The self determinants recognized by human virus-immune T cells can be distinguished from the serologically defined HLA antigens

The self specificity of human influenza virus-immune cytotoxic T cells has been analyzed in order to clarify the relationship between the self antigens that they recognize and the serologically defined HLA-A and -B antigens. Virus-immune effectors from HLA-A2-positive donors were tested on panels of virus-infected target cells from donors who were either HLA-mismatched or matched only for HLA-A2. Virus-immune T cells from 11 out of 11 A2-positive donors lysed all A2-matched virus-infected target cells (and no HLA-mismatched targets), except that each of these effector cells consistently failed to lyse virus-infected target cells from one A2-positive donor (designated M7). Although the A2 specificity of donor M7 could also be distinguished from the A2 antigen of other donors by alloimmune cytotoxic T cells, no differences in the A2 antigen of donor M7 could be defined by extensive serologic analyses. These results indicate that there is a strong but incomplete association between a self antigen recognized by virus-immune T cells and the serologically defined HLA-A2 specificity.     

A requirement for physical linkage between determinants recognized by helper molecules and cytotoxic T cell precursors in the induction of cytotoxic T cell responses

It has long been understood that both antibody and delayed-type hypersensitivity responses are induced through collaborative events in which the determinants recognized by the precursor cells must be physically linked to the determinants recognized by the helper. Although it is clear that the generation of memory cytotoxic T lymphocyte precursors (CTLp) involves linked recognition of determinants, the induction of CTL responses has been viewed as being dependent upon interleukin 2 (IL 2), which could be provided by a helper cell, but independent of requirements for antigen bridging. In this work, we have designed a system that lacks exogenous IL 2 by using as our source of help, antigen-specific helper molecules derived from helper T cells. These soluble helper molecules are uncontaminated by IL 2 and unlike a helper cell, are unable to produce IL 2. Helper molecules specific for chicken red blood cells (Crbc) and for a synthetic polypeptide, poly 18, were tested. Thymocyte responders require a source of help to respond to alloantigens intrinsically expressed on the surface of adherent stimulator cells. To analyze the mechanism whereby the helper molecules acted, we used a system involving recognition of haptenic and carrier determinants that were physically linked by virtue of being located on the same cell surface (intra-structural linkage). Adherent stimulator cells were pulsed with Crbc or poly 18 so that the alloantigens recognized by the thymocyte CTLp (intrinsically expressed class I) were either linked or unlinked to the carrier determinants (Crbc or poly 18) presented by the adherent cells and recognized by the helper molecules. Both types of helper molecule were shown to be antigen-specific in crisscross experiments. The helper molecules specific for Crbc were able to induce the thymocyte CTLp only when both hapten and carrier were present on the same stimulator cell surface. Because we were not able to detect a requirement for H-2-restricted recognition of carrier antigen, this inductive event must be viewed as requiring linked associative recognition of determinants, but being noncognate. In contrast, the helper molecules recognizing poly 18 showed a requirement for both physical linkage of determinants and for H-2 restricted recognition, indicating that the mechanism of induction was cognate in nature. Therefore, we have shown that interactions between CTLp and soluble, antigen-specific, helper cell-derived inductive molecules are similar in nature to those of other T cell precursors and of B cells in the stringent requirement for close physical proximity achieved by linked or cognate recognition of determinants across an antigen bridge.     

Interaction of alpha 1 with alpha 2 region in class I MHC proteins contributes determinants recognized by antibodies and cytotoxic T cells

The structure-function relationship of individual coding regions of class I mouse major histocompatibility complex proteins was studied by a combination of recombinant DNA, gene transfer techniques, and serologic and functional characterization. To examine the role of alpha 1 and alpha 2 regions in antibody and CTL recognition, the third exon of H-2Dd, Kd, and Ld transplantation antigen genes was replaced by the homologous coding region of the Qa-2-coded class I gene, Q6. We have chosen to carry out the exon shuffling experiments between these two different types of class I genes, because they are structurally similar and did not evolve to carry out identical functions. Therefore, it is less likely that the hybrid proteins will fortuitously recreate alpha 1-alpha 2 controlled functionally important determinants. The replacement of H-2 alpha 2 coding region with its Q6 counterpart had different effects on the expression of the three genes. The mutant H-2Dd gene transfected into L cells was expressed at high levels and retained several of the serologic determinants found on parental H-2Dd and Q6 domains. The serologic epitopes on the mutant H-2Kd-transfected cells were detectable at very low levels, whereas the product of the mutant H-2Ld gene could not be identified at all. Analysis of cells transfected with mutant H-2Dd gene with alloreactive and minor antigen(s)-restricted cytotoxic T cells indicated that the hybrid proteins lost the ability to be recognized by T cells. Our data suggest that cytotoxic T cells recognize conformational determinants composed of amino acids from alpha 1 and alpha 2 regions. Alternatively, it could be proposed that T cell recognition sites located in a single alpha 1 or alpha 2 protein region are susceptible to distortion upon alpha 1-alpha 2 interactions. Such susceptibility to conformational changes of the amino-terminal domain of transplantation antigens could be of functional importance for H-2-restricted antigen presentation.     

Effects of fourH-2K mutations on virus-induced antigens recognized by cytotoxic T cells

Lysis of ectromelia- or LCM virus-infected macrophage target cells by virus-specific cytotoxic T cells from mice immunized with the homologous virus occurred only where donors of T cells and target cells shared eitherH-2K orH-2D genes. With both viruses, use of T cell or target cell donors bearing mutations (B6.C-H-2^ba, B6-H-2^bh, B6-H-2^bg1, and B6-H-2^bg2), all of which apparently occurred in the same single genetic element in theH-2K^b region, abolished (H-2^ba) or impaired (H-2^bh,H-2^bg1 andH-2^bg2) lysis in T cell-target cell combinations that shared (apart from the mutations) all other genes in theK, I-A, orI-B regions of theH-2 complex. The data suggest that virus-induced antigenic patterns on infected B6.C-H- 2^ba (mutant) cells are more different antigenically from those on C57BL/6 (wild type) cells than are those on infected cells from the other mutants -B6-H-2^bh, B6-H-2^bg1, and B6-H-2^bg2. (B6.C-H-2^ba× B6 -H-2^bh)F₁ mice behaved like B6-H-2^bh, indicating no complementation, and confirming that theH-2K gene(s) involved in recognition of virus-infected cells by virus-specific T cells behave as a single element. These findings are discussed in relation to the nature of virus-induced antigenic patterns that are recognized by virus-specific cytotoxic T cells.     

Molecular characterization of a subtype of DQw1 recognized by hapten-specific T cells

Previous studies of HLA-restricted antigen recognition by cloned T cells have frequently demonstrated reactivity that did not correlate precisely with the expression of serologically defined HLA specificities. To further explore such discrepancies, we utilized monoclonal antibody (MoAb) blocking, partial NH₂-terminal amino acid sequencing, and Southern blot hybridization techniques to analyze the fine specificity of four autologous trinitrophenyl-specific T cell lines restricted to DR2-linked epitopes. MoAb blocking studies demonstrated that two of these lines recognized determinants on DR molecules while the other two recognized determinants on the same molecule that expresses the DQw1 determinant. However, these latter two lines appeared to recognize a DQw1-related determinant found primarily in association with DR2, but not the other DQw1-associated DR alleles, DR1 and DRw6. To ascertain whether these lines were defining a functional split of DQw1, we performed partial NH₂-terminal amino acid sequencing of the molecules precipitated with a DQw1-specific MoAb (Genox 3.53) from different stimulator lines. The results showed that these T cell lines recognized a subtype of DQw1 that is in linkage disequilibrium with DR2. Moreover, we identified characteristic restriction fragment length polymorphisms with a DQ -specific cDNA that correlated with stimulatory capacity for the DQw1-restricted lines. These results demonstrate that: (1) DQ molecules may provide restriction determinants that are incorrectly assigned to DR molecules on stimulator panel analyses; (2) cloned antigen-specific T cell lines recognize polymorphic regions of class II molecules not distinguished by either conventional typing antisera or xenogeneic MoAb; and (3) the DQw1 epitope(s) is located on a heterogeneous group of DQ molecules that differ from each other in the primary sequence of their chains.Abbreviations used in this paper ATCC American type culture collection; cpm, counts per minute - DNA deoxyribonucleic acid - EBV Epstein-Barr virus - FCS fetal calf serum - MoAb monoclonal antibody - PBMC peripheral blood mononuclear cells - % RAgS percent relative antigen stimulation - RFLP restriction fragment length polymorphism - SDS sodium dodecyl sulfate - S-RPMI supplemented-RPMI - TCL T-cell line - TNP trinitrophenyl     

Differential recognition of the serologically defined HLA-A2 antigen by allogeneic cytotoxic T cells

A comprehensive analysis of human alloimmune cytotoxic T lymphocytes (CTLs) specific for the HLA-A2 antigen identified 11% of HLA-A2 positive cells as outliers. In total, 11 unrelated serologically indistinguishable, but distinguishable by cell-mediated lympholysis (CML) HLA-A2 positive outlier cells were identified. The outlier cells could be subdivided in two subgroups according to reactivity patterns obtained with CTLs directed against the HLA-A2 antigen of outlier cells and their inhibitory capacity in specific competitive inhibition experiments. Thus, the serologically defined HLA-A2 specificity can be divided into at least three subtypes using CTLs specific for the HLA-A2 antigen. Moreover, CTLs specific for an HLA-A2 subtype could be induced when responder cells expressed a different HLA-A2 subtype antigen. On the basis of several family studies, we conclude that the subtype HLA-A2 antigens are inherited in a codominant way.     

Pteridines as a new marker to detect human T cells activated by allogeneic or modified self major histocompatibility complex (MHC) determinants

Evidence is presented that activation of T cells by allogeneic or modified autologous cells leads to the specific production of two distinct molecular compounds of the PT class. One of these two PT has previously been identified as neopterin. The structure of the other is still under investigation. This conclusion was based on the following findings: a) In vitro both PT were produced by T cells proliferating in response to HLA incompatible allogeneic or to autologous EB-virus-transformed or TNP-haptenized cells. b) In vitro stimulation of T cells with mitogenic lectins or protein antigens, of macrophages with zymosan-complement, of B cells with PWM or EB-virus, and of NK effector cells with tumor targets failed to induce comparable release of these PT. c) In vivo increased amounts of both PT were excreted via the urine during allograft rejection and viral infections. It thus appears that the production of these PT is primarily a feature of activated T cells involved in the control of self integrity.     

N protein is the predominant antigen recognized by vesicular stomatitis virus-specific cytotoxic T cells.

The specificity of anti-vesicular stomatitis virus (VSV)-specific cytotoxic T cells was explored with cell lines expressing VSV genes introduced by electroporation. Low levels of nucleocapsid (N) protein were detected on the surface of VSV-infected cells, but N protein could not be detected on the plasma membrane of transfected EL4 cells. Intracellular N protein was detectable by enzyme-linked immunosorbent assay or immunoprecipitation in some of the transfected cell lines but not in others, unless the transfected genes were induced by sodium butyrate. However, all of the stably transfected EL4 cell lines expressing the VSV-Indiana N protein were efficiently lysed by serotype-specific and cross-reactive anti-VSV cytotoxic T cells (CTLs). Primary cross-reactive anti-VSV CTLs appeared to be specific solely for N protein, based on cold-target competition assays using infected and transfected target cells. Cell lines expressing 100- to 1,000-fold less N protein than did VSV-infected cells were efficiently lysed by both primary and secondary anti-VSV CTLs. Cell lines expressing 100-fold less G protein than did VSV-infected cells were not lysed by either population of effectors. Significantly, cold-target competition studies with secondary CTLs demonstrated that N protein-expressing cell lines were more efficient competitors than were VSV-infected cells even though the latter expressed 100- to 1,000-fold more N protein. This was not an artifact of viral infection since infection of the transfected cell lines did not affect their ability to compete. The possibility that cell lines constitutively expressing internal virus proteins present antigen more effectively than infected cells do is discussed.     

Functional clonal deletion of class I-specific cytotoxic T lymphocytes by veto cells that express antigen

Intravenous injection of class I incompatible spleen cells into mice results in a drastic reduction of the recipient's cytotoxic T lymphocyte (CTL) response against the injected, but not against third party, class I antigens when measured in bulk cultures initiated 5 to 6 days after the injection. This specific suppressive effect is partly due to T cells but can also be seen when high numbers of anti-Thy-1 and complement-treated spleen cells of nude mice are injected. Such cells suppressing CTL responses against self histocompatibility antigens are called "veto cells." The precursor frequency of CTL specific for the injected class I antigen is found to be reduced greater than 200-fold at days 5 to 6 after the injection, whereas the frequencies of CTL specific for third party class I antigens are not significantly changed. These results indicate that there is a functional clonal deletion of the CTL recognizing class I incompatible veto cells in vivo. The role of such a veto phenomenon in the induction and maintenance of self tolerance and allograft tolerance is discussed.     

The use of a clonal assay for cytotoxic T lymphocytes to determine the Ly phenotypes of the cytotoxic precursor and effector cells to alloantigens and trinitrophenyl-modified self antigens

Considerable controversy has arisen regarding the Ly phenotypes of cytotoxic T lymphocytes (CTL) and their precursors (CLP) to alloantigens and modified-self antigens. Although there is general agreement that all CTL and their pregenitors express the Ly2 alloantigen, the presence of the LY1 alloantigen on either CTL or CLP is debated. Clonal assays for CLP, capable of detecting single CLP in the absence of accessory cells, have recently been developed. This assay system provides a sensitive means of determining the Ly phenotypes of CLP to alloantigens or trinitrophenyl- (TNP) modified self antigens. Lymph node cells from C57BL/6 (Ly-1.2, 2.2, 3.2) or CBA (Ly-1.1, 2.1, 3.2) mice were treated with anti-Ly serum and complement (C), and the frequencies of CLP of the treated populations to alloantigens or TNP-modified self antigens were determined. We found that the number of CLP reactive to alloantigens or TNP-modified self antigens were greatly reduced after treatment with either anti-Ly-1 or anti-Ly-2 serum and C in both C57BL/6 and CBA mice. In other words, the CLP to alloantigens or TNP-modified self antigens in these 2 strains of mice are Ly 1+2+. We also found that the CTL derived from the Ly1+2+ CLP were also Ly1+2+. The significance of this finding with respect to the cytotoxic repertoire for alloantigens and modified self antigens is discussed.     

Free fatty acids inhibit cytotoxic T lymphocyte-mediated lysis of allogeneic target cells

Short term exposure of murine CTL clones to long chain cis unsaturated free fatty acid (FFA) inhibits alloantigen specific lysis of cognate target cells, whereas long-chain saturated FFA have no effect. Inhibition of lysis occurs when cis FFA is added before or within 10 min after CTL-target cell conjugate formation and thus appears to interfere with lethal hit delivery. Our previous studies have shown that similar treatment with cis FFA inhibits, in CTL, the Ag stimulated increase in intracellular calcium and degranulation, suggesting that inhibition of lysis probably results from perturbation of the CTL signaling pathway. However, inhibition of lysis is probably not due to the inhibition of the rise in intracellular calcium or degranulation, because lysis can occur under conditions in which FFA inhibit degranulation and because cis FFA inhibit calcium-independent killing. Inhibition of lysis is detectable at unbound FFA concentrations less than 1 microM and is generally complete at concentrations less than 5 microM. Although these levels of FFA are somewhat higher than reported for normal physiologic conditions, plasma FFA levels can be elevated into this range in states of stress and disease, suggesting that FFA modulation of the immune response has important physiologic consequences.     

Subcellular localization of perforin and serine esterase in lymphokine-activated killer cells and cytotoxic T cells by immunogold labeling

CTL, NK cells, and lymphokine-activated killer (LAK) cells are cytolytic lymphocytes known to produce a pore-forming protein, named perforin or cytolysin, that lyses target cells by forming large pores on the plasma membrane of the target cell. Other proteins besides perforin are found in the cytoplasmic granules of effector lymphocytes, and these include a family of serine esterases. Ultrastructural immunogold labeling studies with antibodies against perforin and a serine esterase (MTSP-1, also known as granzyme A and SE-1) show that all the granules of LAK cells and a CTL cell line contain perforin and serine esterase. For both LAK cells and CTL, perforin has been located mostly in the fine granular matrix of the granules, whereas gold particles corresponding to serine esterase have been found in both the matrix and the cap regions of the granules. Results from double immunogold labeling indicate that perforin and serine esterase colocalize to the same granules.     

Correlation of Qa-1 determinants defined by antisera and by cytotoxic T lymphocytes

Cytotoxic T lymphocytes (CTL) activated in H-2 identical, Qa-1 disparate mixed leukocyte cultures recognize H-2-nonrestricted target antigens indistinguishable by strain or tissue distribution from serologically defined Qa-1 antigens. Cloned Qa-l-specific CTL define determinants encoded by four Qa-1 genotypes; we used anti-Qa-1 sera in antibody blocking experiments to determine if these determinants reside on molecules recognized by Qa-1-specific antibodies. Antisera containing Qa-1.1-specific and TL-specific antibodies blocked recognition of two CTL-defined determinants associated with Qa-1 ^a . Although both Qa-1 and TL molecules are expressed on activated T cells from appropriate strains, our studies indicated that the CTL recognized Qa-1, not TL. In addition, anti-Qa-1.2 serum inhibited CTL recognition of Qa-1^b- and Qa-1^c-encoded determinants. Qa-1 ^d target cells are unique in that they express determinants recognized by anti-Qa-1^a CTL and by anti-Qa-1^b CTL. Killing of Qa-1 ^d targets by anti-Qa-1^a CTL was not inhibited by anti-Qa-1.1 serum, but was partially inhibited by anti-Qa-1.2 serum. Cytotoxicity of Qa-1 ^d cells by one anti-Qa-1^b CTL clone was inhibited by both anti-Qa-1.2 and anti-Qa-1.1 sera, indicating close association of both serological determinants with the determinants recognized by the CTL. Thus, all of the CTL-defined Qa-1 determinants resided on molecules recognized by Qa-1-specific antibodies, but anti-Qa-1^a CTL and Qa-1.1-specific antibodies did not have identical specificities.Abbreviations used in this paper B6 C57BL/6J - CAB concanavalin A stimulated lymphoblasts - CML cell-mediated lympholysis - CTL cytotoxic T lymphocyte - NMS normal mouse serum - MHC major histocompatibility complex - MLC mixed leukocyte culture - MR maximum release - SMDM supplemented Mishell-Dutton medium - SR spontaneous release     

Mapping of epitopes recognized by alloreactive cytotoxic T lymphocytes using inhibition by MHC peptides

To identify epitopes recognized by alloreactive CTL we have examined H-2Kb-specific CTL for their recognition of synthetic peptides with sequences derived from the native Kb class I molecule. Consecutive nested peptides spanning the immunogenic alpha 1 and alpha 2 domains of Kb were tested for their capacity to inhibit CTL clones in their recognition of cells expressing the native Kb molecule. Inhibition by these peptides was found to be an extremely rare event. One peptide (Kb.111-122) did inhibit recognition by one particular CTL clone, clone 13. Upon further investigation it was observed that clone 13 also recognized peptide Kb.111-122 when presented in the context of the syngeneic MHC molecule, Kd. Considering that residues 111 to 122 are located at the base of the antigen groove, and clone 13 is able to recognize Kb.111-122 when presented by syngeneic target cells, we suggest that inhibition of this CTL clone may be due to MHC restricted, self-presentation of peptide rather than to direct binding of free peptide to the TCR. Taken together, these results suggest inhibition of allospecific CTL by MHC peptides is a rare event at least for Kb recognition. Furthermore, they demonstrate the need for caution when interpreting inhibition by peptide as evidence for recognition by the TCR of the corresponding region on the native molecule.     

MHC-restricted cytotoxic response of chicken T cells: expression, augmentation, and clonal characterization

Major histocompatibility complex (MHC)-restricted cytotoxicity of chicken lymphocytes was studied by using three reticuloendotheliosis virus (REV)-transformed cell lines as targets in 51Cr-release assays. The cell lines, designated RECC-UG5, RECC-UG6, and RECC-UG8, were developed from bone marrow cells of REV-infected line G-B1, line G-B2, and (G-B1 X G-B2)F1 chickens respectively. Effector cells were obtained from spleens of G-B1, G-B2, F1, and F2 chickens 7 days after inoculation of REV. The inbred G-B1 (MHC genotype B13/B13) and G-B2 (MHC genotype B6/B6) lines originate from a common partially inbred line. Initial studies with effector cells from G-B1 and G-B2 chickens showed that significant cytotoxicity occurred only with syngeneic target cells. The degree of cytotoxicity was markedly enhanced by neonatally treating effector cell donors with cyclophosphamide (CY) and delaying virus challenge until the birds were 4 wk old. Augmentation of cytotoxicity was presumed to be due to elimination of bursal-dependent suppressor T cells by CY. The results with spleen cells from REV-inoculated F2 birds clearly showed that cytotoxicity was MHC restricted; i.e., significant lysis only occurred if effector cells and target cells had a common B system antigen. Lysis of RECC-UG5 targets was three to four times higher than lysis of RECC-UG6 targets when effector cells were from heterozygous (B6/B13)F1 and F2 birds. Because these two target cell lines generally showed a similar degree of lysis by effector cells from syngeneic B homozygous birds, the differences obtained with effector cells from B heterozygous birds was most likely due to differences in the number of effector cells with specificity for each target line. Evidence for an additive cytotoxic effect, considered to be due to the lytic activity of two separate T cell clones, was obtained when F1 effector cells were tested with the F1-derived RECC-UG8 targets. The results of other experiments indicated that the effector cells were of T cell lineage and that their activity was probably directed against virus-induced antigens on the transformed target cells.     

Recognition of self and altered self by T cells in autoimmunity and allergy

T cell recognition of foreign peptide antigen and tolerance to self peptides is key to the proper function of the immune system. Usually, in the thymus T cells that recognize self MHC+ self peptides are deleted and those with the potential to recognize self MHC+ foreign peptides are selected to mature. However there are exceptions to these rules. Autoimmunity and allergy are two of the most common immune diseases that can be related to recognition of self. Many genes work together to lead to autoimmunity. Of those, particular MHC alleles are the most strongly associated, reflecting the key importance of MHC presentation of self peptides in autoimmunity. T cells specific for combinations of self MHC and self peptides may escape thymus deletion, and thus be able to drive autoimmunity, for several reasons: the relevant self peptide may be presented at low abundance in the thymus but at high level in particular peripheral tissues; the relevant self peptide may bind to MHC in an unusual register, not present in the thymus but apparent elsewhere; finally the relevant self peptide may be post translationally modified in a tissue specific fashion. In some types of allergy, the peptide+ MHC combination may also be fully derived from self. However the combination in question may be modified by the presence of other ligands, such as small drug molecules or metal ions. Thus these types of allergies may act like the post translationally modified peptides involved some types of autoimmunity.     

Early antigen elimination by cytotoxic T cells results in diminished cellular immune responses to allogeneic tumor

Previous reports have suggested that repeated alloantigenic challenge increases humoral responses to alloantigens, but may cause decreasing cellular responses. We stimulated BALB/c (H-2^d) mice with intraperitoneal EL-4 tumor (H-2^b) and serially assessed cytotoxic responses in spleen and peritoneal lymphocytes using ⁵¹Cr-labeled EL-4 target cells. We observed that cytotoxicity generated in the spleen of nonimmune BALB/c mice was much greater than that in immunized mice; similar peak responses were generated in peritoneal lymphocytes in normal and immunized hosts. Complement-mediated cytotoxicity was not necessary for diminished splenic responses in hyperimmune hosts, for the same phenomenon was seen in the Hzl anti-BL/6 system which is free of humoral responses. Irradiation (2000 rad) of the EL-4 tumor challenge prevented tumor cell proliferation and markedly reduced splenic responses in nonimmune mice. We suggest that cytotoxic cells suppressed further generation of cyto-toxicity; by effecting an early elimination of tumor inoculum, tumor proliferation was abrogated and dose of cellular antigen was, consequently, markedly reduced.     

Xenogeneic serum-induced murine cytotoxic cells. I. The generation of effector components specific for self and allogeneic target cells.

Spleen cells from unimmunized mice cultured in vitro without the intentional addition of exogenous antigen generated cytotoxic effector cells which lysed tumor and mitogen-stimulated blast target cells in a 5-hr chromium release assay. Effectors were generated in fetal bovine serum but not in adult horse serum, although both serum sources supported the generation of allogeneic cytotoxic cells. No correlation was observed between the ability of a serum source to support and generate serum-induced effectors and its ability to support an allogeneic cytotoxic response. The effectors lysed targets which were H-2 matched and those which were not H-2 matched with the cultured spleen cells, although the H-2 matched targets were consistently lysed more efficiently. “Cold” target cell inhibition studies indicated that multiple clones of cytotoxic cells were generated—including effector cells with specificity for self-structures, and particular alloantigens. Possible roles for the xenogeneic serum in the generation of this response are considered, including (a) the provision of essential stimulating and target antigens; (b) the induction of the expression of neoself-determinants; and (c) the possession of mitogenic properties.