CD8	M > F	100
Normal chow	--	26 wk	AA (IHC): CD3	M = F*	174
LDLR−/− mice or AAV-hPCSK9
Atherogenic	12 wk	20 wk	AA (FC): CD3	M > F*	118

AA = aortic arch; AR = aortic root; FC = flow cytometry; IHC = immunohistochemistry; wk = weeks; F=female; M=male.

^*=direct statistical comparison between the sexes.

b. Sex differences in innate immunity in atherosclerosis

The pre-clinical studies of atherosclerosis that include both males and females suggest that plaques are more inflamed in males, with the exception of young ApoE^−/− mice. Specifically, in ApoE^−/− mice fed an atherogenic diet, female mice exhibited similar or more macrophage accumulation compared to male mice by immunohistochemistry (IHC) up to 22 weeks old (Table 5). However, in ApoE^−/− mice over 26 weeks old, plaque inflammation is generally greater in males.^100,113 Similarly, when ApoE^−/− mice were fed normal chow, female mice exhibited more macrophage accumulation at 22 weeks.⁸³ There was no sex difference in macrophage accumulation in the aortic arch at 26 weeks.¹⁷⁴ Male mice exhibited more macrophage accumulation at 32–34 weeks.⁸⁵

In LDLR^−/− mice or the AAV-PCSK9 model that induces LDLR deficiency, macrophages were the same or greater in plaques from male mice (Table 6). Specifically, in LDLR^−/− mice fed an atherogenic diet, males and females did not display any difference in macrophage accumulation in the aortic root at 17 or 23 weeks old by IHC.^122,124 However, male mice treated with AAV-PCSK9 and an atherogenic diet had more CD11b⁺ myeloid cells in aortic arch plaques by flow cytometry at 21 weeks old compared to their female littermates. This discrepancy is unlikely due to the difference in the models in inducing hypercholesterolemia¹⁷² or plaque area.¹⁷⁶ The plasma cholesterol of the LDLR^−/− mice did not differ by sex, and although the plasma cholesterol of female mice administered AAV-PCSK9 frequently tends to be lower than that of male mice,¹³¹ this was not statistically significant in the Moss et al. study.¹¹⁸ One study in white rabbits also found that females have more macrophage accumulation in plaques while another showed no difference. Another important difference between these studies is that most use histology to quantify inflammation whereas the study showing increased inflammation in males used flow cytometry, a more specific and quantitative approach. Indeed, some of the markers used to identify inflammatory cells by IHC (including CD68) have been shown to be expressed on smooth muscle cells in the plaque,¹⁷⁷ making histological identification of leukocytes difficult without lineage tracing.

Of the CD11b⁺ myeloid cells in the aortic arch plaque, male C57BL/6 mice administered AAV-PCSK9 also had a higher proportion of Ly-6C^hi-expressing cells compared to their female counterparts,¹¹⁸ suggesting a more pro-inflammatory myeloid cell phenotype in males. This is consistent with another study showing that male ApoE^−/− mice express higher levels of “M1-like” marker iNOS in the aortic root, whereas female mice express higher levels of “M2-like” marker Ym1 in the brachiocephalic artery.⁹¹ Interestingly, these sex differences in macrophage-associated gene expression were not observed in other vascular regions, suggesting that plaque macrophage phenotype may be specific to the lesional microenvironment of different vascular beds. Although M1- and M2-like markers are generally associated with pro- and anti-inflammatory macrophage phenotypes, respectively, macrophage phenotypic state occurs across a continuous spectrum rather than two polarized states.^178,179 Therefore, these data must be interpreted with caution in both human and animal models.

The fact that male ApoE^−/− mice have smaller aortic root plaques yet male ApoE^−/− mice expressed higher levels of pro-inflammatory macrophage markers further highlights the notion that in mouse atherosclerosis models, plaque inflammatory state, rather than size, may more closely model sex-specific outcomes in humans by virtue of contributing to plaque vulnerability. Differences in macrophage phenotype in the plaque may further contribute to processes that are involved in plaque progression and stability. Regarding foam cell formation, one study found that male and female ApoE^−/− mice have equal numbers of CD45⁺ I-A/I-E⁺ BODIPY⁺ foam cells,¹⁸⁰ suggesting no sex difference in reverse cholesterol transport. Another study found that male ApoE^−/− mice expressed higher levels of MMP-13, MMP-14, and TIMP-3 in the aortic root compared to female mice,⁹¹ suggesting another possible mechanism of increased fibrous cap degradation and susceptibility to plaque rupture in males.

c. Sex differences in adaptive immunity in atherosclerosis

Although very few studies compare adaptive immunity in atherosclerosis by sex, in mouse models, males appear to have more or equal numbers of T cells in plaques compared to females (Table 6). However, we only identified four studies comparing plaque T cells between sexes, all in young mice. Thus, more data are needed to determine whether this sex difference is age-dependent, whether there are sex differences in specific T cell subsets, and whether this sex difference also occurs in larger animal models and in other vascular beds aside from the aorta.

There is also very limited data regarding sex differences in B cells in atherosclerosis. One study in 24-week old ApoE^−/− mice treated with HFD for 16 weeks showed that males had more CD22⁺ staining for B cells in the aortic root compared to females.¹⁰⁰ Additionally, B cell FcγRIIb differentially alters atheroprotective B-1 responses in male and female mice treated with an atherogenic diet.¹⁸¹ More studies are needed for the same reasons stated above. For example, the fetal-derived B-1 subtype is atheroprotective due to its IgM responses against ox-LDL,¹⁸² while the bone marrow-derived B-2 subtype is pro-atherogenic.^183,184

E. Potential Mechanisms of Sex Differences in Vascular Inflammation in Humans and Animal Models

1. Sex Differences in Leukocyte Trafficking

Since the relative protection from heart attack and ischemic stroke in women compared to men is lost after menopause, the effects of sex hormones (particularly estrogen) have been extensively studied in attempt to explain sex differences in atherosclerosis. Multiple studies have shown that ovariectomy increases plaque size and burden and that estrogen supplementation rescues this phenotype in ApoE^−/− mice independently of changes in plasma cholesterol levels.^185,186,187 The plaques in ovariectomized ApoE^−/− mice had a similar percentage of Mac-3⁺ plaque area compared to sham-controls and ovariectomized mice treated with estradiol.¹⁸⁸ This might suggest no impact of estrogen in inflammation. However, because lesion size was increased in ovariectomized mice compared to the other two groups, the absolute number of macrophages in the plaque is likely increased.

A major determinant of leukocyte accumulation in the plaque is their recruitment from the circulation and transmigration into the subintima by mechanisms that have been well characterized.¹⁸⁹ Briefly, local damage to the endothelium and/or systemic inflammatory stimuli result in the upregulation of surface adhesion molecules on endothelial cells. Responding to chemotactic signals, circulating leukocytes initially tether to the endothelium via interactions between their glycoproteins and endothelial P-selectin and E-selectin. These “rolling” leukocytes gradually decrease in velocity, firmly adhere to the endothelium via interactions between integrins on the leukocytes and endothelial ICAM-1 and VCAM-1, and eventually transmigrate into the subintima. Much of our understanding of leukocyte trafficking comes from intravital epifluorescence microscopy studies performed in the cremaster muscle vasculature, and hence such studies were performed exclusively in male mice.^190,191 Using mesenteric vessel intravital microscopy to directly compare males and females, Moss et al. also demonstrated that the number of rolling cells and adherent cells in the mesenteric veins were higher in male C57Bl/6 mice compared to female littermates administered TNFα intraperitoneally.¹¹⁸ The decrease in leukocyte “slow-rolling” and firm adhesion in females correlated with decreased vascular expression of E-selectin and ICAM-1 in vessels from female mice. This sexual dimorphism was also present in a model of chronic vascular inflammation induced by hyperlipidemia in NZW rabbits. Nathan et al. demonstrated that the number of adherent monocytes and degree of transendothelial migration is greater in male rabbits compared to female rabbits in fixed abdominal aortic vessels ex vivo by scanning electron microscopy.¹⁹² Additionally, the number of adherent monocytes was increased in ovariectomized female rabbits compared to sham control females and was attenuated with estradiol replacement, supporting the role for estrogen in attenuating leukocyte trafficking. The following sections review sex differences in the molecular mechanisms involved at each stage of leukocyte recruitment in the pathogenesis of vascular inflammation.

2. Sex Differences in Endothelial Adhesion Molecules

Early lesion development in the ascending aorta of an LDLR^−/− mouse was found to be dependent on P-selectin in males but not in females.¹⁹³ However, both male and female LDLR^−/− mice deficient in P-selectin and E-selectin exhibited significantly attenuated lesions in the proximal aorta after 8 weeks of an atherogenic diet.¹⁹⁴ Young male C57Bl/6 mice express more E-selectin and ICAM-1 protein in the mesenteric veins compared to female mice administered TNFα intraperitoneally.¹¹⁸ In the same study, estrogen reduced the TNFα-induced upregulation of E-selectin and ICAM-1 mRNA and protein in human and mouse endothelial cells. However, the sex of these cells was not noted and may impact the sensitivity to estrogen. The benefits of estrogen on reducing EC adhesion molecule expression does not appear to be stimulus-specific. In addition to the data with TNFα, estrogen has been shown to attenuate lipopolysaccharide-induced upregulation of endothelial ICAM-1¹⁹⁵ and VCAM-1¹⁹⁶ expression as well as IL-1-induced upregulation of endothelial E-selectin, ICAM-1, and VCAM-1 mRNA and protein.¹⁹⁷ These effects were abrogated by the estrogen receptor (ER) antagonist ICI and could not be reproduced by 17α-estradiol, an inactive stereoisomer of estrogen.¹⁹⁷

3. Sex Differences in Circulating Monocytes and Chemotaxis

According to the Atherosclerosis Risk in Communities (ARIC) study, men have more circulating CD14⁺ monocytes compared to women.¹⁹⁸ Similarly, the “non-classical” subset of circulating human monocytes characterized by CD16 are increased in men compared to women.¹⁹⁹ CD16⁺ monocytes have been shown to be associated with endothelial dysfunction in patients with coronary artery disease based on decreased NO bioavailability and increased superoxide production;²⁰⁰ primed to produce atherogenic cytokines like TNFα, IL-1β, and IL-6;²⁰¹ elevated in atherosclerosis;^202,203 and positively associated with carotid IMT.²⁰⁴ Interestingly, estrogen is known to downregulate CD16 expression.²⁰⁵ Human monocytes and macrophages express ERα and ERβ at relatively low levels compared to cells of the lymphoid lineage.²⁰⁶ Thus, it is likely that estrogen has both direct and indirect effects on monocytes/macrophages via direct effects on other leukocytes (T cells) and vascular cells that influence macrophage accumulation and function in the plaque.²⁰⁷

One of the primary chemotactic signals towards sites of inflammation is monocyte chemoattractant protein-1 (MCP-1 or CCL2), which binds the receptor CCR2 on monocytes. A multiple linear regression analysis in 79 patients at increased cardiovascular risk revealed that male sex had a significant interaction with CCR2 levels on circulating monocytes and vascular inflammation independently of traditional cardiovascular risk factors and statin use.²⁰⁸ Estrogen has also been shown to inhibit chemotaxis of human monocytes in response to MCP-1.²⁰⁹ Similar findings have also been found in animal models. On normal chow, ovariectomized NZW rabbits had higher MCP-1 mRNA and protein expression in the descending thoracic aorta compared to ovary-intact controls. MCP-1 expression was also increased in ovariectomized NZW rabbits treated with high-cholesterol diet for 6 weeks. This increase was abolished with supplementation of physiologic levels of estradiol.²¹⁰ Therefore, monocytes in men and male animals may be more primed for chemotaxis towards sites of inflammation via MCP-1, a crucial chemokine in atherosclerosis,^211,212 and the sex difference may be in part regulated by estrogen in females.

4. Sex Differences in Adhesion Molecule Ligands on Immune Cells

Both myeloid cells and T cells initially tether and roll along the endothelium via glycoprotein selectin ligands. In particular, P-selectin glycoprotein ligand 1 (PSGL-1) binds to both endothelial P-selectin and E-selectin, while E-selectin ligand 1 (ESL-1) is specific to endothelial E-selectin. However, there are many other ligands that can bind to selectins and contribute to leukocyte trafficking. PSGL-1 and ESL-1 are constitutively expressed on the surface of circulating immune cells. In a flow cytometry analysis of PBMCs from patients enrolled in the ARIC study, surface PSGL-1 expression on monocytes did not differ by sex. There was also no difference in PSGL-1 expression in human lung macrophages²¹³ and bone marrow biopsy specimens²¹⁴ by immunohistochemistry, although this was evaluated by a visual grading system. Interestingly, surface PSGL-1 expression was higher on granulocytes and lymphocytes from men compared to that of women in the ARIC study. These leukocytes release chemotactic factors and other signals that promote monocyte recruitment or intraplaque macrophage proliferation/survival and thus could contribute to the greater macrophage accumulation observed in plaques of males compared to females. The ARIC study also evaluated PSGL-1 polymorphisms in association with atherosclerosis and ischemic events. In the primary analysis, there was no significant association of the Met62Ile polymorphism with advanced carotid atherosclerosis in the overall patient population.²¹⁵ However, a follow-up study found that there was an association specifically in female patients, but not in male patients.²¹⁶ The Ile allele is hypothesized to modify the affinity of PSGL-1 to P-selectin and thus lead to the increased release of soluble PSGL-1 observed in the circulation.²¹⁷ Further studies are needed to understand the mechanism for the sex difference in the association between this PSGL1 polymorphism and ischemic outcomes.

LFA-1 (CD11b/CD18 or α_Mβ₂) and VLA-4 (CD49d/CD29 or α4β1) are integrins on leukocytes that bind to endothelial ICAM-1 and VCAM-1, respectively, facilitating firm adhesion and subsequent transmigration of leukocytes into the subintima. Interestingly, deficiency of integrin alpha M (α_M) of LFA-1 increased atherosclerosis lesion area, macrophage accumulation, and macrophage proliferation in the aortic root of female, but not male, ApoE^−/− mice.²¹⁸ These sex-specific findings were attributed to increased expression of lipid scavenger receptors class B member 3 (CD36) and macrophage scavenger receptor 1 (MSR1 or CD204) as well as decreased expression of Forkhead box protein M1 (FoxM1), ERα, and ERβ in peritoneal macrophages isolated from female α_M^−/−/ApoE^−/− mice compared to ApoE^−/− mice. Therefore, while LFA-1 promotes vascular inflammation by facilitating leukocyte trafficking, its α_M component may play an atheroprotective role by maintaining ER expression in macrophages, but whether this phenomenon changes with aging has not been studied. Integrin beta 1 (β1) of VLA-4 is absent in the spleen of male mice, but present in female mice.²¹⁹ In fact, testosterone and dihydrotestosterone, but not estrogen nor progesterone, were able to suppress β1 mRNA expression in T cells. Therefore, β1 may regulate sex-specific, VLA-4-mediated, T cell interactions with the vasculature and other immune cells.

F. Conclusions and Limitations in Our Understanding of SABV in Atherosclerosis

In summary, despite the well-known sexual dimorphism in the incidence and complications of atherosclerosis, there are relatively limited data in the clinical and pre-clinical literature to rigorously address mechanisms driving sex as a biological variable in atherosclerosis. In humans, non-invasive imaging and pathological evaluation reveals that males develop atherosclerotic plaques earlier and that plaques from males are more inflamed and have additional unstable features. The overall area burdened by plaque is greater in males while individual plaque stenosis may be greater in females. Importantly, plaque burden, rather than individual stenosis, predicts MI and stroke events. The clinical data also reveal an interaction between sex and age with younger males having more atherosclerosis burden and increased incidence of ischemic events until the seventh decade when women catch up and ultimately surpass males in the incidence of MI in extreme old age.

In the pre-clinical literature, the vast majority of studies do not examine both sexes, but when they do, well-powered direct statistical comparisons between males and females remain relatively rare. In addition to the added cost of such studies, the finding that plaque size is generally larger in young females compared to males in most animal models of atherosclerosis (with the exception of non-human primates) may have contributed to the generalization that such animal models are not useful for studying sex differences. Evaluation of the available data suggests that the larger plaque size in female animals becomes less evident with age with some studies showing greater plaque burden in older male animals. More studies are needed in older animals to determine if aged animals might be a better model to study sex as a biological variable in atherosclerosis.

When sexes are compared in pre-clinical studies, plaque size is the most common measurement. However, most evidence suggests that plaque size does not correlate with plaque rupture in humans (r=0.00, p=0.99).¹⁶⁴ Rather, plaque inflammation and morphology are better surrogates for plaque vulnerability, ischemic events, and mortality. At this time, the number of pre-clinical studies directly comparing plaque inflammation between the sexes is scant. However, the data available suggests that this endpoint may be more reflective of the human condition with males having more inflamed (if smaller) plaques in pre-clinical models. Failure to include both sexes in mechanistic studies of atherosclerosis are also missed opportunities to uncover underlying mechanisms that explain sex differences in this disease and may be translationally relevant. Indeed, one potential contributor to the failure of clinical trials that include both men and women may be the development of therapies targeting a disease mechanism that is validated in only one sex in pre-clinical animal models. Understanding the mechanisms driving sex as a biological variable in atherosclerotic disease is critical to future precision medicine strategies to mitigate what is still the leading cause of death of men and women in the world.

Non-standard Abbreviations and Acronyms

AAV-PCSK9

Adeno-associated virus expressing gain-of-function mutant form (D374Y) of human proprotein convertase subtilisin/kexin type 9

ARIC

Atherosclerosis Risk in Communities study

CCR2

C-C chemokine receptor type 2

CONFIRM

COroNary CT Angiography Evaluation For Clinical Outcomes: An InteRnational Multicenter Registry

CRP

C-reactive protein

CVD

Cardiovascular disease

EC

Endothelial cell

ER

Estrogen receptor

ESL-1

E-selectin ligand 1

FDG

¹⁸F-fluorodeoxyglucose

ICAM-1

Intercellular adhesion molecule 1

IL-1β

Interleukin 1 beta

IMT

Intima-media thickness

IVUS

Intravascular ultrasound

LFA-1

Lymphocyte function-associated antigen 1

MI

Myocardial infarction

NHANES

National Health and Nutrition Examination Survey

NZW

New Zealand White rabbit

OCT

Optical coherence tomography

PBMC

Peripheral blood mononuclear cells

PSGL-1

P-selectin glycoprotein ligand 1

REFINE

Risk Estimation Following Infarction, Noninvasive Evaluation study

SABV

Sex as a biological variable

TNFα

Tumor necrosis factor alpha

VCAM-1

Vascular cell adhesion molecule 1

VLA-4

Very late antigen 4

WWHL

Watanabe Heritable Hyperlipidemic rabbit