Characterization and Comparison of Bone Marrow and Peripheral Blood

CharacterizationBlackwell Publishing Ltd and comparison of bone marrow and peripheral blood mononuclear cells used for cellular therapy in critical leg ischaemia: towards a new cellular product

J.-C. Capiod,1 C. Tournois,2,3 F. Vitry,4 M.-A. Sevestre,5 S. Daliphard,2 T. Reix,5 P. Nguyen,2,3 J.-J. Lefrère1 & B. Pignon3,6 1Laboratory of Haematology, University Hospital of Amiens, Amiens, France 2Laboratory of Haematology, University Hospital of Reims, Reims, France 3EA3801, Institut Fédératif de Recherche 53, Laboratoire d’Hématologie, Centre Hospitalier Universitaire Reims, Hôpital Robert-Debré, Reims Cedex, France 4Department of Methodology, University Hospital of Reims, Reims, France 5Department of Vascular Surgery, University Hospital Amiens, Amiens, France 6Department of Cell Therapy, University Hospital Reims, Reims, France

Background and Objectives Autologous transplantation of either bone marrow (BM) or peripheral blood (PB) mononuclear cells (MNC) induces therapeutic angiogenesis in patients with peripheral arterial occlusive disease. Yet, the precise nature of the cellular product obtained from BM or PB and used in these therapeutic strategies remains unclear. Materials and Methods We have analysed the characteristics of BM-MNC and PB-MNC collected without mobilization and implanted in patients with critical limb ischaemia in a clinical trial of cellular therapy including 16 individuals treated by BM-MNC and eight by PB-MNC. These MNCs were characterized by cell counts, viability assessment and enumeration of leucocyte subsets, CD34 stem and endothelial progenitor cells (EPCs) (CD34+/CD133+/VEGF-R2+) by flow cytometry. Mean fluorescence intensity ratios were determined for CD34, CD133 and VEGF-R2 markers. All analyses were simultaneously performed in two laboratories. Results Accuracy and reliability between both laboratories were achieved. BM-MNCs and PB-MNCs were quantitatively and qualitatively heterogeneous and quite different from each other. Stem cells and EPCs were significantly more present in BM- compared to PB-cell products, but with similar mean fluorescence intensity ratios. A weakly positive correlation was observed between CD34+ cell counts and EPCs levels, confirming the specificity of cell identification. Conclusion A great variability was observed in cell product characteristics according to their origin and also between individuals. These data stress the necessity of optimal Received: 11 August 2008, characterization of cell products especially in multicentric clinical trials. revised 28 October 2008, accepted 12 November 2008, Key words: cellular therapy, cellular product, critical leg ischaemia, endothelial published online 15 December 2008 progenitor cells, therapeutic angiogenesis.

Correspondence: Claire Tournois, EA3801, Institut Fédératif de Recherche 53, Laboratoire d’Hématologie, Centre Hospitalier Universitaire Reims, Hôpital Robert-Debré, 51092 Reims Cedex, France E-mail: [email protected] Statement of equal author contribution: Jean-Claude Capiod and Claire Tournois contributed equally to this study. Conflict of interest: all authors disclose conflict of interest of any kind related to this manuscript. Authors had full access to the data. Abbreviations: 7-AAD, 7-aminoactinomycin; BM, bone marrow; BMI, body mass index; CLI, critical limb ischaemia; ECD, phycoerythrin-texas red; ECs, endothelial cells; EPCs, endothelial progenitor cells; FC, flow cytometry; GFR, glomerular filtration rate; MDRD, modification of diet in renal disease; Med, median; MGG, May Grünwald Giemsa; PB, peripheral blood; PCy5, phycocyanin 5; PCy7, phycocyanin7; VEGF-R2, vascular endothelial growth factor receptor 2.

256

Cell therapy in critical limb ischaemia 257

Table 1 Patient’s characteristics Introduction Several studies have shown that autologous transplantation BM cell products PB cell products of either bone marrow (BM) or peripheral blood (PB) mono- group (n = 16) group (n = 8) nuclear cells (MNC) could be effective in inducing angiogenesis Age (years) 71 [45–84] 60 [37–84] in patients with peripheral arterial occlusive disease [1–6]. Gender Experimental studies have shown that circulating BM- Male 14 (87·5%) 6 (75%) derived endothelial progenitor cells (EPCs) can be mobilized Female 2 (12·5%) 2 (25%) and incorporated into sites of active ischaemia to increase BMI (kg/m2) 24·8 [20·8–30·8] 22·7 [18·5–31·0] postnatal neovascularization [7,8], and that circulating BM- Smoking one current, 12 past, two current, 4 past, derived EPCs are able to differentiate into mature endothelial 3 never 2 never cells (ECs) [9]. The mechanisms regulating the differentiation Disorders of BM-derived EPCs, their mobilization and their homing to Arterial hypertension 11 (69%) 6 (75%) ischaemic sites remain partially unknown [10], but the Hypercholesterolemia 12 (75%) 3 (37·5%) implantation of BM-MNC [11] and of PB-MNC [12] into Diabetes mellitus 6 (37·5%) 1 (12·5%) 2 a ischaemic tissues (through the intramuscular local transplan- GFR (ml/min/1·73 m ) 59 [30–86] 81 [59–172] Haemoglobin (g/l) 132 [104–154] 124 [100–141] tation of these cells) induces collateral vessel formation in White blood cells (109/l) 7·2 [5·0–10·4] 7·4 [5·6–8·9] animals, suggesting that this cell therapy could constitute a new strategy for therapeutic angiogenesis. Quantitative variables were expressed as median and range [min–max]. The questions of which cells and which optimal effective aGFR, Glomerular filtration rate (MDRD or modification of diet in renal quantity to implant are crucial as long as the precise neovas- disease). cularization mechanisms are not completely elucidated. However, in most clinical trials, the exact composition of the cellular product used remains elusive and is obviously highly unilateral CLI and were not suitable candidates for surgery, variable. In fact, several parameters such as the origin of the non-surgical acts or revascularization. Patients with poorly MNC (BM or PB), the health status of the donor and the controlled diabetes mellitus or with a history of malignant preparation procedures [13] may lead to major differences in disorder were excluded. The ethical committee of each this cellular product, which is perhaps destined to become a participating centre approved the protocol. Written informed new ‘transfusion product’. For this reason, it is important to consent was obtained from all patients. Twenty-four patients try to characterize most precisely such a product before the were included: 16 treated with BM-MNC and eight with implementation of future multicentre clinical trials aiming to PB-MNC (Table 1). Patients were consecutively included as fully demonstrate the efficacy and safety of this cellular soon as they presented with appropriate criteria and were not therapy in critical limb ischaemia (CLI). Such an approach selected to receive one or another type of cells. The first 16 would allow choosing the ‘best’ cell product in terms of selected patients were treated with BM-MNC; the following efficacy, to standardize this product by cell characterization ones were treated with PB-MNC. Cells were implanted 1–3 h methods easy to perform, and to contribute to the best knowledge after preparation by 30 intramuscular injections into the of involved mechanisms. We report here our results in gastrocnemius of the ischaemic leg. Total injection volume characterizing BM-MNC and PB-MNC implanted in patients was 30 ml. In this article, we focalized on the two types with CLI in a French clinical trial. This trial was an open cellular products used. bicentric prospective study performed in two academic centres of vascular surgery and tested the safety and efficacy Bone marrow and peripheral blood mononuclear of BM-MNC or non-mobilized PB-MNC implanted in individuals cell samples with CLI. The clinical results will be presented as soon as all patients will have achieved sufficient follow-up. However, an For the preparation of BM-MNC, 500 ml of bone marrow was interim analysis was performed that confirmed the safety and collected under general anaesthesia through multiple was in favour of similar efficacy for both cell products. punctures of the posterior iliac crest using a Jamshidi needle. MNC were isolated using a blood cell separator (Cobe Spectra, version 4, Bone Marrow Processing Program, Gambro BCT, Materials and methods Lakewood, CO, USA). PB-MNC were collected by cytapheresis A prospective clinical trial was initiated in two French academic of one blood mass (5·1 ± 1·1 l) during 90 min with the hospitals in order to establish whether intramuscular same blood cell separator (Cobe Spectra, version 6, auto- implantation of BM-MNC or PB-MNC can induce clinical PBSC program, Gambro BCT). In both cases, the blood cell improvement in patients with CLI. Enrolled patients had separator was programmed to obtain a final volume of 40 ml

258 J.-C. Capiod et al.

MNC concentrate (30 ml for autologous re-injection and CD34+ cell enumeration was also ensured during the first 10 ml for controls and analyses). There was no previous step of EPCs quantification with phycocyanin 5 (PCy5)- mobilization by any haematopoietic growth factor. Blood cell CD34. separator was centralized in one centre (Reims Hospital). Samples of the end-product were collected and diluted Endothelial progenitor cells. Endothelial progenitor cells 1 : 2 in citrate–citric acid–dextrose to avoid cellular aggregates were defined as CD34+/CD133+/VEGF-R2+ cells [15–17]. (ACDA, Baxter Healthcare, Deerfield, IL, USA). A final volume Four-colour labelling was used to assess the co-expression of of 2 ml was dispatched at stable ambient temperature in the CD34, CD133 and VEGF-R2. This allowed the concomitant laboratories of the two participating centres (laboratory A for enumeration of CD34+, CD34+/CD133+ [18], CD34+/ Amiens Hospital; laboratory B for Reims Hospital). Cell VEGF-R2+ cells, and CD34+/CD133+/VEGF-R2+ cells. After a analyses were simultaneously performed in these laboratories, 10-min incubation at room temperature of 2 × 106 BM or PB- within 3 h after preparation of the cell product. MNC per tube with 5 μl of FcR Blocking Reagent (Miltenyi Biotec, Bergisch Gladbach, Germany), the cells were processed for four-colour FC analysis. Filters and photodetectors were Mononuclear cell characterization similar in both laboratories for measuring light emitted by Cell counts FITC, PE and PCy5, and different for the fourth fluorescence Mononuclear cell counts were performed with an XE-2100™ that was phycoerythrin-texas red (ECD) in laboratory A and Sysmex counter (Roche Diagnostics, Meylan, France) in phycocyanin7 (PCy7) in laboratory B. The following anti- laboratory A, and a Gen’s® counter (Beckman Coulter, human monoclonal antibodies were used: ECD-CD45 (Clone Villepinte, France) in laboratory B. Cytospins were prepared J33, Beckman Coulter) in laboratory A and PCy7-CD45 using Cytospin 3 Shandon® cytocentrifuge with a volume of (Clone J33, Beckman Coulter) in laboratory B, PCy5-CD34 50 μl cells adjusted at a concentration of 1 · 109/l, dried and (Clone 581, epitope class III, Beckman Coulter), and PE- stained with May Grünwald Giemsa before examination in VEGF-R2 (R&D System, Wiesbaden, Germany). To identify light microscopy. Cell count was used to calculate positive CD133+ cells, an indirect method was used with a first cell quantification from the percentage of positive cells incubation with biotinylated CD133 (Clone AC133, epitope 1, obtained by flow cytometry (FC) analysis. Miltenyi Biotec) followed by a second with FITC-antibiotin (Miltenyi Biotec). For each sample, seven tubes were prepared Characterization of mononuclear cell in flow cytometry with, respectively, four isotopic controls (one for each A three-colour FC analysis was performed by both laboratories fluorescence wavelength), one for autofluorescence assess- on Epics XL® analysers (Beckman Coulter) to determine the ment and two test tubes for positive cell analysis. Two 30-min proportion of CD2+ cells (T lymphocytes and natural killer incubation periods were performed at 4°C in the dark: first cells), CD19+ cells (B lymphocytes), and CD14+ (monocytes) with CD45, CD34, CD133 and/or VEGF-R2, then with cells in the cell products. PE-CD2/FITC-CD45, PE-CD19/ antibiotin-FITC. After each incubation, the cells were washed FITC-CD45 and PE-CD14/FITC-CD45 (Beckman Coulter) at room temperature in phosphate-buffered saline (PBS; were used. Cell suspensions adjusted at 5 × 105 per tube were Dulbecco ohne Ca/Mg, Biochrom AG seromed®, Berlin, incubated for 10 min, at room temperature, in the dark, with Germany) by centrifugation for 5 min at 200 g. Cell suspensions 10 μl monoclonal antibodies and 20 μl 7-AAD (7-aminoac- were then incubated with Versalyse Lysing Solution® and tinomycin) viability dye (Beckman Coulter). After red cell IOTest 3® Fixative Solution (Beckman Coulter) following lysis with Versalyse Lysing Solution® (Beckman Coulter), a the manufacturer’s instructions. After a final wash with 5 ml count was performed for each relevant marker among PBS and centrifugation for 5 min at 200 g, the cells were 20 × 103 viable (7-AAD negative), CD45+ cells. resuspended in 1 ml PBS with 25 μl IOTest 3® Fixative Solution (Beckman Coulter). The gating strategy included Characterization of stem and progenitor cells in flow exclusion of platelets, debris, nude nuclei (megakaryocytic and cytometry erythroblastic cells) and non-viable cells, and positive selection of CD34+/CD133+, CD34+/VEGF-R2+ and CD34+/CD133+/ CD34+ stem cells. CD34+ stem cell analysis was performed VEGF-R2+ cells among 400 000 relevant cells (more than according to the reference method [14] using PE-CD34 500 000 total events). (Clone 581, epitope class III, Beckman Coulter) and Positive cells quantification was calculated from cell FITC-CD45 anti-human monoclonal antibodies, the 7-AAD counts and percentages of cells identified by FC. viability dye (Beckman Coulter), PE-IgG1 and FITC-IgG1 Mean fluorescence intensity (MFI) ratios were established (Beckman Coulter) as isotypic controls, and StatusFlow using the MFI of each labelled peak vs. that of the peak of Pro™ control as target values. The same protocol as for non-specific labelling obtained with the relevant isotypic quantifying MNC was used among 75 × 103 viable cells. control [19]. Fluorospheres (Flowset®, Beckman Coulter)

Cell therapy in critical limb ischaemia 259 were also used in both laboratories to check for the absence Table 2 Laboratory reliability study (laboratory A–laboratory B) of flow of significant differences between the two instruments in the cytometry (FC) quantification of stem and progenitor cells using the Bland comparable channels. The software for list mode analysis and Altman summarized representation was XL System 2® (Beckman Coulter) and CXP™ (Beckman Coulter). BM and PB cell products (laboratory A–laboratory B)

Cell controls. Two types of cell controls were tested with the Mean Limits of same protocol. First, PB-MNC were purified from three difference biasa agreement CI 95%b samples obtained from healthy patients by density gradient 9 + centrifugation (d = 1·077, Lymphocyte Separation Medium, PE-CD34+ (10 /l) –0·037 [–0·154; 0·079] 9 + Eurobio, Les Ulis, France) and provided a control for PCy5-CD34+ (10 /l) –0·030 [–0·095; 0·036] CD34+/CD133+(109/l) +0·022 [–0·080; +0·124] non-specific labelling. Second, the HEL7 erythroblast cells CD34+/VEGF-R2+ (109/l) –0·00024 [–0·0031; +0·0024] (ATCC® number: TIB-180TM) [20] were used as positive CD34+/CD133+/VEGF-R2+ (109/l) –0·000149 [–0·0034; +0·0031] control for VEGF-R2 and CD34 labelling, respectively, and the HT29 cell line (ATCC® number: HTB-38) for CD133 aBias correspond to the mean difference between laboratory A and labelling [21]. laboratory B. This represents the systematic bias observed between laboratory A and laboratory B for a measure. Intra- and interlaboratory variation. Intralaboratory evaluation bLimit of agreement correspond to the confidence interval (CI) 95% of the of assay reproducibility was realized independently in the mean difference between laboratory A and laboratory B. This represents the two centres. Each laboratory tested 20 samples (11 BM-cell range of value obtained, with 95% CI, by laboratory B for a given value measured by laboratory A. products and nine PB-cell products) from healthy or CLI patients for CD34+, CD34+/CD133+, CD34+/VEGF-R2+ and CD34+/VEGF-R2+ cell counts. A coefficient of variation was calculated from n repetitive measurements for each assay Reliability (n = 4 or 5 for PCy5-CD34+ cell counts; n = 3 for CD34+/ CD133+ cell counts; and n = 2 for CD34+/VEGF-R2+ and PCy5-CD34 labelling being the first step in the characterization CD34+/VEGF-R2+ cell counts). of CD34+/CD133+/VEGF-R2+ EPCs, we compared PCy5- Laboratory reliability was determined using the Bland and CD34+ cell enumeration to PE-CD34+ cell enumeration: the Altman representation [22]. This method led to the determi- mean difference and 95% confidence interval between these nation of a laboratory A and laboratory B mean difference. enumerations were +0·013% [–0·53; +0·56] and +0·006% This difference is considered as the systematic bias a laboratory [–0·25; +0·26] in laboratory A and laboratory B, respectively. has compared to the other laboratory. In a second step, the Enumeration of stem and endothelial progenitor cells by FC Bland and Altman representation allowed us to determine the in both laboratories appeared highly reproducible (Table 2). limit of agreement (95% confidence interval of mean difference) A slightly larger difference was noted for immature EPCs in observed between laboratory A and laboratory B. PB-MNC products, which was found to have a frequency of about 1 × 10–5. Statistical analysis Viability and cell counts in bone marrow and Quantitative variables were expressed as median (Med) and peripheral blood cell products range [min–max]. Comparisons between cell products were performed using the Mann–Whitney test. Paired comparisons The median viability of BM cell products (n = 16) and PB used the non-parametric Wilcoxon test. Significance was set cell products (n = 8) was 95% [91–99] and 98% [94–99], as P < 0·05. The correlation between the different subtypes respectively, without any difference between both laboratories. of stem and progenitor cells used the non-parametric Spearman Cell counts of BM and PB cell products are presented in test. Statistical analyses were performed using SAS software Table 3. Red blood cells, platelets, total nucleated cells, v 8·0 (SAS Institute, Cary, NC, USA). lymphocytes and monocytes were in significantly higher concentration in PB than in BM cell products. The proportion of remaining mature granulocytes was low, with median Results values below 12·7% [2·0–47·5] and without any difference Owing to the fact that we were dealing with very rare events, between BM and PB cell products. As shown in Table 4, FC the reliability in both laboratories was assessed. Then, using analysis confirmed that CD2+ cells (T lymphocytes and this validated strategy of analysis, we characterized the cell natural killer cells), CD19+ B lymphocytes and CD14+ mono- therapy products. cytes were significantly more frequent in PB cell products.

260 J.-C. Capiod et al.

Table 3 Cell counts of BM- and PB-cell productsa (ii) raw MFI values to quantify the cells. As the cell products were concentrated after collection on a cell separator, BM cell products PB cell products platelets, debris, nude nuclei (megakaryocytic and erythroblastic b Med [min–max] Med [min–max] P cells) and non-viable cells were numerous and excluded by gating (Fig. 1a, histogram 1). These events represented a Haematocrit (%) 4·4 [2·4–12·6] 10·0 [6·4–21·0] 0·0012 median of 25·0% (range: 6·9–49·6) of total events, with no Platelets (109/l) 729 [461–1214] 1491 [916–1756] 0·0002 difference between BM and PB cell products. Total nucleated cells (109/l) 37·8 [18·1–71·5] 122·9 [67·7–165·9] 0·0001 Total MNCs (109/l) 32·4 [10·9–52·2] 108·6 [63·1–162·6] 0·0001 Lymphocytes (109/l) 18·0 [6·2–36·5] 57·9 [23·1–125·8] 0·0002 Cell differential and count in bone marrow- and Monocytes (109/l) 5·7 [2·1–12·4] 42·0 [21·2–61·7] 0·0001 peripheral blood-derived products Erythroblasts (109/l) 2·9 [0·4–13·6] – – The population of EPCs (CD34+/CD133+/VEGF-R2+) was a Other cellsc (109/l) 2·2 [0–4·1] – – highly discrete subset with small size and homogenous Mature granulocytes (109/l) 6·2 [1·9–33·9] 8·8 [3·3–31·6] 0·32 structure [forward scatter and side scatter low in FC]: median values observed for 4 × 105 analysed cells were 56 [6–295] a For each sample, the value of each parameter was calculated as the mean of and 4 [1–10] in BM and PB cell products, respectively (gate results from laboratory A and laboratory B; then median, min and max were D in histogram 5, Fig. 1a). These subsets were identified as calculated from the 16 BM- and the eight PB-cell products. Data were small clusters, clearly separable from the background and obtained by combining total cell counts and differentials performed after counting of 400 cells on May Grünwald Giemsa-stained cytospins. absent in the relevant isotypic control. Results of stem and bNon-parametric Mann–Whitney test. progenitor endothelial cell quantification are presented in cOther cells: blasts, immature granulocytes and plasma cells. Table 4. EPCs as well as CD34+, CD34+/CD133+ and CD34+/ VEGF-R2+ cells were significantly higher in BM cell products than in PB cell products. Among CD34+ cells, 55% [27–68] and 49% [21–82], co-expressed CD133+ in BM and PB cell Quantification of stem and progenitor endothelial products, respectively, without any difference between both cells in bone marrow and peripheral blood cell products (P = 0·39). EPCs represented 1·2% [0·2–8·5] and products 1·4% [0·3–8·0] of CD34+/CD133+ cells in BM MNC and PB Stem and progenitor endothelial cells being very rare events MNC products, respectively (P = 0·50). The frequency of (CD34+ cells coexpressing CD133 and VEGF-R2 was only CD34+ cells expressing CD133+ and VEGF-R2+ was only 0·002% of total peripheral blood mononuclear cells [15]), we 0·7% [0·1–3·6] and 0·8% [0·1–3·9] of the total CD34+ in BM used two methods for FC characterization: (i) determination of and PB cell products, respectively (P = 0·52). Among CD34+ the percentage of nucleated cells and quantification of the cells, 0·6% [0·1–3·8] and 1·9% [0·2–3·9] co-expressed absolute number of each category of cells in the cell products; VEGF-R2+ in BM and PB cell products, respectively

Table 4 Flow cytometry characterization of BM- and PB-cell productsa

BM cell products PB cell products Med [min–max] Med [min–max] Pb

CD2+ (109/l) 14·8 [4·6–29·2] 54·2 [21·7–111·8] 0·0001 CD19+ (109/l) 2·8 [1·1–8·0] 7·0 [0·9–20·2] 0·032 CD14+ (109/l) 4·6 [2·0–10·5] 37·4 [12·1–54·0] 0·0001 PCy5-CD34+ %c 2·27 [0·86–4·53] 0·09 [0·06–0·15] – (109/l) 0·78 [0·20–3·04] 0·10 [0·04–0·24] 0·0001 CD34+/CD133+ %c 1·21 [0·23–2·11] 0·04 [0·02–0·11] – (109/l) 0·47 [0·05–1·48] 0·05 [0·02–0·16] 0·0002 CD34+/VEGF-R2+ %c 0·017 [0·002–0·077] 0·002 [0·0003–0·004] – (109/l) 0·0057 [0·0014–0·0335] 0·0023 [0·0004–0·0046] 0·0234 CD34+/CD133+/VEGF-R2+ %c 0·014 [0·002–0·074] 0·001 [0·0003–0·002] – (109/l) 0·0047 [0·0009–0·0276] 0·0011 [0·0002–0·0034] 0·0022 aFor each sample, the value of each parameter was calculated as the mean of results from laboratory A and laboratory B; then median, min and max were calculated for the 16 BM-MNC and the eight PB-MNC products. bNon-parametric Mann–Whitney test. cPercentage of nucleated cells.

Fig. 1 (a) Quantification of endothelial progenitor cells (EPCs) by a four-colour flow cytometry (FC) analysis in mononuclear cells (MNC). (1) Morphometry analysis with side scatter (SS) vs. forward (FS); platelets and broken cells are excluded and relevant MNC were selected in gate K. (2) SS vs. ECD- or PCy7-CD45; relevant MNC were selected by counting 400 000 CD45+/– cells in gate A. (3) Among relevant MNC (AK), total CD34+ cells were selected in gate B. (4) CD34+/ CD133+ cells were selected in gate C. (5) CD34+/CD133+/VEGF-R2+ cells were selected in gate D. (6) and (7) To complete this analysing strategy, two histograms in fluorescence bi-parametric mode were used to show CD34+/VEGF-R2+ cells (histogram 6, gate H2) and CD34+/CD133+/–/VEGF-R2+/– cells (histogram 7). (b) Mean fluorescence intensity (MFI) expression. The mean fluorescence intensity and the MFI ratio (test/isotypic control) were calculated from 100% positive and negative cells [19].

(P = 0·086). In BM cell products, 87% [67–96] of CD34+/ CD34+/CD133+ cell quantifications (r = 0·98, P < 0·0001). VEGF-R2+ cells co-expressed CD133+ as compared to 50% The correlation between CD34+/VEGF-R2+ and CD34+/ [21–86] in PB cell products (P = 0·0009). CD133+/VEGF-R2+ cell concentrations was also high The individual distribution of stem and progenitor cells in (r = 0·94, P < 0·0001). In contrast, the correlation between the 24 cell products is presented on Fig. 2. A large dispersion CD34+ and CD34+/CD133+/VEGF-R2+ cell concentrations of values was observed in both BM and PB cell products. was only weakly positive (r = 0·52, P < 0·01) (Fig. 2a,b). Consequently, in some BM cell products, stem and progenitor cell concentrations were in the range of PB cell products (see, Cell quantification using mean fluorescence intensity for example, the cell product from patient no. 8 for stem cells values and from patient no. 24 for EPCs concentrations) (Fig. 2a,b). When the quantification of the different subsets was Whatever the origin of cell products (BM or PB), a strong expressed as MFI ratios (Fig. 1b), the results were as indicated positive correlation was observed between CD34+ and in Table 5. No significant differences were observed in MFI

Table 5 Mean ﬂuorescence intensity (MFI) expression of BM and PB cell productsa

BM-cell products PB-cell products Med [min–max] Med [min–max] Pb

PCy5-CD34+ MFI ratio 24·8 [6·8–64·9] 31·3 [27·0–41·5] 0·46 CD34+/CD133+ MFI ratio 3·5 [1·7–16·1] 3·9 [2·1–9·4] 0·80 CD34+/CD133+/ MFI ratio 15·7 [5·7–39·5] 18·1 [13·0–51·1] 0·35 VEGF-R2+

aFor each sample, the value of each parameter was calculated as the mean of results from laboratory A and laboratory B; then median, min and max were calculated for the 16 BM-MNC and the eight PB-MNC products. bNon-parametric Mann–Whitney test.

of the three markers (observed in 14 BM MNC samples and in three PB MNC samples), cell products could be classiﬁed in three clusters: low, medium or bright (Fig. 2c). Using MFI expression, the correlation between CD34+ and CD34+/CD133+ was weakly positive according to Spearman’s test (r = 0·46, P < 0·03). However, the correlation between CD34+ vs. CD34+/CD133+/VEGF-R2+ MFI expression was moderately positive (r = 0·61, P < 0·01) (Fig. 2c).

Discussion Precise characterization of cell products used is essential in clinical trials of cell therapy. We report here data on BM and PB cell products implanted in patients with CLI included in such a trial. Owing to the potential importance of EPCs, we focused on enumeration of these cells. Endothelial progenitor cells are rare events [23] and their quantification by FC is challenging. Using strict technical conditions and following published recommendations [24], we showed that accuracy and reliability between laboratories Fig. 2 Progenitor cells among bone marrow (BM)- and peripheral blood can be achieved. The results presented here are considerably (PB) mononuclear cells (MNC). (a) Individual stem cell concentrations strengthened by the fact that they were obtained concomitantly (PCy5-CD34+ and CD34+/CD133+), (b) endothelial progenitor cell in two different laboratories, demonstrating that sticking to concentrations [CD34+/VEGF-R2+ cells and immature endothelial rigorously established standard operating procedure [25] can progenitor cells (EPCs)] and (c) their mean fluorescence intensity (MFI) expression. For each sample, the values displayed are the mean of results provide reliable results in multicentre studies. Furthermore, obtained in laboratory A and laboratory B. BM cell products no. 6 and no. 8 the standard deviation for relatively rare populations is –1/2 came from the same patient. BM cell product no. 16 and PB cell product simply n where ‘n’ does the number of events comprise no. 19 came from the same patient. Numbers matched with the clinical trial the subset. Indeed, when one collects a million events, patient’s inclusion numbers. finding a single event in a gate is not meaningful [26]. Such standard operating procedures and new generation instruments could further improve EPCs quantification, mainly by shortening both preparation and analysis time by allowing expressions between BM and PB cell products; and that for the use of a wider range of conjugates and a faster acquisition either CD34+, CD34+/CD133+ or CD34+/CD133+/VEGF-R2+ of the large numbers of events necessary for reliable enumeration cells. However, MFI ratio comparison showed the presence of of such rare events as EPCs [27]. two types of cell products: without and with a homogeneous Contrary to previously reported studies, PB MNC were MFI ratio, that is, varying in the same way for CD34, CD133 collected in our trial without any previous mobilization (in and VEGF-R2 markers. Among the homogeneous MFI ratios order to avoid possible side-effects reported with the use of

© 2008 The Author(s) Journal compilation © 2008 International Society of Blood Transfusion, Vox Sanguinis (2009) 96, 256–265 Cell therapy in critical limb ischaemia 263 haematopoietic growth factors in patients with arterial the quantification of EPCs showed a large diversity of cellular diseases) [28,29]. In these circumstances, significant differences subpopulations, but their characterization needs further studies. were observed according to the origin of the MNC. As We did not observe any significant difference in the expected, the proportions of both mature and immature cell percentage of CD34+ cells co-expressing CD133 or VEGF-R2 subsets were different: mature cells (including lymphocytes, and of CD34+/CD133+ cells co-expressing VEGF-R2 be- monocytes and platelets) were in higher concentration in PB tween BM and PB cell products. Conversely, the percentage cell products, while EPCs as well as CD34+, CD34+/CD133+, of CD34+/VEGF-R2+ cells co-expressing the CD133 marker CD34+/VEGF-R2+ were significantly higher in BM cell was significantly higher in BM than in PB cell products. This products. Nevertheless, a great interindividual heterogeneity can be explained by the fact that EPCs loose CD133 when was observed: in some BM cell products, progenitor cell leaving BM and maturing [15]. Circulating CD34+/ concentration was as low as in PB cell products. This variability VEGF-R2+ cells could also include mature ECs after their may be related to the patient’s cardiovascular status or the detachment from the vessel wall [40]. However, mature ECs ongoing therapy [30,31]. All these data are to be interpreted cells have been characterized by their morphometry analysis taking into consideration that the mechanisms by which in FC (high size and heterogeneous structure) [41]. We thus MNCs can induce angiogenesis are not yet known. Experimental decided to exclude these cells in our gating strategy. data suggest that implanted cells, especially EPCs, are With either cell product (BM or PB), the correlation incorporated into vascular structures [32]. If this is true, the between CD34+ cells and EPCs concentrations was weakly use of BM MNC, which are rich in stem/progenitor cells, positive. This has to be kept in mind when choosing the best would be preferable. Other studies are in favour of an indirect parameter to characterize a cell product aimed at inducing effect of implanted MNC that may secrete cytokines and/or angiogenesis. Quantification of CD34+ cells is easy and growth factors [33]. PB MNC, even collected without any well-standardized. However, the CD34 marker is not specific previous mobilization, would be as effective as BM MNC. The to angiogenic stem cells. paracrine effect of monocytes has been reported [34]. In order to assess whether or not the epitope density on Platelets are a source of growth factors [35], and products different cell types could influence the results when obtained by apheresis are currently rich in platelets. expressed as a percentage of positive cells, we analysed raw Lymphocytes could also play a role as they release numerous MFI. Using the ratio defined in the material and method factors potentially involved in angiogenesis [36]. Finally, the section, similar MFI ratios for CD34, CD133 and VEGF-R2 quantity of MNC that should be implanted for optimal were observed when comparing BM and PB cell products. efficacy is unknown. However, the MFI ratios were very different between cell There is no consensus regarding the immunophenotype of products, whether the origin was BM or blood. This variability EPCs. As CD45 expression on EPCs is controversial [37,38], may indicate differences between donors. Of interest, we we chose to analyse CD45– and CD45+ cells. For this, noted that MFI ratios in two BM cell products that were aggregated platelets and nude nuclei were excluded. In harvested in the same patient were similar. Besides, MFI accordance to most previous reports, we defined EPCs by the should allow determining the level of expression of CD34, co-expression of CD34+/CD133+/VEGFR-2+. However, it CD133, VEGF-R2 and possible other markers. The next step has recently been reported that these cells, when isolated, are would be to correlate the epitope density of such markers unable to generate endothelial cells in culture [23]. Subsequently, with the functional capacity of cells [42,43]. This approach the co-expression of CD34+/VEGF-R2+ appears as the best should be challenged and validated by clinical trials. combination to define EPCs [39]. In large multicentre trials that will be necessary to prove In our study, the cellular therapy product obtained from the efficacy of cell therapy in CLI, it will be crucial to take BM presents analogies with cellular therapy products used in into account the characterization of the cell products used. CLI in other studies, in terms of total amount of implanted Owing to the fact that the involved mechanisms are not yet BM MNC, of CD34+ cell amount, and of percentage of CD34+ known, mature as well as stem/progenitor cells have to be cells among the BM-MNC [1,3,5]. The cellular therapy considered. If EPCs turned out to be the active cells, then, a products obtained from PB in our study (without previous consensual definition of these cells should be proposed. Other mobilization) was less concentrated, when compared to stem/progenitor cells, such as mesenchymal stem cells [44] cellular therapy products used (after previous mobilization or multipotent adult progenitor cells [45], possibly present in by granulocyte–colony-stimulating factor) in CLI in other cell products, may also be worthwhile to consider. studies, in terms of total amount of implanted PB MNC, of CD34+ cell amount, and of percentage of CD34+ cells among the PB MNC [4,6]. In these studies, the characterization of Acknowledgements cellular therapy products was limited to the quantification of The study is registered on clinicaltrials.gov under the number: total MNC and of CD34+ cells. In our study, the MFI ratio and NCT00533104. Furthermore, the authors appreciated the

© 2008 The Author(s) Journal compilation © 2008 International Society of Blood Transfusion, Vox Sanguinis (2009) 96, 256–265 264 J.-C. Capiod et al. contribution of Sylvie Remy, Valérie Creuza, Catherine Massé, angioblasts, angiogenic ligands, and cytokines. Circulation 2001; and Jacques Vigne. This study was funded by a clinical 104:1046–1052 research hospital programme grant (French ministry of 12 Iba O, Matsubara H, Nozawa Y, Fujiyama S, Amano K, Mori Y, health, PHRC 2003). Kojima H, Iwasaka T: Angiogenesis by implantation of peripheral blood mononuclear cells and platelets into ischemic limbs. Circulation 2002; 106:2019–2025 References 13 Hernandez P, Cortina L, Artaza H, Pol N, Lam RM, Dorticos E, Macias C, Hernandez C, del Valle L, Blanco A, Martinez A, 1 Tateishi-Yuyama E, Matsubara H, Murohara T, Ikeda U, Shintani Diaz F: Autologous bone-marrow mononuclear cell implantation S, Masaki H, Amano K, Kishimoto Y, Yoshimoto K, Akashi H, in patients with severe lower limb ischaemia: a comparison of Shimada K, Iwasaka T, Imaizumi T: Therapeutic angiogenesis for using blood cell separator and Ficoll density gradient centrifuga- patients with limb ischaemia by autologous transplantation of tion. Atherosclerosis 2007; 194:e52–e56 bone-marrow cells: a pilot study and a randomised controlled 14 Sutherland DR, Anderson L, Keeney M, Nayar R, Chin-Yee I: The trial. Lancet 2002; 360:427–435 ISHAGE guidelines for CD34+ cell determination by flow 2 Saigawa T, Kato K, Ozawa T, Toba K, Makiyama Y, Minagawa S, cytometry. International Society of Hematotherapy and Graft Hashimoto S, Furukawa T, Nakamura Y, Hanawa H, Kodama M, Engineering. J Hematother 1996; 5:213–226 Yoshimura N, Fujiwara H, Namura O, Sogawa M, Hayashi J, 15 Peichev M, Naiyer AJ, Pereira D, Zhu Z, Lane WJ, Williams M, Aizawa Y: Clinical application of bone marrow implantation in Oz MC, Hicklin DJ, Witte L, Moore MA, Rafii S: Expression of patients with arteriosclerosis obliterans, and the association VEGFR-2 and AC133 by circulating human CD34(+) cells between efficacy and the number of implanted bone marrow identifies a population of functional endothelial precursors. cells. Circ J 2004; 68:1189–1193 Blood 2000; 95:952–958 3 Higashi Y, Kimura M, Hara K, Noma K, Jitsuiki D, Nakagawa K, 16 Gehling UM, Ergun S, Schumacher U, Wagener C, Pantel K, Otte M, Oshima T, Chayama K, Sueda T, Goto C, Matsubara H, Murohara Schuch G, Schafhausen P, Mende T, Kilic N, Kluge K, Schafer B, T, Yoshizumi M: Autologous bone-marrow mononuclear cell Hossfeld DK, Fiedler W: In vitro differentiation of endothelial implantation improves endothelium-dependent vasodilation in cells from AC133-positive progenitor cells. Blood 2000; 95:3106 – patients with limb ischemia. Circulation 2004; 109:1215–1218 3112 4 Huang P, Li S, Han M, Xiao Z, Yang R, Han ZC: Autologous 17 Quirici N, Soligo D, Caneva L, Servida F, Bossolasco P, Deliliers transplantation of granulocyte colony-stimulating factor- GL: Differentiation and expansion of endothelial cells from mobilized peripheral blood mononuclear cells improves critical human bone marrow CD133(+) cells. Br J Haematol 2001; limb ischemia in diabetes. Diabetes Care 2005; 28:2155–2160 115:186–194 5 Durdu S, Akar AR, Arat M, Sancak T, Eren NT, Ozyurda U: 18 Yin AH, Miraglia S, Zanjani ED, Almeida-Porada G, Ogawa M, Autologous bone-marrow mononuclear cell implantation for Leary AG, Olweus J, Kearney J, Buck DW: AC133, a novel patients with Rutherford grade II–III thromboangiitis obliterans. marker for human hematopoietic stem and progenitor cells. J Vasc Surg 2006; 44:732–739 Blood 1997; 90:5002–5012 6 Kawamura A, Horie T, Tsuda I, Abe Y, Yamada M, Egawa H, 19 Satoh C, Dan K, Yamashita T, Jo R, Tamura H, Ogata K: Flow Iida J, Sakata H, Onodera K, Tamaki T, Furui H, Kukita K, Meguro J, cytometric parameters with little interexaminer variability for Yonekawa M, Tanaka S: Clinical study of therapeutic angiogenesis diagnosing low-grade myelodysplastic syndromes. Leuk Res by autologous peripheral blood stem cell (PBSC) transplantation 2008; 32:699–707 in 92 patients with critically ischemic limbs. J Artif Organs 20 Katoh O, Tauchi H, Kawaishi K, Kimura A, Satow Y: Expression 2006; 9:226–233 of the vascular endothelial growth factor (VEGF) receptor gene, 7 Asahara T, Murohara T, Sullivan A, Silver M, van der Zee R, Li T, KDR, in hematopoietic cells and inhibitory effect of VEGF on Witzenbichler B, Schatteman G, Isner JM: Isolation of putative apoptotic cell death caused by ionizing radiation. Cancer Res progenitor endothelial cells for angiogenesis. Science 1997; 1995; 55:5687–5692 275:964–967 21 Ieta K, Tanaka F, Haraguchi N, Kita Y, Sakashita H, Mimori K, 8 Isner JM, Asahara T: Angiogenesis and vasculogenesis as Matsumoto T, Inoue H, Kuwano H, Mori M: Biological and therapeutic strategies for postnatal neovascularization. J Clin genetic characteristics of tumor-initiating cells in colon cancer. Invest 1999; 103:1231–1236 Ann Surg Oncol 2008; 15:638–648 9 Shi Q, Rafii S, Wu MH, Wijelath ES, Yu C, Ishida A, Fujita Y, 22 Bland JM, Altman DG: Statistical methods for assessing agree- Kothari S, Mohle R, Sauvage LR, Moore MA, Storb RF, ment between two methods of clinical measurement. Lancet Hammond WP: Evidence for circulating bone marrow-derived 1986; 1:307–310 endothelial cells. Blood 1998; 92:362–367 23 Case J, Mead LE, Bessler WK, Prater D, White HA, Saadatzadeh 10 Urbich C, Dimmeler S: Endothelial progenitor cells: characteriza- MR, Bhavsar JR, Yoder MC, Haneline LS, Ingram DA: Human tion and role in vascular biology. Circ Res 2004; 95:343–353 CD34+AC133+VEGFR-2+ cells are not endothelial progenitor 11 Kamihata H, Matsubara H, Nishiue T, Fujiyama S, Tsutsumi Y, cells but distinct, primitive hematopoietic progenitors. Exp Ozono R, Masaki H, Mori Y, Iba O, Tateishi E, Kosaki A, Shintani Hematol 2007; 35:1109 –1118 S, Murohara T, Imaizumi T, Iwasaka T: Implantation of bone 24 Khan SS, Solomon MA, McCoy JP, Jr: Detection of circulating marrow mononuclear cells into ischemic myocardium enhances endothelial cells and endothelial progenitor cells by flow cytometry. collateral perfusion and regional function via side supply of Cytometry B Clin Cytom 2005; 64:1–8

25 Rustemeyer P, Wittkowski W, Jurk K, Koller A: Optimized flow 35 Stellos K, Gawaz M: Platelet interaction with progenitor cells: cytometric analysis of endothelial progenitor cells in peripheral potential implications for regenerative medicine. Thromb Haemost blood. J Immunoassay Immunochem 2006; 27:77–88 2007; 98:922–929 26 Roederer M: How many events is enough? Are you positive? 36 Hur J, Yang HM, Yoon CH, Lee CS, Park KW, Kim JH, Kim TY, Cytometry A 2008; 73:384–385 Kim JY, Kang HJ, Chae IH, Oh BH, Park YB, Kim HS: Identifica- 27 Shaffer RG, Greene S, Arshi A, Supple G, Bantly A, Moore JS, tion of a novel role of T cells in postnatal vasculogenesis: Mohler ER, 3rd: Flow cytometric measurement of circulating characterization of endothelial progenitor cell colonies. Circulation endothelial cells: the effect of age and peripheral arterial disease 2007; 116:1671–1682 on baseline levels of mature and progenitor populations. Cytometry 37 Schomig K, Busch G, Steppich B, Sepp D, Kaufmann J, Stein A, B Clin Cytom 2006; 70:56–62 Schomig A, Ott I: Interleukin-8 is associated with circulating 28 Kawachi Y, Watanabe A, Uchida T, Yoshizawa K, Kurooka N, CD133 + progenitor cells in acute myocardial infarction. Eur Setsu K: Acute arterial thrombosis due to platelet aggregation Heart J 2006; 27:1032–1037 in a patient receiving granulocyte colony-stimulating factor. 38 Timmermans F, Van Hauwermeiren F, De Smedt M, Raedt R, Br J Haematol 1996; 94:413–416 Plasschaert F, De Buyzere ML, Gillebert TC, Plum J, Vandekerck- 29 Kang HJ, Kim HS, Zhang SY, Park KW, Cho HJ, Koo BK, Kim YJ, hove B: Endothelial outgrowth cells are not derived from Soo Lee D, Sohn DW, Han KS, Oh BH, Lee MM, Park YB: Effects CD133+ cells or CD45+ hematopoietic precursors. Arterioscler of intracoronary infusion of peripheral blood stem-cells Thromb Vasc Biol 2007; 27:1572–1579 mobilised with granulocyte-colony stimulating factor on left 39 Fadini GP, Avogaro A, Agostini C: Critical assessment of ventricular systolic function and restenosis after coronary putative endothelial progenitor phenotypes. Exp Hematol 2007; stenting in myocardial infarction: the MAGIC cell randomised 35: 1479–1480; author reply 1481–1482 clinical trial. Lancet 2004; 363:751–756 40 Mutin M, Canavy I, Blann A, Bory M, Sampol J, Dignat-George F: 30 Hill JM, Zalos G, Halcox JP, Schenke WH, Waclawiw MA, Direct evidence of endothelial injury in acute myocardial Quyyumi AA, Finkel T: Circulating endothelial progenitor cells, infarction and unstable angina by demonstration of circulating vascular function, and cardiovascular risk. N Engl J Med 2003; endothelial cells. Blood 1999; 93:2951–2958 348:593–600 41 Erdbruegger U, Haubitz M, Woywodt A: Circulating endothelial 31 Hristov M, Weber C: Endothelial progenitor cells: characterization, cells: a novel marker of endothelial damage. Clin Chim Acta pathophysiology, and possible clinical relevance. J Cell Mol Med 2006; 373:17–26 2004; 8:498–508 42 Ingram DA, Mead LE, Tanaka H, Meade V, Fenoglio A, Mortell K, 32 Yoon CH, Hur J, Park KW, Kim JH, Lee CS, Oh IY, Kim TY, Cho HJ, Pollok K, Ferkowicz MJ, Gilley D, Yoder MC: Identification of a Kang HJ, Chae IH, Yang HK, Oh BH, Park YB, Kim HS: Synergistic novel hierarchy of endothelial progenitor cells using human neovascularization by mixed transplantation of early endothelial peripheral and umbilical cord blood. Blood 2004; 104:2752– progenitor cells and late outgrowth endothelial cells: the role of 2760 angiogenic cytokines and matrix metalloproteinases. Circulation 43 Yoder MC, Mead LE, Prater D, Krier TR, Mroueh KN, Li F, Krasich R, 2005; 112:1618–1627 Temm CJ, Prchal JT, Ingram DA: Redefining endothelial progenitor 33 Zhou B, Liu PX, Lan HF, Fang ZH, Han ZB, Ren H, Poon MC, cells via clonal analysis and hematopoietic stem/progenitor cell Han ZC: Enhancement of neovascularization with mobilized principals. Blood 2007; 109:1801–1809 blood cells transplantaion: supply of angioblasts and angiogenic 44 Park JS, Huang NF, Kurpinski KT, Patel S, Hsu S, Li S: Mechano- cytokines. J Cell Biochem 2007; 102:183–195 biology of mesenchymal stem cells and their use in cardiovascular 34 Rehman J, Li J, Orschell CM, March KL: Peripheral blood repair. Front Biosci 2007; 12:5098–5116 ‘endothelial progenitor cells’ are derived from monocyte/ 45 Reyes M, Dudek A, Jahagirdar B, Koodie L, Marker PH, Verfaillie CM: macrophages and secrete angiogenic growth factors. Circulation Origin of endothelial progenitors in human postnatal bone 2003; 107:1164–1169 marrow. J Clin Invest 2002; 109:337–346