EXPRESSION OF FOREIGN H-2-LIKE ANTIGENS BY A CHEMICALLY-INDUCED MURINE TUMOUR (MCG4)

Alien H-2^k-like antigens were found to be expressed by a methylcholanthrene induced tumour of BALB/c (H-2^d) origin. H-2 specificities of the k haplotype were detected on this tumour by a variety of serological techiques, including ⁵¹Cr-release cytotoxicity, microradioassay and absorption. The antisera employed were conventional polyspecific alloantisera, typing sera with restricted specificty and monoclonal hybridoma-derived anti-H-2^k antibodies. The tumour has a low expression of the private specificty 31, which characterizes K^d molecules, and does not seem to express the private specificities of D^d, K^k and D^k molecules. It appears to express predominantly alien H-2-like antigens which are very similar to but not identical with normal H-2^k molecules.     

H-2d-LIKE SPECIFICITIES ON GARDNER (H-2k) TUMOUR DETECTED IN A RESTRICTED ANTI-TUMOUR REACTION

Cytostatis of the H-2^d tumour LSTRA by H-2-restricted effector lymphocytes was inhibited by antisera against H-2.4 and H-2.31 but not by antisera against public specificities or non-H-2 antigens. The unexpected reaction of the same effector cells against Gardner tumour (H-2^k) was also shown to be inhibited by a combination of antisera against H-2.4 and H-2.31 but not by each antiserum used separately. The inhibitory capacity of these antisera was removed by absorption with B10.D2 but not with B10 lymphocytes. This indicated the presence of H-2^d-like specificities on gardner tumour which could function as self-recognition structures in an H-2K^d and H-2D^d restricted system.     

EQUIVALENT DECREASE OF H-2Kk AND H-2Dk EXPRESSION AFTER VACCINIA VIRUS INFECTION

Expression of H-2K^k and H-2D^k molecules was studied by indirect immunofluorescence on vaccinia virus-infected L_s cells (derived from the L929 cell line), using D-23^b and ASA-21 monospecific alloantisera, respectively directed against the H-2.23 and H-2.32 private specificities of the H-2^k haplotype. Our study demonstrates that the increase in vaccinia virus-induced antigens on the membrane of infected cells (as a function of the multiplicity of infection) is associated with a concomitant decrease in the expression of both H-2K^k and H-2D^k serologically defined private specificities. Absorption experiments of D-23^b and ASA-21 sera, using infected or uninfected L_s cells, also indicate that after virus infection, H-2K^k and H-2D^k private specificities are equally altered in their serological expression. We finally show that this alteration consists in a mere quantitative decrease of H-2 molecules, since the cytolytic capacity of a rabbit anti-H-2 serum, specifically reacting against the constant part of the heavy chains of H-2 molecules, was significantly more absorbed by uninfected cells than by infected cells. We conclude that no preferential decrease in the expression of H-2 K^k and H-2 D^k molecules is induced during cell infection by vaccinia virus.     

Original H-2d and alien H-2k-like antigens of a BALB/c fibrosarcoma as defined by in vitro and in vivo studies of cell-mediated immunity

The present study examines the immunosensitivity and the immunogenicity of both original H-2^d and alien H-2^k-like antigens of the BALB/c (H-2^d) fibrosarcoma C-1 as detected by in vitro and in vivo cell-mediated cytotoxicity (CMC) assays. It was found that ⁵¹Cr-labeled C-1 cells were lysed in vitro by C 57 BL/6 anti-H-2^d lymphocytes. The specificity of this reaction was shown by cold inhibition experiments in which the anti-H-2^d cytotoxic activity on YC8 (H-2^d) targets was inhibited by unlabeled YC8 or C-1 but not by C3UR11 (H-2^k) tumor cells. Both D^d- and K^d-encoded antigens were recognized by appropriate cytotoxic effectors. The immunogenicity of H-2^d antigens of C-1 was revealed by the ability of C 57 BL/6 anti-C-1 lymphocytes to lyse YC8 targets. The expression of H-2^k-like alien alloantigens on C-1 was indicated by the finding that anti-H-2^k cytotoxic T lymphocytes (CTL), generated by culturing BALB/c spleen cells immune to BALB.K (H-2^k), C3Hf (H-2^k) or A (H-2^a = H-2^k/d) tissues with the cells of the same strain used for immunization, lysed C-1 targets. The cytotoxicity of these anti-H-2^k CTL against C 3 UR11 (H-2^k) targets could be specifically inhibited by cold C 3 UR 11 or C-1 cells but not by two other BALB/c tumors. Using recombinant H-2-congenic mice, it was shown that both D^k and K^k antigens were recognized by CTL on C-1 cells. The immunogenicity of the H-2^k-like antigens, however, could not be detected in vitro. In fact, effector spleen cells from BALB/c mice immune to C-1 did not develop any detectable cytotoxicity against C 3 UR 11 targets as assayed either by a direct in vitro test or after in vitro restimulation with C-1 sarcoma cells. A similar experimental design was adopted in Winn assays carried out by mixing spleen cells of BALB/c immune mice with either C-1 or C3 UR 11 targets and injecting the mixtures in BALB/c or hybrid recipients. These in vivo tests revealed the presence of both H-2^d (K^d and D^d) and H-2^k-like (K^k and D^k) antigens on C-1. At variance with the in vitro CMC assays, however, the Winn assay also detected the immunogenicity of the H-2^k alien antigens, since BALB/c anti-C-1 spleen cells were able to significantly reduce the growth of C 3 UR 11 lymphoma cells in (BALB/c × C 3 Hf)F₁ hosts.     

Cell-mediated cytotoxicity against ectromelia virus-infected target cells. III. Role of the H-2 gene complex

The role of the H-2 gene complex in expression of cytotoxicity exerted by specific ectromelia-immune thymus-derived (T) cells against ectromelia-infected target cells was examined. A repertoire of inbred mouse strains (some congenic) including the H-2 haplotypes k, d, b, s, q, the recombinant H-2^a (k|d) and F₁ hybrids (k/b and d/b) were immunized with virus and their spleen cells tested 6 days later, at the peak of the primary response, against H-2^k, H-2^d and H-2^b target cells. Significant specific cytotoxicity occurred only when the immune cell donors and the target cells shared all or part of the same H-2 gene complex. For example, H-2^a (k|d) immune cells killed both H-2^k and H-2^d target cells. There wasno detectable effect of the non-H-2 genetic background, H-2 public specificities, or the M-locus Target cells infected with ectromelia virus exhibited quantitative or qualitative changes (or both) in expression of normal H-2 antigens as indicated by reduced susceptibility to killing by T cells activated against H-2 antigens in mixed lymphocyte culture. These data are consistent with the hypothesis that T cells in this system are responding to virus-induced, specific changes in antigens on infected cells which are controlled by genes in the H-2 complex; these genes seem likely to be those coding for H-2 private specificities, or genes closely linked to them.     

H-2K RESTRICTION OF THE T CELL-MEDIATED LYSIS OF A CHEMICALLY-INDUCED BALB/c FIBROSARCOMA

Previous work has shown that a cytotoxic T lymphocyte (CTL) immune response of syngeneic mice immunized with a chemically-induced BALB/c (H-2^d) fibrosarcoma was directed against an individual tumour-associated antigen. To see whether this reaction was restricted by products of the major histocompatibility complex (MHC), anti-H-2 alloantisera to K or D antigens were used to interfere with the CTL-mediated immune response. Antisera to K^d but not to D^d antigens inhibited the lytic activity of CTL against fibrosarcoma cells. In addition, the study of the CTL response in F₁→ P antitumour immunized chimeric mice showed that antitumour cytotoxicity developed only when F₁ and parental host shared the K^d region. Both experiments strongly indicate that recognition of the individual tumour-associated antigen of the BALB/c fibrosarcoma is restricted by the products of H-2K^d genes.     

THE COMPONENT FRAGMENTS OBTAINED BY ACID DISSOCIATION OF PAPAIN-SOLUBILIZED H-2 MOLECULES*

H-2K^k and H-2D^d molecules were specifically purified from a radioiodinated H-2^a preparation obtained by papain digestion of spleen cell membranes of A/J strain mice. The molecules were isolated by binding to H-2 alloantisera of the corresponding private specificity followed by precipitation with rabbit anti-mouse IgG antiserum. The specifically precipitated radioiodinated H-2K^k and H-2D^d molecules were dissociated by acid treatment into large and small components of about 37,000 and 11,000 daltons respectively. These were separated by gel filtration at acid pH or by gel isoelectric focusing in the presence of 6 M urea. Each component separated by gel filtration of the acid-dissociated H-2 molecules showed a high degree of size homogeneity as determined by sodium dodecyl sulphate-acrylamide gel electrophoresis. Upon gel isoelectric focusing, however, the small components showed two peaks of radioactivity closely located together at pH 7–8, both of which had a restricted pH range, while the large components gave one peak of a relatively wide pH range of pH 5–6. The H-2K^k and H-2D^d molecules gave essentially the same pattern in terms of the numbers and the positions of the radioactivity bands. Under the iodination conditions used the large components of H-2K^k molecules contained more radioactivity than the small components, while the reverse was true in case of H-2D^d molecules. Such a difference was also found with H-2K^k and H-2D^d molecules isolated by use of alloantisera of the respective public specificity. The assay of binding of the isolated components with H-2 alloantisera of defined specificity revealed that the large components retain most of the allospecificities of the parental H-2 molecules. No H-2 allospecificities were found on the small components. The small components showed extensive binding with rabbit antiserum against mouse β₂-microglobin. The same antiserum did not show any binding with the large components. On the other hand, both of the components did bind with rabbit antiserum against papain-solubilized H-2 molecules.     

Tumorigenicity of mouse T lymphoma cells is controlled by the level of major histocompatibility complex class I H-2Kk antigens

We have previously found that an increased tumorigenicity and spontaneous metastatic potential of BW5147-derived T lymphoma cells was associated with a decrease in major histocompatibility complex (MHC) class I H-2K^k antigen expression. This suggested that H-2K^k antigens may control the tumorigenic potential of BW T lymphoma cells. Our current experiments aimed to prove this association by specifically altering H-2K^k expression by gene transfection. Transfected cells expressing a high level of H-2K^k antigens were significantly less tumorigenic and metastatic after subcutaneous inoculation. However, there was selectionin vivo for cells expressing a reduced level of H-2K^k antigens, which concomitantly led to an increased tumorigenicity. These data further confirmed the strong association between H-2K^k expression and tumorigenicity. We subsequently tested whether the immune system is implicated in this phenomenon by inoculating the H-2K^k transfectants into irradiated, immunocompromised recipients. Our results indicate that the reduced tumorigenicity of the BW H-2K^k transfectants is due to an immune rejection mechanism, mediated by CD8⁺ immune effector cells, as revealed byin vivo depletion experiments with anti-CD8 antibodies. Hence, we hereby demonstrated that H-2K^k antigens increased the immunogenicity of BW cells, via a CD8-dependent mechanism, which consequently reduced their tumorigenicity.     

ANALYSIS OF THE B10.D2-H-2d m l MUTANT WITH MONOCLONAL ANTIBODIES

We compared H-2D products of B10.D2 and the mutant B10.D2-H-2^dml spleen cells using two monoclonal antibodies designated D3.179 and S11.7. D3.179 reacts with public specificities and appears to define more than one product as shown by cytotoxicity and binding studies. Strong reactivity was obtained in both binding and cytotoxicity and binding studies. Strong reactivity was obtained in both binding and cytotoxicity to B10.D2 and marginal cytotoxicity and low binding were observed with B10.D2-H-2^dml cells. The second monoclonal antibody S11.7, in contrast, reacts only with H-2D^d bearing strains and gave strong binding to B10.D2-H-2^dml and weaker binding to B10.D2. Taken together, the observations indicate that the mutant not only expresses H-2D, encoded antigens but bear quantitatively greater amounts of some specificities than the parent. A Product of approximately 46,000 daltons was immunoprecipitated from B10.D2 cells by D3.179. Very small amounts of this product were seen in B10.D2-H-2^dml immunoprecipitates.     

SEROLOGICAL AND IMMUNOCHEMICAL STUDIES OF H-2 ALLOSPECIFICITIES ON k36, A SYNGENEIC TUMOUR OF AKR

Expression of H-2 antigenic specificities on K36, a spontaneous leukaemia originating from AKR (H-2^k) mice, was studied by serology and immunochemistry. Two ascites lines of the tumour, as well as a tissue culture adjusted and cloned tumour line, were used in these studies with similar results being obtained. K36 expresses on its cell surface D-region encoded H-2K antigens but does not express K-region encoded H-2K alloantigens. It also expresses on its cell membrane, H-2 specificities of foreign haplotypes not present on normal AKR lymphoid cells. The molecular basis of the H-2D^d specificity on K36 (H-2^k was analysed by immunoprecipitation and polyacrylamide gel electrophoresis. The specificity was shown to be present on a glycoprotein of apparent molecular weight 45,000. However, antisera against the H-2D^d private specificity (H-2.4) precipitate additional glycoprotein of 45,000D and also 70,000D. In tryptic peptide maps of the isolated 45,000D fraction precipitated by anti-H-2.4 serum from radiolabelled K36 glycoprotein, all H-2D^d specific peptides were present in the same quantitative ratio. This is consistent with the structural identity of the foreign H-2D^d from the K36 tumour with normal H-2D^d and supports the hypothesis of a regulator system controlling the H-2 allelism. Under certain circumstances such a system could cause suppression of one and derepression of the other H-2 gene products.     

Cell-mediated anti-tumor response in the RLmale 1 system. I. T nature, H-2Dd involvement and antigenic specificity of the effector cells.

After in vivo priming and in vitro secondary stimulation by the radio-induced RL♂ 1 lymphoma cells, syngeneic BALB/c splenic lymphocytes evidenced a specific but very weak cytolytic activity against RL♂ 1 target cells. Under the same experimental conditions, spleen cells from (B6 x BALB/c)F₁ were at least 100 times more efficient. In both cases, the effector cells were cytolytic T lymphocytes (CTL). Normal H-2D^d antigens of the tumor cell surface were involved in the effector-to-target cell interaction since anti-H-2 or anti-H-2D^d antisera abolished the RL♂ 1 cytolysis by immune syngeneic CTL, whereas anti-H-2K^d were devoid of blocking activity. The testing of a series of lymphoma cells, either virus-induced or not, belonging to different H-2 haplotypes, show that the immune CTL were highly specific for RL♂ 1 target cells. Only with the P815 mastocytoma cells was a weak cross-reactivity detected by both direct tests and competition experiments. Allogeneic (H-2^k) AKR lymphoma cells which share the X1 antigen with RL♂ 1 do not react, possibly due to an H-2 restriction of the CTL activity. The CTL-reacting antigen cannot be definitely identified, since several specificities, including the serologically defined X1 antigen, are possibly involved simultaneously. The Hh antigens which could have accounted for the F₁ anti-parent reaction, appear irrelevant.     

Biochemical purification of detergent-solubilized H-2 alloantigens

A multistep Chromatographic fractionation scheme is described for purifying the H-2K^b and H-2D^b major histocompatibility antigens as isolated from the murine lymphoblastoid cell line EL4 (H-2^b haplotype). The membrane-integrated antigen molecules were solubilized with the non-ionic detergent NP-40 and were purified by gel filtration chromatography, ion exchange chromatography and affinity chromatography with lentil lectin conjugated to Sepharose. During the latter procedure, use of a linear gradient monosaccharide elution effected partial separation of the H-2K^b and H-2D^b antigens. At this stage the H-2 glycoproteins were highly purified based on several criteria. Upon polyacrylamide gel electrophoresis in SDS the major band migrates with a mol. wt of approximately 45,000 daltons corresponding to the mol. wt of antigens obtained by immunoprecipitation. Moreover, near identity of the profiles of the arginine-containing tryptic peptides from chromatographically-purified and immunoprecipitated H-2K^b preparations suggests that a high degree of homogeneity has been achieved in the Chromatographic purification. As is demonstrated, mg quantities of the H-2K^b and the H-2D^b antigens can be isolated and partially separated from each other by this purification scheme thereby opening the way for structural studies of the H-2K and H-2D molecules by a variety of biochemical methods.     

Antigen processing mutant T2 cells present viral antigen restricted through H-2K

Cytotoxic T lymphocytes (CTL) recognize foreign antigens as short peptides presented by class I molecules of the major histocompatibility complex (MHC). T2 cells are profoundly defective in the presentation of endogenously synthesized antigens to CTL due to a deletion of MHC class II-encoded genes for transporters associated with antigen presentation (TAP1/TAP2). Surprisingly, we here demonstrate that T2 cells, after infection with Sendai virus, are readily killed by H-2K^b restricted CD8⁺ T cells. In contrast to classical class I-mediated antigen presentation, the presentation of Sendai virus antigen inT2K^b cells is brefeldin A (BFA) insensitive. The present findings may suggest the presence of an alternative pathway for MHC class I-mediated antigen presentation in T2 cells.     

Monoclonal antibody detection of two classes of H-2Kk molecules

Studies described in this paper indicate that two anti-H-2K^k monoclonal antibodies, namely 27R9 (H-2.25) and 30R3 (H-2.5) recognize different H-2K^k molecules on the surface of lymphocytes. Initial experiments in support of this conclusion were cocapping experiments which showed mutual exclusiveness between H-2K^k molecules which bind either of the two monoclonal antibodies 27R9 (H-2.25) or 30R3 (H-2.5) whereas conventional anti-H-2K^k (H-2.23) alloantiserum binds to both types of H-2 molecules. This result was confirmed by experiments using solubilized H-2 antigen preparations to inhibit antibody binding to spleen cells. Preabsorption of the preparation with one monoclonal antibody did not remove its inhibitory activity for the other monoclonal antibody, and only partially removed its inhibitory activity for the conventional anti-H-2K^k serum. These results suggest that at least two antigenically distinct H-2K^k molecules arc controlled by the H-2K region. Subsequent blocking studies have indicated that the two different molecules are associated, to some extent, in the cell membrane. Furthermore, in ¹²⁵I-protein A radioimmunoassay.each monoclonal antibody was found to bind to only half of the estimated total number of H-2K^k molecules recognized by conventional anti-H-2K^k sera. Several interpretations for the existence of the two classes of H-2K^k molecules are discussed.     

Cytotoxic T cell responses to haptenated cells II. On the role of H-2 genes

Cytotoxic T cell (CTL) responses to cells coupled with high or low doses of the hapten 3-(p-sulfophenyldiazo)-4-hydroxyphenyl acetic acid (SP) were studied. CTL responses to densely coupled cells were obtained with all mouse strains tested. Only mice carrying the K^k allele responded to sparsely coupled cells. Mice expressing I-A^kbut not K^k were nonresponders. (Responder × nonresponder) F₁ hybrids responded in vitro to sparsely coupled cells from responder but not from nonresponder mice. However, responder mice could be cross-primed in vivo with SP-coupled cells from nonresponder mice. CTL responding to sparsely coupled cells were restricted to K^k. Cross-reactive lysis of target cells not carrying the K^k allele was dependent on the hapten density and the H-2 haplotype carried by the target cells.     

Dominant allele-specific regulation of expression of H-2Kk gene products revealed by somatic cell hybridization

A study has been made of the H-2 profiles of the AKR thymoma K36 and two series of somatic cell hybrids derived from it. Despite expressing good amounts of the H-2D ^k product, the H-2K^k antigen was barely detectable in this tumor, approximately only 1% of that expressed by a comparable AKR lymphoma, 339. The isoelectric focusing profile of the H-2K^k product immunoprecipitated from K36 was found to be identical to that from normal AKR lymphocytes. The phenotype of K36 therefore results from regulatory constraints. Fusion ofK36 with normal K-2^k lymphocytes resulted in hybrids expressing the H-2D^k product well but having a drastically reduced expression of the H-2K^k product. Fusion of K36 with normal H-2^b lymphocytes resulted in good expression of the H-2D^k and H-2D^b alleles in addition to good expression of the H-2K^b products. The expression of the H-2K^k antigen remained marginal. The data suggest a suppressive mechanism which is trans-acting and dominant and specific for the H-2K^k allele.     

Cytotoxic T cell responses to haptenated cells III. Isolation and specificity analysis of continuously growing clones

Various procedures were used to derive continuously growing cytotoxic T lymphocyte (CTL) clones from a primary culture containing responder cells from immunized mice and 3-(p-sulfophenyldiazo)-4-hydroxylphenyl acetic acid (SP)-or fluorescein isothiocyanate (FL)-coupled stimulator cells. It seems likely that CTL have to undergo some change, possibly genetic, to be able to grow continuously in T cell growth factor conditioned medium in the absence of any stimulator or filler cells. The most convenient and reliable procedure to generate CTL clones with different specificities was to establish from several aliquots of a primary culture cell populations continuously growing in medium conditioned with T cell growth factor(s). Clones with different specificities segregated in the different populations. SP-and FL-specific CTL clones restricted to H-2K^k, and H-2D^d and two FL-specific CTL clones with no apparent H-2 restriction are described.     

Monoclonal antibody detection of carbohydrate-defined and protein-defined H-2Kk antigens

Three different monoclonal anti-H-2K^k antibodies, 27R9, 30R3 and 11-4 were examined for the biochemical nature of the antigenic determinants they recognize. When these were compared on the basis of their sensitivity to pronase and various glycosidases, 27R9 was shown to bind to protein-defined H-2K^k antigens, while 30R3 and 11-4 bound to H-2 antigens defined by carbohydrate. From sugar inhibition studies, and treatments with specific glycosidases, d-mannose appears to be the immunodominant sugar involved in the antigenic site recognized by 30R3, while several sugars, namely sialic acid, d-mannose and α-and β-linked d-galactose would appear to be components of the antigenic site bound by 11-4. The carbohydrate determinants appear to be present on glycolipid molecules, since both the 30R3 and 11-4 antibodies could be inhibited by glycolipid extracts from spleen cells of the appropriate H-2 haplotype, as well as from several other strains of mice previously shown to be cross-reactive targets for these antibodies. This finding is supported by evidence that the molecule carrying the protein-defined antigen is distinct from that carrying the carbohydrate-defined antigens. The results are discussed in the light of current information on the nature of glycolipid Ia antigens, as well as the role of H-2 antigens in T-cell interactions.     

Regulation of delayed-type hypersensitivity. VII. The role of I-J subregion gene products in the inhibition of delayed-type hypersensitivity to major histocompatibility antigens by specific suppressor T cells

Suppressor T cells for delayed-type hypersensitivity (DTH) to alloantigens are induced by injecting mice intravenously with a high dose of X-irradiated allogeneic cells. Using the same protocol, suppression of DTH directed against H-2 subregion products could also be induced, provided that the H-2 incompatibility between the cells used for the induction of suppression and the recipients includes the I-J subregion genes. Thus, the I-J subregion difference is both necessary and sufficient for the induction of suppression of DTH to the whole or part of the H-2 gene products. The suppression appears to be mediated by antigen-specific suppressor T cells which recognize the allo-I-J molecules and are able to suppress the DTH response to other H-2 subregion gene products in an associative recognition manner. T cells from (B10 × BALB/c)F₁ mice suppressively primed against B10.A(5R) cells (directed against J^kE^k antigens), when adoptively transferred to normal syngeneic recipients, were capable of suppressing the hosts' DTH response to B10.A(4R) cells (K^kA^k antigens) when the recipients were challenged with [B10.A(4R) × B10.A(5R)]F₁ cells (K^kA^k × J^kE^k). The recipients express normal DTH reactivity to B10.A(4R) cells when challenged with a mixture of B10.A(4R) and B10.A(5R) cells (K^kA^k + J^kE^k). These results provide direct evidence that when functioning as alloantigens, the I-J determinants preferentially induce suppressor T cells which specifically impair the immune response to the I-J molecules as well as other H-2 gene products if they are physically associated with the I-J determinants. The role of I-J subregions as the suppressor genes is discussed in terms of the possible application in transplantation immunity and the host's defence against infectious diseases.     

THE H-2 COMPLEX: IMMUNOGENICITY AND ENHANCEMENT STUDIES OF H-2K REGION ALLOANTIGENS

Antigenic differences arising from incompatibilities at the K region of the H-2 complex of the mouse (including the H-2K locus) were studied in different combinations to determine the immunogenicity of K^k, K^q and K^s region products. Only congenic strains and selected recombinant strains were used, so that the only known differences arose from the K region of the H-2 complex. In males, first grafts were rejected in 12–13 days so that, as immunogens, K region antigens were equivalent to those of H-2I. If the H-2 histocompatibility loci are ranked in decreasing order of the immunogenic strength of their antigenic specificities: H-2K = H-2I > H-2D > H-2IC. Second grafts were rejected more rapidly than first skin grafts. After immunization by either lymphocytes or a single skin graft, high cytotoxic antibody titres were achieved, e.g. 1/512, after lymphocyte immunization. However, it was not possible by the passive administration of antiserum to delay graft rejection and produce enhancement by any antisera directed only against K region specificities. In several different combinations studied, it was proved that antigenic differences arising from the I region but not the K region are essential to produce an enhancing antiserum for K + I region differences. In particular, differences incorporating IA and IB are more important than IC differences. However, when differences arise only from the K region, enhancement cannot be produced by any antiserum, whether it contains anti-Ia antibodies or not, presumably as there is no I region incompatibility.