- Email: [email protected]
A deletion mutant defines DQβ variants within DR4 positive DQw3 positive haplotypes
A Deletion Mutant Defines DQfl Variants Within DR4 Positive DQw3 Positive Haplotypes Barbara S. Nepom, Se Jong Kim,* and Gerald T. Nepom
ABSTRACTS: We describe the production of an HLA deletion mutation by radiation mutagenesis of a DR4- and DQw3-homozygous, Dw4- and Dwl4-heterozygous cell line designed to analyze polymorphisms associated with DR4 and DQw3. Southern blot analysis confirms a deletion of class l and class II genes on one haplotype. Variation in DQfl alleles associated with DQw3 was previously described by characteristic RFLP patterns for a DQfl gene. One pattern, which corrdated precisdy with A-10-83 monodonal antibody reactivity (TAIO), defined an allele which we call DQ"3.1". The mutant cell line has lost the polymorphic bands on Southern blots corresponding to the DQ"3.1" allele, while the intact D w l 4 haplotype retains the alternate allele at DQfi which is DQw3-positive, TAlO-negative. These data demonstrate the segregation of two DQw3 positive DQfl allelic variants, both associated with DR4, which can be distinguished on the basis of both RFLP and monoclona] antibody reactivity. ABBREVIATIONS
EBV EDTA FACS FCS FITC SDS
Epstein-Barr virus ethylene diamine tetraacetate fluorescence activated cell sorter fetal calf serum fluorescein-isothyocyanate sodium dodecyl sulfate
IDDM JRA MoAb RFLP
insulin-dependent diabetes mellitus seropositive juvenile rheumatoid arthritis monoclonal antibody restriction endonuclease fragment length polymorphism
INTRODUCTION H L A - D R 4 is a c o m m o n H L A class II specificity which is of particular interest because of its association with a variety o f autoimmune disorders, such as rheumatoid arthritis, seropositive juvenile rheumatoid arthritis (IRA), and insulind e p e n d e n t diabetes mellitus (IDDM). We and others have previously shown that D R 4 represents a public specificity present on multiple different haplotypes. H o m o z y g o u s typing cells, all of which type as DR4, can be divided into at least six different groups on the basis of MLC reactivity [1,2] protein biochemistry [ 3 - 5 ] , and restriction fragment length polymorphisms (RFLP) [6]. In addition, we have described a DQ/3 polymorphism within the DR4-associated D Q w 3 specificity, as defined by serology and RFLP [7]. From Genetic Systems Corporation. Seattle, Washington. *Visiting Scientist, Department of Microbiology, Yonsei Unirersity Collegeof Medicine, Seoul. Korea. Address reprint requests to Barbara Nepom, M.D., Genetic Systems Corp.. 3005 First Avenue. Seattle. WA 98121. ReceivedJanuary 22, 1986; acceptedMarch 25, 1986.
Human Immunology17, 87-93 (1986) © ElsevierSciencePublishingCo., Inc., 1986 52 VanderbiltAve., New York, NY 10017
87
0198-8859/86/$3.50
B.S. Nepom et al.
88
In the present study, we used an EBV-transformed cell line from aJRA patient homozygous for DR4 and DQw3, but heterozygous for Dw4 and Dw14 and for DQ"3.1" and DQ"3.2", to create a haplotype deletion mutant by radiation mutagenesis. Comparison of antibody reactivity and RFLP analysis between the mutant and parental cell lines allows the investigation of polymorphisms present among haplotypes with identical DR and DQ typing but with disparate class II molecules. MATERIALS AND METHODS
Production of deletion mutant. The parental cell line 256 was derived by EBV transformation of peripheral blood lymphocytes from a patient with seropositive JRA, with the genotype A1, B8, DR4, DQw3, (Dw4)/A2, B15, DR4, DQw3, (Dw14) established by family study. 20 x 106 cells were irradiated with 400 rads from a Cesium source. Five days later, 4 x 106 cells remained. After washing with media, half of the remaining cells were treated with antibody P8.1 at 1:10,000 for 1 hr at room temperature, followed by treatment with rabbit complement (Pel-Freeze, Rogers, AR) for 1 hr, also at room temperature. After washing, the cells were resuspended in Iscove's media with 15% fetal calf serum (FCS) and cloned by limiting dilution in round bottom 96-well plates. One hundred thousand irradiated random donor human PBLs/well were used as feeder cells. Samples of wells showing growth were screened by immunofluorescence for expression of various HLA specificities.
Antibodies. Monoclonal antibodies P8.1 (anti-HLA-B8) [8], P4.1 (monomorphic anti-DR) [3], 17.15 (anti-DQw3) [9], and A-10-83 (anti-TA10) [7,10] have been previously described. Antibody 80.1 is an IgM monoclonal which recognizes some B locus alleles, including B15 (Bw62) but not B8 (K. Nelson, personal communication).
Immunofluorescence. After washing, pelleted cells were resuspended in 15 A of antibody in ascites fluid diluted at 1:160, kept on ice for 1 hr, washed, and resuspended in 15 A of affinity purified rabbit anti-mouse IgG conjugated to FITC for 60 min on ice. After washing twice with cold media, cells were fixed in 2% paraformaldehyde and analyzed on a FACS IV (Becton-Dickinson, Oxnard, CA). Of over 200 wells tested, a single well, B6G, was nonreactive with antibody P8.1, so was expanded for further evaluation.
Analysis ofgenomic DNA. RFLP analysis was performed as previously described [6]. Briefly, DNA was purified from EBV-transformed cell lines, digested with Barn HI or Eco RV endonuclease (BRL, Bethesda, MD) for 18 hr at 37°C, then precipitated with ethanol and resuspended in 10 mM Tris, 1 mM EDTA, pH 7.6. Digested DNA was then loaded at 12 /~g/lane on a 0.7% agarose gel and electrophoresed at 30 V for 18 hr. Gels were then denatured and neutralized prior to transfer to nitrocellulose (Schleicher and Schuell) by the method of Southern [11]. After transfer, filters were baked for 18 hr at 80°C, prehybridized at 42°C, and hybridized with nick translated 32P-labeled cDNA probes as previously described [6]. Following 48 hr hybridization at 42°C, blots were washed at 60°C in 1.5 mM NaCl, 0.015 mM Na citrate, and 0.1% SDS three times for 20 min each. Filters were exposed to Kodak XAR-5 film for 2-4 days. The DQ~ probe [12] is a full-length cDNA provided by Dr. P. Peterson. The class I B region probe, P158, is a 2 kb insert from the 3' end of a B7 cDNA inserted in pBR322, provided by Dr. P. A. Biro [13].
Deletion Mutant Defines DQ# Variants
89
Reoclivdy moAb Specificity Parent Mutant
None
--
--
--
P8 I
B8
+
--
815
+
+
80
I
FIGURE 1 FACS analysis of class I expression in deletion mutant B6G. Fluorescence histograms are shown for 4 × 101 ceils of parental cell line 256 (dashed line) and mutant cell line B6G (solid line); cell number is displayed on the vertical axis and relative fluorescence intensity using a log scale amplifier on the horizontal axis. Antibodies P8.1 (anti-B8) and P80.1 (anti-B 15) are described in the text.
RESULTS Analysis o f Cell S u r f a c e E x p r e s s i o n o f Class I and Class I I M o l e c u l e s on M u t a n t and Parental Lines Expression of H L A - B antigens on the parent and mutant cell lines was assessed by immunofluorescence and FACS analysis. Figure 1 shows the results using two different polymorphic class I antibodies: P8.1, which is specific for HLA-B8, and 80.1, which recognizes H L A - B 15 but not B8. Both cell lines are positive for the expression of B15, while B8 is clearly present on the parental cell line 256 but is completely negative on the mutant B6G, indicating that expression of this class I product has been lost on the mutant line. An M o A b recognizing the Al specificity, which is present on the B8 haplotype, was initially negative on the mutant while positive on the parent, but after several weeks in tissue culture the mutant became positive as well, suggesting that an early expression defect was later corrected (data not shown). After screening against a panel of informative class II MoAbs, as shown in Figure 2, results demonstrate that D R molecules, as recognized by the monomorphic D R antibody P4.1, are present on both the mutant and the parental cells. Because this antibody gives a broad flat peak on FACS analysis, quantitative differences in expression of D R molecules are not apparent from these curves. Similarly, the D Q w 3 specificity, as assessed by reactivity with M o A b 17.15, was strongly postive on both cell lines. Interestingly, however, A-10-83, a "TA10" reactive M o A b which splits the D Q w 3 specificity into two subgroups, was positive on the parental line 256 but completely negative on mutant B6G. Microcytotoxicity studies with this antibody confirm these results, with A-10-83 strongly killing cell line 256 while being completely negative on mutant B6G. Cell line B 6 G therefore appears to have lost expression of some class I and class II H L A specificities associated with the H L A - B and H L A - D Q regions. Reochv,ty moAb Specificity
~ ~
,
~
~
Parent Mutant
None
--
P4 ]
DR
+
+
1715
DOw3
+
+
+
-
A-10-83 DO"31"
FIGURE 2 FACS analysis of class II expression in deletion mutant B6G. The parental line 256 is represented by dashed lines, mutant B6G by solid lines. Antibodies P4.1 (anti-DR), 17.15 (anti-DQw3), and A-10-83 (anti-DQ"3.1") are described in the text.
90
B.S. Nepom et al.
R F L P Analysis o f Class I and Class II H L A Genes in M u t a n t B 6 G In order to confirm that the lack of expression described above was due to genetic deletion we performed Southern blot analysis of the two cell lines. Panel A in Figure 3 shows the restriction fragment length polymorphism (RFLP) pattern of the parental line in the left-hand lane compared with the mutant line in the right lane. Cell lines were digested with Eco RV and probed with a class I B region probe, P158. Similar bands are seen in both cases, with the exception of the 8.3 kb band seen hybridizing strongly in the parent line and lost in the mutant line. This 8.3 kb band is characteristic of the HLA-B8 allele in B8 homozygous cell lines, and thus indicates deletion of this allele in the mutant B6G. In order to investigate the genomic basis for the class II mutations suggested by loss of T A 1 0 reactivity, we performed RFLP analysis on the D Q region. We have previously described two alleles present at DQ/3 which represent variation within the DQw3 specificity, which we call DQ"3.1" and DQ"3.2" [7]. Southern blots distinguish between these alleles using a number of different restriction enzymes ([7], and unpublished). When genomic D N A is digested with Barn HI, the D Q " 3 . 1 " variant has a pair of bands at 6.9 and 3.7 kb, while the DQ"3.2" variant has instead a 12.0 kb band. Both alleles also have a constant 10.2 kb band on RFLP analysis representing another D Q locus, often called "DX/3" [14]. As seen in Panel B of Figure 3, the parental line 256, shown on the left, is heterozygous for both the D Q " 3 . 1 " and DQ"3.2" alleles. However, the mutant B6G, shown on the right, retains the 12.0 kb band characteristic of the DQ"3.2" allele, but has lost the 6.9 and 3.7 kb pair of bands, demonstrating that the DQ"3.2" allele is retained while the DQ"3.1" allele has been lost.
P kb -
~
~
~
,
30
B
kb -8.3
~ 10. 2
-3.7
:
1.85
FIGURE 3 RFLP analysis of class I and class II regions of deletion mutant B6G. Panel A shows a Southern blot of DNA from the parental and mutant cell lines digested with Eco RV and probed with a B region probe, p158. The 8.3 kb band, characteristic of the HLA-B8 specificity, is absent in mutant B6G. Panel B shows DNA from both cell lines digested with Barn HI and probed with a DQ/3 cDNA probe. The 6.9 and 3.7 kb bands defining the DQ"3.1" allele have been lost in the mutant, while it retains the 12.0 kb band characteristic of the alternative DQ"3.2" allele.
Deletion Mutant Defines DQ/3 Variants
91
Through the combination of both FACS analysis and RFLP analysis of this mutant cell line, we can describe the genotype of the mutant as shown schematically in Figure 4. Although the break points are not known exactly, we postulate that at least the DQ"3.1" through the B8 genes have been deleted. DISCUSSION HLA deletion mutant cell lines have proven extremely useful in the genetic analysis of the HLA system. Heterozygosity, even among some presumed "homozygous typing cells," has made the study of individual HLA loci and alleles somewhat confusing. Pious et al. [15], Kavathas et al. [16,17], and others have developed the techniques of radiation and chemical mutagenesis to produce mutant cell lines with hemizygous deletions, point mutations, and expression defects in the HLA region, allowing precise genetic mapping as well as evaluation of regulation and expression. We have been interested in utilizing these techniques to analyze HLA-DR4related haplotypes, which are of particular interest because of their association with autoimmune disorders such as rheumatoid arthritis, seropositive JRA, and IDDM. We and others have previously shown that DR4 is a public specificity present on multiple different haplotypes, which can be distinguished both phenotypically by MLC reactivity [1,2] and biochemically by analysis of DR surface molecules [3-5]. Like HLA-DR4, the DQw3 specificity has also been shown to be a public specificity present on different haplotypes. Giles et al., for example, have demonstrated that the molecules bearing the MB3 specificity (now known as DQw3) differ structurally between DR4- and DR5-related haplotypes [18]. Similarly, Ishikawa et al. have shown structural polymorphism among MB3 light chains from MB3-positive cells differing at DR (DR4,5,6, and 9) as well as within a single DR type (DR4,Dw4 and DR4,LD"KT2") [19]. In addition, we have recently described two allelic variants of DQ/3 genes corresponding to the DR4associated DQw3 specificity [7]. In the first variant, which we call DQ"3.1", 6.9 and 3.7 kb polymorphic bands are seen on RFLP analysis when cells are digested with Barn HI and probed with DQ/3. In the second variant, DQ"3.2", a 12.0 kb DQ/3 band is present instead. All cells showing the DQ"3.1" allele are recognized by TA10-reactive MoAb A-10-83 [7]. In the present study, we have created a haplotype deletion mutant by radiation mutagenesis from a cell homozygous for DR4 and DQw3, but heterozygous for Dw4 and Dw14 as well as for DQ"3.1" and DQ"3.2", in order to study DR4associated polymorphisms at the genomic level. We have shown, by surface
FIGURE 4 Schematicdiagram of the presumed haplotypes of the parental and mutant cell lines. The break points of the mutant's deletion are not precisely mapped, but the deletion is presumed to extend from at least the DQ through the B regions. DQ Parental 256
DR
B
A
--"51"--4--8--
1
--"5.2"--4--15--2
Mutant B6G
--~ . . . . . . . . . . . . . . . . . . . . --~'5.2"--
4 - - 15 - -
(Ow4) (Dw14)
~,
1-2- -
(Dw14)
92
B.S. Nepom et al. binding studies, that the specificities B8 and DQ"3.1" present on one haplotype are no longer expressed in the mutant line. This lack of expression is due to genetic deletion, as shown by loss of polymorphic bands in RFLP patterns of both the B8 and D Q " 3 . 1 " alleles on the mutant cell line. Although we do not yet have polymorphic genomic markers informative for the H L A - D R region on that same haplotype, the simplest explanation for the data is that this is a single deletion extending from at least the H L A - D Q through the HLA-B regions. The loss of the polymorphic 6.9 and 3.7 kb bands in the mutant B 6 G as demonstrated in the Southern blot in Figure 3B, in concert with the loss of reactivity to MoAb A-10-83, effectively maps the TA10 specificity to the DQw3associated haplotype which we call 3.1, and directly demonstrates the allelic nature of the D Q " 3 . 1 " and DQ"3.2" variants. Restriction fragment length polymorphisms (RFLP) have previously been used to discriminate among different HLA specificities [20], and sometimes among multiple haplotypes within a single serologically defined specificity [6,21]. However, these genomic polymorphisms frequently do not correlate with serologic reactivity. The description of RFLP patterns identifying the DQ"3.1" and DQ"3.2" alleles and the precise correlation of the DQ"3.1" RFLP with reactivity to MoAb A-10-83 marks one of the first examples of a genomic polymorphism within a single HLA specificity encoding an expressed, serologically defined product. The ability to distinguish among different DR4- and DQw3-related haplotypes has been of considerable interest in recent years. Analysis of variations identified by MLC [1,2], 2D gel electrophoresis [ 3 - 5 ] , serology [7], and RFLP [6,7] have clearly shown that different DRI3 and DQ/3 alleles consitute at least six different DR4 positive haplotypes. One of the questions of considerable functional importance is whether these specific haplotypes can confer susceptibility to different DR4-associated diseases. For example, we have shown that the DR4 haplotype associated with I D D M represents a different haplotype from that associated with seropositive JRA [22]; others have come to a similar conclusion for I D D M and adult rheumatoid arthritis [23]. Additional structural and functional analyses such as those described here using deletion mutants will elucidate the various DR4associated class II gene products contributing to such haplotypic variation.
ACKNOWLEDGMENTS
We thank Drs. K. Nelson, H. Maeda, P. Antonelli, and P. Martin for providing MoAbs used in these studies. We would also like to thank Margery Moogk for technical assistance and Holly Chase for manuscript preparation.
REFERENCES 1. Reinsmoen NL, Bach FH: Five HLA-D clusters associated with HLA-DR4. Hum Immunol 4:249, 1982. 2. Nose Y, Sato K, Nakagawa N, Kondoh K, Inouye H, Tsuji K: HLA-D clusters associated with DR4 in the Japanese population. Hum Immunol 5:199, 1982. 3. Nepom BS, Nepom GT, Mickelson E, Antonelli P, Hansen JA: Electrophoretic analysis of human HLA-DR antigens from HLA-DR4 homozygous cell lines: correlation between/3-chain diversity and HLA-D. Proc Natl Acad Sci USA 80:6962, 1983. 4. Nepom GT, Nepom BS, Antonelli P, Mickelson E, Silver J, Goyert SA, HansenJA: The HLA-DR4 family of haplotypes consists of a series of distinct DR and DS molecules. J Exp Med 159:394, 1984.
Deletion Mutant Defines DQ/3 Variants
93
5. Groner JP, Watson AJ, Bach FH: Dw/LD-related molecular polymorphism of DR4 /3-chains. J Exp Med 157:1687, 1983. 6. Holbeck SL, Kim SJ, Silver J, Hansen JA, Nepom GT: HLA-DR4-associated haplotypes are genotypically diverse within HLA. J Immunol 136:637, 1985. 7. Kim SJ, Holbeck SL, Nisperos B, Hansen JA, Maeda H, Nepom GT: Identification of a polymorphic variant associated with HLA-DQw3 characterized by specific restriction sites within DQB. Proc Natl Acad Sci USA 82:8139, 1985. 8. Antonelli P, Nisperos B, Braun M, Hansen JA: Recognition by a murine monoclonal antibody of a unique epitope specific for the human alloantigen HLA-B8. Hum Immunol 11:11, 1984. 9. Hansen JA, Martin P, Kamoun M, Nisperos B, Thomas ED: A supertypic HLA-DR specificity (DR4 + 5) defined by a murine monoclonal antibody. Hum Immunol 2:103, 1981. 10. Maeda H: Mouse monoclonal antibody detects a new polymorphic Ia determinant other than HLA-DR antigen: a possible allele of DC-1. Tissue Antigens 23:163, 1984. 11. Southern EM: Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98:503, 1975. 12. Larhammar D, Schenning L, Gustafson K, Wiman K, Cleasson L, Rask L, Peterson PA: Complete amino acid sequence of an HLA-DR antigen-like/3 chain as predicted from the nucleotide sequence: similarities with immunoglobulins and HLA-A, -B, and -C antigens. Proc Natl Acad Sci USA 79:3687, 1982. 13. Biro PA, Pan J, Sood AK, Kole R, Reddy VB, Weissman SM: The major histocompatibility complex. Cold Spring Harbor Quant Biol 1982 XLVII:1082, 1983. 14. Holbeck S, Nepom, GT: Analysis of HLA DQ/3 allelic variation using synthetic oligonucleotide probes, submitted for publication. 15. Pious D, Soderland C, Gladstone P: Induction of HLA mutations by chemical mutagens in human lymphoid cells. Immunogenetics 4:437, 1977. 16. Kavathas P, Bach FH, DeMars R: Gamma ray-induced loss of expression of HLA and glyoxalase I alleles in lymphoblastoid cells. Proc Natl Acad Sci USA 77:4251, 1980. 17. Kavathas P, DeMars R, Bach FH: Hemizygous HLA mutants oflymphoblastoid cells: a new cell type for histocompatibility testing. Hum Immunol 1:317, 1980. 18. Giles RC, Hurley CK, Capra JD: Primary structural variation among serologically indistinguishable DS antigens: the MB3-bearing molecule in DR4 cells differs from the MB3-bearing molecule in DR5 cells. J Immunol 133:1, 1984. 19. Ishikawa N, Kasahara M, Ikeda H, Ogasawara K, Hankin S, Takenouchi T, Wakasaka A, Kikuchi Y, Aizawa M: Demonstration of structural polymorphism among MB3 light chains by two-dimensional gel electrophoresis. J Immunol 134:417, 1985. 20. Dausset J, Cohen D: HLA at the gene level. In: ED Albert, MP Baur, WR Mayr, Eds. Histocompatibility testing 1984. Berlin, Springer-Verlag, 1984, p. 22. 21. Bosch ML, Schreuder GMT, Spits H, Termijtelen A, Tilanus MGJ, Giphart MJ: Polymorphism within the HLA-DRw6 haplotype I. Restriction fragment length variation and its correlation with serology. J Immunol 134:3212, 1985. 22. Nepom BS, Palmer J, Kim SJ, Hansen JA, Holbeck SL, Nepom GT: Specific genomic markers for the HLA-DQ subregion discriminate between DR4 positive IDDM and DR4 positive JRA. J Exp Med, in press. 23. Festenstein H, Awad J, Hitman GA, Cutbush S, Groves AV, Cassell P, Oilier W, Sachs JA: New HLA-DQ RFLP associations with rheumatoid arthritis and insulin dependent diabetes mellitus, submitted for publication.
ABSTRACTS: We describe the production of an HLA deletion mutation by radiation mutagenesis of a DR4- and DQw3-homozygous, Dw4- and Dwl4-heterozygous cell line designed to analyze polymorphisms associated with DR4 and DQw3. Southern blot analysis confirms a deletion of class l and class II genes on one haplotype. Variation in DQfl alleles associated with DQw3 was previously described by characteristic RFLP patterns for a DQfl gene. One pattern, which corrdated precisdy with A-10-83 monodonal antibody reactivity (TAIO), defined an allele which we call DQ"3.1". The mutant cell line has lost the polymorphic bands on Southern blots corresponding to the DQ"3.1" allele, while the intact D w l 4 haplotype retains the alternate allele at DQfi which is DQw3-positive, TAlO-negative. These data demonstrate the segregation of two DQw3 positive DQfl allelic variants, both associated with DR4, which can be distinguished on the basis of both RFLP and monoclona] antibody reactivity. ABBREVIATIONS
EBV EDTA FACS FCS FITC SDS
Epstein-Barr virus ethylene diamine tetraacetate fluorescence activated cell sorter fetal calf serum fluorescein-isothyocyanate sodium dodecyl sulfate
IDDM JRA MoAb RFLP
insulin-dependent diabetes mellitus seropositive juvenile rheumatoid arthritis monoclonal antibody restriction endonuclease fragment length polymorphism
INTRODUCTION H L A - D R 4 is a c o m m o n H L A class II specificity which is of particular interest because of its association with a variety o f autoimmune disorders, such as rheumatoid arthritis, seropositive juvenile rheumatoid arthritis (IRA), and insulind e p e n d e n t diabetes mellitus (IDDM). We and others have previously shown that D R 4 represents a public specificity present on multiple different haplotypes. H o m o z y g o u s typing cells, all of which type as DR4, can be divided into at least six different groups on the basis of MLC reactivity [1,2] protein biochemistry [ 3 - 5 ] , and restriction fragment length polymorphisms (RFLP) [6]. In addition, we have described a DQ/3 polymorphism within the DR4-associated D Q w 3 specificity, as defined by serology and RFLP [7]. From Genetic Systems Corporation. Seattle, Washington. *Visiting Scientist, Department of Microbiology, Yonsei Unirersity Collegeof Medicine, Seoul. Korea. Address reprint requests to Barbara Nepom, M.D., Genetic Systems Corp.. 3005 First Avenue. Seattle. WA 98121. ReceivedJanuary 22, 1986; acceptedMarch 25, 1986.
Human Immunology17, 87-93 (1986) © ElsevierSciencePublishingCo., Inc., 1986 52 VanderbiltAve., New York, NY 10017
87
0198-8859/86/$3.50
B.S. Nepom et al.
88
In the present study, we used an EBV-transformed cell line from aJRA patient homozygous for DR4 and DQw3, but heterozygous for Dw4 and Dw14 and for DQ"3.1" and DQ"3.2", to create a haplotype deletion mutant by radiation mutagenesis. Comparison of antibody reactivity and RFLP analysis between the mutant and parental cell lines allows the investigation of polymorphisms present among haplotypes with identical DR and DQ typing but with disparate class II molecules. MATERIALS AND METHODS
Production of deletion mutant. The parental cell line 256 was derived by EBV transformation of peripheral blood lymphocytes from a patient with seropositive JRA, with the genotype A1, B8, DR4, DQw3, (Dw4)/A2, B15, DR4, DQw3, (Dw14) established by family study. 20 x 106 cells were irradiated with 400 rads from a Cesium source. Five days later, 4 x 106 cells remained. After washing with media, half of the remaining cells were treated with antibody P8.1 at 1:10,000 for 1 hr at room temperature, followed by treatment with rabbit complement (Pel-Freeze, Rogers, AR) for 1 hr, also at room temperature. After washing, the cells were resuspended in Iscove's media with 15% fetal calf serum (FCS) and cloned by limiting dilution in round bottom 96-well plates. One hundred thousand irradiated random donor human PBLs/well were used as feeder cells. Samples of wells showing growth were screened by immunofluorescence for expression of various HLA specificities.
Antibodies. Monoclonal antibodies P8.1 (anti-HLA-B8) [8], P4.1 (monomorphic anti-DR) [3], 17.15 (anti-DQw3) [9], and A-10-83 (anti-TA10) [7,10] have been previously described. Antibody 80.1 is an IgM monoclonal which recognizes some B locus alleles, including B15 (Bw62) but not B8 (K. Nelson, personal communication).
Immunofluorescence. After washing, pelleted cells were resuspended in 15 A of antibody in ascites fluid diluted at 1:160, kept on ice for 1 hr, washed, and resuspended in 15 A of affinity purified rabbit anti-mouse IgG conjugated to FITC for 60 min on ice. After washing twice with cold media, cells were fixed in 2% paraformaldehyde and analyzed on a FACS IV (Becton-Dickinson, Oxnard, CA). Of over 200 wells tested, a single well, B6G, was nonreactive with antibody P8.1, so was expanded for further evaluation.
Analysis ofgenomic DNA. RFLP analysis was performed as previously described [6]. Briefly, DNA was purified from EBV-transformed cell lines, digested with Barn HI or Eco RV endonuclease (BRL, Bethesda, MD) for 18 hr at 37°C, then precipitated with ethanol and resuspended in 10 mM Tris, 1 mM EDTA, pH 7.6. Digested DNA was then loaded at 12 /~g/lane on a 0.7% agarose gel and electrophoresed at 30 V for 18 hr. Gels were then denatured and neutralized prior to transfer to nitrocellulose (Schleicher and Schuell) by the method of Southern [11]. After transfer, filters were baked for 18 hr at 80°C, prehybridized at 42°C, and hybridized with nick translated 32P-labeled cDNA probes as previously described [6]. Following 48 hr hybridization at 42°C, blots were washed at 60°C in 1.5 mM NaCl, 0.015 mM Na citrate, and 0.1% SDS three times for 20 min each. Filters were exposed to Kodak XAR-5 film for 2-4 days. The DQ~ probe [12] is a full-length cDNA provided by Dr. P. Peterson. The class I B region probe, P158, is a 2 kb insert from the 3' end of a B7 cDNA inserted in pBR322, provided by Dr. P. A. Biro [13].
Deletion Mutant Defines DQ# Variants
89
Reoclivdy moAb Specificity Parent Mutant
None
--
--
--
P8 I
B8
+
--
815
+
+
80
I
FIGURE 1 FACS analysis of class I expression in deletion mutant B6G. Fluorescence histograms are shown for 4 × 101 ceils of parental cell line 256 (dashed line) and mutant cell line B6G (solid line); cell number is displayed on the vertical axis and relative fluorescence intensity using a log scale amplifier on the horizontal axis. Antibodies P8.1 (anti-B8) and P80.1 (anti-B 15) are described in the text.
RESULTS Analysis o f Cell S u r f a c e E x p r e s s i o n o f Class I and Class I I M o l e c u l e s on M u t a n t and Parental Lines Expression of H L A - B antigens on the parent and mutant cell lines was assessed by immunofluorescence and FACS analysis. Figure 1 shows the results using two different polymorphic class I antibodies: P8.1, which is specific for HLA-B8, and 80.1, which recognizes H L A - B 15 but not B8. Both cell lines are positive for the expression of B15, while B8 is clearly present on the parental cell line 256 but is completely negative on the mutant B6G, indicating that expression of this class I product has been lost on the mutant line. An M o A b recognizing the Al specificity, which is present on the B8 haplotype, was initially negative on the mutant while positive on the parent, but after several weeks in tissue culture the mutant became positive as well, suggesting that an early expression defect was later corrected (data not shown). After screening against a panel of informative class II MoAbs, as shown in Figure 2, results demonstrate that D R molecules, as recognized by the monomorphic D R antibody P4.1, are present on both the mutant and the parental cells. Because this antibody gives a broad flat peak on FACS analysis, quantitative differences in expression of D R molecules are not apparent from these curves. Similarly, the D Q w 3 specificity, as assessed by reactivity with M o A b 17.15, was strongly postive on both cell lines. Interestingly, however, A-10-83, a "TA10" reactive M o A b which splits the D Q w 3 specificity into two subgroups, was positive on the parental line 256 but completely negative on mutant B6G. Microcytotoxicity studies with this antibody confirm these results, with A-10-83 strongly killing cell line 256 while being completely negative on mutant B6G. Cell line B 6 G therefore appears to have lost expression of some class I and class II H L A specificities associated with the H L A - B and H L A - D Q regions. Reochv,ty moAb Specificity
~ ~
,
~
~
Parent Mutant
None
--
P4 ]
DR
+
+
1715
DOw3
+
+
+
-
A-10-83 DO"31"
FIGURE 2 FACS analysis of class II expression in deletion mutant B6G. The parental line 256 is represented by dashed lines, mutant B6G by solid lines. Antibodies P4.1 (anti-DR), 17.15 (anti-DQw3), and A-10-83 (anti-DQ"3.1") are described in the text.
90
B.S. Nepom et al.
R F L P Analysis o f Class I and Class II H L A Genes in M u t a n t B 6 G In order to confirm that the lack of expression described above was due to genetic deletion we performed Southern blot analysis of the two cell lines. Panel A in Figure 3 shows the restriction fragment length polymorphism (RFLP) pattern of the parental line in the left-hand lane compared with the mutant line in the right lane. Cell lines were digested with Eco RV and probed with a class I B region probe, P158. Similar bands are seen in both cases, with the exception of the 8.3 kb band seen hybridizing strongly in the parent line and lost in the mutant line. This 8.3 kb band is characteristic of the HLA-B8 allele in B8 homozygous cell lines, and thus indicates deletion of this allele in the mutant B6G. In order to investigate the genomic basis for the class II mutations suggested by loss of T A 1 0 reactivity, we performed RFLP analysis on the D Q region. We have previously described two alleles present at DQ/3 which represent variation within the DQw3 specificity, which we call DQ"3.1" and DQ"3.2" [7]. Southern blots distinguish between these alleles using a number of different restriction enzymes ([7], and unpublished). When genomic D N A is digested with Barn HI, the D Q " 3 . 1 " variant has a pair of bands at 6.9 and 3.7 kb, while the DQ"3.2" variant has instead a 12.0 kb band. Both alleles also have a constant 10.2 kb band on RFLP analysis representing another D Q locus, often called "DX/3" [14]. As seen in Panel B of Figure 3, the parental line 256, shown on the left, is heterozygous for both the D Q " 3 . 1 " and DQ"3.2" alleles. However, the mutant B6G, shown on the right, retains the 12.0 kb band characteristic of the DQ"3.2" allele, but has lost the 6.9 and 3.7 kb pair of bands, demonstrating that the DQ"3.2" allele is retained while the DQ"3.1" allele has been lost.
P kb -
~
~
~
,
30
B
kb -8.3
~ 10. 2
-3.7
:
1.85
FIGURE 3 RFLP analysis of class I and class II regions of deletion mutant B6G. Panel A shows a Southern blot of DNA from the parental and mutant cell lines digested with Eco RV and probed with a B region probe, p158. The 8.3 kb band, characteristic of the HLA-B8 specificity, is absent in mutant B6G. Panel B shows DNA from both cell lines digested with Barn HI and probed with a DQ/3 cDNA probe. The 6.9 and 3.7 kb bands defining the DQ"3.1" allele have been lost in the mutant, while it retains the 12.0 kb band characteristic of the alternative DQ"3.2" allele.
Deletion Mutant Defines DQ/3 Variants
91
Through the combination of both FACS analysis and RFLP analysis of this mutant cell line, we can describe the genotype of the mutant as shown schematically in Figure 4. Although the break points are not known exactly, we postulate that at least the DQ"3.1" through the B8 genes have been deleted. DISCUSSION HLA deletion mutant cell lines have proven extremely useful in the genetic analysis of the HLA system. Heterozygosity, even among some presumed "homozygous typing cells," has made the study of individual HLA loci and alleles somewhat confusing. Pious et al. [15], Kavathas et al. [16,17], and others have developed the techniques of radiation and chemical mutagenesis to produce mutant cell lines with hemizygous deletions, point mutations, and expression defects in the HLA region, allowing precise genetic mapping as well as evaluation of regulation and expression. We have been interested in utilizing these techniques to analyze HLA-DR4related haplotypes, which are of particular interest because of their association with autoimmune disorders such as rheumatoid arthritis, seropositive JRA, and IDDM. We and others have previously shown that DR4 is a public specificity present on multiple different haplotypes, which can be distinguished both phenotypically by MLC reactivity [1,2] and biochemically by analysis of DR surface molecules [3-5]. Like HLA-DR4, the DQw3 specificity has also been shown to be a public specificity present on different haplotypes. Giles et al., for example, have demonstrated that the molecules bearing the MB3 specificity (now known as DQw3) differ structurally between DR4- and DR5-related haplotypes [18]. Similarly, Ishikawa et al. have shown structural polymorphism among MB3 light chains from MB3-positive cells differing at DR (DR4,5,6, and 9) as well as within a single DR type (DR4,Dw4 and DR4,LD"KT2") [19]. In addition, we have recently described two allelic variants of DQ/3 genes corresponding to the DR4associated DQw3 specificity [7]. In the first variant, which we call DQ"3.1", 6.9 and 3.7 kb polymorphic bands are seen on RFLP analysis when cells are digested with Barn HI and probed with DQ/3. In the second variant, DQ"3.2", a 12.0 kb DQ/3 band is present instead. All cells showing the DQ"3.1" allele are recognized by TA10-reactive MoAb A-10-83 [7]. In the present study, we have created a haplotype deletion mutant by radiation mutagenesis from a cell homozygous for DR4 and DQw3, but heterozygous for Dw4 and Dw14 as well as for DQ"3.1" and DQ"3.2", in order to study DR4associated polymorphisms at the genomic level. We have shown, by surface
FIGURE 4 Schematicdiagram of the presumed haplotypes of the parental and mutant cell lines. The break points of the mutant's deletion are not precisely mapped, but the deletion is presumed to extend from at least the DQ through the B regions. DQ Parental 256
DR
B
A
--"51"--4--8--
1
--"5.2"--4--15--2
Mutant B6G
--~ . . . . . . . . . . . . . . . . . . . . --~'5.2"--
4 - - 15 - -
(Ow4) (Dw14)
~,
1-2- -
(Dw14)
92
B.S. Nepom et al. binding studies, that the specificities B8 and DQ"3.1" present on one haplotype are no longer expressed in the mutant line. This lack of expression is due to genetic deletion, as shown by loss of polymorphic bands in RFLP patterns of both the B8 and D Q " 3 . 1 " alleles on the mutant cell line. Although we do not yet have polymorphic genomic markers informative for the H L A - D R region on that same haplotype, the simplest explanation for the data is that this is a single deletion extending from at least the H L A - D Q through the HLA-B regions. The loss of the polymorphic 6.9 and 3.7 kb bands in the mutant B 6 G as demonstrated in the Southern blot in Figure 3B, in concert with the loss of reactivity to MoAb A-10-83, effectively maps the TA10 specificity to the DQw3associated haplotype which we call 3.1, and directly demonstrates the allelic nature of the D Q " 3 . 1 " and DQ"3.2" variants. Restriction fragment length polymorphisms (RFLP) have previously been used to discriminate among different HLA specificities [20], and sometimes among multiple haplotypes within a single serologically defined specificity [6,21]. However, these genomic polymorphisms frequently do not correlate with serologic reactivity. The description of RFLP patterns identifying the DQ"3.1" and DQ"3.2" alleles and the precise correlation of the DQ"3.1" RFLP with reactivity to MoAb A-10-83 marks one of the first examples of a genomic polymorphism within a single HLA specificity encoding an expressed, serologically defined product. The ability to distinguish among different DR4- and DQw3-related haplotypes has been of considerable interest in recent years. Analysis of variations identified by MLC [1,2], 2D gel electrophoresis [ 3 - 5 ] , serology [7], and RFLP [6,7] have clearly shown that different DRI3 and DQ/3 alleles consitute at least six different DR4 positive haplotypes. One of the questions of considerable functional importance is whether these specific haplotypes can confer susceptibility to different DR4-associated diseases. For example, we have shown that the DR4 haplotype associated with I D D M represents a different haplotype from that associated with seropositive JRA [22]; others have come to a similar conclusion for I D D M and adult rheumatoid arthritis [23]. Additional structural and functional analyses such as those described here using deletion mutants will elucidate the various DR4associated class II gene products contributing to such haplotypic variation.
ACKNOWLEDGMENTS
We thank Drs. K. Nelson, H. Maeda, P. Antonelli, and P. Martin for providing MoAbs used in these studies. We would also like to thank Margery Moogk for technical assistance and Holly Chase for manuscript preparation.
REFERENCES 1. Reinsmoen NL, Bach FH: Five HLA-D clusters associated with HLA-DR4. Hum Immunol 4:249, 1982. 2. Nose Y, Sato K, Nakagawa N, Kondoh K, Inouye H, Tsuji K: HLA-D clusters associated with DR4 in the Japanese population. Hum Immunol 5:199, 1982. 3. Nepom BS, Nepom GT, Mickelson E, Antonelli P, Hansen JA: Electrophoretic analysis of human HLA-DR antigens from HLA-DR4 homozygous cell lines: correlation between/3-chain diversity and HLA-D. Proc Natl Acad Sci USA 80:6962, 1983. 4. Nepom GT, Nepom BS, Antonelli P, Mickelson E, Silver J, Goyert SA, HansenJA: The HLA-DR4 family of haplotypes consists of a series of distinct DR and DS molecules. J Exp Med 159:394, 1984.
Deletion Mutant Defines DQ/3 Variants
93
5. Groner JP, Watson AJ, Bach FH: Dw/LD-related molecular polymorphism of DR4 /3-chains. J Exp Med 157:1687, 1983. 6. Holbeck SL, Kim SJ, Silver J, Hansen JA, Nepom GT: HLA-DR4-associated haplotypes are genotypically diverse within HLA. J Immunol 136:637, 1985. 7. Kim SJ, Holbeck SL, Nisperos B, Hansen JA, Maeda H, Nepom GT: Identification of a polymorphic variant associated with HLA-DQw3 characterized by specific restriction sites within DQB. Proc Natl Acad Sci USA 82:8139, 1985. 8. Antonelli P, Nisperos B, Braun M, Hansen JA: Recognition by a murine monoclonal antibody of a unique epitope specific for the human alloantigen HLA-B8. Hum Immunol 11:11, 1984. 9. Hansen JA, Martin P, Kamoun M, Nisperos B, Thomas ED: A supertypic HLA-DR specificity (DR4 + 5) defined by a murine monoclonal antibody. Hum Immunol 2:103, 1981. 10. Maeda H: Mouse monoclonal antibody detects a new polymorphic Ia determinant other than HLA-DR antigen: a possible allele of DC-1. Tissue Antigens 23:163, 1984. 11. Southern EM: Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98:503, 1975. 12. Larhammar D, Schenning L, Gustafson K, Wiman K, Cleasson L, Rask L, Peterson PA: Complete amino acid sequence of an HLA-DR antigen-like/3 chain as predicted from the nucleotide sequence: similarities with immunoglobulins and HLA-A, -B, and -C antigens. Proc Natl Acad Sci USA 79:3687, 1982. 13. Biro PA, Pan J, Sood AK, Kole R, Reddy VB, Weissman SM: The major histocompatibility complex. Cold Spring Harbor Quant Biol 1982 XLVII:1082, 1983. 14. Holbeck S, Nepom, GT: Analysis of HLA DQ/3 allelic variation using synthetic oligonucleotide probes, submitted for publication. 15. Pious D, Soderland C, Gladstone P: Induction of HLA mutations by chemical mutagens in human lymphoid cells. Immunogenetics 4:437, 1977. 16. Kavathas P, Bach FH, DeMars R: Gamma ray-induced loss of expression of HLA and glyoxalase I alleles in lymphoblastoid cells. Proc Natl Acad Sci USA 77:4251, 1980. 17. Kavathas P, DeMars R, Bach FH: Hemizygous HLA mutants oflymphoblastoid cells: a new cell type for histocompatibility testing. Hum Immunol 1:317, 1980. 18. Giles RC, Hurley CK, Capra JD: Primary structural variation among serologically indistinguishable DS antigens: the MB3-bearing molecule in DR4 cells differs from the MB3-bearing molecule in DR5 cells. J Immunol 133:1, 1984. 19. Ishikawa N, Kasahara M, Ikeda H, Ogasawara K, Hankin S, Takenouchi T, Wakasaka A, Kikuchi Y, Aizawa M: Demonstration of structural polymorphism among MB3 light chains by two-dimensional gel electrophoresis. J Immunol 134:417, 1985. 20. Dausset J, Cohen D: HLA at the gene level. In: ED Albert, MP Baur, WR Mayr, Eds. Histocompatibility testing 1984. Berlin, Springer-Verlag, 1984, p. 22. 21. Bosch ML, Schreuder GMT, Spits H, Termijtelen A, Tilanus MGJ, Giphart MJ: Polymorphism within the HLA-DRw6 haplotype I. Restriction fragment length variation and its correlation with serology. J Immunol 134:3212, 1985. 22. Nepom BS, Palmer J, Kim SJ, Hansen JA, Holbeck SL, Nepom GT: Specific genomic markers for the HLA-DQ subregion discriminate between DR4 positive IDDM and DR4 positive JRA. J Exp Med, in press. 23. Festenstein H, Awad J, Hitman GA, Cutbush S, Groves AV, Cassell P, Oilier W, Sachs JA: New HLA-DQ RFLP associations with rheumatoid arthritis and insulin dependent diabetes mellitus, submitted for publication.