Ultrastructural localization of nonspecific esterase activity in guinea pig and human monocytes, macrophages, and lymphocytes

Using either hexazotized pararosaniline or new fuchsin as coupling agents, we investigated the ultrastructural localization of alpha- naphthyl acetate esterase (ANAE) activity in guinea pig bone marrow and peritoneal exudates, and in human peripheral blood cells. DFP- inhibitable ANAE activity was present on the cell surface of lymphocytes, monocytes, macrophages, neutrophils, eosinophils, basophils, megakaryocytes, platelets, and blasts. Demarcation lines in megakaryocytes and the perinuclear cisternae in normoblasts were also positive. In addition, lymphocytes, monocytes, and macrophages displayed ANAE activity associated with cytoplasmic-vesicle clusters (CVC). Reaction product was always present on the cytoplasmic surfaces of these vesicles and in the adjacent cytoplasm; vesicle interiors were invariably ANAE-negative. Small lymphocytes generally had a single large paranuclear ANAE-positive CVC, whereas mononuclear phagocytes had multiple discrete foci of similar appearing ANAE-positive CVC that sometime became confluent. ANAE activity was also found in the Gall bodies of human lymphocytes and in coated vesicles of macrophages. Cytoplasmic ANAE activity was increased in oil-induced guinea pig peritoneal macrophages. Both surface and cytoplasmic esterase activities had a neutral pH optimum. An identical distribution of reaction product was observed when alpha-naphthyl butyrate was employed as substrate. The function of these esterases, and their relation to known surface and cytoplasmic neutral proteases, awaits further investigation.     

Large granular lymphocytes in human peripheral blood: ultrastructural and cytochemical characterization of the granules

Large granular lymphocytes (LGL) are defined as nonadherent mononuclear cells with cytoplasmic azurophilic granules, avid receptors for the Fc portion of IgG, and cytotoxic functions (NK or ADCC activities). In the present study, the granules of LGL isolated from human peripheral blood have been analyzed by enzyme cytochemistry and electron microscopy. It had been found that: (1) in the single cells, granules at different stages of maturation could be detected: in addition, packaging of the granules took place in the proximity of the Golgi apparatus, which is similar to that seen in secretory cell types. (2) Acid phosphatase (AP) was observed within the granules and the vesicles located in the Golgi area: the Golgi apparatus identified through its thiamine pyrophosphatase-positivity was consistently negative for AP. (3) Alpha naphthyl-acetate esterase (ANAE) activity was localized in the granules as well as on the membrane of LGL and monocytes. (4) The ANAE activity of LGL was of the monocytic and not of the lymphocytic type, as shown by NaF inhibition. (5) The LGL granules, although identifiable as primary lysosomes, were not involved in the process of phagocytosis, since LGL failed consistently to ingest latex particles or opsonized red cells.     

Ultrastructural localization of acid alpha-naphthylacetate esterase in human normal and neoplastic lymphocytic and monocytic cells and in hairy cells

The ultrastructural localization of acid alpha-naphthylacetate esterase (ANAE) activity was studied in normal and neoplastic human monocytic and lymphocytic cells including hairy cells. Normal monocytes and malignant monocytic cells from cases of AML-M4 and -M5 (FAB-classification) displayed fundamentally the same ANAE staining pattern with reaction products localized to the external surface of the plasma membrane and to small vesicles close to this membrane, more rarely to larger intracytoplasmic vesicles containing endocytosed material or cellular debris. The enzyme activity of the neoplastic cells increased with morphological differentiation. Certain normal blood lymphocytes and T cell-derived CLL cells showed the reaction product associated with paranuclearly located clusters of vesicular structures, extending from the membranes into the adjacent cytoplasm. Gall bodies were often found in the vicinity and were invariably positive for ANAE. Now and then, plasma membrane activity was present, but never as pronounced as in monocytes. Hairy cells demonstrated a pattern of reaction very similar to that of monocytic cells, whereas B cell-derived CLL cells and lymphoblasts of T- and common type ALL were generally non-reactive.     

Identification of ceruloplasmin receptors on the surface of human blood monocytes, granulocytes, and lymphocytes

Membrane receptors for ceruloplasmin (CP) were identified on all human blood leukocytes (granulocytes, monocytes, and lymphocytes) by a visual probe and 125I-CP binding. To synthesize the visual probe, amide-modified submicron-sized polystyrene latex minibeads, activated with glutaraldehyde, were covalently bound to CP. Incubation of this probe with human leukocytes, either fractionated or unfractionated, led to its binding, which was visualized on individual cells by using electron microscopy. At 4 degrees C, only surface binding occurred, but internalization also occurred at 37 degrees C. The binding was completely inhibited in the presence of excess nonderived CP, indicating the specificity of the binding. Incubation of fractionated leukocytes with 125I-CP also led to its specific binding to all three fractions. Scatchard analysis indicated the highest number of receptors for granulocytes and the lowest for lymphocytes. The binding affinity was lowest, however, for granulocytes, with monocytes showing the highest affinity. These data, indicating active uptake of CP by blood leukocytes, may reflect the requirement of leukocytes for copper that can be derived from CP. CP may also serve other functions within the cells.     

Size distribution, electronic recognition, and counting of human blood monocytes

During a study on the separation of human blood monocytes from lymphocytes, a method was developed to recognize and count monocytes by electronic means. Lightscattering (Cytograf, Bio/Physics), and changes in electrical resistance (Channelyzer, Coulter) were used to size mononuclear leukocytes directly in cell suspensions. Both methods revealed a size distribution profile in which two populations of mononuclear leukocytes could be distinguished. The largest cells were virtually eliminated after phagocytosis of iron particles. We confirmed that these cells were monocytes by three different criteria: the intracellular lysozyme activity, the number of phagocytes, and the percentage of cells with kidney-shaped nuclei. The highly significant correlations we found showed that monocytes could be recognized and counted by electronic sizing. For this method, purified mononuclear leukocyte preparations had to be used, since the presence of erythrocytes, platelets, and polymorphonuclear cells interfered. Statistical analysis revealed that electronic sizing permitted discrimination of differences in monocyte content of 4.5%, with a probability of 95%. It was calculated that this sensitivity of electronic monocyte counting was about three times higher than the sensitivity of microscopic methods. Since 100,000 cells can be sized within a few seconds, not only the efficiency of the preparation but also minor changes in the size of monocytes and lymphocytes introduced during the isolation can be followed.     

Acid hydrolase activities in B cell chronic lymphocytic leukaemia lymphocytes: correlation of cytochemical reactions with immunological phenotype

The cytochemical reactions of 5 acid hydrolases, alpha-naphthyl acetate esterase (ANAE), acid phosphatase (AP), beta-glucuronidase, beta-glucosaminidase and dipeptidylaminopeptidase IV (DAP IV) were investigated in lymphocytes from 30 patients with B cell chronic lymphocytic leukaemia (B-CLL). Based on ANAE and AP reactivities, 4 cytochemically distinctive subgroups were identified: Group 1: AP and ANAE less than 50% positive lymphocytes (5 cases); Group 2: AP greater than 50%, ANAE less than 50% positive lymphocytes (11 cases); Group 3: AP less than 50%, ANAE greater than 50% positive lymphocytes (7 cases); Group 4: AP and ANAE greater than 50% positive lymphocytes (7 cases). beta-Glucuronidase displayed similar patterns of reactivity to AP. beta-Glucosaminidase activity was observed in the majority of lymphocytes in most patients, whereas DAP IV activity was present in less than 20% of lymphoid cells. The study failed to establish any relationship between cytochemical grouping and patients' clinical status, peripheral lymphocyte counts, E or mouse rosette values, light or heavy chain cellular immunoglobulin (Ig) class. Attempts to correlate acid hydrolase and Ig heavy chain isotype expression, putative markers of B cell maturation, were unsuccessful and indicate that within the narrow spectrum of B cell differentiation seen in B-CLL these characteristics are unrelated.     

Induction of TNF-alpha and MnSOD by endotoxin: effect of E5531, a synthetic non-toxic lipid A analog

E5531, a synthetic lipid A analog, has been shown to inhibit endotoxin (lipopolysaccharide, LPS)-induced tumor necrosis factor-alpha (TNF-alpha) production by human monocytes and murine macrophages. Whether it also inhibits LPS induction of manganese superoxide dismutase (MnSOD) is not clear. In the current study, we demonstrated that E5531, while having no effect on TNF-alpha and MnSOD mRNAs by itself, markedly inhibited LPS- and lipid A-, but not TNF-alpha-, induced increases in TNF-alpha and MnSOD mRNAs in human monocytes. In contrast, E5531 at concentrations and conditions that markedly inhibit LPS-induced increases in TNF-alpha and MnSOD mRNAs, and TNF-alpha production by human monocytes, had no effect on murine peritoneal macrophages. These results demonstrate that E5531 is a potent LPS antagonist in human monocytes. However, it does not show antagonist action against LPS in murine macrophages in the range of concentrations tested, suggesting that E5531 is a more potent antagonist in humans than in mice.     

Effects of thyrotropin and cholera toxin on the thyroidal adenylate cyclase-adenosine 3',5'-monophosphate system

The present experiments examined the relationship between cholera toxin and TSH stimulation of the adenylate cyclase system in bovine thyroid tissue. Preincubation of thyroid slices for 20 min at 4 C with a maximal concentration of cholera toxin (100 microgram/ml) did not impair the subsequent stimulation of cAMP by submaximal amounts of TSH (1 mU/ml) during a 5-min incubation at 37 C. Incubation of cholera toxin or TSH with mixed gangliosides, followed by the addition of thyroid slices resulted in inhibition of the cholera toxin but not the TSH stimulation of cAMP formation. Previous exposure of thyroid slices to TSH induced refractoriness to subsequent stimulation of cAMP formation by TSH, but the response to cholera toxin was unchanged. NAD is necessary for cholera toxin, but not TSH, stimulation of adenylate cyclase. In the absence of NAD, cholera toxin inhibited the effect of maximal concentrations of TSH and prostaglandin E1 on adenylate cyclase activity but had no effect on NaF stimulation. In the presence of NAD, the stimulation of adenylate cyclase activity of bovine thyroid plasma membranes by a maximal amount of TSH was not influeced by maximal amounts of cholera toxin. Cholera toxin had a biphasic action on the binding of [125I]iodo-TSH, with low concentrations enhancing and high concentrations inhibiting binding. TSH augmented the binding of [125I]iodo-cholera toxin over the range of 1-100 mU/tube. Cholera toxin at 10 microgram/ml maximally inhibited binding. In addition to the requirement for ribosylation of adenylate cyclase, the present results indicate that the mechanisms of action of TSH and cholera toxin on cAMP formation are different.     

Hydrogen peroxide and hydroxyl radical mediation of activated leukocyte depression of cardiac sarcoplasmic reticulum. Participation of the cyclooxygenase pathway

Human peripheral blood leukocytes, activated by phorbol myristate acetate, disrupt canine sarcoplasmic reticulum calcium transport, in vitro, by an oxygen-derived free radical mechanism. Activated leukocytes significantly depress Ca++ uptake activity and Ca++ -stimulated, Mg++ -dependent ATPase activity. The depression is completely inhibited by sodium-azide (0.1 mM) or the combination of superoxide dismutase (10 micrograms/ml) and catalase (10 micrograms/ml). Exogenous hydrogen peroxide (0.441-4.41 mM) uncoupled Ca++ uptake activity from ATP hydrolysis, and this effect was inhibited by catalase. Mannitol alone did not inhibit the effects of activated leukocytes, but superoxide plus mannitol (20-100 mM) resulted in normal ATPase activity, while Ca++ uptake remained depressed. In the presence of indomethacin and ibuprofen, activated leukocytes depressed Ca++ uptake and had no effect on ATPase activity. 2-Amino-methyl-4-t-butyl-6-iodophenol (MK-447) further depressed Ca++ uptake and partially inhibited the effect on ATPase activity. Indomethacin plus catalase completely inhibited the effects of activated leukocytes on cardiac sarcoplasmic reticulum. We conclude, first, that activated leukocytes depress canine cardiac sarcoplasmic reticulum Ca++ transport by an oxygen-free radical mechanism with the generation of hydrogen peroxide and hydroxyl radical. In addition to the classical membrane NADPH oxidase system, significant oxygen radical generation can occur through the cyclooxygenase pathway of arachidonic acid metabolism, and seems to be responsible for the generation of the hydroxyl radical.     

Separation of Lymphocytes, Polymorphonuclear Leukocytes and Monocytes on Glass Columns, Including Tissue Culture Observations

A procedure was described for the separation of lymphocytes, polymorphonuclear (PMN) leukocytes and monocytes on Garvin’s glass bead columns. Dextran-sedimented leukocyte suspensions were added to columnsand incubated at 37 C. Lymphocytes were eluted with fresh serum. Thecolumns were then washed completely free of erythrocytes, platelets andlymphocytes while PMN leukocytes and monocytes continued to adhere. PMNleukocytes and monocytes were released from the glass with EDTA. Monocytes appeared in the effluents somewhat later than the PMN leukocytes,permitting a separation.

A fresh serum, heat labile, Ca^{+ +}- and Mg^{+ +}-requiring PMN leukocyte and monocyte adherence promoting factor was demonstrated andfound to be essential to the procedure.

The viability of column-separated cells was shown by their non-stainingwith trypan blue, motility, phagocytic ability, oxygen consumption, andsurvival or development in tissue culture.

In tissue culture, macrophages developed only from monocytes, whereasNowell’s blast-like, dividing, phytohemagglutinin cells were produced onlyin cultures containing lymphocytes.

Submitted on October 3, 1963 Accepted on December 12, 1963     

Stimulation of cAMP accumulation and superoxide production in human neutrophils and monocytes

The effect of sodium fluoride (NaF) on superoxide generation and cyclic adenosine monophosphate (cAMP) levels in human neutrophils and monocytes was investigated. NaF (greater than 10 mM) stimulated superoxide (O2-) production in both cell types in a time dependent manner. NaF (0.5 to 20 mM) increased cAMP levels by 1.5- to 3.-fold in both neutrophils and monocytes. Increases in cAMP levels were time-dependent; the maximal level was attained within 5 minutes after the addition of NaF, and cAMP levels remained elevated for up to 10 minutes. Only high concentrations of NaF (10 and 20 mM) increased both cAMP levels and O2- production. Therefore, a direct role of cAMP in O2- generation is not likely. It is speculated that since NaF (greater than 10 mM) can complex with extracellular Ca++, and thus reduce free Ca++ concentration required for O2- generation, a NaF-dependent increase in cAMP may restore cytosolic free Ca++ by mobilizing intracellular stores of Ca++. Further, in view of the proposed involvement of a phosphorylation-dephosphorylation mechanism in the regulation of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, we speculate that NaF, by inhibiting phosphoprotein phosphatase activity, may indirectly activate the NADPH oxidase system and thus superoxide generation.     

Effects of calcium-regulating hormones and drugs on human monocyte chemiluminescence

Chemiluminescence (CL) associated with phagocytosing monocytes has been used as an index of the oxygen dependent metabolic activity of these cells. Because of the relationships between monocytes and cells involved in bone resorption, we studied the effects on human monocyte CL of hormones and drugs active in mineral metabolism. Two-hour preincubations of monocytes with human PTH-(1-34), bovine PTH, or prostaglandin E2 caused significant decreases in peak CL during phagocytosis stimulated by the addition of latex particles. Similar studies with dichloromethylene diphosphonate, ethane hydroxydiphosphonate, or salmon calcitonin (sCT) caused significant increases in CL, whereas preincubation with human CT at the same molar concentration did not. CL was also decreased after preincubations with methylprednisolone or bacterial endotoxin. The effect of bovine PTH was dose dependent to concentrations as low as 10 ng/ml, was not fully present after a shorter 1-min preincubation with the hormone, and differed in an otherwise identical system using polymorphonuclear leukocytes instead of monocytes. The production of hydrogen peroxide by phagocytosing monocytes was also significantly affected by each of the drugs and hormones. The direction and magnitude of these changes were similar to those in CL experiments, except for the effects of sCT. These studies relate the oxygen-dependent function of phagocytes to mediators of bone resorption and provide a new system for studying the effects of hormones and drugs on the cellular elements of bone and blood.     

Peritoneal fluid and plasma levels of human macrophage colony-stimulating factor in relation to peritoneal fluid macrophage content

The peritoneal fluid (PF) of women with infertility (especially in the presence of endometriosis) contains increased numbers of leukocytes, 90% to 95% of which are macrophages. The high numbers of peritoneal macrophages presumably result from an influx of blood monocytes into the peritoneum, and/or from local proliferation of peritoneal macrophages. Once in the peritoneal cavity, monocytes differentiate into tissue macrophages. Mononuclear phagocyte proliferation and differentiation are influenced by different cytokines, including macrophage colony-stimulating factor (M-CSF). The purpose of this study was to determine the relationship of M-CSF levels in human PF and plasma to the macrophage content, and to the patient diagnoses. Mean concentrations of PF M-CSF were higher than plasma levels (2.44 +/- 0.13 v 0.95 +/- 0.06 ng/mL, respectively). The mean concentrations of plasma M-CSF did not differ in samples from women of different diagnostic groups (normal, peritoneal adhesions, endometriosis, inactive pelvic inflammatory disease, uterine fibroids, and idiopathic infertility), but the PF concentration was slightly higher in normal women. The absolute (total) amount of PF M-CSF in normal women was lower than in those of the other diagnostic groups. The total amount of PF M-CSF in all women correlated closely with the total number of peritoneal macrophages. The tubal patency status (open versus closed) did not influence the plasma and PF concentrations of M-CSF, nor the PF absolute amount of M-CSF. The PF M-CSF may have come from peritoneal macrophages, fibroblasts, mesothelial cells, or endothelial cells. PF M-CSF may play important roles in the proliferation and/or the differentiation of peritoneal mononuclear phagocytes.     

Phagocytosis of Candida albicans by human leukocytes: opsonic requirements.

The kinetics of phagocytosis of Candida albicans by human polymorphonuclear leukocytes was studied. The basis for these studies was a phagocytic assay with use of C. albicans radiolabeled with [3H]adenine. After incubation of leukocytes with C. albicans, extracellular C. albicans was separated from phagocytes by centrifugation through Ficoll-Hypaque suspensions (specific density, 1.175 g/cm3). Recovery of leukocytes by this technique was greater than or equal to 85%. The initial rate of phagocytosis was more rapid than that previously reported for bacteria. Ethylenediaminetetraacetate, vinblastine, ethylmaleimide, NaF, and ice bath temperature completely inhibited phagocytosis. Colchicine had no effect, and NaN3 was partially inhibitory. Pooled sera possessed low titers (greater than or equal to 1:40) of heat-stable opsonins. The opsonic activity of pooled sera was shown to depend primarily upon complement activated through both the alternative and classical pathways. Decomplemented hyperimmune sera were opsonic at high dilutions (greater than or equal to 1:160), and complement amplified the initial rate of ingestion seen with hyperimmune sera.     

Lactoferrin receptors in normal and leukaemic human blood cells

The binding of 125I-lactoferrin to a variety of peripheral blood cells was examined. In the concentration range from 10(-9) mol/l to 10(-6) mol/l a receptor-like binding of lactoferrin was observed in monocytes as well as in polymorphonuclear leucocytes. At the low concentrations of lactoferrin in plasma (about 10(-9) mol/l) the cellular binding to monocytes was about 10 times higher than the binding to polymorphonuclear leucocytes and lymphocytes. A major result from a kinetic analysis was a lower apparent binding affinity to polymorphonuclear leucocytes (KD about 2 X 10(-7) mol/l) than to lymphocytes and monocytes (KD about 2 X 10(-8) mol/l). Studies in leukaemic cells showed that lymphocytes from patients with chronic lymphocytic leukaemia bound lactoferrin to the same small extent as normal lymphocytes. In contrast, a larger component of binding with high affinity (KD about 2 X 10(-8) mol/l) could be demonstrated to lymphoblasts as well as to myeloblasts.     

Cytochemical Profile of Human Haematopoietic Biopsy Cells and Derived Cell Lines

Twenty-three human haematopoietic cell lines, normal and mitogen stimulated peripheral blood lymphocytes and tumour material from fresh leukaemias, myelomas and lymphomas were investigated with a panel of cytochemical reactions. Normal and mitogen stimulated lymphocytes, non-neoplastic lymphoblastoid cell lines (LCL), lymphoma lines with B-lymphocyte characteristics, chronic lymphocytic leukaemia and fresh lymphocytic lymphomas reacted weakly or negatively with all stains. T-lymphocyte acute leukaemia lines were PAS and alpha-naphtyl acetate esterase positive. Myeloma lines and fresh myelomas were strongly beta-glucoronidase positive. A histiocytic lymphoma cell line was strongly esterase positive with naphtol AS-D acetate esterase inhibited by NaF. The three fresh histiocytic lymphomas, however, reacted as the lymphocytic lymphomas suggesting a lymphoid origin. A myeloid leukaemia line was strongly positive for acid phsophatase. No major disagreement was noted between the reactivity of established neoplastic lines and the corresponding fresh biopsy cells indicating an unaltered qualitative expression of enzyme production after prolonged in vitro culture.     

Hormone receptors: VI. On the nature of the binding of glucagon and insulin to human circulating mononuclear leukocytes.

Several characteristics of the binding of insulin and glucagon to human circulating mononuclear leukocytes have been studied. Functional analysis (latex bead ingestion) revealed that cell mixtures, as prepared according to Boyum and used generally in studies of insulin resistance in humans, consist of 20-29% phagocytic monocytes, with the remainder being lymphocytes. Partial separation of monocytes from lymphocytes on columns of Sephadex G-10, followed by correlation of insulin binding with cell type, confirms that the monocyte is the binding species. Insulin influenced neither glucose uptake nor the further conversion of glucose to lipids and CO2 by the leukocytes. The transport of alpha-aminoisobutyrate, a nonmetabolizable amino acid, into these cells was also unaffected by insulin. Monocyte/lymphocyte mixtures specifically bound glucagon and prostaglandin E1. At physiological concentrations of these hormones, steady states were reached in 15 min and 45 min, respectively. In contrast to the 8-10-fold increases in cellular cyclic AMP produced by prostaglandins, the effect of glucagon was very small but apparently real. Under appropriate preincubation conditions, sodium azide and iodoacetamide inhibited phagocytosis and insulin binding in parallel. The binding of glucagon was unaffected by these agents. Although both antimycin A and actinomycin D inhibited phagocytosis of the monocytes, only the former inhibited insulin binding; there was only a slight effect on glucagon binding. We would conclude that the binding of insulin to human circulating monocytes, although reflective of insulin resistance in diabetes mellitus and obesity, may not be to traditional receptors. In contrast, the binding of glucagon to lymphocyte/monocyte mixtures may be to function-linked receptors.     

Human chronic lymphocytic leukaemia. Surface markers and activation of lymphocytes.

Cell surface markers and the responses of lymphocytes to T- and B-cell mitogens were studied in 10 patients with CCL. T cells were identified as cells rosetting with sheep red blood cells (SRBC), and S-Ig was used as a marker for B lymphocytes. Most cells from all patients had a detectable amounts of S-Ig, and the percentage of cells rosetting with SRBC was low in all cases. Of the lymphocytes from these patients, 3-74% (mean 33%) were positive for the acid esterase (ANAE), which has been claimed to be a T-cell marker. However, some patients had cells that were positive for both S-Ig and ANAE. Acid esterase staining is therefore not a valid T-cell marker in chronic lymphocytic leukaemia. In cultures containing the T-cell mitogen leucoagglutinin (LA) and the T- and B-cell mitogen pokeweed mitogen (PWM) the reactivity of the lymphocytes was low. The cells responded vigorously to the T- and B-cell mitogen protein A (PA); however, the response was serum-dependent, being strong in a culture medium containing foetal calf serum (FCS), but impaired in the presence of human AB serum. Only 1 patient had cells that responded to the B-cell mitogen LPS.     

Human Chronic Lymphocytic Leukaemia Surface Markers and Activation of Lymphocytes

Cell surface markers and the responses of lymphocytes to T- and B-cell mitogens were studied in 10 patients with CLL. T cells were identified as cells rosetting with sheep red blood cells (SRBC), and S-Ig was used as a marker for B lymphocytes. Most cells from all patients had detectable amounts of S-Ig, and the percentage of cells rosetting with SRBC was low in all cases. Of the lymphocytes from these patients, 3–74 % (mean 33 %) were positive for the acid esterase (ANAE), which has been claimed to be a T-cell marker. However, some patients had cells that were positive for both S-Ig and ANAE. Acid esterase staining is therefore not a valid T-cell marker in chronic lymphocytic leukaemia. In cultures containing the T-cell mitogen leucoagglutinin (LA) and the T- and B-cell mitogen pokeweed mitogen (PWM) the reactivity of the lymphocytes was low. The cells responded vigorously to the T- and B-cell mitogen protein A (PA); however, the response was serum-dependent, being strong in a culture medium containing foetal calf serum (FCS), but impaired in the presence of human AB serum. Only 1 patient had cells that responded to the B-cell mitogen LPS.