APPLICATIONS OF ISO-IMMUNE CYTOLYSIS USING RADIOLABELLED TARGET CELLS

FDA/PI flow cytometry assay of complement-mediated cytotoxicity of antibodies generated during xenotransplantation

BACKGROUND

A flow cytometry complement-mediated cytotoxicity assay (FCCA) using fluorescein diacetate (FDA) and propidium iodide (PI) to measure antibody-dependent toxicity is useful to determine the success of xenotransplant organs. We evaluated the validity of different mathematical models as a measure of cytotoxicity in FCCA.

METHODS

Sera from untreated baboons (n = 7) and from immunosuppressed animals (n = 5) undergoing different xenotransplantation protocols with pig organs were tested by endogenous FCCA and a similar assay also using exogenous complement, and the results were compared with those of a complement-dependent hemolytic assay to detect anti-pig antibodies (APHA). The influence of PI/FDA staining and the use of several mathematical models were analyzed.

RESULTS

For both groups of animals, we observed high correlations between the endogenous and exogenous FCCA pathways and between calculations based on PI and FDA staining. Of the four mathematical models tested--the Von Krogh equation, two exponential models, and area under the curve--the Von Krogh equation was the most appropriate in terms of goodness of fit and concordance with APHA.

CONCLUSIONS

FDA/PI FCCA is useful to measure endogenous and exogenous complement-mediated cytotoxicities, and it has advantages related to identification of potential new xenoantibodies. Although all four mathematical models produced acceptable solutions, the Von Krogh equation was the best option.

Properties of purified papain-solubilized rat AgB antigens and reactivity of a xenoantiserum against the isolated antigens

Rat AgB transplantation antigens were isolated after papain digestion of spleens from the inbred strain Hooded Lister. Both subunits of the AgB antigens were present in the purified material. Some physical characteristics of the antigens have been determined. An antiserum, raised in a rabbit, against the purified material reacted exclusively with AgB antigens on splenocytes but detected novel structures on both adult and embryonic fibroblasts. These structures, antigenically related to AgB antigens, were not detected on plasmacytoma or hepatoma cells, nor did they display any antigenic similarity with rat beta 2-microglobulin. Radioimmunoassays specific for the AgB antigen heavy chain and for beta 2-microglobulin, respectively, were used to estimate the contents of these antigens in several tissues. Spleen and thymus exhibit the largest density, while brain is almost devoid of these antigens.

Cloning, characterization, and functional studies of a nonintegrin platelet receptor for type I collagen

A cDNA (1.6 kb) encoding a platelet protein receptor that binds type I collagen has been isolated from a human bone marrow cDNA library by using a degenerate oligonucleotide probe derived from the amino acid sequence of a CNBr fragment of the purified receptor. Computer search revealed that this cDNA represents the coding sequence of a unique protein. Using the prokaryotic expression system pKK 223-3-65 cDNA, a 54-kD recombinant protein was obtained and purified to apparent homogeneity. In an eukaryotic expression vector (pcDNA3-65 cDNA), a 65-kD protein was identified that was recognized by monoclonal anti-65 kD antibody (anti-65m). The recombinant protein binds to type I, but not to type III collagen by affinity column chromatography. The binding of the recombinant protein to type I collagen-coated Petri dishes is inhibited by anti-65m in a dose-dependent manner. The pcDNA3-65 cDNA-transfected nonadherent T cells express the protein, allowing them to attach to a type I collagen matrix, and are inhibited by anti-65m in a dose-dependent manner. Like the receptor protein purified from platelet membranes, the recombinant protein inhibits type I collagen-induced platelet aggregation and the adhesion of [14C]serotonin-labeled platelets to type I collagen in a dose-dependent manner. The recombinant protein neither binds to type III collagen-coated Petri dishes nor inhibits type III collagen and ADP-induced platelet aggregation, indicating specificity for type I collagen.

Effects of insulin and transferrin on the generation of lymphokine-activated killer cells in serum-free medium

We have developed a serum-free medium designated RDSF for the generation of LAK cells based on RD6F medium, which was originally developed as a serum-free medium for the growth of myeloma and hybridoma cells. The cytotoxic activity of LAK cells generated in RDSF against Raji, K562 and oral cancer cells, is 3-4 times that of LAK cells generated in medium containing 10% human type AB serum. RDSF medium consisted of nutrient mixture supplemented with transferrin, 2-aminoethanol, 2-mercaptoethanol, sodium selenite and interleukin-2. In this study, we have found that insulin which has been shown to be the most important polypeptide hormone in serum-free media for animal cells, inhibited the generation of cytotoxic activity of LAK cells cultured from peripheral blood lymphocytes. In addition, we found that transferrin was an essential component for the growth and generation of LAK cells in serum-free culture. These results suggest that RDSF may be useful in adoptive immunotherapy for cancer as well as for studying factors involved in the growth and differentiation of LAK cells.

Use of [35S]methionine-labelled rat lymphoblasts in micro-cytotoxic and limiting dilution assays

When only limited numbers of effector cells are available for in vitro T cytotoxic determinations, standard assays cannot be performed. 51Cr is still the most commonly used marker of target cells in cytotoxicity assays but since the incorporation of this marker is low, especially in non-tumor cells such as lymphoblasts, larger numbers of both target and effector cells are required. Here we report the use of [35S]methionine-labelled rat ConA blasts in cytotoxic, micro-cytotoxic and limiting dilution assays. Regardless of whether [35S]methionine or 51Cr targets were employed, cytotoxic activities were identical when large numbers of target cells (10(4)) were used. The high uptake of [35S]methionine by ConA blasts (9 cpm/cell) permitted the use of a small number of target cells without any loss of sensitivity. Therefore, the number of effectors and targets required was dramatically reduced, especially with high E : T ratios such as 100 : 1. The use of low number of [35S]methionine-labelled rat ConA blasts as targets was also suitable for the measurement of alloreactive T cell precursor frequencies. This technique illustrates the possibility of studying T cytotoxicity in animal species lacking tumor target cell lines under experimental conditions where the availability of effector cells is limited and the optimal use of such cells becomes critical.

The use of non-radioactive chromium as an alternative to 51Cr in NK assay

A novel method to measure target cell cytolysis based on the use of 'cold', non-radioactive chromium and on the determination of metal release by graphite furnace atomic absorption spectroscopy (FAAS) is proposed. Natural killer (NK) assays were performed by labelling target cells with chromium as Na2CrO4, and results were compared with those obtained by conventional overnight labelling with 51Cr of targets killed by the same effectors. The cytotoxic capacity of peripheral blood lymphocytes from healthy subjects was evaluated, and NK activity measured with both methods showed a good agreement at each of the tested effector to target cell ratios (between 100:1 and 1:1), with a high and significant coefficient of correlation (r = 0.931, p < 0.0001). The selection of the appropriate Cr concentrations for labelling target cells took into account both the sensitivity of our instrumentation and the possible toxic effects of the metal. A study of the effects of Cr on the cell line (K562) which is usually employed as a target in NK tests showed that Cr could have a detrimental effect on cellular function, with significant numbers of cells with depolarised mitochondria and reduced DNA synthesis after 24 h incubation using Cr levels higher than 15 mumol/l (780 micrograms/l). The method proposed here has a number of advantages, including the use of a non-radioactive tracer, limited costs, high sensitivity and reproducibility, and the possibility of storing samples. In addition, the technique uses a fixed Cr concentration which is known to be non toxic.

Single and dual labelling of cytotoxic target cells. Comparison of three radioactive tracers, [35S]methionine, [75Se]selenomethionine, and chromium-51

[75Se]selenomethionine (75SeM) has been shown to provide several advantages over Na(2)51CrO4 (51Cr) labelling of metabolizing target cells: high labelling efficiency and low spontaneous release of 75SeM-labelled target cells permit improved monitoring of cytotoxicity due to extended effector/target ratios in short- and long-term assays. Unfortunately, 75SeM will soon be difficult to obtain. Therefore we studied the suitability of [35S]methionine (35SM) as a substitute for 75SeM. Furthermore, we explored the potential of dual labelling of suspension target cells applying combinations of 35SM and 51Cr or 75SeM and 51Cr. 35SM is a suitable substitute for 75SeM retaining most of the advantages of 75SeM labelling. Although considerably higher labelling of cells is possible we obtained the most efficient labelling with 100-400 kBq/ml of 35SM or 75SeM resulting in a relatively high uptake (3-15 cpm/cell) and very low spontaneous release (1-2%/h) up to 24 h. This permits short- and long-term cytotoxic assays and the use of low numbers of target cells (1 x 10(3)) providing increased cytotoxic sensitivity with reduced amounts of effector cells. Suitable dual labelling of target cells with 35SM plus 51Cr or 75SeM plus 51Cr documented convincingly identical release kinetics for 35SM and 75SeM but partially discordant ones for 51Cr. Depending on the target cell used dual labelling permits discrimination and monitoring of different cytotoxic or release mechanisms in cellular cytotoxicity.

Toxicity of Pulpispad using four different cell types

The cytotoxic effect of a zinc oxide-eugenol-based paste (Pulpispad) was evaluated in vitro after setting for 1 day and 1 week. Target cells were L929 cells, gingival, periodontal ligament and pulpal fibroblasts. The material was incubated with the cells for 4 and 24 hours, and its toxicity was evaluated with the 51Cr-release method. Pulpispad was highly cytotoxic to all cell lines even after setting for 1 week. The use of Pulpispad is not recommended for future clinical application. The various responses among the four cell lines indicated that diploid cell lines can, under certain circumstances, be less sensitive than aneuploid cell lines. It is therefore suggested that in the evaluation of biomaterials the choice of cell lines should be carefully considered, as they can display varying sensitivities.

The expression of murine Qa region gene product(s) in L cell transformants

The cosmid H3.5, containing genes mapping to the murine H-2 Qa region, was used to transfect L cells by the calcium phosphate co-precipitation method. The resultant transfected cells expressed a Qa-like determinant as detected by an immune serum raised against the transfectant cells and Qa specific monoclonal antibodies. Two-dimensional gel analysis revealed the expression of a class I-like heavy chain with a similar molecular mass to the Qa2 antigens of the positive strain B10 and B10.A but with a different isoelectric point. The cosmid H3.5 spans 40 kb of DNA and contains at least one complete Qa region gene which encodes the Qa-like determinant detected in this study.

Utilization of murine peripheral blood lymphocytes for H-2 typing

We adapted the NIH Standard Protocol for HLA-A, B, C typing to perform murine H-2 typing. The assay is direct, measuring the cytotoxicity of the antiserum/cell/complement reaction with a supravital dye. This method is advantageous because it: utilizes peripheral blood lymphocytes (PBL) obtained from the tail vein; uses microliter volumes of antiserum; is practical because the formalin fixed reactions need not be read immediately; involves standard and inexpensive cytotoxicity techniques; is easily interpreted and is readily reproducible.

Expression of murine H-2Kb histocompatibility antigen in cells transformed with cloned H-2 genes

Cosmids containing H-2 histocompatibility antigen genes of the H-2b haplotype have been isolated. One of these genes expresses a 45,000 molecular weight protein, indistinguishable from H-2Kb when introduced into mouse L cells. These H-2Kb transformed L cells can be killed by allospecific anti-H-2Kb cytotoxic T cells. Moreover, when infected with influenza virus, they can be killed by an H-2Kb-restricted, influenza virus-specific cytotoxic T cell line. These results show that expression of the H-2Kb gene product on the L-cell surface is sufficient to make it a target for specific T-cell killing.

Genetic regulation of the expression of H-2-K.5 antigen of erythroid cells by H-2K end

Difference in the amounts of H-2K.5 antigens present on erythrocytes was observed, using quantitative absorption method, among several B 10 congenic strains. However, no such variation was seen on the lymphocytes and lung cells. The variability in the amount of this antigen on the erythrocyte surface was primarily dependent on the haplotype of the H-2K end in the B 10 recombinant strains examined. This suggests that the regulator of H-2 expression on erythrocytes is in this region. Further genetic analysis confirmed that this regulation functions in the cis-position. Finally, the H-2K.5 antigenic activity was expressed on the reticulocytes of all strains tested, but appeared lower in the type 2 group indicating that the regulation begins early in erythropoietic differentiation and eventually results in a complete loss of detectable activity.

HLA-A2 as a target for cell-mediated lympholysis: evidence from immunoselected HLA-A2 negative mutant cell lines

Cloned mutants of the human B lymphoblastoid cell line SB have been isolated using mutagenesis with ethyl methanesulfonate followed by negative selection with an anti-HLA-A2 serum and complement. Absorption analysis with 125I Staphylococcus aureus protein A binding to antibody sensitized cells. HLA typing, and immune precipitation analysis showed the mutants to be serologically identical to the SB parent except for the loss of HLA-A2. When tested as target cells for cell-mediated lympholysis by cytotoxic T lymphocytes generated in the mixed lymphocyte response, the SB and mutant cell lines demonstrated comparable susceptibility when the putative targets were HLA antigens other than HLA-A2. However, when compared for susceptibility to lysis by cytotoxic T lymphocytes considered to be HLA-A2 specific, the SB parent was effectively killed whereas little or no killing of the HLA-A2 mutants was observed. The results provide a new line of evidence that HLA antigens recognized by antibody can also be the true molecular targets for cytotoxic T lymphocytes.

Cytotoxic effect of anti-lymphocytic globulin on human lymphoid cells

The cytotoxic activity of two anti-lymphocytic globulin (ALG) preparations was investigated by measurement of 51Cr release from human peripheral blood mononuclear cells and tonsil lymphocytes incubated with ALG in the presence of complement. Both types of ALG had a greater complement-dependent cytotoxic effect on tonsil lymphocytes than on peripheral blood mononuclear cells. The ALG preparations were more cytotoxic for a population of tonsil lymphocytes enriched in non-T cells than for one enriched in T cells. This finding does not support the view that ALG has a selective action on T lymphocytes.

Some structural properties of thymus leukemia antigen (TL) solubilized with detergent

Plasma membranes were isolated from the leukemia cell line ASL1w and extracted with detergent (DOC). DOC solubilized more TL activity than could be detected on isolated membranes. However, extraction of membranes with LDS or EDTA solubilized only 17% and 4%, respectively, of the activity. This indicated that TL was not loosely associated with the membrane but rather was integrated into the lipid bilayer. At low concentrations of DOC (0.05%), TL was found to be largely aggregated and was also prone to autolysis. Neither aggregation nor autolysis was observed at a higher DOC concentration (0.5%). The apparent molecular weight of TL in 0.5% DOC was determined by Sephadex G-200 chromatography to be about 65,000-70,000. Digestion of a 0.5% DOC extract of TL with either papain or trypsin produced a fragment of TL of about 35,000 molecular weight. These fragments were similar in size to a fragment produced by autolysis. These data suggested that a region of the TL molecule was very prone to proteolytic attack. The 35,000 molecular weight proteolytic fragments bound specifically to lentil lectin affinity columns, which indicated that they retained at least part of the carbohydrate present on the native molecule.

Frequency distribution of histocompatibility-2 antigenic specificities in the Japanese wild mouse genetically remote from the European subspecies

Fifty-eight Japanese wild mice, Mus musculus molossinus, collected from twenty-eight localities were surveyed for twelve H-2 antigens using the haemagglutination method. Significantly higher frequencies of H-2.3 and H-2.5 specifities and relatively lower frequencies of the other public specificities were observed. This was confirmed by examining four specificities, H-2.3, 5, 13 and 23, in 370 mice. Quantitative absorption of the alloantisera by erythrocytes and spleen lymphocytes of molossinus mouse revealed definite absorption of H-2.3, 5 and 8 antigens, though their antigenic strength was apparently weaker than the controls, B10 congenic mice. Comparative assay of the strength of H-2.5 antigen in either homozygous and heterozygous conditions distinctly eliminated the possibility of gene-dose effect for the reduced strength of the public antigens in molossinus mouse. To explain this, divergence time between molossinus and domesticus was computed based on the allelic frequency data already reported on ten loci in both subspecies. It is roughly 1.5 x 10(6) years, which could allow for the occurrence of considerable genetic changes in many public antigens, except those conserved through selection.

Problems associated with the use of merthiolate as a preservative in anti-lymphocytic globulin

The cytotoxic properties of 2 anti-lymphocytic globulin (ALG) preparations were investigated in vitro by measuring the release of 51Cr from labelled human peripheral blood mononuclear cells, tonsil lymphocytes and Chang cells, incubated with different concentrations of ALG. One of the ALG preparations showed non-selective cytotoxicity in the absence of complement. Evidence was obtained to suggest that this effect was due to merthiolate (sodium ethylmercurithiosalicylate) which had been added to the ALG as a preservative during manufacture. The mercury concentration in the ALG was found to be greater than that stated by the manufacturers. It is conceivable that the clinical use of such as ALG preparation might lead to mercury accumulation in the tissues, with resulting toxic effects. The whole question of the use of merthiolate in the preparation of sera for administration to human subjects needs to be reconsidered.

Incorporation of rat histocompatibility (AgB) antigens into liposomes, and their susceptibility to immune lysis

The purification of detergent-solubilized rat histocompatibility antigens is described. The antigen may readily be incorporated into synthetic vesicles (liposomes) which appear to be unilamellar and between 0.1 microbeter and 0.2 micrometer in diameter. The addition of specific antibody leads to the agglutination and precipitation of the liposomes. An enzyme, horseradish peroxidase, may be incorporated into the trapped aqueous phase of the liposome, and its release by antibody and complement can be detected.

The preparation of liposomes bearing human (HLA) transplantation antigens

The incorporation of HLA antigens into liposomes is described. At and above the lipidphase-transition temperature, between 50 and 80% of the added antigenic activity may be incorporated into liposomes in the form of multilamellar vesicles. The antigen appears to be asymmetrically orientated with respect to the lipid bilayer.

Reactions and crossreactions of a rabbit anti-H2 antigen serum

A rabbit antiserum raised against highly purified, papain-solubilized H-2d antigens contained two sets of non-crossreacting antibodies directed against each one of the two H-2 antigen subunits. The antiserum recognized only 12,000 and 47,000 dalton polypeptide chains when splenocyte membrane glycoproteins were analysed. Among the molecules precipitated with the rabbit antiserum all H-2K and D antigens were present. In addition to regular H-2K and D antigens minor amounts of material with the typical H-2 antigen subnit structure, but lacking alloantigenic determinants, were precipitated by the antiserum. These 'non-H-2 antigens' were produced in relatively greater amounts by T-cells than by B-cells. Both sets of antibodies in the rabbit antiserum reacted with the TL antigens demonstrating that there is an immunological crossreactivity between the classical alloantigenic H-2 antigen chain and the alloantigenic TL antigen chain. The F9 cell line, believed to represent cells at the morula stage, display H-2 antigen-like structures as revealed by the rabbit antiserum.

A highly sensitive and reproducible microcytotoxicity assay for demonstrating cytotoxic antibodies to cell surface antigens

An assay is described, in which target cells are suspended in diluted rabbit serum (as a source of complement) and pipetted into the wells of Falcon No. 3034 Microtest plates, to which diluted antiserum (or control serum) has been added. After 18--24 h of incubation, the plates are stained with crystal violet and counted using an electronic Quantimet image analyzer. This technique could reproducibly detect complement-dependent cytotoxic antibodies to target cell H-2 antigens and to MuLV gp70 on cell lines which expressed this antigen, at serum dilutions up to 1:160,000. The cytotoxic antibodies could be removed by absorption of sera with cells carrying the respective antigens.

Tumour-specific complement-dependent serum cytotoxicity against a chemically induced rat hepatoma

A short-term 51Cr release test was used for the detection of complement-dependent cytolytic activity of syngeneic serum for transplanted aminoazo dye-induced rat hepatoma cells in suspension. Serum samples from rats bearing intraperitoneal implants of one hepatoma (hepatoma D23) were specifically cytotoxic for hepatoma D23 target cells, although this activity was not detected in sera from donors bearing subcutaneous tumour grafts. Other sera containing demonstrable IgG antibodies reactive in membrane immunofluorescence tests with individually distinct tumour-specific antigens or tumour-associated embryonic antigens were not cytotoxic for hepatoma D23 cells; and even though serum from donors carrying intraperitoneal tumour grafts contained tumour-specific IgG antibody, the complement-dependent reactivity was confined to the 19s region of fractionated tumour-bearer serum. These findings are discussed in relation to the development of humoral responses in the tumour-bearing host and with regard to the significance of the availability of an objective and reproducible assay for measuring humoral responses directed against the tumour-specific antigens associated with chemically induced rat tumours.

Inhibitory effects of plasma from uraemic patients on human mononuclear phagocytes cultured in vitro

Human mononuclear phagocytes were cultured in plasma from uraemic patients. The presence of uraemic plasma during the engulfment or digestion of 125I-labelled Candida albicans did not inhibit these functions in mononuclear phagocytes cultured for 8 days under normal condition. When normal human macrophages were cultured in the presence of uraemic plasma for 2-4 days, a marked detachment of the cells from the glass coverslips was registered. The phagocytic function of the remaining cells was impaired. Creatinine, urea and methylguanidine in concentrations higher than those usually measured in plasma from uraemic patients did not influence the functional properties of the cells. The inhibitory effect of uraemic plasma on the mononuclear phagocytes is suggested as an explanation for the increased frequency of infections in uraemic patients.

HLA-B specificities and w4, w6 specificities are on the same polypeptide

The human alloantigenic specificities w4 and w6, which are products of a diallelic system genetically associated with the HLA-B locus, have been solubilized by papain digestion of membranes from the lymphoblastoid cell line, RPMI 4265. The w4 and w6 specificities copurified with the HLA-B locus products, HLA-B7 and HLA-B12. Sequential immunoprecipitation experiments were performed using alloantisera to HLA-B7, HLA-B12, w4 and w6, and a purified HLA-B7, B12, w4, w6 antigen pool labeled with 125I-Bolton-Hunter reagent. These experiments demonstrated directly that HLA-B7 and w6, which are genetically associated with each other, are different antigenic determinants on the same molecule, while HLA-B12 and w4, also genetically associated, are distinct antigenic determinants on a second molecule. Arguments are presented which suggest that the HLA-B determinants and w4, w6 determinants are in fact on the same polypeptide, and the genetic implications of the findings are discussed.

Studies with B-cell allo- and hetero-antisera: parallel reactivity and special properties

The expression of B-cell antigens on various cell populations was studied through the use of human alloantisera and with heteroantisera raised to preparations of the alloantigen bearing molecules isolated from B-cell lines. The allo-and hetero-antisera competed with each other in blocking experiments and gave closely parallel results, reacting with normal and leukemic B lymphocytes, monocytes, E-rosette-negative acute lymphatic leukemias, all acute and certain chronic myelogenous leukemias, and a minor population of cells in fetal spleen and liver. These highly immunogenic surface components appeared to comprise the dominant B- cell specific plasma membrane determinants. Neither type of antiserum reacted with any but a minor population of normal or pokeweed-mitogen-transformed T cells, fetal thymic lymphocytes, E-rosette-positive acute lymphatic leukemias, or Sezary-cell leukemia. Through the use of these antisera evidence was obtained that Fc-receptor-bearing Ig-negative lymphocytes were divisible into two groups according to the presence or absence of the B-cell antigens. Both hetero- and allo-antisera blocked binding of immune complexes or antibody-coated ox erythrocytes to Fc receptors on B cells. F(ab')2 fragments of the heteroantibodies strongly inhibited antibody-dependent cell-mediated killing.

Quantitative isolation of mouse thymus leukemia antigen, TL

This report describes the purification of TL from papain digests of a tumor line (ASL1) and of an established cell line (L251A). Through the use of gel filtration and ion-exchange chromatography, the TL was purified approximately 100-fold with respect to the original digest. It was noted that the TL isolated from ASL1 had a specific activity 1.4 times higher than that isolated from L251A. The reason for this anomaly is unexplained. However, this work indicates that classical methods of protein chemistry can be used in the purification of these membrane components which are present in only small amounts on the cell surface.

HLA antigens: rabbit antisera reacting with all A series or all B series specificities

Papain-solubilized HLA antigens giving only two bands of 34 000 mol.wt. and 11 000 mol.wt. by sodium dodecyl sulfate (SDS) gel electrophoresis and isolated from the cultured human lymphoblastoid cell line RPMI 4265, have been used to prepare antisera in rabbits. Antisera were raised against soluble products of the A (or 1 st) series of HLA, bearing the determinant HLA-A2 (A2 substance), or against a mixture of products of the B (or 2nd) series of HLA, bearing the determinants HLA-B7 and HLA-B12 (B7-12 substances). Rabbit antisera to A2 substance reacted primarily with A2 substance on Ouchterlony analysis, showing an apparent spur of cross-reactivity with B7-12 substances. Rabbit antisera to B7-12 substances reacted primarily with B7-12 substances, giving a spur of cross-reactivity with the A2 material. Neither antiserum precipitated beta-2microglobulin. Both types of sera reacted with membrane molecules of 43 000 mol.wt. and 11 000 mol.wt. by immune precipitation and gel electrophoresis in SDS of detergent-solubilized radiolabeled membranes from cultured cell lines, as predicted for sera directed towards HLA antigens. F(ab')2 fragments of the antibodies blocked the complement-mediated cytotoxicity of all HLA alloantisera tested for human peripheral blood lymphocytes. After absorption with B7-12 substances, F(ab')2 fragments of antisera to A2 substance only blocked the cytotoxicity of HLA alloantisera to A series specificities. After absorption with A2 substance, F(ab')2 fragments of rabbit antisera to B7-12 substances only blocked the cytotoxicity of HLA alloantisera to B series specificities. The results prove the existence of shared antigenic determinants between all members of the same series. These findings support the genetic evidence that A series HLA antigens are allelic products of a single locus, while B series HLA antigens are allelic products of a separate locus, by establishing some invariance of structure, presumably amino acid sequence, between members of the same series. The apparent cross-reactivity between A2 substance and B7-12 substances, and the ability of the unabsorbed F(ab')2 preparations to block the cytotoxicity of all HLA alloantisera, suggests that some determinants are common to both HLA loci. This may be considered to support the hypothesis that the two loci arose by gene duplication.