Increased spontaneous secretion of rheumatoid factor by intestinal lamina propria mononuclear cells from Crohn's disease but not ulcerative colitis patients.

Increased levels of rheumatoid factors (RF) have been observed in the serum of Crohn's disease but not ulcerative colitis patients, and have been proposed to relate to an increased state of intestinal lymphocyte activation. We have therefore examined the spontaneous in vitro secretion of RF by intestinal lamina propria mononuclear cells (MNC) isolated from specimens from control and inflammatory bowel disease (Crohn's disease, ulcerative colitis) patients. Normal intestinal lamina propria MNC spontaneously secrete rheumatoid factors of different isotypes during 14 days of in vitro culture (9.7 ng/ml IgA RF, 11.6 ng/ml IgM RF and 64.6 ng/ml IgA anti-Fc (IgG)). In matched studies intestinal MNC isolated from normal large bowel exhibited significantly greater levels of RF synthesis and secretion in vitro than normal small bowel intestinal MNC. A large increase in spontaneous RF secretion was observed from Crohn's disease intestinal MNC (21.4 ng/ml IgA RF, 21.4 ng/ml IgM RF, and 108.15 ng/ml IgA anti-Fc (IgG)), when compared with normal controls. The amount of RF secreted was dependent on the amount of inflammatory activity of the bowel specimens, from which the MNC were isolated (198.3 ng/ml of IgA anti-Fc(IgG) from involved versus 50.0 ng/ml from matched non-involved tissue). Ulcerative colitis MNC released decreased amounts of RF (7.1 ng/ml IgA RF, 6.2 ng/ml IgM RF, and 42.3 ng/ml IgA anti-Fc(IgG)). These observations using isolated intestinal MNC may explain the findings of RF changes in the sera of inflammatory bowel disease patients. Our observations support the hypothesis of a heightened state of activation in normal intestinal lamina propria MNC, which is further increased in active Crohn's disease. The dissimilarities observed between Crohn's disease and ulcerative colitis may indicate fundamental differences in disease pathophysiology and will lead to further studies exploring intestinal immunoregulatory properties of RF.

Increased activation of isolated intestinal lamina propria mononuclear cells in inflammatory bowel disease.

Normal human lamina propria lymphocytes are in a heightened state of activation compared with peripheral blood with regard to cell-surface activation antigen expression (transferrin receptor, interleukin-2 receptor, 4F2) and the increased spontaneous secretion of immunoglobulins in vitro. This study evaluates the cell-surface expression of activation-associated antigens in different subpopulations of isolated colonic lamina propria mononuclear cells in inflammatory bowel disease. In pilot studies using three-color flow cytometry, autofluorescence was observed that was emitted by unstained lamina propria mononuclear cells, which interfered with both the sensitivity and the specificity of the analyses. Because a major portion of the intestinal lymphocyte populations of interest were autofluorescent, a method to remove autofluorescence signals was developed by designing a computer program for the subtraction of autofluorescence from the emissions of each individual cell. This technique increases both the sensitivity and specificity of flow-cytometric analyses of intestinal lamina propria mononuclear cells. Using fluorescence-activated cell-sorter analyses with subtraction of autofluorescence on a single-cell basis, increased expression of lymphocyte activation antigens (interleukin-2 receptor, transferrin receptor, 4F2) was found on the cell surface of isolated intestinal B cells, T cells, CD4+ T cells, and CD8+ T cells in both Crohn's disease and ulcerative colitis. Therefore, markedly increased intestinal lymphocyte activation is a major immunological alteration in inflammatory bowel disease and includes all lymphocyte subpopulations investigated in this study. In addition, 5-aminosalicylic acid, which is used for the treatment of intestinal inflammation in inflammatory bowel disease, inhibits the expression of cell-surface activation antigens on mitogen-activated peripheral blood lymphocytes in a dose-dependent manner. These observations suggest that lymphocyte activation may play an important role in underlying immune processes that lead to chronicity and perpetuation of inflammatory bowel disease and may implicate an additional mechanism for the therapeutic action of 5-aminosalicylic acid.

Human rib bone marrow mononuclear cells spontaneously synthesize and secrete IgE in vitro.

We have examined spontaneous secretion of IgE by human rib bone marrow mononuclear cells (MNC). Bone marrow MNC from nine out of 12 rib specimens synthesized and secreted substantial amounts of IgE during 14 days of in vitro culture. The 14-day supernatants from these bone marrow MNC contained a mean of 2589 pg/ml of IgE (n = 12) with a maximum production of 15,408 pg/ml of IgE compared with small amounts of IgE (80-200 pg/ml) produced by similarly cultured normal and inflammatory bowel disease intestinal lamina propria MNC. Using two rib specimens, time-course studies revealed spontaneous secretion of IgE to be minimal during the first 2 days of culture (152 pg/ml), followed by a steady increase between days 4 (517 pg/ml) and 14 (3588 pg/ml). The addition of pokeweed mitogen resulted in 72% suppression of spontaneous IgE production by bone marrow MNC. The bone marrow MNC isolated from the ribs consisted of 22% Leu12+ (B) cells of which 3.2% were surface IgE positive. Staining for cytoplasmic immunoglobulin revealed 1% of the bone marrow MNC to be cytoplasmic IgE+. The presence of IgE-bearing and IgE-secreting MNC in human bone marrow is consistent with the observation that allergen-specific IgE-mediated hypersensitivity is adoptively transferred by human bone marrow transplantation and demonstrates the usefulness of human bone marrow MNC for examination of IgE secretory and regulatory events.

Bovine IgG1, but not IgG2, binds to human B cells and inhibits antibody secretion.

We previously observed that milk-derived bovine IgG, but not serum-derived bovine IgG, strongly inhibits antibody secretion by pokeweed mitogen (PWM)-stimulated human peripheral blood mononuclear cells (PBMC). Bovine milk contains a greater percentage of IgG1 (90%) than does bovine serum (53%). To determine whether bovine IgG subclasses have different functional capabilities, we have examined the effects of bovine IgG1 and IgG2 subclasses upon not only antibody secretion but also mitogenesis by human PBMC. Both bovine IgG subclasses markedly inhibited PWM-stimulated mitogenesis. However, only bovine IgG1, and not IgG2, inhibited antibody secretion during a 14-day in vitro culture period. Also, antibody secretion was inhibited following a 24-hr preincubation of human PBMC with bovine IgG1, but not with IgG2. To determine whether these differences corresponded to specificities of human Fc gamma receptors on subsets of mononuclear cells, fluorescence-activated cell sorter (FACS) analyses were performed. Both bovine IgG subclasses bound to human monocytes. However, only bovine IgG1 bound to human B cells, and bovine IgG1 bound more avidly to human B cells than did human IgG. One model to explain these findings is that inhibition of mitogenesis may be due to the binding of both bovine IgG1 and IgG2 subclasses to monocytes; whereas subclass-specific inhibition of antibody secretion may result from the selective binding of bovine IgG1, but not bovine IgG2, to B cells. The observation that bovine IgG1 has a greater avidity for human B lymphocyte Fc receptors than human IgG may have important implications for future studies of Fc gamma receptors on human leucocytes.

Normal human intestinal B lymphocytes. Increased activation compared with peripheral blood.

The state of activation of normal human intestinal mononuclear cells obtained from transplant donors was studied. Compared with PBMC, freshly isolated intestinal mononuclear cells expressed significantly more cell surface activation antigens on both B and T lymphocytes. Intestinal mononuclear cells contained significant numbers of immunoglobulin secreting cells immediately after cell separation. This population included CD5-positive B cells that secreted predominantly IgA. Cells from the large bowel consistently revealed higher numbers of IgA secreting cells than cells from the small bowel. Thus, intestinal B cells are markedly activated in vivo compared with PBMC and this increased activation correlates with increased spontaneous antibody secretion. B cells from the large intestine are more highly activated and secrete more antibody than do cells from the small intestine. The intestinal lamina propria lymphoid compartment exhibits a heightened state of activation that may be important for its distinct role in mucosal defense.

Alterations in serum immunoglobulin G subclasses in patients with ulcerative colitis and Crohn's disease.

We have examined the concentration of immunoglobulin G (IgG) subclass antibodies in the sera of 27 patients with ulcerative colitis and 21 patients with Crohn's disease as well as in 11 normal controls and 11 patients with systemic lupus erythematosus. In comparison with a control mean serum IgG1 concentration of 5173 micrograms/ml, patients with ulcerative colitis exhibited a significantly increased mean serum concentration of 7924 micrograms/ml (p less than 0.05), whereas patients with Crohn's disease had a near normal mean serum IgG1 level of 5898 micrograms/ml. In contrast, control sera had a mean IgG2 level of 2477 micrograms/ml and ulcerative colitis sera had a similar IgG2 level of 2269 micrograms/ml, whereas Crohn's disease sera had a significantly increased mean IgG2 level of 5111 micrograms/ml (p less than 0.05). Patients with systemic lupus erythematosus, like those with ulcerative colitis, had a markedly elevated serum IgG1 level of 15,594 micrograms/ml (p less than 0.001) without a significantly increased IgG2 serum level (3271 micrograms/ml). Neither ulcerative colitis nor Crohn's disease sera exhibited altered levels of IgG3 or IgG4. These data show that alterations in IgG subclass concentrations occur in the sera of patients with active, untreated inflammatory bowel disease, similar to the previously noted changes in the IgG subclasses secreted by lymphocytes from involved inflammatory bowel disease intestinal specimens.

Inhibition of antibody secretion by 5-aminosalicylic acid.

We have examined the effects of sulfasalazine and its metabolites sulfapyridine and 5-aminosalicylic acid on antibody secretion by normal peripheral blood and intestinal mononuclear cells. Sulfasalazine and 5-aminosalicylic acid both inhibited pokeweed mitogen-stimulated secretion of immunoglobulins (Igs) A, G, and M by peripheral blood mononuclear cells in a dose-dependent manner, whereas sulfapyridine had little effect. Sulfasalazine and 5-aminosalicylic acid also inhibited spontaneous secretion of IgA by intestinal mononuclear cells, but sulfapyridine did not. Sulfasalazine inhibited pokeweed mitogen-stimulated lymphocyte proliferation, while 5-aminosalicylic acid and sulfapyridine exhibited minimal inhibition. Sulfasalazine was toxic for peripheral blood mononuclear cells, whereas 5-aminosalicylic acid and sulfapyridine were not toxic. Thus, the inhibition of antibody secretion by sulfasalazine was due to direct toxicity. On the other hand, 5-aminosalicylic acid, the therapeutically active component of sulfasalazine, was neither toxic nor antiproliferative, and appeared to exert its effects on metabolic pathways directly related to antibody synthesis. The calculated ID50 values of 5-aminosalicylic acid for antibody secretion were 1.35 mM for IgA and 1.05 mM for IgG, concentrations that are achieved in the colons of treated individuals. Indomethacin did not inhibit antibody secretion at pharmacologically relevant concentrations. 5-Aminosalicylic acid mediated inhibition of antibody secretion may play a role in inflammatory bowel disease by stopping antibody-mediated memory events involved in the induction or perpetuation of the disease process.

Human lamina propria lymphocytes bear homing receptors and bind selectively to mucosal lymphoid high endothelium.

It has been hypothesized that the selective recognition of tissue-specific endothelial cell molecules helps determine the in vivo distribution of lymphoid effector cells by controlling the extravasation of their circulating precursors. Here we report (a) immunofluorescence studies of the cell surface phenotype of human lamina propria lymphocytes (LPL), including staining with monoclonal antibody Hermes-1, which defines a 90-kDa lymphocyte surface glycoprotein involved in recognition of high endothelial venules (HEV); and (b) functional analyses of the ability of LPL to bind to HEV in frozen sections of mucosal lymphoid tissues (appendix or Peyer's patch) vs. peripheral lymph nodes. Essentially all LPL bear the Hermes-1 antigen, over 90% at levels comparable to those of circulating PBL. As a population, LPL display a quantitative preference for adherence to mucosal HEV, binding 0.8-1.5 times as well as PBL to mucosal HEV, but only 0.1-0.5 times as well to HEV in peripheral lymph nodes. Of particular interest was the behavior of the lymphoblast fraction, which typically constituted 3-7% of LPL. These cells, defined by size, consisted of a mixture of T cells and surface IgA+ blasts. One hundred percent were Hermes-1 bright, and they bound 4-8 times more efficiently to mucosal HEV than PBL while failing to bind detectably to lymph node HEV. LPL binding to mucosal HEV involves the gp90 Hermes, since the monoclonal anti-gp90 antibody, Hermes-3, and a polyclonal anti-gp90 antiserum inhibit the binding of small LPL and of LP blasts. The remarkable efficiency and specificity of binding by LP blasts may reflect retention of homing properties of the blood-borne precursors of these blasts and is discussed in relation to the capacity of immunoblasts in mesenteric nodes and in thoracic duct lymph to traffic selectively to mucosal lymphoid and extralymphoid sites. The demonstration of organ-specific endothelial cell recognition by LP lymphoblasts provides considerable support for the concept that selective interactions with endothelium play an important role in directing the distribution of activated lymphocyte subsets in vivo.

Antibody secretion by human intestinal mononuclear cells from normal controls and inflammatory bowel disease patients.

Inflammatory bowel disease (IBD) intestinal mononuclear cells (MNC) exhibit decreased spontaneous IgA secretion with an increased percentage of monomeric IgA and IgA subclass 1 in both ulcerative colitis and Crohn's disease patients. When compared with control intestinal MNC, a marked increase in spontaneous secretion of IgG is observed from IBD MNC. The greatest increase in spontaneous IgG secretion is seen with ulcerative colitis intestinal MNC, due to the secretion of large amounts of IgG subclass 1. Crohn's disease intestinal MNC have increased IgG subclass 2 secretion. Similar differences in IgG subclass concentrations also occur in the sera of active, untreated, IBD patients. Therefore, major alterations occur with regard to spontaneous antibody secretion of IgA and IgG subclasses in IBD. Because intestinal MNC comprise a unique immunologic compartment, it will be important to better understand the regulatory mechanisms, effector capabilities, and inducing antigens involved in intestinal IgA and IgG subclass secretion in IBD.

Bovine milk IgG, but not serum IgG, inhibits pokeweed mitogen-induced antibody secretion by human peripheral blood mononuclear cells.

Bovine milk IgG markedly inhibits the pokeweed mitogen (PWM)-induced secretion of immunoglobulins from human peripheral blood mononuclear cells. Heat-aggregated bovine milk IgG is even more inhibitory, demonstrating significant inhibition when levels as low as 5-9 micrograms/ml are continuously present in the in vitro 14-day culture system. However, bovine serum IgG, regardless of its state of aggregation, and control proteins have little effect on PWM-induced secretion of human IgG, IgA, and IgM. In a similar fashion, goat milk IgG, especially when aggregated, inhibits human antibody secretion whereas goat serum IgG does not. Inhibition appears to be mediated by Fc gamma receptors since F(ab')2 fragments of milk-derived bovine IgG do not inhibit PWM-induced antibody secretion. The continuous presence of bovine milk IgG is not essential since preincubation of milk IgG with PWM and human mononuclear cells for 24 hr also results in inhibition of human immunoglobulin secretion. In examining potential mechanisms of inhibition, it was found that bovine milk IgG, bovine serum IgG, and another chitin-containing protein, bovine thyroglobulin, each caused a small and equal inhibition of the binding of 125I-labeled PWM to human mononuclear cells, yet only the milk IgG inhibited antibody production. These studies raise the question of whether bovine milk IgG might modulate the human immune system in vivo.

Spontaneous secretion of IgG subclasses by intestinal mononuclear cells: differences between ulcerative colitis, Crohn's disease, and controls.

Spontaneous IgG and IgG subclass secretion patterns by isolated intestinal mononuclear cells (MNC) from control and inflammatory bowel disease (IBD) specimens were examined. Intestinal MNC from IBD specimens spontaneously secreted more total IgG than did control intestinal MNC. This increased spontaneous IgG secretion by ulcerative colitis intestinal MNC was primarily due to markedly increased production of IgG1. Slightly increased secretion of IgG3, but not IgG2 by ulcerative colitis intestinal MNC was present when compared with control and Crohn's disease intestinal MNC. In contrast, Crohn's disease intestinal MNC exhibited increased spontaneous secretion of all the IgG subclasses examined, with IgG2 being predominant.

Altered patterns of secretion of monomeric IgA and IgA subclass 1 by intestinal mononuclear cells in inflammatory bowel disease.

Intestinal mononuclear cells (MNC) from patients with inflammatory bowel disease (IBD) demonstrated altered patterns of spontaneous secretion of immunoglobulin A (IgA). Control intestinal MNC had markedly high spontaneous secretion of IgA compared to IBD intestinal MNC. Control intestinal MNC secreted predominantly dimeric IgA (only 31% monomeric IgA), whereas IBD intestinal MNC secreted increased amounts (43%-53%) of monomeric IgA. Control intestinal MNC secreted 61% IgA subclass 1 (IgA1), whereas IBD intestinal MNC secreted 71%-74% IgA1. Intestinal MNC from involved portions of resected specimens secreted more of both monomeric IgA and IgA1 than MNC from uninvolved areas of the same bowel. Therefore, intestinal MNC from involved IBD intestinal specimens secrete less total IgA but high percentages of monomeric IgA and IgA1 compared to control intestinal MNC. This could be caused by increased homing of monomeric IgA- and IgA1-producing cells into involved intestine, or the in situ proliferation of monomeric IgA and IgA1 precursor cells resulting from immunoregulatory alterations. These observations may represent the normal mucosal IgA immune response to infectious agents or inducing factors of either a primary or a secondary nature.

Inhibition of cell-mediated cytotoxicity by 2-cyclohexene-1-one: evidence for a role for glutathione and/or glutathione-protein interactions in cytolysis.

In order to explore the role of glutathione in cell-mediated cytotoxicity, we have examined the effect of the sulphydryl-reactive and glutathione-depleting agent 2-cyclohexene-1-one on antibody-dependent cellular cytotoxicity, spontaneous cell-mediated cytotoxicity, and cell-mediated lympholysis by human peripheral blood mononuclear cells. 2-Cyclohexene-1-one significantly inhibited (P less than 0.001) both antibody-dependent and spontaneous cell-mediated cytotoxicity using three different cell-line targets, at three different killer:target cell ratios (10:1, 25:1 and 50:1). Using K-562 cell-line targets, spontaneous cell-mediated cytotoxicity was inhibited by 2-cyclohexene-1-one with an ID50 of 0.71 X 10(-4) M-1.48 X 10(-4) M, while antibody-dependent cellular cytotoxicity was less sensitive to inhibition, and required slightly higher concentrations of 1.48 X 10(-4) M-3.98 X 10(-4) M to achieve 50% inhibition. Similar results were seen with human colon tumour cell-line and Chang liver cell-line cells as targets. Maximal inhibition occurred when 2-cyclohexene-1-one was added to the cytotoxicity assay 60 min prior to, at the start of, or within the first 60 min of a 4-hr assay; inhibition of cytotoxicity occurred with pretreatment of effector cells; and no inhibition of cytotoxicity was observed with pretreatment of target cells. Both the allogeneic mixed leucocyte reaction and cell-mediated lympholysis were also significantly inhibited (P less than 0.001) by 2-cyclohexene-1-one. These studies demonstrate that 2-cyclohexene-1-one is an effective inhibitor of cell-mediated cytotoxicity and suggest that glutathione, specific glutathione-protein interactions, or protein-bound sulphydryl groups are involved in allowing cells to carry out cytolysis.

Suppression of in-vitro antibody production by pokeweed mitogen-stimulated human bone-marrow mononuclear cells. Evidence for a soluble suppressor substance.

The spontaneous synthesis and secretion of immunoglobulin by human bone-marrow mononuclear cells (MNC) in vitro, as well as its suppression by the addition of pokeweed mitogen (PWM), were previously reported by this laboratory. In the present study we demonstrate that this suppression is mediated by a soluble substance elaborated by marrow MNC stimulated with PWM. Marrow MNC were pulsed for 1 hr with PWM, washed and recultured for 7 days in media without PWM. The culture supernatants were collected by centrifugation and filter sterilized before addition to fresh marrow MNC in the in vitro antibody synthesis assay. The 14-day assay culture supernatants were then subjected to a solid phase radioimmunoassay to determine the immunoglobulin content. The suppressor substance was non-specific as to immunoglobulin isotype and was not genetically restricted. Suppressor activity was diminished by heating the supernatants at 56 degrees for 1 hr. The activity could be elaborated by cells subjected to 1000 R or 2000 R before or after 1-hr incubation with PWM. While the addition of PWM anytime during the culture period would suppress IgA production at the level produced up to that time, the suppressor substance only suppressed IgA production when added during the first 4 days of culture. The addition of indomethacin had no effect on the suppressor activity indicating that the activity was not mediated by prostaglandin. Including human fibroblast interferon or hydrocortisone in the assay cultures had no effect on IgA production or its suppression by PWM. We concluded that the lectin-induced suppression was mediated by a marrow-derived suppressor substance (MDSS).

Synthesis and secretion of IgA, IgM, and IgG by peripheral blood mononuclear cells in human disease states, by isolated human intestinal mononuclear cells, and by human bone marrow mononuclear cells from ribs.

We have examined the secretion of IgA, IgM, and IgG by isolated human intestinal MNC, human bone marrow MNC from rib specimens, and peripheral blood MNC from patients with CD, UC, SLE, and HSP. "Normal" control intestinal MNC exhibited high spontaneous secretion of IgA, whereas intestinal MNC from UC and CD patients exhibited only modest increases in IgA secretion. Peripheral blood MNC from patients with CD, UC, SLE, and HSP exhibited markedly elevated spontaneous secretion of immunoglobulins in general and IgA in particular. Pure human bone marrow MNC exhibited high spontaneous secretion of IgA with modest amounts of IgG and normal IgM being secreted. The addition of PWM to cultures in which high spontaneous synthesis and secretion of immunoglobulins was seen, resulted in no further enhancement, and in some instances suppression, of antibody secretion. In patients with autoimmune disease, there appeared to be dual immunoregulatory defects, one involving a lack of normal T-suppressor cell functional capabilities for spontaneous antibody synthesis, and the other the presence of PWM activable T-suppressor cells. In human bone marrow, we have identified MNC that secrete suppressor factors in the presence of PWM and that are capable of inhibiting antibody synthesis and secretion. Column separation using Sephacryl S-300 revealed that the IgA secreted by "normal" control intestinal MNC is predominantly dimeric, whereas the IgA secreted by human bone marrow MNC is predominantly monomeric. Furthermore, mucosal MNC from patients with CD and uninvolved intestine from patients with UC exhibited patterns similar to control intestinal MNC, being predominantly dimeric IgA with some monomeric IgA secreted. By contrast, intestinal MNC from patients with UC had a decreased proportion of dimeric IgA and increased proportion of monomeric IgA, thus indicating that IgA precursor B-cells may have migrated into the intestine from extraintestinal sites, or that the normal dimeric IgA-secreting cells in the intestine had begun secreting increased proportion of monomeric IgA as well. These studies indicate that homing patterns and/or immunoregulation of IgA-secreting cells are altered in human intestine, bone marrow, and autoimmune disease states.

Immunoglobulin synthesis by peripheral blood mononuclear cells in minimal change nephrotic syndrome.

The spontaneous and pokeweed mitogen (PWM)-induced immunoglobulin synthesizing activities of circulating mononuclear cells (MNC) from minimal change nephrotic syndrome systemic (MCNS) patients in relapse (N = 13) were compared with those of patients in remission (N = 9), patients with active systemic lupus erythematosus (SLE, N = 9), and healthy controls (N = 17). Cumulative amounts of IgM, IgG, and IgA secreted over a 12-day culture period were determined in a solid phase radioimmunoassay. Mean levels of spontaneous immunoglobulin production in control cultures did not exceed 370 ng/ml. In contrast unstimulated IgM, IgG, and IgA synthesis among MCNS patients in relapse averaged 588, 1258, and 2665 ng/ml, respectively. The majority of patients exhibited synthetic activities that equalled or exceeded those of patients with active SLE. Spontaneous immunoglobulin production declined by 80 to 97% in three patients restudied in stable remission. A fourth patient with frequent relapses maintained high rates of synthesis in remission as well as in relapse. PWM stimulation increased immunoglobulin production in patients in remission and controls but failed to increase or suppressed immunoglobulin secretion in SLE patients and patients in relapse. These results suggest that MNC from MCNS patients in relapse are reversibly activated in vivo.

Marked in vitro spontaneous secretion of IgA by human rib bone marrow mononuclear cells.

Mononuclear cells (MNC) isolated from bone marrow curettaged from human ribs spontaneously produced significant amounts of Iga in 14-day in vitro culture. When culture supernatants were tested for each immunoglobulin isotype by radioimmunoassay at days 7, 14, 21, and 28, the peak of spontaneous IgA production was reached by day 14. Culturing marrow MNC in the presence of the polyclonal activator pokeweed mitogen (PWM) ablated this spontaneous IgA activity, and did not stimulate the production of IgG or IgM. In contrast, peripheral blood MNC produced significant amounts of all three immunoglobulin isotypes studied in the presence of PWM, but minimal immunoglobulin spontaneously. PWM stimulated both marrow and peripheral blood MNC to proliferate as demonstrated by increased DNA synthesis. In view of the opposite effects of PWM on immunoglobulin production by peripheral blood and bone marrow MNC, PWM may be stimulating a different functional or developmental subpopulation of cells in each case. We conclude that the bone marrow is a major site of IgA production in the human, that cells within the bone marrow have a blastogenic response to PWM in vitro, and that spontaneous production of IgA by human marrow MNC is inhibited by PWM.