Studies on thymus function. I. Cooperative effect of thymic function and lymphohemopoietic cells in restoration of neonatally thymectomized mice.

Immunological restoration of 45-day old, neonatally thymectomized C3Hf mice by treatment with humoral thymic function (thymoma grafts, thymus or thymoma in diffusion chambers) ranges from 0 to 12% and is difficult to acheive. When small numbers (5-20 x 10(6)) of young adult lymphohemopoietic cells, ineffective by themselves, are given in association with humoral thymic function, a cooperative effect is observed and restoration ranges from 30 to 60%. With a particular cell dosage (20 x 10(6)), effectivity for cooperation with thymic function was the following in decreasing order: spleen, lymph nodes, thoracic duct cells, bone marrow, blood leukocytes, thymus, and Peyer's patch cells. Comparable results were obtained using spleen, thymus, and hemopoietic liver from newborn donors in association with thymic function. For similar cell dosages, newborn thymus cells were more effective than adult thymus in their cooperative effect with thymic function. Dispersed thymus cells in association with young adult bone marrow or newborn hemopoietic liver cells showed no synergism for the cooperative effect with thymic function in the present model. Using hemiallogeneic cells (F(1) hybrid into parent) it was possible to show that restoration was mediated by proliferative expansion of the injected cells. This was indicated by specific tolerance to tissues of the other parental strain and by cellular chimerism, especially of lymphoid tissues, as indicated by chromosome markers and absence of significant numbers of immunocompetent cells of host origin. A population of paritally differentiated cells of hemopoietic origin, termed postthymic, sensitive to humoral activity of the thymus and present in the lymphohemopoietic tissues of adult and newborn mice is postulated to explain our results. These cells are postthymic and thymus dependent in the sense that they already received thymic influence, probably through traffic, and are incapable of self-renewal in absence of the thymus. Sensitivity to humoral activity of the thymus is characterized by proliferative expansion and/or a differentiative process eventually leading to larger numbers of competent cells.

Studies on thymus function. II. Cooperative effect of newborn and embryonic hemopoietic liver cells with thymus function.

Significant immunological restoration of 45-day old, neonatally thymectomized C3Hf mice was obtained by the cooperation of syngeneic newborn or embryonic hemopoietic liver cells with thymic function. Thymic function or cells alone are almost ineffective or restore approximately 10% of the animals. Newborn liver cells are effective in association with thymus grafts or humoral thymic function (thymoma grafts and thymus or thymomas in diffusion chambers). Embryonic liver cells are ineffective, even in large numbers, when associated with humoral thymic function. On the other hand, embryonic liver cells are effective in the cooperative effect only in association with viable thymus grafts, preferably syngeneic, whether the grafts were placed subcutaneously, intraperitoneally, or under the kidney capsule. Dispersed viable thymic cells are ineffective in association with embryonic liver cells. Cells capable of cooperating with humoral thymic function start to appear to embryonic liver by day 19-21 of gestation and are detectable until day 5-6 postbirth. Embryonic hemopoietic liver cells from 12 to 18 days of gestation contain cells capable of cooperation only with viable free thymus grafts and not with humoral thymic function. A prethymic cell population of partially differentiated cells of hemopoietic origin, insensitive to humoral activity of the thymus but requiring thymic stroma and traffic through the thymus is postulated to explain our results. This population of prethymic cells can become postthymic through this process and eventually develop into competent cells. Postthymic cells are characterized by their sensitivity to humoral activity of the thymus and by their wide distribution in the lymphohemopoietic tissues of newborn and young adult mice.

Carcinogen-induced tumors of the thymus. IV. Humoral influences of normal thymus and functional thymomas and influence of posthymectomy period on restoration.

A progressive decrease of the restoring effectivity of syngeneic or allogeneic thymus and functional thymoma grafts was observed when the treatment of neonatally thymectomized mice was delayed. Early treatment (5-20 days postthymectomy) was effective, while the number of restored animals was markedly decreased after late treatment (30-50 days postthymectomy). Similar results were obtained with subcutaneous or intraperitoneal thymus grafts and with thymus grafts within cell-impenetrable diffusion chambers. After the onset of the postthymectomy-wasting syndrome the only successful treatment was the implantation of multiple thymus grafts. On the other hand, single thymus grafts, thymoma grafts, or thymus or thymoma within diffusion chambers were ineffective. When spleen cells from 5-day old or 45-day old neonatally thymectomized animals were given in association with thymoma grafts, only the cells derived from the 5-day old thymectomized mice proved effective in restoring wasted thymectomized hosts. These results suggest that a population of cells sensitive to the action of the thymus decreases progressively with time in the absence of thymic function.

Studies on thymus function. 3. Duration of thymic function.

The immune functions of neonatally thymectomized C3Hf mice exposed only temporarily to thymus function show a progressive decay with time in the absence of the thymus. The immune responses studied at different ages in the range of 100-600 days were: first-set rejection of H-2-compatible and incompatible skin allografts, second-set rejection of skin allografts, capacity of spleen cells to produce graft-versus-host reactions in F(1) hybrids, resistance to infection with mouse hepatitis virus, and response of spleen cells to phytohemagglutinin in vitro. These long-term studies had the purpose of determining the duration of the restoration induced by thymus function when the mice were exposed only temporarily to it. Different models were used but the two basic ones were: (a) mice grafted intraperitoneally at 15 days of age with a syngeneic thymus that was removed surgically at 10, 20, or 30 days after grafting, and (b) mice grafted at 15 days of age with allogeneic strain A thymoma or C57BL thymus, these representing situations in which there is spontaneous rejection of the restoring graft. In all the experimental models used, the animals were restored when tested at 100 days of age, but progressively became immunologically incapacitated at 200-300 days of age. From the more controlled experiments in which the restoring thymus graft was removed surgically, the following conclusions can be drawn. (a) A short exposure to a thymus graft can produce restoration of immune functions in neonatally thymectomized mice, but this restoration is not self-sustaining in the absence of the thymus and declines progressively with age. The decline usually starts at 200-300 days of age. (b) This was especially clear in experiments in which the same animal was tested twice in its lifetime for capacity to produce graft-versus-host reactions; these animals were competent at 100 days and became incompetent at 400 days of age. (c) The shortest period of thymic exposure studied was 10 days; if vascularization of the graft is taken into account, 2-3 days of thymic function are sufficient to produce restoration. (d) The immune decay observed in the thymectomized animals exposed temporarily to thymus was more profound than the physiological decay of immunity observed in control animals of similar age. (e) Of all the tests studied, the response of spleen cells to phytohemagglutinin was to be preserved the longest in animals exposed only temporarily to thymic function. The present results were interpreted in accordance with our previous findings indicating that a population of postthymic cells can be developed by temporary exposure of neonatally thymectomized animals to thymic function, but that this population is not self-sustaining in the absence of thymus and progressively decays by physiological attrition.

Antigen-specific helper factor in man.

Supernates of tetanus toxoid (TT) antigen-stimulated human T cells were studied for the presence of an antigen-specific T-cell helper factor (ASF). Supernates were circulated over an immunosorbent column consisting of insolubilized TT antigen. The material which bound to the column was eluted with 3 M NaCNS and was shown to contain a factor which in the presence of TT-induced specific IgG anti-TT antibody synthesis in autologous B cells without causing readily detectable proliferation. ASF activity was partially inhibited by antisera directed against the B-cell alloantigens of the ASF donor. Immunosorbent columns containing such antisera removed ASF activity. Immunosorbent columns containing antisera to human immunoglobulin heavy chain determinants did not remove ASF activity; whereas immunosorbent columns containing rabbit idiotypic antiserum directed against anti-TT antibodies completely removed ASF activity. ASF was destroyed by treatment with proteolytic enzymes; its molecular weight was estimated by Sephadex G-100 gel column chromatography to be between 25,000 and 75,000 daltons.

Association of HL-A 5 and immune responsiveness in vitro to streptococcal antigens.

Lymphocytes, from randomly selected individuals having normal immune function, when incubated in vitro with varying concentrations of streptococcal antigens, responded in three ways: (a) response over the entire antigen concentration range, i.e., responders; (b) low response to only the highest antigen concentrations; and (c) no response at any antigen concentration. Frequency distribution analysis of these groups indicated that a significant association occurred between the ability to respond and HL-A 5.

The immune response to hepatitis B vaccine in humans: inheritance patterns in families.

We have recently shown that the human antibody response to the hepatitis B virus surface antigen (HBsAg) vaccine is major histocompatibility complex (MHC) associated. In studies of nonresponders to the vaccine, we found an increased incidence of individuals homozygous for human histocompatibility leukocyte antigen (HLA) proteins associated with the extended (conserved) haplotype [HLA-B8,SC01,DR3]. In later prospective vaccination trials, we showed that none of five individuals homozygous for this haplotype developed more than 1,300 radioimmunoassay (RIA) units of antibody (mean, 467 RIA units), while all heterozygotes made at least 2,500 RIA units (mean antibody level, 15,608 units). Our results suggested that [HLA-B8,SC01,DR3] lacks an immune response gene for HBsAg, and that response is inherited in a dominant fashion. To provide further evidence for this hypothesis, we have now analyzed the results of HBsAg immunization in families. 43 members of 10 families were immunized with the hepatitis B vaccine, including seven families where at least one member bore the haplotype [HLA-B8,SC01,DR3], and three families where one member had already received, but failed to respond to, the vaccine. In two of these three families, the presence of [HLA-B8,SC01,DR3] was subsequently found. Of nine MHC-identical sibling pairs in the study, both members of eight pairs had similar antibody responses (five nonresponder and three responder pairs). In all families with such sibling pairs, including the discordant pair, rank-ordering members by antibody level demonstrated that no relative's value came between the sibling pair values. Furthermore, of nine [HLA-B8,SC01,DR3]-haplotype-homozygous individuals, six were nonresponders, and two others had only low-normal responses. [HLA-B8,SC01,DR3]-heterozygous family members always had higher levels of antibody than their homozygous relatives. Linkage analysis of nonresponse to HLA haplotypes revealed a maximum likelihood LOD (logarithm of the odds) score of 6.3 at a recombination fraction of 0.1. The MHC association with lack of antibody response to HBsAg was not seen with tetanus immunization, where 1 of 20 HBsAg responders and 1 of 21 poor or nonresponders had tetanus titers of less than 1:512; both tetanus nonresponders were [HLA-B8,SC01,DR3] heterozygotes. Our results indicate that: (a) response to the HBsAg vaccine is MHC linked, and inherited in a dominant fashion; (b) an abnormal or missing immune response (Ir) gene for HBsAg is a characteristic of most examples of the extended haplotype [HLA-B8,SC01,DR3]; and (c) other haplotypes also have abnormal or missing Ir genes for HBsAg.

Phospholipase C-gamma 1 association with CD3 structure in T cells.

Recently, we and others have reported tyrosine phosphorylation of phospholipase C-gamma 1 (PLC gamma 1) enzyme after CD3 activation of T cells, and have proposed that PLC gamma 1 mediates signal transduction through the T cell receptor (TCR/CD3). Here, using immunoblotting and immune complex PLC assays, we show that CD3 stimulation of Jurkat cells induces the association of PLC gamma 1 enzyme with CD3 complex. PLC activity is also found to co-precipitate with the CD3 zeta chain from activated cells. In addition, in vitro PLC assays show that CD3 activation leads to about 10-fold stimulation of PLC gamma 1 activity. These results, along with the observation that Jurkat cells preferentially express PLC gamma 1, indicate that PLC gamma 1 participates in CD3 signaling.

Postthymectomy wasting associated with autoimmune phenomena. I. Antiglobulin-positive anemia in A and C57BL-6 Ks mice.

1. Mice of A and C(57)BL/6 Ks strains, thymectomized at birth acquire wasting disease in 84.1% (A) and 77.1% (C(57)BL/6 Ks) of the cases. There is no sex predelection. 2. Anemia in these animals is characterized by shortened red cell survival and increased fragility to hypotonic salt solutions. Among thymectomized A mice reticulocytosis is absent and extramedullary hematopoiesis is found in the spleen in the presence of bone marrow hypoplasia for the erythroid and lymphocyte series. 3. Positive antiglobulin tests of the red cells were observed in all the thymectomized C(57)BL/6 Ks (7/7) and 71.2% of the A strains (62/87). Normal mice do not show positive Coombs' tests. 4. The globulin coat on the A strain consists of IgM, whereas beta(1C) and IgG are not detectable. By contrast, red cell coats of NZB mice developing spontaneous autoimmune hemolytic anemia show IgM and beta(1C), but these erythrocytes do not react with anti-gamma chain antibodies. Another difference in the globulin coats of the two types of erythrocytes is that the IgM on NZB red cells has available light chain determinants but these are apparently hidden or absent in the case of sensitized erythrocytes. The difference in antibody coating, association with a component of complement in one but not the other, suggests a different mechanism for the immune surface phenomenon in each instance. 5. Anemia in NZB mice is associated with reticulocytosis while that in thymectomized A mice is not. 6. Thymectomy appears to initiate a chain of events leading to a series of autoimmune phenomena which may be due to alteration in host response consequent to loss of thymic tissue and thymic dependent functions or alternatively to infection to which increased susceptibility exists as a result of thymic extirpation.

Polymorphic Hh genes in the HLA-B(C) region control natural killer cell frequency and activity.

We demonstrated earlier that individuals homozygous for conserved major histocompatibility complex (MHC)-extended haplotypes have low natural killer (NK) activity as measured by cytolysis of the K562 tumor cell lines. In the present study, we investigated the segregation and MHC linkage of NK activity in families in which MHC haplotypes of human histocompatibility leukocyte antigens (HLA)-A, -C, and -B, complotype, and DR specificities are known. In two informative families, low activity was inherited as a recessive trait linked to the MHC. By using individuals homozygous for specific fragments of extended haplotypes or for HLA-B alleles, we found that the HLA-C and -B and not the complotype or HLA-DR region contains genes controlling NK activity. The majority of the unrelated individuals with low NK activity were homozygous or doubly heterozygous for HLA-B7 (Cw7), B8 (Cw7), B44 (Cw5), B18, or B57 (Cw6). Thus, these alleles form one complementation group designated NKB1. Another less frequent group, NKB2, was also identified, and consisted of individuals homozygous for B35 (Cw4). NK activity was correlated with the number of circulating NK (CD16+ CD56+) cells. Individuals homozygous for the NKB complementation groups have fewer circulating NK cells than individuals heterozygous for these alleles and alleles of other complementation groups, possibly explaining the low activity of cells in these subjects. These findings suggest that during the maturation of NK cells there is NK cellular deletion in donors homozygous for NKB genes resulting in low NK cell numbers and activity.

Linkage of pemphigus vulgaris antibody to the major histocompatibility complex in healthy relatives of patients.

Pemphigus vulgaris (PV) is an autoimmune disease caused by high concentrations of antibody to an epidermal cadherin. The disease is associated with two kinds of HLA-DR4, DQ8 haplotypes dominantly distributed among Jewish patients, and these plus DR6, DQ5 haplotypes in non-Jewish patients. Low levels of the PV antibody were found in 48% of a total of 120 asymptomatic parents, children, and siblings of 31 patients, thus exhibiting dominant inheritance. The inheritance of these low levels of antibody in asymptomatic relatives was linked to the major histocompatibility complex with a highly significant logarithm of the odds score of 9.07, almost always to a DR4 or DR6 haplotype of the patient. Disease appears to occur in susceptible individuals with low levels of antibody when a second factor, either environmental or genetic, induces high levels, sufficient to produce blisters.

Peripheral human T cells sensitized in mixed leukocyte culture synthesize and express Ia-like antigens.

We have studied the modulation of Ia-like antigens on the surface membrane of human T cells responding in a one-way mixed leukocyte culture. A heterologous antiserum, (anti-p23,30), which is specific to HLA-D-related antigens and which is unreactive with normal peripheral T cells or thymocytes, was found to bind significantly to all T cells transformed in mixed leukocyte culture (MLC) as determined by indirect immunofluorescence on a fluorescence-activated cell sorter 1. Furthermore, cytotoxic T cells responsible for cell-mediated lympholysis were shown to react with anti-p23,30, whereas their unactivated progenitors did not. Immunoprecipitation and sodium dodecyl sulfate polyacrylamide gel electrophoresis of a radioactive 29,000 and 34,000 dalton complex from MLC-primed T cells labeled with [35S]methionine indicated that allosensitized T cells synthesized these HLA-D-related antigens.

Major histocompatibility complex susceptibility genes for dermatitis herpetiformis compared with those for gluten-sensitive enteropathy.

Dermatitis herpetiformis (DH) shares some clinical features and major histocompatibility complex (MHC) markers with gluten-sensitive enteropathy (GSE). We compared MHC haplotypes in 27 patients with DH, 35 patients with GSE, and normal controls. As in GSE, the frequencies of two extended haplotypes, [HLA-B8, SC01, DR3] and [HLA-B44, FC31, DR7], were increased in patients with DH. Distributions of fragments of extended haplotypes, consisting of some but not all of the elements of complete extended haplotypes, were analyzed to attempt to localize a susceptibility gene. Besides complete extended susceptibility haplotypes, (DR3, DQ2) and (DR7, DQ2) fragments were most common in GSE. In contrast, DH showed only a few such fragments but many instances of the fragment (SC01). The differences in distribution of these fragments in the two diseases were highly significant (P < 0.002). HLA-DQ2 and DR3 had the highest odds ratios for GSE, but the highest odds ratio for DH was for the complotype SC01. These findings suggest that the MHC susceptibility gene for DH is between class II and complotype regions, closest to the complotype, whereas that for GSE is in the class II region.

The cellular basis for lack of antibody response to hepatitis B vaccine in humans.

We had previously obtained evidence that among normal subjects the humoral antibody response to hepatitis B surface antigen (HBsAg) was bimodally distributed with about 14% of subjects producing less than 1,000 estimated radioimmunoassay RIA units. From the study of major histocompatibility complex (MHC) markers in the very poor responders who produced less than 36 estimated RIA units of antibody, it appeared that there was an excess of homozygotes for two extended haplotypes [HLA-B8, SC01, DR3] and [HLA-B44, FC31, DR7]. This finding suggested that a poor response was inherited as a recessive trait requiring nonresponse genes for HBsAg on both MHC haplotypes and was strengthened by finding a much lower antibody response among prospectively immunized homozygotes for [HLA-B8, SC01, DR3] compared with heterozygotes. In the present study, we have analyzed the cellular basis for nonresponse to this antigen by examining antigen-specific proliferation of T cells from responders and nonresponders in the presence and absence of autologous CD8+ (suppressor) cells. Peripheral blood cells from nonresponders to HBsAg failed to undergo a proliferative response to recombinant HBsAg in vitro, whereas cells from responders proliferated vigorously. This failure of cells from nonresponders to proliferate was not reversed in cell mixtures containing CD4+ and antigen-presenting cells devoid of CD8+ cells. There was no difference between responders and nonresponders with respect to the number of circulating T cells or their subsets, or the proliferative response to mitogens such as pokeweed or phytohemagglutinin or another antigen, tetanus toxoid. Our results indicate that our HBsAg nonresponding subjects have a very specific failure in antigen presentation or the stimulation of T helper cells, or both. Our evidence is against specific immune suppression as the basis for their nonresponsiveness. The failure of antigen presentation or T cell help is consistent with recessive inheritance of nonresponsiveness and suggests that response is dominantly inherited.

Tubules of globoid leukodystrophy: a right-handed helix.

Morphologic similarity between the cytoplasmic tubules of globoid leukodystrophy and Gaucher's disease (as demonstrated by thin sectioning, negative staining, and shadowing techniques) and their resemblance to negatively stained beef cerebroside are presented as evidence favoring the accumulation of cerebrosides in globoid cells. The tubules of globoid leukodystrophy have a 60-angstrom periodic banding similar to the tubules of Gaucher's disease. A right-handed helical twisting of the tubules is observed in both diseases.

Acute lymphoblastic leukemic cells with T (thymus-derived) lymphocyte markers.

Five of nine children with acute lymphoblastic leukemia had lymphoblasts that bound sheep erythrocytes or reacted with antiserum to thymocytes, suggesting involvement of T (thymus-derived) cells. When lymphoblasts from all patients were examined by immunofluorescence they were found to lack a marker for B (bone marrow or bursa-equivalent) cells, that is, the presence of surface immunoglobulins.

Possible role of natural killer cells in pemphigus vulgaris - preliminary observations.

Pemphigus vulgaris (PV) is an autoimmune blistering disease that affects the skin and multiple mucous membranes, and is caused by antibodies to desmoglein (Dsg) 1 and 3. Natural killer (NK) cells have a role in autoimmunity, but their role in PV is not known. NK cells in the peripheral blood leucocytes (PBL) of 15 untreated Caucasian patients with active PV were studied and compared with healthy controls for the expression of major histocompatibility complex (MHC) class II and co-stimulatory molecules. CD56+ CD16- CD3- NK or CD56+ CD16+ CD3- NK cells from the PBL of PV patients co-express MHC class II and co-stimulatory molecule B7-H3 without exogenous stimulation. CD4+ T cells from the PBL and perilesional skin of PV patients were co-cultured with CD56+ CD3- NK cells from the PBL of the same patients; in the presence of Dsg3 peptides underwent statistically significant proliferation, indicating that NK cells functioned as antigen-presenting cells. Supernatants from these co-cultures and serum of the same patients with active PV had statistically significantly elevated levels of interleukin (IL)-6, IL-8 and interferon-gamma, compared with controls indicating that the NK cells stimulated CD4+ T cells to produce proinflammatory cytokines. In these experiments, we present preliminary evidence that NK cells may play a role in the pathobiology of PV.

Strategies and technical challenges in allele level Class II typing in 2578 bone marrow transplantation donor-recipient pairs.

Human leukocyte antigen typing of 2578 donor-recipient pairs whose transplantation was facilitated by the National Marrow Donor Program allowed for an in-depth analysis of the accuracy of high-volume allele level testing data. The methods employed provided allele level typing at DRB1/3/5, DQA1, DQB1, DPA1, and DPB1 using sequence-specific oligonucleotide probe hybridization (SSOPH), polymerase chain reaction (PCR) restriction fragment length polymorphism analysis, sequence specific PCR, and direct sequence-based typing (SBT). Each typing was independently tested by two laboratories in Phase 1, and in subsequent phases targeted samples were typed in duplicate by SBT to monitor typing quality. Comparison with prior transplant center typing was also evaluated. SSOPH detected discrepancies ranged from 0.6% at DPB1 to 5.1% at DQB1 in Phase 1. The majority of discrepancies, 62%, resulted from human error such as sample handling, result interpretation, or clerical errors. Alleles that are frequently discrepant have been identified in this predominantly white population.

The histocompatibility system in juvenile, insulin-dependent diabetic multiplex kindreds.

We have histocompatibility (HLA) genotyped 24 families with two or more juvenile, insulin-dependent, ketosis-prone diabetic siblings. This criterion for family selection was used to obtain a homogeneous form of diabetes within a sibship, because diabetes appears to be a genetically heterogeneous disease. 58 diabetic and 53 nondiabetic sibs and 40 parents were studied. 55% of the diabetic pairs were concordant for both HLA haplotypes (expected 25%), 40% were concordant for one haplotype (expected 50%), and 5% were discordant for both haplotypes (expected 25%). These values are significantly different from the expected values (P < 0.001). On the other hand, the inheritance of haplotypes among the nondiabetic sibs in these families was not significantly different from the expected mendelian segregation. When comparing 20 pairs of HLA identical (sharing two haplotypes) with 15 pairs of haploidential (sharing one haplotype) diabetic sibs for the intrapair difference in age of onset of disease, we found that the HLA identical sibs were significantly more concordant for age of onset (3.9 yr difference) than the haploidential (7.3 yr difference) (P < 0.05). The same type of analysis for the difference in seasonal incidence in months revealed that the HLA indentical sibs were more concordant (1.8 mo difference) than the haploidentical sibs (3.2 mo difference) (P < 0.025). Furthermore, the HLA identical diabetic sibs were more likely to develop diabetes in the winter months (78%) than the haploidentical diabetic sibs (21%). No particular HLA haplotype or antigen seemed to be associated with any particular clinical feature. These data are compatible with the theory of genetic heterogeneity of juvenile, insulin-dependent diabetes. It is suggested that there are one or more diabetes response genes in the HLA region playing an important role in the pathogenesis of juvenile, insulin-dependent diabetes in the families studied here. It is, however, possible that other genes, not associated with the HLA complex, may play an etiologic role in some cases of juvenile, insulin-dependent diabetes, resulting in lack of association between HLA and some forms of diabetes.

Studies of human sera with cytotoxic activity.

42 human sera showing in vitro cytotoxic activity of restricted or broad HL-A specificities with test human lymphocytes were studied for the molecular and immunoglobulin class of cytotoxic antibody using sucrose gradient separations, DEAE-cellulose chromatography, and Sephadex G-200 gel filtration. Sera originated from patients with previous multiple pregnancies (19), multiply transfused patients (8), subacute bacterial endocarditis (4), systemic lupus (2), and human umbilical cord sera (9). In 32 of 42 instances, predominant cytotoxic activity was found in high molecular weight gradient fractions; however, DEAE chromatographic separations revealed cytotoxic activity in initial buffer fractions containing primarily gammaG globulin. Gradient separations of cytotoxic activity within initial gammaG DEAE fractions showed localization of cytotoxicity only in high molecular weight materials. Confirmation of high molecular weight gammaG cytotoxic activity was obtained by resistance to mercaptoethanol treatment, abolition of activity after absorption only with specific anti-gammaG antisera, and by the finding that high molecular weight cytotoxic activity in gradients or gel filtration run at neutral pH 7.4 became 7S when separations were rerun at an acidic pH of 4.0. Such 7S activity again became rapidly sedimenting when the same fractions were again rerun in gradients at neutral pH.19S gammaM cytotoxic activity was documented in a panel of 15 human sera containing anti-"I" cold agglutinins. In this instance the cytotoxic activity appeared to be related to the cold agglutinin antibody since it was mercaptoethanol sensitive and could be demonstrated in monoclonal antibody eluates containing primarily gammaM. No in vitro protection again cytotoxic effect was demonstrated when isolated human 19S anti-gamma-globulins were aded to the lymphocytotoxic assay system. With the exception of cold agglutinins, human cytotoxic antibodies appear to be primarily gammaG producing in vitro lymphocyte killing either as 7S gammaG globulin or as rapidly sedimenting aggregates or complexes of gammaG molecules.