SEARCH

SEARCH BY CITATION

Keywords:

  • Blood cells;
  • Cell culture;
  • Cell transplantation;
  • Chondrogenesis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

Using a variety of cell separation techniques and cultivation conditions, circulating, adherent, connective tissue, clonogenic cells were found in just 3 donors out of 66, demonstrating that these precursors are extremely rare in postnatal human blood. Contrary to humans, guinea pig blood shows much more reproducible connective tissue colony formation; it was therefore chosen to study the differentiation potential of adherent blood-derived clonogenic cells. Out of 22 single colony-derived strains of various morphologies, only 5 spindle-shaped strains showed extensive proliferative capacity in vitro. None of these strains formed bone upon in vivo transplantation, whereas two strains formed cartilage in high-density pellet cultures in vitro. Both chondrogenic strains included cells expressing aggrecan, whereas nonchondrogenic strains did not. Out of four polyclonal strains studied, one formed both cartilage and abundant bone accompanied by hematopoiesis-supporting stroma. Evidently, there are cells in adult guinea pig blood capable of both extensive proliferation and differentiation toward cartilage: circulating chondrogenic precursors. Although some of these cells lack osteogenic potential and therefore represent committed chondrogenic precursors, others may be multipotential and consequently belong to the family of skeletal stem cells. This is the first demonstration of postnatal circulating chondrogenic precursors, as well as of precursor cells with chondrogenic but not osteogenic potential.

Disclosure of potential conflicts of interest is found at the end of this article.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

Transformation of blood leukocytes into connective tissue cells belongs to the oldest and most controversial concepts in cell biology. One of the highly contentious issues within this concept is the interpretation of human peripheral blood cultures where macrophages and their derivatives can acquire notoriously diverse morphological and tinctorial phenotypes often displaying a transitory fibroblast-like morphology [1, 2]. Consequently, subsets of these cells have frequently been mistakenly identified as fibroblast-like cells [3, [4], [5], [6]7], thus incorrectly suggesting an abundance of adherent connective tissue progenitors in human blood. However, when rigorous criteria are applied, including precise immunophenotyping, the ability to form discrete, compact colonies [8] and to undergo multiple passages generating fibroblast-like strains [9], and the capacity for connective tissue formation upon in vivo transplantation in diffusion chambers [10, [11], [12]13] or for bone formation upon in vivo transplantation in an open system [14], true adherent connective tissue precursors are either totally absent [2, 15, [16]17] or very rare [14] in adult human blood. In the first part of this study, we used a variety of cell separation techniques and cultivation conditions to establish blood cell cultures from a significantly extended and diversified group of human donors. Despite all the variables, adherent colony formation was observed in cultures from just 3 donors out of 66, demonstrating that circulating connective tissue precursors are absent in the vast majority of human donors and extremely scarce in the remaining few.

Unlike humans, blood cells of several animal species, such as mouse, rabbit, and, in particular, guinea pig, show much more reproducible formation of adherent connective tissue colonies [8, 14, 18, 19]. In the second part of the study, the guinea pig model was chosen to further analyze the differentiation potential of adherent strains of blood-derived cells in a search for circulating progenitors of solid-phase connective tissues. It was shown that some single colony-derived and polyclonal strains of circulating guinea pig cells differentiated to form cartilage, thus demonstrating for the first time the existence of postnatal circulating chondrogenic precursors.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

Human Model

Human Blood Specimens.

Human blood specimens were obtained with informed consent per institutionally approved protocols. Peripheral blood from 66 donors was received either as whole blood (34 donors) or as one of the following blood products provided by the Blood Services Section, Division of Transfusion Medicine, Clinical Center, NIH: buffy coat, or total leukocyte fraction (28), leukapheresis product, or leukocytes minus granulocytes (3), monocyte fraction (1). Blood was received from 39 males and 27 females of the following age groups: 0–9 years old—7 (including one newborn, three children from 4–6 years old, and three children from 7–9 years old); 10–19 years old—7; 20–29 years old—7; 30–39 years old—10; 40–49 years old—14; 50–59 years old—11; 60–69 years old—5; 70 years old and older—5. Thirty-seven donors were healthy volunteers (including one who received granulocyte colony-stimulating factor [G-CSF] injections at 10 μg/kg for 5 days); from 19 patients, blood was taken within several hours after they had suffered fractures of one or more bones; four patients had fibrous dysplasia of bone or McCune-Albright syndrome (Online Mendelian Inheritance in Man [OMIM] #174800); one had fibrodysplasia ossificans progressiva (OMIM #135100); four had juvenile dermatomyositis with calcinosis; one had oncogenic osteomalacia. Most cultures were established within several hours after blood collection, with the exception of trauma patients, whose blood was sent overnight so that the cultures were prepared on the next day. The total number of nucleated cells ranged from 11.5 × 106 to 1,690 × 106 per blood specimen.

Blood Cell Preparation.

Blood cells were plated either without further separation or following additional separation steps. For the former approach, whole blood or blood products were diluted with Hanks' balanced saline solution (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) and pelleted at 1,200 rpm for 30 minutes, where the supernatant containing anticoagulant was discarded, and the cells were resuspended in culture medium. Alternatively, blood mononuclear cells were isolated by centrifugation over Ficoll-Paque (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) following the manufacturer's protocol. Otherwise, whole blood or blood products were allowed to undergo sedimentation for 3 hours at room temperature or overnight at +4°C, after which plasma and leukocyte layers were carefully aspirated and cells were pelleted. Finally, for some samples, erythrocyte lysis was conducted by mixing up to 1 × 108 pelleted nucleated cells with 5 ml of ammonium chloride buffer: 144 mM ammonium chloride (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), 17 mM Tris (MP Biomedicals, Aurora, OH, http://www.mpbio.com), pH 7.2, for 5–10 minutes at 37°C with slow rotation, followed by adding equal volume of α-minimal essential medium (αMEM) (Invitrogen) and centrifugation.

Blood Cell Cultures.

Blood cells were plated at 0.1–6.5 × 106 nucleated cells per cm2 into T-75, T-162, or T-175 flasks or 60-mm, 100-mm, or 150-mm dishes (BD Labware, Lincoln Park, NJ, http://www.bd.com; Corning Inc., Corning, NY, http://www.corning.com; Nalge Nunc International, Naperville, IL, http://www.nalgenunc.com). Alongside regular tissue culture vessels, those with modified surfaces (Primaria; BD Labware), or those that were industrially precoated with either human fibronectin or rat collagen type I (BD Labware), were used. Alternatively, the vessels were coated with either 0.1% porcine gelatin (Sigma) or 50 μg/ml human fibronectin (Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com) for 1–5 hours before cell plating. Most cultures were prepared in the medium consisting of αMEM, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin sulfate (Invitrogen), and 20% fetal bovine serum (FBS; Equitech-Bio Inc., Kerrville, TX, http://www.equitech-bio.com) (medium 1), either with or without 10−8 M dexamethasone (Sigma), 10−4 M l-ascorbic acid phosphate magnesium salt n-hydrate (AscP; Wako Chemical, Osaka, Japan, http://www.wako-chem.co.jp/english), 10−4 M 2-mercapthoethanol (Invitrogen), and 10% human serum (Pel-Freez, Rogers, AK, http://www.invitrogen.com). For additional cultures, several other medium formulations were used: medium 2: X-vivo 15, a serum-free medium for lymphocyte growth (Cambrex, Walkersville, MD, http://www.cambrex.com); medium 3 (medium designed for embryonic stem cells): Dulbecco's modified Eagle's medium (DMEM) (Invitrogen), l-glutamine, penicillin, streptomycin sulfate, 0.1 M nonessential amino acids (Invitrogen), 10−4 M 2-mercapthoethanol, 1,000 U/ml recombinant human leukemia inhibitory factor (R&D Systems Inc., Minneapolis, http://www.rndsystems.com), 20% FBS; medium 4 (medium formulated for neuronal cells): neurobasal medium (Invitrogen), l-glutamine, penicillin, streptomycin sulfate, 20% FBS, supplement N-2 or G-5 (Invitrogen); medium 5 (medium designed for human “mesodermal progenitor cells” [20]): 58% low glucose DMEM, 40% MCDB 201 (Sigma), 2% FBS, 1% ITS+ (BD Biosciences, San Diego, http://www.bdbiosciences.com), penicillin, streptomycin sulfate, 10−9 M dexamethasone, 10−4 M AscP, 10 ng/ml epidermal growth factor, and 10 ng/ml platelet-derived growth factor-BB (Sigma). The media were replaced 1–3 days after plating and once per week thereafter. To some cultures, guinea pig bone marrow cells irradiated with 6,000 R were added as feeder after the first medium replacement at 2.8–3 × 107 nucleated cells per T-75 flask [21]. The cultures were thoroughly inspected under the inverted microscope twice per week. If no adherent connective tissue colonies were found, the cultures were followed for 9–25 weeks, fixed with absolute methanol, stained with saturated water solution of methyl violet (Sigma), and yet again assessed for adherent colony formation.

Guinea Pig Model

All procedures were performed in accordance with a National Institute of Dental and Craniofacial Research approved small animal protocol (number 97-024).

Blood Collection.

The blood from adult 5–8-month-old Hartley male guinea pigs was collected by either cardiac puncture or cannulation of either inferior vena cava or aortic root. The first portion (approximately 0.5 ml) of blood obtained by cannulation was discarded. The procedures were performed under the anesthesia achieved by intraperitoneal injection of ketamine (Fort Dodge Animal Health, Fort Dodge, IA, http://www.wyeth.com/divisions/fort_dodge.asp) at 140 mg/kg body weight and xylazine (Phoenix Scientific Inc., St. Joseph, MO, http://www.psiqv.com) at 7 mg/kg body weight. Blood was collected into sterile syringes or tubes containing sodium heparin (Fisher Scientific International, Hampton, NH, http://www.fisherscientific.com) at a final concentration of 100 U/ml.

Blood Cell Cultures.

Freshly obtained whole blood was plated into T-75 flasks or 150-mm dishes (BD Labware) at 0.5–2.2 × 105 nucleated cells per cm2 in culture medium consisting of αMEM, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin sulfate, 20% FBS, 10−8 M dexamethasone, and 10−4 M AscP. Cells were cultured at 37°C in an atmosphere of 100% humidity and 5% CO2. Complete medium replacement was performed one day after plating, at which stage most erythrocytes were removed, and twice per week thereafter. Adherent colonies were released on days 14–18 with two consecutive treatments with 1× trypsin-EDTA (Invitrogen). Polyclonal strains were generated by detaching all colonies simultaneously and plating them together into a T-75 flask. Single colony-derived strains were generated in a manner previously described [22]. Each individual colony was separately detached using a polystyrene cloning cylinder (Sigma) and S/P High Vacuum Grease (Baxter Healthcare Corp., McGaw Park, IL, http://www.baxter.com). Cells were then plated into a single well of a 6-well plate (Nunc, Rochester, NY, http://www.nuncbrand.com). Subsequent passages were performed at 70%–90% confluence.

Cultures of Guinea Pig Bone Marrow Stromal Cells.

Bone marrow single cell suspensions were prepared from femora, tibiae, and humeri of Hartley male guinea pigs. Polyclonal bone marrow stromal cell (BMSC) strains were generated as described earlier [21] and used as positive controls in differentiation experiments.

Phenotypic Characterization.

Polyclonal and single colony-derived strains of blood-derived adherent cells and polyclonal strains of BMSCs were plated in eight-chamber glass slides (Nalge Nunc International) at 0.12 × 105 cells per cm2. Histochemical stainings were performed by standard protocols for α-naphthyl acetate esterase, acid phosphatase, and alkaline phosphatase (kits 91-A, 387-A, and 86-C, respectively; Sigma). Immunohistochemical analysis was performed when the cells reached 70% confluence, usually on the next day after plating, with a panel of antibodies (Table 1) as described previously [23]. Bound antibody was localized by consecutive application of peroxidase-conjugated Broad Spectrum Poly HRP Conjugate and AEC (red) Single Solution (Zymed Labs Inc., San Francisco, CA, http://www.zymed.com). Slides were then counterstained with hematoxylin and mounted in Histomount (Zymed Labs). Slides were viewed using an Axioplan 2 (Carl Zeiss, Jena, Germany, http://www.zeiss.com), and images were acquired using a DMC-1 digital camera (Polaroid, Cambridge, MA, http://www.polaroid.com) and formatted using Photoshop 6.0 Software (Adobe Systems Inc., San Jose, CA, http://www.adobe.com).

Table Table 1.. Immunophenotypic characterization of polyclonal and single colony-derived strains of guinea pig blood-derived adherent cells and of polyclonal strains of guinea pig bone marrow stromal cells
Thumbnail image of
In Vivo Transplantation Assay.

Osteogenic potential of cell strains was analyzed by in vivo transplantation assay. Four polyclonal and five single colony-derived strains of guinea pig blood-derived adherent cells (passages 2–6 and 4, respectively), as well as several polyclonal strains of guinea pig BMSCs (passages 2–4), were transplanted as previously described [22, 24]. Briefly, 1.5–3.5 × 106 cells in 1 ml of culture medium were mixed with 40 mg of sterile hydroxyapatite/tricalcium phosphate ceramic powder (particle size 0.5–1.0 mm; generously provided by Zimmer Inc., Warsaw, IN, http://www.zimmer.com) in a 1.8-ml cryotube (Nunc). After a 90-minute incubation at 37°C with slow rotation (25 rpm), the particles with attached cells were collected by brief centrifugation. A secondary matrix was formed by consecutively mixing the particles with 15 μl of mouse fibrinogen (3.2 mg/ml in phosphate-buffered saline) and 15 μl of mouse thrombin (25 U/ml in 2% CaCl2; Sigma) such that a fibrin gel was formed around the particles. The transplants were placed in subcutaneous pockets in the back of 8–15-week-old immunodeficient female beige mice (bg-ν/ν-xid; Harlan Sprague Dawley, Indianapolis, http://www.harlan.com). Three or more transplants per strain were generated, and each experiment was repeated at least twice. The transplants were recovered 8 weeks later, fixed in 4% phosphate-buffered formalin (Sigma), and either embedded in glycol methacrylate or demineralized in 10% EDTA, pH.8.0 (Quality Biological, Gaithersburg, MD, http://www.qualitybiological.com), and embedded in paraffin; 5- or 7-μm sections were prepared and either stained with hematoxylin and eosin (paraffin sections) or Goldner's modified trichrome stain (glycol methacrylate).

In Vitro Pellet Culture Assay.

Chondrogenic potential of four polyclonal and five single colony-derived strains of guinea pig blood-derived adherent cells (passages 3–5 and 3, respectively) and of polyclonal strains of guinea pig BMSCs (passage 3) was analyzed using a technique modified from an earlier report [25]. Aliquots of 4 × 105 cells were pelleted in 15 ml of polypropylene conical tubes (Corning) and cultured at 37°C in an atmosphere of 100% humidity and 5% CO2. Culture medium consisted of DMEM with high glucose (Invitrogen), 1% ITS+, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin sulfate, 10−7 M dexamethasone, 50 μg/ml AscP, 2 mM pyruvate (Invitrogen), and 10 ng/ml human recombinant transforming growth factor-β1 (R&D Systems). Medium was replaced every other day. Three or more cultures per strain were generated, and each experiment was repeated at least twice. Cultures were harvested on day 21, fixed in 4% phosphate-buffered formalin, demineralized in 10% EDTA, and embedded in paraffin. Seven-μm sections were cut near the middle of the pellets and stained with toluidine blue or immunohistochemically for collagen type II. For the latter staining, HistoMouse-SP Bulk Kit (Zymed Labs) was used following the manufacturer's recommendations. To facilitate antibody access, sections were predigested with 0.1 U/ml chondroitinase ABC (Seikagaku, Tokyo, http://www.seikagaku.co.jp/english) in Tris-acetate buffer followed by pepsin (Digest-All Kit; Zymed Labs). Primary monoclonal antibody against collagen type II (clone 6B3; Chemicon, Temecula, CA, http://www.chemicon.com) was followed consecutively by a biotinylated secondary antibody, streptavidin-peroxidase, and hydrogen peroxide/chromogen solution (Zymed Labs).

Statistical Analysis

One-way analysis of variance was performed and post-test comparison was done using the Bonferroni multiple comparison test (InStat; GraphPad Software Inc., San Diego, http://www.graphpad.com). Differences were considered statistically significant at p < .05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

Human Cultures

In cultures of human blood, colonies and fields of adherent polymorphic cells often developed that included a broad array of morphologies, ranging from small round cells and slender spindle-shaped cells (“fibroblastoid cells”) to large multinuclear cells; these diverse morphological types are shown in Figure 1A–1C. The abundance of these cells varied from one donor to another and was strongly influenced by the medium composition, with the mixture of dexamethasone and AscP inhibiting their appearance. These cells could persist in primary cultures for many months (Fig. 1C), but they neither reached confluency nor survived beyond the first passage. In previous studies, such cells were shown to exhibit hematopoietic markers, to form osteoclast-like cells, and to represent monocytes/macrophages [26, [27]28]. Contrary to this pattern, adherent colonies consisting of uniform fibroblast-like cells appeared in only four cultures (Fig. 1D, 1E). The colonies grew rapidly at the periphery and formed dense central areas. Using a diversified group of blood donors, a number of cell separation techniques, several culture substrata, and a variety of culture media, only 3 donors were identified among 66 studied whose blood yielded one or two adherent connective tissue colony-forming cells; blood from the remaining 63 donors, or 95.5%, demonstrated no such cells. One positive donor (HBl-2) was a healthy 60-year-old male whose blood cells were received as a buffy coat; this donor yielded two colonies, one in a regular flask and another in a fibronectin-coated flask [14]. Another donor (HBl-40) was a healthy 28-year-old male who had undergone G-CSF-induced stem cell mobilization, and his blood cells were received as a leukapheresis product; a single colony was formed in a fibronectin-coated flask (Fig. 1D). The third donor (HBl-53) was a 47-year-old female with multiple bone fractures whose whole blood was received; a single colony was formed in a fibronectin-covered flask (Fig. 1E). All four colonies were formed in cultures of cells plated in medium 1 with dexamethasone and AscP without further cell separation. No adherent connective tissue colonies were formed in cultures from younger donors or from donors suffering from several pathological conditions featuring bone and connective tissue involvement.

thumbnail image

Figure Figure 1.. Adherent cell growth in primary cultures of human peripheral blood cells. (A–C): Nonfibroblastic cells (macrophages and their derivatives). (D–E): Colonies of fibroblast-like connective tissue cells. (A, B, D, E): Live cultures. (C): Fixed culture stained with methyl violet.

Download figure to PowerPoint

From each of the four colonies, a single colony-derived strain of adherent connective tissue cells was subsequently generated by multiple passages. The strains derived from donor HBl-2 were characterized earlier; both of them displayed immunophenotype characteristics of fibroblastic/smooth muscle/weakly osteogenic cells and were capable of adipocytic conversion in vitro, whereas only one of two strains formed bone upon in vivo transplantation [14]. We were unable to characterize chondrogenic potential of human single colony-derived strains of circulating adherent connective tissue cells because sufficient numbers of these cells could only be received by passage 3, and by this passage even polyclonal strains of human BMSCs (positive controls) largely lost their ability to form cartilage in preliminary pellet culture experiments.

Guinea Pig Cultures

Discrete colonies of adherent fibroblast-like cells were reproducibly formed in cultures of guinea pig blood from all animals examined, with no exception. The colony-forming efficiency per 1 × 106 nucleated cells was 1.1–3.9 (mean ± SEM, 2.7 ± 0.6) for blood obtained by cardiac puncture and 2.6–8.6 (5.3 ± 1.0) for blood obtained by cannulation (p > .05). By combining all colonies grown in a flask, four polyclonal (multicolony-derived) strains were established from four guinea pigs. When individual colonies were detached and propagated separately, 22 clonal (single colony-derived) strains were generated from three other animals.

Characterization of Polyclonal Strains

Each of the polyclonal strains was capable of prolonged proliferation through six or more passages and generated millions of cells. The strains were comprised of uniform spindle-shaped fibroblastic cells (Fig. 2A), staining variably (from negative to strongly positive) for α-naphthyl acetate esterase (Fig. 2B), positive for acid phosphatase (Fig. 2C), and negative for alkaline phosphatase (not shown). Immunophenotypically, all four strains were similar to each other and to guinea pig BMSCs: variably positive for markers associated with fibroblasts (fibronectin, collagens type I and III), osteogenic cells (osteonectin, osteopontin, bone sialoprotein), and smooth muscle cells and negative for markers of hematopoietic cells, macrophages, and endothelial cells (Table 1).

thumbnail image

Figure Figure 2.. Characterization of polyclonal strains of guinea pig blood-derived adherent cells. (A): Live confluent culture, unstained. (B, C): Cultures stained for α-naphthyl acetate esterase (B) and acid phosphatase (C). (D–J): Eight-week-old transplants of polyclonal strains. New bone was scant in the transplants of strain 1 (D), abundant in the transplants of strain 4 (G), and absent in the transplants of strains 2 (E) and 3 (F), where just fibrous tissue could be observed surrounding ceramic particles. (H–J): Higher magnification of the transplants of strain 4; (H) and (I) represent the same area viewed under conventional (H) and polarized light (I). In (I), collagen bundles were organized in parallel patterns consistent with lamellar-like structure of the new bone. (D–I): Demineralized sections stained with hematoxylin and eosin. (J): Undemineralized section stained with Goldner's trichrome. Osteoblasts (yellow arrow) deposited unmineralized osteoid (red arrow, red) that, upon mineralization, became bone matrix (green). Abundant bone was accompanied by hematopoietic marrow that included adipocytes and megakaryocytes (blue arrow). Abbreviations: a, adipocyte; b, bone; c, ceramic particles; h, hematopoietic marrow; ft, fibrous tissue.

Download figure to PowerPoint

In Vivo Transplantation Assay

Eight weeks post-transplantation, polyclonal strain 1 formed very scant bone limited to a few scattered trabeculae (Fig. 2D), whereas polyclonal strains 2 and 3 formed fibrous tissue surrounding the ceramic particles with no evidence of bone (Fig. 2E, 2F). Polyclonal strain 4 formed abundant bone (Fig. 2G) comparable in volume with bone developed in the transplants of guinea pig BMSCs (not shown). Bone formed by strain 4 had lamellar-like, well mineralized structure, was deposited on the surfaces of ceramic particles by mature osteoblasts, and contained embedded osteocytes (Fig. 2H–2J). In transplants with large amounts of bone, adipocytes and hematopoietically active marrow including megakaryocytes could be observed in close proximity to the bone surfaces (Fig. 2J). Earlier, it was demonstrated by in situ hybridization using guinea pig-specific DNA sequence that bone formed in such transplants was of donor origin [14].

Pellet Culture Assay

By day 21, pellet cultures of polyclonal strain 4 dramatically increased in size and developed chondroid morphology with cells encased in chondrocytic lacunae that were intermittently aligned in columnar patterns. The lacunae were surrounded by substantial amounts of matrix that demonstrated robust metachromatic staining, indicating the abundance of sulfated proteoglycans, and strong uniform deposition of collagen type II protein (Fig. 3D, 3E). Cultures of polyclonal strains 1, 2, and 3 showed neither a size increase, metachromasia, nor chondroid morphology (Fig. 3A–3C). Under identical conditions, guinea pig BMSCs developed cultures similar to those of strain 4 (Fig. 3F). These findings indicate that polyclonal strain 4 and BMSCs, but not the three other polyclonal strains of guinea pig blood-derived adherent cells, demonstrated the capacity to undergo chondrogenic differentiation in vitro.

thumbnail image

Figure Figure 3.. Twenty-one-day-old pellet cultures of four polyclonal strains of guinea pig blood-derived adherent cells and of a polyclonal strain of guinea pig bone marrow stromal cells (BMSCs). Cultures of strains 1 (A), 2 (B), and 3 (C) showed neither size increase nor metachromatic staining. To the contrary, cultures of strain 4 (D, E) and of BMSCs (F) demonstrated drastic increase in size, strong metachromasia with toluidine blue (D, F), cartilage-like morphology, and positive staining for collagen type II (E). (A–D, F): Staining with toluidine blue. (E): Immunostaining for collagen type II.

Download figure to PowerPoint

Isolation of Single Colony-Derived Strains

Adherent colonies formed in cultures of guinea pig blood were of three morphological types: spindle-shaped, fibroblast-like; flattened, fibroblast-like; and polygonal (Fig. 4A–4C). The proportion of colonies of each type varied in cultures from individual animals where spindle-shaped colonies always constituted the most prevalent morphology. In this study, 30 large colonies from three guinea pigs, representing all morphological types, were chosen for further propagation; 22 of these colonies exhibited proliferation in the first passage cultures: 6 spindle-shaped, 4 flattened, 11 polygonal, and 1 intermediate. However, only nine single colony-derived strains from two animals continued proliferation and reached a cell density sufficient to perform the second passage: five spindle-shaped and four polygonal. At this passage, each strain generated from 0.15 × 106 to 0.86 × 106 cells. Only five of the nine strains had a high enough proliferation potential necessary to attain near confluency in the second passage cultures, completing over 20 population doublings and generating several million cells required for two differentiation assays and for immunostaining. All of these five strains were of the spindle-shaped type: strain 10 (4.0 × 106 cells at the third passage) and strain 11 (17.6 × 106), both from arterial blood of one animal; strain 21 (9.5 × 106) and strain 23 (3.3 × 106), both from venous blood of the same animal; strain 26 (2.7 × 106), from venous blood of another guinea pig.

thumbnail image

Figure Figure 4.. Morphological types of adherent colonies in cultures of guinea pig blood and chondrogenic differentiation of single colony-derived strains of guinea pig circulating adherent cells. (A–C): Three morphological types of adherent colonies in cultures of guinea pig blood: spindle-shaped, fibroblast-like (A); flattened, fibroblast-like (B); and polygonal (C); live cultures. (D–H): Twenty-one-day-old pellet cultures of five single colony-derived strains of guinea pig circulating adherent cells. Cultures of strains 10 (D) and 11 (E) demonstrated chondrogenic differentiation, whereas cultures of strains 21 (F), 23 (G), and 26 (H) did not. Staining with toluidine blue.

Download figure to PowerPoint

In Vivo Transplantation and Pellet Culture Assays of Single Colony-Derived Strains

Contrary to two polyclonal strains, none of the five single colony-derived strains formed any bone 8 weeks after in vivo transplantation. The transplants of all strains uniformly revealed fibrous tissue adjacent to ceramic particles (not shown). After 21 days in pellet cultures, two strains, 10 and 11, demonstrated chondrogenesis (Fig. 4D, 4E) similar to that observed in cultures of polyclonal strain 4 and of guinea pig BMSCs. Conversely, three other strains, 21, 23, and 26, showed no signs of chondrogenic differentiation (Fig. 4F–4H).

Characterization of Single Colony-Derived Strains

Immunophenotypically, the five strains shared with BMSCs variably positive staining for markers associated with fibroblasts (fibronectin, collagens type I and III), osteogenic cells (osteonectin, osteopontin), and smooth muscle cells (α-smooth muscle actin) (Table 1; Fig. 5A–5F). Interestingly, for most of the markers, the degree of variability between the single colony-derived strains was higher than between polyclonal strains of blood-derived cells (Table 1). Although all five single colony-derived strains were negative for collagen type II (Table 1), two chondrogenic strains, 10 and 11, contained low numbers of cells strongly to moderately positive for another chondrocytic marker, aggrecan (Fig. 5G, 5H; Table 1). In this respect, these two blood-derived strains were similar to guinea pig BMSCs (Fig. 5L) and opposite to three nonchondrogenic blood-derived strains, 21, 23, and 26, in which aggrecan-positive cells were totally absent (Fig. 5I–5K; Table 1).

thumbnail image

Figure Figure 5.. Immunophenotype of single colony-derived strains of guinea pig circulating adherent cells. (A–F): Staining for osteonectin (A), osteopontin (B), collagen type I (C), collagen type III (D), fibronectin (E), and α-smooth muscle actin (F). (G–L): Staining for aggrecan: whereas strain 10 (G), strain 11 (H), and guinea pig bone marrow stromal cells (L) contained a minority of strongly and moderately stained cells, strains 21, 23, and 26 (I–K) were totally negative. Counterstaining with hematoxylin; the 50-μm scale bar applies to all pictures. Abbreviations: Col, collagen; FN, fibronectin; ON, osteonectin; OP, osteopontin; SMA, α-smooth muscle actin.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

Adherent connective tissue clonogenic cells can be found reproducibly and at a relatively high concentration in blood from every adult guinea pig. In this study, starting with individual adherent colonies, both polyclonal and single colony-derived strains of guinea pig circulating cells were generated and analyzed. Out of five single colony-derived strains with high proliferation potential, none displayed osteogenic differentiation, and two strains demonstrated chondrogenic differentiation confirmed by established morphological and biochemical criteria. The circulating clonogenic cells that initiated the two strains were therefore capable of both extensive proliferation and differentiation toward cartilage but not toward bone. These findings may be compared with the differentiation potential previously described for BMSC clones. When mouse and rabbit clones were transplanted in vivo within diffusion chambers, approximately one-third formed bone and cartilage, whereas the remaining clones only formed fibrous tissue [29, 30]. By in vitro assay, a number of human clones were restricted to osteogenic differentiation, where most clones exhibited bipotential (osteo-adipo or osteo-chondro), with a smaller proportion demonstrating a capacity for multipotential (osteo-chondro-adipo) [31, 32]. Thus, no pure chondrogenic clones have ever been reported, either in vivo or in vitro, among BMSCs from three different species where all clones capable of chondrogenic differentiation were also found to undergo osteogenic differentiation. Osteogenesis has been considered a default pathway for multipotential BMSCs, all of which are reversibly committed to osteogenesis and yet capable of entering other differentiation pathways [33, 34]. Our findings represent, for the first time, the existence of precursor cells with chondrogenic but not osteogenic potential. Apparently, in contrast to BMSCs, at least some guinea pig blood-derived, clonogenic, adherent cells are chondrogenic but lack osteogenic potential. Therefore, within the scope of the conventional (orthodox) BMSC differentiation repertoire [34], these cells can be considered committed chondrogenic precursors.

Out of four polyclonal strains of guinea pig circulating adherent cells, one strain proved to be both osteogenic and chondrogenic. The extent of both bone and cartilage formed by the strain was comparable with that achieved by guinea pig BMSCs under similar conditions. Furthermore, this strain demonstrated the capacity to develop into myelosupportive reticular stroma with associated adipocytes within the transplants. These results imply that some highly proliferative guinea pig blood-derived adherent strains may contain multipotential skeletal stem cells with the ability for generating at least four stromal cell lineages: osteoblasts/osteocytes, chondrocytes, adipocytes, and reticular cells. Although it cannot be excluded that polyclonal strains with broad differentiation potential are mixtures of several strains of committed precursors, it seems more likely that such strains indeed represent progeny of circulating multipotential skeletal stem cells.

The four polyclonal strains of guinea pig blood-derived adherent cells invariably consisted of uniform, spindle-shaped, fibroblastic cells. Immunophenotypically, the strains were similar to each other and to guinea pig BMSCs, all sharing a fibroblastic/osteogenic/smooth muscle, nonhematopoietic, nonendothelial signature. The expression of osteogenic markers, such as alkaline phosphatase, osteonectin, osteopontin, and bone sialoprotein, was similar in all polyclonal strains studied and not predictive of their in vivo osteogenic potential. Earlier, analysis of monolayer cultures of polyclonal strains of human BMSCs also revealed that the levels of procollagen type I, alkaline phosphatase, and osteopontin were not predictive of osteogenic potential of the strains [35], whereas the levels of osteopontin and bone sialoprotein mRNA could predict the ability to form bone in vivo [36].

In contrast to the uniform morphology of polyclonal strains of guinea pig circulating cells, individual blood-derived adherent colonies demonstrated three distinct morphological types, consisting of either spindle-shaped, flattened, or polygonal cells. Previously, adherent colonies derived from mouse and rabbit blood were also found to display three morphological types analogous to those described here for guinea pig cultures [14]. Despite pronounced morphological differences, all colony types from all species examined shared similar immunostaining patterns: positive for markers associated with fibroblastic, osteogenic, and smooth muscle cells and negative for endothelial, hematopoietic, and monocyte/macrophage markers [14]. In this study, among guinea pig colonies of various morphologies, only certain colonies composed of spindle-shaped cells were able to proliferate extensively. The proliferative advantage of spindle-shaped colonies may explain why polyclonal strains generated by combining and expanding colonies of all types uniformly displayed spindle-shaped morphology. Immunophenotypically, single colony-derived strains of spindle-shaped cells demonstrated a higher degree of variability than polyclonal strains. Most importantly, the two single colony-derived strains possessing chondrogenic potential, similarly to highly chondrogenic guinea pig BMSCs, included cells strongly to moderately positive for aggrecan, whereas three nonchondrogenic strains had no such cells. Likewise, chondrogenic polyclonal strains of human BMSCs were shown to constitutively express the aggrecan message [37]. Aggrecan is the major cartilage proteoglycan, deposition of which is considered a hallmark of chondrogenesis [38]. It may be hypothesized that under nonchondrogenic conditions, a small fraction of chondrogenic cell populations expresses aggrecan constitutively, and thus the expression of aggrecan in monolayer cultures may be predictive of chondrogenic potential.

In this study, the differentiation potential of adherent strains of circulating guinea pig cells was examined using in vivo transplantation assay and in vitro pellet culture assay. The use of transplantation assays under defined experimental conditions [22, 24, 39, 40] has become a valuable standard for delineating cells' osteogenic potential: not only does it analyze the osteogenic function in vivo, it also provides the most convincing result, the formation of histologically proven bone tissue [34]. The pellet (micromass) culture provides an assay of similar significance for chondrogenesis: culturing cells in artificially condensed conditions and inducing them with a chondrogenic cytokine results in the generation of a three-dimensional structure reminiscent of true hyaline cartilage [25, 41]. These two assays are more physiologically relevant and provide more convincing results than widely used monolayer culture differentiation assays.

When studying connective tissue transformation of blood cells, the major concern has been the possibility of contaminating blood samples with cells from the blood vessel or heart walls [10, 13, 42, [43], [44]45]. An important, although not definitive, result was demonstration that the concentration of adherent colony-forming cells in guinea pig blood was independent of the number of cardiac punctures used to obtain the blood [46, 47]. The cannulation technique has been widely accepted as the safest way of avoiding contamination, even more so when accompanied by discarding the first portion of blood [13, 42, 45]. In the present study, all recommended precautions were taken, and yet the concentration of adherent colony-forming cells in guinea pig blood obtained by cannulation was not lower than in blood obtained by cardiac puncture. Our result lends additional support to the view that adherent colony-forming cells do not come from the walls of either heart or blood vessels but are genuinely present in circulation.

Contrary to guinea pigs, our data demonstrate that circulating adherent connective tissue precursors are totally absent in the great majority of human donors, adults, adolescents, and children likewise, normal as well as pathological, and are extremely scarce in a few donors, in agreement with earlier observations [2, 14, [15], [16]17]. The finding of such precursors in only 3 donors out of 66 may suggest that, in postnatal humans, these cells enter the circulation only in very rare, special circumstances, such as following bone fractures or G-CSF-mediated hematopoietic stem cell mobilization. Contrary to this, circulating adherent connective tissue precursors, some of them demonstrating multilineage differentiation potential, were shown to be more prevalent in preterm fetuses and in full-term umbilical cord blood [26, 28, 48, [49], [50], [51], [52], [53], [54], [55], [56]57], although the latter finding has been challenged [27, 28, 58, 59]. Additionally, circulating nonadherent osteogenic precursors were recently found in postnatal humans [60, 61]. If these results are further substantiated, with both rigorous characterization of the osteogenic cells and corroboration of donor origin of bone formed upon their in vivo transplantation, the relationship between circulating adherent and nonadherent connective tissue precursors will be of significant interest.

Our results are the first indication that cells with chondrogenic potential can circulate in postnatal animals. Although the origin of circulating chondrogenic precursors, their destination in the bloodstream, and their functional role remain unknown, the same holds true for circulating connective tissue precursors in general. It was suggested that the fate of circulating connective tissue precursors may be defined on their arrival to a local tissue by its specific cues [62], implying that all such precursors are multipotential. According to our data, at least some blood-derived connective tissue precursors are already committed while in circulation. If such circulating chondrogenic precursors can home to cartilage, they might participate in osteochondral regeneration contributing to the high ability of articular cartilage of some mammals to self-repair.

Previous studies have demonstrated that the concentration of adherent colony-forming cells varies in different parts of guinea pig circulation, where it appears to be highest in the inferior vena cava, intermediate in heart, and lowest in renal vein [8]. In our study, chondrogenic and nonchondrogenic clonogenic precursors were found in arterial and venous guinea pig blood, respectively. The limited number of analyzed strains precludes us from drawing precise conclusions based on this observation. However, since human blood is generally drawn from veins, this could diminish even further the chances, already very low, of finding chondrogenic precursor cells in human blood. Nevertheless, the search for these elusive cells in postnatal human circulation needs to be continued in order to determine their role in physiological chondrogenesis and their possible application for the treatment of osteoarthritis.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

This study demonstrates that adherent, clonogenic, connective tissue cells can be found in blood from only very few postnatal human donors, and even in them with extremely low frequency. In adult guinea pig blood where such cells are much more common, both committed chondrogenic precursors and multipotential skeletal stem cells have been identified and characterized. These data represent a step toward our understanding of circulating skeletal precursors and, more generally, of circulating connective tissue precursors.

Disclosures of Potential Conflicts of Interest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

The authors indicate no potential conflicts of interest.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures of Potential Conflicts of Interest
  9. Acknowledgements
  10. References

The authors are indebted to Zimmer, Inc., for its gift of hydroxyapatite/tricalcium phosphate ceramic powder, to Dr. Larry W. Fisher for his gift of antibodies, and to Drs. Susan F. Leitman, Elizabeth J. Read, Michael T. Collins, Eileen M. Shore, and Lisa G. Rider for providing us with human blood. This research was supported by the Intramural Research Program of the National Institute of Dental and Craniofacial Research, NIH, Department of Health and Human Services. M.H.M. is currently affiliated with the Department of Surgery, University of California, San Francisco; A.I.L. is currently affiliated with the Department of Orthopedic Surgery, Johns Hopkins University, Baltimore; N.Z. is currently affiliated with the Center for Musculoskeletal Research, University of Rochester, School of Medicine and Dentistry, Rochester, New York; S.G. is currently affiliated with the Mesenchymal Stem Cell Group, Division of Haematology, Institute of Medical and Veterinary Science, Adelaide, South Australia, Australia.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures of Potential Conflicts of Interest
  9. Acknowledgements
  10. References