Transcriptional Regulation of Bim by FOXO3a and Akt Mediates Scleroderma Serum-Induced Apoptosis in Endothelial Progenitor Cells

Abstract

Background

Endothelial progenitor cells (EPCs) contribute to vascular regeneration/repair, and may thus protect against scleroderma vasculopathy. We aimed to determine (i) whether circulating EPCs were reduced in scleroderma; and (ii) whether scleroderma sera could induce EPC apoptosis and, if so, what the underlying apoptotic signaling pathway was.

METHODS and RESULTS

Circulating EPC levels were quantified in 54 patients with scleroderma and 18 healthy controls by colony forming unit assay and flow cytometry, which revealed markedly decreased EPC levels in scleroderma patients relative to healthy subjects. Substantial apoptosis was detected in EPCs after culturing in the presence of scleroderma sera, compared to normal sera. Intriguingly, depletion of IgG fraction from SSc sera completely abolished the apoptotic effects. Furthermore, SSc sera inhibited the activation/phosphorylation of Akt, which in turn suppressed the phosphorylation and degradation of forkhead transcription factor FKHRL1 (FOXO3a), resulting in the upregulation of apoptotic protein Bim. siRNA mediated FOXO3a and Bim knockdown substantially reduced scleroderma serum-induced EPC apoptosis. Importantly, Bim expression and baseline apoptosis were increased in EPCs freshly isolated from SSc patients relative to that obtained from healthy subjects.

Conclusions

Scleroderma serum-induced EPC apoptosis is mediated chiefly by the Akt-FOXO3a-Bim pathway, which may account, at least in part, for the decreased circulating EPC levels in SSc patients.

INTRODUCTION

Systemic sclerosis (SSc) or scleroderma is a chronic, multisystem connective tissue disease affecting the skin and various internal organs ¹ . The disease is characterized by the triad of vascular damage, fibrosis, and autoimmunity ² . Although the relationship between these three pathologic features is not fully understood, the observation that Raynaud’s phenomenon and other disturbances in the peripheral vascular system precede the onset of fibrosis in SSc ³ , raises the possibility that fibrosis in SSc may represent a default pathway resulting from vascular failure. Hence, understanding the pathobiology of SSc vascular disease is the key in dissecting the disease pathogenesis and developing novel therapeutic strategies.

The circulatory damage in SSc primarily affects the microvasculature and small and medium sized arteries, resulting in chronic underperfusion and ischemia in affected skin and internal organs ^2,3 . These changes were expected to trigger an extensive compensatory angiogenic response. Numerous studies, however, indicate that angiogenesis—a process mediated chiefly by endothelial progenitor cells (EPCs) derived from the bone marrow—is impaired in SSc ² , despite elevated levels of vascular endothelial growth factor (VEGF), a potent angiogenic factor, in the peripheral blood ⁴ . Any defect in the EPC supply within and mobilization from the bone marrow or excessive destruction of EPCs upon their mobilization by immune system could potentially impede vascular regeneration/repair. We recently demonstrated that transplantation of SSc skin into severe combined immunodeficient (SCID) mice elicited robust angiogenesis—an effect that was significantly stronger than normal skin grafting—and that the neovascular cells were almost exclusively derived from the recipients, perhaps originating from the bone marrow⁵. These data indicate that SSc skin is capable of releasing factors that mobilize EPCs from the bone marrow, and narrow the possibilities for angiogenesis/vascular repair defects in SSc down to EPC depletion within the bone marrow and/or EPC destruction in the circulation.

FOXO3a is a member of FOXO subfamily of Forkhead transcription factors, which regulate cell survival and apoptosis ⁶ . In the presence of survival factors, Akt, the survival kinase, is phosphorylated and activated, which in turn phosphorylates FOXO3a, leading to association with 14-3-3 proteins, nuclear exclusion, and retention and degradation of FOXO3a in the cytoplasm. Conversely, the presence of apoptosis factors inhibits Akt phosphorylation, which leads to FOXO3a dephosphorylation, nuclear translocation, and activation of FOXO target genes, including the proapoptotic genes, Bim and Puma, resulting in cell apoptosis ⁶ .

Contradictory results have been reported regarding the levels of circulating EPCs in SSc⁷ ,⁸ . Although SSc sera are capable of inducing mature endothelial cell (EC) apoptosis, little is known regarding the effects of SSc sera on EPCs. In this study, we examined the levels of circulating EPCs in five cohorts of patients: 1) early stage diffuse cutaneous SSc (dcSSc), 2) intermediate/late stage dcSSc, 3) early stage limited cutaneous SSc (lcSSc), 4) intermediate/late stage lcSSc, and 5) healthy controls. We found significant reduction of circulating EPC levels in all disease phenotypes compared to healthy subjects. We provide evidence that SSc, but not normal, sera induce EPC apoptosis via the Akt-FOXO3a-Bim pathway, which may account, at least in part, for the reduction of circulating EPCs in SSc.

PATIENTS and METHODS

The authors had full access to the data and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written. The following provides description for each of the methods used in the study. Detailed description is found in the online supplemental information.

Subject Characteristics

Fifty-four patients [14 men and 40 women, aged 56.9 ± 12.6 years, median disease duration of 8.5 (0.17-40) years with SSc as defined by the American College of Rheumatology SSc classification criteria ⁹ presenting to Duke Hospital for treatment were included in this study. Patients with overlap connective tissue syndromes, previous myocardial infarction, stroke, valvular or congenital heart disease, congestive heart failure, diabetes mellitus, hypertrophic cardiomyopathy, dyslipidemia, history of smoking, or history of arterial hypertension before the diagnosis of SSc were excluded. Patients were classified into four groups (Supplemental Table 1), based on SSc disease type (dcSSc or lcSSc) and duration (from the onset of the first non-Raynaud’s symptom of scleroderma). They are: 1) 12 patients with early stage lcSSc—sclerosis confined to areas distal to the elbows or knees and above the clavicles (disease duration <5 years); 2) 13 patients with intermediate/late stage lcSSc (disease duration ≥5 years); 3) 10 patients with early stage dcSSc—sclerosis extending proximal to the elbows or knees with or without truncal involvement (disease duration <3 years); and 4) 19 patients with intermediate/late stage dcSSc (disease duration ≥3 years). Disease stages were defined as suggested by LeRoy EC et al ¹⁰ . Eighteen health subjects undergoing diagnostic cardiac catheterization at Duke University Medical Center showing no evidence of coronary artery diseases were used as controls in the study.

Peripheral Blood Collection

After obtaining informed consent from subjects, an average of 30 ml peripheral blood was collected from each patient.

Colony Forming Unit Assay

Colony forming assay was performed as described by Hill et al ¹¹ .

Flow Cytometry Analysis

Mononuclear cells were isolated by density centrifugation as described in the online supporting material for Colony Forming Unit Assay.

Immnoaffinity Depletion of IgG

Sera were heated for 30 minutes at 56°C before protein G adsorption.

Isolation of CD133⁺ EPCs and Cell Culture

Human umbilical cord blood cells (UCB) were obtained from the Carolina Cord Blood Bank under protocols approved by the Institutional Review Board at Duke University Medical Center.

Detection and Quantification of Apoptosis

CD133⁺ EPCs isolated from UCB or human bone marrow were incubated with 10% sera from SSc patients, or normal healthy subjects, or IgG-depleted SSc and normal sera for 48 hours.

RNA Interference

To silence FOXO3a and Bim gene expression, we performed transfection of siRNA duplex using Lipofectamine™ and PLUS™ (invitrogen).

TaqMan Real-Time RT-PCR (TRT-PCR)

RNA was isolated from UCB-derived CD133⁺ EPCs treated with sera for 48 hours using RNeasy Mini kit (Qiagen).

Immunoblotting Analysis

Cell extracts were separated by SDS-PAGE in non-reducing conditions. Proteins on the gel were then transferred to a nitrocellulose membrane.

Data Analysis

Data are expressed as mean ± SEM of three separate triplicate culture experiments unless otherwise indicated. Multiple group comparison was done by one-way ANOVA for data with normal distribution and by Kruskal-Wallis test for data with non-normal distribution. Post hoc test was done with the Dunnett’s t test for data with normal distribution or Mann-Whitney U test for data with non-normal distribution, if the p value for the overall comparison was significant (<0.05). Two group comparisons were performed with Student’s t test for data with normal distribution. A multinomial logistic regression analysis was performed to determine the association between EPC counts and each variable. Analyses were performed in SPSS for Windows version 16.0 (SPSS, Chicago, IL).

RESULTS

Patient Characteristics

The patient characteristics are presented in Supplemental Table 1. The mean age of the lcSSc cohort tended to be higher than that of the dcSSc cohort. More than 60% of subjects were female in all SSc patient groups. Factors that may impact on EPCs ¹¹ , including pulmonary hypertension, statin therapy, and corticosteroid therapy, were recorded. These factors were taken into account in the analyses and interpretation of the EPC data.

Reduced Circulating EPCs in SSc patients

We first used colony forming unit assay ¹¹ —an established method at the time when this study was conducted—to measure circulating EPCs. We now know that EPCs detected by this method correspond to early EPCs ¹² . We nevertheless were able to determine the endothelial nature of the cells within the colony. Indeed, these cells stained positive with multiple endothelial-specific markers, in particular VWF and VEGFR2 (Figure 1). The number of circulating early EPCs was dramatically reduced in lcSSc group (n = 26) and dcSSc group (n = 29) as compared with healthy control group (1.1 ± 0.45 for lcSSc and 1.5 ± 0.46 for dcSSc versus 26.9 ± 2.2 for the control group, p = 0.0001), when both early and intermediate/late stage diseases were analyzed as a whole (Figure 1). The difference between lcSSc and dcSSc cohorts was not statistically significant (p = 0.863). It was previously shown that EPC level was higher in early stage than in late stage SSc patients ⁸ . To further study the changes of early EPC levels relative to SSc disease duration, we stratified and analyzed our data obtained in the substantial number of patients in each of the subgroups. Interestingly, there was no significant difference in the levels of circulating early EPCs in early versus intermediate/late lcSSc (p = 0.25), or early versus intermediate/late dcSSc (p = 0.88). All SSc subgroups had lower early EPC levels than healthy subjects (Figure 1G). We also analyzed the complicating effects of pulmonary arterial hypertension (PAP), statin and corticosteroid therapies on early EPC levels. dcSSc patients with PAP (n = 17) had lower early EPCs levels than the dcSSc patients without PAP (n = 12, 0.75 ± 0.28 versus 3.04 ± 0.97, p = 0.041). The presence of PAP did not affect early EPC levels in lcSSc patients. No significant effects of gender, statin and corticosteroid therapies on circulating early EPC levels were observed by multinomial logistic regression analysis.

Figure 1. Decreased Early EPCs in SSc patients by CFU measurements.

The CFU is characterized as a central cluster of round cells surrounded by spindle-shaped cells radiating from the periphery (A). Cells from the colony stain positive for VEGFR2 (red, B), and vWF (green, C), and DAPI nuclear stain (blue, D); Overlay of the three colors (yellow, E) confirms the endothelial nature of these cells. The number of early EPCs is substantially and significantly decreased in lcSSc ( ) and dcSSc ( ) patients as compared with healthy ( ) subjects (F). The overall group comparison was assessed by one-way ANOVA. Dunnett’s test was performed for the comparisons between disease groups and healthy controls. **p≤0.0001. There are no differences between early (∎) and intermediate/late stage () within SSc subgroups (G).

Since there is controversy regarding the definition and enumeration of circulating EPCs, in particular with the use of colony forming assay, to confirm the above findings and to provide a robust analysis of EPCs that may be phenotypically different from the early EPCs identified by colony forming assay, flow cytometry was used to determine the number of EPCs co-expressing CD133/CD34, CD133/VEGFR2 or CD133/CD34/ VEGFR2 in the peripheral blood as shown in Supplemental Figure 1. Circulating CD133⁺/CD34⁺ , CD133⁺/VEGFR2⁺ and CD133⁺/CD34⁺/VEGFR2⁺ cell levels were significantly lower in the lcSSc and dcSSc patients, compared to the healthy control subjects (Figure 2). There was no significant difference between lcSSc and dcSSc cohorts. The difference between early and intermediate/late lcSSc and between early dcSSc and intermediate/late dcSSc was not statistically significant. Moreover, there was no significant difference for the various EPC phenotypes among the different SSc patient cohorts (data not shown). The presence or absence of PAP, statin therapy and corticosteroid therapy did not affect the levels of any of the EPC phenotypes (data not shown). These data indicate that decreased circulating EPCs may be an integral part of the SSc vascular disease that often precedes the onset of fibrosis, thus implicating a potential role of EPC reduction in the pathogenesis of SSc—a notion that is consistent with the prevailing view in the field ¹³ .

Figure 2. Decreased EPCs in SSc patients by FACS analysis.

The levels of circulating EPCs, as defined by CD133⁺ /CD34⁺ cells ( ), CD133⁺ /VEGFR2⁺ cells ( ), or CD133⁺/CD34⁺/VEGFR2⁺ ( ) cells, are substantially decreased in lcSSc and dcSSc patients compared to control group. There is no significant difference for various EPC phenotypes between the two patient groups. Significance was determined by one-way ANOVA followed by Dunnett’s test to compare the level of EPCs in different disease groups versus the control group. *p<0.05, **p≤0.0001.

Induction of Apoptosis in EPCs by SSc Sera

To determine if excessive EPC destruction by autoantibodies present in the sera contributed to the reduction of circulating EPCs in SSc patients, we examined the apoptotic effects of SSc sera on EPCs in vitro. CD133⁺ EPCs isolated from human UCB or bone marrow were cultured in the presence of each serum samples from lcSSc (n = 5), dcSSc (n = 5) patients, and healthy control (n = 5) for 48 hours. At the concentration of 10%, all lcSSc and dcSSc samples showed rather consistent pro-apoptotic effects, whereas serum samples from healthy subjects had no effects on EPC apoptosis, relative to FBS. Hence, in the following experiments, pooled sera from 5 lcSSc and 5 dcSSc samples (SSc sera) and pooled sera from 5 health subjects (normal sera) were used. SSc sera dose-dependently induced UCB-derived EPC apoptosis as detected by DNA fragmentation, whereas normal sera had no apoptotic effects (Figure 3A). A similar degree of apoptosis was observed in bone marrow-derived EPCs when treated with SSc sera, but not with normal sera (data not shown). Thus, in the following experiments, we used exclusively EPCs isolated from UCB, because of its ready availability to us. The apoptosis findings in UCB-derived EPCs were further confirmed by TUNEL and caspase 3/7 assay, which showed that substantially more EPCs underwent apoptosis in the presence of 10% SSc sera than in normal sera. Remarkably, depletion of IgG from SSc sera almost completely abolished the apoptosis-inducing effect of SSc sera, whereas no difference was observed with normal sera with and without IgG depletion (Figure 3B). The IgG depletion involved two steps, heat inactivation at 56°C for 30 minutes and protein G binding. To exclude the possibility that inactivation of complements, which are sensitive to heat treatment, might account for the lack of apoptosis-inducing effects observed with IgG-depleted SSc sera, we treated EPCs with SSc and normal sera with heat inactivation, but without IgG depletion. As shown in Figure 3B, heat inactivation alone did not change the effects of SSc and normal sera on EPC apoptosis. These observations exclude the involvement of complements in SSc serum-induced EPC apoptosis and indicate that EPCs may be destructed upon their mobilization from the bone marrow by heat-resistant factors, including autoantibodies, present in the IgG fraction of SSc sera.

Figure 3. SSc sera induce apoptosis in EPCs.

SSc sera dose-dependently induce UCB-derived EPCs to undergo apoptosis, as determined by DNA fragmentation ELISA assay (A). The apoptosis effect of 10% SSc sera versus normal sera or FBS is confirmed by caspase 3/7 assay; remarkably, depletion of IgG from SSc sera, but not heat inactivation, abolishes the apoptosis-inducing effect of SSc sera (B). Furthermore, EPCs are 5.5 fold more sensitive to SSc serum-induced apoptosis than HMVECs (B). TUNEL labeling also shows substantial EPC apoptosis in the presence of 10% SSc sera: Nuclear staining by Hoechst 3342 (C, F), TUNEL staining (D, G), and overlay (E, H). Significance was determined by one-way ANOVA followed by Dunnett’s test to compare the caspase activity in different conditions versus the normal serum treatment. *p<0.05, **p≤0.0001.

EPCs Are More Susceptible Than Mature Endothelial Cells to SSc Sera

Anti-EC antibodies, found in the serum of SSc patients, are known to induce EC apoptosis both in vitro and in vivo² . Hence, it has been assumed that increased endothelial apoptosis is the earliest microvascular abnormality in SSc that precedes fibrotic changes. Yet, the deficiency of EPC-mediated repair and its contribution to SSc vascular lesion formation relative to endothelial injury has not been studied before. To address this important question, we treated human microvascular endothelial cells (HMVECs) with sera obtained from lcSSc and dcSSc patients and compared with EPCs treated in the same fashion. As expected, SSc sera induced apoptosis in a significant number of HMVECs, compared with normal sera or FBS. Intriguingly, ECs were 5.5 fold less susceptible than UCB-derived EPCs to the toxic factors present in the SSc sera (Figure 3B). These data suggest that the defective repair, mediated at least in part by serum-induced circulating EPC apoptosis, might be more important than vascular damage in SSc vascular lesion formation.

SSc Sera Regulate Akt-FOXO3a-Bim Pathway in EPCs

FOXO transcription factors are phosphorylated by Akt which leads to cytoplasmic retention and impairment of FOXO nuclear transcriptional activity ⁶ . FOXO3a has been shown to mediate apoptosis in glioma ¹⁴ , and FOXO4 was involved in H₂O₂ -induced apoptosis in EPCs isolated from peripheral blood through the induction of proapoptotic protein Bim ¹⁵ . To determine which FOXO proteins in UCB-derived EPCs were most affected by SSc sera, we performed Western blotting for FOXO1, FOXO3a, and FOXO4 in cells treated with SSc and normal sera. Endogenous levels of all three FOXO proteins were detected in cells cultured in FBS and normal sera. The expression of FOXO3a was markedly increased in cells treated with SSc sera, whereas the expression of FOXO1 and FOXO4 was essentially unchanged (Figure 4), pointing to a potential role of FOXO3a in regulating EPC apoptosis. Hence, in the following experiments, we focused on FOXO3a.

Figure 4. SSc sera induce FOXO3a expression in EPCs.

Endogenous levels of FOXO1, FOXO3a and FOXO4 expression are detected in UCB-derived EPCs. SSc sera markedly induce the expression of FOXO3a, but not FOXO1 and FOXO4.

We examined whether SSc sera with and without IgG depletion affected the phosphorylation and expression of Akt and FOXO3a. As shown in Figure 5, Akt and FOXO3a were constitutively phosphorylated in EPCs treated with normal or IgG-depleted SSc sera. Treatment of these cells with SSc sera resulted in dephosphorylation of both Akt and FOXO3a. SSc serum treatment also increased the level of total FOXO3a, but did not affect the expression level of total Akt in EPCs. In contrast, normal sera with and without IgG depletion, IgG-depleted SSc sera and FBS had no effects on total Akt and FOXO3a expression (Figure 5).

Figure 5. SSc sera inactivate Akt, leading to the activation of FOXO3a and Bim expression.

Western blotting was performed for phospho-FOXO3a (p-FOXO3a), total FOXO3a, phospho-Akt (p-Akt), total Akt, Bim, Puma and β-actin. SSc sera inhibit the activation and phosphorylation of Akt, which leads to the dephosphorylation and activation of FOXO3a, resulting in upregulation of Bim and, to a much lesser degree, Puma. IgG depletion abolishes the effects of SSc sera on Akt, FOXO3a, Bim, and Puma to a degree similar to that observed with normal sera and FBS.

We then examined the factors downstream of Akt-FOXO3a signaling, namely Bim and Puma. The expression of Bim protein was significantly increased in EPCs treated with SSc sera, as compared with normal sera. The inducing effect of SSc sera on Bim expression was almost completely abolished by IgG depletion. A similar but much less dramatic effect was observed in Puma expression in response to SSc and normal serum treatment (Figure 5). These changes were further confirmed by real time RT-PCR. The mRNA levels for Bim and, to a much lesser degree, for Puma were higher in EPCs treated with SSc sera than that observed in the presence of normal sera, or IgG depleted SSc sera (Figure 6). Importantly, the level of Akt dephosphorylation correlated with that of FOXO3a dephosphorylation, and the induction of Bim expression. These data indicate that factors present in SSc sera, particularly in the IgG fraction, may inactivate Akt, which, in turn, dephosphorylates and activates FOXO3a, resulting in the induction of proapoptotic proteins Bim and, to a lesser degree, Puma. Collectively, these findings suggest that SSc serum-induced EPC apoptosis might be mediated by the Akt-FOXO3a-Bim pathway.

Figure 6. SSc Sera Induce Bim Expression at the mRNA Level.

TaqMan real-time RT-PCR was performed for Bim (A) and Puma (B). SSc serum treatment results in upregulation of Bim and, to a much lesser degree, Puma. IgG depletion abolishes the effect of SSc sera on Bim and Puma expression. Significance was determined by one-way ANOVA followed by Dunnett’s test to compare Bim/Puma expression in different conditions versus the normal serum treatment. *p<0.05, **p≤0.0001.

FOXO3a Silencing Reduces SSc Serum-induced Bim Expression and Apoptosis in EPCs

To confirm the role of Akt-FOXO3a-Bim pathway in SSc serum-induced EPC apoptosis, we performed gene-silencing experiments. EPCs were transfected with a mixture of 3 sets of siRNA duplexes specific for FOXO3a or a scrambled siRNA control (Figure 7A, B). The cells were then cultured in the presence of SSc sera, normal sera, or FBS. FOXO3a siRNA markedly reduced endogenous FOXO3a expression (Figure 7C). Furthermore, SSc sera failed to induce the expression of total FOXO3a and Bim expression in FOXO3a siRNA-transfected cells, as compared with scrambled siRNA transfection (Figure 7C, D). By contrast, the effect of FOXO3a knockdown on the endogenous expression and SSc serum-induced induction of Puma was, if anything, trivial (data not shown). We then determined whether the reduction in FOXO3a and Bim levels resulting from FOXO3a siRNA treatment could inhibit SSc serum-induced apoptosis in EPCs. We repeated the siRNA knockdown experiments and measured DNA fragmentation using ELISA. As expected, apoptosis in FOXO3a siRNA-transfected EPCs in the presence of SSc sera was suppressed to a level similar to that observed in cells treated with FBS and normal sera (Figure 7E). These data indicate that FOXO3a is crucial in the activation of Bim expression and the induction of EPC apoptosis.

Figure 7. siRNA-Mediated FOXO3a knockdown Reduces SSc Serum-Induced Bim Expression and Apoptosis in EPCs.

EPCs were transfected with three FOXO3a–specific siRNAs (50nM) or a scrambled siRNA (50nM) control: FITC-labeled siRNA (green, A), and overlay of siRNA and nuclear staining (red) by propidium iodide (B). FOXO3a siRNAs (siFOXO3a), but not scrambled siRNA control (siControl), markedly suppress endogenous and SSc serum-induced FOXO3a and Bim protein expression at 48 hours (C). Bim expression at mRNA level is also reduced by FOXO3a knockdown (D). FOXO3a siRNAs protect SSc serum-induced EPC apoptosis, whereas control siRNA has no effects (E). Significance was tested by one-way ANOVA followed by Dunnett’s test to compare the level of Bim expression or apoptosis in different culture conditions versus the control siRNA in the presence of SSc sera. *p<0.05, **p<0.0001.

Bim Silencing Reduces SSc Serum-induced EPC Apoptosis

To further confirm that Bim serves as the important factor downstream of Akt-FOXO3a signaling in mediating SSc serum–induced EPC apoptosis, CD133⁺ EPCs were transfected with Bim-specific or control scrambled siRNA. As shown in Figure 8 A, transfection with Bim siRNA decreased Bim mRNA expression by 65% as compared with control siRNA. The level of apoptosis in Bim siRNA-transfected EPCs was approximately 1/3 of that seen in scrambled siRNA-transfected EPCs in the presence of SSc sera (Figure 8B). Thus, Bim siRNA transfection in EPCs resulted in a similar degree (about 2/3) of inhibition in Bim expression and apoptosis. These findings highlight the critical role of Bim in mediating SSc serum-induced EPC apoptosis. Our data, however, cannot exclude the potential minor contributions by other proapoptotic pathways, shown to be regulated by FOXO3a, including JNK activation and FLIP downregulation, to EPC apoptosis induced by SSc sera ¹⁶ .

Figure 8. Bim is critical in regulating EPC apoptosis.

Bim siRNA (siBim), but not control siRNA (siControl), transfection reduces Bim expression (A) and inhibits SSc serum-induced EPC apoptosis (B). CD133⁺ EPCs isolated from lcSSc and dcSSc patients show increased endogenous Bim expression (C) and enhanced baseline apoptosis (D) relative to those obtained from healthy subjects. Values were obtained from experiments in triplicates and repeated at least twice. Significance was tested by one-way ANOVA followed by Dunnett’s test or unpaired student’s t test to compare the levels of Bim expression and apoptosis in SSc patients versus controls. *p<0.05, **p≤0.0001.

Increased Bim Expression and Basline Apoptosis in SSc Circulating EPCs

Since treatment with SSc sera induced the dephosphorylation of Akt and activation of FOXO3a, resulting in the upregulation of Bim expression and apoptosis in UCB-derived EPCs, we reasoned that circulating EPCs derived from SSc patients should have increased Bim expression and baseline apoptosis, due to constant exposure to SSc sera. To test this possibility and to establish the missing link between the activation of Akt-FOXO3a-Bim pathway in vitro and the decreased EPC levels in SSc patients in vivo, we measured the expression levels of Bim expression in CD133⁺ EPCs isolated from dcSSc and lcSSc patients relative to healthy subjects using real-time RT-PCR. Due to the low frequency of CD133⁺ EPCs in the peripheral blood, we pooled the cells isolated from the peripheral blood of 9 dcSSc and 9 lcSSc patients and 9 healthy subjects into 3 tubes per group. As shown in Figure 8C, Bim expression was significantly and substantially increased in EPCs obtained from dcSSc and lcSSc patients, as compared with those isolated from healthy subjects. Remarkably, freshly isolated CD133⁺ EPCs from SSc patients showed increased baseline apoptosis as compared with normal EPCs when cultured in EGM containing 10% FBS for 18 hours (Figure 8D). These data indicate that data generated from our in vitro study are highly relevant to the pathophysiological changes in vivo and support the notion that Akt-FOXO3a-Bim axis plays a central role in the increased EPC apoptosis in SSc patients.

DISCUSSION

Several lines of evidence indicate that progressive vascular disorder is a primary event in the pathogenesis of SSc ² . Indeed, the observation that Raynaud’s phenomenon precedes the diagnosis of SSc by months or years has prompted the speculation that fibrosis in SSc may represent a default pathway resulting from vascular failure ³ . Previous studies related to the pathogenesis of SSc vascular disease have focused almost exclusively on vascular injury. Specifically, endothelial apoptosis is thought to be an initiating event in SSc vascular lesion formation ¹⁷ . Since the identification of circulating EPCs in 1997 ¹⁸ , the field of vascular biology has experienced a significant paradigm shift with the introduction of EPC-induced vascular repair. Indeed, the availability of progenitor cells is closely linked to the initiation and development of several vascular disorders, including atherosclerosis ¹⁹ . Thus, characterizing the vascular repair aspect, namely EPCs, in SSc would likely shed light on the understanding of the pathogenesis of SSc vascular lesions. In this report, we provide data demonstrating substantial depletion of circulating EPCs in SSc patients, regardless of the disease subtypes and the different methodologies used to measure EPCs.

Conflicting data have been reported regarding the levels of EPCs in SSc patients. Kuwana et al ⁷ , demonstrated a reduced number of CD34⁺/CD133⁺/VEGFR2⁺ EPCs with impaired function in SSc patients. Their findings are consistent with our observations. Very recently, Avouac et al ²⁰ , however, showed that the level of the same type of EPCs was increased in SSc patients, relative to healthy subjects. Although all these studies used flow cytometry to detect EPCs, different gating may have contributed to the discrepancy. Furthermore, there appear to exist two different types of EPCs—early and late EPCs—according to their time-dependent appearance, when they are isolated using cell culture-based methods ¹² . The two types of EPC have different morphology, gene expression profiles, and survival behaviors. Both types of EPCs, however, contribute to neovasculogenesis in vivo; early EPCs mainly secrete angiogenic cytokines, whereas late EPCs supply a sufficient number of endothelial cells ²¹ . Interestingly, we demonstrated that early EPCs were markedly decreased in SSc patients. In contrast, Avouac et al ²⁰ found that the levels of late EPCs were not different between SSc patients and healthy subjects. The pathogenetic significance of the decreased early EPCs versus unchanged late EPCs remains to be determined in SSc.

The initial stages of SSc are generally not accessible for analysis in man. Hence, we classified the dcSSc and lcSSc disease cohorts into early and intermediate/late stages, hoping that the early stage could reflect some of the pathobiological changes occurring in the initial stages of SSc. In contrast to the findings made by Del Papa et al ⁸ showing that CD45⁻/CD133⁺ EPC levels were decreased in intermediate/late stage but increased in early stage SSc patients, we found that EPCs were substantially and consistently decreased in both early and intermediate/late stage SSc patients. These data, coupled with the observation that bone marrow precursors play an important role in inducing angiogenesis in SSc skin grafts in SCID mice ⁵ and numerous reports supporting the critical role of EPCs in the initiation of other vascular diseases, such as atherosclerosis, indicate that the lack of circulating EPCs may also be important in the initiation of SSc vascular lesion formation. The discrepancy between our data and that of Del Papa et al ⁸ may be due to differences in EPC definition and measurements employed in these two studies. Specifically, EPCs were defined as CD45⁻/CD133⁺ cells in the Del Papa paper, whereas several EPC phenotypes, including CD133⁺/CD34⁺, CD133⁺/VEGFR2⁺, and CD133⁺/CD34⁺/VEGFR2⁺ cells and those determined by the colony-forming assay, were investigated in our study.

Endothelial cell apoptosis induced by anti-endothelial cell antibodies (AECA) is thought to be one of the earliest steps in the vascular pathology of SSc ²² . The integrity of the endothelium is assured by effective EC repair under steady state conditions, which is mediated by the proliferation of adjacent ECs and, more importantly, by circulating EPCs. It has been documented that, in the setting of atherosclerosis, disease risk factors diminish the supply of EPCs needed to maintain the homeostasis of the cardiovascular system, tilting the balance of vascular injury and repair in favor of injury and atherosclerosis progress ¹⁹ . To determine if autoimmunity would cause EPCs to undergo apoptosis, resulting in impaired repair capacity in SSc, we examined whether EPCs were sensitive to autoantibody-induced cell death. Indeed, using multiple assays, we demonstrated, for the first time, substantial EPC apoptosis when the cells were treated with sera from patients with SSc. Furthermore, depletion of IgG from SSc sera abolished EPC apoptosis, suggesting the involvement of autoantibodies in this process. These findings indicate that SSc serum-induced EPC apoptosis may serve as a major putative mechanism for the decreased circulating EPC levels in SSc. In addition to inducing apoptosis, SSc sera may also exert inhibitory effects on EPC proliferation. Given the greater sensitivity of EPCs than ECs to SSc sera, it seems that defective repair may be more important than vascular damage in SSc vascular lesion formation,

Little is known about the molecular signaling pathway underlying EPC apoptosis, particularly in response to SSc serum exposure. It has been shown that FOXO proteins are important regulators for the fate of hematopoietic stem cells ⁶ . The Akt-FOXO3a-Bim axis has been implicated in ² glioma cell apoptosis ¹⁴ . Furthermore, FOXO4-dependent expression of Bim has been shown to mediate reactive oxygen species-induced EPC apoptosis ¹⁵ . Studies in other cell systems showed that FOXO3a could directly modulate Bim expression and induce apoptosis ²³ . We surveyed FOXO1, FOXO3a and FOXO4, and found that, although all three FOXO proteins were present in UCB-derived EPCs, only FOXO3a was affected by SSc sera. Consistent with previous observations showing the cause-and-effect relationship between Akt and FOXO3a, we showed that both Akt and FOXO3a were dephosphorylated in EPCs in the presence of SSc sera. Furthermore, SSc serum treatment upregulated the expression level of FOXO3a, Bim, and induced EPC apoptosis. Intriguingly, IgG depletion from SSc sera abolished these effects. These data indicate that Akt-FOXO3a-Bim axis plays an important role in mediating SSc serum-induced EPC apoptosis.

To further establish the central role of FOXO3a and Bim in mediating SSc serum-induced EPC apoptosis, we performed siRNA knockdown experiments. As expected, transfection of EPCs with FOXO3a-specific siRNA silenced FOXO3a expression and abolished SSc serum-induced up-regulation of Bim. Furthermore, the silencing of either FOXO3a or Bim substantially reduced SSc serum-induced EPC apoptosis. Importantly, EPCs isolated from SSc patients display increased Bim expression and baseline apoptosis. Taken together, our data support the notion that SSc serum-induced EPC apoptosis via the Akt-FOXO3a-Bim signaling pathway may be responsible for reduced EPC levels in SSc.

Supplementary Material

01

Supplemental Figure 1 Flow cytometry gating and analysis for peripheral blood mononuclear cells stained (PBMCs) with APC-CD133, PE-Cy7-CD34, PE-VEGFR2 and 7AAD.

Forward scatter/side scatter (FSC/SSC) flow cytometric plot of the PBMCs is shown in (A). Viable cells are identified by gating on FSC/SSC plot as the 7AAD-negative population (B). CD133 positive cells (upper portion) are gated on viable cells (C). EPCs are identified by double positivity for CD34/CD133 (upper right quadrant, D). Gating and analysis for CD133/CD34/VEGFR2 triple positive cells are shown in (E). Triple positive cells are located in the upper right quadrant.

Supplemental Table1. Characteristics of Patients

Supplemental Table 2. SiRNA oligonucleotides

Clinical Perspective.

The etiology of systemic sclerosis (SSc) is still not completely understood, but appears to be autoimmune. Immune activation targets mature endothelial cells (ECs), resulting in vascular integrity breakdown and ensuing fibrosis. Endothelial progenitor cells (EPCs) are bone marrow–derived nonleukocyte cells that participate in vascular repair and homeostasis. It has been shown that injury of ECs, not only induces a cascade of proinflammatory events, contributing to vascular lesion formation, but also stimulates the mobilization of EPCs from the bone marrow, mediating vascular repair. Hence, the availability of circulating EPCs plays a critical role in maintaining the integrity and functional activity of the endothelial monolayer and in vasculogenesis. Several lines of evidence indicate that increased demand for vascular repair in the context of repeated injury could exhaust the supply of EPCs in the bone marrow, interrupting the balance between vascular repair and injury. In this report, we provide evidence showing, for the first time, that the factors in the peripheral blood of SSc patients that cause EC injury may also damage EPCs. Moreover, probably owing to the lack of protective mechanisms in these immature cells, EPCs are more sensitive to the toxic factors than ECs, implicating excessive EPC destruction in the pathogenesis of SSc. Importantly, we have identified Akt - FOXO3a - Bim pathway to mediate EPC apoptosis. Although much work needs to be done to determine the exact factors in the SSc serum causing EPC apoptosis, targeting the Akt - FOXO3a - Bim pathway may be considered as a venue for future therapies.