CD34 is a transmembrane protein that is strongly expressed on hematopoietic stem/progenitor cells (HSCs); despite its importance as a marker of HSCs, its function is still poorly understood, although a role in cell adhesion has been demonstrated. To characterize the function of CD34 antigen on human HSCs, we examined, by both inhibition and overexpression, the role of CD34 in the regulation of HSC lineage differentiation. Our results demonstrate that CD34 silencing enhances HSC granulocyte and megakaryocyte differentiation and reduces erythroid maturation. In agreement with these results, the gene expression profile of these cells reveals the upregulation of genes involved in granulocyte and megakaryocyte differentiation and the downregulation of erythroid genes. Consistently, retroviral-mediated CD34 overexpression leads to a remarkable increase in erythroid progenitors and a dramatic decrease in granulocyte progenitors, as evaluated by clonogenic assay. Together, these data indicate that the CD34 molecule promotes the differentiation of CD34+ hematopoietic progenitors toward the erythroid lineage, which is achieved, at least in part, at the expense of granulocyte and megakaryocyte lineages.
Disclosure of potential conflicts of interest is found at the end of this article.
CD34 is a highly glycosylated transmembrane protein, a member of the sialomucin family of surface antigens. CD34 is strongly expressed on hematopoietic stem/progenitor cells (HSCs), and its expression shows a progressive and rapid decline as HSCs undergo differentiation . Since its discovery, CD34 has become the most widely used marker for the enrichment of human hematopoietic progenitors, but despite its importance, its function and regulation are not completely understood .
The CD34 molecule is a 115-kDa type I transmembrane glycoprotein with a protein backbone of 40 kDa that shows no significant sequence homology to any other known protein . The heavily glycosylated extracellular region of CD34 is similar to that of leukosialin (CD43), a hematopoietic cell-specific sialoglycoprotein expressed by leukocytes . The cytoplasmic domain bears a consensus site for tyrosine phosphorylation and two potential targeting sites for protein kinase C, but it lacks intrinsic enzymatic activity . Studies on CD34 function suggest that it may play a role in cell adhesion and signal transduction in hematopoietic stem/progenitor cells. In fact, the cross-linking of CD34 with specific monoclonal antibodies (MoAbs) induces F-actin reorganization, resulting in increased tyrosine phosphorylation through the activation of the nonreceptor-type tyrosine kinases Syk and Lyn at the capping site . Moreover, some reports have shown that engagement of particular CD34 epitopes by anti-CD34 MoAbs induces actin polymerization and homotypic adhesion in KG1a and CD34+ cells, indicating that CD34 is capable of transducing signals throughout the cell membrane [7, –9].
The natural ligand for CD34 has not yet been identified in the hematopoietic microenvironment, whereas CD34 expressed on murine high-endothelial venule cells in the lymph nodes binds to the lymphocyte homing receptor l-selectin . Recent experiments indicate that CD34 expressed on endothelial cells may play a role in leukocyte adhesion and homing during the inflammatory process. These studies raise the interesting possibility that CD34+ hematopoietic stem/progenitor cells may localize to the bone marrow (BM) as a result of binding to an l-selectin-like protein . It is possible that CD34 antigen mediates the interaction between HSCs and the BM stromal lectins, favoring the binding of locally released growth factors and influencing the compartmentalization of stem/progenitor cells. Furthermore, it has been suggested that CD34 backbone may act as a scaffold for the attachment of specific glycans, which in turn mediate progenitor cell binding to stromal lectins or to other components of the hematopoietic niche within the bone marrow .
CD34 knockout (KO) mice are viable, develop normally, and show a typical hematopoietic profile of adult blood; however, these mice have subtle abnormalities in hematopoiesis . Cheng et al. reported that hematopoietic development in embryoid bodies deleted for the CD34 gene is delayed in both erythroid and myeloid differentiation; moreover, hematopoietic progenitors from the yolk sac, the fetal liver, and the adult BM are decreased in CD34-null mice, and these progenitors are unable to expand in liquid cultures in response to hematopoietic growth factors . These results strongly suggest that CD34 antigen is involved in the proliferation and/or maintenance of progenitor cells during both embryonic and adult hematopoiesis. Furthermore, previous data from our laboratory demonstrate that induction of CD34 antigen during HSC maturation correlates with cell cycle activation and expression of a large variety of differentiation markers [13, 14]. These data suggest that CD34 is essential for critical HSC functions, such as adhesion, differentiation, and proliferation, in the hematopoietic microenvironment. Despite the efforts made so far, the functional role of CD34 antigen in the hematopoietic system remains unclear. To better characterize the function of this glycoprotein on human hematopoietic stem/progenitor cells, we examined, both by short interfering RNA (siRNA)-mediated gene silencing and by retroviral-mediated overexpression, the role of CD34 antigen in the regulation of HSC lineage differentiation.
In this study, we show that CD34 interferes with the differentiation capacity of hematopoietic progenitor cells, and in particular, CD34 favors differentiation toward the erythroid lineage, which is achieved at least in part at the expense of the granulocyte and megakaryocyte lineages. Thus, our results provide the first demonstration of a functional role for CD34 antigen in hematopoietic stem/progenitor cell differentiation.
Materials and Methods
CD34+ Stem/Progenitor Cells Purification
After donors' informed consent was received, human CD34+ cells were purified from umbilical cord blood (CB) samples, collected after normal deliveries, according to the institutional guidelines for discarded material. These CBs are unsuitable for transplantation because of insufficient cellularity. Mononuclear cells were isolated by Ficoll-Hypaque (Lymphoprep; Nycomed Pharma, Oslo, Norway, http://www.nycomed.com) gradient separation and washed twice with phosphate-buffered saline, and then CD34+ cells were separated using magnetic cell sorting procedure (EasySep Human CD34+ positive selection kit; StemCell Technologies, Vancouver, BC, Canada, http://www.stemcell.com). CD34+ cell purity assessed by flow cytometry was always >95%.
Transfection and siRNA Treatment
Nucleofection of CD34+ stem/progenitor cells was performed according to the protocol provided by the manufacturer, with main modifications (Amaxa, Cologne, Germany, http://www.amaxa.com). Twenty-four hours prior to electroporation, the primary CD34+ hematopoietic cells were seeded in 24-well plates at a density of 5 × 105 cells per milliliter in Iscove's modified Dulbecco's medium (IMDM) (Gibco, Grand Island, NY, http://www.invitrogen.com) containing 20% human serum (Lonza, Milano, Italy, http://www.lonza.com), stem cell factor (50 ng/ml), Flt3-ligand (Flt3-l) (50 ng/ml), thrombopoietin (TPO) (20 ng/ml), interleukin (IL)-6 (10 ng/ml), and IL-3 (10 ng/ml) (all from R&D Systems Inc., Minneapolis, http://www.rndsystems.com). CD34+ cells were subjected to three cycles of nucleofection every 24 hours. Briefly, for each electroporation, 5 × 105 cells were gently resuspended in 100 μl of Human CD34+ Nucleofector Solution (Amaxa), mixed with 5 μg of a mixture of four siRNAs targeting CD34 mRNA (supplemental online Table 1) (SMARTpool; Dharmacon, Inc., Lafayette, CO, http://www.dharmacon.com), and pulsed with the program U-01. As a control, in a separate sample, a nontargeting siRNA (siCONTROL Non-Targeting Pool; Dharmacon) was used. Immediately after nucleofection, cells were transferred into prewarmed fresh medium in 24-well plates. Cells were analyzed 24, 48, 72, and 96 hours after the last nucleofection for CD34 antigen expression.
To achieve an optimal expansion and differentiation of the primary CD34+ hematopoietic stem/progenitor cells, a serum-free liquid culture was performed after the last nucleofection by seeding CD34+ cells in 12-well plates at a density of 2 × 105 cells per milliliter in IMDM (Gibco) containing 20% BIT serum substitute (bovine serum albumin, insulin, and transferrin; StemCell Technologies), SCF (50 ng/ml), Flt3-ligand (Flt3-l) (50 ng/ml), IL-6 (10 ng/ml), and TPO (10 ng/ml) (all from R&D Systems).
Retroviral Vector Construction and Packaging
The CD34 cDNA sequence was generated by reverse transcription-polymerase chain reaction (PCR) amplification of total RNA extracted from CB CD34+ cells using CD34 direct (5′-TTT TTT AAG CTT ATG CTG GTC CGC AGG GGC GCG-3′) and CD34 reverse primers (5′-TTT TTT GAA TTC TCA GGG TTC CAG CTC CAG CCT TTC TCC-3′). PCR was carried out using a proofreading thermostable DNA polymerase, and the amplified fragment was inserted in the pCR2.1 T/A cloning vector (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), where it was fully sequenced to exclude polymerase-induced mutations. Full-length CD34 cDNA sequences were then EcoRI-excised and cloned in the EcoRI site of LXIΔN vector , resulting in the construction of LCD34IΔN retroviral vector.
Packaging lines for the described constructs were generated by transinfection in the ecotropic Phoenix and amphotropic GP+envAm12 cells, as previously described . Viral titers were assessed by flow cytometry analysis of a truncated version of low-affinity nerve growth factor receptor (ΔLNGFR) expression percentage upon infection of KG1 cells and CD34+ hematopoietic progenitors.
Hematopoietic Cell Transduction and Purification
Transduction of human CB CD34+ progenitors was performed after 48 hours of preactivation by two cycles of infection (12 hours each) with viral supernatant in the presence of polybrene (8 μg/ml) on retronectin-coated plates (10 μg/cm2) followed by 48 hours postincubation in the liquid culture conditions described above. Transduced CD34+ cells were subsequently purified by an immunomagnetic procedure using mouse anti-human p75-NGFR MoAb and tiny fluorescence-activated cell sorting-compatible magnetic nanoparticles in a column-free magnetic system (EasySep “Do-It-Yourself” Selection Kit; StemCell Technologies) following the manufacturer's guidelines.
Viability Assay and Cell Cycle Analysis
Viability measurement was assessed by trypan blue exclusion assay 12 hours after each nucleofection, as well as after two cycles of infection and immunomagnetic purification . In a Neubauer chamber, at least 100 cells were microscopically analyzed in duplicate for viability. The mean percentage of living cells of the two analyses was calculated.
Cell cycle analysis was performed 48 hours after the last electroporation according to Nicoletti et al. . The values relative to viability assay and cell cycle analysis are reported as mean ± 2 SEM from five independent experiments.
Morphological and Immunophenotypic Analysis
Transduction efficiency of CD34+ cells was monitored by flow cytometric analysis (Epics XL; Coulter Electronics Inc., Hialeah, FL, http://www.beckmancoulter.com) of ΔLNGFR expression using a murine anti-human p75-NGFR MoAb (BD Biosciences, San Diego, http://www.bdbiosciences.com) and fluorescein isothiocyanate (FITC) rabbit anti-mouse IgG (Dako, Copenhagen, Denmark, http://www.dako.com). The CD34 antigen was monitored using a mouse anti-CD34 MoAb conjugated to FITC (Miltenyi Biotec, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com).
Differentiation of CD34+ cells was monitored by morphological analysis of May-Grunwald-Giemsa-stained cytospins and by flow cytometric analysis of CD41, CD66b, and glycophorin A (GPA) surface antigen expression, performed at days 10, 12, and 14 after the last nucleofection or postinfection. The images were captured with an Axioskop 40 microscope (Carl Zeiss MicroImaging Inc., Thornwood, NY, http://www.zeiss.com/micro) by means of an AxioCam HRc digital camera (Carl Zeiss MicroImaging) and Axiovision software, version 3.1 (Carl Zeiss MicroImaging).
The following MoAbs were used for flow cytometric analysis: FITC-conjugated mouse anti-human CD41 MoAb, phycoerythrin (PE)-conjugated mouse anti-human CD66b MoAb, and PE-conjugated mouse anti-human GPA MoAb (all from Becton, Dickinson and Company, Mountain View, CA, http://www.bd.com). Myeloperoxidase (MPO) intracellular staining was performed using a FIX & PERM cell permeabilization kit (Caltag Laboratories, Burlingame, CA, http://www.caltag.com) and FITC-conjugated mouse anti-human MPO (Caltag Laboratories). The values relative to morphological and immunophenotypic analysis are reported as mean ± 2 SEM from 13 independent experiments.
Methylcellulose Clonogenic Assay
Human colony-forming units (CFU) were cultured in methylcellulose as previously described . CD34+ cells were plated in triplicate in MethoCult GF H4434 “Complete” Methylcellulose Medium (StemCell Technologies) containing a cocktail of the following recombinant human cytokines: SCF (50 ng/ml), granulocyte-macrophage colony-stimulating factor (GM-CSF; 10 ng/ml), IL-3 (10 ng/ml), and erythropoietin (3 U/ml). After 14 days of culture at 37°C in a humidified atmosphere with 5% CO2, erythroid burst-forming units (BFU-E), erythroid colony-forming units (CFU-E), granulocyte colony-forming units (CFU-G), macrophage colony-forming units (CFU-M), granulocyte-macrophage colony-forming units (CFU-GM), and granulocyte/erythrocyte/macrophage/megakaryocyte colony-forming units (CFU-GEMM) were scored. The values are reported as mean ± 2 SEM from 13 independent experiments.
Collagen Clonogenic Assay
Megakaryocyte colony-forming units (CFU-MK) were assayed by using a commercial megakaryocyte assay detection kit (MegaCult-C; StemCell Technologies). CD34+ cells were cultured at a density of 2.5 × 103 cells per milliliter in collagen-based medium, and 0.75 ml of the suspension was seeded in each chamber of a double-chamber slide. This collagen-based system contained a medium supplemented with 1.1 mg/ml collagen, 1% bovine serum albumin, 0.01 mg/ml bovine pancreatic insulin, 0.2 mg/ml human transferrin (iron-saturated), and the following human recombinant cytokines: 50 ng/ml TPO, 10 ng/ml IL-3, and 10 ng/ml IL-6. The chamber slides were incubated at 37°C for 12 days and then fixed for 20 minutes in 1:3 methanol/acetone. Megakaryocyte colonies were stained using a primary monoclonal anti-CD41 (GPIIb/IIIa) antibody and then identified by using an alkaline phosphatase/naphthol detection system (all from StemCell Technologies). The cells were counterstained with Evans Blue, causing the cell nuclei to turn blue, regardless of lineage. CD41-positive colonies were scored as CFU-MK. The values are reported as mean ± 2 SEM from 13 independent experiments.
RNA Extraction and Microarray Analysis
Total cellular RNA was extracted 48 hours after the last nucleofection or 3 days postinfection from 0.5 × 106 cells of each sample using the RNeasy Micro kit (Qiagen, Valencia, CA, http://www1.qiagen.com) following the protocol supplied by the manufacturer. Disposable RNA chips (Agilent RNA 6000 Nano LabChip kit; Agilent Technologies, Waldbrunn, Germany, http://www.agilent.com) were used to determine the concentration and purity/integrity of RNA samples using an Agilent 2100 Bioanalyzer. RNAs originating from three different experiments were pooled to obtain at least 2 μg per sample. One-cycle target labeling assays, as well as the Affymetrix Human HG-U133A GeneChip array hybridization, staining, and scanning, were performed, using Affymetrix standard protocols (Affymetrix, Santa Clara, CA, http://www.affymetrix.com) . The GeneChip Operating Software (GCOS) absolute analysis algorithm was used to determine the amount of a transcript mRNA (signal), whereas the GCOS comparison analysis algorithm was used to compare gene expression levels between two samples.
Present genes were selected as the sequences showing the detection call present and signal >100 at least in one sample. Differentially expressed genes were selected as the sequences showing a change call increased or decreased and signal log ratio ≥1 or ≤−1 in the pairwise comparisons between CD34 siRNA-treated (CD34siRNA) and Mock-tranfected (MOCK)/Negative Control (NegCTR) or LCD34IΔN CD34+ and MOCK/LXIΔN CD34+ cells. The gene list passing this filter was selected as “changing genes.” All data have been deposited in the Gene Expression Omnibus MIAME-compliant public database at http://www.ncbi.nlm.nih.gov/geo (supplemental online Table 3). To validate microarray data, we selected a set of differentiation-related genes, which were monitored by real-time quantitative (QRT) PCR performed as reported in the supplemental online data.
The statistics used for data analysis were based on the two-tailed Student t test for comparison of averages in paired samples. In silencing experiments, MOCK versus NegCTR, MOCK versus CD34siRNA, and NegCTR versus CD34siRNA were compared; similarly, in overexpression experiments, MOCK versus LXIΔN, MOCK versus LCD34IΔN, and LXIΔN versus LCD34IΔN were compared. Data were analyzed using Excel 2007 (Microsoft, Redmond, WA, http://www.microsoft.com), and p < .05 was considered significant.
CD34 Gene Silencing in CD34+ Stem/Progenitor Cells
To investigate the role of CD34 in the hematopoietic differentiation, we silenced its expression in CD34+ stem/progenitor cells using an siRNA approach. To maximize siRNA transfection efficiency, we used the Nucleofector technology (Amaxa) by optimizing CD34+ cell nucleofection protocol in terms of number and frequency of electroporations, as well as cell culture conditions pre- and post-transfection, as detailed in Materials and Methods. In a set of 13 independent experiments, CD34+ cells were transfected with a mixture of four siRNAs targeting CD34 mRNA (supplemental online Table 1) and with a nontargeting siRNA as a NegCTR.
All the experiments were performed starting with CD34+ cells freshly isolated from CB, with cell viability greater than 90%. Trypan blue exclusion assay showed a slightly decreased viability in siRNA-transfected cells compared with controls; however, there were no statistically significant differences in viability among MOCK, NegCTR, and CD34siRNA cells (supplemental online Fig. 1A).
The expression level of CD34 on control cells (MOCK and NegCTR) and CD34siRNA was assessed by immunofluorescence analysis at 24, 48, 72, and 96 hours after the last nucleofection. Flow cytometric analysis showed a significant decrease in CD34 in CD34siRNA cells compared with control samples, as detailed in Figure 1. The data demonstrate that 48 hours after the last nucleofection, siRNA-treated cells showed the minimum level of CD34 expression and the maximum difference from controls (Fig. 1). Consequently, we have chosen day 2 after the last nucleofection for seeding the cells in liquid or semisolid medium after gene silencing. CD34 inhibition remained stable in transfected cells up to 4 days after last nucleofection (Fig. 1).
Biological Effects of CD34 Gene Silencing on Proliferation and Differentiation Capacity of CD34+ Cells
To study the effect of CD34 inhibition on CD34+ cell proliferation, we performed a flow cytometry analysis of cell cycle using a propidium iodide staining technique. Cell cycle analysis 48 hours after the last nucleofection showed no differences in cell cycle distribution between controls and CD34siRNA cells (supplemental online Fig. 2). The effects exerted by siRNA-mediated CD34 gene silencing were also assessed at days 10, 12, and 14 of serum-free liquid culture after the last nucleofection by both morphological and flow cytometric analyses.
Flow cytometric analysis of lineage differentiation markers, performed at day 12 of liquid culture after the last nucleofection, showed that CD34siRNA cells had higher expression of granulocyte markers CD66b and MPO, as well as a significant increase in the megakaryocyte marker CD41; on the contrary, the erythroid marker GPA showed a significant decrease in CD34siRNA compared with NegCTR and MOCK cells (Fig. 2A, 2C).
Morphological analysis, performed at day 14 of liquid culture after the last nucleofection (Fig. 2B), revealed that CD34siRNA cells had a strong enrichment in granulocyte and megakaryocyte precursors at different stages of maturation, whereas the erythroid elements were poorly represented compared with MOCK and NegCTR. Differential cell count and statistical analysis are reported in Figure 2D.
To better characterize the role of CD34 in stem/progenitor cell differentiation, cells were plated in methylcellulose-based medium at day 2 after the last nucleofection in a set of 13 independent experiments. According to the lack of interference of CD34 silencing with the proliferative activity, clonogenic assay did not show differences in total colony number, demonstrating that CD34 antigen does not influence the clonogenic capacity of CD34+ stem/progenitor cells (supplemental online Table 2). On the other hand, the methylcellulose-based clonogenic assay showed a significant increase in the percentage of CFU-G in CD34siRNA cells, coupled to a decrease in BFU-E (Fig. 3A). Absolute numbers of colonies are reported in supplemental online Table 2. These results indicate that the early silencing of CD34 enhances granulocyte differentiation at the expense of erythroid differentiation.
Next, we examined the effect of CD34 silencing on megakaryocyte differentiation using a collagen-based culture system that supports megakaryopoietic progenitor growth in vitro. For this purpose, controls and CD34siRNA cells were plated in collagen-based medium at day 2 after the last nucleofection in a set of 13 independent experiments. The results obtained, reported in Figure 3B, clearly indicate that CD34 knockdown induced a remarkable increase in CFU-MK.
Gene Expression Profile of CD34-Silenced Cells
To investigate changes in gene expression induced by CD34 gene silencing, we performed microarray analysis of mRNA expression profile in control (MOCK and NegCTR) and CD34siRNA cells, using the Affymetrix HGU133A GeneChip array. Microarray analysis was performed on pooled RNA derived from three independent experiments at 2 days after the last nucleofection.
Analysis of microarray data showed that the transcripts increased in CD34siRNA cells are genes involved mainly in granulocyte and megakaryocyte differentiation. As shown in Figure 4, known regulators of granulocyte commitment, such as CCAAT enhancer binding protein (C/EBP)α  and C/EBPδ , and markers of granulocyte differentiation, such as MPO, elastase 2 (ELA2), and RNASE2 , were upregulated in CD34siRNA cells. Similarly, CD34siRNA cells showed an increased expression of transcription factors involved in megakaryocyte commitment, such as FLI1  and MKL1 , and an increased expression of genes coding for platelet surface proteins, such as CD9 , GP1BB , ITGA5 , ITGA2B (CD41), and ITGB3 (CD61) .
Among downregulated genes, we found erythroid transcription factors (KLF1, TAL1, and TCF3) , globin chains (HBA1 and HBA2) , erythroid cytoskeleton proteins (EPB4.2 and ERAF) [30, 31], and surface antigens associated with erythroid differentiation (RHAG, GYPA, and GYPC) [32, –34]. As a whole, these microarray data provide strong molecular support for the biological effect promoted by CD34 gene silencing on the differentiation capacity of hematopoietic stem/progenitor cells. To confirm the microarray data, we carried out, in three independent experiments, a TaqMan QRT-PCR analysis on a validation set selected among the differentially expressed genes between CD34siRNA versus MOCK/NegCTR samples; gene selection was also made on the basis of their biological function both as master regulators of lineage commitment (KLF1 and CEBPα) and as markers of differentiated cells (MPO, ELA2, HBA1, HBA2, and RHAG). All genes analyzed by QRT-PCR showed the same expression pattern as that obtained by microarray analysis (supplemental online Fig. 3A).
CD34 Overexpression in CD34+ Stem/Progenitor Cells
To confirm the results obtained by RNA interference, we overexpressed human CD34 cDNA in human CD34+ cells using retroviral gene transfer. To this end, we constructed the LCD34IΔN retroviral vector expressing full-length CD34 cDNA and ΔLNGFR as reporter gene, in the context of a bicistronic transcript driven by the viral LTR .
All the experiments were performed starting from freshly isolated CB CD34+ cells, with cell viability greater than 90%. No statistically significant differences among MOCK, LXIΔN, and LCD34IΔN cell viability after two cycles of infection were found. Trypan blue analysis indicated a slight decrease in percentage of viability in cells purified by immunomagnetic system compared with nonpurified cells (supplemental online Fig. 1B).
In a set of 13 independent experiments, gene transfer efficiency, assessed by flow cytometric analysis of ΔLNGFR positivity, ranged from 15% to 30%. Transduced cells were then purified by nerve growth factor receptor (NGFR) expression at day 3 postinfection. Flow cytometric analysis performed on transduced/NGFR-purified cells revealed that the CD34 transgene was highly expressed at the protein level up to 10 days postinfection (Fig. 5).
The data demonstrate that the maximum of the transgene expression level, together with a CD34 expression of at least 60% in controls, was reached at day 3 postinfection (Fig. 5). Consequently, we chose day 3 postinfection for seeding the cells in liquid or semisolid medium.
Biological Effects of CD34 Overexpression in CD34+ Cells
Biological effects promoted by CD34 overexpression were assessed by morphological and flow cytometric analysis after serum-free liquid culture. Flow cytometric analysis of lineage differentiation markers performed at day 12 postinfection revealed in LCD34IΔN-transduced cells a decreased expression of granulocyte markers CD66b and MPO. In contrast, expression of the erythroid marker GPA was upregulated (Fig. 6A, 6C).
Morphological analysis performed on transduced CD34+ cells at day 14 postinfection showed that the LCD34IΔN sample was mainly enriched in erythroblasts, whereas it was poor in mature granulocytes (Fig. 6Biii), as compared with untransduced (Fig. 6Bi) and LXIΔN control cells (Fig. 6Bii). Differential cell count and statistical analysis are reported in Figure 6D.
To better characterize the role of the CD34 antigen in stem/progenitor cell differentiation, transduced and untransduced CD34+ cells were plated in methylcellulose at day 3 postinfection. The results obtained from clonogenic assay, reported in Figure 6E, clearly indicate that CD34 overexpression induces a remarkable increase in BFU-E erythroid progenitors and a significant decrease in CFU-G progenitors. Cells transduced with the LXIΔN “empty” vector exhibited levels of clonogenic progenitors that were superimposable to those of the untransduced control. Values are reported as mean ± 2 SEM from 13 independent experiments. Absolute numbers of colonies are reported in supplemental online Table 2.
Gene Expression Profile of CD34-Transduced Cells
To investigate the molecular phenotype of CD34-transduced cells, we assessed the transcriptome profile using the Affymetrix HGU133A GeneChip array. This expression profile was compared with that of the untransduced control and LXIΔN empty vector-transduced cells. Microarray analysis was performed on RNA pooled from three independent experiments at day 3 postinfection. Using the filtering procedure described in Materials and Methods, we identified a number of genes significantly modulated in CD34-transduced cells versus control cells. As depicted in Figure 7, CD34-transduced cells showed an increased expression of erythroid marker proteins (ALAS2  and SCL2A1 ), globin genes (HBA1, HBA2, HBB, HBD, HBE1, HBG1, and HBZ) , erythroid transcription factors (TAL1, KLF1, and GATA1) , erythroid cytoskeleton proteins (EPB49 and ANK1) [37, 38], and surface antigens associated with erythroid differentiation (RHAG and GYPA) [32, –34]. On the contrary, we found many markers of granulocyte differentiation, such as granule proteins (AOAH, AZU1, BP1, CTSG, DEFA1, DEFA4, ELA2, EPX, MPO, PRG2 and 3, PRTN3, S100A12, EPX, and RNASE2 and 3) , the granulocyte colony-stimulating factor receptor (CSF3R) , a neutrophil chemokine receptor (CX3CR1) , and transcription factors of granulocyte commitment (CEBPα and CEBPδ) [19, 20], among downregulated genes. The microarray data were validated in three independent experiments by QRT-PCR on the same validation set assessed for CD34 inhibition experiments (MPO, ELA2, CEBPA, KLF1, HBA1, HBA2, and RHAG). As shown in supplemental online Fig. 3B, we found a consistent expression pattern among QRT-PCR and microarray data.
Since its discovery, CD34 has become the most widely used marker for the enrichment of human hematopoietic progenitor cells, but despite its importance, its function remains largely unclear . Because of the expression of CD34 on hematopoietic progenitor cells from the earliest multipotent stem cell to the more differentiated progenitors, it has previously been hypothesized that this sialomucin may have a role in the physiology of these cells [13, 14].
Data from knockout mice suggest that CD34 expression is important for proliferation and/or maintenance of hematopoietic progenitor cells both in the embryo and in the adult . In fact, the analysis of embryonic stem (ES) cells derived from CD34 KO mice showed a significant delay in both erythroid and mono/macrophage differentiation that could be reversed by transfection of the mutant ES cells with a construct expressing CD34. In addition, the colony-forming activity of hematopoietic progenitors is significantly reduced in adult CD34-deficient animals; furthermore, CD34-deficient progenitors also appear to be unable to expand in liquid cultures in response to hematopoietic growth factors . Moreover, several pieces of experimental evidence demonstrate that CD34 has signal-transducing capacity and is involved in cell adhesion, causing actin polymerization and homotypic adhesion in KG1a and CD34+ cells [6, , –9].
On the basis of these premises, we decided to investigate the role of CD34 antigen in the regulation of human hematopoietic stem cell differentiation. The role played by the CD34 antigen was assessed by two different experimental approaches: gene silencing by means of siRNAs and overexpression by retrovirally mediated gene transfer.
The results obtained by RNA interference clearly showed that early downregulation of CD34 caused a remarkable increase in the percentage of granulocyte colonies coupled with a decrease in the percentage of the erythroid ones. Moreover, the megakaryocyte differentiation assay showed a significant increase in the number of megakaryocyte colonies after CD34 gene silencing. In agreement with clonogenic assay results, CD34-silenced cells were highly enriched in megakaryocyte and granulocyte precursors at different levels of maturation, whereas erythroblasts were poorly represented as compared with control samples. In addition, increased expression of granulocyte markers CD66b and MPO and of the megakaryocyte marker CD41, as well as a decrease in the erythroid marker GPA, was observed in CD34-silenced cells.
These results strongly suggest that CD34 expression may have a role in regulating hematopoietic progenitor cell differentiation. As CD34 gene silencing induces an increase in granulocyte and megakaryocyte precursors coupled to a decrease in the erythroid precursors, we can conclude that CD34 expression seems to support the erythroid lineage at the expense of the granulocyte and megakaryocyte lineages.
To determine the effects of CD34 gene silencing on the expression profile of hematopoietic progenitor cells, we performed a set of gene expression profiling experiments on CD34-silenced cells. CD34 gene silencing induces the expression of both master regulators (C/EBPα and C/EBPδ) [19, 20] and markers of granulocyte differentiation (MPO, ELA2, and RNASE2) , in addition to genes involved in megakaryocyte commitment (FLI1 and MKL1) [22, 23] and genes coding for platelet surface proteins (CD9 and CD61 [24, 27]).
Moreover, CD34 gene silencing causes the downregulation of several genes associated with erythroid differentiation, such as transcription factors (KLF1 and TAL1) , globin genes , cytoskeleton proteins (EPB4.2 and ERAF) [30, 31], and surface antigens (RHAG, GYPA, and GYPC) [32, 33, 34]. Together, these data demonstrate that CD34 gene silencing induces the expression of a large number of genes, leading to granulocyte and megakaryocyte differentiation, as well as causing downregulation of genes involved in erythroid differentiation.
To confirm the results obtained by RNA interference, we overexpressed the human CD34 cDNA in human CD34+ hematopoietic stem/progenitor cells by retroviral gene transfer. Consistent with results obtained by RNA interference, CD34 overexpression induces erythroid differentiation at the expense of granulocyte differentiation, as assessed by clonogenic assay and flow cytometric analysis.
Taken together, our data provide a clear demonstration that CD34 is able to affect the differentiation capacity of hematopoietic stem/progenitor cells. In particular, we showed that CD34 promotes differentiation of hematopoietic progenitors toward the erythroid differentiation lineage, which is achieved, at least in part, at the expense of the granulocyte and megakaryocyte lineages, in accordance with the data on ES cells from CD34 KO mice . To our knowledge, this is the first demonstration of a functional role for CD34 in the hematopoietic stem/progenitor cell differentiation.
Our data clearly demonstrate that the expression of CD34 antigen is able to exert a strong influence on the differentiation program of hematopoietic progenitor cells; however, the molecular basis underlying the role played by CD34 in HSC differentiation remains to be elucidated. The lack of information about the natural ligand for CD34 and the signal transduction pathway activated by CD34 in human HSCs makes it difficult to understand how this molecule works to affect stem/progenitor cell differentiation.
Stimulation of CD34 has been shown to activate Fyn and Lyn tyrosine kinases, resulting in actin polymerization . In addition, several reports have demonstrated a role for these tyrosine kinases in cell adhesion and in hematopoietic progenitor cell differentiation. Karur et al. recently demonstrated that Lyn kinase promotes erythroid progenitor cell expansion and supports the subsequent late-stage erythroid development . At the same time, a study have demonstrated a selective activation of Fyn and Lyn kinases following TPO stimulation during megakaryocyte differentiation . In this way, CD34 could play a role in modulating HSC differentiation by modifying the activation status of such cellular kinases and influencing the adhesion properties of hematopoietic progenitor cells. It has recently been hypothesized by Bullock et al.  that the capacity of CD34 to induce homotypic aggregation may be the mechanism by which CD34 negatively regulates cell proliferation and promotes differentiation. On the whole, CD34 antigen may affect the interaction between HSCs and the stromal cells in the hematopoietic endothelial niche, which regulates HSCs maintenance and differentiation [44, –46].
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
This work was funded by the Italian Ministry of University & Research (Progetti di Ricerca di Interesse Nazionale) 2005 (Contract 2005050779) and 2006 (Contract 2006057308) and by Associazione Italiana per la Ricerca sul Cancro (AIRC) 2005 and 2006. E.B. is the recipient of a fellowship from the Regional Program for Industrial Research, Innovation and Technological Transfer-ERGENTECH. S.S. and R.Z. contributed equally to this study.