Albumin associated with purified pig lymphocyte plasma membrane.

Plasma-membrane preparations purified from pig lymphocytes contained a major polypeptide component of mol.wt. about 68 000. This component was identified as pig albumin by the following comparisons with authentic pig serum albumin: (a) co-migration when analysed by sodium dodecyl sulphate/polyacrylamide-gel electrophoresis under reducing and non-reducing conditions; (b) identical isoelectric points; (c) similar "fingerprints" of arginine-containing tryptic peptides; (d) reactivity with anti-(pig albumin) serum. The albumin was bound tightly to the plasma membrane. Biosynthetic labelling of pig lymphocytes under a variety of conditions failed to provide evidence that albumin was synthesized by lymphocytes, suggesting that the plasma-membrane-associated albumin was of extraneous origin. Radiolabelled pig serum albumin, however, failed to bind to the plasma-membrane fraction when added before cell disruption. Although lymphocyte plasma membrane preparations from other species possessed a polypeptide of about 68 000 mol.wt., this was judged not to be albumin on the basis of electrophoretic mobility under non-reducing conditions; also, no polypeptide was precipitated by anti-albumin sera. It is concluded that pig lymphocyte plasma-membrane preparations possess albumin which, although firmly attached, was probably of extraneous origin. This association appeared not to be common to lymphocytes from other species.     

Human T lymphocyte mitogenesis in response to the B oligomer of pertussis toxin is associated with an early elevation in cytosolic calcium concentrations

Pertussis toxin was found to serve as a mitogen in the human T lymphocyte, an effect which could be mimicked by its resolved binding component, the B oligomer. The mechanism of action of this component appeared to involve a rapid and sustained elevation of cytosolic calcium levels, as monitored by fura-2 fluorescence. The source of mobilized calcium was predominantly extracellular, suggesting that the binding of the B oligomer to the T cell plasma membrane in some way elicited calcium channel activation. Notably, the influx of calcium was not observed with cholera toxin, an AB toxin lacking mitogenic effects on the human T lymphocyte.     

Endotoxin-induced T lymphocyte proliferation

The lymphocyte response to endotoxin (LPS) has been attributed largely to the action of this agent as a polyclonal activator of B lymphocytes. In this study we found that a cloned murine interleukin 2-dependent cytotoxic T cell line, CT 6, proliferates in response to LPS, thus providing the first evidence that T cells can be stimulated directly by LPS. The response was dose and time dependent and was blocked by polymyxin B, an inhibitor of LPS-induced mitogenesis. The fact that this is a cloned T cell line, free of other potentially contaminating lymphoid cell types, precludes the possibility that this proliferation is due to contaminating B lymphocytes or is mediated by macrophage-derived products such as interleukin 1. Moreover, highly purified splenic T lymphocyte populations (purified by negative/positive selection or by a rigorous column purification procedure) contain a small subpopulation (approximately 3%) of T cells that proliferate in response to LPS. This population is missing in the endotoxin-hyporesponsive C3H/HeJ mouse. As was observed in the CT 6 line, proliferation of splenic T cells in response to LPS was inhibited by polymyxin B. Furthermore, treatment of LPS-stimulated T cells with anti-T cell antibodies plus complement blocks the uptake of 3H-thymidine by these cultures. Exogenous interleukin 1 failed to stimulate the T cell cultures comparably to LPS and therefore cannot account for the degree of stimulation observed. These findings support and extend previous findings that suggested a role for an endotoxin-sensitive T cell population in the induction of certain responses, such as LPS-induced adjuvanticity of the lymphocyte-dependent LPS induction of macrophage procoagulant activity.     

Conditions affecting the immunosuppressive properties of human alpha-fetoprotein.

The hypothesis that alpha-fetoprotein (AFP) is immunosuppressive in vitro was tested with immunoabsorbent purified human cord AFP in human lymphocyte cultures. Albumins were purified identically from cord plasma and pooled adult plasma. No preparations were suppressive in phytohemagglutinin-stimulated cultures. Both AFP and human cord albumin (HCA) produced 50% suppression of allogeneic lymphocyte cultures at 50 to 100 microgram/ml final concentrations, whereas adult and commercial albumins were not suppressive. When AFP was eluted from the immunoabsorbent column in 0.15 M NaCl rather than the conventional 0.5 M NaCl, activity was greatly diminished or lost, but could be restored by dialysis against 0.5 M KCl. Previously, inactive lots of HCA also demonstrated the same phenomenon. It appears, therefore, that the immunosuppressive activity of AFP (and albumin) may depend both on its source and the procedure by which it is isolated. Whether analogous conditions occur in vivo is unknown at this time.     

Characterization of a plasma membrane-resident albumin-binding protein associated with the proliferation of estrogen-target,serum-sensitive cells

Estrogens control the proliferation of their target cells through a receptor-mediated pathway. Recently presented evidence suggests that estradiol cancels the proliferative inhibition exerted by human albumin (HA) and recombinant human albumin (rHA) on estrogen-target serum-sensitive cells (indirect-negative hypothesis). We postulate that this mechanism requires the presence of a plasma membrane estrogen receptor (mER) and a plasma membrane albumin-binding protein (mABP). Direct evidence confirming the presence of mERalpha in MCF7 cells has recently been presented. Herein, we now show that Western blot analysis of purified T47D membrane proteins with the C542 ERalpha specific monoclonal antibody also revealed specific, multiple M(r) mERs (67, 110, and 130k M(r)). In addition, Western blot analysis with an ABP antiserum revealed a potential 60k M(r) ABP in both MCF7 and T47D plasma membrane extracts. No such evidence was observed in similar extracts from ER-negative, serum-insensitive MDA-MB231 cells. Ligand blot analysis of similar plasma membrane extracts with bovine serum albumin confirmed the presence of a 60k M(r) ABP in MCF7 and T47D cells; again, no such evidence was observed in comparable extracts from MDA-MB231 cells. Fluorescence and confocal microscopy of MCF7 cells fixed in 2.0% paraformaldehyde/0.1% glutaraldehyde identified specific membrane ABP antigenic sites by immunocytochemistry. Serum-insensitive MDA-MB231 cells fixed and labeled similarly did not exhibit this mABP. These results suggest that the proposed mABP is expressed only in serum-sensitive estrogen-target cells and is not expressed in cells insensitive to the proliferative inhibition of HA and rHA. Also, the present data suggest that the proposed mABP may be the recognition mechanism by which both HA and rHA inhibit MCF7 and T47D cell proliferation.     

Human macromolecular insoluble cold globulin (MICG). II. Immunologic definition of T cell and null cell MICG and the biologic effect of antiserum to MICG.

Thymus cell-derived macromolecular insoluble cold globulin (T-MICG) is a 225,000-dalton protein, selectively synthesized in human T cells. Null cell-derived macromolecular insoluble cold globulin (N-MICG) is a 185,000-dalton protein, synthesized in null cells, and antigenically distinct from T-MICG. Evidence to support these conclusions was provided by using isolated cell preparations that were radiolabeled, lysed in desoxycholate, and precipitated with monospecific antiserum to each component. These studies demonstrated that antiserum to T-MICG precipitated a 225,000 dalton protein from PBL and T cells, but not from B or null cells. Antiserum to N-MICG reacted with a 185,000 dalton protein present in PBL and null cells, but not with lysates from either T or B cells. The plasma membrane distribution of these proteins was shown by absorption of antiserum to T + N-MICG with either isolated T or null cells. Antibody-induced cytotoxicity and immunofluorescence confirmed the cell surface location of T and N-MICG. Divergent biologic effects of these antisera were also noted. Antiserum to T-MICG inhibited T cell rosette formation and the one-way mixed lymphocyte reaction, although anti-N-MICG antiserum had no such effect. The potential importance of these proteins is discussed.     

Constitutive production of a factor supporting B lymphocyte differentiation by a T cell hybridoma

To investigate the role of soluble T cell products during B cell differentiation more fully, we have produced T cell hybridomas by the fusion of normal helper T cells with the T cell lymphoma BW5147. In this report we describe the production of one such hybrid, 14G3, the subclone 14G3.1F2, and the functional activity of the constitutive product. The hybrid supernatant acts exclusively in antigen-nonspecific, but antigen-dependent, promotion of B cell differentiation. It is optimally effective in the presence of small amounts of EL4 supernatant. It does not itself contain any detectable IL 2 or BCGF or interferon activity, however. 14G3.1F2 activity is probably an important component of the conventional TRF preparations produced by mixed lymphocyte populations, and will be useful in further dissection of the contributions of different soluble T cell products to B cell differentiation.     

Distribution and partial purification of a liver membrane protein capable of inactivating cytosol enzymes.

1. The inactivation of cytosol enzymes in liver extracts was carried out by several subcellular fractions, with plasma membranes having the highest specific activity. Rough and smooth microsomal fractions were both active, whereas lysosmal inactivation capacity appeared to be derived entirely from contaminating plasma-membrane fragments. 2. Inactivation capacity in liver fractions was derived from parenchymal cells. Of the non-liver cells tested, plasma membranes from H35 hepatoma cells were able to inactivate glucose 6-phosphate dehydrogenase (EC 1.1.1.49), adipocyte "ghosts" showed slight activity and erythrocyte and reticulocyte "ghosts" were inactive. 3. Liposomes prepared from pure lipids with net negative, positive or neutral charge did not possess inactivation capacity. 4. Liver plasma-membrane inactivation capacity was destroyed by heating at 50 degrees C. 5. Inactivation factor solubilized from membranes by trypsin plus Triton X-100 treatment was partially purified by (NH4)2SO4 fractionation, gel filtration, ion-exchange chromatography and hydroxyapatite chromatography. 6. Partially purified inactivation factor analysed by gel electrophoresis gave a major protein band that co-migrated with capacity for inactivation of glucose 6-phosphate dehydrogenase. 7. It is concluded that inactivation factor is a membrane protein whose intracellular distribution and other properties are consistent with a possible role for this activity in the initial step of protein degradation.     

Specific antigen-binding and antibody-secreting lymphocytes associated with the erythrocyte autoantibody responses of NZB and genetically unrelated mice.

The receptor characteristics as well as incidence of antigen-binding lymphocytes (ABL) or B and T cell classes with membrane receptors specific for the exposed (X) and cryptic (HB) murine erythrocyte autoantigens were examined in NZB and nine control strains of mice. Whereas only NZB and NZB hybrid mice synthesize anti-X autoantibody pathogenetically implicated in the genetically determined autoimmune hemolytic anemia, the NZB as well as control strains synthesize the ubiquitous anti-HB anti-erythrocyte autoantibody. By utilizing immunocytoadherence assays, maximum numbers of specific ABL of both B and T lymphocyte classes were optimally demonstrated at erythrocyte:lymphocyte ratios of 20:1 and after lymphocyte fixation at 56 degrees C for 20 min. Surface membrane receptor specificity was established by inhibition with semi-purified soluble X or HB autoantigen. Inhibition of immunocyto-adherence with class specific antisera to mouse immuno-globulins demonstrated that the receptors on both B and T cells were of IgM class. Specific receptors regenerated in vitro after trypsinization which excluded the role of cytophilic antibody in the immunocytoadherence reactions. B lymphocyte ABL reactive with the X autoantigen were demonstrable in NZB, NZB hybrid, and control mice. Only in NZB and NZB hybrid mice, strains that uniformly synthesize anti-X autoantibody, were X ABL of T lymphocyte class demonstrated. The presence and incidence of T lymphocyte X ABL is compatible with the expression of a single dominant gene carried by the NAB strain. The incidence of B lymphocyte X ABL increased with age, suggesting proliferation of this cell population. HB ABL of both B and T lymphocyte classes were observed in all strains, concordant with the ubiquitous presence of humoral anti-HB autoantibodies. Differentiation of precursor B cells are evaluated by PFC assay of cells secreting specific autoantibodies. Anti-X PFC were observed only in NZB and NZB hybrid mice; and the observed frequency suggested that less than 3.5% of the specific ABL were differentiated for the secretion of anti-X autoantibody. Anti-HB PFC were observed in all strains and represented as high as 11.8% of specific ABL. Genetic determination of the anti-X anti-erythrocyte autoantibody response does not prescribe the presence of precursors of the antibody-forming cell, but rather appears to influence regulation of the differentiation of these cells. These data suggest that circumvention of immunologic tolerance to this specific erythrocyte autoantigen may occur at the level of the T lymphocyte; or alternatively, that T lymphocytes as well as B lymphocytes, are induced to proliferate and differentiate in the NZB strain.     

Characterization of the bovine lens plasma membrane substrates for cAMP-dependent protein kinase

cAMP-dependent protein kinase, derived from either calf lens or bovine heart, promotes the phosphorylation of three lens plasma membrane proteins of molecular mass 28 kDa, 26 kDa and 18 kDa. Correlation of the maximal level of phosphorylation of these components with the Coomassie blue staining intensity of fractionated lens membranes suggests that the phosphorylation of the 28 kDa and 18 kDa components may be approximately stoichiometric. The protein kinase substrates could be dephosphorylated by a cardiac sarcoplasmic-reticulum-bound protein phosphatase activity. The 26 k Da component comigrated with MP26, the major lens membrane component that has been localized to the lens fiber cell junction. Treatment of phosphorylated lens membranes with chymotrypsin did not suggest that any of the three major phosphorylated components was derived from the partial proteolysis of a larger phosphoprotein. After electrophoretic separation of phosphorylated proteins, treatment with N-chlorosuccinimide confirmed that there was little similarity in the structure of the three phosphoproteins. Chymotrypsin did, however, reveal a cryptic phosphorylation site in a 22 kDa fragment that appeared to be derived from MP26. Treatment of phosphorylated membranes with reducing agents resulted in the disappearance of the 28 kDa phosphorylated component and the appearance of a new phosphorylated component of 18 kDa; neither MP26 nor the original 18 kDa component was affected by such treatment. It is not clear whether the original 18 kDa phosphoprotein, present in unreduced samples, is the same as that generated with reducing agents from the 28 kDa phosphorylated lens membrane component.     

Effect of purified phospholipases on the binding of tetrodotoxin to axon plasma membrane

Summary The role of phospholipids in the binding of [³H] tetrodotoxin to garfish olfactory nerve axon plasma membrane was studied by the use of purified phospholipases. Treatment of the membranes with low concentrations of either phospholipase A₂ (Crotalus adamanteus andNaja naja) or phospholipase C (Bacillus cereus andClostridium perfringens) resulted in a marked reduction in tetrodotoxin binding activity. A 90% reduction in the activity occurred with about 45% hydrolysis of membrane phospholipids by phospholipase A₂, and with phospholipase C the lipid hydrolysis was about 60–70% for a 70–80% reduction in the binding activity. Phospholipase C fromB. cereus andCl. perfringens had similar inhibitory effects. Bovine serum albumin protected the tetrodotoxin binding activity of the membrane from the inhibitory effect of phospholipase A₂ but not from that of phospholipase C. In the presence of albumin about 25% of the membrane phospholipids remained unhydrolyzed by phospholipase A₂. It is suggested that these unhydrolyzed phospholipids are in a physical state different from the rest of the membrane phospholipids and that these include the phospholipids which are directly related to the tetrodotoxin binding component. It is concluded that phospholipids form an integral part of the tetrodotoxin binding component of the axon membrane and that the phospholipase-caused inhibition of the binding activity is due to effects resulting from alteration of the phospholipid components.     

Dynamic properties of ankyrin in T lymphocytes: colocalization with spectrin and protein kinase C beta

Ankyrin is a well characterized membrane skeletal protein which has been implicated in the anchorage of specific integral membrane proteins to the spectrin-based membrane skeleton in a number of systems. In this study, the organization of ankyrin was examined in lymphocytes in relation to T cell function. Light and electron microscope immunolocalization studies revealed extensive heterogeneity in the subcellular distribution of ankyrin in murine tissue-derived lymphocytes. While ankyrin can be localized at the lymphocyte plasma membrane, it can also be accumulated at some distance from the cell periphery, in small patches or in a single discrete, nonmembrane-bound structure. Double immunofluorescence studies demonstrated that ankyrin colocalizes with spectrin and with the signal transducing molecule protein kinase C beta (PKC beta) in tissue-derived lymphocytes, suggesting a functional association between these molecules in the lymphocyte cytoplasm. In addition, T lymphocyte activation-related signals and phorbol ester treatment, both of which lead to PKC activation, cause a rapid translocation of ankyrin, together with spectrin and PKC beta, to a single Triton X-100-insoluble aggregate in the cytoplasm. This finding suggests a mechanism for the reported appearance of PKC in the particulate fraction of cells after activation: activated lymphocyte PKC beta may interact with insoluble cytoskeletal elements like ankyrin and spectrin. Further evidence for a link between the subcellular organization of these proteins and PKC activity is provided by the observation that inhibitors of PKC activity cause their concomitant redistribution to the cell periphery. The dynamic nature of lymphocyte ankyrin and its ability to accumulate at sites distant from the plasma membrane are properties which may be unique to the lymphocyte form of the molecule. Its colocalization with PKC beta in the lymphocyte cytoplasm, together with its redistribution in response to physiological signals, suggests that structural protein(s) may play a role in signal transduction pathways in this cell type. Our data support the conclusion that ankyrin is not solely involved in anchorage of proteins at the plasma membrane in lymphoid cells.     

Studies on human porin. IV. The primary structures of "Porin 31HM" purified from human skeletal muscle membranes and of "Porin 31HL" derived from human B lymphocyte membranes are identical.

We report on the purification of "Porin 31HM" from the crude plasma membrane fraction of human skeletal muscle. Furthermore, all tryptic peptides of the molecule were purified and characterized by different methods. The alignment of the peptides with the complete primary structure of the human B lymphocyte plasma membrane-derived "Porin 31HL", published by us recently (Kayser, H. et al. (1989) this Journal 370, 1265-1278), proved both structures to be completely identical. Our data demonstrate that porin fractions from crude plasma membranes of different human cell types do not show any variation on the primary structure level.     

The precise radioimmunoassay of adenosine: minimization of sample collection artifacts and immunocrossreactivity.

Anti-adenosine antibodies were produced in rabbits immunized with N6-carboxymethyladenosine conjugated to methyl albumin. 125I-N6-Aminobenzyladenosine was synthesized and used as a high-specific-activity, high-affinity ligand. A radioimmunoassay (RIA) was developed that can detect 6.25 nM (312.5 fmol) of underivatized adenosine and cross-reacts less than 0.02% with adenine nucleotides and guanosine and not at all with 1 mM inosine. The sensitivity of the RIA can be increased to a detection limit of 0.125 nM (6.25 fmol) by derivitizing samples with benzyl bromide to form N6-benzyladenosine. The assay was adapted to an automated RIA procedure. Assay precision was increased by: (i) inhibiting slight adenosine deaminase activity present in anti-sera; (ii) treating buffers and albumin used in the RIA with charcoal to remove contaminating adenosine; and (iii) correcting for a small but variable component of immunoreactivity not attributable to adenosine. A second antibody prepared with a 2',3'-disuccinyladenosine-albumin conjugate was also found to detect some non-adenosine-mediated immunoreactivity in plasma samples. Immunointerference in human plasma was eliminated in samples treated with ZnSO4/Ba(OH)2 or partially purified over C18 Sep Paks to remove nucleotides and assayed after sample benzylation or succinylation. Human blood was mixed with a novel "stop" solution that was optimized to inhibit adenosine formation from AMP by greater than 99% and to inhibit adenosine uptake into red cells and degradation by greater than 94%. Human plasma/stop solution was assayed by RIA and HPLC with equivalent results.     

Cellular distribution of p68, a new calcium-binding protein from lymphocytes.

A Ca2+-binding protein of mol. wt. 68 000 ( p68 ) is a major component of a Nonidet P-40 insoluble fraction of human and pig lymphocyte plasma membrane. An affinity-purified rabbit antibody has been produced against p68 and used to study its cellular distribution. The antibody stained fixed and permeabilised human B lymphoblastoid cells, peripheral blood lymphocytes and sections of human tonsil. Whole cells, however, were not stained, indicating that the protein was not represented at the cell surface. This assignment was consistent with the detection of p68 in immunoprecipitates from biosynthetically- but not surface-labelled cells. It is concluded that p68 is located on the cytoplasmic face of the plasma membrane. Subcellular fractionation experiments confirmed that p68 was largely membrane-bound in lymphocytes, although a small soluble fraction (approximately 10% of the total) was detected. Sub-fractionation of lymphocyte membranes revealed that p68 was associated not only with the plasma membrane but also with other endomembrane systems. As judged by immunoprecipitation, p68 was present in a variety of cultured cell lines of both lymphoid and non-lymphoid origin. p68 demonstrated a diffuse distribution in fixed and permeabilised fibroblasts which did not correspond to the distribution of either microfilaments or intermediate filaments. However, in detergent-extracted cells the protein was localised in a lamina-like network. A similar immunofluorescent staining pattern has recently been observed for spectrin-related proteins in the detergent-resistant cytoskeleton of fibroblasts. It is suggested that p68 is part of a sub-membranous cytoskeletal complex not only in lymphocytes but also in other cell types.     

Lymphocyte membrane alpha-1-acid glycoprotein: a cellular synthesis during lymphocyte activation

Soluble alpha 1 acid-glycoprotein is considered an "acute phase protein" with an inhibitory effect on lymphocyte activity; it has recently been shown that a lymphocyte modulatory variant of alpha 1 acid-glycoprotein has a positive role on T cell activation. It is not clear whether the presence of this glycoprotein on lymphocyte membranes is due to an endogenous production or to a passive uptake of soluble alpha 1 acid-glycoprotein by its carbohydrate moiety. Our data show an increase of membrane alpha 1 acid-glycoprotein both in peripheral blood lymphocyte and T-enriched lymphocytes after phytohemagglutinin stimulation. Peripheral blood lymphocyte enzymatic treatment by neuraminidase does not affect alpha 1 acid-glycoprotein expression while pronase digestion induces a strong decrease of alpha 1 acid-glycoprotein positive lymphocytes and a resynthesis after phytohemagglutinin stimulation. Furthermore, the presence of alpha 1 acid-glycoprotein was prevalently, found on helper/inducer lymphocytes. These data support the hypothesis of a synthesis of alpha 1 acid-glycoprotein by T lymphocytes during their activation process.     

Human B cell differentiation. II. Suppression by T cells of T-dependent and T-independent plasma cell maturation.

In vitro human plasma cell generation induced by both T-dependent (PWM) and T-independent (NWSM) mitogens was found to be suppressed by peripheral blood lymphocytes preincubated with human aggregated IgG. T cells, but not B lymphocytes, were able to mediate the suppressive activity; since aggregated (Fab)'2 fragments were found unable to generate suppressor cells, it was concluded that the suppressor cell was a T lymphocyte bearing Fcgamma receptors. These cells appeared to be largely radiosensitive. In most cases the proliferative responses remained unchanged. Since NWSM-induced activation is not dependent on the presence of T cells, these results show that, at least in this case, T cells exert their suppressor function directly on B lymphocytes. Whether PWM-induced B cell differentiation is suppressed by the same mechanism or/and by inactivation of T helper lymphocytes remains under investigation.     

The lymphocyte plasma membrane: locus of control in the immune response.

The lymphocyte plasma membrane is the locus of events which control the immune response. T and B lymphocytes, which mediate cellular and humoral immunity respectively, show distinctive plasma membrane morphologies and cell surface receptors. The dynamic state of these plasma components is emphasized by their lateral mobility in the fluid plane of the membrane, as well as variation in their structure or expression as the lymphocyte proliferates and differentiates in response to stimulation by antigen or mitogens. The best understood membrane glycoproteins are surface membrane immunoglobulins that serve as antigen receptors on B cells, and the histocompatability-beta2 microglobulin complex that has an immunoglobulin-like structure. Other less well defined surface structures showing modulation during the cell cycle may affect growth regulation of proliferating lymphocytes. Some of these are shared by fetal and neoplastic cells. Major theories of lymphocyte signaling are discussed, and the early events in lymphocyte activation are reviewed. While a complete model encompassing all these early events is not yet possible, the central issues can be usefully discussed in term of receptor-transducer-effector concepts derived by strong parallels from a knowledge of hormone-membrane interactions.     

Analysis of lymphocyte proteins from New Zealand black mice by two-dimensional gel electrophoresis

Splenic lymphocyte proteins from New Zealand Black (NZB) mice, which spontaneously develop autoimmune disease, and several control strains were analyzed by two-dimensional polyacrylamide gel electrophoresis. A number of strain- and age-related differences were observed, among which was the persistently elevated synthesis of two peptides, 12.5 KD and 10.5 KD, by spleen cells from older NZB mice. Although synthesis of these peptides was moderately high in young NZB and control mice, it diminished with age in control mice. These proteins were found in the cytoplasm and were not expressed on the plasma membrane nor secreted into the medium. Production of these proteins was restricted to B and null cells; T cells did not synthesize these peptides. These proteins appear to be indicators of disease activity, because their increased synthesis was associated with lymphocyte subset alterations associated with the onset of overt autoimmune disease in NZB mice.