Preleukemic expression of TL antigens in x-irradiated C57BL/6 mice.

Anomalous appearance of TL (thymus-leukemia) antigens is a characteristic feature of radiation-induced leukemias of C57bl/6 mice. We now report that thymocytes of irradiated C57BL/6 mice express TL antigens long before the development of overt leukemia. Thus, TL is a marker for preleukemic changes occurring during radiation leukemogenesis. Low levels of murine leukemia virus (MuLV)-related antigens are also detected on preleukemic thymocytes. Comparative tests on individual mice show no direct correlation between TL and MuLV antigen expression.

The G-IX system. A cell surface allo-antigen associated with murine leukemia virus; implications regarding chromosomal integration of the viral genome.

This report concerns a cell surface antigen (G(IX); G = Gross) which exhibits mendelian inheritance but which also appears de novo in cells that become productively infected with MuLV (Gross), the wild-type leukemia virus of the mouse. In normal mice, G(IX) is a cell surface allo-antigen confined to lymphoid cells and found in highest amount on thymocytes. Four categories of inbred mouse strains can be distinguished according to how much G(IX) antigen is expressed on their thymocytes. G(IX) (-) strains have none; in the three G(IX) (+) categories, G(IX) (3), G(IX) (2), and G(IX) (1), the amounts of G(IX) antigen present (per thymocyte) are approximately in the ratios 3:2:1. A study of segregating populations derived mainly from strain 129 (the prototype G(IX) (3) strain) and C57BL/6 (the prototype G(IX) (-) strain) revealed that two unlinked chromosomal genes are required for expression of G(IX) on normal lymphoid cells. The phenotype G(IX) (+) is expressed only when both genes are present, as in 129 mice. C57BL/6 carries neither of them. At one locus, expression of G(IX) is fully dominant over nonexpression (G(IX) fully expressed in heterozygotes). At the second locus, which is linked with H-2 (at a distance of 36.4 +/- 2.7 units) in group IX (locus symbol G(IX)), expression is semidominant (50% expression of G(IX) in heterozygotes); gene order T:H-2:Tla:G(IX). As a rule, when cells of G(IX) (-) mice or rats become overtly infected with MuLV (Gross), an event which occurs spontaneously in older mice of certain strains and which also commonly accompanies malignant transformation, their phenotype is converted to G(IX) (+). This invites comparison with the emergence of TL(+) leukemia cells in TL(-) mouse strains which has been observed in previous studies and which implies that TL(-) --> TL(+) conversion has accompanied leukemic transformation of such cells. So far the only example of G(IX) (-) --> G(IX) (+) conversion taking place without overt MuLV infection is represented by the occurrence of GCSA(-):G(IX) (+) myelomas in BALB/c (GCSA:G(IX) (-)) mice. Unlike the other Gross cell surface antigen described earlier, GCSA, which is invariably associated with MuLV (Gross) infection and never occurs in its absence, G(IX) antigen sometimes occurs independently of productive MuLV infection; for example, thymocytes and some leukemias of 129 mice are GCSA(-):G(IX) (+), and MuLV-producing sarcomas may be GCSA(+):G(IX) (-). The frequent emergence of cells of G(IX) (+) phenotype in all mouse strains implies that the structural gene coding for G(IX) antigen is common to all mice. There is precedent for this in the TL system, in which two of the Tla genes in linkage group IX appear to be ubiquitous among mice, but are normally expressed only in strains of mice carrying a second (expression) gene. It is not yet certain whether either of the two segregating genes belongs to the MuLV genome rather than to the cellular genome. This leaves the question whether MuLV may have a chromosomal integration site still debatable. But there is a good prospect that further genetic analysis will provide the answer and so elucidate the special relationship of leukemia viruses to the cells of their natural hosts.

Isoantigens of the H-2 and Tla loci of the mouse: interactions affecting their representation on thymocytes.

H-2 and TL isoantigens of the mouse are specified by the closely linked genetic loci H-2 and Tla. A. study of their representation on thymocytes was performed in order to reveal any interactions between the determinant genes or their products affecting the synthesis or disposition of these components of the thymocyte surface. The method employed was quantitative absorption of cytotoxic antibody by viable thymocytes. The phenotypic expression of TL antigens was found to reduce the demonstrable amount of certain H-2 antigens to as little as 34% of the quantity demonstrable on TL- thymocytes. A reduction was observed in all three H-2 types tested, (H-2(b), H-2(a), and H-2(k)). As antigenic modulation (change of TL phenotype from TL+ to TL-, produced by TL antibody) is known to entail a compensatory increase in H-2(D) antigen, it is concluded that the TL phenotype, rather than the Tla genotype, influences the surface representation of H-2 antigens. The two known TL+ phenotypes of thymocytes (TL.2 and TL.1,2,3) depress H-2 equally. The H-2 specificities affected are those determined by the D end of the E-2 locus, which is adjacent to Tla; antigens of the K end, which is distal to Tla, are not depressed. The reduction of demonstrable H-2 antigen on the thymocytes of TL+ x TL- progeny is half that of thymocytes of TL+ x TL+ progeny and the reduction affects equally the products of both H-2 alleles (cis and trans in relation to Tla), indicating that the mechanism of H-2 reduction by TL is extrachromosomal. Whether it involves diminished synthesis of H-2 or steric masking by TL at the cell membrane is unknown, but in either case the reciprocal relation of TL and H-2(D) antigens implies that they probably occupy adjacent positions on thymocytes and that the gene order, H-2(K): H-2(D):Tla is reflected in cell surface structure. Extrachromosomal interaction, apparently involving control of synthesis, occurs also within the TL system of antigens. Thymocytes of TL.2 x TL.1,2,3 progeny express the full homozygous quantity of antigens TL.1 and TL.3 (but not of TL.2), in contrast to the half-quantity present in thymocytes of TL- x TL.1,2,3 progeny. Another example of interaction is implicit in the finding that thymocytes of TL-1,2,3 x TL.1,2,3 progeny have more TL.2 antigen than thymocytes of TL.2 x TL.2 progeny, but in this instance there is nothing to indicate whether the mechanism is chromosomal or extrachromosomal. Thus the quantitative surface representation of at least some H-2 and TL antigens is influenced by the cellular complement of H-2:Tla genes as a whole. Comparison of H-2 heterozygous thymocytes with H-2 homozygous thymocytes in quantitative absorption tests shows (a) more than the expected 50% of each parental-type H-2 antigen on heterozygous cells, and (b) a greater suppression of H-2 by TL in H-2 heterozygotes in comparison with H-2 homozygotes. Both results may be explained on the basis of differences in the density of H-2 antigenic sites and consequent differences in the efficiency of absorption of H-2 antibody. These considerations may be useful in other contexts, e.g. in estimating the representation of Rh antigens on the red cells of human subjects homozygous and heterozygous for Rh components.

Thymus-leukemia (TL) antigens of the mouse. Analysis of TL mRNA and TL cDNA TL+ and TL- strains.

A thymus-leukemia (TL)-specific probe, pTL1, has been generated from a TL-coding gene of BALB/c mice. Multiple species of TL mRNA were detected in TL+ cells by Northern blot analysis with pTL1, and different Tla haplotypes could be distinguished on the basis of characteristic patterns of TL mRNA. No TL-related message was found in normal or leukemic TL- cells, including thymocytes from Tlab mice. However, TL mRNA could be detected in TL+ leukemias occurring in Tlab mice. A cDNA library has been made from ASL1 (a TL+ leukemia of A mice [Tlaa]), and pTL1+ clones have been sequenced. At least three structurally distinct TL genes are expressed in ASL1. Sequence comparison of TL genes from three Tla haplotypes indicates that TL genes are highly conserved (greater than 90% homology) and are more distantly related to H-2 genes. Several polyadenylation sites have been found in the 3' untranslated region of TL genes, and differential polyadenylation contributes to the size heterogeneity of TL transcripts. The predicted amino acid sequence of TL products indicates that TL and H-2 are similar in domain structure and disulfide bonds, but differ in glycosylation sites and in cytoplasmic domain sequences.

Mouse isoantigens: separation of soluble TL (thymus-leukemia) antigen from soluble H-2 histocompatibility antigen by column chromatography.

Mouse H-2 histocompatibility antigen has been extracted, solubilized, and partly purified from the cells of an A strain spontaneous leukemia carrying TL (thymus-leukemia) antigens. H-2 and TL. 1, 2, 3 activities were measured by inhibition of the cytotoxic effect of the corresponding isoantibodies. TL activity was associated with the H-2 active fraction obtained by solubilization and fractionation by gel filtration. TL specificity was largely separated from H-2 antigen by subsequent chromatography on DEAE Sephadex as an adjacent component in a series of fractions. The soluble H-2 antigen prepared from the leukemia cells was tested for most of the specificities determined by H-2(a) with no exceptional results. TL. 1, 2, 3 activities, measured as each component separately, were located in approximately the same position; there is no clear indication yet whether the three TL specificities are separable from one another. It appears that in addition to the close genetic linkage between the H-2 and TL loci, and their reciprocal interaction in producing H-2 and TL antigens, these antigens exhibit some similarity at the chemical level.

Relation of GIX antigen of thymocytes to envelope glycoprotein of murine leukemia virus.

Expression of Gix surface antigen on thymocytes is an inherited mendelian train of certain strains of mice. We report here the following new findings: (a) Gix antigen was found free in the serum of Gix+ mouse strains. (b) Expression vs. nonexpression of Gix antigen was invariably correlated with presence or absence of the group-specific antigen of Murine leukemia virus (MuLV) gp69/71 in the serum of mice of inbred and segregating populations. (c) Gix antigen could be removed from normal Gix+ mouse serum by precipitation with antiserum to MuLV gp 69/71. (d) Anti-gp69/71 serum was weakly cytotoxic for Gix+ thymocytes, and partially blocked the cytotoxic activity of Gix antibody for Gix+ thymocytes. (e) Purified AKR virus absorbed Gix activity, and disruption of the virions did not increase their absorbing capacity. These serological data indicate that Gix antigen is a constituent of gp69/71, the glycoprotein which is the major component of the MuLV envelope. On present evidence, Gix antigen is represented in intact virions and is probably accessible to Gix antibody.

Age-related changes in cell surface antigens of preleukemic AKR thymocytes.

Thymocytes from preleukemic AKR mice aged 5-6 mo have an altered pattern of cell surface antigens. The expression of four MuLV-related antigens on the cell surface (GIX, GCSA, gp70, p30) is markedly increased in comparison to 2-mo-old AKR mice and approximates the heightened levels of these antigens found on thymic leukemia cells. H-2 and Thy-1 alloantigens also show characteristic modifications in relation to age and leukemia development. In contrast to the high Thy-1/low H-2 levels on 2-mo-old AKR thymocytes, thymocytes from 6-mo-old mice and thymic leukemia cells frequently show a low Thy-1/high H-2 surface phenotype. As thymocytes from mouse strains with a low incidence of leukemia do not show these changes, they appear to represent a stage in the conversion of normal cells to leukemia cells.

Antigenic modulation. Loss of TL antigen from cells exposed to TL antibody. Study of the phenomenon in vitro.

Antigenic modulation (the loss of TL antigens from TL+ cells exposed to TL antibody in the absence of lytic complement) has been demonstrated in vitro. An ascites leukemia, phenotype TL.1,2,3, which modulates rapidly and completely when incubated with TL antiserum in vitro, was selected for further study of the phenomenon. Over a wide range of TL antibody concentrations modulation at 37 degrees C was detectable within 10 min and was complete within approximately 1 hr. The cells were initially sensitized to C' by their contact with antibody, thereafter losing this sensitivity to C' lysis together with their sensitivity to TL antibody and C' in the cytotoxic test. The capacity of the cells to undergo modulation was abolished by actinomycin D and by iodoacetamide, and by reducing the temperature of incubation to 0 degrees C. Thus modulation apparently is an active cellular process. Antigens TL. 1,2, and 3 are all modulated by anti-TL.1,3 serum and by anti-TL.3 serum. This modulation affects all three TL components together, even when antibody to one or two of them is lacking. aAnti-TL.2 serum does not induce modulation and in fact impairs modulation by the other TL antibodies. The influence of the TL phenotype of cells upon the demonstrable content of H-2 (D region) isoantigen, first shown in cells modulated in vivo, has been observed with cells modulated in vitro. Cells undergoing modulation show a progressive increase in H-2 (D region) antigen over a period of 4 hr, with no change in H-2 antigens of the K region. Restoration of the TL+ phenotype of modulated cells after removal of antibody is less rapid than TL+ --> TL- modulation and may require several cell divisions.

Source and hormone-dependence of Gix-gp70 in mouse serum.

The gp70 family of glycoproteins is distinguished by the role of these molecules as constituents of C-type viral envelopes and also as Mendelian cellular constituents expressed independently of virus production. The source of G(IX)-gp70 in the serum of 129 strain mice, which are not overt producers of virus, could not be traced to any organ or tissue that is known to be G(IX)-positive by serological tests. Hematopoietic tissues were excluded as source of serum G(IX)-gp70 by tests with reciprocal radiation chimeras made from 129 and 129-G(IX)(-) donors and recipients. Thymus and spleen were excluded because excision of these organs did not affect levels of G(IX)-gp70 in the serum. The serum of young adult 129 males contains roughly four times as much G(IX)-gp70 as adult 129 females and the levels rise in both sexes with increasing age. Castration of 129 males reduced the level of serum G(IX)-gp70 to that of females, and the level was fully restored by testosterone. Thus the epididymis and seminal fluid, though rich in G(IX)-gp70, do not contribute significant amounts of G(IX)-gp70 to the serum. The level of G(IX)-gp70 in the serum of testosterone-treated females, though more than double that of untreated females, did not reach the level of normal males, under the conditions tested. This may signify that G(IX)-gp70 production by males is subject to imprinting by testosterone in early life. Evidently the main source of serum G(IX)-gp70 is a tissue or organ that is common to males and females, is directly or indirectly responsive to testosterone, and has not so far been identified serologically as G(IX)- positive.

New mutant and congenic mouse stocks expressing the murine leukemia virus-associated thymocyte surface antigen GIX.

For several reasons the G(IX) antigen (1) has a prominent place in current work on murine leukemia virus (MuLV): In the prototype G(IX+) mouse strain 129, the G(IX) trait is mendelian, and is expressed selectively (though not exclusively) on thymocytes. Thus, expression of this cell surface component is under the control of cellular genes and is subject to the controls governing the differentiation of T lymphocytes (2). Although the 129 mouse produces no demonstrable leukemia virus such as that found in the AKR strain, it was soon realized that G(IX) antigen must in some way be related to MuLV, because productive infection with MuLV is frequently associated with appearance of G(IX) antigen on cells that are genotypically G(IX-), most notably on MuLV-infected rat cells, or cells that belong to other differentiation pathways (1). The basis of this connection between G(IX) and MuLV has recently become clear from the demonstration that G(IX) is one of MuLV envelope. Therefore, our working hypothesis is that the presence of G(IX) is one of the antigens present on gp69/71 (3,4), the major glycoprotein component of the MuLV envelope. Therefore, our working hypothesis is that the presence of G(IX) antigen always denotes the presence of gp69/71 (though not all variants of gp69/71 need necessarily carry G(IX)). Study of the circumstances under which G(IX) is expressed on the cell surface is thus potentially a powerful approach to understanding how the expression of C-type viral genomes is controlled. Such studies are greatly facilitated by the availability of mutant and congenic strains of inbred mice which differ from the nonmutant or partner strains only with respect to one or another manifestation of the viral genome. It is for this reason that we record here (Table I) some details of two G(IX) mutant and two G(IX) congenic stocks derived in our colonies at Memorial Sloan-Kettering Cancer Center (MSKCC). In addition, to these four strains, Table I includes data for the three relevant partner strains, and for strain AKR, for comparison. These eight strains all differ from one another with respect to one or more MuLV-related traits.

Relationship of infectious murine leukemia virus and virus-related antigens in genetic crosses between AKR and the Fv-1 compatible strain C57L.

In a further genetic study of murine leukemia virus (MuLV) and its components we examined the backcross C57L X (C57L X AKR). This population was selected because strains AKR and C57L are both Fv-1n, and the restriction which the Fu-1b allele imposes on the output of virus was thereby obviated. The segregants were scored for three characters: (a) infectious Gross-AKR-type MuLV (V), in the tail; (b) group-specific antigen indicative of p30 internal viral protein, in spleen; and (c) GIX antigen, now thought to be indicative of gp69/71 viral envelope glycoprotein, on thymocytes. Our conclusions are: (a) It is confirmed that the AKR mouse has two unlinked chromosomal genes, Akv-1 and Akv-2, each of which can independently give rise to the life-long high output of MuLV that is characteristic of AKR mice. (b) Of the eight phenotypes that could possibly be derived from segregation of the three pairs of independent alternative traits, seven were observed, but on progeny testing only three were shown to reflect stably heritable genotypes; these were V+p30+GIX+ and V-p30-GIX- (the parental types) and V-p30+GIX+. A third, newly identified AKR gene, designated Akvp, segregating independently of Akv-1 and Akv-2, also determines expression of p30 and GIX but in this case independently of XC-detectable MuLV. (c) The four remaining observed phenotypes, which did not breed true on progeny testing, involved mostly antigen-negative parents yielding antigen-positive progeny; it is likely that these discrepancies represented suppression of phenotype by a maternal resistance factor.

Spontaneous autoimmunization to GIX cell surface antigen in hybrid mice.

The GIX antigen expressed on the thymocytes of GIX+ mice is a type-specific constituent of glycoprotein gp70, which forms the major envelope component of murine leukemia virus. In the prototype GIX+ mouse strain 129, this glycoprotein is a Mendelian character expressed independently of virus production. In the intact thymocyte plasma membrane, part of this glycoprotein, bearing group-specific (gs) antigen, is inaccessible to antibody. The moiety bearing the type-specific GIX determinant is accessible to GIX antibody, which may be an important factor in determining the consequences of autoimmune responses involving GIX. Previously, all attempts to induce GIX antibody in mice had failed. We now find that the hybrid mouse (B6-GIX+ X 129) spontaneously produces substantial amounts of GIX antibody, presumably of the IgM class appearing as early as 2 mo of age. The specificity of the GIX natural mouse antibody is the same as that recognized by the conventional GIX typing serum produced in rats ("anti-NTD"). As neither parent strain produces appreciable GIX antibody, we surmise that this autoimmune response requires two dominant genes, each parent contributing a high-response allele to the hybrid. These can be envisaged as two immune response loci, controlling different immunocompetent cells which must cooperate to produce GIX antibody. Production of GIX antibody by the hybrids increases progressively with age. This is accompanied by decreased expression of GIX antigen on their thymocytes. We attribute this to antigenic modulation. Antibody to gs antigen of gp70 is also found in autoimmune (B6-GIX+ X 129) hybrids but not in either parent strain. We are investigating evidence of a pathological autoimmune syndrome in these hybrids. The special interest of this syndrome is that it presumably signifies the consequences of autoimmunization to a single C-type virus component, expressed without significant virus production, in a mouse with no evident genetic predisposition to such disease in the absence of that antigen.

G(AKSL2): a new cell surface antigen of the mouse related to the dualtropic mink cell focus-inducing class of murine leukemia virus detected by naturally occurring antibody.

Normal mouse sera were tested for cytotoxic antibody to surface antigens of cultured monolayer cells infected with AKR-derived ecotropic MuLV, xentropic MuLV, or dualtropic MCF 247 MuLV. Antibody to ecotropic MuLV-infected cells was found in a proportion of C57BL/6, C3Hf/Bi, AKR-Fv-1b, and (C3Hf/Bi X AKR)F1 mice, but not AKR or (AKR X C3Hf/Bi)F1 mice. Antibody to xenotropic MuLV-infected cells was virtually restricted to C57BL/6 mice. Antibody to MCF 247-infected cells was found in all strains tested, including AKR mice. Absorption analysis of (C3Hf/Bi x akr)f1 and AKR-Fv-1b sera with selective reactivity for MCF 247-infected cells showed that these sera recognize distinctive antigens on MCF 247-infected cells that are not present on ecotropic or xenotropic MuLV-infected cells. The transplantable AKR spontaneous leukemia AKSL2 was found to be uniquely sensitive to the cytotoxic action of naturally occurring antibody to MCF 247-related antigens and absorption tests with AKSL2 as the target cell and sera from a single AKR-Fv-1b mouse have permitted the definition of a new MuLV-related cell surface antigen, which has been designated G(AKSL2). Thymocytes from young mice of high leukemia-incidence strains (AKR, C58, and PL) express G(AKSL2), whereas thymocytes from 12 other strains do not. In AKR mice, the antigen is expressed in higher amounts on cells from thymus and bone marrow than on spleen cells. All AKR spontaneous leukemias tested express G(AKSL2), as did three MuLV-induced leukemias arising in G(AKSL2)- strains. Five X-ray-induced leukemias of G(AKSL2)- strains were G(AKSL2)-, as were MuLV+ and MuLV- chemically induced sarcomas. In the limited survey conducted to date, natural antibody to G(AKSL2) has been restricted to strains expressing G(AKSL2) in their normal tissues: AKR, AKR congenic mice AKR-Fv-1b and AKR hybrid mice (C3Hf/Bi x akr)f1 and (C57BL/6 X AKR)F1. In vitro G(AKSL2) induction tests involving MuLV infection of cultured monolayer cells showed that 8 of 12 newly isolated dualtropic MuLV shared the property of G(AKSL2) induction with the prototype MCF MuLV, MCF 247. Of the 12 ecotropic MuLV tested, only the N-tropic MuLV isolated from a leukemia originally induced by Passage A Gross virus induced G(AKSL2). The xenotropic and amphotropic MuLV isolates tested lacked G(AKSL2) inducing activity. Recognition of the g(aksl2) system provides a way to trace the origin and natural history of a class of dualtropic MCF MuLV in the mouse and to determine whether natural antibody to G(AKSL2) plays a role in AKR leukemogenesis.

Cell surface antigens of murine leukemias induced by radiation leukemia virus. Recognition of individually distinct cell surface antigens by cytotoxic T cells on leukemias expressing crossreactive transplantation antigens.

The specificity of transplantation immunity and T cell cytotoxicity against leukemias induced by RadLV was examined. Subcutaneous inoculation of two RadLV leukemias induced in BALB/c mice, BALBRVB and BALBRVD, resulted in initial tumor growth in CB6F1 mice, followed by complete tumor regression. Mice that had rejected leukemias BALBRVB or BALBRVD were subsequently challenged with various tumors of BALB/c origin. The growth of all five RadLV leukemias tested, and of one radiation-induced leukemia, was significantly inhibited. Another radiation-induced leukemia, a methylcholanthrene-induced sarcoma, and a leukemia induced by the Moloney leukemia virus, were not inhibited. The results indicate that RadLV leukemias share cell surface antigens that induce transplantation immunity in vivo. Cytotoxic lymphocytes were generated by coculturing spleen cells from mice that had rejected leukemia BALBRVB or BALBRVD with the corresponding leukemia cells. Direct tests and inhibition tests showed that such cytotoxic cells recognized individually specific antigens on leukemias BALBRVB and BALBRVD, distinct from the shared antigens detected in transplantation experiments. The effector cells in cytotoxicity assays were Thy-1+, Lyt-1+,-, Lyt-2+, and Lyt-3+ T cells.

Relation of chromosome 4 (linkage group 8) to murine leukemia virus-associated antigens of AKR mice.

Genes specifying or controlling the expression of G(IX) (cell surface), GCSA (cell surface), and gs (internal viral) antigens are located in chromosome 4 (linkage group [LG] VIII) of the AKR mouse. All three antigens may exhibit mendelian inheritance, mice being antigen positive or antigen negative, but each may also appear in leukemic cells of mice whose inherited genotype was antigen negative. The G(IX)-determining gene in LG VIII of AKR mice apparently is equivalent to Gv-1, which determines expression of the same antigen in 129 strain mice, but which in the latter strain is located in LG IX. As the estimated distance of Gv-1 from H-2 in 129 mice is considerable (37 units) further tests are now indicated to assess the possibility of pseudolinkage in this case. The Fv-1 locus, also located in LG VIII, influences the mouse's titer of MuLV, and might thereby be thought to regulate the G(IX) and gs phenotypes of AKR backcross segregants. But the data indicate a discrete LG VIII locus for G(IX), since expression of this antigen is mendelian and independent of infectious virus titer. Since the G(IX) and GCSA phenotypes of AKR backcross segregants were invariably concordant, these two antigens must be specified or controlled by closely linked genes, and the latter also is presumably independent of virus titer. The question as to what extent expression of gs antigen in the segregants is secondary to virus production is undecided.

A survey of the humoral immune response of cancer patients to a panel of human tumor antigens.

Evidence is growing for both humoral and cellular immune recognition of human tumor antigens. Antibodies with specificity for antigens initially recognized by cytotoxic T lymphocytes (CTLs), e.g., MAGE and tyrosinase, have been detected in melanoma patient sera, and CTLs with specificity for NY-ESO-1, a cancer-testis (CT) antigen initially identified by autologous antibody, have recently been identified. To establish a screening system for the humoral response to autoimmunogenic tumor antigens, an enzyme-linked immunosorbent assay (ELISA) was developed using recombinant NY-ESO-1, MAGE-1, MAGE-3, SSX2, Melan-A, and tyrosinase proteins. A survey of sera from 234 cancer patients showed antibodies to NY-ESO-1 in 19 patients, to MAGE-1 in 3, to MAGE-3 in 2, and to SSX2 in 1 patient. No reactivity to these antigens was found in sera from 70 normal individuals. The frequency of NY-ESO-1 antibody was 9.4% in melanoma patients and 12.5% in ovarian cancer patients. Comparison of tumor NY-ESO-1 phenotype and NY-ESO-1 antibody response in 62 stage IV melanoma patients showed that all patients with NY-ESO-1(+) antibody had NY-ESO-1(+) tumors, and no patients with NY-ESO-1(-) tumors had NY-ESO-1 antibody. As the proportion of melanomas expressing NY-ESO-1 is 20-40% and only patients with NY-ESO-1(+) tumors have antibody, this would suggest that a high percentage of patients with NY-ESO-1(+) tumors develop an antibody response to NY-ESO-1.

TL antigen as a transplantation antigen recognized by TL-restricted cytotoxic T cells.

In contrast to broadly expressed classical class I antigens of the major histocompatibility complex, structurally closely related TL antigens are expressed in a highly restricted fashion. Unlike classical class I antigens, TL antigens are not known to be targets of cytotoxic T cells or to mediate graft rejection. Whereas classical class I antigens function as antigen-presenting molecules to T cell receptors (TCR), the role of TL is yet to be defined. To elucidate the function of TL, we have derived transgenic mice expressing TL in most tissues including skin by introducing a TL gene, T3b of C57BL/6 mouse origin, driven by the H-2Kb promoter. By grafting the skin of transgenic mice, we demonstrate that TL can serve as a transplantation antigen and mediate a TCR-alpha/beta+ CD8+ cytotoxic T cell response. This T cell recognition of TL does not require antigen presentation by H-2 molecules. Furthermore, we show that C57BL/6 F1 mice develop CD8+ T cells that are cytotoxic for C57BL/6 TL+ leukemia cells, providing further support for the concept that aberrantly expressed nonmutated proteins such as TL can be recognized as tumor antigens.

G(RADA1): a new cell surface antigen of mouse leukemia defined by naturally occurring antibody and its relationship to murine leukemia virus.

A new cell surface antigenic system of the mouse, designated G(RADA1), is described. The antigen is defined by cytotoxic tests with the A strain X-ray-induced leukemia RADA1 and naturally occurring antibody from random-bred Swiss mice and can be distinguished from all other serologically detected cell surface antigens of the mouse. Absorption tests indicate that G(RADA1) is present in the normal lymphatic tissue and leukemias of mouse strains with high spontaneous leukemia-incidence, e.g., AKR, C58, and C3H/Figge. Low leukemia-incidence strains, e.g., C57BL/6, BALB/c, and A lack G(RADA1) in their normal tissues, but a proportion of leukemias and solid tumors arising in these strains are G(RADA1)+. The relation of G(RADA1) to MuLV is shown by G(RADA1) appearance after MuLV infection of permissive cells in vitro; four of five N-tropic MuLV isolates, one of four B-tropic MuLV, and none of four xenotropic MuLV induce G(RADA1). Two MCF MuLV, thought to represent recombinants between N-ecotropic and xenotropic MuLV, also induce G(RADA1). Serological and biochemical characterization indicates that G(RADA1) is a type-specific determinant of the gp70 component of certain MuLV. The presence of natural antibody to RADA1 in various mouse strains and the emergence of G(RADA1)+ leukemias and solid tumors in mice of G(RADA1)- phenotype suggest widespread occurrence of genetic information coding for this antigen.

A cell surface antigen of the mouse related to xenotropic MuLv defined by naturally occurring antibody and monoclonal antibody. Relation to Gix G(rada1), G(aksl2) systems of MuLV-related antigens.

A new cell surface antigen of the mouse related to xenotropic murine leukemia virus (MuLV) is described. The antigen, designated G(erld), is defined by cytotoxic tests with the B6-x-ray-induced ERLD and naturally occurring antibody. G(erld) is distinct from all previously defined cell surface antigens. Monoclonal antibody with the same specificity has been developed. Inbred mouse strains are classified as G(erld)+ or G(erld)- according to the presence of absence of the antigen on lymphoid cells. G(erld)+ strains differ with regard to quantitative expression of G(erld) on normal thymocytes. The emergence of G(erld)+ tumors in G(erld)- strains indicates the presence of genes coding for the antigen even in strains not normally expressing the antigen. G(erld) has the characteristic of a differentiation antigen in normal mice. In G(erld)+ strains, high levels of the antigen are found on thymocytes with lower levels being detected on cells of spleen, lymph nodes and bone marrow. No G(erld) was detected in brain or kidney or on erythrocytes. The segregation ratios for G(erld) expression on thymocytes in backcross and F2 mice of crosses between G(erld)+ (B6, 129, and B6-Gix+) and G(erld)- (BALB/c) strains suggest that G(erld) expression is controlled by a single locus in B6, by two unlinked loci in 129, and by three unlinked loci in B6-Gix+ mice. Induction of the antigen by MuLV infection of permissive cells in vitro indicates that G(erld) is closely related to xenotropic and dualtropic MuLV; all xenotropic and dualtropic MuLV tested induced the antigen, whereas the majority of ecotropic and the two amphotropic MuLV failed to do so. As dualtropic MuLV are thought to be recombinants between ecotropic and xenotropic MuLV sequences, G(erld) coding by dualtropic MuLV may signify the contribution of the xenotropic part in the recombinational event. Serological and biochemical characterization indicates that G(erld) is related to the gp 70 component of the MuLV envelope. The relation of G(erld) to the previously defined gp 70-related cell surface antigens (Gix, G(rada), and G(aksl2) is discussed, particularly with regard to their characteristics as differentiation antigens, the genetic origin of dualtropic MuLV, and the leukemogenicity of MuLV.

Identification of NY-ESO-1 epitopes presented by human histocompatibility antigen (HLA)-DRB4*0101-0103 and recognized by CD4(+) T lymphocytes of patients with NY-ESO-1-expressing melanoma.

NY-ESO-1 is a member of the cancer-testis family of tumor antigens that elicits strong humoral and cellular immune responses in patients with NY-ESO-1-expressing cancers. Since CD4(+) T lymphocytes play a critical role in generating antigen-specific cytotoxic T lymphocyte and antibody responses, we searched for NY-ESO-1 epitopes presented by histocompatibility leukocyte antigen (HLA) class II molecules. Autologous monocyte-derived dendritic cells of cancer patients were incubated with recombinant NY-ESO-1 protein and used in enzyme-linked immunospot (ELISPOT) assays to detect NY-ESO-1-specific CD4(+) T lymphocyte responses. To identify possible epitopes presented by distinct HLA class II alleles, overlapping 18-mer peptides derived from NY-ESO-1 were synthetized and tested for recognition by CD4(+) T lymphocytes in autologous settings. We identified three NY-ESO-1-derived peptides presented by DRB4*0101-0103 and recognized by CD4(+) T lymphocytes of two melanoma patients sharing these HLA class II alleles. Specificity of recognition was confirmed by proliferation assays. The characterization of HLA class II-restricted epitopes will be useful for the assessment of spontaneous and vaccine-induced immune responses of cancer patients against defined tumor antigens. Further, the therapeutic efficacy of active immunization using antigenic HLA class I-restricted peptides may be improved by adding HLA class II-presented epitopes.