Human Umbilical Cord Perivascular (HUCPV) Cells: A Source of Mesenchymal Progenitors



We describe the isolation of a nonhematopoietic (CD45, CD34, SH2+, SH3+, Thy-1+, CD44+) human umbilical cord perivascular (HUCPV) cell population. Each HUCPV cell harvest (2–5 × 106, depending on the length of cord available) gave rise to a morphologically homogeneous fibroblastic cell population, which expressed α-actin, desmin, vimentin, and 3G5 (a pericyte marker) in culture. We determined the colony-forming unit-fibro-blast (CFU-F) frequency of primary HUCPV cells to be 1:333 and the doubling time, which was 60 hours at passage 0 (P0), decreased to 20 hours at P2. This resulted in a significant cell expansion, producing over 1010 HUCPV cells within 30 days of culture. Furthermore, HUCPV cells cultured in nonosteogenic conditions contained a subpopulation that exhibited a functional osteogenic phenotype and elaborated bone nodules. The frequency of this CFU-osteogenic subpopulation at P1 was 2.6/105 CFU-F, which increased to 7.5/105 CFU-F at P2. Addition of osteogenic supplements to the culture medium resulted in these frequencies increasing to 1.2/104 and 1.3/104 CFU-F, respectively, for P1 and P2. CFU-O were not seen at P0 in either osteogenic or non-osteogenic culture conditions, but P0 HUCPV cells did contain a 20% subpopulation that presented neither class I nor class II cell-surface major histocompatibility complexes (MHC−/−). This population increased to 95% following passage and cryopreservation (P5). We conclude that, due to their rapid doubling time, high frequencies of CFU-F and CFU-O, and high MHC−/− phenotype, HUCPV cells represent a significant source of cells for allogeneic mesenchymal cell-based therapies.


Since it was first used to treat a patient with Wiskott-Aldrich syndrome [1], bone marrow (BM) has been the most common source of cells for cell-based therapies. The mesenchymal population of BM is targeted for a variety of therapeutic approaches affecting a wide range of tissues, including those of the musculoskeletal system: bone [24], cartilage [58], and tendons and ligaments [912]. BM cell therapy has also been suggested for repair of the myocardium [1317] and is being pursued clinically for applications in hematology and oncology such as aplastic anemia [1820] and malignant lymphoma [21]. Following encouraging results in nonobese diabetic/severe combined immunodeficient mice [22, 23], Koç et al. [24] have shown beneficial clinical outcomes by coinfusion of culture-expanded mesenchymal cells with hematopoietic stem cells in patients treated with high-dose chemotherapy for solid tumors. Other promising therapeutic approaches include mesenchymal stem cells (MSCs) as carriers of the therapeutic genes [25] or the infusion of allogenic BM for the treatment of osteogenesis imperfecta (OI) [26, 27]. In the latter, BM was from an HLA (human leukocyte antigen)–identical or single mismatched sibling. However, since immune rejection [28] and donor number limitations [29] are major constraints to common use, there is an acute need to find alternative cell sources for such cell-based therapies. As cells are a fundamental requirement for tissue engineering [30], cell sourcing also remains a major challenge for human tissue-engineering strategies.

One potential alternative source of mesenchymal cells became feasible with the report by McElreavey et al. [31] of the culture of cells from Wharton's jelly (WJ), the primitive connective tissue of the human umbilical cord (UC), first described by Thomas Wharton in 1656 [32]. Thus, Naughton et al. [33] and Purchio et al. [34] derived “prechondrocytes,” from explant cultures of UC WJ, and Mitchell et al. [35], using a similar approach, reported that the fibroblast-like cells of WJ could be induced to differentiate into “neural-like” cells expressing neuron-specific enolase (NSE), as well as other neural cell markers. Romanov et al. [36], using a different approach, enzymatically digested mesenchymal precursor cellsfrom the UC vasculature endothelial surface, and Kadner et al. [37, 38] minced either UC vessels or whole cord to derive an autologous cell source of myofibroblasts for cardiovascular tissue engineering. Chacko and Reynolds [39] described the cells residing in WJ as “smooth muscle cells,” but Takechi et al. [40] refined the description to “myofibroblasts” after in situ labeling of vimentin, desmin, α-actin, and myosin, which has been recently confirmed by Kadner et al. [38].

The human UC is embryologically derived at day 26 of gestation, and it grows to form a 30- to 50-cm-long helical organ at birth. Given this expansion, during the 40 weeks of gestation, there must be a mesenchymal precursor cell population within the UC that gives rise to the WJ connective tissue. We postulated that these cells would most likely be located closest to the vasculature, and thus to their source of oxygen and nutrients. Consequently, we reasoned that human umbilical cord perivascular (HUCPV) cells, which were either discarded, or not specifically isolated, in the previously described studies, should contain a subpopulation that, when isolated, would be capable of exhibiting a functional mesenchymal phenotype.

Thus, we report herein a novel harvesting protocol designed to isolate HUCPV cells and show that the resultant cell population possesses a high frequency of colony-forming unit-fibroblast (CFU-F)–deriving cells [41] that proliferate and differentiate rapidly to form bone nodules (BNs). Furthermore, we show that the isolated cell population includes an expanding subpopulation that expresses neither class I nor class II major histocompatibility (MHC) antigens, suggesting a potential role as a human allogeneic cell source for cell-based therapies.

Materials and Methods

Ethical approval was obtained from both the University of Toronto and Sunnybrook and Women's College Health Sciences Centre, Toronto. The UCs from consenting full-term caesarian section patients were provided immediately upon delivery in a previously supplied vessel containing 80% α-MEM (Gibco Burlington, ON, Canada no. 12571)) and 20% antibiotics (penicillin G at 167 units/ml; Sigma Oakville, ON, Canada no. P-3032), gentamicin (50 μg/ml; Sigma no. G-1397), and amphotericin B (0.3 μg/ml; Sigma no. A9528). Pieces of cord, 4–5 cm long, were dissected by first parting the epithelium of the UC section along its length to expose the underlying WJ. Each vessel, with its surrounding WJ matrix, was pulled away, and the ends of each dissected vessel were tied together with a suture creating “loops” that were placed into a 50-ml tube (Falcon Mississauga, ON, Canada no. 352070) containing a solution of 1 mg/ml collagenase (Sigma no. C-0130) with phosphate buffered saline (PBS). The remaining two vessels were dissected in a similar fashion, then looped and placed in the collagenase solution. After 18–24 hours, the loops were removed from the suspension, which was then diluted with PBS to reduce the viscosity of the suspension and centrifuged. Following the removal of the supernatant, the cells were resuspended in 10 ml PBS and counted using a hemocytometer. The suspended cells were run through an Easy Sep magnetic bead conjugated-CD45 depletion protocol (Stem Cell Technologies [Vancouver, Canada] no. 18259) to remove any hematopoietic cells, then observed by flow cytometry for expression of CD45 and cell-surface antigens (see below). Finally, the cells were plated in T-75 cm2 tissue culture polystyrene dishes (Falcon no. 353136) in supplemented medium (SM) (75% α-MEM, 15% fetal bovine serum [FBS]; Stem Cell Technologies no. S13E40), and 10% antibiotics, which was changed every 2 days.

Subculture and Cell Seeding

At day 7, adherent cells, judged 80%–90% confluent by phase contrast microscopy, were passaged using 0.1% trypsin solution (Gibco no. 27250-042). They were then plated in T-75 tissue culture polystyrene flasks at 4 × 103 cells/cm2 in SM.

Antibody Staining

HUCPV cells were prepared for antibody staining following culture for 5 days on four-well glass chamber slides (Labtek no. 0107-0005). The cells were fixed in 3.7% formalin for 5 minutes, permeabilized by incubation with 100% methanol for 2 minutes at room temperature, and washed three times in 2% FBS/PBS. They were blocked with 10% FBS/PBS for 60 minutes, then incubated for a further 60 minutes with the following primary mouse-anti-human antibodies (1 μ1/100 μ1 PBS): α-smooth muscle actin, desmin, vimentin (all Sigma), neuron-specific enolase (Cymbus Biotechnology, Eastleigh, U.K.), and a 3G5 monoclonal antibody to microvascular pericytes (kind gift from Dr. A. Canfield, Manchester, U.K.). The cells were then washed three times in 2% FBS/PBS and incubated with two drops of Alexa Fluor 488 goat anti-mouse immunoglobulin G (IgG) 2mg/mL secondary antibody (Molecular Probes [Eugene, OR] no. A-11001) for 20 minutes, then washed again three times in 2% FBS/PBS. The cells were finally counterstained with nuclear Hoechst 33258 (observed as blue fluorescence). The primary antibody was omitted to produce negative controls. The labeled samples were mounted on glass slides, and positive staining was observed as green fluorescence.

Limiting Dilution and CFU-F Assays

Dilutions of 1 × 105, 5 × 104, 2.5 × 104, 1 × 104, 5 × 103, and 1 × 103 HUCPV cells were seeded onto six-well tissue culture plates (Falcon no. 353046) and fed every 2 days with SM. The number of colonies, comprising >16 cells, were counted in each well on day 10 of culture and confirmed on day 14. CFU-F frequency, the average number of cells required to produce one colony, was consequently determined to be 1 CFU-F per 333HUCPVcellsplated. Based on this frequency, the unit volume required to provide 333 HUCPV cells (done in triplicate from each of three cords) was calculated, and eight incremental unit volumes of HUCPV cells were seeded into individual wells on six-well plates. Again, colonies comprising >16 cells (CFU-Fs) were counted on day 10 of culture to assay CFU-F frequency with incremental seeding.

Cell Proliferation Assay

To obtain the cell-proliferation growth curve, aliquots of 4 × 104 P2 HUCPV cells were plated into five six-well tissue culture polystyrene dishes. On days 1 through 5 of culture, one of the six-well plates was trypsinized, and the cells were counted. The total number of live cells was obtained at each time point by staining with 0.4% Trypan blue (Sigma no. T8154). Mean doubling time of the HUCPV was calculated using the obtained cell counts from day 1 through day 5, and the procedure was repeated with cells from three separate cords.

Doubling time of the HUCPV cells for passages 1–9 was determined by seeding 3 × 105 cells into T-75 flasks, which were fed with SM every 2 days, then trypsinized and counted using a hemocytometer (live cells were identified by Trypan blue [0.4%] exclusion) after 4 days. Mean doubling time was calculated from day 0 to day 4 for three separate cords.

Flow Cytometry

Test cell populations of >1 × 105 cells were washed in 2% FBS/PBS and suspended in 2% FBS/PBS with saturating concentrations (1:100 dilution) of the following conjugated mouse-anti-human antibodies: HLA-A, B, C-phycoerythrin (PE) (MHC I), HLA-DR, DP, DQ-fluorescein isothiocyanate (FITC) (MHC II), CD45-PE, CD34-PE, CD235a (Glycophorin A), CD90-PE (Thy-1), CD44-PE, CD106-FITC (VCAM-1), CD117-PE (c-kit), and CD123-PE (IL-3) (all BD Biosciences, San Jose, CA), and the following unconjugated antibodies: HLA-G, CD105 (SH2), CD73 (SH3), Oct3 (all BD Biosciences), and STRO-1 (hybridoma cell line secreting STRO-1 antibody was a kind gift from Dr. S. Gronthos, Adelaide, Australia) for 30 minutes at 4°C. Unconjugated primary antibodies were treated with a goat-anti-mouse-FITC–conjugated secondary antibody (BD Biosciences) for 20 minutes at 40°C after washing with 2% FBS/PBS. The cell suspensions were then washed twice with 2% FBS/PBS and resuspended in 2% FBS/PBS for flow cytometric analysis (XL; Beckman Coulter, Miami, using the ExpoADCXL4 software (Beckman Coulter). Positive staining was defined as the emission of a fluorescence signal that exceeded levels obtained by >99% of cells from the control population stained with matched isotype antibodies (FITC-conjugated and PE-conjugated mouse IgG1κ monoclonal isotype standards), which was confirmed by positive fluorescence of human BM samples. For each sample, at least 10,000 list mode events were collected. All plots were generated in EXPO 32 ADC Analysis software.

Serially passaged HUCPV cells (0.5–1 × 106) were also assayed for expression of MHC I and MHC II cell-surface antigens by flow cytometry. Additional aliquots of 1 × 106 serially passaged HUCPV cells were frozen using an isopropanol freezing container (Nalgene [Rochester, NY] cat. no. 5100 0001) and stored at −150°C for 1 week in a 90% FBS, 10% dimethyl sulfoxide (DMSO) solution (Sigma D-2650, lot no. 11K2320). After 1 week of cryopreservation, the HUCPV cells were thawed and analyzed (∼2.5 × 105 cells) by flow cytometry (see above), by gating on the live cell population and observing them for expression of MHC I and MHC II cell-surface antigens.

Bone Nodule Assay

At the weekly passage, aliquots of 4 × 103 cells per cm2 were plated on 35-mm tissue culture polystyrene dishes in osteogenic medium comprising SM with osteogenic supplements (OSs) (dexamethasone at 10−8M; Sigma no. D-8893), β-glycerophosphate (at 5 mM; Sigma no. G-9891), and L-ascorbic acid (at 50 μg/ml; Sigmano. A-4544)). Control cultures were maintained in SM without OS. Cultures were re-fed every 2 days for a period of 7 days. The cultures were maintained until BNs were observed (usually after 3–6 days), at which point the cultures were re-fed once with SM containing 9 μg/ml tetracycline (Sigma no. 7660) [42], then fixed after 24 hours in Karnovsky fixative (25% by volume 8% paraformaldehyde, 10% by volume 25% glutaraldehyde, 50% by volume 0.2 M cacodylate buffer, 15% by volume distilled H2O), and prepared for analysis by light microscopy, phase contrast microscopy, fluorescence microscopy, and scanning electron microscopy (SEM).


Figure 1A shows the SEM appearance of the perivascular WJ matrix which, by routine hematoxylin and eosin light microscopy (not shown), was seen to possess a relatively homogeneous distribution of cells. The harvested cells exhibited a morphologically homogeneous “fibroblast-like” appearance (Fig. 1B) with a stellate shape and long cytoplasmic processes extending between 100 and 300 μm. These cells labeled positively for α-actin, desmin, vimentin, and the 3G5 monoclonal antibody (not shown), but we found no evidence of NSE.

Figure Figure 1..

(A): Scanning electron microscopy of an umbilical artery that has been excised from a human umbilical cord as part of the HUCPV cell harvesting procedure. The white dotted line represents the outer margin of the vessel and thus illustrates the perivascular tissue from which the HUCPV cells are harvested. (B): HUCPV cells display a fibroblastic morphology (field width = 660 μm). Abbreviation: HUCPV, human umbilical cord perivascular.

CFU-F and CFU-O Expansion

The digestion procedure yielded an average of 2–5 × 106 HUCPV cells per UC (depending on the length of UC harvested, which can vary from 10–30 cm). Normalized to a unit length of cord, this represents a harvesting yield of 2.5–25 × 104 cells/cm of cord and a harvesting efficiency of 100% since every cord yielded cells (n = 72). Counting the number of cell colonies at passage 0 (P0) established a CFU-F frequency of 1:333 (±0.83). Seeding multiples of this number of cells demonstrated an increase in CFU-F frequency with increasing cell-seeding densities (Fig. 2), indicative of some paracrine signaling between HUCPV cells, which may potentiate CFU-F formation.

Figure Figure 2..

A nonlinear increase in CFU-F frequency is observed with serially increasing cell-seeding densities compared with the expected linear CFU-F frequency. This difference may indicate a paracrine signaling mechanism between human umbilical cord perivascular cells (n = 6). Error bars denote standard deviation. Abbreviation: CFU-F, colony-forming unit-fibroblast.

P0 through P7 HUCPV cells demonstrated a decreasing doubling time of 59.4 ± 42.4 hours (P0) to 19.71 ± 12.4 hours (P2), and this remained approximately constant until P8 (Fig. 3), by which time over 50 population doublings had already been achieved. The HUCPV cells demonstrated a growth curve with an initial lag phase (0–24 hours) and subsequent log phase (24–120 hours) (Fig. 3 insert). Figure 4 shows that from day 0 to the end of the second passage (30 days of culture) the number of HUCPV cells increased from 6.6 × 103 to 1.4 × 1010. Within this CFU-F population, frequencies of CFU-O were determined to be 2.6/105 CFU-F and 0.75/105 CFU-F in the absence of OSs, and 1.20/104 CFU-F and 1.29/104 CFU-F at P1 and P2, respectively, with the addition of OS. No BNs were found in P0 cultures in either osteogenic or nonosteogenic conditions. Thus, after 30 days of culture, 1.8 × 106 CFU-O cells were resident in the whole CFU-F population in OS conditions.

Figure Figure 3..

Doubling time of HUCPV cells with successive passaging, demonstrating increasing proliferation to a 20-hour doubling time from P2 to P7, and increase after P8 (n = 3). Insert: Proliferation of P2 HUCPV cells from 0–120 hours, illustrating a normal growth curve with a lag phase of 0–24 hours and a log phase of 24–120 hours (n = 3). Error bars denote standard deviation. Abbreviation: HUCPV, human umbilical cord perivascular.

Figure Figure 4..

Frequency of CFU-F and CFU-O cells with successive passaging of human umbilical cord perivascular cells in the presence and absence of OSs (n = 4). Error bars denote standard deviation. Abbreviation: CFU-F, colony-forming unit-fibroblast; CFU-O, colony-forming unit-osteogenic; OS, osteogenic supplement.

Bone Nodule Formation

Passaged HUCPV cells in the presence of OS demonstrated markers of osteogenic expression within 4–5 days of culture. Colonies of cells with high alkaline phosphatase (ALP) expression that was positive for mineralization with von Kossa staining were indicative of osteogenic differentiation. The colonies were characterized by an accumulation of fibroblast-like cells in direct contact with one another. The colonies expanded in size, to between 300 and 800 μm in diameter (Fig. 5A) and approximately 100 μm in height (Fig. 5B). The cells bordering the nodules (Fig. 5C) were of a fibroblastic morphology, while those toward the interior of a nodule were more polygonal. Ultraviolet fluorescence of the tetracycline-labeled nodules (Fig. 5D) illustrated the variation of mineralization associated with their structure. Mineralization appeared to be relatively heavy in the middle of the nodule, as seen by an intense fluorescence, while the periphery of the nodule had less fluorescence intensity.

Figure Figure 5..

Tetracycline-labeled bone nodules observed by (A) phase microscopy and (B) fluorescence microscopy (FW = 832 μm). (C): A similar nodule seen by scanning electron microscopy (FW = 590 μm). (D): A bone nodule sectioned horizontally (parallel to culture dish surface) and stained with Masson trichrome (FW = 720 μm). Note the cells surrounded by abundant collagenous extra-cellular matrix. Abbreviation: FW, field width.

Figure 5D illustrates a demineralized Masson trichrome-stained transverse section of a BN. The areas in blue represent the collagen that makes up the bulk of the BN in which were embedded round nucleated cells, putatively identified as osteocytes.

Flow Cytometric Analysis

All analyzed HUCPV cells labeled positively for CD105 (SH2), CD73 (SH3), CD90 (Thy-1), and CD44, but negatively for CD45, CD34, CD235a (glycophorin A), CD106 (VCAM1), CD123 (IL3), SSEA-4, HLA-DR, DP, DQ (MHC II), HLA-G, and Oct4 (Table 1). HUCPV cells did not label with the hybridoma-derived STRO-1 antibody, although the latter did label a 35% subpopulation of a human BM positive control. Subpopulations of HUCPV cells labeled positively for other cell-surface proteins, including 15% CD117 (c-kitlow) and 75% HLA-A, B, C (MHC Ilow).

Table Table 1.. Flow cytometry results of human umbilical cord perivascular cells labeled for several cell-surface and intracellular markers. Data gained from a total of 11 umbilical cords in which n ≥ 3.
  1. a

    Data refer to cells at P0 through P5, except for a (P3) and b (P1 through P5). Abbreviations: ++, =98%; –, =1.5%; MHC, major histocompatibility complex.

CD105 (SH2)++
CD73 (SH3)a++
CD90 (Thy1)++
CD117 (c-kit)15% +
MHC I75% +
CD106 (VCAM1)
D123 (IL-3)
CD235a (glycophorin A)b

Figure 6 illustrates the MHC I/II (MHC− /−)expression of serially passaged HUCPV cells and cryopreserved HUCPV cells. The input cell population contained 20.8% ± 3.1% which were MHC−/−. This subpopulation increased to 31.2% ± 1.7% at P5. Following cryopreservation, HUCPV cells demonstrated an increased MHC−/− population, rising from 65.2% ± 5.4% at P0 to 96.0% ± 3.9% at P5. Upon rapid thawing of the frozen aliquots of cells in a 37°C water bath, cell survival at P0 was 49.2% ± 23.8 (n = 12), while thawing of cells from P1 through P9 resulted in a survival of 62.6% ± 19.7 (n = 30).

Figure Figure 6..

Flow cytometry results of MHC−/− expression on HUCPV cells with serial passaging, and the change of MHC−/− expression with cryopreservation of serially passaged HUCPV cells, which reached 95% at P5 (n = 3). Error bars denote standard deviation. Abbreviations: HUCPV, human umbilical cord perivascular; MHC, major histocompatibility complex.


The method we describe depends on isolation of the UC vasculature and enzymatic digestion of its perivascular tissue to rapidly harvest a highly proliferative HUCPV population. This is a significantly different approach from previous reports in which the vasculature and its surrounding tissue have been discarded [31, 34, 35]. This distinction in the isolation procedure not only provided a cell harvest more rapidly than hitherto described but may also explain the lack of NSE-positive cells and the low number (15%) of c-kitlow-expressing cells in our cultures compared with those described by Mitchell et al. [35]. Indeed, contrary to the latter authors, we found no evidence of a neural phenotype even if we cultured the cells in neural inductive medium, consisting of low-serum media cultures with basic fibroblast growth factor, butylated hydroxyanisole, and DMSO (not shown). Recently, Kogler et al. [43] also reported the isolation of somatic stem cells from umbilical cord blood (UCB). However, they achieved a mesenchymal cell harvest in only 94 of 233 cord bloods (40.3%), while we harvest cells from every cord received (100%). Furthermore, the frequency of the MSC-like cells derived from UCB was 1:200 million, while the harvesting method reported herein results in a CFU-F frequency of 1:300. Therefore, we conclude that our harvesting procedure is considerably more consistent and yields a greater number of relevant cells than can be achieved from UCB.

Although neither pre- nor post-CD45 sorted isolates of HUCPV cells demonstrated CD45 expression, we nevertheless negatively sorted for the CD45 population to eliminate any possible contamination by hematopoietic precursors from UCB, and we tested the resultant population for a series of markers that are characteristic of embryonic and mesenchymal phenotypes (see below). Thus, immunohistochemistry demonstrated the presence of three specific cytoskeletal markers—α-actin, desmin, andvimentin—which correlates with the in situ characterization of WJ cells by Takechi et al. [40], Kobayashi et al. [44], and Kadner et al. [37, 38]. Furthermore, due to their reactivity with the 3G5 monoclonal antibody [45], HUCPV cells appear to be similar to another perivascular mesenchymal precursor, the pericyte [4649]. In addition, flow cytometry illustrated that HUCPV cells present several cell-surface antigens commonly found on BM-derived so-called MSCs. Although no STRO-1 expression was observed, the cells were SH2, CD44, and Thy-1 positive. Thy-1 is commonly associated with cells of hematopoietic origin, but we were careful to exclude hematopoietic contamination during harvesting. Thy-1 is also known to be expressed in connective tissue and various fibroblast and stromal cell lines [50], including multipotent adult progenitor cells (MAPCs) [51]. A small sub-population that expressed c-kitlow was also present, and this contrasts with MAPCs [51], which show no c-kit presence. As discussed above, Mitchell et al. [35] demonstrated “very high” expression of c-kit on cells extracted from WJ, which also expressed NSE even in uninduced culture conditions. In contrast, our HUCPV cells exhibited spontaneous BN formation in non-osteogenic culture conditions. These differences suggest that our harvesting protocol resulted in a cell population that is distinct from those described by both Mitchell et al. [35] and Kadner et al. [37, 38], who showed no differentiated phenotype other than the myofibroblast markers found in WJ cord cells in situ.

We found that the harvested HUCPV cell population was highly ALP positive (not shown) and, in addition to a subpopulation that can spontaneously elaborate BNs after P0, contains a subpopulation that may be induced to express an osteogenic phenotype and elaborate bone matrix in culture by the addition of dexamethasone. Notably, CFU-O frequency in the OS+ cultures was twice that of the OS cultures. Committed osteoprogenitors have been described as progenitor cells restricted to osteoblast development and bone formation [52]. Since we are unaware of any reported pathologies associated with mineralization of the UC, we suggest that it is the culture conditions—environment and manipulation—that are causing this restricted induction of these early osteoprogenitors. However, committed BM-derived populations have also been shown to give rise to both adipogenic [53] and chondrogenic [54] lineages; thus, we may reasonably expect, through further culture manipulation, to derive these, and other, mesenchymal phenotypes.

The frequency of 1 CFU-F per 333 HUCPV cells, shown by the limiting dilution assay, is significantly higher than that observed in neonatal BM, which has been shown to possess approximately 1 MSC per 10,000 BM stromal cells [55]. Our results show that these CFU-F–derived HUCPV cells proliferate rapidly in culture, demonstrating a changing doubling time during the first 30 days of culture of approximately 60, 30, and 20 hours for P0, P1, and P2, respectively (average 33.5 hours). In contrast, a 36-hour doubling time has been reported in ongoing cultures of human embryonic stem cells [56], and a longer, 60-hour, average doubling time can be calculated for the first 30 days of adult BM culture (from the 21- to 36-day data reported by Suva et al. [57]). The latter, 4-day doubling, corresponds to the report of Bruder et al. [58], who showed that, on average, MSCs achieved two population doublings for each 9-day culture from passages 1 through 10. Thus, HUCPV cells represent a population of cells that can be rapidly expanded for potential clinical applications.

The rapid doubling time of HUCPV cells raises the question of whether a therapeutic mesenchymal cell dose could be achieved more rapidly than from currently employed marrow sources. With an average infusion of 4.3 × 109 nucleated cells, Horwitz et al. [26] injected 1.7 × 105 MSCs (based on 1 MSC : 2.5 × 104 mononuclear BM cells [59, 60]) that successfully ameliorated the condition of three patients with OI. As a result, if approximately 2 × 105 MSCs are required for a therapeutic dose (TD), Figure 4 illustrates that a single such dose can be derived from HUCPV cells within 10 days of harvest—and 1,000 TDs after 24 days of culture expansion. This compares favorably with the expansion of MAPCs that, given the data of Reyes et al. [61], require 14 days to establish a culture containing approximately 104 cells with a doubling time of 48 hours, which would result in a single TD within 22 days—and 1,000 TDs after 42 days of culture expansion.

Furthermore, our data show an increase, with both passage and, particularly, freeze-thawing, of a HUCPV population that expresses neither class I nor class II MHC antigens (MHC−/−). Although the majority of cells at P0 were MHC I positive, the MHC−/− phenotype increased modestly from 20%–30% during the first five passages. Specifically, in the 15%–30% of the HUCPV cells that survived vitrification, the MHC−/− phenotype increased considerably to 65% at P0, 90% at P3, and 95% at P5. While these percentages of potentially allogeneic cells are unattainable in adhesion-dependent BM-derived cells, which retain class 1 expression [62], they also exceed those recently published for BM-derived cells expanded in noncontact suspension conditions [63].

Since Horwitz et al. [27] have shown that systemically infused marrow-derived mesenchymal populations have a clear clinical potential, our selective and rapid proliferation of MHC−/− HUCPV cells is of particular clinical relevance. Although some authors have found little evidence of an immunogenic response using allogeneic [64], and indeed xenogenic [65], MSC therapy, it has been noted that histocompatibility of the cell source is a significant hurdle to be addressed for safe and effective application of cell-based therapies [66]. In this context, an MHC−/− cell population as described herein may represent a promising avenue to surmount the potential hazards of alloreactive T cells or a host immune response leading to graft versus host disease.


Although the in vivo function of HUCPV cells still needs to be studied, we believe these cells represent a population of normal, rapidly expandable, MHC−/− cells that can potentially generate multiple therapeutic doses of cells for cell-based therapies, and thus they represent a significant alternative to BM in the treatment of pathologies associated with the connective tissues of the human body.


All work was carried out at the Institute of Biomaterials and Biomedical Engineering, University of Toronto. We thank Feryal Sarraf for preparation of histology samples, Elaine Cheng and Jane Ennis for the UC harvests and cell culture, Lorraine Hanoun for the cell proliferation data collection, and Cheryl Smith and Joanna Vergidis for assistance with the flow cytometry analysis. Our work was financially supported through an Ontario Research and Development Challenge Fund (ORDCF) grant to J.E.D.