Search

We have examined the capacity of highly purified human CD34+ marrow cell isolates from unrelated, HLA-mismatched donors to establish in vitro hematopoiesis on recipient marrow stromal cells in 2-stage hematopoietic long-term marrow cultures (H-LTMC). HLA-typing of both peripheral blood mononuclear cells and CD34+ marrow cells was performed for both HLA class I and HLA class II antigens for eight healthy individuals. Significant antigenic mismatches for these molecules ranged from three to six antigens for each recipient-donor pair. Comparison of MHC antigen expression by peripheral blood cells and CD34+ marrow cell isolates confirmed the presence of identical HLA-A, -B, and -C, and -DR specificities on the surface of these cells. Typing of -DQ specificities, however, was not consistently reactive on CD34+ cells. The > or = 20% plating efficiency of purified CD34+ cells for BFU-E, CFU-GM, and CFU-MIX allowed us to use inoculum doses of 10(3), 10(4), and 10(5) cells to determine the efficiency of allogeneic CD34+ cells in achieving in vitro engraftment and the establishment of hematopoiesis in H-LTMC. Engraftment of adherent BFU-E, CFU-GM, and CFU-MIX was equally efficient for autologous and allogeneic CD34+ cells. In vitro hematopoiesis reflected by the cumulative recoveries of progenitor cells over time was also equivalent for allogeneic and autologous CD34+ cells. These results demonstrate that highly purified, HLA-mismatched CD34+ marrow cells proliferate and establish in vitro hematopoiesis as efficiently as autologous cells in marrow derived stromal cell cultures and confirm that interactions between stromal cells and highly purified CD34+, DR-, and CD34+, DR+ marrow cell isolates are not MHC-restricted.

Primary autologous as well as allogeneic and xenogeneic stroma will support human stem cell proliferation and differentiation for several months. In the present study, we investigated the capacity of porcine microvascular endothelial cells (PMVECs) together with combinations of cytokines (granulocyte-macrophage colony-stimulating factor [GM-CSF] + stem factor [SCF], interleukin-3 [IL-3] + SCF + IL-6, and GM-CSF + IL-3 + SCF + IL-6) to support the expansion and development of purified human CD34+ bone marrow cells. In short-term cultures (7 days), the greatest expansion of nonadherent hematopoietic cells and clonogenic progenitors was seen with CD34+ cells in direct contact with PMVEC monolayers (PMVEC contact), followed by PMVEC noncontact and liquid suspension cultures, respectively. Maximal expansion of nonadherent cells (42-fold) and total CD34+ cells (12.6-fold) occurred in PMVEC contact cultures treated with GM-CSF + IL-3 + SCF + IL-6, with similar increases in the number of granulocyte-macrophage colony-forming units (CFU-GM), CFU-mix, erythroid burst-forming units (BFU-E), CFU-blast and CFU-megakaryocyte (CFU-Mk) progenitor cells. Moreover, the number of CD34+ CD38- and CD34+ CD38+ cells increased 148.1-fold and 8.0-fold, respectively. Replating studies show that cells from day 7 dispersed blast cell colonies generated on cytokine-treated PMVEC monolayers have a high replating potential for multilineage progenitor cells. In long-term PMVEC contact cultures, CD34+ cells seeded onto PMVEC monolayers with GM-CSF + IL-3 + SCF + IL-6 showed a total calculated expansion of over 5,000,000-fold of nonadherent cells over 35 days in culture. Maximal clonogenic cell production was observed at day 28, with 6,353-fold for total CFC and comparable increases for CFU-GM, CFU-mix, CFU-blast, BFU-E, and CFU-Mk. The total number of CD34+ cells increased 2,584-fold at day 28. Furthermore, the extended growth kinetics of these cultures indicates that these phenotypically primitive progenitor cells are also functionally expanded on PMVEC monolayers. These results support the hypothesis that direct contact with a PMVEC monolayer supports the initial expansion of hematopoietic progenitor cells with a high replating potential and, possibly, a more primitive phenotype (CD34+, CD34+/CD38-).

This report presents the results of studies using long-term bone marrow cultures (LTBMC) of human bone marrow cells to investigate the effect of HIV-1 on in vitro hematopoiesis. Confluent stromal cell layers established from human bone marrow cells were irradiated to eliminate residual hematopoietic progenitor cells and exposed to HIV-1ADA or to HIV-1IIIB, monocytotropic and lymphocytotropic strains of HIV-1, respectively. A productive infection did not develop in cultures exposed to HIV-1IIIB but did for cultures exposed to HIV-1ADA as there was a progressive increase in HIV-1 p24 antigen. Stromal cell layers infected with HIV-1ADA were also cocultured with autologous CD34+ bone marrow cells. Four days, 1, 2, and 3 weeks later, the number of colony-forming units granulocyte/macrophage (CFU-GM) in non- and HIV-infected LTBMC was determined. The number of CFU-GM increased during the first week in both non- and HIV-infected LTBMC. One week after the coculture of CD34+ cells with stromal cell layers infected with HIV-1ADA, the number of CFU-GM in six out of eight experiments was reduced compared to noninfected control LTBMC. In those six experiments, the number of CFU-GM was 53 +/- 6% standard error of the mean (SEM) of the number in noninfected LTBMC. A reduced number of CFU-GM was observed in the nonadherent fraction of HIV-infected LTBMC for at least 2 weeks. These results demonstrate that some cells in the stromal cell layers of LTBMC were targets for HIV-1 and that HIV-infected stromal cell layers suppressed or delayed the production of CFU-GM.

To determine the fate of Fc epsilon RI+ cells on fibroblasts in vitro, human bone marrow derived CD34+ cells were cultured in the presence of recombinant human interleukin 3 and recombinant human hematopoietic stem cell factor for 3 weeks, and Fc epsilon RI+ cells were purified by immunomagnetic selection. This enriched Fc epsilon RI+ cell population consisted of 92-94% basophils and 3-5% mast cells as determined by morphologic, immunohistochemical, and ultrastructural criteria. The Fc epsilon RI+ cells were then cocultured with 3T3 fibroblasts. Basophils decreased markedly by 1 week and were absent from cocultures by 2-3 weeks, while the mast cell numbers on the fibroblast monolayers remained constant. Ultrastructural examination of cocultures at 2 days demonstrated phagocytosis of basophils by fibroblasts. By 1 week, phagocytosed basophil membranes and granules gave fibroblasts the superficial appearance of mast cells by toluidine blue staining. Mast cells surviving in cocultures could be distinguished from granule-containing fibroblasts by IgE surface labeling and by ultrastructural demonstration of tryptase-positive granules. Thus, while mast cells remain viable in coculture with 3T3 fibroblasts, basophils do not survive and are internalized and degraded by the fibroblast monolayer.

Retroviral mediated gene transfer into stem cells has been proposed as therapy for many inherited hematopoietic diseases. Deficiency of the enzyme adenosine deaminase (ADA) results in depletion of T lymphocytes, causing severe combined immunodeficiency syndrome (SCIDS). In this report, we describe retroviral mediated gene transfer of a murine ADA cDNA into Rhesus monkey hematopoietic stem cells. Immunoselected CD34+ bone marrow cells were exposed to medium containing the ADA retrovirus during culture on a stromal cell line engineered to express the transmembrane form of stem cell factor. After infusion of autologous, transduced cells into irradiated recipients, gene transfer was observed in all three monkeys. The ADA provirus was detected in 2% of circulating granulocytes and T cells from 100 days post-transplantation to longer than 1 year and in B cells from 250 days post-transplantation and beyond. Mouse ADA activity was detected in peripheral blood cells at approximately 3% the activity of monkey ADA. Thus, we have shown gene transfer into repopulating cells that contribute to all hematopoietic lineages with persistent gene expression. These data provide support for the use of stem cell targeted gene transfer for therapy of ADA deficiency.

Human marrow cells that express the CD34 antigen but lack CD33 are able to initiate sustained, multilineage in vitro hematopoiesis in long-term Dexter cultures and are believed to include the primitive stem cells responsible for effecting long-term hematopoietic reconstitution in vivo following marrow transplantation. In studies described in this report we investigated the effects of a novel anti-CD33 immunotoxin on the clonogenic potential of normal human CD34+ marrow cells and on the ability of these cells to initiate hematopoiesis in two-stage Dexter cultures (long-term marrow cultures, LTMC). This immunotoxin (anti-CD33-bR), shown previously to kill both clonogenic myelogenous leukemia cells and normal mature myeloid progenitor cells (granulocyte-macrophage colony-forming units, CFU-GM), consists of an anti-CD33 monoclonal antibody conjugated to purified ricin that has been modified by blocking the carbohydrate binding domains of the ricin B-chain to eliminate nonspecific binding. For our studies, normal CD34+ human marrow cells were isolated from the light-density (less than 1.070 g/ml) cells of aspirated marrow by positive selection with immunomagnetic beads linked to the monoclonal antibody K6.1. These cell isolates were highly enriched with both multipotential and lineage-restricted clonogenic, hematopoietic progenitors (mixed lineage colony-forming units, CFU-Mix; CFU-GM; and erythroid burst-forming units, BFU-E) which constituted greater than or equal to 20% of the cells. Recovery of clonogenic progenitors from these CD34+ cell preparations, following treatment with anti-CD33-bR (10 nM), was reduced by greater than or equal to 85% for CFU-GM and 20%-40% for CFU-Mix and BFU-E. However, the capacity of these cells to initiate hematopoietic LTMC was preserved. Indeed, the production of high proliferative potential (HPP) CFU-GM, BFU-E, and CFU-Mix in cultures seeded with 10(5) anti-CD33-bR-treated CD34+ marrow cells was substantially greater than that observed in LTMC seeded with equivalent numbers of untreated CD34+ cells. Moreover, concentrations of long-term culture initiating cells in CD34+ cell isolates, quantified by a limiting dilution technique, were found to be increased following anti-CD33-bR treatment. These findings support the potential usefulness of anti-CD33-bR for in vitro marrow purging or in vivo treatment to eliminate CD33+ leukemic clones, while sparing normal CD34+/CD33- stem cells that support normal hematopoiesis and hematopoietic reconstitution in vivo.

To define conditions for improved efficiency of retroviral-mediated gene transfer and expression in primate progenitor cells, four rhesus monkeys were treated with a 200 mg/kg intravenous bolus of 5-fluorouracil (5-FU). The kinetics of hematopoietic suppression and recovery were assessed in peripheral blood, bone marrow mononuclear cells, and bone marrow cells fractionated in an albumin density gradient. Bone marrow mononuclear cells were transduced with N2, a retroviral vector carrying the bacterial neomycin phosphotransferase gene (NPT), which confers resistance to the otherwise toxic neomycin analogue, G418. Circulating colony-forming units-granulocyte-macrophage (CFU-GM) disappeared at 2 days. CFU-GM, transducible CFU-GM, CD34+ cells, and the percent of cells in cycle decreased at 3 days in unfractionated bone marrow cells and in a light density population known to be enriched for these progenitors and for stem cells. NPT activity in the light-density fraction, marginally detectable before treatment, disappeared at 3 days as well. At day 7 the CFU-GM plating efficiency, the CD34+ cell content, and the percentage of cells in cell cycle began to increase in the light-density fraction. The NPT assay became faintly positive again but the CFU-GM were not yet transducible, implying that it was an earlier progenitor population that was dividing and differentiating. By day 15, there was a marked rebound in all of the progenitors measured, and transduction efficiency assessed by G418R CFU-GM and NPT assay rebounded to several times pretreatment levels. The data suggest that CFU-GM are optimally transduced at 15 days but that earlier progenitors are more likely cycling and transducible before 5 days, a time when a gene transfer experiment would probably have the best chance to succeed.

Myeloid progenitor cells were highly purified from normal human bone marrow by positive immunoselection with high-affinity monoclonal antibodies linked to magnetic beads and were successfully infected in vitro with the human immunodeficiency virus type 1 (HIV-1). From 99 to 100 percent pure bone marrow cells expressing the CD34 phenotypic marker were obtained. These cells were devoid of mature myeloid or T cell surface and intracellular markers as analyzed by immunohistochemical staining and flow cytometry. HIV-1 particles were detected by supernatant reverse transcriptase activity and transmission electron microscopy 40 to 60 days after infection. Viral particles were predominantly observed assembling and accumulating from within intracellular membranes, while phenotypically the cells were observed to have differentiated into CD4+ monocytes. These studies have important implications in understanding the pathogenesis of HIV-1 as well as the possible cause of certain of the observed hematologic abnormalities in HIV-1 infection. They also indicate that the bone marrow may serve as a potentially important reservoir of HIV-1 in the body.

The Cowan I strain of the bacterium Staphylococcus aureus has been used as an adsorbent for antibodies complexed with radiolabeled antigens from cell lysates. This application is advanced as a superior alternative to other methods of immune precipitation for the isolation of antigens. It exploits the high adsorption capacity for IgG molecules by protein A molecules on the cell walls of certain strains of staphylococci, along with the advantageous sedimentation properties of the bacteria. The interaction of immune complexes with the adsorbent was defined initially using a model system of bovine serum albumin with a high excess of rabbit anti-bovine serum albumin antibodies (IgG). The uptake of immune complexes under these conditions was extremely rapid, occurring within seconds, whereas maximum binding of free IgG was much slower. In addition, once bound the complexed antigen could not be displaced from the adsorbent either by large amounts of normal IgG or by extra free antibody. Antigen could be eluted almost completely from the inert adsorbent for analytic or preparative purposes with a variety of solvent systems, such as the detergent SDS in combination with urea and high temperature, and neutral salts with strong lyotropic salting in properties. The efficacy of the protein A-antibody adsorption technique was tested in direct comparisons with a conventional double antibody precipitation method for the isolation of mouse lymphocyte IgM. The bacterial adsorbent not only had a distinct advantage in speed of antigen isolation, but analyses by polyacrylamide gel electrophoresis in SDS also revealed consistently higher antigen recoveries, lower levels of background radioactivity, and an absence of other cell components which may nonspecifically bind to and complicate analyses using conventional immune precipitates.

A 6 1/2-month-old male with severe combined immunodeficiency (SCID) had a low percentage and number of T cells (11%; 241/mm3) and a high percentage and number of B cells (52%; 1187/mm3) and null cells (37%; 868/mm3). In vitro studies were performed to determine if this child's primary defect involved differentiation of both T and B lymphocytes or if failure of B lymphocytes to differentiate into immunoglobulin producing cells was secondary to T lymphocyte abnormalities. Immunoglobulin production by lymphocytes in response to polyclonal mitogens (pokeweed mitogen and foetal calf serum) was measured by pulse-labelling cells with 3H-leucine and then precipitating cytoplasmic and secreted immunoglobulins with polyvalent anti-human immunoglobulin and S. aureus (Cowan strain I) protein A. The patient's lymphocytes did not synthesize immunoglobulins in vitro in response to mitogens. They did not suppress synthesis of immunoglobulins by normal lymphocytes. However, addition of normal purified T cells, which themselves did not synthesize immunoglobulins, enabled the patient's B lymphocytes to become immunoglobulin synthesizing and secreting cells. Gamma, mu, and light chains were secreted. This suggests that the primary abnormality was in the T-cell axis at the level of lymphoid stem cells or prothymocytes and that failure of B lymphocytes to become immunoglobulin-producing cells was secondary to this defect.

Procedures are detailed for the rapid isolation of representative cell membrane antigens with protein A-bearing staphylococci as an adsorbent for IgG antibodies complexed with the antigens. Cell surface membrane proteins were radioiodinated and solubilized in nonionic detergent. Specific antisera were subsequently added and the immune complexes precipitated by addition of the staphylococcal adsorbent and low speed centrifugation. The antigens isolated included surface immunoglobulins from mouse and human lymphocytes, human beta-microglobulin and HL-A alloantigens, mouse H-2 alloantigens, and the murine leukemia virus glycoprotein gp 70. Rabbit, sheep, goat, and mouse antisera were all effective for the specific phase of the precipitation reaction. The surface membrane immunoglobulins of mouse splenic lymphocytes and human peripheral blood lymphocytes differed with respect to class composition and protein A reactivity. Mouse lymphocyte surface immunoglobulins were nonreactive with protein A, whereas a high proportion of human lymphocyte surface immunoglobulins of different classes bound directly to the staphylococci. In sequential immunoprecipitation studies the prior isolation of one antigen had no appreciable effect on the subsequent recovery of another antigen. Adsorption of antigen-antibody complexes is quantitative when protein A sites are provided in excess over antiserum IgG sites, and this obviates the need for equivalence point titrations for optimal precipitation necessary with alternative double antibody techniques.

Mice have recently been shown to have serum IgD. We have used affinity chromatography to partially purify mouse serum IgD, and have prepared a rabbit antiserum against this mouse Ig class. This antiserum, once adsorbed by IgD-depleted mouse serum bound to Sepharose, was isotype specific as determined by radioimmunoassay and fluorescence activated cell sorter analysis, and bound to the same splenic lymphocyte surface molecules as a hybridoma produced monoclonal anti-mouse delta antibody.