Reduced humoral and cellular cytotoxic sensitivity in major histocompatibility variants of the YAC (Moloney) lymphoma

Previously we have reported that two sublines of the YAC lymphoma selected for reduced expression of H-2^a and Moloney-virus determined cell-surface (MCSA) antigens are, in contrast to YAC, allotransplantable in H-2-incompatible recipients, and resistant to rejection by preimmunized semisyngeneic hosts. A third YAC variant with reduced MCSA but unchanged H-2-antigen expression, was not allotransplantable and showed only a slight decrease in its immunosensitivity in preimmunized semisyngeneic hosts in vivo. This suggested that H-2-antigen expression may be more important than MCSA expression for recognition and rejection by semisyngeneic mice. We have now tested the sublines expressing low H-2^a for their in vitro sensitivity to humoral and cell-mediated lysis. — The variants were more resistant than YAC to complement lysis by anti-H-2^a, anti-MCSA, anti-Thy 1.2 and antispecies sera. Absorption tests with antispecies serum indicated that the decreased cytolytic sensitivity of the variants was not related to the concentration of the relevant antigens, which was similar to that of the original YAC tumor. As expected from the low amount of H-2^a the variants showed a decreased sensitivity to the killing effect of allogeneic cytotoxic T lymphocytes (CTL). They were also lysed to a lesser extent than YAC by semisyngeneic CTL, probably directed against virally determined antigens. However, they were also less sensitive to lysis by natural killer (NK) cells, although NK lysis is probably unrelated to MHC expression. In conclusion, our selection for reduced H-2-expression appears to have resulted in the isolation of variants with a generally increased resistance to various humoral and cell-mediated lytic functions.On leave from Instituto Venezolano de Investigaciones Cientificas (IVIC), Caracas, Venezuela.     

Antibody-induced redistribution of normal and tumor associated surface antigens

Redistribution (capping) of normal and tumor-associated surface antigens was studied on murine and human cells by the indirect membrane immunofluorescence (MIF) technique. The capping of H-2 isoantigens was compared on normal mouse T-lymphocytes and on YAC cells, a Moloney leukemia virus (MLV) induced lymphoma. H-2 and Moloney virus induced cell surface antigen (MCSA) capping was compared on three YAC lines with different MCSA concentrations. H-2 and tumor-associated surface antigen capping was compared on two polyoma induced sarcoma lines and five methylcholanthrene induced sarcoma lines. In the human system, IgM-capping was compared on normal lymphocytes and on the Burkitt lymphoma derived Daudi line. Capping of HL-A and the Epstein-Barr virus (EBV) determined membrane antigen (MA) was compared on the Burkitt lymphoma derived line Maku and on EBV-superinfected Daudi cells. H-2 antigens on normal murine cells capped more promptly and on a larger fraction of the cell population on the various tumor cells. Surface associated IgM showed a better capping on normal lymphocytes than on Daudi cells. All tumor associated antigens except MCSA, showed good capping. MCSA was almost completely refractory to capping. Increasing concentrations of MCSA appeared to inhibit the capping of H-2 on the YAC sublines with different concentrations of MCSA. The polyoma induced ascites sarcoma (SEWA) did not cap either with regard to H-2 or the polyoma determined surface antigen.     

Spontaneous loss and subsequent stimulation ofH-2 expression in clones of a heterozygous lymphoma cell line

The expression of H-2K^k antigens in a (C3H × DBA/2)F₁ lymphoma cell line growing in vitro was investigated with monoclonal antibodies specific for a public antigen of theH-2K ^k region (H-2.m3) in fluorescence analysis and microcytotoxicity assays and in cell-mediated cytotoxicity with allogeneically stimulated effector cells. Estimates of relative levels of H-2K^k-antigen expression obtained by the different methods were highly correlated. The uncloned, unselected population gradually lost H-2K^k surface antigen expression under culture conditions. This was due to the appearance of H-2K^k negative variants. Fifteen cloned sublines of a population enriched for cells expressing antigen H-2.m3 in the fluorescence activated cell sorter contained either two distinct populations, one consisting of H-2.m3 negative and one of H-2.m3 positive cells, or consisted of H-2.m3 negative cells only. The expression of the H-2.m3 determinant of H-2K^k paralleled that of other serological H-2K^k determinants and of H-2K^k target determinants for cell-mediated cytotoxicity. In nearly all clones where two populations could be detected, the proportion of H-2.m3 negative cells increased with time in culture. The amounts of H-2K^k antigen expressed by the clones appeared not to be correlated to the amounts of H-2D^k antigens on the cell surface as judged by cell-mediated cytotoxicity.In at least one clone and in the uncloned population, H-2K^k-antigen expression detectable by fluorescence analysis could be stimulated by growing the cells in the peritoneal cavities of (C3H × DBA/2)F₁ mice or by adding mouse interferon preparations to the cell cultures. The increase in susceptibility to cell-mediated lympholysis of cells grown in vivo paralleled the increase inH-2 expression detected by fluorescence. In contrast, cells growing in the presence of interferon in vitro showed reduced sensitivity to lysis by alloreactive lymphocytes, although H-2 antigens were strongly expressed as measured by fluorescence.     

Role of H-2 antigens in the host response to methylcholanthrene-induced tumors

H-2 loss variant sublines of a sarcoma (M-AS), induced by methylcholanthrene in an (A × A.SW)F₁ mouse, were used to study the role of the MHC products in the recognition of MC-TSTA. The two reciprocal variant sublines (M-A and M-S) were found to express the TSTA of the original tumor as shown by cross-reactions in graft rejection experiments performed in (A × A.SW)F₁ mice. In the A/Sn and A.SW mice the presence of the reciprocal parental H-2 antigens on the immunizing cells decreased the response against the tumor antigens. An admixture of lymphocytes derived from hyperimmune mice inhibited the outgrowth of the tumor cells. The growth inhibition was mediated by T cells and was H-2 restricted. Cells derived from hyperimmune syngeneic mice inhibited the outgrowth of the variant subline used for immunization but had no effect on the reciprocal variant subline.     

The genetic control of natural killer cell activity and its association with in vivo resistance against a moloney lymphoma isograft

Spleens of normal young mice of certain strains contain lymphocytes that can kill strain A-derived YAC-1 lymphoma cells in a⁵¹Cr release cytotoxic assay in vitro. We have previously classified mouse genotypes as high or low reactors, according to their responses in this test. In vivo resistance to small numbers of YAC ascites lymphoma cells is correlated with in vitro cytolytic activity. In vitro and in vivo tests were carried out on the same individual (A x C57BL)F₁ x A backcross mice. Natural in vitro killer cell activity appeared to be under polygenic control, including a strong H-2-linked factor. No linkage was found with five different isozyme loci, with theIg-l locus or with C5 serum activity. Also in vivo resistance showed strong linkage with theH-2 complex. In (A x CBA)F₁ x A backcross mice, a weak linkage was found with the coat color locusC. There was a correlation between in vitro killer activity and in vivo resistance in the same backcross mice. In vivo resistance was particularly strong in mice that combined theH-2 ^b -linked resistance factor with a high cytolytic activity in vitro.     

Induction of high-grade anti-tumor immunity by use of a recombinant H-2Kb/avian erythroblastosis virus erbB gene transfectant

Summary The recombinant H-2K^b-erbB gene, encoding for a part of the H-2 class I antigen and the kinase domain of the V-erbB peptide, was successfully introduced into murine mastocytoma P815 variant P1.HTR cells, which resulted in low but significant cell-surface expression of the hybrid gene product. When the chimeric gene transfectant was inoculated into the CDF₁ mice, it soon grew but regressed thereafter. The tumorigenicity of this transfectant was lower than the H-2K^b gene transfectant that expressed the H-2K^b antigen at a comparable level. These CDF₁ mice that had received the chimeric gene transfectant obtained a high-grade anti-tumor immunity against the challenge of a high dose of parental tumor. Corresponding to these observations, anti-tumor cytotoxic T lymphocytes, which lyse parental P1.HTR cells but not syngeneic L1210 or NS-1 tumor cells, were developed in the peritoneal cavity of mice that had been inoculated with the transfectant and parental tumor. Definite antibody activity binding to parental P1.HTR tumor cells was also demonstrated in the sera of these mice, precipitating 40-kDa, 74-kDa and 98-kDa molecules from the surface of the radiolabeled P1-HTR tumor cells. The results suggested that the chimeric H-2-erbB gene transfectant efficiently triggers both cellular and humoral anti-tumor immune responses.     

H-2 antigen variants in a cultured heterozygous mouse leukemia cell line

An H-2 antigen variant, referred to as ?4 + 31 clone 1, was selected by its resistance to an anti-H-2D^d antiserum (BALB.G anti-BALB/c.H-2 ^g antiH-2 ^d ). When tested by cell-mediated cytolysis, this variant was found to be sensitive to cytolytic T lymphocytes raised in the same donor-host combination as that used in raising the antiserum. Further CML characterization of this variant, reported here, indicates that the cell line is in fact resistant to anti-H-2D^d killer cells raised in a more restricted immunization, viz. BALB.G anti-BALB/cH-2 ^db ,H-2 ^g anti-H-2 ^db . It is, however, sensitive to cytolytic cells raised in (BALB.B xBALB/c-H-2db) F¹ H-2 ^b /H-2 ^db ) against the BALB/c strain. These results suggest that the variant does not express H-2D^d itself, but probably expresses CML target antigens that are missing in theH-2 ^db mutation. This in vitro-isolated variant might thus be the complementary mutation to the in vivoobtainedH-2 ^db mutation.     

An H-11-linked gene has a parallel effect on Leishmania major and L. donovani infections in mice

The courses of visceral infection following intravenous injection of Leishmania donovani amastigotes, or lesion growth following subcutaneous injection of L. major promastigotes, were examined in B10.129(IOM) (H-2 ^b, H-11 ^b) mice and compared with disease profiles observed in congenic C57BL/10ScSn(=B10) (H-2 ^b, H-11 ^a) and B10.D2/n (H-2 ^d, H-11 ^a) mice, and in BALB/mice. Possession of alternative alleles at H-11 and closely linked loci transformed the normal curing/healing phenotype of B 10 mice into a characteristically different noncuring/nonhealing phenotype affecting both visceral and subcutaneous infections in B10.129(10M) mice. In reciprocal radiation bone marrow chimeras made between the congenic B10 and B10.129(10M) strains, both cure and noncure phenotypes were transferable with the donor hematopoietic system. Although it was possible to demonstrate transfer of suppression with T-enriched spleen cells from day 61 L. donovani-infected B10.129(10M) donor mice into 550 rad syngeneic recipients, the pretreatment of mice with sublethal irradiation did not, as in the earlier studies of Scl-controlled L. major nonhealing or H-2-controlled L. donovani noncure phenotypes, have a clear or consistent prophylactic effect. Together with the progressive disease profile observed even for L. donovani at low parasite doses this suggests that, despite their ability to develop initial delayed-type hypersensitivity reactions to parasite antigen early in L. major infection, B10.129(10M) mice possess some inherent defect in ability to mount a cell-mediated response effective at the level of macrophage antileishmanial activity in vivo even when suppressor T cells are not generated. Further elucidation of this characteristically different noncuring/nonhealing phenotype may provide important insight into common events involved in the development of the cell-mediated immune response to both visceral and subcutaneous forms of leishmaniasis.     

H-2 control of expression of an idiotype shared by normal B cells and a B-cell lymphoma

The spleens of normal B10,H-2 ^a H-44^b p/Wts (2 ^a 4 ^b ) mice; contain cells which, in response to mitogen stimulation, secrete hemolytic antibody specific for a determinant present on both sheep and bromelain-treated mouse erythrocytes. These cells were found to be Ly-1 positive. Approximately 50% of these cells bear surface immunoglobulin (sIg) with the same idiotype as the sIg of a 2^a4^b-derived B-cell lymphoma, CH12. Backcross analysis revealed H-2 control of the frequency of the idiotype-positive B cell. The regulatory gene did not correlate with the Igh-1 allotype, and analysis of 22 inbred mouse strains mapped the gene to the I-E subregion. Surprisingly, only strains homozygous for E ^k expressed the idiotype, and expression was a recessive trait. Possible mechanisms for this control of idiotype expression and its relation to lymphomagenesis are discussed.Abbreviations used in this paper 2 ^a4^b B10.H-2 ^aH-4^bp/Wts - Br-MRBC bromelain-treated mouse erythrocytes - C complement - LPS lipopolysaccharide W - pfc plaque-forming cells - sIg surface immunoglobulin - SRBC sheep erythrocytes - Ts T suppressor.     

H-2 antigen variants in a cultured heterozygous mouse leukemia cell line

In previous studies on the emergence of H-2 antigen loss variants from an H-2^b/H-2^d heterozygous mouse leukemia cell line, it has been shown that stable variants could be obtained after a single-step selection procedure with antibody and complement. Selection against the K-end alleles revealed an asymmetry in the types of variant obtained. This paper deals with the results of selection against theD-end alleles. Once again, an asymmetry is noticed. Selection against the H-2D^d gene product results in the isolation of stable variants that have lost H-2D^d, either alone or in combination with the linked H-2K^d antigens. Selection against H-2D^b, on the other hand, has not, so far, resulted in the isolation of any stable variants. These experiments have not demonstrated unequivocally the mechanism by which these H-2 antigen loss variants are generated. However, certain preliminary considerations suggest the possibility that these variants may not, in fact, be true genetic mutants. The alternative -that these variants arose through some nongenetic, but stable mechanism — must be considered seriously.     

A novel MHC class I-related molecule expressed on mouse thymic stroma cells and mature lymphocytes

A monoclonal antibody (mAb) TP-3 has been established by immunizing rats with the BALB/c mouse thymic epithelial cell line TEL-2. The TP-3 antigen is expressed on stroma cells of thymus, spleen, and lymph node in syngeneic BALB/c mice (H-2 ^d ). This antigen is also expressed at a low level on the cell surface of immature thymocytes, and at a high level on mature T and B cells. In allogeneic mice such as C57BL/6 (H-2 ^b ) or C3H (H-2 ^k ), no cells expressed the TP-3 antigen. Using H-2 congenic mice, reactivity with mAb TP-3 was found to map to a region of H-2D ^d L ^d or between D ^d and Qa, suggesting that TP-3 is a major histocompatibility complex (MHC) class I antigen. However, immunoprecipitation analysis indicated that this antigen is not identical to the classical mouse class I molecules in terms of molecular size, antigenicity, and tissue distribution.     

Identification of a new CML target antigen controlled by a gene associated with theQa-2 locus

BALB/cBy anti-BALB/cJ spleen cells were tested in a secondary cellmediated lympholysis assay. The effector cells generated displayed a positive cytotoxic effect against Con A lymphoblasts from only those strains that were typed serologically as having theQa-2 ^a allele. Confirmation that the target antigen is controlled by a locus closely associated with or identical toQa-2 was obtained by the findings that target cells from B6.K2 (Qa-2 ^a,Qa-3 ^a) mice were lysed by the effector cells, while those from theQa-2, 3 congenic strain B6.K1 (Qa-2 ^b,Qa-3 ^b) were not. The fact that target cells from aQa-2-positive/Qa-3-negative strain (DBA/1,Qa-2 ^ai,Qa-3 ^b) were killed indicates that the target antigen is controlled, at least in part, by theQa-2 locus, not the Qa-3.There is no observedH-2 genetic restriction for this cytotoxic effect, since target cells which have theQa-2 ^a allele but differ from the stimulator cells at theH-2K, D, andI regions were lysed efficiently.     

H-2-linked genes determine the level of the primary in vitro anti-Mls response

The level of cell proliferation and interleukin-2 (IL-2) production observed in an anti-Mls mixed lymphocyte reaction between spleen cells from H-2 compatible, Mls incompatible mouse strains is determined by the H-2 haplotype of the mouse combination. Thus, while AKR (H-2 ^k) spleen cells stimulated strong M1s^a responses in H-2^k responder cells, AKR H-2b spleen cells stimulated no or negligible M1s^a responses in responder cells from H-2 ^bmouse strains. This effect was observed at the levels of IL-2 production and cell proliferation. The magnitude of the response observed using F₁ (H-2 ^k/H-2 ^b) responder cells was found to be a function of stimulator rather than responder cells. The poor stimulatory capacity of AKRH-2 ^bspleen cells was also shown not to be due to the loss of the stimulatory Mls ^aallele during the construction of the congenic strain from AKR and C57BL/6 parental strains. Using stimulator cells from a second series of congenic mice, we found H-2 ^b(strain DLLP) again to represent a poorly Mls^a stimulatory H-2 haplotype. In addition, H-2^q (DBA/1) cells displayed very poor Mls^a stimulatory potential while H-2^d (D1.C) cells were efficient Mls^a stimulators. Again the effect was shown to be at the level of the stimulator cells. In toto, our findings indicate that the H-2 ^kand H-2 ^dhaplotypes encode strong Mls^a stimulatory potential while the H-2 ^band H-2 ^qhaplotypes determine poor Mls^a stimulatory potential in primary in vitro responses, measured as cell proliferation and IL-2 production.Abbreviations used in this paper: CTL cytotoxic T lymphocyte - IL-1 interleukin-1 - IL-2 interleukin-2 - MLR mixed lymphocyte reaction - NMS normal mouse serum     

Study of H-2 mutations in mice. IV. A comparison of the mutants M505 and Hz1 by skin grafting and serological techniques

The line B6.M505 is congenic with C57BL/6JY and carries a mutant form of theH-2 ^b haplotype designatedH-2 ^bd . The mutant site 505 was located by the F₁ tests in theK end of theH-2 gene complex. The M505 mice are histoincompatible with the B6.C(Hz1) line (haplotypeH-2 ^ba ) carrying another mutation in theK end ofH-2 ^b . Inability of M505 to complement Hz1 in tests with B6 skin grafting is considered as an evidence that the same gene was altered by both mutations. The gained H antigens of two mutants can cross-react in vivo as revealed by accelerated rejection of Hz1 skin grafts by B6 recipients presensitized with M505 spleen cells. The lost antigenic determinants are not identical as shown by accelerated rejection of B6 skin grafts by Hz1 hosts preimmunized with M505 spleen cells. Absorptions of the antiserum ASY-015, (d×a) anti-i, anti-H-2.33 with M505 spleen cells did not clear forH-2 ⁱ ,H-2 ^b andH-2 ^ba , and absorptions with Hz1 did not clear forH-2 ⁱ ,H-2 ^b , andH-2 ^bd . These results show that changes of histocompatibility determinants may be accompanied by loss of some haptenic determinants in the Hz1 and M505 mutations.     

Influence of the H-2 u haplotype on immune function in F1 hybrid mice

Recently we reported that antigen-primed T cells from (H-2 ^u × H-2 ^s)F₁ and (H-2 ^u × H-2 ^q)F₁ mice responded poorly in vitro to antigen in the context of antigen-presenting cells of the non-H-2 ^u parent. It was suggested that this effect might be due to unbalanced expression of parental antigens in the F₁ hybrid with the result that the non-H-2^u A antigens were greatly reduced or absent in these mice. If this were the case, non-H-2^u Ia-A cells might be expected to stimulate a mixed lymphocyte reaction (MLR) when cultured with Fl responder cells. When tested, (SJL × PL)F₁ responder cells reacted strongly to SJL stimulator cells. There was no significant reaction to PL stimulator cells. The use of major histocompatibility complex (MHC) congenic mice showed the stimulatory antigens to be associated with the MHC. The MLR could be blocked significantly by monoclonal A-specific antibody of the appropriate specificity. When a monoclonal antibody reactivewith a private epitope associated with A^s was used to probe for the presence of A^s on the surface of (SJL × PL)F₁ spleen cells, no antigen could be detected, indicating loss or alteration of this antigen. These findings suggest that an alteration of the expression of the parental A^s molecule may be responsible for this phenomenon.Abbreviations used in this paper APC antigen-presenting cells - BSS balanced salt solution - CTL cytotoxic T lymphocyte - IL-2 interleukin-2 - MHC major histocompatibility complex - MLR mixed lymphocyte reaction - T₂ suppressor T lymphocyte     

Class I MHC molecules rather than other mouse genes dictate influenza epitope recognition by cytotoxic T cells

Influenza nucleoprotein (NP) is an important target antigen for influenza A virus cross-reactive cytotoxic T cells (Tc). Here we examine the NP epitope recognized by cloned and polyclonal BALB/c Tc and the genetics of this recognition pattern. We can define NP residues 147–161 as the epitope seen in conjunction with K ^d , the only H-2^d class I responder allele for NP restriction. H-2 ^d /H-2 ^b F₁ mice (C57BL × DBA/2) primed by influenza infection lyse only H-2^d target cells treated with peptide 147–161 while H-2^b targets are recognized only after treatment with NP residues 365–379 (previously found to be recognized by D^b restricted Tc cells). Tc cell recognition of NP peptide 147–161 is entirely dictated by expression of K ^d and not by other B10 or OH background genes of congenic mice. Restriction of a unique NP sequence by each responder class I major histocompatibility complex (MHC) allele suggests that antigen and class I MHC interact for Tc recognition.     

H-2-associated differences in replicated strains of mice divergently selected for body weight

A random-bred strain (Q) was established and divided into six replicates. Each replicate was divergently selected for 6-week weight (for over 30 generations) and each had an unselected control. We have investigated the H-2 haplotype of individual mice of the 18 selected Q strains to determine whether selection for size had also selected for H-2 or H-2-linked genes. From the results it appeared that only the H-2 ^b and H-2 ^q haplotypes were present in the foundation stock. A large number of individuals of the six small sublines were of H-2 bhaplotype, while the majority of those of the six large sublines were of the H-2 ^q haplotype. Individuals in the six control strains were H-2 ^b , H-2 ^q or both (i. e., H-2 heterozygotes and/or H-2 recombinants). These results suggest that control of body size is associated with H-2 or an H-2-linked gene(s).     

Thirteen new chromosome-7 congenic lines

Thirteen new congenic lines have been produced which have chromosome-7 segments introduced from different strains onto the C57BL/10Sn background. Sublines B10.P(61NX)C,D, and E received chromosome-7 segments from P/J, B10.CE(62NX) from CE/J, B10.SEC(64NX)A,C,E, and F from SEC/1Re, B10.SM(65NX) from SM/J, B10.WB(66NX) from WB/Re, B10.A(67NX) from A/SnGrf, B10.AKR(68NX) from AKR/SnGrf, and B10.K(69NX) from C3H.K. Isograft testing indicated that three sublines, B10.P(61NX)D, B10.CE(62NX)B, and B10.WB(66NX)B are histoisogenic, i.e., histocompatible within each line. With the exception of B10.A(67NX), B10.AK(68NX), and B10.K(69NX), which have not been isografted, the remaining sublines showed residual heterozygosity on isografting. The three histoisogenic lines have undergone F₁ testing and have been found to possess theH-4 ^a allele and new and distinct alleles at theH-1 locus. They have been designated B10.P(61NX)-H-4^a H-1 ^d , B10.WB(66NX)-H-4^a H-1 ^e , and B10.CE(62NX)-H-4^a H-1 ^f . Direct exchange of grafts has indicated the following genotypes: B10.A(67NX)-H-4^a H-1 ^b , B10.AK(68NX)-H-4^a H-1 ^b , and B10.K(69NX)-F-4^a H-1 ^b . The B10.SEC(64NX) and B10.SM(65NX) sublines have not been typed completely forH-4 andH-1. F ₁ testing or direct exchange of skin grafts indicated that B10.P(61NX)-H-4^a H-1 ^d , B10.WB(66NX)-H-4^a H-1 ^e , B10.A(67NX)-H-4^a H-1 ^b B10.AK(68NX)-H-4^a H-1 ^b and B10.K(69NX)-H-4^a H1 ^b possess nonon-H-1 histocompatibility differences from the G57BL/10 background.     

Use of hybridoma-resistant variant cell lines to study H-2D b/FMR-specific cytotoxic T cells

An H-2D ^b b heterozygous tumor cell line and a variant subclone bearing a mutant gene product were used to analyze the H-2D^b specificity of cytotoxic T lymphocytes (CTL) generated during a Moloney murine sarcoma virus (MSV) infection. When the mutant cells were used as targets for MSV-specific CTL, the amount of cell lysis, compared with that seen with the nonmutant parental cells, was drastically decreased. However, cells of the mutant clones remained susceptible to allogeneic CTL specific for the nonmutant H-2D^b molecule. The mutant cells also did not differ from the parent cells in their level of viral antigen expression. Biochemically the parental and mutant molecules were similar but not identical. The data indicate that minor alterations of the H-2 antigens caused by somatic mutation may prevent virus-infected cells from being recognized as targets by CTL.     

Trans-regulatory genes affect Slpa and Slpo expression and act in a tissue-specific manner

The plasma protein C4 and its androgen-dependent homologue Slp are encoded by genes located in the mouse major histocompatibility complex, H-2. The C4 and Slp protein levels and liver mRNA levels are influenced by non-H-2-linked regulatory genes. The allele-specific regulation of C4 expression and the androgen regulation of Slp expression are manifest only in some of the tissues where these genes are expressed. Therefore, we studied tissues in which the effects of the non-H-2 regulatory genes are apparent. We show that these genes only affect the Slp expression in tissues where it is androgen-dependent. This indicates that the non-H-2 regulatory genes most likely act through interaction with the androgen regulation of Slp expression. We also show that liver cells of mice with the Slp ^o allele, which do not produce Slp protein, do express Slp mRNA; this expression is also androgen-induced and regulated by non-H-2 genes. Thus, both the Slp ^a and Slp ^o alleles appear to be regulated in the same way.