Substances involved in the natural resistance of fish to infection–A review

Natural‘antibodies’are substances found in the blood of animals that have not been immunised against infective agents. However, exposure to these agents or to cross-reacting antigens may well have taken place. Fish contain naturally-occurring, relatively nonspecific, lectin-like proteins or glycoproteins, which are distinct from immunoglobulins, and which react with a wide variety of antigens and may confer some degree of immunity against natural infection. In most cases the cause of the antigenic stimulus is not obvious although the formation of these‘antibodies’may have been brought about by exposure to various micro-organisms. Many of these antibody-like molecules behave in a similar manner to immune antibodies or immunoglobulins and cross-react with specific carbohydrate moieties on the cell walls of bacteria, erythrocytes and certain other cellular antigens, due to the presence of similar antigenic determinants. It is difficult to ascribe an appropriate definition to the term‘natural antibody’. In fish, these‘antibodies’have been so designated on the basis of functional rather than structural criteria. Such naturally-occurring, low grade, antibody-like‘immune’substances include‘acute phase’proteins, lysozyme and chitinase, interferon, agglutinins, lysins, complement and properdin, precipitins, and non-immunoglobulin, lectin-like molecules. In addition to the above non-immunoglobulin materials, natural immunoglobulins identifiable as IgM have also been reported in fish. Furthermore, mucus contains many biochemical agents capable of reaction against infective organisms and thus providing the host with an immediate or a first line of defence mechanism. This review compiles some of the relevant information in the literature concerned with natural‘immune’substances, present in the serum and mucus of fish, involved in protection against pathogens. Wherever possible the basic physicochemical properties of these substances are indicated and their potential immunobiological functions discussed.     

Analysis of the B-G antigens of the chicken MHC by two-dimensional gel electrophoresis

The B-G antigens of the chicken major histocompatibility complex (MHC) have been analyzed by high resolution two-dimensional (2-D) gel electrophoresis. Monoclonal antibodies recognizing a widely shared B-G determinant were used for immunoprecipitating the B-G antigens from radioiodinated, detergent-solubilized erythrocyte membrane preparations. The B-G antigens produce a variety of patterns on 2-D gels. The number of polypeptides within a B-G pattern varies among haplotypes from single polypeptide arrays showing slight microheterogeneity to complex patterns which contain as many as four or five polypeptide arrays differing in relative mobility and isoelectric point. Many of the patterns, but not all, include a polypeptide of Mr =48 kd focusing near pH 6.9. At present it is not understood whether the multiple polypeptides within some B-G patterns represent the expression of multiple B-G genes or whether they are the result of modifications of single gene products during biosynthetic processing. 2-D gel analyses were also used to confirm the assignment of the same B-G haplotype in several different inbred flocks and the fate of the B-G antigens in two B system recombinant haplotypes. The 2-D gel patterns of these highly polymorphic antigens provide evidence for a complexity of the B-G locus not previously demonstrated. This technique may serve to define more objectively the diverse chicken MHC haplotypes which are now recognized and characterized only by serological techniques using alloantisera and monoclonal antibodies with varying cross-reactivities.     

Antigens similar to major histocompatibility complex B-G are expressed in the intestinal epithelium in the chicken

A monoclonal antibody directed against the erythrocytic B-G antigens of the major histocompatibility complex (MHC) of the chicken, an antiserum raised against purified erythrocytic B-G protein, and a cDNA probe from the BeckmanB-G subregion were used to look for evidence of the expression ofB-G genes in tissues other than blood. Evidence has been found in northern hybridizations, in immunoblots, and in immunolabeled cryosections for the presence of B-G-like antigens in the duodenal and caecal epithelia. Additional B-G-like molecules may be expressed in the liver as well. The BG-like molecules in these tissues appear larger and somewhat more heterogeneous than the B-G antigens expressed on erythrocytes. Further characterization of these newly recognized B-G-like molecules may help to define a function for the enigmatic B-G antigens of the MHC. al. 1977; Miller et al. 1982, 1984; Salomonsen et al. 1987; Kline et al. 1988), and in the multiplicity of B-G restriction fragment patterns found in genomic DNA from different haplotypes (Goto et al. 1988; Miller et al. 1988; Chaussé et al. 1989). The B-G antigens have contributed, together with the B-F (class I) and B-L (class II) antigens, to the definition of over 27 B system haplotypes in experimental flocks (Briles et al. 1982). Yet the function of the B-G antigens remains entirely unknown. No mammalian counterparts have been identified, although the possibility remains that there may be similar antigens among the blood group systems of mammals. In an effort to define a function of the B-G antigens, a recently cloned B-G sequence (Miller et al. 1988; Goto et al. 1988) and antibodies to the B-G polypeptides (Miller et al. 1982, 1984) were used to examine other tissues for evidence of B-G expression.     

Differential expression of cell surface antigens of canine keratinocytes defined by monoclonal antibodies

Nine monoclonal antibodies (MAb) directed against cell surface antigens of canine keratinocytes define distinct keratinocyte subpopulations owing to the differential expression of these antigens during the process of differentiation and depending on the tissue location of the cells. There was distinct antigenic heterogeneity between the different layers of stratified squamous epithelium and between stratified squamous epithelial of different tissue origin. Two MAb reacted only with antigens expressed by esophageal mucosa. Three MAb bound to antigens on keratinocytes of the suprabasilar and granular layers of stratified squamous epithelia, and they crossreacted with the transitional epithelial cells of the urinary tract. Two MAb reacted with antigens only expressed on differentiated cells, superficially located in the stratified squamous epithelium. The use of these MAb as markers for keratinocytes in studies on the characterization and differentiation of keratinocytes, as well as in tumor diagnosis and allograft transplantation, is discussed.     

Quantitation of changes in cell surface determinants during skeletal muscle cell differentiation using monospecific antibody

The differentiation of skeletal muscle is characterized by recognition, alignment, and subsequent fusion of myoblast cells at their surfaces to form large, multinucleated myotubes. Monoclonal antibodies were used to investigate anti-genie changes in the cell surface membrane specific for various stages of myogenesis. Chick embryonic skeletal muscle cells were cultured in vitro to the desired stage of differentiation and then injected into BALB/c mice. Spleen cells from the immunized mice were hybridized with NS-1 or P3 8653 mouse myeloma cells. Hybrid cell clones were selected in HAT medium and screened using an indirect radioimmunoassay for the production of monoclonal antibodies specific to myogenic cell surfaces. Target cells for the radioimmunoassay included three stages of myogenesis (myoblasts, midfusion myoblasts, and myotubes) and chick lung cells as a control for polymorphic antigens. Sixty-one clones were obtained which produced antibodies specific for myogenic cells. Thirty-five of these clones were generated from mice immunized with midfusion myoblast stages of myogenesis and 26 were obtained from mice immunized with the later myotube stage of myogenesis. Quantitative measurements by RIA of myogenic determinants per cell surface area on each target cell type revealed that most of the determinants decrease during myogenesis when midfusion myoblasts are used as the immunogen. When myotube stages are used as the immunogen, more determinants increase with cell differentiation. Therefore, the most common pattern of determinant change is for them to be present at all stages of myogenesis but to vary quantitively through development. There are determinants unique to each stage of myogenesis and marked quantitative differences within a cell stage for each determinant.     

Antibodies to nonpolymorphic determinants of the Thy-1 molecule inhibit T cell proliferation

The effect of anti-Thy-1 monoclonal antibodies on murine mixed lymphocyte reactions and concanavalin A-induced mitogenesis were investigated. It is demonstrated that rat antibodies against nonpolymorphic determinants of the murine Thy-1 antigen inhibited cell proliferation in the absence of complement. In contrast, antibodies against polymorphic determinants of Thy-1 had no effect on T cell activation. Inhibition of T cell proliferation did not depend on the isotype of the blocking antibody, because both IgM and IgG antibodies against monomorphic determinants were inhibitory, whereas IgM or IgG antibodies against allotypic determinants were inactive. In addition, the blocking activity could not be attributed to the xenogeneic (rat) origin of the antibodies to nonpolymorphic Thy-1 determinants, because rat anti-Thy-1.2 antibodies had no effect on cell activation. Thus, the efficacy of anti-Thy-1 antibodies as T cell inhibitors was determined by the antibody specificity. The suppressive mechanism of anti-Thy-1 antibodies was effective throughout the entire course of mixed lymphocyte reactions. Addition of antibodies at any time point during the first 90 hr of a 120-hr mixed lymphocyte culture resulted in significant suppression of the proliferative response. However, in some cases an early enhancement preceded suppression of the response. The modulation of proliferative responses by anti-Thy-1 did not result from a nonspecific mitogenic effect of the antibodies on T lymphocytes, because no effects were observed when antibodies were added to responder cells alone. These results suggest that the Thy-1 molecule, or a molecule that is located on the cell membrane in close proximity to the Thy-1 antigen, is involved in the activation of T lymphocytes.     

The high background immune reactivity of mice to polymorphic determinants on xenogeneic erythrocytes: theoretical and practical implications

Background responses have been assessed by fusing lipopolysaccharide- (LPS) stimulated spleen cells from unimmunized mice with MOPC 315.43 myeloma cells and screening the hybrids for the production of antibody against chicken red blood cells (CRBC). Clones specific for CRBC represented about 1% of total hybrid clones (1000 to 5000 clones were obtained per mouse). The majority of the anti-CRBC clones (greater than 95%) secreted antibody against polymorphic CRBC determinants (present on CRBC from some but not all chickens) rather than species-specific determinants present on all CRBC. Some of the polymorphic determinants were linked to the B locus (the MHC of the chicken) and some were non-B antigens. The relative amount of these 2 categories varied slightly according to the mouse strain. These results agree well with the specificities of natural mouse antibody and rosette-forming spleen cells. The response of immunized mice against CRBC and human RBC was also selective for polymorphic determinants. These results have considerable importance for the use of xenogeneic RBC as "standard" antigens, and are interpreted in terms of a model for the advantages of genetic polymorphism as a protection against antigen mimicry by parasites.     

Class IgG,IgM and IgA antibodies againstStaphylococcus aureus antigens in human serum and saliva

Using the ELISA method antibodies against the sonicate, teichoic acid (TA) and exoproducts ofStaphylococcus aureus were determined in sera and saliva of healthy individuals. Main serum antibodies against all the antigens used were shown to be class IgG antibodies. However, antigens of the sonicate stimulated significantly even the systemic IgA response. In the saliva class IgA antibodies predominated, but IgG antibody levels against TA and exoproducts approached the level of IgA antibodies. Levels of IgM antibodies against all antigens tested were low in both the serum and saliva which corresponds with the anamnestic type of response. On the basis of these results one may assume that not only IgG, but also IgA antibodies are important in the systemic immunity against staphylococcal infection and in the immunity of mucous membranes; besides IgA, even class IgG antibodies play an important role.     

Antibodies against PfEMP1, RIFIN, MSP3 and GLURP are acquired during controlled Plasmodium falciparum malaria infections in naïve volunteers

Antibodies to polymorphic antigens expressed during the parasites erythrocytic stages are important mediators of protective immunity against P. falciparum malaria. Therefore, polymorphic blood stage antigens like MSP3, EBA-175 and GLURP and variant surface antigens PfEMP1 and RIFIN are considered vaccine candidates. However, to what extent these antibodies to blood stage antigens are acquired during naive individuals' first infections has not been studied in depth. Using plasma samples collected from controlled experimental P. falciparum infections we show that antibodies against variant surface antigens, PfEMP1 and RIFIN as well as MSP3 and GLURP, are acquired during a single short low density P. falciparum infection in non-immune individuals including strain transcendent PfEMP1 immune responses. These data indicate that the immunogenicity of the variant surface antigens is similar to the less diverse merozoite antigens. The acquisition of a broad and strain transcendent repertoire of PfEMP1 antibodies may reflect a parasite strategy of expressing most or all PfEMP1 variants at liver release optimizing the likelihood of survival and establishment of chronic infections in the new host.     

Expression of monoclonal antibody-defined cell surface antigens during rat brain development

Using single-cell suspensions of mechanically dissociated, prenatal BDIX-rat brain cells (13th, 15th, and 21st days after fertilization) for immunization, we have established a collection of 37 monoclonal antibodies (Mabs) directed against neural cell surface determinants. The developmental-stage-dependent expression of cell-surface antigens recognized by these Mabs was analyzed both on plasma membranes isolated from whole brains of BDIX rats (prenatal days 13-22 and adults) using an indirect 125I solid-phase radioimmunoassay, and on intact BDIX-rat brain cells (prenatal days 13-22) using a fluorescence-activated cell sorter. Different types of developmental stage-dependent profiles of Mab binding were found, these being indicative of the presence of neural cell surface determinants whose expression increases, decreases, or does not change with brain development. Some of the Mab-binding profiles showed transient changes as a function of developmental stage. These Mabs are currently being used for the characterization, reproducible identification, and isolation of neural cell subpopulations of the developing rat brain, with the aim of investigating the cell type dependence and developmental (differentiation) stage dependence of malignant transformation following pulse exposure to the carcinogen N-ethyl-N-nitrosourea at defined stages of brain development.     

A malaria vaccine based on the polymorphic block 2 region of MSP-1 that elicits a broad serotype-spanning immune response

Polymorphic parasite antigens are known targets of protective immunity to malaria, but this antigenic variation poses challenges to vaccine development. A synthetic MSP-1 Block 2 construct, based on all polymorphic variants found in natural Plasmodium falciparum isolates has been designed, combined with the relatively conserved Block 1 sequence of MSP-1 and expressed in E.coli. The MSP-1 Hybrid antigen has been produced with high yield by fed-batch fermentation and purified without the aid of affinity tags resulting in a pure and extremely thermostable antigen preparation. MSP-1 hybrid is immunogenic in experimental animals using adjuvants suitable for human use, eliciting antibodies against epitopes from all three Block 2 serotypes. Human serum antibodies from Africans naturally exposed to malaria reacted to the MSP-1 hybrid as strongly as, or better than the same serum reactivities to individual MSP-1 Block 2 antigens, and these antibody responses showed clear associations with reduced incidence of malaria episodes. The MSP-1 hybrid is designed to induce a protective antibody response to the highly polymorphic Block 2 region of MSP-1, enhancing the repertoire of MSP-1 Block 2 antibody responses found among immune and semi-immune individuals in malaria endemic areas. The target population for such a vaccine is young children and vulnerable adults, to accelerate the acquisition of a full range of malaria protective antibodies against this polymorphic parasite antigen.     

Reactivity of monoclonal islet cell antibodies on normal hyperplastic and neoplastic thyroid C-cells

Summary Thyroid C-cell reactivity to 15 monoclonal antibodies raised against a series of pancreatic islet cells (H[human]ISL, B[bovine]ISL and R[rat]ISL) was evaluated using an indirect immunoperoxidase technique on frozen thyroid sections. Of the monoclonal anti-islet cell antibodies, five reacted specifically with bovine C-cells or human hyperplastic and neoplastic C-cells but not with follicular cells. Two monoclonal antibodies of the bovine series showed strong immunoreactivity with C-cells and only a weakly positive immunostaining of follicular cells. Five monoclonal antibodies reacted with both thyroid C-cells and follicular cells, whereas 3 monoclonal anti-islet cell antibodies did not stain any cell type of the thyroid. In human medullary carcinomas, calcitonin- and somatostatin-producing neoplastic cells were immunoreactive with the same monoclonal antibodies as were normal human C-cells. The protein bands identified by the monoclonal antibodies in human medullary carcinomas had the same molecular weight as those from pancreatic islet extracts. Our study demonstrates the presence of similar differentiation antigens on thyroid C-cells and pancreatic islet cells; this further illustrates common modes of differentiation and specialisation of these embryologically different members of the dispersed neuroendocrine system. The crossreactivity of seven of the monoclonal antibodies investigated with follicular epithelium of the thyroid suggests the existence of common antigenic determinants in different endocrine organs and may partly explain the multiple organ autoimmune response found in patients with polyendocrine diseases.     

The immune response to Schistosoma mansoni infections in inbred rats. VI. Regulation by T cell subpopulations

These studies assess the roles of subpopulations of T lymphocytes in inducing and modulating resistance to Schistosoma mansoni. CDF rats were depleted of RT 7.1+ (anti-Pan-T), W3/25+ (anti-T helper/inducer), or OX8+ (anti-T suppressor) cells by the in vivo administration of monoclonal antibodies (mAb). The development of parasites and immunity to challenge by S. mansoni were compared with results in undepleted normal and congenitally athymic rats. Discrete subpopulations of T lymphocytes were adoptively transferred to ascertain effects upon parasite development and the protective immune response. In vitro studies, involving utilizing cocultivation of cell subpopulations +/- cyclosporin A, were utilized to dissect mechanisms. Depletion of T lymphocytes by anti-RT7.1 mAb and anti-W3/25 mAb resulted in augmented initial worm development, suboptimal resistance, and decreased antibody and delayed-type hypersensitive reactivity directed against schistosome antigens. Depletion with OX8 mAb produced opposite effects. The adoptive transfer of T cell subpopulations produced concordant results with T cell regulation expressed B cell-dependent effector mechanisms. The coadoptive transfer of cells resulted in the suppression of resistance afforded by the W3/25+ cells by OX8+ cells, which could be augmented in vitro by cyclosporin A. Thus, protective immunity to S. mansoni in rats is regulated by discrete subpopulations of T lymphocytes. The findings suggest the possibility of selective immune regulation of resistance based on the manipulation of specific T cell subpopulation.     

Characterization of two distinct disulfide-linked B-G molecules in the chicken

Alloantisera specific for B-G antigens recognized a complex of molecules of apparent molecular weights of 90 and 98 Kd under nonreducing conditions and molecules of 40, 44, and 48 Kd under reducing conditions on both embryo- and adult-derived peripheral red blood cells (RBC). The chicken B-G molecules produced a unique two-dimensional "diagonal" pattern. Two antisera permitted the characterization of the complex B-G molecular profile as a homodimer composed of 48-Kd subunits and as a heterodimer composed of 40- and 44-Kd subunits. A rabbit antiserum produced against B-G molecules preferentially recognized the 48-Kd reduced molecules, suggesting that the 90-Kd molecule was a homodimer composed of two 48-Kd molecules. One B-G reagent was capable of recognizing only the 98-Kd nonreduced B-G molecule that gave rise to 40- and 44-Kd molecules under reducing conditions, suggesting that the 98-Kd molecule was a heterodimer composed of 44- and 40-Kd subunits. Adult chicken B-G2 molecules produced a variety of two-dimensional isoelectric focusing/sodium dodecyl sulfate-polyacrylamide gel electrophoresis (IEF/SDS-PAGE) patterns depending on the characteristics of the reagent employed in the immunoprecipitation. B-G molecules were immunoprecipitated from primitive and definitive chicken RBCs but not from any nonerythroid cells tested. B-G molecules were not expressed by avian erythroblastosis virus (AEV)-transformed erythroleukemia cells, nor were they induced to appear with butyric acid-induced erythroid differentiation.     

MHC-like molecules in some nonmammalian vertebrates can be detected by some cross-reactive xenoantisera

Rabbit antisera raised to human and chicken MHC molecules were used to immunoprecipitate cross-reactive molecules from biosynthetically and cell surface-labeled spleen and/or blood cells of representative vertebrate species. Five major points emerged: 1) There were many nonspecific cross-reactions using these techniques, so various criteria were developed to distinguish these from true MHC-like molecules. 2) Only very small subpopulations of immunogen-specific antibodies cross-reacted with MHC-like molecules in other nonmammalian species. These subpopulations were different for each species and even within a species, sometimes being so limited as to behave like alloantisera. This led to a very scattered pattern of true cross-reactions that sometimes failed to reflect the properties of the bulk antibody population. 3) Antisera containing antibodies to class II beta- and class I alpha-chains cross-reacted better and more widely than those to B-G, class II alpha and, in general, beta 2-microglobulin. 4) Some cross-reactive antibodies were clearly directed to epitopes on the surface of the mature heterodimers, but many seemed to recognize nonlinear cryptic determinants, presumably in the contact regions between the chains. These latter antibodies recognized biosynthetic intermediates and also a variety of unusual cell surface MHC-like molecules present in reptile and amphibian, but absent in the mammal and chicken cells tested. These included E homodimers whose relationship to chicken B-G molecules is unknown. 5) MHC-like molecules were identified in a bird, three reptiles, and two amphibians, but no molecules with the expected properties were found with these reagents in any of the fish tested.     

Antibodies against PfEMP1, RIFIN,MSP3 and GLURP Are Acquired during Controlled Plasmodium falciparum Malaria Infections in Na?ve Volunteers

Antibodies to polymorphic antigens expressed during the parasites erythrocytic stages are important mediators of protective immunity against P. falciparum malaria. Therefore, polymorphic blood stage antigens like MSP3, EBA-175 and GLURP and variant surface antigens PfEMP1 and RIFIN are considered vaccine candidates. However, to what extent these antibodies to blood stage antigens are acquired during naive individuals'' first infections has not been studied in depth. Using plasma samples collected from controlled experimental P. falciparum infections we show that antibodies against variant surface antigens, PfEMP1 and RIFIN as well as MSP3 and GLURP, are acquired during a single short low density P. falciparum infection in non-immune individuals including strain transcendent PfEMP1 immune responses. These data indicate that the immunogenicity of the variant surface antigens is similar to the less diverse merozoite antigens. The acquisition of a broad and strain transcendent repertoire of PfEMP1 antibodies may reflect a parasite strategy of expressing most or all PfEMP1 variants at liver release optimizing the likelihood of survival and establishment of chronic infections in the new host.     

Cross-reactivity between human and murine lymphocyte antigens: III. Reactivity of H-2 allo- and xenoantisera with human lymphoid cells

H-2 alloantisera and antimouse lymphocyte xenoantisera react with 14%–100% of human lymphocytes from a panel of at least 80 unrelated people. Population and family studies did not reveal HL-A specificity of such lymphocytotoxic antibodies but indicated that the antibodies are directed against polymorphic antigenic determinants inherited in association with HL-A antigens. H-2 allo- and xenoantisera absorbed with human lymphoid cells and a panel of platelets bearing all the known HL-A specificities were still cytolytic when tested against murine lymphocytes, suggesting that only a small proportion of the heterogeneous population of H-2 antibodies react with human lymphocytes. On the other hand, HL-A alloantisera could be absorbed by lymphocytes from certain murine strains. These results suggest that the crossreactivity between human and murine lymphocytes is caused by antigens common to several HL-A (or H-2) types or by antigens linked to HL-A but not identical with them.     

MHC coded B-G homodimer and heterodimer heterogeneity among different chicken lines

Chicken lines were classified into six distinct groups based on expression of B-G molecules by peripheral red blood cells (RBC). In addition to the previously reported 48 kD subunits of homodimeric B-G molecules, subunits of 60, 40, and possibly 20 kD were detected in certain of the chicken lines. Several of the chicken lines express the previously reported 40 and 44 kD subunits of heterodimeric B-G molecules; however, B21B21 chickens expressed 44 kD subunits only and B5B5 and B13B13 chickens did not express detectable levels of any heterodimeric-related molecules. These studies provide further evidence for the polymorphic nature of the B-G antigens.     

The immunobiology of T cell responses to Mls-locus-disparate stimulator cells. II. Effects of Mls-locus-disparate stimulator cells on cloned, protein antigen-specific, Ia-restricted T cell lines

Cloned, protein antigen-specific, Ia-restricted T cell lines frequently (approximately 20%) also respond strongly to stimulator cells from strains expressing stimulatory alleles at the chromosome 1-encoded Mls-locus. Furthermore, such responses are blocked by monoclonal antibodies specific for Ia antigens expressed by the stimulator rather than the responder cells. However, such responses show no specificity for polymorphic determinants on Ia molecules, although in such responses, as in primary and secondary T cell responses to stimulating Mls-locus alleles, I-E molecules appear to play a central role. These results, combined with the unique immunobiology of the primary T cell proliferative response to Mls-locus-disparate stimulator cells, suggest to us that this response involves the interaction of the receptor on T cells for antigen:self Ia with a relatively nonpolymorphic region of Ia glycoproteins. This hypothesis is supported by the observation that a monoclonal antibody to the T cell receptor will inhibit both responses, although the response to Mls-locus-disparate stimulators appears to be more sensitive to these antibodies. We propose that the interaction of the T cell receptor with Ia is stabilized by a cell interaction molecule encoded or regulated by the Mls-locus gene product permitting the T cell receptor:Ia glycoprotein interaction to lead to T cell activation.     

“Natural” Antibodies to chicken MHC antigens are present in mice,rats, humans,alligators and allogeneic chickens

In the absence of any deliberate immunization, mice, rats, humans and alligators all have detectable titers of antibody against chicken red blood cells (CRBC's). Remarkably, this antibody is directed predominantly against private or public determinants of MHC proteins on the CRBC's, and little or no antibody is directed against species-specific determinants on MHC or other proteins, including other polymorphic blood group antigens. In chickens, natural antibody can be detected against CRBC's from all chickens differing at the MHC locus, but natural antibodies against other polymorphic antigens are not detected. Using a rosette-forming cell (RFC) assay, we have also shown that a large percentage of mouse spleen cells will rosette with chicken erythrocytes, and that the majority of these RFC's also recognize polymorphic antigens.