During the past decade, human full-term cord blood (HTCB) has been considered to be an attractive hematopoietic stem cells (HSC) source not only for transplantation purposes, but also for basic research investigators studying useful marker to identify immature stages of HSC differentiation (1–3). Compared with mobilized peripheral blood stem cells and marrow HSCs, CB-HSCs are essentially characterized by a series of advantages such as less-stringent HLA-matching requirements between donors and recipients, absence of ethical concerns as well as risks for the mother and the infant, low Epstein Barr virus and cytomegalovirus contaminations, immediate availability of frozen units, and maintenance of ethnic balance by targeted collection programs (1).
Although hTCB transplantation will go on growing thanks to the advances in HSC basic biology knowledge and new approaches for ex vivo expansion of hematopoietic progenitor cell populations, little is known about the possible use of human early preterm cord blood (hEPCB) for hematopoietic purposes (4, 5).
In the last 10 years, just a few reports showed the characterization of CD34Pos HSCs residing in early preterm (17–24 weeks) CB (5, 6). Cervera et al. (7) offered a four-color flow cytometry (FCM) analysis of HSC and lymphocyte compartments present in preterm samples with a gestational age range between 26 and 36 weeks. Therefore, to better define the immunophenotypic distinctiveness of human early preterm cells, we studied 19 hEPCB and 17 hTCB samples using a multidimensional FCM approach, according to Tung et al.'s (8) recommendations.
In this perspective, the purpose of our study was to (i) determine the frequency of CD34Pos CD45Dim cells; (ii) correlate the frequency of CD34PosCD45Dim with the entire gestational époque; (iii) analyze the CD34PosCD45Dim population for a series of activation structures, adhesion molecules, cytokine-receptors, and efflux-related proteins, such as CD38 (ADP-ribosyl-cyclase), CD29 (β1-integrin), CD31 (PECAM-1), CD71 (Transferrin receptor), CD90 (Thy-1), CD133 (Prominin-1), CD117 (c-kit), CD135 (Flt3), CD200 (OX2), CD243 (Multidrug Resistance protein-1, MDR1) and CD338 (ABCG-2); (iv) define the antigenic pattern of undifferentiated (CD38Neg) and committed (CD38Pos) HSCs by an ad hoc polychromatic multidimensional analysis; and (v) compare the distribution of monocytes, natural killer cells, and T- and B-lymphocytes between the two kinds of CB samples.
MATERIALS AND METHODS
Collection of Human CB Samples
According to Shields and Andrews (9), samples were defined as hTCB if they were from fetuses between 37 and 41 weeks of gestation and as hEPCB if they were obtained from fetuses with less than 28 weeks of gestation.
In this study, 17 hTCB samples were obtained during normal delivery, whereas 19 hEPCB specimens were drawn in the course of prenatal karyotypic diagnosis in fetuses with suspected critical alterations. Karyotype results were available in 12 of 19 fetuses and revealed only two cases with trisomy 18, along with 10 normal karyotypes. Written informed consent was obtained from all 36 mothers before the study.
About 5 ml of hTCB was drawn via sterile venipuncture into vacutainers containing sodium heparin (Becton Dickinson [BD], San Jose, CA; catalog no. 367876] from umbilical veins of 17 normal newborns. About 4 ml of hEPCB was drawn via sterile venipuncture into vacutainers containing sodium heparin (BD catalog no. 367876) by an in utero ultrasound-guided procedure from 19 early preterm fetuses. Within 24 h from the collection, cord blood (CB) mononuclear cells (MNCs) were separated by Ficoll density gradient centrifugation and immediately cryopreserved with 10% of dimethylsulfoxide in liquid nitrogen. After a rapid thaw at 37°C, cells were washed once with total fetal bovine serum (FBS) and resuspended in phosphate-buffered saline with 10% FBS just before the application of immunophenotyping procedures. Cellular viability was assessed each time after thawing by both Trypan blue dye exclusion and analysis of light scatter properties in FCM, and it was never less than 90%.
Monoclonal Antibodies Staining
We studied CB cells by polychromatic FCM and by analyzing the expression of a series of hematopoietic antigens. The tubes contained the following antibody cocktails:
In particular, BD provided all monoclonal antibodies (MoAbs) used except for anti-CD338-APC and anti-CD133-PE MoAbs obtained from R&D (Minneapolis, MN) and Miltenyi Biotec (Bergisch Gladbach, Germany), respectively.
For all antibody staining experiments, at least 2 × 105 cells of each CB sample were incubated at 4°C for 30 min with the appropriate amount of MoAbs, following the manufacturer's instructions. Afterward, the mixtures were diluted 1:20 in ammonium chloride (NH4Cl) lysing solution, then incubated at room temperature for 5 min, and finally washed with staining media, composed of Dulbecco phosphate-buffered saline supplemented with 10% FBS, and analyzed using an unmodified BD FACSCantoII flow cytometer (BD) that was set up according to published guidelines (10). Cells stained with irrelevant monoclonal immunoglobulin reagents conjugated with the same fluorochromes as in the six-colors combinations were used as negative controls and analyzed as fluorescence-minus-one controls.
CaliBRITE beads (BD catalog no. 340486) were used as quality controls across the study as described (11, 12). Appropriate amount of beads was used following the manufacturer's instructions. Daily control of CaliBRITE intensity showed no change in instrument sensitivity throughout the study. The relative voltage range for each detector was assessed una tantum by the “eight-peak” technology (Rainbow Calibration Particles, BD catalog number 559123) at the beginning of the study.
Compensation was set in FACS-DiVa (BD) software, and samples were analyzed compensated. Later on, samples were acquired within 30 min on the FACSCantoII, and a minimum of 100,000 events was recorded for each monoclonal combination. Levels of CD antigen expression were displayed as median fluorescence intensity (MFI). The software used for cytometric analysis was the FACS-DiVa (BD), and, throughout this study, we adopted the “logicle” (or “bi-exponential”) display method as suggested by Herzenberg et al. (13). In particular, all data of this study were plotted on “logicle” axes, thus incorporating the useful features of logarithmic display and providing accurate visualization of populations with fluorescence values at or below zero (13).
A CD45 vs. CD71 dot-plot was used as a tool for distinguishing erythroid and nonerythroid cells (14). A P1 gate was imposed on CD45Neg cells, and cells of interest were derived by a reverse gate (“not-P1”). The presence of CD71 allowed us to accurately impose the gate on CD45Neg cells, including, along with debris, all erythroblasts (Fig. 1).
Descriptive statistics included medians, lowest and highest values, and 95% confidence interval for the medians. Statistical significance was assessed by Mann–Whitney nonparametric test for unpaired data, and Wilcoxon test was used for paired data.
Identification of HSC Population
A multidimensional gating strategy was adopted to select HSCs in terms of CD34PosCD45Dim events (Fig. 1). First, we imposed a P1 gate on CD45Neg cells in a CD45 vs. CD71 dot-plot to exclude erythroblasts and cell debris remaining after the erythrocyte lysis step (Fig. 1A). Subsequent analysis was done on not-P1 events (“reverse gate” strategy). Because the gate was accurately designed, CD45Dim cells entirely fell within not-P1 fraction. The combination CD45, CD71, and CD34 represents a well-established tool for analysis of erythroid and nonerythroid cells (14). CD34Pos cells were then selected in a CD34 vs. SSC dot-plot (Fig. 1B). Furthermore, according to Brocklebank and Sparrow (15), “true” HSCs were identified as CD34PosCD45Dim events in a CD45 vs. CD34 dot-plot (Fig. 1C) and also evidenced as live cells on the basis of light scatter properties (Fig. 1D). Starting from the above dot-plots, we determined the percentage of CD34PosCD45Dim cells in not-erythroid MNCs (not-P1 gate) from each sample.
hEPCB Samples Show a Higher Percentage of HSCs Compared With hTBC
For each sample, 100,000 events were analyzed, and, according to previous studies (16–18), we observed a higher median percentage of CD34PosCD45Dim cells within not-erythroid MNC compartment in hEPCB as against hTCB samples (Figs. 2A and 2B), being the median value of 1.7% (25th–75th percentiles = 1.1–3.3; n = 19) and 0.3% (25th–75th percentiles = 0.2–0.4; n = 17) (P < 0.0001), respectively (Fig. 2C).
CD34PosCD45Dim HSC Percent Correlates With the Gestational Week
A regression analysis was used to determine the relationship between the gestational age and the frequency of CD34PosCD45Dim cells inside not-erythroid MNC population. In particular, we subdivided early preterm samples into two gestational age blocks (16th–20th and 21st–27th week of gestation), and, for the first time, we evidenced an inverse relationship between HSC% and gestational age during the period between the 16th and 20th week, with a R2 value of 0.91 and a P value <0.001 (Fig. 3). Interestingly, we did not find any correlation during the 21st–27th week period (R2 = 0.02, Fig. 3).
Intensity Analysis of Surface Antigens in CD34PosCD45Dim HSC Population
CD34PosCD45Dim HSCs were screened for CD29, CD31, CD38, CD71, CD90, CD117, CD133, CD135, CD200, CD243, and CD338 expression in terms of median value of MFI (Table 1). Thanks to this panel we revealed significant differences between hEPCB and hTCB samples as to CD38 and CD135 molecules. In particular, CD38 was mainly expressed by hTCB compared with hEPCB cells, the MFI being 1816 (25th–75th percentiles = 1598–2826, n = 17) and 855 (25th–75th percentiles = 482–1301, n = 19) (P < 0.0001), respectively. As to CD135, it was preferentially expressed, with low fluorescence intensity, by hTCB progenitor cells (P = 0.0006). The other antigens tested did not exhibit significant differences between the two kinds of samples.
Table 1. Immunophenotype comparison between hEPCB and hTCB HSCs
25TH, 75TH PERCENTILES
25TH, 75TH PERCENTILES
Analysis of surface antigens on CD34PosCD45Dim hematopoietic stem cells detected in hEPCB and hTCB samples. Values are expressed as median value of median fluorescence intensity. Negative values were due to bi-exponential analysis.
hEPCB and hTCB Samples Are Enriched in Distinctive HSCs Subpopulations
The screening of CD34PosCD45Dim HSCs compartment was performed for the above-mentioned molecules also in terms of median percentage of positive cells, and significant differences were obtained only for CD38 and CD243 when comparing hEPCB and hTCB. In particular, we found hEPCB considerably richer in immature HSCs compared with the full-term counterparts as evidenced by the correspondent dot-plot analysis (Figs. 4A and 4B). Specifically, within CD34PosCD45Dim cells, the median percentage of CD38Neg events was 16.4% (25th–75th percentiles = 4.6–24.5, n = 19) and 2.6% (25th–75th percentiles = 1.3–3.4, n = 17) in hEPCB and hTCB, respectively (P < 0.0001, Fig. 4C).
CD243 was displayed by a rare subset of HSCs (Figs. 4D and 4E) prevalently belonging to full-term samples. We were able to find median values of 4.7% (25th–75th percentiles = 2.3–11.7, n = 17) and 0.95% (25th–75th percentiles = 0.5–3.4, n = 19) (P < 0.01) of CD243Pos HSCs in hTCB and hEPCB cells, respectively (Fig. 4F).
Intensity Analysis of Surface Antigens on Undifferentiated (CD34PosCD45DimCD38Neg) and Committed (CD34PosCD45DimCD38Pos) HSCs in hEPCB and hTCB Samples
To delve into the antigenic mosaic of undifferentiated HSCs, we subdivided hEPCB and hTCB CD34PosCD45Dim fractions in undifferentiated (CD34PosCD45DimCD38Neg) and committed (CD34PosCD45DimCD38Pos) subsets. The antigens for which we found significant differences were CD29, CD31, CD71, CD117, CD133, and CD135 (Table 2). Table 2 reports data analysis expressed as median value of MFI and 25th and 75th percentiles.
Table 2. Immunophenotype of immature and committed HSCs
CD34PosCD45DimCD38Neg, MEDIAN MFI (25TH, 75TH PERCENTILES )
CD34PosCD45DimCD38Pos, MEDIAN MFI (25TH, 75TH PERCENTILES)
Data concerning the immunophenotype matching between undifferentiated and committed HSCs residing in the same type of sample are presented. Negative values are justified by the application of bi-exponential analysis.
1,395 (1,012, 1,885)
3,223 (2,569, 4,499)
1,242 (914.5, 2,005)
1966 (1,545, 3,444)
4,696 (3,373, 6,429)
5,755 (3,559, 9,330)
5,841 (4,615, 7,772)
5,034 (4,122, 6,363)
49 (6, 633)
636 (224, 1,263)
85 (−47, 232)
440 (167, 1,297)
689 (451, 985)
847 (646, 1,927)
902 (101, 1,755)
2,063 (1,107, 2,644)
1,326 (1,074, 1,900)
287.5 (23.25, 727.3)
1,330 (990, 1,563)
420 (115, 933)
−40 (−99.50, −23.5)
65.00 (24.00, 80.00)
94.50 (−227.5, 221.5)
174 (99.25, 234.8)
Significant differences were evidenced on the basis of Wilcoxon matched pairs test when comparing immunophenotypes between undifferentiated and committed HSCs from the same type of samples.
CD34PosCD45DimCD38Pos HSCs group in both kinds of CB samples expressed higher amounts of CD29, CD71, and CD135, compared with CD34PosCD45DimCD38Neg cells. CD31 was predominantly displayed in CD38Pos cells but only in hEPCB samples. By contrast, CD117 was predominantly displayed in CD38Pos cells of the hTCB group of specimens. Finally, CD133 was preferentially expressed in CD38Neg compared with CD38Pos population, but, only in early samples, the difference was significant.
Multiparametric Analysis of hEPCB and hTCB HSCs: Subpopulation Analysis
Based on the combined analysis of CD71, CD34, CD45, and other three specificities (see Materials and Methods section), we identified some interesting subpopulation patterns. In particular, the CD34PosCD45Dim HSCs analysis via CD133 vs. CD38 dot-plot showed three different subpopulations (Fig. 5A and 5B), i.e., CD133PosCD38Neg, CD133PosCD38Pos, and CD133NegCD38Pos. We evidenced a significant enrichment of CD133PosCD38Neg subpopulation selected inside CD34Pos CD45Dim compartment among early preterm samples (Fig.5C), with the median percentage being 17.4% (25th–75th percentiles = 2.9–23.7, n = 6) and 0.7% (25th–75th percentiles = 0.0–2.6, n = 7) (P = 0.008) in hEPCB and hTCB cases, respectively.
Analysis of MNCs Populations: Lymphocytes and Monocytes
The percentage of T- and B-cells was determined in the two groups of CB samples as CD45PosCD3Pos and CD45PosCD19Pos events, respectively. We found T-cells principally displayed in hTCB specimens (P < 0.05), whereas, as reflection of this finding, B-cells were more evidenced in hEPCB group (P = 0.0028) (Table 3). Moreover, natural killer cells and monocytes, selected, respectively, as CD45PosCD7 PosCD56Pos and CD45PosCD14Pos events, showed no significant differences between the two CB groups (Table 3).
Table 3. Analysis of lymphocytes subsets and monocytes in hEPCB and hTCB samples
25TH, 75TH PERCENTILES (%)
25TH, 75TH PERCENTILES (%)
The percentage of T-cells is higher in hTCB compared with hEPCB; conversely, the B-cell compartment is principally displayed in hEPCB compared with hTCB samples. No significant differences were evidenced regarding the percentage of NK cells and monocytes in the two kinds of samples.
Multidimensional Analysis of B-Cells: hEPCB Samples Display a Higher Percentage of Immature B-Lymphocytes
We screened eight hEPCB and 11 hTCB samples selecting immature B-cells as CD10PosCD20Pos events within the CD45PosCD19Pos fraction (Figs. 6A and 6B). Interestingly, hEPCB samples showed a higher percentage of immature B cells compared with hTCB specimens being of 48.2% (25th–75th percentiles = 29.1–60.9) and 19% (25th–75th percentiles = 11.5–24.7), respectively (P < 0.05) (Fig. 6C).
To more precisely evaluate B-cell development stages, we analyzed CD34, CD19, and CD10 sequential expression on the CD45Dim/Pos population in eight hEPCB and 11 hTCB samples. In particular, we compared the percentages of CD34Pos/CD19Pos/CD10Pos pro-B, CD34Neg/CD19Pos/CD10Pos pre-B, and CD34Neg/CD19Pos/CD10Neg mature B-cells between the two kinds of samples. Our findings evidenced that pro- and pre-B cells were principally displayed in hEPCB samples, whereas the latter population was mainly showed in hTCB samples (P = 0.014, data not shown).
The definition of HSC antigenic mosaic by multidimensional FCM is one of the key tools available to study stem cell biology (5, 7, 16–20). We decided to exploit this technology to propose a polychromatic FCM characterization of HSCs from long-term cryopreserved hEPCB samples and to compare their immunophenotypic pattern with counterparts from cryopreserved full-term CB. Regarding to long-term cryopreservation, current practice indicates that standard cryopreservation protocols of CB freezing in 10% DMSO in controlled-rate freezers and storage below −135°C give an average of 80% recovery of multipotent hematopoietic progenitor cell subpopulation, as demonstrated by functional studies in terms of colony forming unit assay (21, 22).
According to scientific reports using three- (5, 16–18) or four-color (7) FCM, we confirmed the higher percentage of CD34PosCD45Dim HSCs residing in hEPCB samples compared with the hTCB counterpart (P < 0.0001).
In addition, we studied the correlation between percentages of CD34PosCD45Dim cells and gestation age. Surprisingly, we observed that the percentage of human HSCs decreased linearly during the gestational period between the 16th and 20th week, whereas no significant difference appeared among the samples with the highest gestational age. There is a partial discrepancy between the present results and earlier findings, suggesting a complete absence of a linear regression in samples ranged between 17th and 21st week of gestation (9). However, our data can be in agreement with the fetal hematopoietic mechanism theory (16). In fact, although during the late first trimester of pregnancy, no hematopoietic progenitors are still detectable in fetal circulation (16), the gestational period between the 16th and 20th week corresponds to the well-described developmental phase in which fetal circulation shows the highest concentration of HSC homing from fetal liver to marrow.
The identification of HSCs with high repopulating ability is one of the principal objectives of clinical transplantation (23). One of the first molecules able to permit the identification of long-term repopulating stem cells (LTRSCs) is CD38. In particular, the CD34PosCD38Neg fraction was identified as the human hematopoietic compartment containing LTRSCs, whereas the CD34PosCD38Pos as the part enriched for cells capable of providing short-term and intermediate stage hematopoiesis in humans (24). It is worth noting that our data not only demonstrated the highest percentage of CD34PosCD45 DimCD38Neg cells in hEPCB but also the favored expression, in terms of fluorescence intensity, of CD38 at the surface of HSCs of hTCB samples.
In an attempt to better describe the functional attitudes of CB-HSC compartments, we screened the CD34PosCD45Dim population also for membrane efflux-related proteins. We found an interesting result as concerns CD243 (MDR1); in particular, in case of hTCB samples, a small subpopulation characterized by CD34PosCD45DimCD243Pos immunophenotype was observed. The expression of CD243 on this subset of HSCs could be explained considering the physiological function of CD243 as a key molecule involved not only in inducing multidrug resistance but also in protecting hematopoietic cells from undergoing apoptosis for a period sufficiently long to enable successive cytokine-supported cell differentiation (25).
As regards CD135 (FLT3), its significant higher expression on hematopoietic progenitor cells of hTCB in terms of MFI can be explained considering the increase of cells committed to the myelomonocytic lineage in late pregnancy as demonstrated by Rappold et al. (26) who observed the strong activity of FLT3 ligand on full-term CB myeloid precursor cells as well as a higher expression of CD135 on full-term CB monocytes.
In this study, our results contain the first evidence that the CD34PosCD45DimCD38NegCD133Pos cellular fraction is principally displayed in hEPCB samples compared with the hTCB. Our findings could make hEPCB samples a potentially better source of highly immature HSC population, typically displayed in second trimester CB and potentially able to give long-term repopulating activity (27). On the other hand, even if it is clear that the quality of a CB unit is dependent on the level of its total nucleated cells content (28), the detection and, therefore, the biological knowledge of such a rare hEPCB CD34PosCD45DimCD38NegCD133Pos subpopulation could prove useful for basic stem cell research as well as for potential manipulation in ex vivo clinical-related expansion protocols (24, 29). We are confident that technical advancements in hEPCB stem cells handling will facilitate clinical studies, positively influencing improvement in engraftments rates in the near future.
In this report, we also offered an immunophenotypic comparison between undifferentiated and committed HSCs residing in the same sample group. Of note, this matching was able to point out considerable differences as regards the expression of adhesion molecules, transferrin, and tyrosine kinase receptors. In particular, although CD34PosCD45 DimCD38Pos HSCs from hEPCB and hTCB were clearly positive for CD29, CD71, and CD135, a marked expression of CD31 and CD117 is restricted to CD34PosCD45DimCD38Pos hEPCB and hTCB cells, respectively. Although the physiological role of these peculiar antigenic assortments characterized by homing (30–33) and signal transduction receptors (26, 34, 35) is fully known on adult HSCs, an important task in the near future will be to elucidate by in vivo system the hEPCB committed HSCs self-renewal capacity, thus assessing their possible critical value to sustain human long-term hematopoiesis.
In this study, we found a higher percentage of T-cells in the hTCB group of samples. This result could be elucidated considering the maturation process of cortical thymocytes that lead to T-cells increase during pregnancy followed by a decline in the late childhood (36, 37).
In contrast, the highest absolute count of B-lymphocytes was found in fetuses, followed by a gradual decrease up to birth. In fact, before 28 weeks of gestation, a steady increase in the number of B-cells is documented, but maturation occurs mainly in late pregnancy (38).
We studied the B-cell lineage maturation on the basis of sequential expression of specific surface antigens such as CD10, CD19, CD20, CD34, and CD45. In doing so, we first evidenced a higher percentage of CD45PosCD19PosCD10 PosCD20Pos immature B-cells in hEPCB compared with the full-term counterparts. Afterward, the CD34, CD19, and CD10 sequential expression on the CD45Dim/Pos population helped us to determine the different stages of B-cell development. In particular, we observed that the CD34Pos/CD19Pos/CD10Pos pro-B and CD34Neg/CD19Pos/CD10Pos pre-B cell populations increased in hEPCB compared with hTCB samples. These findings are in agreement with recent literature on B-cell development (39, 40).
In conclusion, we are aware that hematopoietic and immunogenetic properties of third-trimester CB samples have not yet been exploited systematically, presumably because of their lack of easy availability. Nevertheless, with the efficient ex vivo expansion of HSC as well as stem cell transduction techniques, CB from developing fetuses might serve as a target for in utero gene therapy for genetic diseases amenable to prenatal diagnosis or for in utero or postnatal autologous stem cell transplantation.
Taken together, our data indicate that hEPCB samples may satisfy the requirements to be considered as a potential starting place for planning future cell therapy studies with a view to estimate their possible availability in the wide field of transplantation medicine.
The authors gratefully acknowledge the support of CEINGE-Biotecnologie Avanzate di Napoli Biobank staff for providing human full-term cord blood samples.