0 1991 S. Karger AG. Basel

Vox Sang 1991;61:106-110

oo42-9oo7/9~o612-nio~$ 2 . 7 ~ 1

Quantitation of Soluble HLA Class I1 Molecules by an Enzyme-Linked Immunosorbent Assay Ulrike Westhofa, Friedrich I? Thinnes", Hilde Gotz b, Hans Grosse-Wildea "Department of Immunogenetics, University Hospital of Essen, Medical School, FRG; bDepartment of Immunochemistry, Max-Planck-Institute of Experimental Medicine, G8ttingen. FRG

Abstract. In order to quantify soluble HLA-DR,DQ,DP molecules (sHLA-RQP) an enzyme-linked immunosorbent assay was developed utilizing two monoclonal antibodies specific for HLA-DR,DP (Tii35) and HLA-DQ (Tii22) gene products, respectively. Highly purified HLA class I1 molecules isolated from a lymphoblastoid cell line were used for calculation of exact sHLA-RQP protein values. Circadian variations of sHLA-RQP plasma levels were studied in 7 healthy probands showing no significant deviations; measurements in 4 probands at intervals between 4 and 6 weeks revealed that sHLA-RQP levels remain relatively stable. The population analysis of 209 unrelated, HLA-typed healthy donors resulted in an average protein concentration of 1.53k 2.44 pglml plasma for sHLA-RQP. Four out of 209 probands (= 1.9%) had no detectable sHLA-RQP. Significant associations of high or low sHLA-RQP levels to particular HLA-DR or -DQ specificities were not observed. However, plasma derived from HLA-DR9 positive had the highest and from HLA-DR8 positive donors the lowest mean sHLA-RQP values. By comparing HLA identical with two-haplotypedifferent siblings we found no evidence that sHLA-RQP plasma levels are under genetic control of the HLA complex or closely linked genes. Furthermore, soluble HLA class I plasma concentrations in 100 probands analyzed showed no correlation to those of sHLA-RQP.


Human MHC antigens of class I1 type are heterodimeric glycoproteins consisting of an a-chain with approximately 36 kilodaltons and a noncovalently bound P-chain with approximately 29 kilodaltons molecular weights [l, 21. More than 15genes are now identified belonging to the human MHC class11 region (HLA-D region), some of which are not expressed (pseudogenes). Well characterized are three series of multiallelic gene products, the HLADR, HLA-DQ, and HLA-DP specificities [3,4]. These HLA class I1 antigens are typically found on membranes of B-lymphocytes or monocytes and are the physiological ligands or receptors for the CD4 antigens present e. g. on T helper cells [5]. As already shown more than 20years ago for HLA class I [6,7], the gene products of the class I1 region exist not only in a membrane-anchored form but also as soluble

glycoproteins. The first descriptions of soluble MHC class I1 antigens were given by Callahan et al. [8] and Parish et al. [9] studying sera of different congenic mouse strains. In man, Wilson et al. [lo] were 1979the first to detect serum soluble HLA class I1 antigens. Thereafter Russo et al. [ll], Herlyn et al. [12], Stevenson et al. [13], and Thompson et al. [14] tried to quantify soluble class I1 antigens. These studies, however, revealed that if at all only few sera of healthy donors contained measurable amounts of class I1 molecules. Having recently established an enzyme-linked immunosorbent assay (ELISA) for soluble HLA class I molecules (sHLA-ABC) [15,22] we proceeded to develop a sensitive ELISA for soluble class I1 molecules (sHLA-RQP). In the following we report on the quantitative analysis of sHLARQP plasma levels in unrelated HLA-typed healthy individuals and members of families and on the comparison of sHLA-ABC with sHLA-RQP levels.


Quantitation of Soluble HLA Class I1 (sHLA-RQP)

Materials and Methods Probands Heparinized venous blood from 209 unrelated healthy individuals and from members (n = 59) of 12 families were typed for HLA-A,B,C and HLA-DR,DQ antigens according to the NIH guidelines [16]. Plasma Samples EDTA containing (1 mg/ml) blood samples were centrifuged at 1,400g for 10 min. The supernatant plasma was collected and tested either directly or stored at -80°C not longer than 2 years until use. Just before assay, the plasma was centrifuged at 14,OOOg for 2 min to avoid platelet contamination. ELISA Determination U-shaped Maxisorp ImmunoplatesTM (Nunc, Wiesbaden, cat. No. 442404) were coated overnight at 4°C with 100 yl of an 1:l (v/v) mixture of the HLA-DR, DP-specific monoclonal antibody Tii35 [17] and the HLA-DQ-specific monoclonal antibody Tii22 [HI, kindly provided by Dr. A. Ziegler, Berlin, priory diluted 1:500 with phosphatebuffered saline (PBS). The plates were washed 3 times with PBS +0.02% Tween 20 and blocked by 300 yl PBS + 1% bovine serum albumine (Merck, Darmstadt, cat. No. 12018) per well for 30 min at room temperature (RT). After 3 further washings 100 yI of undiluted EDTA-plasma were incubated for 60 min at RT, followed by 5 washings with PBS +0.02% Tween 20. The bound HLA class I1 molecules were detected by polyclonal rabbit antisera specific for HLA class I1 (D 77533 and E 77534), kindly supplied by Dr. J. J. Neefjes and Dr. H. L. Ploegh, Amsterdam. Prior to use both antisera were diluted 1:1,000 with PBS, mixed 1:l (v/v), and 100yI of this mixture were incubated for 60 min at RT. After 3 washing steps the plates were treated with 100 yVwell of alkaline phosphatase-conjugated goat antirabbit-Ig (Dianova, Hamburg, cat. No. 111-055-045)for 60min at RT, prior diluted 1:1,000 with PBS. The plates were washed 10 times and, after addition of the substrate (Sigma Chemie, Deisenhofen, cat. No. 104-0),developed for 60 rnin in the dark. The optical density (OD) was measured in a Microplate AutoreaderTM(Biotech Instruments, Vermont, Nebr.) at 405 nm wavelength. In order to convert the OD values into sHLA-RQP protein concentrations, highly purified HLA class I1 molecules were tested. For that membrane-bound class I1 antigens were solubilized and enriched from an HLA homozygous lymphoblastoid cell line as described previously [19]. The class I1 molecules were purified further by gel filtration, using an HiLoad'r" 16/60 column (Pharmacia LKB, Freiburg, cat. No. 171068-01).The protein concentrations of the relevant fractions obtained were determined by the method of Lowry et al. [20]. Figure 1 shows the geometric dilution of a probe containing 17.0yg/ml of sHLA-RQP demonstrating a range of sensitivity between 0.5 and 9.0 Nml. Samples with concentrations higher than 9 yg/rnl sHLA-RQP were diluted 1:2 with PBS to reach the sensitivity range of the ELISA established. In all tests the same plasma sample (sHLA-RQP level 2.6 yg/ml) was taken as reference allowing the conversion of individual O D units into sHLARQP protein concentrations. Optimal experimental conditions were defined by analyzing various antibody dilutions and incubation times. The average of background values was 0.234 f 0.045 O D and the specific OD values were corrected for the actual background. The intraassay (triplicates) and interassay variations were 4.2 and 9.5% respectively.



Fig. 1. Standard curve for the quantitation of sHLA-RQP, demonstrating a range of sensitivity between 0.5 and 9.0 yg/ml. Fig. 2. Plasma concentrations of 7 probands determined at 3-hour intervals beginning at 8.00a. m.

Statistical Analysis The significance of associations between sHLA-RQP plasma concentrations and individual HLA antigens was calculated by the general linear model test (SAS Institute Cary, N.C.).


In a first series we studied the circadian variations of sHLA-RQP levels in 7 healthy probands. Figure 2 depicts the values obtained starting at 8.00 a.m. followed by 4 con-



Fig. 3. Distribution of soluble class I1 plasma concentrations (pg/ml) among 209 unrelated healthy individuals.

Table 1. Mean values (fSD) of sHLA-RQP plasma levels in rela. tion to HLA-DR and HLA-DQ types of the donors HLA


Plasma levels Pml


40 50 44 45 61 39 50 12 5 2

1.59k2.01 1.81k 1.89 1.45k2.03 1.57+ 1.43 1.76k2.15 1.62k0.45 1.44k1.34 0.9620.54 2.79k2.74 2.24k2.71

DRw52 DRw53

156 100

1.7022.13 1.57k1.48

DQwl DQw2 DQw3

111 80 95

1.67k1.89 1.41k1.62 1.59k1.64

secutive measurements at 3-hour intervals. There was no evidence for a circadian rhythm of sHLA-R . The standard deviation of the median was 9.9% only. addition plasma samples from 4 of the 7 probands were easured at intervals of 4-6 weeks [data not shown]; there was no significant fluctuation of sHLA-RQP levels. Pilot experiments showed that in principle also serum can be assayed instead of EDTA plasma. However, the comparison between plasma and serum revealed an approximate 30% reduction of sHLA-RQP values in the latter.


Thereafter we determined the individual sHLA-RQP plasma levels in 209 unrelated probands. Their distribution is summarized in figure 3. The mean value of sHLA-RQP was 1.53 k 2.43 yglml. In 4 probands (1.9y0) we could not detect measurable amounts of sHLA-RQP, whereas 10 out of 209 individuals (4.8%) had values 24.5 yglml. The highest sHLA-RQP level was recorded in 1individual as 11.0 yglml. Since all probands were typed for HLA we analyzed the influence of particular HLA-DR or -DQ antigens present in the individuals tested with regard to mean sHLA-RQP levels as shown in table 1. The lowest mean value was found in HLA-DR8-positive (n = 12) and the highest mean value for sHLA-RQP in HLA-DR9 (n = 5) individuals. These differences, however, did not reach statistical significance. The phenotypes of the 4 sHLA-RQP ‘deficient’ healthy donors were HLA-DR~,~;DQw~,-/HLA-DR~,~;DQ w~/HLA-DR~,-;DQw~,-/HLA-DRw~,~;DQw~,-. Because of the high standard deviations of the mean values for each HLA-DR specificity we asked whether particular HLA-DR phenotypes had lower standard deviations. This, however, is ruled out by the data given in table 2. As soluble HLA class11 molecules are encoded by HLA-DR,DQ,DP their plasma levels my be regulated by these linked genes. We therefore analyzed 12 families comprising 59 healthy members. Among these, 18 pairs represented HLA genotypically identical and 20 pairs two-haplotype-different siblings. The distribution of the sHLARQP levels revealed that a selection for MHC identity did not result in a significant correlation of sHLA-RQP values between these pairs. From a previous study the sHLA-ABC levels in plasma samples of 100 out of the 209 probands were known [15].

Quantitation of Soluble HLA Class I1 (sHLA-RQP)


Table 2. Mean values ( f SD) of sHLA-RQP plasma levels in relation to different HLA-DR phenotypes



Plasma levels Pdml

DR 1.2 DR 1.5 DR 1,7 DR 2.4 DR 2,5 DR 3,w6 DR4J DR 5,w6 DR5,7

10 10 10 14 14 11 10 10 12

1.57f 1.03 1.98f3.58 1.02f0.82 1.5251.09 3.42f3.29 2.49f3.10 1.31f0.96 0.97 f0.80 2.68f2.68

Fig.4. Comparison of the sHLA-ABC and sHLA-RQP plasma levels in 100 healthy individuals.

Figure 4 compares the sHLA-ABC and the sHLA-RQP values of these probands showing no correlation at all between both plasma markers.

Discussion To our knowledge this is the first comprehensive report on the successful quantitation of soluble HLA class I1 molecules in the majority (98%) of healthy individuals. Previously published attempts ranged from nondetectable amounts to a maximum of approximately 27% of healthy probands having detectable sHLA-RQP in their plasma

[lo-141. In patients suffering from hematological diseases like leukemia or multiple myeloma Herlyn et al. [12] and Wilson et al. [lo] found increased levels of soluble class I1 antigens. Increased sHLA-RQP levels were reported by Thompson et al. [14] in patients suffering from graft-versushost disease after allogeneic marrow transplantation. The ELISA described is simple and well reproducible (intra- and interassay variations are 4.2 and 9.5%, respectively) and possesses a sensitivity between 0.5 and 9.0 pg sHLA-RQP/ml (fig. 1). As already shown for sHLA-ABC plasma levels [15], there was no significant circadian variation or fluctuation at 4- to 6-week intervals, respectively. In contrast to the well-known association between HLA-A9 and significant higher sHLA-ABC levels, we could not find any particular HLA-DR or HLA-DQ specificity correlated with significantly elevated or decreased sHLA-RQP plasma concentrations. This situation may change if immunoassays are developed quantitating separately the three distinct MHC class I1 gene products utilizing the respective specific monoclonal antibodies of if the probands analyzed are also typed for their HLA-DP specificities. At least at the membrane level HLA-DQ and HLADP gene products appear to be expressed at lower quantity than HLA-DR [21,22]. As already shown for sHLA-ABC [W], also soluble MHC class I1 plasma levels are not exclusively governed by the MHC, since HLA-identical and HLA-different offspring showed similar interindividual variations of sHLARQP. The same was true when sHLA-ABC versus sHLARQP levels were compared in a total of 100 healthy probands. If sHLA levels are at all under a genetic control, other gene(s) must be considered to be operative. Candidates are the structural genes for a- and/or y-interferon known to modulate the membrane expression of class I and class I1 molecules, which are localized at chromosome 9 and 12, respecively [23]. But not only production or secretion of sHLA can govern the individual plasma levels. Equally or even more important could be a variation in the metabolism or degradation of sHLA in various organs, e.g. the kidneys. From pilot studies we know that urine from healthy donors contains only very low or nondetectable amounts of sHLA, which would mean that in men these molecules are retained by the kidneys. Although the exact biological function or significance of soluble HLA molecules in particular of sHLA-RQP are not known, the sensitive ELISA descibed will allow the study of diseases, where activation and/or proliferation of immune cells together with modulation of MHC class I1 metabolism play a central role in the pathogenesis.



References I Benacerraf B: Role of MHC gene products in immune regulation. Science 1981;212:1229-1238. 2 Kaufman JF, Auffray C, Korman AJ, et al: The class I1 molecules of the human and murine major histocompatibility complex. Cell 1984;36:1-13 3 Gorga JC, Horejsi V, Johnson DR, et al: Purification and characterization of class I1 histocompatibility antigens from a homozygous human B cell line. J Biol Chem 1987;262:16087-16094. 4 Shackelford DA, Strominger JL: Analysis of the oligosaccharides on the HLA-DR and DC1 B cell antigens. J Immunol 1983:130: 274-282. 5 Biddison WE, Rao PE, Talle MA, et al: Possible involvement of the OKT4 molecule in T cell recognition of class I1 HLA antigens. Evidence from studies of cytotoxic T lymphocytes specific for SB antigens. J Exp Med 1982;156:1065-1076. 6 Charlton RK, Zmijewski CM: Soluble HL-A7 antigen: Localization in the P-lipoprotein fraction of human serum. Science 1970; 170:636-637. 7 Van Rood JJ, van Leeuwen A, von Santen MCT: Anti-HL-A2 inhibitor in normal serum. Nature 1987:226:366-367. 8 Callahan GN, Ferrone S, Poulik MD, et al: Characterization of Ia antigens in mouse serum. J Immunol 1976;117:1351-1355. 9 Parish CR, Chilcott AB, McKenzie IFC: Low molecular weight Ia antigens in normal mouse serum. I. Detection and production of a xenogeneic antiserum. Immunogenetics 1976;3:113-128. 10 Wilson BS, Indiveri F,Pellegrino MA, et al: Production and characterization of DR xenoantisera: Use for detection of serum DR antigens. J Immunol 1979;122:1967-1971. 11 Russo C, Pellegrino MA, Ferrone S: A double-determinant immunoassay to quantitate human Ia antigens. Transplant Proc 1983: 15:57-59. 12 Herlyn M, Lange B, Bennicelle J, et al: Increased levels of circulating HLA-DR antigens in sera of patients with acute lymphoblastoid leukemia. Leuk Res 1984;8:323-334. 13 Stevenson FK, George AJT, Walters MT, et al: Analysis of soluble HLA class I1 antigenic material in patients with immunological diseases using monoclonal antibodies. J Immunol1986;86:187-190. 14 Thompson S, Wareham M, Pearson ADJ, et al: A preliminary study of serum class I1 levels in healthy individuals and bone marrow transplant patients. Clin Chim Acta 1989;185:45-52.

15 Doxiadis I, Westhoff U, Grosse-Wilde H: Quantification of soluble HLA class I gene products by an enzyme linked immunosorbent assay. Blut 1989;59:449-545. 16 Terasaki PL, McClelland J: Microdroplet assay of human serum cytotoxins. Nature 1964;204:99&1000. 17 Pawelec GP, Shaw S, Ziegler A, et al: Differential inhibition of HLA-D or SB-directed secondary lymphoproliferative responses with monoclonal antibodies detecting human Ia-like determinants. J Immunol 1982:129:1070-1075. 18 Wernet P, Ziegler A, Shaw S, et al: Monoclonal antibodies against Ia-like antigens inhibiting HLA-D and/or SB-directed secondary lymphoproliferative responses. Transplant Proc 1983;15:94-98. 19 Thinnes FP, Egert G, Gotz H, et al: Primlrstruktur der menschlichen Histokompatibilitatsantigene der Klasse I1 (HLA-D). I. Isolierung, Reinigung und Charakterisierung des alp-Ketten-Komplexes einer homozygoten, lymphoblastoiden B-Zell-Linie, H2LCL (HLA-A3,3; B7,7; D w 2 , ~ 2 ;DR2,2; MT1,l; DC1,l; MB1 ,l.) Hoppe Seylers Z Physiol Chem 1984;365:1277-1289. 20 Lowry OH, Rosenbrough NJ, F a n AL, et al: Protein measurement with the folin phenol reagent. J Biol Chem 1951;193:265-275. 21 Lee JS: Regulation of HLA class I1 gene expression; in Dupont B (ed.): Immunobiology of HLA. New York, Springer, 1989, vol2, pp 49-62. 22 Westhoff U, Doxiadis I, Beelen UW, et al: Soluble class I concentrations and GvHD after allogeneic marrow transplantation. Transplantation 1989;48:891-893. 23 Trent JM,Olson S, Lawn RM: Chrosmosomal localization of human leukocyte, fibroblast, and immune interferon genes by means of in situ hybridization. Proc Natl Acad Sci USA 1982; 79:7809-7813.

Received: November 6, 1990 Revised manuscript received: February 5,1991 Accepted: February 5,1991 Dr. Hans Grosse-Wilde Institut f i r Immungenetik Universitatsklinikum Essen Virchowstrasse 171 D-W-4300 Essen 1 (FRG)

Quantitation of soluble HLA class II molecules by an enzyme-linked immunosorbent assay.

In order to quantify soluble HLA-DR,DQ,DP molecules (sHLA-RQP) an enzyme-linked immunosorbent assay was developed utilizing two monoclonal antibodies ...
895KB Sizes 0 Downloads 0 Views