Correlation of Different Circulating Endothelial Progenitor Cells to Stages of Diabetic Retinopathy: First In Vivo Data

Abstract

purpose. To investigate vasculogenic circulating progenitor cells (CPCs), endothelial progenitor cells (EPCs), and mature EPCs in patients with type 1 diabetes mellitus (T1DM) with or without diabetic retinopathy (DR).

methods. A case-control study comparing 90 patients with T1DM with and without DR was performed. Patients were studied and staged for retinopathy according to the Early Treatment of Diabetic Retinopathy Study (ETDRS) classification. Ninety patients were included: 30 without DR (control [CO]), 30 with mild nonproliferative DR (mNPDR), 10 with moderate-severe NPDR (msNPDR), 10 with mild-moderate proliferative diabetic retinopathy (mmPDR), and 10 with high-risk PDR (hrPDR). CPCs (CD34/CD133), EPCs (CD34/CD133/CD309), and mature EPCs (CD34/CD133/CD309/CD31) were enumerated by flow cytometry.

results. EPCs were reduced in mNPDR (114 ± 66; P < 0.001) and msNPDR (77 ± 40; P = 0.042) compared with CO (244 ± 115). In contrast, EPCs were unchanged in mmPDR (248 ± 155) compared with CO. Strikingly, EPCs were augmented in hrPDR (389 ± 124) compared with all other stages. Numbers of undifferentiated progenitor cells (CPCs) did not differ among CO, mmPDR, and hrPDR. Augmentation (3×) of mature EPCs in hrPDR (325 ± 118; P < 0.001) compared with CO (100 ± 49) but against all other stages of DR was observed. The percentage of mature EPCs/EPCs was augmented in an ETDRS classification-dependent manner.

conclusions. In patients with T1DM with DR, EPCs undergo stage-related regulation. In nonproliferative retinopathy, a reduction of EPCs was observed, and in proliferative retinopathy, a dramatic increase of mature EPCs was observed.

Macrovascular and microvascular diseases are the main causes of morbidity and mortality in patients with diabetes. ¹ In patients with type 1 diabetes mellitus (T1DM), diabetic retinopathy (DR) is the primary microvascular complication of long diabetes duration and inadequate glucose management. ^{2 3} In contrast to advances in other areas of diabetes treatment, approximately 25% of patients with T1DM develop DR. ^{4 5} Furthermore, advanced DR is still the leading single cause of blindness in the industrialized world. ^{3 5} Thus, the principal mechanisms of pathogenesis and disease progression of DR in T1DM must be further elucidated. ^{6 7 8}  

In longstanding diabetes with insufficient glycemic control, hyperglycemia, elevated HbA1C levels, insulin resistance, and hypertension are positively correlated with progressive diabetic retinopathy. ^{2 9} The conversion from nonproliferative DR (NPDR) to proliferative DR (PDR) was associated with chronic inflammatory reactions and immune phenomena, as indicated by higher white blood cell counts, ¹⁰ inflammatory markers, ¹¹ and increased endothelial leukocyte adhesion in patients with diabetes. ¹²  

Primary pathohistologic changes in DR are thickening of capillary basement membrane, loss of microvascular pericytes, and formation of microaneurysms. ¹³ Increased vascular permeability and progressive microvascular closures lead to ischemia and tissue hypoxia, with final retinal neovascularization triggered by different local angiogenic factors. ^{2 9 14 15 16 17} Such end-stage DR involves the activation and proliferation of retinal vascular endothelial cells ¹⁸ leading to vision-threatening complications, among them secondary glaucoma, recurrent vitreous bleeding, and tractional retinal detachment. ⁵  

Many trials have described an important role of locally activated vascular endothelial growth factor (VEGF), a cytokine promoting endothelial cell growth and permeability, ^{2 9 15} in end-stage DR. VEGF mobilizes bone marrow–derived circulating endothelial progenitor cells (EPCs) ^{19 20} and serves as a chemoattractant protein for EPCs. It has been demonstrated that in ischemic tissue, circulating EPCs go directly to the site of neovascularization to differentiate and induce new vessel formation, ²¹ leading to a so-called postnatal vasculogenesis rather than isolated local angiogenesis. This phenomenon is assumed to play a role in tumor vascularization and PDR. Interestingly, a number of investigators have reported an impact of progenitor/stem cells in choroidal neovascularization in patients with age-related macular degeneration. ^{22 23 24}  

Similarly, preliminary studies in patients with type 2 diabetes mellitus (T2DM) have shown that EPCs are elevated in vivo in proliferative DR (PDR) ^{25 26} compared with non-PDR. ²⁵ More primitive progenitor cells have been found to be augmented ²⁵ or unchanged ²⁶ in vivo in DR of T2DM. Thus far, in T1DM, only an augmented clonogenic potential of mixed precursor cells has been demonstrated in vitro. ²⁷ Thus, an in vivo study of circulating progenitor cells (CPCs), EPCs, and mature EPCs was performed in patients with T1DM carefully investigated for different stages of DR. 

Patients and Methods

Patients

The study was approved by the institutional ethics committee and complies with the Declaration of Helsinki, ²⁸ including current revisions, and with the Good Clinical Practice guidelines. ^{29 30} The procedures followed were in accordance with institutional guidelines; all subjects gave written informed consent before the study. 

Ninety patients with T1DM were studied and staged for DR according to the Early Treatment of Diabetic Retinopathy Study (ETDRS) classification. ^{31 32 33} Patients with T1DM were consecutively enrolled at the outpatient clinic of the Department of Ophthalmology, Rudolf Foundation Clinic Vienna, Vienna, Austria. Patients who had retinal abnormalities other than DR or who underwent panretinal laser treatment, vitrectomy, or anti-VEGF therapy in the past 12 months were excluded. Patients who had systemic conditions such as cardiovascular disease, malignant disease, hematologic disorder, or estimated glomerular filtration rate less than 60 mL/min or who underwent treatment with erythropoietin were excluded. Smoking status did not differ among the five patient groups (P = 0.958). On average, 35.6% of patients were current or former smokers (n = 32). 

All patients were treated according to the Guidelines of the American Diabetes Association. When necessary, antihypertensive drug treatment or statins were added. Although patients with advanced retinopathy needed ACE inhibitors, ARB, or statins more often, the differences did not reach levels of significance. Ninety percent of our patients received statins for 1 year before EPC examination. Because of financial reimbursement, most (78%) of them were administered simvastatin 40 mg; other dosages and statins were used in the other patients. However, we did not observe statistical differences. As baseline characteristics, current age, age at manifestation, diabetes duration, glycosylated hemoglobin (HbA1c), lipid parameters, body mass index (BMI), and blood pressure were assessed in all patients. 

Ophthalmic Examination, Assessment, and Staging of Diabetic Retinopathy

DR grading was performed according to the standardized ETDRS retinopathy severity scales. ^{31 32 33} Before retinal photography, the patient’s pupils were dilated with tropicamide 1%; this procedure was repeated if the pupils did not reach at least 5 mm in diameter. Color retinal slides with a suitable 40° retinal camera were taken of two fields in both eyes. For the central field, the center of the optic disc was positioned at the nasal end of the horizontal meridian of the field of view; for the nasal field, the optic disc was positioned 1 disc diameter from the temporal edge of the field on the horizontal meridian. Two pseudostereoscopic images with a horizontal shift of 3° to 5° were then taken of each field in both eyes, resulting in eight images of each patient. The overlap of the fields recorded a retinal view of approximately 75° horizontally by 40° vertically; therefore, clinically significant lesions of diabetic retinopathy were easily detectable. Slides were separately screened by two experienced graders in the same center, according to a validated protocol. ³⁴ In patients with advanced disease (moderate-severe NPDR and PDR), fluorescein angiography and optical coherence tomography were also performed. For grading classification of the T1DM patients, the following stages were differentiated: no retinopathy, meaning no definite retinopathy (CO; n = 30); mild nonproliferative DR, defined by small hemorrhages and/or hard exudates (mNPDR; n = 30); moderate-severe NPDR, defined by moderate-severe bleeding, hard exudates, venous beadings, or intraretinal microvascular abnormalities (msNPDR; n = 10); mild-moderate proliferative diabetic retinopathy, defined by small neovascularizations on optic disc (NVD) or elsewhere (mmPDR; n = 10); and high-risk PDR, defined by large NVD or vitreous hemorrhage with retinal neovascularizations (hrPDR; n = 10). 

Because it has been reported that EPCs are reduced in patients with diabetes, we used a group of nondiabetic subjects (NODS; n = 30) who were matched for age (P = 0.714), sex (P = 0.841), and body mass index (P = 0.529) to investigate differences in numbers of CPCs, EPCs, and mature EPCs between CO and NODS (all P values were derived using the independent Student’s t-test). 

Flow Cytometry of Bone Marrow–Derived Circulating Progenitor Cells

Different surface markers have been used to discriminate progenitor cells involved in vascular biology. ^{35 36 37} Progenitor cells (derived from the human bone marrow) must be differentiated from circulating endothelial cells, which are enriched after desquamation from the vessel wall in states of vessel damage. ³⁸ The most primitive progenitor cell for vascular repair, the hemangioblast, also termed circulating progenitor cell (CPC), was described to be CD34⁺ and CD133⁺. More specific and differentiated endothelial progenitors, the EPCs, were found to be CD34⁺, CD133⁺, and CD309⁺. ^{36 37} EPCs are further distinguished as early and late or resting and mature, depending on their expression of adhesion molecules, such as CD31, enabling them to participate in vascular repair. Thus, one can discern between CPCs, EPCs, and mature EPCs (mat-EPCs) because of their surface receptors. For practical reasons and to study different subsets, we used CD34/CD133 coexpression for all progenitor cells in peripheral blood with possible angioblastic potential (CPCs). We used CD34/CD133/CD309 coexpression for the classic endothelial progenitor cells (EPCs) ³⁹ and CD34/CD133/CD309/CD31 coexpression for the identification of EPCs that express surface molecules with adhesion potential. We chose CD31 on the basis of literature that reported expression on EPC, which is not yet a fully matured endothelial cell. Additionally, we enumerated CD34⁺/CD133⁺/CD309⁺ cells, which were CD31⁻, thus resembling a nonmature state (nonmat EPCs). 

Immediately after blood sampling, 100 μL whole heparinized blood specimens were stained with saturating concentrations of monoclonal antibodies (mAbs): FITC-conjugated anti-CD31 mAb (Becton Dickinson [BD] Biosciences, San Jose, CA; clone WM59), phycoerythrin (PE)-conjugated anti-CD309 (R&D Systems, Minneapolis, MN; clone 89106), PECY5-labeled anti-CD34 (BD Biosciences; clone 581), and APC-conjugated anti-CD133 (Miltenyi Biotec, Bergisch Gladbach, Germany; clone AC133) mAbs or corresponding isotype controls. Because of the different antibody conjugations we used, we might have found different cell numbers than other groups. Anti-CD34 mAb is used by most groups in fluorescein isothiocyanate (FITC), though the general recommendation for stem cell enumeration by the International Society of Hematotherapy and Graft Engineering ^{40 41 42} is to use anti-CD34 in a strong (reddish) color such as PE, to prevent slipping of cells. All experiments were performed with 30 minutes of cleaning time of the flow cytometer before acquisition (standard operating procedure at our institution). The lyse-and-wash method reduced the debris; in our analysis, it accounted for up to 12% of all events in all acquisitions obtained. After the lyse-and-wash procedure, white blood cells (WBCs) were instantly acquired (FACSCalibur; BD Biosciences). The acquisition goal was 1 × 10⁶ events. If fewer events were detectable, results were processed statistically only if at least 5 × 10⁵ events could be obtained. Progenitor cells were counted by flow cytometry and expressed in absolute numbers per 10⁶ white blood cells (events in the leukocyte gate), to enable a comparison. One measurement was made for each patient. To identify EPCs, a gating strategy including four gates has been applied: in the given example, 237 EPCs were identified out of 1 × 10⁶ events. A negative control was applied with phosphate-buffered saline (PBS) but without antibody (see 1 Fig. 2 ). 

Statistical Analysis

Data are presented as mean ± SD. Differences between patient groups were analyzed by Student’s unpaired t-test or ANOVA, as appropriate. To identify determinants of levels of CPCs, EPCs, and mat-EPCs, univariate correlation analysis was performed. In addition, variables associated with levels of CPCs, EPCs, and mat-EPCs were included in multivariate regression analysis. An α-level of P < 0.05 (two-tailed) was considered statistically significant. All statistical analyses were performed with a statistical software package (SPSS 15.0; SPSS Inc., Chicago, IL). 

Results

Patient History and Laboratory Baseline Characteristics

Baseline characteristics for all patients were as follows (mean ± SD): age, 41 ± 12 years; diabetes duration, 21 ± 2 years; HbA1c, 7.8 ± 1.5 rel%. Patients were well matched. Similarly, no significant differences between NPDR and PDR concerning age (42.5 ± 13.5 vs. 43.0 ± 8.9 years; P = 0.884), diabetes duration (23.3 ± 10.8 vs. 25.8 ± 11.7; P = 0.458), manifestation age of diabetes (18.8 ± 12.2 vs. 19.1 ± 12.4; P = 0.920), LDL cholesterol (104 ± 28 vs. 107 ± 55; P = 0.836), and HDL cholesterol (75 ± 21 vs. 63 ± 20; P = 0.070) were found. Additional baseline characteristics among the five groups (CO, mNDPR, msNDPR, mmPDR, hrPDR) are shown in Table 1 . According to ANOVA, patient groups did not differ for glucose control or insulin regimen. 

Funduscopy of Diabetic Eyes

Respective examples of fundus photographs are shown in Figure 1 . The first funduscopic view (Fig. 1A)shows no signs of diabetic retinopathy (CO); the second slide (Fig. 1B)represents mNPDR with small hemorrhages and singular hard exudates in the macula. The third example (Fig. 1C)shows msNPDR, with larger hemorrhages and hard exudates in three to four quadrants and small cotton wool spots as a sign of mild ischemia. The fourth view (Fig. 1D)demonstrates features of mmPDR, mild signs of neovascularization outside the optic disc, and panretinal laser scars. The last view (Fig. 1E)is from a patient with hrPDR; the funduscopic picture is blurred by recurrent vitreous hemorrhage caused by neovascularizations covering more than one-third of the optic disc area. 

Enumeration of Bone Marrow–Derived Progenitor Cells

CPCs, EPCs, and mat-EPCs were enumerated as described. In Figure 2 , respective examples of CPC, EPC, and mat-EPC numbers in different stages of retinopathy are shown. In patients with T1DM, CPCs, EPCs, and mat-EPCs (absolute numbers per 1 × 10⁶ white blood cells) were significantly reduced in early diabetic retinopathy; EPCs and mat-EPCs then showed significant stage-related increases with more advancing stages of disease. 

Compared with the control group (CO; 1176 ± 473), CPCs decreased in the mNPDR group (592 ± 324; P < 0.001) and in the msNPDR group (510 ± 264; P < 0.001) and then gradually increased to levels of controls in the mmPDR group (1207 ± 779; P = 0.021 compared with msNPDR) and finally to 1155 ± 370 in the hrPDR group (Table 2) . 

Numbers of EPCs primarily declined with advancing stages of retinopathy: CO (244 ± 115) compared with mNPDR (114 ± 66; P < 0.001); msNPDR (77 ± 40; P < 0.001) compared with CO. In contrast, in patients with proliferative disease, no change compared with CO in patients with mmPDR (248 ± 155) was observed. In hrPDR, clinically characterized by heavy neovascularization, a significant increase in EPCs (89 ± 124; P = 0.002) but not CPCs was found compared with CO (P < 0.001), msNPDR (P = 0.038), and mmPDR (Table 2) . 

Levels of mat-EPCs decreased from 100 ± 49 in the controls to 62 ± 41 (P = 0.002 compared with CO) in the mNPDR group and 45 ± 24 (P = 0.241 compared with CO) in the msNPDR group. Strikingly, highly significant increases to 158 ± 119 (P = 0.015 compared with msNPDR) in the mmPDR group and to 325 ± 118 (P = 0.006 compared with mmPDR) in the hrPDR group were found (Table 2) . 

We were able to confirm previous findings of reduced EPCs in patients with diabetes compared with healthy subjects. We found a significant reduction of EPCs (244 ± 115 vs. 353 ± 162; P = 0.002) and mat-EPCs (100 ± 49 vs. 146 ± 56; P < 0.001) but not CPCs (1176 ± 473 vs. 1213 ± 372; P = 0.713) in CO compared with NODS. 

Correlation and Association of Progenitor Cells with Baseline Characteristics

In our study, CPC levels were univariately significantly correlated with diabetes duration (R = −0.254; P = 0.023), HbA1c (R = 0.250; P = 0.027), and triglyceride levels (R = 0.249; P = 0.030). After multivariate regression and confounding analysis, diabetes duration remained the strongest predictor for CPC (β = −0.254; P = 0.023) but was significantly confounded by HbA1c and triglyceride. 

EPCs were univariately significantly correlated with HbA1c (R = 0.366; P < 0.001), total cholesterol (R = 0.271; P = 0.018), and triglyceride levels (R = 0.375; P < 0.001). After multivariate regression and confounding analysis, triglyceride remained the strongest predictor for EPC (β = 0.375; P < 0.001) but was significantly confounded by HbA1c and total cholesterol. 

In univariate and multivariate regression analysis, mat-EPCs were significantly associated only with the level of glycemic control, the HbA1c (β = 0.288; P = 0.011). 

Correlation of the Percentage of mat-EPCs with the Classification of Diabetic Retinopathy

To determine the association with differentiation of progenitor cells, the percentages of EPCs/CPCs and mat-EPCs/EPCs were calculated (Table 2) , providing insight into changes in differentiation of bone marrow–derived progenitor cells responsible for vascular repair. The percentage of EPCs/CPCs, estimating the proportion of resting EPCs, was only weakly affected by different stages of retinopathy. In hrPDR only, a significantly increased percentage of EPCs/CPCs against all other stages was found. In contrast, the percentage of mat-EPCs/EPCs, thereby estimating the relation of mature EPCs to resting EPCs, was augmented in a classification-dependent manner (Table 2 ; Fig. 3 ). The relation of nonmature EPCs (CD34/CD133/CD309 triple⁺ cells, which were CD31⁻) to resting EPCs inversely decreased in the same classification-dependent manner (Fig. 3) . 

Discussion

We are the first to report a correlation of bone marrow–derived progenitor cells with DR severity in patients with T1DM. In this in vivo study, we could demonstrate that in patients with DR, EPCs show stage-related changes in numbers. Patients with advanced stages of nonproliferative DR experienced significant reductions of EPCs and mat-EPCs. The more undifferentiated progenitor cells (CPCs) showed no significant difference in the various DR stages. In proliferative DR, strong increases were found in EPC and mat-EPC numbers compared with all other stages. Additionally, the percentage of mat-EPCs/EPCs (estimating the relation of mat-EPCs to resting EPCs) was augmented in a modified ETDRS classification-dependent manner. In contrast, the percentages of nonmat-EPCs/EPCs (estimating the relation of nonmature EPCs to resting EPCs) inversely decreased in a classification-dependent way. Thus, our findings strongly argue for an involvement of EPCs and especially mat-EPCs but not CPCs in the pathophysiology of DR of T1DM. 

Recent studies in patients with T2DM have investigated the potential role of bone marrow–derived progenitor cells in PDR ^{25 26} compared with non-PDR. ²⁵ Lee et al. ²⁵ measured CD117⁺ and CD34⁺ cells in the peripheral blood by flow cytometry. They were the first to show that CD34⁺ cells may be increased in patients with NPDR and PDR compared with patients with diabetes but without DR. However, by measuring CD34⁺ cells, undifferentiated progenitor cells without a vascular potential might have been enumerated. Lee et al. ²⁵ described an increase of CD117⁺ cells in NPDR and PDR; however, CD117 is not expressed on mature and active EPCs or on mature endothelial cells. 

Interesting observations have been reported by Fadini et al., ²⁶ who investigated EPCs in patients with T2DM with and without DR and with and without peripheral arterial occlusive disease (PAOD). They showed a different involvement of EPCs in those two severe complications. Similarly, EPCs were augmented in PDR but reduced in PAOD, thus arguing for an involvement of progenitor cells in both diseases, at least in patients with T2DM. 

In T1DM, information regarding EPCs and DR is scarce. In a preliminary study, Asnaghi et al. ²⁷ investigated the clonogenic potential of EPCs in 34 subjects with and without T1DM with and without DR. They found a significant correlation of EPC colony-forming unit (CFU) assays with HbA1c levels. However, the CFU assays used for EPC investigations must be interpreted carefully because several papers ^{43 44} have shown that the culture of unseparated whole peripheral blood in a CFU assay might be misleading: the 10% primary monocytes in the assays express most tested endothelial genes and proteins at even higher levels than the supposed EPC progeny. ⁴³ Additionally, LDL uptake, lectin binding, and CD31/CD105/CD144 expression are inherent features of monocytes, making them phenotypically indistinguishable from putative EPCs. Thus, monocytes and their progeny can phenotypically mimic EPC in various experimental models. ⁴³ Little is known about the variability of EPC numbers in peripheral blood under nonstress conditions. 

To our knowledge, this is the first study to examine (in vivo) systemic EPCs in relation to advancing stages of DR in patients with T1DM. Vasculogenic progenitor cells have been hypothesized to be deleted in chronic angiopathy (diabetic more than nondiabetic) and to be augmented in acute ischemia. ^{20 45 46 47 48} Several cytokines have been thought to be responsible for the chronic decrease and for the acute increase of EPCs in various in vivo animal and in vitro human studies. ^{20 45 46 47 48} However, the mediators responsible for the recruitment of EPCs are still under investigation. VEGF seems to play a major role in the attraction of EPCs in acute ischemia in macrovascular lesions. Additionally, elevated levels of VEGF have been reported in the ocular fluid of patients with DR. ⁴⁹ However, VEGF might not be responsible for the systemic recruitment of EPCs from the bone marrow. Although Lee et al. ²⁵ found an increase of systemic VEGF in proliferative and nonproliferative diabetic retinopathy, those VEGF levels did not correlate with the severity of DR. Wasada et al. ⁵⁰ demonstrated increased VEGF levels in patients with T2DM compared with controls, but those levels were associated with diabetic nephropathy rather than retinopathy. ⁵⁰ In the EUCLID ⁵¹ study (patients with T1DM), mean VEGF concentrations were not significantly different in patients with different stages of retinopathy. However, apart from VEGF, estrogen, erythropoietin, ⁵² IL-18, stromal-derived factor-1, ⁵³ nitric oxide, ⁵⁴ and asymmetric dimethyl arginine have been reported to be involved in the mobilization, proliferation, chemoattraction, or conservation (or any combinations thereof) of EPCs. Data regarding changes in the latter proteins are scarce and must be evaluated in future prospective studies. 

The exact mechanism for the elevation of EPC levels in patients with T1DM and PDR is unknown. We suspect that an unknown mediator produced locally, within the retina, exerts the recruitment of EPCs systemically. 

Those high EPC levels seem to be harmful in patients with advanced DR and to lead to major ocular complications, such as vitreous bleeding, proliferative vitreoretinopathy, and secondary glaucoma. High EPC ²⁶ numbers might be a possible explanation for the diabetic paradox ⁹ : limited revascularization (collateral vessel formation) in diabetic hearts and limbs but increased neovascularization in advanced DR. A high number of structurally impaired EPCs might therefore be the reason for the disastrous end-stage complications in diabetic retinopathy, as suggested by Tepper et al. ⁵⁵ and Caballero et al. ⁵⁶ Additionally, differential expression of growth factors/cytokines between tissues/body compartments may also contribute to preferential homing of EPCs to the retina, offering an explanation for the paradox. Further randomized, prospective studies are clearly needed to illuminate this hypothesis. 

In summary, our findings argue for a strong involvement of EPCs in the pathophysiology of retinopathy of T1DM. To be precise, the relation of mature EPCs to resting EPCs was augmented in a relatively strict classification-dependent manner. Thus, measuring EPC levels in patients with diabetic retinopathy may facilitate staging. 

We do not know, however, whether high EPC levels are the cause or the consequence of advanced DR. Moreover, we still cannot predict the rapidity of disease progression in different patients with PDR. This knowledge would be of great benefit because it has been shown that the severity of retinopathy in patients with T1DM is significantly associated with the amount of risk for cardiovascular disease. ^{57 58} More and larger studies are needed to clarify whether measuring EPC levels is helpful for early identification of patients at risk in the diagnostic or therapeutic routine management of diabetic eye disease. 

³  These authors contributed equally to the work presented here and should therefore be regarded as equivalent authors.

 Supported by the Austrian Diabetes Association, Grant Research Price 2006.

 Submitted for publication January 16, 2008; revised June 17 and August 6, 2008; accepted October 20, 2008.

 Disclosure: S. Brunner, None; G.-H. Schernthaner, None; M. Satler, None; M. Elhenicky, None; F. Hoellerl, None; K.E. Schmid-Kubista, None; F. Zeiler, None; S. Binder, None; G. Schernthaner, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 2.

 

Respective examples of EPCs in patients with T1DM and different stages of diabetic retinopathy. Row 1: negative control of a representative patient with PBS but without antibody. Row 2: ungated dot plots of 1 million WBCs. Row 3: EPCs in a patient without DR. Row 4: EPCs in a patient with msNPDR (note the reduction of EPCs). Row 5: EPCs in a patient with hrPDR (note the total increase of EPCs and the augmentation of CD31 expression by EPCs.

Table 1.

 

Baseline Characteristics According to Different Stages of Diabetic Retinopathy

Retinopathy Stages	CO	mNPDR	msNPDR	mmPDR	hrPDR	P *
Height (cm)	170 ± 9	169 ± 9	165 ± 10	168 ± 9	169 ± 6	0.636
Weight (kg)	72 ± 12	72 ± 14	66 ± 13	78 ± 16	77 ± 11	0.372
BMI (kg/m2)	24.9 ± 3.2	24.9 ± 3.5	24.2 ± 3.0	27.7 ± 6.6	26.9 ± 3.1	0.227
Pulse (s)	83 ± 14	79 ± 13	77 ± 13	80 ± 11	76 ± 14	0.708
Systolic blood pressure (mm Hg)	118 ± 17	125 ± 19	128 ± 24	132 ± 18	126 ± 14	0.279
Diastolic blood pressure (mm Hg)	77 ± 10	77 ± 11	78 ± 11	77 ± 8	79 ± 6	0.992
Total cholesterol (mg/dL)	215 ± 92	196 ± 32	207 ± 41	157 ± 25	249 ± 50	0.064
HDL cholesterol (mg/dL)	64 ± 14	73 ± 18	85 ± 29	66 ± 22	59 ± 20	0.051
Triglyceride (mg/dL)	166 ± 217	95 ± 47	99 ± 32	104 ± 53	197 ± 88	0.226
Statins (% of patients)	90.0	96.6	80.0	90.0	70.0	0.514
Basis insulin (IU)	25 ± 13	23 ± 10	18 ± 8	20 ± 6	17 ± 15	0.335
Bolus insulin (IU)	24 ± 10	21 ± 7	15 ± 8	21 ± 11	14 ± 12	0.084
Total insulin (IU)	45 ± 23	43 ± 14	33 ± 13	36 ± 18	33 ± 24	0.318

*   P was determined by ANOVA.

Figure 1.

 

Respective examples of fundus photographs following ETDRS severity scale. (A) No signs of DR. (B) mNPDR. (C) msNPDR. (D) mPDR. (E) High-risk PDR.

Table 2.

 

Number of Progenitor Cells According to Different Stages of DR

Retinopathy Stages	CO	mNPDR	msNPDR	mmPDR	hrPDR	P *
CPC (n)	1176 ± 473	592 ± 324	510 ± 264	1207 ± 779	1155 ± 370	<0.001
EPC (n)	244 ± 115	114 ± 66	77 ± 40	248 ± 155	389 ± 124	<0.001
mat-EPC (n)	100 ± 49	62 ± 41	45 ± 24	158 ± 119	325 ± 118	<0.001
EPC (% CPC)	21 ± 4	19 ± 1	15 ± 0	22 ± 4	34 ± 1	<0.001
mat-EPC (% EPC)	42 ± 12	54 ± 17	60 ± 12	65 ± 20	83 ± 9	<0.001

*   P determined by ANOVA.

Figure 3.

 

Correlation of percentages of mat-EPCs and nonmat-EPCs with classification of DR. Left: increasing percentages of mat-EPCs (identified by the CD31 expression on EPCs) with advancing stages of DR. Right: reduced numbers of nonmat-EPCs with increasing severity of retinopathy.

The authors thank the Department of Ophthalmology for expendable items, especially photographic articles.