Characterization of the Adherent Cells Developed in Dexter-Type Long-Term Cultures from Human Umbilical Cord Blood

Authors


Abstract

We have previously shown that when human umbilical cord blood (UCB) cells are cultured in standard Dexter-type long-term cultures (D-LTC), adherent cells develop forming a discrete net on the bottom of the culture flask. The identity of such cells, however, has not been defined. Accordingly, the major goal of the present study was to characterize the adherent cells developed in standard UCB D-LTC. Cultures were established from 14 UCB samples and from nine bone marrow (BM) samples, as controls. Both UCB and BM cultures were initiated with the same number of mononuclear cells (MNC) (2.5 × 106 MNC/ml). After three weeks in culture, adherent cell numbers in UCB D-LTC were 24%-30% of the numbers found in BM cultures. More than 90% of the adherent cells in UCB D-LTC expressed the acid phosphatase enzyme, whereas no alkaline phosphatase-positive cells were observed. This was in contrast to BM D-LTC, in which alkaline and acid phosphatase were expressed by 60%-75% and 20%-45% of the adherent cells, respectively. Immunochemical analysis showed that CD61 (osteoclast marker) and Factor VIII (endothelial cell marker) were not expressed by the adherent cells developed in UCB cultures. Interestingly, the majority of such cells expressed CD1a (dendritic cell marker), CD14, CD68 and CD115 (antigens mainly expressed by macrophagic cells). When the cultures were supplemented with the recombinant cytokines epidermal growth factor, basic fibroblast growth factor, platelet-derived growth factor or granulocyte-macrophage colony-stimulating factor (GM-CSF), only GM-CSF had a significant positive effect on adherent cell number. In order to test for some functional properties of the adherent cells developed in culture, production of stem cell factor (SCF), interleukin 6 (IL-6) and tumor necrosis factor-α (TNF-α) was assessed. IL-6 and TNF-α showed elevated levels in UCB D-LTC, whereas SCF levels were always below detection. Finally, analysis of fibroblast progenitors (fibroblast colony-forming units [CFU-F]) showed that these cells were present in BM samples (6 CFU-F/105 MNC) and were totally absent in UCB samples. Taken together, the results of the present study indicate that the vast majority of the adherent cells developed in standard UCB D-LTC belong to the macrophage lineage and that fibroblasts seem to be absent. Interestingly, the high proportion of CD1a+ cells suggests that dendritic cells are also present in these cultures.

INTRODUCTION

Umbilical cord blood (UCB) has been recognized as a rich source of primitive hematopoietic stem/progenitor cells (HSPC) [1-3]. Studies by different groups of investigators have demonstrated that UCB-derived HSPC possess very high proliferation and expansion potentials, which exceed those from their bone marrow (BM) counterparts. This has been shown both in vitro in liquid cultures supplemented with recombinant cytokines [4-6], and in vivo in animal models, in which immunodeficient mice are transplanted with human HSPC [7, 8].

Interestingly, when Dexter-type long-term cultures (D-LTC) are established from human UCB, HSPC proliferation is only transiently sustained. Indeed, several studies have shown that, compared to BM D-LTC, progenitor cell levels in UCB D-LTC are significantly lower throughout the entire culture period and reach undetectable levels several weeks earlier [9-11]. These findings have been explained by the fact that, in UCB D-LTC, a confluent stromal adherent cell layer always fails to form, leaving HSPC without a competitive hematopoietic microenvironment and a source of stimulatory cytokines [12]. It is noteworthy, however, that a single study has been published showing data that differ from the results described above. Ye and colleagues reported on the establishment of confluent adherent cell layers from human UCB, that contained stromal cells and that were capable of supporting HSPC growth for several weeks [13]. Although the culture conditions they used in their study seemed to be very specific and are not used by most of the investigators, their results raise the question as to whether stromal elements are present in UCB.

As mentioned above, in our studies we have not been able to develop confluent stromal cell layers in UCB D-LTC [11]. Instead, we consistently observe the presence of single adherent cells that in certain areas of the culture form a discrete net. The number of such cells is usually 25%-50% of the numbers of adherent cells found in D-LTC from BM. To date, however, we have not established the identity of such adherent cells. That is to say, it is not clear if nonhematopoietic stromal cells are present in these cultures or if the adherent cells observed are all of hematopoietic origin (i.e., macrophages). Thus, the main goal of the present study was to identify the adherent cells developed in standard UCB D-LTC and to characterize some of their functional properties.

MATERIALS AND METHODS

Cell Collection

UCB cells, collected according to institutional guidelines, were obtained from 14 normal full-term deliveries from the “La Raza” and “20 de Noviembre” Hospitals (Mexico City). Nine BM samples were obtained from the iliac crest of BM transplant donors, at the “Bernardo Sepulveda” Hospital, National Medical Center (Mexico City). These procedures have been approved by the ethical committee of the National Medical Center.

Cell Processing

Buffy coat cells, both from UCB and BM, were obtained by centrifugation (400 g for 7 min) and low-density mononuclear cells ([MNC]; <1.077 g/ml) were isolated using Ficoll-Paque Plus (Pharmacia Biotech; Uppsala, Sweden). Cells were then resuspended in Iscove's modified Dulbecco's medium (IMDM) supplemented with 2% fetal bovine serum ([FBS]; StemCell Technologies Inc. [STI]; Vancouver, BC, Canada). Total numbers of nucleated and viable cells were determined with a hemocytometer, using Turck's solution and trypan blue stain, respectively.

D-LTC

D-LTC were established as previously described [11]. Low-density MNC were resuspended in LTC medium (STI) at a final concentration of 2.5 × 106 cells per ml. The LTC medium composition is as follows: alpha medium supplemented with 12.5% horse serum, 12.5% FBS, 0.2 mM inositol, 20 μM folic acid, 10–4 M 2-mercaptoethanol, 2 mM L-glutamine, and freshly dissolved hydrocortisone to yield a final concentration of 10–6 M. The cell suspension was loaded into 24-well plates (1 ml/well) and incubated at 37°C in an atmosphere of 5% CO2 in air. After three days cultures were transferred to a different incubator and maintained at 33°C. Four days later (seven days after initiation of the culture) one-half of the supernatant and nonadherent cells were removed from the wells and replaced with fresh culture medium. The cultures were processed in this manner on week 2. From week 3, culture medium change was done by removing 100% of the volume and replacing it with fresh culture medium. On certain weeks one of a number of parallel cultures was sacrificed for evaluation of the adherent cells. These were detached with a cell scraper after trypsinization (i.e., 0.25% trypsin containing 0.1 mM EDTA was added and the cultures were incubated at 37°C for 10 min; the action of trypsin was stopped by adding one-half volume of FBS). The cells were then resuspended in IMDM with 2% FBS and total numbers of nucleated and viable cells were determined with a hemocytometer, using Turck's solution and trypan blue stain, respectively.

In order to test for the effects of recombinant cytokines on the development of the adherent cells in culture, D-LTC from UCB were also established in the presence of recombinant cytokines that have been shown to have an effect on stromal cells; these included recombinant human (rHu) epidermal growth factor (rHuEGF; 50 ng/ml), rHu basic fibroblast growth factor (rHubFGF; 50 ng/ml), rHu platelet-derived growth factor (rHuPDGF; 12.5 ng/ml) or rHuGM-CSF (10 ng/ml). Cytokines were added weekly during medium change. GM-CSF (Molgramostim; Novartis/Schering-Plough; Basel, Switzerland) was a gift from Novartis Farmaceutica SA de CV, Mexico. All other cytokines were purchased from R&D Systems (Minneapolis, MN).

Identification of Adherent Cells

This was performed by histochemical and immunochemical analyses. On weeks 3, 5 and 7, adherent cells were detached (as described above) and used for “cytospin” preparations. For histochemical analysis, alkaline phosphatase (a fibroblast marker) and acid phosphatase (a macrophage marker) expression were determined using naphtol AS-biphosphate as a substrate, as previously described [14, 15]. Immunochemical studies consisted of the determination of the expression of the following antigens: CD1a (dendritic cell marker) [16]; CD14 (monocyte marker) [17]; CD61 (osteoclast marker) [18]; Factor VIII (endothelial cell marker) [19]; CD68 (macrophage marker) [20] and CD115 (M-CSF receptor) [21]. Antigen determination was performed on days 7, 14, 21, and 28. In most cases mouse antihuman monoclonal antibodies were used together with a biotin-coupled antimouse antibody and a streptavidine-peroxidase amplifier system (DAKO; Carpinteria, CA). CD115 expression was determined by using a rat antihuman monoclonal antibody (Calbiochem; La Jolla, CA). Observations were made with a light microscope (Carl Zeiss, ICS Standard 25).

Fibroblast Colony-Forming Unit (CFU-F) Assay

CFU-F were assayed according to the method described by Castro-Malaspina et al. [22]. MNC were inoculated at 3 × 105 cells/ml, in 35-mm petri dishes, containing 1 ml of IMDM and 20% FBS. The cultures were incubated at 37°C in an atmosphere of 5% CO2 in air. After three days the nonadherent cells were removed and the medium was changed. The cultures were returned to the incubator for seven more days. At the end of the period, the medium was discarded and the adherent cells were stained with Wright-Giemsa. Clones of >50 fibroblasts were scored as fibroblastic colonies. In UCB samples, CFU-F assays were also established in the presence of rHuEGF (50 ng/ml), rHuFGF (50 ng/ml), rHuPDGF (12.5 ng/ml) or rHuGM-CSF (10 ng/ml). BM CFU-F assays were also established in the presence of rHuGM-CSF.

Cytokine Levels

On weeks 3, 5 and 7 the levels of stem cell factor (SCF), interleukin 6 (IL-6) and tumor necrosis factor-α (TNF-α), present in D-LTC supernatants, were determined by enzyme-linked immunoabsorbent assay (ELISA) using commercial kits from R&D Systems. The limit of detection of each ELISA kit was as follows: SCF = 4.0 pg/ml; IL-6 = 0.09 pg/ml; TNF-α = 0.18 pg/ml.

Statistics

Statistical analysis was performed by using Student's t-test.

RESULTS

CFU-F Assays

As a first approach to the identification of the adherent cells developed in UCB D-LTC, we wanted to test the possibility that fibroblast progenitor cells may be present in UCB (if this were the case, such cells could give rise to mature fibroblasts in culture). Thus, CFU-F assays were performed in all 14 freshly obtained UCB samples. CFU-F were also determined in normal BM samples. As shown in Table 1, 3, Table 3.-8 CFU-F per 105 MNC were observed in BM samples; similar numbers were observed when the assays were established in the presence of rHuGM-CSF. In contrast, all UCB samples tested showed no fibroblast progenitors at all. Even if the cultures were established in the presence of rHuEGF, rHubFGF, rHuPDGF or rHuGM-CSF, CFU-F levels were always below detection.

Adherent Cell Numbers in D-LTC

On weeks 3, 5 and 7 total adherent cell numbers were determined both in UCB and BM D-LTC. As shown in Table 2, the numbers observed were similar to those reported previously by us [11]. At all time points analyzed, the levels of adherent cells in BM D-LTC were 3.5- to 4.1-fold higher than in UCB cultures. As expected, adherent cell layers in BM D-LTC consisted of confluent stromal layers (not shown). In contrast, and confirming our previous observations, adherent cells in UCB cultures consisted of single cells distributed throughout the well. Most of them were macrophage-like cells and others had a spindle shape (Fig. 1).

Figure Figure 1..

Appearance under the inverted microscope of a representative (week 3) D-LTC established from UCB.The figure shows a phase contrast photograph (400×) of the adherent cells.

Table Table 1.. Fibroblast progenitor cell content in umbilical cord blood and bone marrow
 CFU-F/105MNC
CytokinesUCB(n= 14)BM(n= 9)
  1. a

    Results represent median (range) of the indicated number of experiments. BD = below detection; ND = not done.

 BD6 (3-8)
EGFBDND
bFGFBDND
PDGFBDND
GM-CSFBD5 (4-8)
Table Table 2.. Adherent cell number in D-LTC from UCB and BM
WeekUCB(n= 4)BM(n= 6)
  1. a

    Results represent mean ± SD of the indicated number of experiments.

  2. b

    *Significantly lower (p < 0.05) than in BM.

3122,625 ± 70,268*430,167 ± 236,511
573,750 ± 34,913*304,166 ± 78,833
796,188 ± 50,128*328,333 ± 137,143
Table Table 3.. Cytokine production by adherent cells present in D-LTC from UCB
WeekSCFIL-6TNF-α
  1. a

    Results (mean ± SD of four experiments) correspond to pg/ml.

  2. b

    BD = below detection.

3BD93 ± 2360 ± 7
5BD64 ± 2047 ± 18
7BD79 ± 497 ± 2

In order to test the possibility of increasing adherent cell numbers in UCB D-LTC by stimulation with recombinant cytokines, UCB D-LTC were established in the presence of cytokines that have been shown to have a positive effect on the growth of stromal cells. Accordingly, cultures were supplemented with rHuEGF, rHubFGF, rHuPDGF or rHuGM-CSF, added on day 0 and weekly during medium change. As shown in Figure 2, among all the cytokines tested, only rHuGM-CSF had a significant effect on increasing adherent cell numbers. It is noteworthy, however, that the levels observed in rHuGM-CSF-supplemented cultures (2.3 ± 0.41-fold higher than in untreated cultures) were still lower than the levels found in standard BM cultures (3.8 ± 0.3-fold higher than in UCB D-LTC).

Figure Figure 2..

Effect of recombinant cytokines on the number of adherent cells developed in standard UCB-derived D-LTC.Results correspond to the fold increase compared to untreated cultures. Cytokines were added weekly during medium change at the following concentrations: EGF, 50 ng/ml; bFGF, 50 ng/ml; PDGF, 12.5 ng/ml; GM-CSF, 10 ng/ml. GM-CSF was the only cytokine that had a significant effect (p < 0.05) on adherent cell number.

Histochemical Analysis

Analysis of alkaline and acid phosphatase expression, both in UCB and BM cultures, showed that in the former, more than 90% of the adherent cells expressed acid phosphatase, whereas we could not detect any alkaline phosphatase positive cell (not shown). This was also observed in cultures that were established in the presence of rHuEGF, rHubFGF, rHuPDGF or rHuGM-CSF (not shown). In contrast, in BM D-LTC, alkaline phosphatase-positive cells comprised 60%-75% of the adherent cells, whereas acid phosphatase-positive cells corresponded to 20%-45% (not shown).

Immunochemical Analysis

During four weeks of culture, adherent cells in UCB D-LTC were analyzed in terms of the expression of CD1a, CD14, CD61, CD68, CD115 and Factor VIII. At all time points analyzed, we were unable to detect any cell expressing either CD61 or Factor VIII, markers for osteoclasts and endothelial cells, respectively (not shown). In contrast, expression of the dendritic cell marker CD1a and the monocyte-macrophage markers CD14, CD68 and CD115 was observed in a significant proportion of the adherent cells. As shown in Figure 3, by the first week of culture CD1a was expressed by 58% of the adherent cells, CD14 was expressed by 40%, CD68 by 83% and CD115 by 62% of the cells. After four weeks in culture, the levels of CD1a+ cells went up to 84%; CD14+ cell levels increased to 60%; CD68+ cell levels remained around 85% and CD115+ cell levels showed a slight increase (73% of the adherent cells).

Figure Figure 3..

Immunochemical analysis of the adherent cells developed in UCB D-LTC.Results correspond to the relative proportion of adherent cells expressing CD1a, CD14, CD68 or CD115.

Cytokine Production

In order to assess the capacity of the adherent cells developed in UCB D-LTC to produce hematopoietic cytokines, we quantified the levels of SCF, IL-6 and TNF-α present in culture supernatants on weeks 3, 5 and 7. The levels of these cytokines were also assessed in D-LTC from BM. As shown in Table 3, at all time points analyzed SCF levels in UCB cultures were below detection. This was in sharp contrast to the results obtained in supernatants from BM D-LTC, in which SCF levels were between 436 and 579 pg/ml. IL-6 levels, on the other hand, were higher in supernatants from UCB cultures than in their BM counterparts. Indeed, in the former cultures, IL-6 mean levels were 64-93 pg/ml, whereas in BM cultures the levels were 27-46 pg/ml. Finally, TNF-α levels were also higher in supernatants from UCB D-LTC (7-60 pg/ml) than in BM culture supernatants (<11 pg/ml). It is noteworthy, however, that TNF-α levels decreased to almost undetectable levels throughout the culture period.

DISCUSSION

It has been widely demonstrated that when BM cells are cultured according to the method described by Dexter et al. [23] and Coulombel et al. [24], known as D-LTC, a confluent adherent layer develops, which consists of stromal cells and their products [12]. Previous studies from our group have shown that when UCB cells are cultured under these conditions, confluent stromal cell layers fail to form; instead, individual adherent cells develop, forming a subtle net that physically interacts with the hematopoietic cells [11]. Interestingly, such an adherent layer clearly differs from the one in BM D-LTC, both in total cell numbers and organization. The identity of UCB adherent cells, however, has not been established. One possibility is that these cells are all of hematopoietic origin (e.g., macrophages); a second possibility is that nonhematopoietic stromal cells (e.g., fibroblasts) are present among such adherent cells. In trying to address this issue, we have conducted the present study, in which the major goal was to identify and characterize the adherent cells developed in UCB D-LTC.

In D-LTC established from BM, fibroblasts comprise >50% of the adherent cells developed in culture [12]. Therefore, we asked the question of whether fibroblasts were present in our UCB cultures. A whole series of experiments indicated that this was not the case. First, we observed that whereas BM samples contained a mean of 6 CFU-F per 105 MNC, UCB samples showed no CFU-F at all. This was observed even in assays in which recombinant cytokines (rHuEGF, rHubFGF, rHuPDGF or rHuGM-CSF) were added to the cultures. Secondly, when adherent cells developed in UCB D-LTC were analyzed for the expression of alkaline phosphatase (a fibroblast marker), we were unable to detect any alkaline phosphatase-positive cell. This was in contrast to BM cultures, in which alkaline phosphatase was present in 60%-75% of the adherent cells. Finally, we could not detect production of SCF, a fibroblast product, in UCB cultures. In contrast, BM adherent cells produced significant amounts of this cytokine. Taken together, these results indicate that, under the culture conditions used in our study, fibroblasts were not present among the adherent cells developed in UCB D-LTC.

In a previous study, Ye and colleagues [13] reported on the development of confluent stromal adherent layers in UCB D-LTC. These adherent cell layers contained a high proportion of fibroblasts and were able to produce detectable, although low, amounts (20-35 pg/ml) of SCF. In their study the authors maintained the cultures at 37°C for the whole culture period and used customized glass coverslips, that had a very different surface (no detectable granular particles or coarse bump structures) from that of standard glass coverslips or standard plastic surface. These culture conditions clearly differ from the ones that we used. Thus, it is possible that these methodological differences account for the differences observed in terms of the adherent cells developed in D-LTC.

As for fibroblasts, the presence of osteoclasts and endothelial cells in UCB D-LTC could not be demonstrated by our immunochemical analysis. These results are in contrast to those by Nieda and colleagues, who recently reported on the identification of endothelial cells in cultures established from UCB [25]. It is noteworthy, however, that there were important methodological differences between both studies that could explain the differences in the results. On the one hand, Nieda et al. established their cultures from UCB-derived CD34+ cells, whereas we started our cultures with low-density, MNC. On the other hand, they supplemented their cultures with rHu IL-2; we tested several cytokines (rHuEGF, rHubFGF, rHuPDGF and rHuGM-CSF), but not IL-2.

The results of the present study indicate that the vast majority of the adherent cells developed in standard UCB D-LTC are of hematopoietic origin and belong to the monocyte/macrophage lineage. Indeed, more than 90% of the cells expressed acid phosphatase, and the antigens CD14, CD68 and CD115 were expressed by a mean of 50%, 83% and 70% of the adherent cells, respectively. In keeping with these observations, we found significant production of IL-6 and TNF-α (two macrophage products) in UCB culture supernatants. It is noteworthy that the IL-6 levels observed in our study (mean levels of 64-93 pg/ml) were similar to those reported by Ye et al. (68-103 pg/ml; [13]), and that these levels were higher than those observed in BM D-LTC.

Also of interest was the fact that TNF-α levels in UCB cultures were significantly higher (four- to sixfold) than in BM D-LTC. This result is particularly interesting because it may suggest that the macrophages developed in BM and UCB cultures differ in their ability to produce TNF-α. That is to say, if we assume that 80% of the adherent cells in UCB D-LTC were macrophages, their average number would be 78,016/well (approximate mean number in weeks 3, 5 and 7; taken from Table 2). If we assume that 30% of the adherent cells present in BM D-LTC were macrophages (based on our results of acid phosphatase expression), their mean number would be around 106,266/well. Thus, UCB macrophages would produce significantly higher levels of TNF-α compared to higher numbers of BM-derived macrophages. This notion, however, should be taken with caution, since we are not certain of the actual number of macrophages in each type of culture. Studies would be necessary in which purified BM and UCB macrophages were cultured under similar conditions, to assess and compare their capacity to produce and secrete TNF-α and other cytokines.

We found that CD1a, a dendritic cell marker, was expressed by a high proportion of the adherent cells developed in UCB D-LTC. This was observed even at week 1, on which 43%-73% of the cells were CD1a+. Interestingly, the numbers of CD1a+ cells increased significantly during the culture period. In this regard, it is important to mention that TNF-α, which is present at significant levels in UCB cultures, has been shown to be a key factor in dendritic cell development [26, 27]. Thus, CD1a+ cells developed in our cultures may result from the high TNF-α levels in culture supernatants.

Although it is known that dendritic cells may express the CD68 antigen, CD14 and CD115 seem to be exclusively expressed by monocyte/macrophages [28]. Thus, our results indicate that a significant proportion of the adherent cells developed in UCB D-LTC possess features of both macrophages and dendritic cells. In the present study, we did not attempt to purify and characterize such a cell population; however, our observation may be of relevance, since it has been previously reported on the existence of immature bipotent CD34+ cells with the capacity to give rise to both macrophages and dendritic cells [29]. Purification, characterization and in vitro culture of dendritic cells have become issues of significant relevance, due to their role as antigen-presenting cells and their potential application in immunological therapy [30-32]; thus further studies aimed at the biological characterization of UCB-derived dendritic cell populations should be encouraged, since UCB seems to be an excellent source of such cells.

Acknowledgements

This work was partially supported by grant no. 0122P-M9506 from the National Council of Science and Technology (CONACYT; Mexico) and by grant no. FP0038-366 from the Mexican Institute for Social Security (IMSS).

The authors would like to thank Dr. Veronica Sánchez (Gynecology Hospital, CMN La Raza) and Dr. Fernando Escobedo (CMN 20 de Noviembre) for making UCB samples available for this study. Drs. Enrique Gómez-Morales and Elizabeth Sánchez are also thanked for providing normal BM samples.

Ancillary