Erythroid Progenitor Cells Expanded from Peripheral Blood without Mobilization or Preselection: Molecular Characteristics and Functional Competence

Background Continued development of in-vitro procedures for expansion and differentiation of erythroid progenitor cells (EPC) is essential not only in hematology and stem cell research but also virology, in light of the strict erythrotropism of the clinically important human parvovirus B19. Methodology/Principal Findings We cultured EPC directly from ordinary blood samples, without ex vivo stem cell mobilization or CD34+ cell in vitro preselection. Profound increase in the absolute cell number and clustering activity were observed during culture. The cells obtained expressed the EPC marker combination CD36, CD71 and glycophorin, but none of the lymphocyte, monocyte or NK markers. The functionality of the generated EPC was examined by an in vitro infection assay with human parvovirus B19, tropic for BFU-E and CFU-E cells. Following infection (i) viral DNA replication and mRNA production were confirmed by quantitative PCR, and (ii) structural and nonstructural proteins were expressed in >50% of the cells. As the overall cell number increased 100–200 fold, and the proportion of competent EPC (CD34+ to CD36+) rose from <0.5% to >50%, the in vitro culture procedure generated the EPC at an efficiency of >10 000-fold. Comparative culturing of unselected PBMC and ex vivo-preselected CD34+ cells produced qualitatively and quantitatively similar yields of EPC. Conclusions/Significance This approach yielding EPC directly from unmanipulated peripheral blood is gratifyingly robust and will facilitate the study of myeloid infectious agents such as the B19 virus, as well as the examination of erythropoiesis and its cellular and molecular mechanisms.


Introduction
The basic mechanisms of stem cell proliferation and differentiation leading to erythropoiesis are well established. In vitro studies on this topic have been carried out with progenitor cells obtained not only from bone marrow, but also from foetal liver and peripheral blood [1][2][3][4][5][6]. The erythropoietic growth factors affect the progenitors in all these locations [3], and many procedures have been undertaken to reproduce the erythroid maturation in vitro including initial selection of the CD34+ cells [7][8][9][10][11], adherence depletion [1,3,12,13] and phased culturing [6,12,14]. In vitro culture of selected CD34+ cells following G-CSF mobilization of peripheral blood stem cells (PBSC) was recently shown to yield a homogenous population of erythroid progenitor cells fulfilling the strict host cell specificity and growth requirements of the erythrotropic parvovirus B19 [15][16][17]. The resulting CD36+ cells were generated with a defined combination of growth factors [7].
Parvovirus B19 comprising three major genotypes [18] belongs to the Parvoviridae family, genus Erythrovirus [17] and replicates selectively in erythroid progenitor cells at BFU-E and CFU-E stages [13,19]. For this restriction, both in vitro investigations and clinical studies of this virus have been greatly hampered by the unavailability of fully permissive cell cultures. The ex vivo-derived CD36+ cell cultures [16] could thereby be of significant academic and practical utility via propagation of this clinically important virus, which hitherto has been done at best in primary cultures and semipermissive cell lines [11,13,[20][21][22][23][24][25][26][27][28][29]. By using the same distinct growth factor composition [7,16], we show that the methodology for obtaining erythroid stem cells can be markedly simplified, as performed without mobilization or CD34+ cell preselection, directly from ordinary samples of peripheral blood mononuclear cells (PBMC).

Results
Cell Culture 56105 freshly separated PBMC from the blood samples of staff members or the buffy coats of healthy donors were cultured in the presence of growth factors favouring the differentiation of pluripotent stem cells toward the erythroid lineage [7,16]. In parallel, an equal amount of cells was cultured without the growth factors. On days 4 or 5 the first signs of cell clustering were seen in the cultures exposed to the growth factors (Fig. 1A). At this stage, the cells were split 1:5 and cultured for 5 additional days. From the initial 56105 cells their number increased to (5-10)6107 on day 10 (Fig. 1C); the overall increment was (1-2)6102-fold. In phenotype, the cells with the growth factors showed a progressive increase both in size and number, and particularly in extent of clustering (Fig. 1A) reaching a maximum by day 10 (Fig. 1C). In the absence of growth factors, apoptotic bodies and cell debris abounded, until on day 10 no living cells were seen (Figs. 1B-C).

Flow Cytometry
PBMC were analysed by flow cytometry immediately following separation (day 0) and after 10 days of culture with the growth factors. The cells were analysed for the main lymphocytic, monocytic and erythroid marker expression patterns. During this time a profound increase in erythroid cells was observed, with a decrease in lymphocytes and monocytes ( Figs. 2A-B). On day 0, only ,0,25% of the cells were CD34+. Specifically, the day 0 cell population comprised primarily lymphocytes and monocytes, .50% CD3+ (CD4+ or CD8+), ,10% CD56+, and ,20% CD14+ (Fig. 2B). All these values approached 0 on day 10. On the other hand, the erythroid progenitor cells that were barely detectable on day 0 increased to ,95-100% by day 10, as indicated by CD71 and glycophorin expression patterns ( Figs. 2A-B). Importantly, on day 10, CD36 positivity reached with some donors ,100% of the cells. On day 0 this marker showed variable expression, in coexpression with the monocyte marker CD14 [30]. Of note, some of our donors presented with mononuclear CD36+ CD142 cells, as recently identified [31]. As with CD36, on day 0, also the proportion of globoside Pexpressing cells showed variability among the donors, with a highest value of .20% (Fig. 2B), while on day 10 of culture nearly every cell co-expressed this B19 virus receptor along with CD36 (Fig. 2D).

Comparison of Erythroid Progenitor Cells Generated from Unselected PBMC and Selected CD34+ Cells
In order to determine which cell type(s) represented the growth factor target(s), we followed in parallel cultures from unselected PBMC and from pure populations of preselected CD34+ cells. The cultures were analyzed by flow cytometry on days 0 and 10. Differentiation of the CD34+ cells into erythroid progenitors fully comparable with those from the unselected PBMC could be verified. Indeed, as opposed to the very low fraction of CD34+ cells in the PBMC on day 0 ( Figs. 2A-B), the percentages of glycophorin+ and CD36+ erythroid progenitors were $70% and $80%, respectively, from the cultures of unselected PBMC and selected CD34+ cells (Fig. 2C).

B19 Virus Permissivity of the Generated Erythroid Progenitor Cells
The erythroid progenitor cells expanded from peripheral blood were analysed for their functional competence by a virological approach. As the cells harboured both globoside P antigen and CD36 (Fig. 2D), they were considered suitable for a B19 parvovirus in vitro infection assay.
In vitro infection. Both of the procedures performed [16,32] turned out comparable in all the downstream analyses. Furthermore, we observed no difference in any of the B19 infection parameters between the cells obtained from B19 seropositive and seronegative donors.
Nucleic acid analyses. DNA and RNA were extracted from the infected and uninfected cells at 2, 24 and 48 hrs, and real-time PCR and RT-PCR were performed. The contiguous primers annealing to the common exon of the B19 genome were used for both DNA and RNA detection, the latter after DNase treatment. DNA was quantified by interpolation on a standard curve obtained with serial dilutions of plasmid DNA containing the coding region of the B19 genome. An overall increment of 3 logs of the DNA copy numbers was observed at 24-48 hrs post infection (Fig. 3A). Our assessment of the total B19 mRNA signal ( Fig. 3C) took into account both the amount of DNA amplified by PCR (Fig. 3B) in absolute numbers and the extent of background DNA signal obtained by RT-PCR in the absence of reverse transcriptase. In RNA detection, the spliced VP transcripts, corresponding to the bands of 148 and 268 bp, were seen in agarose gel electrophoresis (Fig. 3D) following amplification with the non-contiguous primers [33].
Protein expression. The erythroid progenitor cells were analyzed for both structural (VP2) and nonstructural (NS1) proteins of the B19 virus, and in both native and denaturing conditions (Fig. 4). Immunofluorescence staining was performed on the infected and uninfected cells fixed at 2 and 48 hrs. At 48 hrs post-infection .50% of the cells were positive for VP2 and ,50% for NS1, by contrast to 0% at 2 hrs post infection (Fig. 4A). Correspondingly, in Western blotting a strong VP2 band (58 kDa) was obtained from the cells lysed at 48 hrs postinfection, as opposed to none from the negative control cells (Fig. 4B).

Extent of Increase of Erythroid Progenitor Cells
Considering the proportion of hemopoietic CD34+ stem cells initially present in the unselected PBMC ( Figs. 2A-B) and the expansion of overall cell number during culture (Fig. 1C), we first calculated the relative fold increase of the cells expressing the erythroid progenitor markers (CD34 vs. CD36 and CD71 along with glycophorin) (Fig. 5: grey bars; Table 1: rows 1, 2, 3). Moreover, the immunofluorescence results, in which .50% of the cells infected on day 10 expressed the B19 antigens, were taken into account together with the aforementioned variables, to calculate the fold increase of B19 permissive cells. As shown in Figure 5 (white bar) and Table 1 (row 4), (i) the overall cell count increased 100-200 fold; and concomitantly with this (ii) the proportion of EPC multiplied from ,0.5% to .50%. I.e. it can be concluded that, during the in vitro culture our B19 permissive erythroid cells expanded .10 000-fold.

Discussion
For in vitro erythropoiesis many different approaches have been taken, utilizing pluripotent stem cells from a variety of sources   [1][2][3][4]6,9,34]. We used unfractionated PBMC obtained from ,1 ml of peripheral blood of ordinary blood donors in the absence of bone marrow stem cell mobilization or of ex vivo priming of any other type, and selectively expanded the contained stem cells by in vitro culture in the presence of a defined combination of growth factors favouring erythropoietic commitment and expansion [7]. Indeed, our approach did not include any cell preselection, neither positive, for CD34+ [8][9][10][11], nor negative, against CD3-CD14- [13], nor adherence depletion [1,3], and not even phased culturing [6,12,14] except for a single passage. Our straightforward approach yielded the desired population of erythroid progenitor cells comparable in purity and yield to the previously introduced approach including mobilization and preselection. Our comparison between selected CD34+ stem cells and unselected PBMC verified qualitatively and quantitatively similar yields of EPC. By this robust approach the initially diverse PBMC evolved in vitro, with natural selection toward a notably uniform population of colony-forming cells phenotypically characteristic of erythroid progenitors [6,8,12,14,35]. Of the cells cultured for 10 days nearly 100% expressed the marker combination of erythroid progenitors, CD36, CD71 and glycophorin, and were devoid of the lymphocyte, monocyte or NK markers CD4, CD8, CD3, CD14, CD56.
The aforementioned markers define best the immunophenotype of erythroid progenitor cells, as shown in a large number of studies  [6,8,12,14,[35][36][37]. In those studies glycophorin and the transferrin receptor CD71 have been identified as the reference molecules of erythroblasts. Although a well-established marker for erythroid progenitors [2,16], CD36 on its own is far from cell type specific, as it occurs within the PBMC in monocytes, macrophages and even platelets, and participates in a variety of functions [30,38]. However, we consider it appropriate to include CD36 as an indicator of erythropoietic commitment in concert with glycophorin and CD71 [10,16], which were almost absent at the beginning yet occurred in nearly every cell at the end of our expansive culturing.
The expanded EPC were verified to be functionally competent by permissivity to B19 parvovirus infection. The B19 in vitro infection procedures employed here have been previously used in the documentation of functionality of mobilized, preselected (CD34+) cells [16,32]. Our EPC obtained bypassing those steps were similarly susceptible to B19 virus infection, in terms of DNA replication (increase of 3 logs), mRNA transcription and protein production (both NS and capsid). Neither the presence of antiviral antibodies nor virus-specific T cells [39] in the donor blood affected the virological outcome.
Previous studies have addressed the role of accessory cells in establishing ideal conditions for the erythropoietic proliferation and commitment [1,3,6,36,40,41]. In a comparative study with different cell lineages [36], cellular diversity has turned out to be essential compared with selected pluripotent cells. Our main finding, obtained with the novel procedure, is in line with this. The identification of the factors accounting for the differentiation within this condensed in vitro population is an interesting subject of further investigation.
Our procedure for EPC generation is gratifyingly robust. This straightforward technique can be foreseen to facilitate B19 virus basic research both in its molecular and cellular aspects, as well as to simplify the measurement of B19 virus neutralizing immunity [39,[42][43][44][45][46]. The latter studies have been hampered greatly by the absence of facile cell culture methodology for this clinically important virus [17,47].

Ethics Statement
All the samples used during this study were obtained following written informed consent from the donors. This study was approved by the Ethical committee of the Helsinki University Central Hospital.

Blood Samples and Cells
Peripheral blood samples were obtained into Vacutainer tubes from nonsymptomatic staff members. Leukocyte-enriched buffy coats of healthy blood donors were obtained from the Finnish Red Cross Blood Transfusion Service, Helsinki, Finland. From these preparations peripheral blood mononuclear cells (PBMC) were isolated by Ficoll-Paque centrifugation and two PBS washes [39,42].

Cell Culture
From the PBMC, erythroid progenitor cells (EPC) were generated via culture in an expansion medium containing erythropoietic growth factors [7,16]. Briefly, 56105 PBMC were cultured in MEM (HyClone, Logan, UT) supplemented with the serum substitute BIT 9500 (StemCell Technology, Vancouver, BC, Canada), diluted 1:5 for a final concentration of 10 mg/ml bovine serum albumin, 10 mg/ml rhu insulin, and 200 mg/ml iron-saturated human transferrin, enriched with 900 ng/ml ferrous sulfate (Sigma, St. Louis, MO, USA), 90 ng/ml ferric nitrate (Sigma), 1 mM hydrocortisone (Sigma), 3 IU/ml rhu erythropoietin (Janssen-Cilag, Espoo, Finland), 5 ng/ml rhu IL-3 (R&D Systems, Minneapolis, MN, USA), and 100 ng/ml rhu stem cell factor (Peprotech, London, UK). The cells were maintained at  Figure 2 and the cell counts of Figure 1C and Table 1. White bar: foldincrease of B19 virus permissive EPC; percentage of cells expressing capsid proteins at 48 hrs post-infection (Fig. 4A) Rows 1-3: Fold increase of cells expressing erythroid markers CD36, CD71 and glycophorin on day 10 of culture; the values are based on CD34 + fraction (%) on day 0 (0.22%) and on the expansion factor of 150, expressed as an average fold increase throughout the culture (Fig. 1C). The first column indicates the values obtained during the flow cytometry analysis of the three markers (Fig. 2B). Row 4: Fold increase of B19 virus permissive cells (row 4); the values are based on CD34 + fraction (%) on day 0 (0.22%) and on the expansion factor of 150, expressed as an average fold increase throughout the culture (Fig. 1C). The first column indicates the percentage of cells expressing capsid proteins (Fig. 4A) For comparison, from some of the blood samples, CD34+ cells were first isolated by CD34 Microbead Kits (Miltenyi Biotech, Bergish Gladbach, Germany) containing magnetic beads conjugated to CD34 antibody, and were then cultured as described.

Parvovirus B19 Permissivity Study
Serology. All the PBMC donors' sera were studied for the presence of B19 virus IgG and IgM antibodies as described [43].
In vitro infection. Cell cultures of day 10 were infected at 103 genome equivalents per cell with B19 virus from a high-titer viremic, B19-seronegative donor plasma [48]. Two different in vitro infection procedures were used in parallel [16,32]. In one, 26104 cells in 10 ml volumes on a 96-well plate were treated with 10 ml of diluted virus for 2 hrs at 4uC, washed and maintained at a final concentration of 26104/100 ml, at 37uC. In the other, 26106 cells in 500 ml volumes on a 24-well plate were treated with the virus at 4uC, under agitation, washed and maintained at a final concentration of 26105/ml, at 37uC. For control, mock-infected cultures, without addition of virus, were maintained in parallel with the infected ones. Aliquots of the test and control cultures were harvested at 2, 24 and 48 hrs post infection for DNA, RNA and protein analyses.
Nucleic acid analyses. DNA and RNA were extracted from the infected and uninfected cells at various times post infection, as described [33,49]. Briefly, 105 cells were lysed with Proteinase K (100 mg/ml; Fermentas Finland, Helsinki, Finland) and 1% SDS in 100 mM NaCl, 10 mM EDTA, 10 mM Tris-Cl pH 7.5 for 1 h at 55uC, and the nucleic acids were extracted by phenolchloroform followed by ethanol/Na-acetate precipitation. RNA was selectively extracted with ToTally RNA purification kit (Ambion, Austin, TX, USA) and treated with DNase TurboDNAfree (Ambion). A constant amount of both extracts, corresponding to ,104 cells per reaction, was amplified, in parallel for viral DNA and mRNA, by the Light Cycler (Roche Applied Sciences, Basel, Switzerland) by using Quantitect SYBRgreen PCR and Quantitect SYBRgreen RT-PCR Kits (Qiagen, Hilden, Germany), respectively. The contiguous primers, corresponding to the common central exon of the viral genome and its thermal profile have been described [33,49]. The absolute quantity of viral DNA was determined by interpolation using a standard curve constructed with serial dilutions (108-100 geq) of a plasmid containing the viral coding region (GenBank AY504945). Standards were included in each assay and linearity of the curve was confirmed. The RT-PCR was performed in the presence (RT+) and absence (RT-) of reverse transcriptase with the primer pair used for DNA amplification. The actual signal due to mRNA was obtained taking into account the Ct values [RT+] and [RT2], and the Ct corresponding to the DNA template. The relative increase at the various time points was then calculated. For additional evidence of RNA specificity of detection, noncontiguous primers (HR6) for spliced transcripts [33] were used. Specificity of the real-time PCR and RT-PCR results was confirmed by melting curve analysis and agarose gel electrophoresis of the amplicons for the contiguous and the non contiguous primer pairs, respectively. For extraction control, PCR for beta-actin DNA was performed in parallel [50].
Western Blotting of lysates from ,36104 infected or uninfected cells was performed as described [42] with reducing 7.5%DS-PAGE gels, by using a monoclonal B19 virus capsid antibody (NovoCastra Laboratories, Wetzlar, Germany) followed by antimouse IgG-HRP (DakoCytomation). The corresponding isotype control antibody (R&D Systems) was applied for specificity control.