The association between baseline circulating progenitor cells and vascular function: The role of aging and risk factors

Abstract

Background:

To investigate the cross-sectional and longitudinal relationships between vascular function and circulating progenitor cell (CPC) counts with respect to aging and exposure to risk factors.

Methods:

In 797 adult participants, CPCs were enumerated by flow cytometry as CD45^med mononuclear cells expressing CD34 epitope and its subsets co-expressing CD133, and chemokine C-X-C motif receptor 4 (CXCR4+). Arterial stiffness was evaluated by tonometry-derived pulse wave velocity (PWV) and microvascular function was assessed as digital reactive hyperemia index (RHI).

Results:

In cross-sectional analyses, for every doubling in CD34+ cell counts, PWV was 15% higher and RHI was 9% lower, after adjusting for baseline characteristics and risk factors (p for all < 0.01). There were significant CPC-by-age-by-risk factor interactions (p <0.05) for both vascular measures. Among younger subjects (< 48 years), CPC counts were higher in those with risk factors and vascular function was better in those with higher compared to those with lower CPC counts (p for all < 0.0l). In contrast, in older participants, CPCs were not higher in those with risk factors, and vascular function was worse compared to the younger age group. A lower CPC count at baseline was an independent predictor of worsening vascular function during 2-year follow-up.

Conclusion:

A higher CPC count in the presence of risk factors is associated with better vascular function among younger individuals. There is no increase in CPC count with risk factors in older individuals who have worse vascular function. Moreover, a higher CPC count is associated with less vascular dysfunction with aging.

Background

Arterial stiffness and microvascular dysfunction are markers of subclinical vascular disease that are associated with long-term cardiovascular morbidity and mortality.^1–7 Arterial stiffness can be evaluated as pulse wave velocity (PWV), with a higher PWV indicating stiffer arteries.⁸ A lower reactive hyperemia index (RHI), a measure of digital microvascular dilator dysfunction, measured using peripheral arterial tonometry, reflects microvascular dysfunction.^9,10 Aging and exposure to atherosclerotic risk factors such as obesity, hypertension, type 2 diabetes mellitus, and tobacco smoking cause vascular injury and subsequent adverse cardiovascular events.^6,11–13 The natural innate response to ischemia or vascular trauma involves recruitment of bone marrow-derived circulating progenitor cells (CPCs) to the sites of injury where they contribute to repair and regeneration.^14–19 Both animal and human studies have demonstrated that CPCs can facilitate the repair of injured blood vessels and prevent age-related vascular remodeling.^20,21 We and others have demonstrated that CPCs can be accurately phenotyped in humans as CD34-expressing mononuclear cells that possess the potential to differentiate into hematopoietic, endothelial, and nonhematopoietic lineages and contribute to vascular and myocardial regeneration.^22–26 Dual expression of CD34 and CD133, a 5-transmembrane antigen of primitive stem cells, identifies an early CPC-enriched subpopulation.²⁷ Finally, co-expression of the chemokine (C-X-C motif) receptor 4 (CXCR4) permits homing of CPCs to hypoxic milieu and further characterizes the subpopulation of CD34+ CPCs with the capacity for homing and tissue repair.²⁸

Age is an important determinant of CPC activity. We have previously shown that in younger individuals exposed to cardiovascular risk factors, there is no reduction in CPC counts, whereas older subjects with risk factors have lower CPC counts.²⁹ However, the age-dependence of the associations between vascular function, risk factors, and CPCs is unknown. Our aim was to investigate the relationship between vascular function and CPC counts with respect to aging and exposure to cardiovascular risk factors. We hypothesized that a relationship exists between vascular function and CPCs and that this relationship is modulated by age and exposure to cardiovascular risk factors. We also hypothesized that lower CPC counts at baseline would be associated with further worsening of the vascular function over time.

Methods

Study population

We recruited participants from Emory Clinical Cardiovascular Research Institute and the Emory/Georgia Institute of Technology Predictive Health Institute between 2006 and 2013.^30,31 None of the participants had a documented history of cardiovascular disease. Subjects were free of any acute illness and had normal performance status. Presence of hypertension, hypercholesterolemia, and diabetes mellitus were defined according to the Seventh Report of the Joint National Committee on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure; Adult Treatment Panel III; and American Diabetes Association criteria, respectively.^32–34 All subjects signed an informed consent that was approved by the Emory and Georgia Tech institutional review boards. Vascular testing and blood draws were performed after an overnight fast.

Subjects were followed with comprehensive evaluations up to 3 years. A total of 521 subjects completed a follow-up visit at 2 or 3 years. During each follow-up visit, physical measurements, blood samples, and vascular tests were remeasured. Informed consent was obtained from all subjects and all study protocols were approved by the Institutional Review Board of Emory University.

Circulating progenitor cell (CPC) measurements

We measured CPC counts using flow cytometry as previously described.^35,36 Briefly, after an overnight fast, venous blood samples were collected into EDTA tubes and processed within 4 hours.³⁶ Peripheral blood mononuclear cell subsets enriched for CPCs were enumerated using flow cytometry as CD45^med cells co-expressing CD34, CD133, and/or CXCR4 epitopes. Venous blood measuring 300 μL was incubated with 7 μL of fluorochrome-labeled monoclonal antihuman mouse antibodies (BD Biosciences, Puerto Rico), 15 μL PerCP-CD45 (BD Biosciences), 3 μL PE-Cy7-conjugated anti-CXCR4 (clone 12G5) (eBioscience, Puerto Rico), and 10 μL APC-CD133 (Miltenyi, Cambridge, MA, USA) in the dark for 15 minutes. Afterwards, 1.2 mL of ammonium chloride lysing buffer was added to lyse red blood cells, followed by the addition of 1.2 mL of staining medium (phosphate-buffered saline (PBS) with 3% heat inactivated serum and 0.1% sodium azide) to stop the lysing reaction. Samples were then centrifuged at 1500 rpm for 5 minutes, followed by washing with PBS. In the next step, cells were suspended in 500 μL of staining medium and mixed, and run in the flow cytometer within 4 hours (BD FACSCanto II Flow Cytometry System; BD Biosciences). Prior to flow cytometry, 100 μL of AccuCheck Counting Beads (Invitrogen, Waltham, GA, USA; catalog number PCB100) was added to act as an internal standard for direct estimation of the concentration of target cell subsets. A minimum of 2.5 million events were acquired from the cytometer. Flow cytometry data were analyzed with FlowJo software (Treestar, Inc., Ashland, OR, USA). The absolute mononuclear cell count was estimated using a Coulter AcT/Diff cell counter (Beckman Coulter, Pasadana, CA, USA) as the sum of lymphocytes and monocytes. CD45^med cells were selected to exclude CD45^bright and CD45− (negative) cells. Excluding these rare CD45^bright cells helps eliminate lymphoblasts. Exclusion of CD45− cells eliminates nonhematopoietic progenitors like mesenchymal or osteoprogenitor cells as these are typically CD45−. The coefficients of variation of the cell types were CD34+ 2.9%, CD34+/CD133+ 4.8%, and CD34+/CXCR4+ 6.5%.

Vascular function measurements

All studies were performed on subjects at rest after an overnight fast with participants resting in the supine position in a quiet, temperature-controlled environment set at 22°C.^9,37,38

Arterial stiffness.

Arterial stiffness, estimated using carotid-femoral PWV, was measured using pressure waveforms at carotid and femoral arterial sites using tonometry and electrocardiogram gating, as previously described.^8,39,40 All measurements were performed using the Sphygmocor device (AtCor Medical, Sydney, Australia) that records sequential high-quality pressure waveforms at peripheral pulse sites using a high-fidelity tonometer. Velocity measured as distance/time in meters/second was calculated using the ‘foot-to-foot’ method that measures the interval between the R wave on the electrocardiogram and the foot of the recorded pressure waveform at each site, whereas distance between the sites was measured manually by the operator. Reproducibility studies in our laboratory on nine subjects on consecutive days have demonstrated a coefficient of variation of 3.8% for PWV.⁴¹

Peripheral arterial tonometry.

Digital peripheral arterial tonometry was used to measure microvascular function during reactive hyperemia (Endo-PAT; Itamar Medical, Caesarea, Israel) in the tip of the index finger.^5,42 Pulse volume amplitude was measured at rest and during reactive hyperemia that was elicited by the release of an upper arm blood pressure cuff inflated to suprasystolic pressure for 5 minutes. RHI was calculated as the ratio of the post- to preocclusion pulse volume amplitude of the tested arm, divided by the post- to preocclusion ratio of the control arm (the average pulse volume amplitude over a 1-minute interval starting 1 minute after cuff deflation divided by the average pulse volume amplitude at baseline measured for 1 minute before cuff inflation).

Statistical analysis

Baseline characteristics are summarized as means ± SD or as counts and proportions for continuous and categorical variables, respectively. Differences between groups were assessed using ANOVA and t-test for continuous variables and χ² or Fisher’s exact test for categorical variables where appropriate. Wilcoxon signed rank test was used to compare CPCs between subgroups of interest. CPC counts were log₂ transformed prior to regression analyses. Linear regression analyses were performed to determine the association between vascular function measures at baseline and their changes during follow-up (outcome variable) and log-transformed CPC levels, and number of risk factors. All models were adjusted for baseline characteristics (age, sex, and race), and cardiovascular risk factors (smoking status, obesity, history of hypertension, history of diabetes, and history of hyperlipidemia). The interaction of CPC counts with age and risk factors were tested by entering the interaction term in the regression models. All analyses were conducted using Stata 14 (StataCorp LP, College Station, TX, USA). All p-values less than 0.05 were considered as statistically significant.

Results

Table 1 describes the characteristics of the 797 study subjects who were further sub-divided into groups based on the median age of the cohort (younger and older than 48 years). Older participants had a higher prevalence of hypertension, hyperlipidemia, and diabetes and lower CD34+ and CD34+/CD133+ cell counts compared to those < 48 years of age (online supplemental Figure 1). Also, individuals ⩾ 48 years of age had higher PWV compared to the younger participants (Table 1). The relationship between CPC counts and age was not significantly different between men and women (p = 0.36 for interaction; online supplemental Figure 2).

Table 1.

Baseline characteristics of the study cohort.

Characteristic	All participants (n = 797)	Age < 48 y (n = 398)	Age ⩾ 48 y (n = 399)	p-value
Demographics
Mean age, years (SD)	48 (11)	38 (5)	57 (7)	< 0.001
Female (%)	511 (64.0)	267 (66.1)	244 (61.8)	0.20
Black (%)	184 (23.0)	95 (23.8)	89 (22.3)	0.51
Risk factors (%)
Smoking	40 (5.0)	20 (5.0)	20 (5.1)	0.92
Obesity (BMI > 30)	199 (25.2)	98 (24.6)	101 (25.7)	0.72
Hypertension	256 (32.1)	101 (25.1)	155 (39.2)	< 0.001
Hyperlipidemia	132 (16.5)	24 (6.0)	108 (27.3)	< 0.001
Diabetes	77 (9.6)	27 (6.7)	50 (12.7)	0.004
CPC counts (cells/μL), median (IQR)
CD34+	2.06 (1.28–3.07)	2.58 (1.32–3.35)	2.01 (1.27–2.99)	0.02
CD34+/CD133+	0.87 (0.54–1.38)	1.03 (0.60–1.54)	0.84 (0.53–1.36)	0.01
CD34+/CXCR4+	0.81 (0.48–1.32)	0.83 (0.51–1.30)	0.80 (0.48–1.34)	0.64
Vascular function, mean (SD)
RHI	1.94 (0.6)	2.08 (0.6)	1.92 (0.6)	0.26
PWV, m/s	7.1 (1.2)	6.4 (1.0)	7.6 (1.4)	< 0.001

Bold indicates a statistically significant value.

BMI, body mass index; CPC, circulating progenitor cells; PWV, pulse wave velocity; RHI, reactive hyperemia index.

Relationship between CPCs, risk factors, and age

Figure 1 shows the significant associations between the increasing number of risk factors and higher CPC counts. In a multivariable analysis adjusted for age, sex, and race, the number of risk factors (hypertension, obesity, diabetes, smoking history, and hyperlipidemia) was independently associated with higher CPC counts for all CPC subsets in the entire cohort (Figure 1, panels A–C). For every unit increase in the number of risk factors, the CPC counts increased by 21%, 24.5%, and 21.4% for the CD34+, CD34+/CD133+, and CD34+/CXCR4+ subsets, respectively. A significant interaction was present in the relationship between all CPC counts and risk factors with respect to age (p for all < 0.01). Among participants < 48 years of age, a higher burden of risk factors was associated with higher CPC counts, whereas among older participants there were no significant differences in the levels of all CPC subsets between those with or without risk factors after sex and race adjustment (Figure 1, panels D–F).

Figure 1.

Relationship between the number of risk factors and CPCs (panels A and D = CD34+ cells; B and E = CD34+/CD133+ cells; C and F = CD34+/CXCR4+ cells) in the entire cohort. (A–C) These panels demonstrate associations between the number of risk factors and CPC subsets in the entire cohort. (D–F) These panels demonstrate associations between the number of risk factors and CPC subsets stratified by age.

Black lines indicate associations among those < 48 years old; gray lines indicate associations among those ⩾ 48 years old.

All models were adjusted for baseline characteristics (age, sex, and race).

Relationship between risk factors, vascular function, and CPCs

There was no significant correlation between PWV and RHI in the entire cohort (r = 0.29, p = 0.43). This finding is consistent with previous reports among individuals without a history of cardiovascular disease.⁴³ Linear associations were found between increasing number of risk factors and worsening vascular function (higher PWV and lower RHI) after adjusting for age, sex, and race (Figure 2).

Figure 2.

Relationship between vascular function and the risk factors. All models were adjusted for baseline characteristics (age, sex, and race). Risk factors included smoking, obesity, and history of hypertension, diabetes, and hyperlipidemia.

Dashed lines represent 95% CI.

PWV, pulse wave velocity; RHI, reactive hyperemia index.

Multivariate regression models were created to identify the determinants of PWV and RHI. Increasing age, male sex, Black race, and history of hypertension, smoking, and hyperlipidemia were independently associated with a higher PWV (β = 0.36, p < 0.001; Table 2). Male sex, Black race, and history of hypertension were associated with a lower RHI (β = −0.03, p = 0.02; Table 2). In addition to these factors, higher CPC counts were also found to be independently associated with worse vascular function (higher PWV and lower RHI) in the entire cohort. For every doubling in CD34+ cells, PWV was 15% higher and RHI was 9% lower.

Table 2.

Multivariate determinants of vascular function.

Variables	PWV β (95% CI)	p-value	RHI β (95% CI)	p-value
Model 1: Risk factors a
Age	0.04 (0.03, 0.05)	< 0.001	−0.008 (−0.012, 0.004)	0.21
Male sex	0.46 (0.66, 0.25)	< 0.001	−0.10 (−0.20, −0.008)	0.03
Black race	0.61 (0.38, 0.85)	< 0.001	−0.20 (−0.31, −0.09)	< 0.001
Smoking	0.76 (0.34, 1.17)	< 0.001	0.03 (−0.16, 0.23)	0.70
Obesity (BMI > 30)	0.13 (−0.08, 0.36)	0.22	−0.01 (−0.11,0.09)	0.83
Hypertension	0.11 (0.02, 0.33)	0.02	−0.11 (−0.21, −0.01)	0.02
Hyperlipidemia	0.33 (0.02, 0.57)	0.03	0.01 (−0.11,0.14)	0.85
Diabetes	−0.09 (−0.42, 0.23)	0.57	0.001 (−0.15, 0.16)	0.90
Models 2–4: Risk factors + individual CPC subsets
CD34+, cells/mL, per 100% increase	0.15 (0.12, 0.15)	0.009	−0.09 (−0.15, −0.01)	0.01
CD34+/CD133+, cells/mL, per 100% increase	0.14 (0.11, 0.14)	0.012	−0.06 (−0.12, −0.01)	0.03
CD34+/CXCR4+ cells/mL, per 100% increase	0.16 (0.13, 0.21)	< 0.001	−0.08 (−0.20, −0.03)	< 0.001

Progenitor cells counts (CPC) were log-transformed (base 2) for the purpose of multivariable analysis.

^aThe β and CIs reported for the demographics and clinical characteristics are derived from the model not incorporating any CPC counts. All models were adjusted for baseline characteristics (age, sex, and race), and risk factors (smoking status, obesity, history of hypertension, history of diabetes, and history of hyperlipidemia).

Bold indicates a statistically significant value.

BMI, body mass index; CPC, circulating progenitor cells; PWV, pulse wave velocity; RHI, reactive hyperemia index.

To investigate whether the relationship between CPC counts and vascular function was affected by age and risk factors, linear regression models were created using interaction terms. There were significant CPC-by-age-by-risk factor interactions (p < 0.05) for both vascular measures, indicating that the association between PWV or RHI and CPCs differed by both age and presence or absence of risk factors. In the younger group, both PWV and RHI were worse in those with risk factors (p for all < 0.01; Figure 3). Further, in the younger age group, in the presence of risk factors, the group with high CPC counts had better vascular function compared to those with low CPC counts (p for all < 0.05; Figure 3). These differences were not present in the older group, in whom no significant differences in PWV and RHI were observed between either those with or without risk factors, or among those with high versus low CPC counts (Figure 3).

Figure 3.

Relationship between CPCs and measures of vascular function stratified by age and risk factors. Panel 1 demonstrates the relationships between PWV, risk factors, and CPC levels in younger (1A) and older (1B) individuals. Panel 2 demonstrates the relationships between RHI, risk factors, and CPC levels in younger (2A) and older (2B) individuals. For both PWV and RHI, vascular function was worse in those with risk factors compared to those without risk factors (p for all < 0.01) in the younger group (1A, 2A). Also, in the younger age group with risk factors (1A, 2A), the group with high CPC counts had better vascular function compared to those with low CPC counts (p for all < 0.05). These differences were not present in the older group (1B, 2B), in whom no significant differences in PWV and RHI were observed.

*p < 0.05 comparing high vs low CPC numbers with those with risk factors.

CPC, circulating progenitor cells; PWV, pulse wave velocity; RHI, reactive hyperemia index.

CPCs and change in vascular function at follow-up

A total of 521 individuals completed a follow-up visit at a median of 2 years (IQR 1–3). There were no significant differences between those who completed their follow-up visit and the remaining 276 participants who did not have follow-up visits (online supplemental Table 1). Among those who completed their follow-up visit, vascular function worsened among subjects with one or more risk factors (n = 300) at baseline (PWV 7.1–7.4, p = 0.01; RHI 1.97–1.84, p = 0.03), but no significant changes were observed for subjects without risk factors (n = 221) at baseline (PWV 7.0–7.1, p = 0.41; RHI 1.98–1.97, p = 0.56). As shown in Table 3, history of smoking and hyperlipidemia were independently associated with worsening of PWV during follow-up, whereas history of hypertension was the only risk factor associated with worsening of RHI during follow-up. Additionally, a lower CPC count at baseline was independently associated with worse vascular function during follow-up (Table 3). Subjects with lower than median CPC counts at baseline had deterioration of their vascular function during follow-up, whereas those with higher CPC counts had no change in vascular function (Figure 4). These associations were independent of age, sex, and race (RHI: β = −0.21, p = 0.005; PWV: β = 0.19, p = 0.01).

Table 3.

Multivariate determinants of changes in vascular function during follow-up.

Variables	PWV β (95% CI)	p-value	RHI β (95% CI)	p-value
Model 1: Risk factors a
Age	0.004 (−0.007, 0.01)	0.44	−0.006 (−0.01,0.007)	0.08
Male sex	0.04 (−0.26, 0.34)	0.78	−0.12 (−0.29, −0.02)	0.02
Black race	0.13 (−0.18, 0.45)	0.40	0.05 (−0.13, 0.23)	0.56
Smoking	0.51 (0.01, 1.04)	0.03	−0.08 (−0.39, 0.21)	0.57
Obesity (BMI > 30)	0.04 (−0.34, 0.24)	0.73	0.006 (−0.15, 0.17)	0.93
Hypertension	0.06 (−0.35, 0.23)	0.67	−0.11 (−0.28, −0.05)	0.01
Hyperlipidemia	0.40 (0.02, 0.79)	0.038	−0.17 (−0.40, 0.04)	0.11
Diabetes	0.15 (−0.59, 0.27)	0.47	−0.14 (−0.39, 0.10)	0.10
Models 2–4: Risk factors + individual CPC subsets at baseline
CD34+, cells/mL, per 100% increase	−0.14 (−0.28, −0.07)	0.009	0.09 (0.02, 0.14)	0.01
CD34+/CD133+ cells/mL, per 100% increase	−0.12 (−0.31, −0.01)	0.01	0.06 (0.01,0.11)	0.01
CD34+/CXCR4+ cells/mL, per 100% increase	−0.17 (−0.33, 0.08)	< 0.001	0.08 (0.01,0.12)	0.008

Progenitor cells counts (CPC) were log-transformed (base 2) for the purpose of multivariable analysis.

^aThe β and 95% CIs for the demographics and clinical characteristics are derived from the model not incorporating any progenitor cell counts.

Model 1 adjusted for baseline demographics (age, sex, and race), and baseline risk factors (smoking status, obesity, history of hypertension, history of diabetes, and history of hyperlipidemia).

Model 2 adjusted for variables in Model 1 and CD34+ cells.

Model 3 adjusted for variables in Model 1 and CD34+/CD133+ cells.

Model 4 adjusted for variables in Model 1 and CD34+/CXCR4+ cells.

Bold indicates a statistically significant value.

BMI, body mass index; CPC, circulating progenitor cells; PWV, pulse wave velocity; RHI, reactive hyperemia index.

Figure 4.

Changes in vascular function during follow-up for subjects with high (⩾ 2.04 cells/mL) compared to low (< 2.04 cells/mL) number of CD34+ CPCs at baseline.

CPC, circulating progenitor cells; PWV, pulse wave velocity; RHI, reactive hyperemia index.

Discussion

In a large and diverse cohort of individuals including healthy subjects and those with risk factors for cardiovascular disease, we demonstrate, firstly, that CPC counts are higher with increasing risk factor burden in the younger participants. However, in older individuals, no stimulation of CPCs is observed with increasing risk factor burden. Secondly, in younger individuals, those who had one or more risk factors and high CPC counts had preserved vascular function compared to those with similar risk factor burden but with low CPC counts. These associations were not present in the older group. Finally, subjects with higher CPC counts at baseline were protected against worsening of vascular function during a 2-year follow-up period, whereas those with low CPC counts had deterioration of vascular function over time. These findings imply that at a younger age, exposure to risk factors leads to stimulation of CPC release from the bone marrow, which is protective against further vascular damage. However, such repair mechanisms may be lost in older individuals who are unable to increase CPC mobilization in response to risk factors, and this results in worse vascular function.

Earlier studies have indicated that bone marrow-derived CPCs contribute to the maintenance of normal endothelial function by homing to the sites of vascular injury where they contribute to repair and regeneration of the vascular wall.^22,44,45 Experimental studies have demonstrated that a progressive decline in progenitor cells with aging contributes to vascular injury from risk factors and to the development of atherosclerosis.^44–46 The atheroprotective properties of CPCs is compromised by increasing age. In fact, treatment with CPCs isolated from young mice prevents atherosclerosis progression in ApoE^−/− mice, whereas treatment with progenitor cells from older mice is ineffective.⁴⁵ We and others also have demonstrated a lower frequency of circulating CD34+ cells in humans with aging.^29,47 In the present study, we demonstrate that mobilization of CPCs in the presence of, and in response to, risk factors is observed only among younger subjects, whereas among older individuals there no increase in CPCs with increasing risk factor burden. We further show that higher CPC counts at baseline protect individuals against further deterioration of vascular function, whereas those with low CPC counts show further decline in their vascular function over time. These findings suggest that there is an age-related exhaustion of CPC mobilization which plays a critical role in vascular aging and development of vascular disease.

With increasing age and exposure to risk factors, degenerative changes in the extracellular matrix of the vascular medial and adventitial layers results in greater arterial stiffness.⁴⁸ This in turn leads to the development of left ventricular hypertrophy, diastolic dysfunction, and eventually results in greater cardiovascular morbidity and mortality.^2,3,49–69 Traditional and metabolic cardiovascular risk factors have also been shown to be associated with impairment in microvascular function measured as RHI.^6,70,71 Lower RHI correlates directly with coronary microvascular dysfunction in patients with early atherosclerosis,¹ and is predictive of adverse cardiovascular events.⁷ Our results demonstrate that in addition to traditional risk factors, CPC counts are independent determinants of both arterial stiffness and microvascular function, with higher levels of CPCs being associated with protection against further vascular deterioration.

Numerous studies have shown the prognostic value of CPCs, with lower CPC counts indicating worse cardiovascular outcomes in a variety of clinical scenarios.^35,72–76 However, the mechanisms of these protective effects remain unclear. Our findings indicate that a lower CPC count is associated with worsening of large arterial and microvascular function over time. Together with the previous experimental data,⁴⁵ these findings suggest that there is a causal relationship between vascular health and CPCs, and also explains the reason why those with lower CPC counts are at higher risk of future cardiovascular events.

Study strengths and limitations

The major strength of our study includes the large number of male, female, and ethnically diverse participants with broad cardiovascular risk profiles who were studied and had detailed phenotyping of several CPC subsets and vascular function. Our study was limited by the single-center nature of the study and the fact that most participants had few cardiovascular risk factors. Finally, our cut-off age of 48 was selected based on the median age of the cohort and not a priori and therefore should be considered as hypothesis-generating.

Conclusion

In conclusion, we found that among younger individuals, increasing risk factor burden is associated with worsening of vascular function and higher levels of circulating CPCs. This is not observed among older individuals, suggesting that older age is associated with a failure of stimulation of CPCs in the presence of risk factors and this may result in worse vascular function than in the younger subjects with similar risk factor burden. A low CPC count predicted worsening of vascular dysfunction during follow-up, a change that was not observed in those with higher CPC counts. These findings suggest that exhaustion in the endogenous regenerative capacity with aging may be precipitated by extended exposure to risk factors, resulting in worsening of vascular function and adverse outcomes. Novel cell-based therapeutic approaches that restore or stimulate CPCs in older individuals may potentially prevent age-related vascular dysfunction and atherosclerosis progression.