Determination of the frequency of HLA antibody secreting B-lymphocytes in alloantigen sensitized individuals

Abstract

Sera from prospective transplant patients are usually screened for HLA antibodies prior to transplantation, but presently available tests do not permit quantification of the humoral alloantigen directed response. We adapted a culture system for isolated human B-lymphocytes to assay the secretion of HLA-antibodies on a single cell basis. B-cell supernatants were screened for HLA antibodies by complement dependent cytotoxicity. The assay assigns precursor frequencies for HLA-alloantibody secreting B-lymphocytes (BCPFs), and simultaneously allows for dissection of the humoral alloantigen directed response into its monoclonal components. The lymphocytes of 15 HLA-seropositive multiparous women that were used to validate the assay, were found to contain HLA-BCPFs ranging from 0 to 123 per 10⁶ B-lymphocytes (mean: 43 ± 45 per 10⁶ B-lymphocytes). The HLA-specificities of antibodies in the B-cell supernatants were in agreement with serum specificities. Genuine HLA reactivity of B-cell supernatants was confirmed using an ELISA with purified HLA class I antigens. When applied to lymphocytes of patients on transplant waiting lists, the present assay may enable the unraveling of serum specificities in their components, thus supplementing HLA antibody serum screening data.

INTRODUCTION

Donor specific HLA antibodies are an obstacle for organ transplantation, as they may cause hyperacute rejection [1–4]. Hence, serum HLA antibodies considerably limit the access of alloantigen sensitized patients to the pool of donor organs, thus increasing waiting times [5].

The immune status of sensitized individuals is usually examined by screening sera for the presence of HLA-antibodies in complement dependent cytotoxicity (CDC) [6], ELISA based methods [7,8] and flow cytometry [9]. Yet, the only measure of quantification has classically been the determination of serum antibody titer. Serum HLA antibody is the product of all HLA directed plasma cells combined, but antibody productivity may vary from cell to cell and from one individual to another. We attempted to improve the estimation of the magnitude of the allo-directed humoral response by developing a B-cell precursor frequency assay (BCPF). For the cellular compartment, the advent of cytotoxic T lymphocyte precursor [10] and T-helper cell [11] precursor frequency assays has enabled the quantification of the alloantigen directed T-lymphocyte response in transplant patients as well as in multiparous women [12]. Clinically, the magnitude of these responses has been correlated with graft survival in, e.g. renal [13,14], corneal [15] and cardiac [16] transplantation. Likewise, the availability of a culture system for B-lymphocytes and the analysis of their secreted products at clonal level will permit a more detailed characterization of the humoral compartment of the allo-response. Possibly, determination of the characteristics of HLA antibody specificities secreted by individual B-lymphocytes, and of the proportion of B-lymphocytes involved, will be an auxilliary instrument for the determination of unacceptable mismatched antigens on donor organs for prospective transplant patients with high panel reactive antibody (PRA) values.

Activation of B-lymphocytes in vitro through CD40-engagement has been shown to induce proliferation, differentiation and concomitant secretion of immunoglobulins in various culture systems using anti-CD40 antibodies [17,18], CD40L transfectants [19] and soluble trimeric CD40L [20]. Culture of B-lymphocytes, in the presence of the CD40L expressing mouse thymoma cell line EL4B5 [21], allows testing of supernatants for the presence of specific antibodies. The feasibility of determining specific BCPFs with this system has been shown for several antigenic systems: P. falciparum specific antigens in patients suffering from malaria infections [21] mycobacterial heat shock protein in RA patients [22], rheumatoid factor in RA patients [23], and A and B antigens of the AB0 bloodgroup system [24].

In the present study we examined peripheral B-lymphocytes, derived from alloantigen sensitized individuals, for their ability to produce HLA-antibodies in culture. Culture in limiting dilution format thus enabled the calculation of HLA-specific BCPF values.

MATERIALS AND METHODS

Subjects

The subject population consisted of 15 (multi) parous women with serum HLA (MHC-class I) antibodies, as determined by CDC against panels of 51 HLA-typed cells (Table 1) and 2 healthy non transfused males, without CDC reactive antibodies (Table 5). Correlation coefficients (with Yates correction) for combined serum HLA antibody specificities were determined using GraphPad InStat version 3·00 for Windows 95 (GraphPad Software, San Diego, CA,USA). Informed consent was obtained for blooddonations from both categories of individuals, under guidelines of the local Medical Ethics Committee. To enable person-to-person comparison, the multiparous women were mainly selected for the presence of HLA-A2 antibodies (13/15 women) in their sera. The two remaining women had HLA-B5 + B35 and HLA-A1 + B27 serum antibodies, respectively.

Table 1.

Characteristics of multiparous females and spouses

Donor	Donor HLA typing	Spouse's HLA-typing	Last child's HLA-typing	Serum antibody specificity*	R†	Interval ‡	Grav §
1	A3,A68(28),B51(5),B7	A2,A29(19),B62(15),B35	A2,A68(28),B51(5),B35	A2 + B35 + B62 + B57	0·58	9	2
2	A1,-,B7,B8	unknown	unknown	A2 + A28	0·96	2	1
3	A2,A24(9),B7	A2,-,B51(5),B7	A2,-,B7,B51(5)	B5 + B35 + B53 + B70 + B62 + B57	0·53	2	2
4	A1,A3,B7,B8	unknown	A2,A3,B7,B13	A2 + A28 + B17	0·89	3	1
5	A3,-,B51(5),B35	A2,-,B7,B62(15)	A2,A3,B51(5),B62(15)	A2 + A28 + B62 + B17	0·86	7	2
6	A3,-,B7,B35	A1,-,B27,B35	A1,A3,B7,B27	A1 + B27 + B22 + B44 + B49	0·32	9	unknown
7	A11,A29(19),B44(12),B56(22)	A2,-,B18,B60(40)	A2,A29,B44(12),B60(40)	A2 + A28 + B40	0·89	2	1
8	A23(9),A11,B44(12),B55(22)	unknown	unknown	A2 + A28	0·85	3	unknown
9	A3,A11,B14,B55(22)	unknown	A2,A11,B51(5),B55(22)	A2 + A28 + B5 + B35 + B53 + B49	0·63	2	1
10	A1,A3,B7,B62(15)	A2,A3,B7,-	A1,A2,B7,B62(15)	A2	0·96	6	3
11	A1,A25(10),B8,B18	A2,-,B44(12),B60(40)	A1,A2,B8,B44(12)¶	A2 + A28 + B12 + B40 + B17	0·88	14	3
11	A1,A25(10),B8,B18	A2,-,B44(12),B60(40)	A1,A2,B8,B44(12)¶	A2 + A28 + B12 + B40	0·88	20	3
12	A1,A32(19),B35,B38(16)	unknown	A1,2,B44(12),B35	A2 + A28	0·92	5	2/3
13	A1,A3,B8,B35	A1,A2,B7,-	A2,A3,B7,B35	A2	0·84	62	2
14	A3,A34(10),B7,B62(15)	A2,-,B8,B60(40)	A2,A3,B7,B8	A2 + B8	0·89	2	1
15	A3,-,B7,-	unknown	A2,A3,B7,-	A2	0·56	2	2/3

^*antibodies in the serum of the same date of cell collection for BCPF expts, determined by CDC (except for donors 5:serum data 4 months after cell collection, and 10, 3 months after cell collection).

^†Coefficient of correlation (with Yates correction for small numbers).

^‡interval, in months elapsed between last pregnancy and PBL collection.

^§number of gravidities prior to PBL collection (2/3 denotes delivery of twins).

^¶typing of 2nd child given; 3rd child not available for HLA-typing.

Table 5.

B-cell culture of non-transfused males

Donor	Donor HLA typing	Reactivity‡	Frequency (× 106)	GOF†
16	A2,A32,B35,B60	pol	8	4
		pan	23	30*
17	A1,A3,B7,B8	pol	13	23*
		pan	30	41*

^†GOF (goodness-of-fit); if value is followed by

^*the frequency value is statistically not significant.

^‡reactivities: pol, polymorphic; pan, panreactive.

Cells

Mononuclear cells were isolated from heparinized blood by Ficoll-Isopaque sedimentation and cryopreserved until use. All subjects, and where informative, their spouses and children were serologically HLA-typed. Additionally, cryopreserved mononuclear cell suspensions of HLA-typed individuals were used as panel cells for screening B-lymphocyte supernatants for HLA antibodies by CDC. B-lymphocytes were isolated with anti-CD19 DynaBeads (Dynal, Oslo, Norway) and released with the appropriate Detach-A-Bead (Dynal) solution according to the manufacturer's instructions. The purity of CD19⁺ enriched B-lymphocytes of 2 individuals was determined by flow cytometry with FITC-and PE-labelled mouse Mabs for CD3, CD19, and CD20 (Becton and Dickinson Immunocytometry Systems, San Jose, CA, USA). Isolated B-lymphocyte fractions contained 94% CD19⁺ CD20⁺ B-lymphocytes and 2% CD3⁺ T-lymphocytes for one and 97% B-lymphocytes en 1% T-lymphocytes for the other individual.

Cell culture

All cultures were done in Iscove's Modified Dulbecco's medium (Gibco/Life Technologies (Breda, the Netherlands) with 10% FCS (Gibco) and 50 µm 2-mercaptoethanol (Sigma, St Louis, MO,USA). Irradiated (50 Gy) mouse thymoma cell line EL4.B5 cells (kindly provided by Dr R Zubler, Geneva) were seeded at 50,000/well in 96 flat bottom plates on day −1. On day 0, CD19⁺ lymphocytes were seeded in 2 limiting dilution series (4000–250/well and 3–0·3/well with 96 or 48 wells for each dilution) on the EL4.B5 loaded wells, in the presence of 5% T-lymphocyte supernatant (T-SN). This T-SN was produced by culturing E-rosette enriched T-lymphocytes for 36 h in the presence of 5 µg/ml Phytohemagglutinin (Murex, Dartford, UK) and 10 ng/ml Phorbol 12-Myristate-13 acetate (Sigma). These T-lymphocytes were taken from a male blood donor who was serologically typed as HLA-A3, A26, B27,-, Cw1,-. For some individuals, the proliferation of B-lymphocytes cultured at 1000 cells/well in the EL4.B5 system was measured at day 6/7 by overnight incorporation of ³H-TdR (1 µCi/well, Amersham, Little Chalfond, UK).

CDC

Supernatants of cultured B-lymphocytes were harvested on day 10, and screened for the presence of HLA-antibodies by a modified version of the CDC. Briefly, microtest trays were filled with 2 µl supernatant under oil. Then, 1 µl of suspensions of Carboxyfluorescein diacetate (Biofine, Leiden, the Netherlands) loaded lymphocytes at 5 × 10⁶/ml were added and incubated for 1 h. Small informative panels (size 5–8 HLA typed individuals, selected on the serum reactivity of the B-lymphocyte donor), including autologous lymphocytes were used. Complement (Bioscope, Leiden, the Netherlands) was added at 5 µl and allowed to react for 2 h. Cell suspensions (1 µl) and complement (5 µl) were dispensed using a LambdaJet (One Lambda, Canoga Park, CA, USA). After addition of 5 µl of an EDTA–Propidium Iodide (Sigma)–ink (Leitz, Wetzlar, Germany) solution, percentages of dead lymphocytes were read on a Leitz Patimed (Leitz) and calculated using Patimed-Software (version 4·0) from raw 2-colour fluorescence data. Percentage of dead cells used as cut-off value for positivity was set at 20%. Monoclonal antibody FK5, recognizing a pan leucocyte marker, served as positive control and fresh culture medium as negative control. To rule out any influence of compounds (PHA/PMA or cellular products such as soluble HLA) in the T-SN on antibodies present in the B-lymphocyte supernatants, this T-SN was added 1:1 to dilution series of one IgG and two IgM human monoclonal HLA-antibodies with specificities HLA-A11, B13 and A2/A28, respectively, and found not to affect the CDC titers of these monoclonal antibodies (not shown).

ELISA for determination of IgG and IgM

Plates (Greiner, Alphen a/d Rijn, The Netherlands) were coated overnight with a goat anti-IgG or anti-IgM (Jackson Labs, Westgrove, PA, USA) diluted in 10 mm Tris pH 9·0, then were blocked with 2% bovin serum albumin (BSA, Sigma) in 0·025% Tween-20 (Sigma) in PBS (PBS-T). Fifty µl of supernatants, or standard human serum (CLB, Amsterdam, The Netherlands) in a serial dilution, were incubated for 60 min at 37°C. After washing with PBS-T, biotin labelled goat-anti-IgM or anti-IgG (Biosource, Camarillo, CA, USA) diluted in 1% BSA/PBS-T was incubated for 60 min at 37°C. After extensive washing, streptavidine horseradish peroxidase (Pierce, Rockford, IL, USA), diluted in 1% BSA in PBS-T was added and incubated for 60 min at 37°C. A colour reaction was obtained with 4·6 mm 2,2′-azine-bis(3-ethylbenzthiazoline-6-sulphonic acid) (ABTS, Sigma) in a citric acid/PBS buffer at pH 4·2. The reaction was stopped with 250 mm oxalic acid (Sigma) and measured at 450 nm. Supernatants showing OD values of 2× negative control were considered positive for the presence of IgG or IgM. The sensitivities of the ELISA were 5 ng/ml for IgG and 10 ng/ml for IgM.

ELISA for HLA antibodies

The QuikScreen test (GTI, Brookfield, WI, USA) was used for the detection of HLA class I reactive antibodies in 50 µls of B-lymphocyte supernatants, according to the manufacturers instructions with the alkaline phosphatase conjugated goat anti human IgM,G,A reagent provided. Cut-off value for positivity was set at twice the OD value of culture medium.

Statistics

B-cell Precursor frequencies (BCPF) and goodness of fit (GOF) were calculated [25] and related to number of input B-lymphocytes.

RESULTS

B lymphocyte culture

In 4/4 multiparous females tested we found good proliferation of B-lymphocytes as shown by ³H-TdR incorporation (mean 54 955 cpm; s.d. 18 568 cpm). B-lymphocytes from 3 multiparous females, seeded at low densities (3–0·3/well), showed maturation to Ig producing cells with frequencies 0·70, 0·45 and 0·57 as determined by testing supernatants in ELISAs for IgM and IgG. These frequencies are somewhat lower than the value of 0·95 reported by Wen et al. [21], likely caused by lower ELISA sensitivities.

Antibody characteristics in multiparous females

Three types of CDC reactive antibodies were detected in supernatants of the B-lymphocytes cultured at high densities (Table 2): (a) polymorphic reactivity, i.e. reactivity with one or more cells of the panel but not the autologous cell; (b) alloantigen specific reactivity (e.g. anti HLA-A2) as a part of (a) where panel composition allowed for assignment of specificities and (c) pan-reactivities (i.e. reactivity with all cells of a panel including the autologous cells). Those B-lymphocyte supernatants that were included as positives usually contained sufficient antibodies to cause 80% or greater cell lysis per test well.

Table 2.

HLA antibodies in B-cell supernatants of multiparous females

		Polymorphic antibodies		Antigenic specific antibodies			Panreactive antibodies
Donor	Antigenic stimuli†	Frequency (× 106)	GOF‡	HLA specificity	Frequency (× 106)	GOF	Frequency (× 106)	GOF
1	A2/B62	106	0·33	A2	18	0·096	239	10·9
				B62	81	0·488
2	unknown	24	0·88	A2	12	1·4	0
3	B51	14	1·1	B51	7	3·8	132	19·9*
4	A2/B13	5	7·9	A2	4	8·2	16	527*
5	A2/B62	93	2·9	B62	68	6·3	0
6	A1/B27	26	1·6	A1	13	0·986	38	66·3*
7	A2/B60	61	6·0	B40	27	4·5	14	4·7
				A2	14	1·2
8	unknown	25	4·5	A2	16	1·4	0
9	A2/B51	101	23·6*	A2	19	10·6*	138	78·2*
				B51	nd
				A24	22	13·1*
10	A2	25	81·4*	A2	4	0·99	222	104·7*
11¶	A2/B44	222	24·5*	A2	95	7·7	167	60·5*
				A2/A28	41	16·2*
				B60	20	12·5*
12	A2/B44	46	n/a §	A2	11	n/a	46	n/a
13	A2/B7	147	15·6*	0			556	17·7*
14	A2/B8	0		0			189	67·1*
15	A2	0		0	128	1·8

^†as deduced from HLA-typing of last child, when available.

^‡GOF; goodness-of-fit, if value is followed by

^*the frequency value is statistically not significant.

^§experiment done at one density only.

^¶at 14 months postpartum.

B-cell precursor frequencies in multiparous females

Polymorphic antibodies were present in the B-cell supernatants of 13/15 multiparous women. Analysis of the limiting dilution data yielded BCPF values with statistically adequate goodness of fit for 9 of these individuals (Table 2). When polymorphic antibodies were considered in more detail, alloantigen specific antibodies were present in the B-cell supernatants of 12 women, yielding BCPF values with GOF in 10 women. The specificities of alloantigen reactive antibodies were all in agreement with serum antibody specificities shown in Table 1. In one subject, 13, alloantigen specificity could not be assigned to any of the polymorphic antibodies observed in the B-cell supernatants.

Furthermore, panreactive antibodies were observed in the B-cell supernatants of 12 women (Table 2), although with adequate GOF in only 2 women. The culture of B-cells of 2 subjects, 14 and 15, produced panreactive but no polymorphic antibodies, despite the presence of clearcut allo antibodies in their sera. In B-cell cultures of 10 women, both panreactive antibodies and polymorphic antibodies were present.

To examine intertest variability, we studied one subject's cells from two bleeding dates in multiple experiments: two separate experiments from one bleed of donor 11 (expt 1 and expt 2), yielded reproducible results (Table 3) with anti-HLA-A2 BCPF-values of 95/10⁶ and 119/10⁶, respectively, while the polymorphic supernatants, as wells as the anti-B60 containing supernatants of that date showed BCPF values without adequate GOF. Anti HLA-A2/A28 reactivities were also observed. A bleed of 6 months later of this subject, also tested twice (expt 6 and expt 7), yielded reproducible results for the polymorphic supernatants which now reached adequate GOF. At the specific antibody level, all supernatants showed BCPF values with adequate GOF, but these values showed decreased intertest reproducibility.

Table 3.

HLA antibodies in B-cell supernatants of donor RO

		Polymorphic antibodies		Antigenic specific antibodies			Panreactive antibodies
Expt.	Interval†	Frequency (× 106)	GOF‡	HLA–specificity	Frequency (× 106)	GOF‡	Frequency (× 106)	GOF‡
1	14	222	24·5*	A2	95	7·7	167	60·5*
				A2/A28	41	16·2*
				B60	20	12·5*
2	14	280	16·2*	A2	119	5·4	386	83·6*
				A2/A28	86	10·7
				B60	25	10·9*
6	20	132	9·1	A2	33	1·9	0
				A2/A28	26	6·3
				B60	41	6·9
7	20	118	2·07	A2	77	6·4	0
				A2/A28	7	4·1
				B60	14	7·1

^†Time in months elapsed since the third pregnancy.

^‡GOF; goodness-of-fit, if value is followed by

^*the frequency value is statistically not significant.

Taken together, the CDC data on 11 multiparous women who showed polymorphic BCPF values with adequate goodness of fit, ranging from 0 to 123/10⁶, yielded a mean BCPF value of 43 (± 45 sd.) per 10⁶ B-lymphocytes.

Antibodies secreted by B-lymphocytes are HLA specific

To further confirm the HLA reactivity of the antibodies contained in the supernatants and to rule out the presence of autoantibodies that sometimes obscure CDC serum data, we analysed 68 randomly chosen supernatants of donor 11 (Table 3, expt 1, B-cell seeded at 4000/well) by ELISA (QuikScreen). Twenty-six of 44 CDC polymorphic reactive wells also were positive by ELISA (Table 4), and 16 of 24 CDC negative wells also were negative by ELISA. The poor correlation coefficient, r = 0·215, of this test is caused by incongruent antibodies of both types.

Table 4.

Comparison of ELISA and CDC reactivities of B-cell supernatants of individual 11

	CDCpos	CDCneg	Total
ELISA pos	26	8	34
ELISA neg	18	16	34
Total	44	24	68

R = 0·215.

Non sensitized individuals

In 1 of 2 males we found polymorphic reactivity with statistically adequate goodness of fit, but the reactivities were just above the CDC background (Table 5).

DISCUSSION

The study of cells involved in solid organ rejection has mainly focused on the T-cell compartment, however, the presence in serum of specific HLA antibodies routinely serves as the main contraindication for transplantation. While the antibody repertoire of highly immunized patients awaiting transplantation is studied by screening complex sera, the relative contributions within the B-cell compartment in the production of these antibodies are unknown. Estimation of these HLA-specific B-cell precursor frequencies may be useful for fine tuning immunosuppressive regimens directed towards the B-cell compartment. To test whether specific HLA-B-cell precursor frequencies are within measurable range, we studied women who had developed HLA antibodies upon delivery.

The present study is the first demonstration of the ability of in vitro cultured B-lymphocytes of some individuals to secrete HLA-antibodies in quantities that are detectable by CDC and ELISA. The serum HLA antibody components are reflected in the antibody specificities found in the B-cell supernatants. This assay thus can replace cumbersome absorptions to unravel a composite of specificities in a serum. We confirmed the HLA class I specificity for the majority of CDC reactive antibodies in one individual, using ELISA for binding to isolated HLA-class I molecules. A comparison of these assays also revealed CDC negative, but HLA binding antibodies, as well as the converse, CDC positive and polymorphic, non HLA binding antibodies. The former type of antibody has been implicated in graft rejection [26] while the latter probably reflects low affinity antibodies which escape ELISA detection, yet may be important for cell activation and graft rejection [27,28]. Incidentally, the present study underscores the importance of testing patients' sera in both binding and functional assays, with the demonstration of those three types of antibodies within one individual as the products of single B-lymphocytes. This type of study may also give a view of the heterogeneity within one individual of the B-lymphocyte clones with HLA-A2 specificity, which cannot be detected when testing sera for HLA-A2 reactivity. Besides polymorphic antibodies, the B-cells of many individuals produce panreactive antibodies as well. These may be natural antibodies of low affinity that are reactive with multiple tissues, and are known to be present in normal individuals [29].

The present data also demonstrate that the frequency of HLA antibody producing B-cells can be estimated accurately in most cases. Data from two independently performed B-lymphocyte cultures from a single bleeding of one multiparous female, 11, showed intertest reproducibility of the BCPF values. A bleeding of 11 from a later date showed BCPFs in a lower range, also tested twice. However, the magnitude of BCPF values of the multiparous women could not be correlated with interval since last pregnancy or with the number of pregnancies. Lack of correlation between BCPF value and serum antibody titre (for autoantigens Ro and LA) has been reported [30].

Humoral responses in other antigenic systems have been studied in the EL4.B5 culture setup, thus allowing comparison of our data with data from these systems. In B and A (of the red blood cell AB0 system) individuals, B-lymphocytes producing anti A and B antibodies were reported to occur with a mean BCPF value of 86/10⁶ [24]. In contrast, Rheumatoid Factor (RF) producing B-lymphocytes occur at 322/10⁶ in normal individuals, but are significantly higher in RA patients: 3816/10⁶ [23], and even as high as 3% [22]. The range of HLA-polymorphic BCPF values found in the present study 43 (± 45 sd.)/10⁶, coincides with the range found in the AB0 system. B-lymphocytes producing panreactive antibodies occur at higher frequency (Table 2), albeit in almost all cases without adequate goodness of fit.

Despite the absence of cytotoxic HLA antibodies in their sera, we found low but distinct BCPF reactivity in two nonimmunized males, but with low CDC percentages just exceeding background. We tested their supernatants for the presence of CDC reactive antibodies with panels of 8 cell suspensions, and no reaction pattern was found more than once.

Peripheral blood of multiparous females was used as source of B-lymphocytes in this study, although it is known, at least for pregnancy in the rat, that the majority of alloantibodies is formed in the spleen [31]. This indicates that with the present assay we are accessing only a fraction of lymphocytes capable of alloantibody formation. In contrast, the anatomical location of alloantibody producing B-lymphocytes, formed as a consequence of incompletely matched grafts is unknown. Regardless of the nature of the immunizing event, it is now widely accepted that efficient generation of antibody requires clonal amplification of B-lymphocytes, somatic mutation, selection of mutants with high affinity and subsequent differentiation. Differentiation to either memory cells or to plasma cells takes place in the follicles of lymphoid organs where B-lymphocytes have migrated and have formed germinal centers. In what quantities and in which stage B-cells then migrate to the circulation is not known. Equally unknown is to what extent the amount of B-cells present in peripheral blood is proportional to the B-cells present in the lymphoid organs. In this light it is intriguing that some (2/15) multiparous women show no polymorphic antibodies in B-lymphocyte culture supernatants and 2/15 show polymorphic antibodies, yet do not show alloantigen specific antibodies, despite the presence of serum antibodies with clear-cut HLA specificity. This suggests that these HLA-A2 specific B-lymphocytes have homed to lymphoid organs, due to prolonged absence of alloantigen in these women. On the other hand, the considerable frequency, in the circulation, of B-cells producing panreactive antibodies, while cytotoxic antibodies reactive with autologous cells are normally absent from the serum, suggests the alleviation in vitro of suppressive cells which in vivo keep these circulating B-cells in check. It also suggests different homing characteristics for B-cells producing these two categories of antibodies.

Application of the BCPF assay, to the lymphocytes of prospective transplant patients with high PRA values, may serve two purposes.

(i) When supernatants are analysed with informative panels, the presence of certain specificities will aid in the identification of donororgans that should be avoided for transplantation thus supplementing data that are derived from testing sera against panels for acceptable mismatch identification.

(ii) Once patients have been identified with high BCPF values, inclusion of immunosuppressive agents (drugs, intravenous immunoglobulins) in the B-lymphocyte culture system will reveal the efficiency of these drugs at suppressing all, or a fraction, alloantibody secreting B-lymphocytes [32–34].

At the laboratory level, the results of BCPF testing will aid in the identification of HLA antibody producing individuals, with a sufficiently high BCPF for establishing anti HLA secreting hybridomas. Indeed, we have successfully developed 3 hybridomas from donor 11 secreting monoclonal antibodies with divergent specificities. Furthermore this test will be suitable for assessing the efficiency of in vitro immunization protocols for determination of the capacity to produce HLA specific antibodies in the absence of an in vivo primary immunization event.

In summary, we developed a test for the assay by CDC of supernatants of cultured B-lymphocytes. The results of these tests yield BCPF values with statistical goodness of fit for the majority of individuals tested.