Production of human bone marrow-derived suppressor factor. Effect on antibody synthesis and lectin-activated cell proliferation

Natural suppressor cells resident in normal human bone marrow (BM) exert potent suppressor activity on in vitro antibody responses and other immune functions. A suppressor-enriched population of BM cells can constitutively produce a soluble mediator with similar suppressor activity and kinetics as the suppressor cells. This novel BM-derived suppressor factor (BDSF) suppresses human in vitro primary antibody responses as well as lectin-activated proliferative responses. The mediator (BDSF) has a Mr of less than 1.5 kDa, contains a lipid component, and is insensitive to indomethacin treatment. The BM cells producing the factor bear the HNK-1 surface marker but not T, B, or macrophage markers. The ability of BDSF to suppress Ag-dependent IgM responses during the inductive phase makes it an ideal molecule with the potential to regulate early immune and hemopoietic events within the BM compartment.     

Stimulation of hematopoiesis and bone marrow suppressor cells by the subcutaneous injection of linoleic acid

The effects of injection of linoleic acid into C57Bl/6 mice on hematopoietic and immunological parameters were examined. Administration of linoleic acid stimulated hematopoiesis as it increased spleen weight and cellularity, increased the number of bone marrow and splenic granulocytic-monocytic progenitor cells, and increased the colony stimulating factor activity in the serum of the treated mice. Associated with the hematopoietic stimulation in linoleic acid-treated mice was a decline in the spleen cell blastogenic responses and the appearance of bone marrow suppressor cells which were inhibitory to normal spleen cell blastogenesis. The linoleic acid-induced bone marrow suppressor cells resembled cells of the monocyte lineage in that they were sensitive to treatment with L-leucine methyl ester, partially sensitive to treatment with anti-Ia antibodies and complement, and their suppressor activity was minimized by indomethacin, a prostaglandin synthesis inhibitor. These results suggest that administration of linoleic acid results in hematopoietic stimulation and, concurrently, in the appearance of suppressor cells in the bone marrow. The bone marrow suppressor cells resemble immature cells of the monocyte lineage and appear to mediate their suppressive effects through the production of prostaglandins.     

A soluble factor produced by bone marrow natural suppressor cells blocks interleukin 2 production and activity

We previously reported that a population of Fc gamma-receptor+ (Fc gamma R+) suppressor cells present in normal unstimulated rabbit bone marrow inhibited the growth of autologous rapidly proliferating bone marrow cells devoid of Fc gamma R. It is now reported that the Fc gamma R+ bone marrow cells produced a soluble, nondialyzable suppressor factor(s) (SF) which blocked the proliferation of Fc gamma R- bone marrow cells. In addition, the Fc gamma R+ cells and SF significantly inhibited spleen cell proliferation in response to concanavalin A (Con A), phytohemagglutinin, and pokeweed mitogen. The bone marrow SF exhibited a dose-dependent suppression of the growth of IL-2-dependent T lymphocytes in the presence of IL-2. SF also completely blocked the production or release of IL-2 by Con A-stimulated T cells. Thus, these bone marrow natural suppressor cells produced a soluble factor, which regulated the growth of rapidly proliferating bone marrow cells and also regulated T cell reactivity by modulating IL-2 production and activity.     

Cytokine regulation of bone marrow natural suppressor cell activity in the suppression of lymphocyte function.

Natural suppressor (NS) cells, which nonspecifically suppress immune responses, are present in the spleen following exposure to radiation, chronic graft-versus-host disease, or cancer and in normal bone marrow. A model system is described which allows the study of cytokines activating and inhibiting NS cells, cytokines mediating NS activity, and NS effects on cytokine synthesis. Recombinant interleukin-3 (rIL-3) and granulocyte-macrophage colony-stimulating factor (rGM-CSF) efficiently activated NS cells present in normal bone marrow and were effective at concentrations as low as 5 U/ml. At high concentrations, GM-CSF, but not IL-3, did not activate NS cells. Recombinant interferon-gamma (rIFN-gamma) blocked the activation of bone marrow NS cells by rIL-3, but did not down-regulate NS cells once activated. The NS cells secreted one or more soluble suppressor factors, which blocked IL-2 synthesis and also inhibited IL-2-dependent T cell proliferation in the presence of excess IL-2.     

Immunoregulatory activity of human bone marrow. Identification of suppressor cells possessing OKM1, SSEA-1, and HNK-1 antigens

Human bone marrow contains natural regulatory cells capable of suppressing the in vitro primary IgM response of normal tonsillar cells. The suppression is mediated by non-T cells possessing Fc receptors, OKM1, SSEA-1, and HNK-1 antigens on their surface. The suppression was abrogated by treatment of bone marrow cells (BMC) with anti-HNK-1 or anti-SSEA-1 antisera and complement. Furthermore, BMC depleted of HNK-1+ cells could respond in a primary in vitro antibody response when provided with accessory T cells and macrophages from tonsillar cells. Our findings support the idea that HNK-1+ and HNK-1- BMC populations act antagonistically in the regulation of antibody synthesis. Further, the finding of HNK-1+, SSEA-1+, and OKM1+ suppressor cells in human bone marrow may represent a precursor phenotype of mature natural killer cells with potent immunoregulatory activity.     

Migration of natural suppressor cells from bone marrow to peritoneal cavity by live BCG

Delayed-type hypersensitivity (DTH) responses were suppressed in mice inoculated with bone marrow cells from mice that had been injected with 10(8) colony-forming units (CFU) of live BCG. Upon analysis of this DTH-suppression by the use of a macrophage migration inhibition (MI) assay, the in vitro correlate of DTH, suppressor macrophages in the peritoneal cavity were found to play an important role in DTH suppression. However, neither suppression of DTH nor production of suppressor macrophages was observed in mice inoculated with bone marrow cells from mice that had been injected with methotrexate (MTX), a folic acid antagonist, and 10(8) CFU of live BCG. Moreover, suppressor cells against the MI activity of peritoneal exudate cells from BCG cell wall-immunized mice existed in bone marrow cells from normal mice, natural suppressor (NS) cells, and they were sensitive to MTX. In addition, these NS cells phagocytized carbonyl iron particles, were adherent to Sephadex G-10, and had Fc receptors, but they had no B or T cell markers, suggesting that these cells belonged to a macrophage compartment. From this evidence, we hypothesized that the origin of suppressor macrophages in the peritoneal cavity induced by live BCG injection was MTX-sensitive NS cells in bone marrow, and that these NS cells were stimulated by a small dose of live BCG trapped in bone marrow after i.v. injection of a high dose of live BCG and migrated from bone marrow to the peritoneal cavity.     

Tissue distribution of immunoglobulin-secreting cells in normal and IgA-deficient chickens.

A reverse hemolytic plaque assay was employed to enumerate lymphoid cells actively secreting either immunoglobulin (Ig)G, IgA, or IgM in the small intestine, lungs, and lymphoid organs of normal and IgA-deficient chickens. In normal birds, intestinal lamina propria lymphocytes were proportionately richest in cells secreting IgG and IgA whereas the bone marrow was richest in IgM-secreting cells. The highest ratio of IgA to IgG secreting cells was also found in the lamina propria lymphocytes of the intestine (0.9), followed by the IgA to IgG ratios in the intestinal epithelium (0.31), and the lungs (0.19). The IgA to IgG ratios in the bone marrow (0.08) and the spleen (0.02) were considerably lower. Thus, both the intestine and the lungs were relatively enriched in cells actively secreting IgA. These IgA-secreting cells are the likely source of the IgA found in such quantities in intestinal and respiratory secretions. The tissue distribution of Ig-secreting cells was also studied in two generations of birds with experimentally induced IgA deficiency. There was a striking diminution of IgA-secreting cells in all tissues, including the intestine and lungs, whereas cells secreting IgG and IgM were normal or increased. The lack of IgA-secreting cells in these birds represents the effects of donor suppressor T cells having specificity for IgA.     

B lymphocyte precursors in human bone marrow: an analysis of normal individuals and patients with antibody-deficiency states.

Lymphoid cells containing cytoplasmic IgM but lacking stable surface IgM are believed to be the direct precursors of B lymphocytes. We have characterized these pre-B cells in the bone marrow of normal individuals and patients with a variety of immunoglobulin deficiencies or hematologic disorders by using immunofluorescence and autoradiography. Pre-B cells comprised 5.8 +/- 5.7% of lymphoid cells in normal bone marrow. Eleven patients with infantile X-linked agammaglobulinemia (X-LA) lacked B lymphocytes but had a normal frequency (3.8 +/- 3.6%) of bone marrow pre-B cells. A smaller proportion of marrow pre-B cells from patients with X-LA were engaged in spontaneous DNA synthesis than was found for normal controls. In individuals other than the group with X-LA, the number of circulating B cells was positively correlated with the frequency of marrow pre-B cells. These results indicate that patients with X-LA have a defect in maturation of pre-B cells, and suggest that some patients with acquired B lymphocyte deficiency may have lost the capacity to generate pre-B cells from stem cells.     

Interferon is the suppressor of hematopoiesis generated by stimulated lymphocytes in vitro

Lymphocytes that inhibit hematopoiesis may have a pathogenic role in some forms of bone marrow failure, and lymphocyte-mediated suppression may also be important in the normal regulation of bone marrow function. We have investigated the mechanism of in vitro suppression of hematopoiesis by T cells by using the methylcellulose colony culture system. Total peripheral blood T cells and separated subpopulations of helper (OKT4+) and suppressor (OKT8+) cells that have been stimulated by exposure to lectin suppress autologous colony formation by bone marrow myeloid (CFU-C) and erythroid (BFU-E) progenitor cells. Medium conditioned by these cells is also inhibitory, indicating that the suppressor activity is a soluble factor. A strong correlation existed for the concentration of interferon and the degree of hematopoietic suppressor activity in these supernatants; both activities peaked at days 3 to 5 of incubation and had sharply declined by day 7. Interferon production was enhanced by exposure of lymphocytes to sheep red blood cells during the rosetting procedure. Specific antiserum and a monoclonal antibody directed against gamma-(immune) interferon abrogated the inhibitory activity for hematopoiesis produced by lectin-stimulated T cells; an antiserum to alpha-interferon was generally much less effective in neutralizing activity. We infer from these results that gamma-interferon is the mediator of hematopoietic suppression generated by lectin-treated T-cells.     

Induction of suppressor cells by a tumor-derived suppressor factor

Murine fibrosarcomas produce a factor that activates suppressor cells to inhibit expression of delayed-type hypersensitivity (DTH) responses to dinitrochlorobenzene (DNCB). This tumor-derived suppressor factor (TDSF) was partially purified by preparative isoelectric focusing of spent medium and 3 M KCl extracts of cultured methylcholanthrene-induced and spontaneous fibrosarcomas of C3H/He mice. Incubation of 1 micrograms/ml of a fraction, isoelectric pH less than 2.9, with normal syngeneic spleen cells for 1-6 hr at 37 degrees C induced suppressor cells that inhibited the primary DTH response to DNCB upon intraperitoneal transfer to normal C3H/HeJ mice. TDSF was not present in extracts of either syngeneic embryonic fibroblasts or normal spleen cells or in medium conditioned by normal peritoneal exudate cells but was present in 3 M KCl extracts of and the spent medium from four different cultured murine fibrosarcomas. TDSF activity was not restricted at the major histocompatibility complex. The suppressor cells inhibited the efferent limb of the DTH response because (1) hyporesponsive recipients of TDSF-treated spleen cells had splenic effector T cells capable of transferring DTH to DNCB into naive secondary recipients and (2) the ability of Lyt 1+,2- effector Tdth cells to transfer a secondary DTH response to DNCB was inhibited by co-incubation with macrophages or Lyt 1-,2+ T cells treated with TDSF. Preliminary biochemical analysis suggested that TDSF was an RNA- protein complex. Thus, several murine fibrosarcomas produced a soluble factor that activated splenic suppressor cells to depress the immune response to nonneoplastic antigens. These suppressor factors represent a novel group of regulatory molecules which may be ribonucleoprotein complexes.     

Nature of the target cells in bone marrow B-suppressor action studied on a simplified model of suppression activity registration

The addition of bone marrow cells to spleen cells and lymph node cells stimulates by mitogents, but not to fibroblast-like cells, leads to a significant reduction of DNA synthesis in mixed cultures in vitro. The suppression effects appears only in two days and the suppressor cell activity is the stronger, the intensive is the target cell proliferation. It is shown that intact bone marrow cells can suppress the lipopolysaccharide-activated bone marrow cell proliferation in vitro. A conclusion may be draw that cells of the lymphoid system serve as target cells for the bone marrow suppressor cells, and the role of these lymphoid system cells is to control immunogenesis processes by suppressing the target cell proliferation activity in the bone marrow.     

The ontogeny and distribution of B cells in normal and mutant immune-defective CBA/N mice: two-parameter analysis of surface IgM and IgD

A dual-laser fluorescence-activated cell sorter was utilized to study the distribution of the surface IgM and IgD on individual B cells of normal and immune-defective CBA/N mice. Cells from different lymphoid organs and from developing mice were studied. Two major populations of cells were seen. Those with low densities of surface IgM and intermediate-high densities of surface IgD were relatively or totally absent from the bone marrow, spleens, and lymph nodes of adult, immune-defective (CBA/N x DBA/2)F1 male mice, and developed late in ontogeny in the lymphoid organs of normal F1 female mice. By contrast, the second major population, with intermediate-high surface IgM and low surface IgD, was found in highest frequency in the lymphoid organs of immature mice, the bone marrow of adult mice, and the lymphoid organs of F1 male mice compared to F1 female mice at any age. These two major populations of B cells were further subdivided into five groups of cells to better define the surface IgM and IgD characteristics of developing B cells of immune-defective and normal mice. The relationship of these groups of cells to populations defined by other criteria are discussed.     

Stimulation of suppressor cells in the bone marrow and spleens of high dose cyclophosphamide-treated C57Bl/6 mice

Systemic administration of a single dose (300 mg/kg) of cyclophosphamide (Cy) induced the appearance of a population of suppressor cells in the bone marrow and spleens of mice. Suppressor cells were assayed by their capacity to inhibit the concanavalin A (Con A) blastogenesis or the mixed-lymphocyte response of normal C57Bl/6 spleen cells. Cy-induced bone marrow (Cy-BM) suppressor cells were present as early as 4 days following Cy therapy and their activity gradually decreased over the next 2 weeks. Cy-induced splenic (Cy-Sp) suppressor cells were maximally present on Days 6 through 10 following Cy therapy. Studies were performed to characterize the suppressor cells of bone marrow obtained 4 days after Cy treatment and of normal bone marrow (N-BM). Some suppressor activity was present in normal bone marrow. N-BM suppressor cells resembled cells of the monocyte/macrophage lineage in that they were slightly adherent to Sephadex G-10, sensitive to L-leucine methyl ester (LME), and insensitive to treatment either with anti-T-cell antibody and complement or with anti-immunoglobulin antibody and complement. Their suppressive activity was abrogated by incubation with either indomethacin or catalase. Cy-BM suppressor cells were also resistant to treatment with anti-T-cell and anti-immunoglobulin antibody and complement but were not adherent to Sephadex G-10 and not sensitive to LME. Their suppressive activity was partially eliminated by indomethacin alone or in combination with catalase. We conclude that Cy chemotherapy induces the appearance of a population of immune suppressive cells and that these cells appear first in the bone marrow and subsequently in the spleen.     

Rabbit bone marrow suppressor cells block the production or release of a soluble bone marrow growth factor

Fc gamma-receptor (Fc gamma R)-bearing cells from normal rabbit bone marrow suppressed the constitutive proliferation rates of the remaining, Fc gamma R-, cells. In the absence of the suppressive influences of Fc gamma R+ cells, cells in the Fc gamma R- population spontaneously elaborated a soluble growth factor (GF) which induced the proliferation of unseparated bone marrow cells. To examine regulation by Fc gamma R+ bone marrow cells, graded numbers of the Fc gamma R+ cells were mixed with constant numbers of the FcR- cells. At 24 hr, supernates were collected and tested for GF activity. The Fc gamma R+ suppressor cells efficiently and in a dose-dependent fashion blocked GF production or release. The GF was nondialyzable and relatively heat stable. Supernates with GF activity also had colony-stimulating factor activity, but were negative in assays modified from murine interleukin 1 (IL-1) and IL-2 assays. Regulation of GF production or release represents a new function for bone marrow suppressor cells.     

Suppression of cell-mediated immune responses after total lymphoid irradiation (TLI). I. Characterization of suppressor cells of the mixed lymphocyte reaction

Total lymphoid irradiation (TLI) was administered to (BALB/c X C57BL/6)F1 mice in eight daily doses of 200 rad (total 1600 rad). Spleen cells isolated from mice after treatment with TLI do not respond to alloantigens in vitro in a one-way mixed lymphocyte reaction (MLR), but normal reactivity recovers after approximately 2 mo. Radioresistant, antigen-nonspecific suppressor cells are documented in the spleens of TLI-treated mice immediately after radiotherapy, but suppressive capacity gradually disappears within 30 days. After TLI, the spleen is repopulated with large cells, the proportion of which is greatest at a time when theta-bearing cells are still depleted. Radioresistant suppression is mediated predominantly by the large cell subset and is thymus independent. Suppressor function can be abolished by lethal physicochemical procedures including formaldehyde fixation, multiple freeze-thawing, and heating to 56 degrees C, and it cannot be conferred by supernatants of TLI-suppressed MLR suspensions. Suppression cannot be overcome by adding various cell factors including T cell growth factor (TCGF) and lymphocyte-activating factor (LAF), nor is it affected by a prostaglandin inhibitor. Equally potent radioresistant suppressive activity is documented by co-culturing cells derived from other sources enriched in large, immature hematopoietic cells, including fetal liver cells and bone marrow cells obtained from normal and congenitally athymic mice. The presence of a large cell population and MLR suppressor function is also documented in the spleens of mice treated with single dose or fractionated doses of lethal whole body irradiation, followed by reconstitution with bone marrow cells obtained from normal mice. The data suggest that MLR suppressor cells, which are large, immature and predominantly radioresistant, can be induced after a short and well-tolerated TLI regimen.     

Inhibition of adoptive secondary responses by lymphoid cell populations

The ability of normal and tolerant lymphoid cells to inhibit the adoptive secondary response was investigated in order to delineate the influence which the host can exert on a memory cell population as the host recovers from the effects of irradiation. LBN rats were irradiated with 500R whole body X-irradiation, reconstituted with either normal or tolerant lymphoid cells, then they were given immune spleen cells and challenged with soluble antigen (DNP-BGG). Cells capable of suppressing the adoptive secondary responses were shown to possess the following characteristics: 1) in nonimmune donors they were found in the greatest concentration in lymph nodes, followed by spleen and bone marrow and were practically absent in the thymus; 2) their numbers were not increased in donors rendered tolerant by long term treatment with deaggregated DNP-BGG plus cyclophosphamide nor in donors given large doses of DNP-BGG 48 to 72 hr before sacrifice; 3) in animals rendered tolerant by long term antigen treatment alone some enhancement of suppressor function was seen; 4) the suppressor cells could be shown among both glass wool adherent and nonadherent cells and 5) the nonantigen specific suppressor cells did not affect the kinetics of antibody formation nor the affinity of the antibody which was produced. These results are discussed in terms of the nature and source of the suppressor cell populations and their relevance to the control of secondary responses in intact animals.     

IgM rheumatoid factor autoantibody and immunoglobulin-producing precursor cells in the bone marrow of humans

The natures of the IgM rheumatoid factor (RF)-, IgM-, and IgG-secreting cells in the human bone marrow as compared to the peripheral blood, have been investigated by (1) response to the polyclonal B-cell activator, the Epstein-Barr virus (EBV), (2) sensitivity to the S-phase specific antimetabolite hydroxyurea, (3) presence of the BA-1 and Ia antigens on the cell surface, and (4) cell size, as determined by counter flow elutriation. The EBV-inducible bone marrow IgM-RF precursors derived from medium to large B cells that were inhibited by hydroxyurea pretreatment. The marrow total IgM response derived from small to medium size cells, and was only partially inhibited by hydroxyurea. Hydroxyurea had no effect on IgM-RF or IgM synthesis by peripheral blood cells. These results indicate that the marrow EBV-induced IgM-RF response is not representative of the response by peripheral blood cells, moreover; the marrow RF secreting response arises from a dividing cell pool that may represent newly generated autoreactive B cells.     

Murine bone marrow IgA responses to orally administered sheep erythrocytes

Specific immunization protocols have been established for the induction of murine bone marrow IgA responses to the T cell-dependent (TD) antigen sheep red blood cells (SRBC). Systemic immunization, either i.p. or i.v., followed by a second injection, induced splenic IgM and IgG responses and a bone marrow IgM response. No significant IgA responses were observed in either lymphoid tissue compartment. Oral immunization with SRBC by gastric intubation for 2 days, followed 1 wk later by an i.p. injection of SRBC resulted in a splenic IgA plaque-forming cell (PFC) response, but did not elicit a bone marrow IgA response. Repeated daily gastric intubation of SRBC to C3H/HeN and C3H/HeJ mice led to the previously reported pattern of systemic unresponsiveness in C3H/HeN mice and good anamnestic type IgM, IgG, and IgA splenic anti-SRBC PFC responses in the C3H/HeJ strain upon parenteral challenge. Oral administration of SRBC for 14 days to C3H/HeN mice, followed by systemic SRBC challenge, resulted in diminished splenic PFC responses of all isotypes, whereas gastric intubation of SRBC for 28 days led to complete systemic unresponsiveness to antigen in C3H/HeN mice. Interestingly, the repeated oral administration of SRBC resulted in significant bone marrow IgA PFC responses upon i.p. challenge in both C3H/HeN and C3H/HeJ mouse strains. The bone marrow IgA responses were clearly dependent upon chronic oral exposure to SRBC, because gastric intubation with SRBC for 2 consecutive days/wk for 10 wk also induced bone marrow and splenic IgA anti-SRBC PFC responses in C3H/HeN mice. These results suggest that memory B cells reside in the bone marrow of orally immunized mice and can yield anamnestic-type responses to challenge with the inducing antigen. The memory cells may arise in the Peyer's patches of the gut and migrate to the bone marrow. The possibility that the bone marrow is a component of the common mucosal immune system in mammals is suggested by this study.     

The effect of bone marrow depletion on prostaglandin E-producing suppressor macrophages in mouse spleen

The i.p. injection of Corynebacterium parvum (CP) into CBA/J mice effected increases in macrophage colony-forming cells (M-CFC) when spleen cells were cultured with L cell culture filtrate as a source of colony-stimulating factor. Significant increases in phagocytic macrophages (M phi) with Fc receptors for IgG2a and IgG2b immune complexes were additionally noted among the spleen cells in these mice. These M phi effectively inhibited Con A-induced lymphocyte proliferation, probably reflecting a 10-fold increase above normal controls in prostaglandin E to 47 ng/3 X 10(6) spleen cells/ml. To determine whether the suppressor M phi are immediate derivatives of splenic M-CFC, we tried to induce suppressor M phi by the injection of CP into mice depleted of bone marrow M-CFC by the earlier administration of the bone-seeking isotope, 89Sr. This procedure reduced M-CFC in the bone marrow to less than 1% of normal for more than 30 days. Monocytes in the blood fell to 5% of normal by day 10 and were 30% on day 30. Levels of resident peritoneal M phi showed relatively little change in this period. By contrast, splenic M-CFC increased to 20-fold higher than the "cold" 88Sr controls. CP-induced suppressor M phi activity, however, was sharply reduced in 89Sr marrow-depleted mice on day 10, despite the striking increase in M-CFC. There was a threefold increase in the number of phagocytic M phi binding IgG2a immune complexes, with no significant increase in IgG2b binding M phi. The kinetics of recovery of suppressor M phi activity showed that on days 20, 30, and 50 after 89Sr injection the activities reached 20%, 30%, and 70% of the "cold" control, respectively, and correlated with the recovery of significant levels of M-CFC in the bone marrow. Taken together, these observations suggest that splenic M-CFC are not an immediate source of PGE-suppressor M phi in vivo. It appears more likely that the CP-inducible suppressor M phi, in particular, originate from radiosensitive bone marrow cells or require for differentiation a microenvironment provided by bone marrow cells. The data also suggest that the expression of the Fc gamma 2b receptor and of suppressor activity by CP-induced splenic M phi are related phenomena.     

Selective induction of autoantibody secretion in human bone marrow by Epstein Barr virus

The bone marrow is an important site for B lymphocyte differentiation and antibody synthesis in animal and man. However, few experiments have examined directly its immunologic functions in humans. In the present experiments, we have induced bone marrow B lymphocytes from human donors with degenerative arthritis of varying ages to secrete two autoantibodies, IgM and anti-IgG (rheumatoid factor) and IgM anti-human thyroglobulin (Tg), by stimulation with the polyclonal B cell activator Epstein Barr virus (EBV). The EBV-stimulated bone marrow cells secreted significantly more IgM anti-IgG (p less than 0.01) and IgG anti-Tg (p less than 0.01) than matched, identically treated peripheral blood cells. Bone marrow cultures from donors over the age of 60 yr, particularly females, produced more rheumatoid factor than cultures from younger donors (p less than 0.01). The EBV-inducible autoantibodies were immunospecific as demonstrated by adsorption studies. A potential pathogenic role in the inflammatory process was suggested by the finding that the EBV-inducible IgM anti-IgG autoantibodies were capable of activating the classical complement pathway as assessed by the cleavage of C4. These results indicate that the human bone marrow is a selective reservoir for EBV-inducible autoantibody precursor B lymphocytes, and that the size of the reservoir increases with age.