Journal of Immunological Methods, 130 (1990) 73-79 Elsevier
73
JIM05586
Measurement of alloantibody by flow cytometry Peter L. Leenaerts 1.2, Dirk De R u y s s c h e r 2, Michel V a n d e p u t t e 2 a n d M a r k W a e r 1,2 I Division of Nephrology, and " Rega Institute for Medical Research, University of Leuven, Louvain, Belgium (Received 14 September 1989, accepted 12 February 1990)
A new method based on flow cytometry has been developed to determine IgG alloantibodies in the serum of immunized rodents. The method utilizes target lymphoid cells, diluted serum and labeled anti-mouse or anti-rat IgG antibodies. In serum from highly immunized animals alloantibodies could be demonstrated up to a dilution of (3 x 104)- 1 which makes the test approximately 300 times more sensitive than a simultaneously performed complement-dependent cytotoxicity assay (CDCA). Moreover a linear relationship between the amount of alloantibody and the log value of the mean fluorescence of the target cells was found. This linearity permits the comparison of alloantibody production between individual samples by comparing the mean fluorescence values. The reproducibility of the assay was excellent since the coefficients of variation for all dilutions were less than 15%. By using two-color fluorescence the method can discriminate alloantibodies directed against class I and class II MHC antigens. Finally, in addition to its high sensitivity and good reproducibility, the method was found to be at least twice as fast as CDCA. Key words: AIIoantibody; Flow cytometry; Major histocompatibility complex antigen
Introduction Many techniques, ranging from precipitation and electrophoresis methods to ELISA assays, have been reported for the measurement of antibodies against soluble antigens (e.g., Hudson and Hay, 1980). However, the number of techniques to determine allo- or xenoantibodies against mem-
Correspondence to: P.L. Leenaerts, Stanford University Hospital, Division of Nephrology, Stanford, CA 94305, U.S.A. (Tel.: (415) 723-8322). * Supported by the National Fund for Scientific Medical Research of Belgium (N.F.W.G,O.). Abbreviations: CDCA, complement dependent cytotoxicity assay; EA, erythrocyte antibody coated; FC, flow cytometry; FITC, fluorescein isothiocyanate; PE, phycoerythrin; SD, standard deviation; CV, coefficient of variation.
brane-bound antigens, is much more limited. Indirect assays, such as EA (erythrocyte antibodycoated) rosette inhibition (Asfar et al., 1985; Jones et al., 1988) and indirect hemagglutination (Power et al., 1987), have been used for the detection of alloantibodies against red blood cells. Other indirect methods have utilized radiolabeled protein A molecules (Dorval et al., 1975) or iodinated antibodies against surface antigens (Morris et al., 1975). The latter methods suffer from the disadvantages associated with the use of radioactive materials. A cellular enzyme-linked immunospecific assay (CELISA) was developed by Morris et al. (1982) and was claimed to be more sensitive than a cytotoxicity assay. Despite its higher sensitivity and quantitative potential the method is not widely used (Cunningham et al., 1987; AI-Muzairai et al., 1989). The most frequently used technique for the detection of alloantibody is the complement-de-
0022-1759/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
74 pendent cytotoxicity assay. Whether a direct microcytotoxicity test (Terasaki et al., 1978) or an indirect antiglobulin test (Johnson et al., 1972) is utilized the basic principle remains the same. In the direct method, cells coated with complementfixing antibodies are directly lysed by complement whereas in the indirect method complement fixing antiglobulins are used to detect membrane-bound antibodies. Lysis is visualized using a vital dye exclusion test. CDCA is thus an on-off phenomenon and the amount of antibody bound to the cell surface is not directly measured. Quantification of alloantibodies can, however, be expressed as the serum dilution that still results in 80 or 50% cell lysis. Since complement activation, and thus cell lysis, requires the binding of a relatively large amount of antibody to the cell surface, it is not surprising that the sensitivity of this method is not very high. Sachs et al. (1971) showed that, even using strong xenogeneic antisera, the limit of detection of xenoantibodies was reached using a dilution of 1/128 or 1/256. In order to improve the sensitivity of assays for the detection of anti-HLA antibodies in kidney recipients, Garovoy et al. (1983) introduced a flow cytometry cross-matching method. Whether or not the increased sensitivity is relevant to clinical transplantation is a controversial issue (Garovoy et al., 1985; lwaki et al., 1987; Thistlewaite et al., 1987; Lazda et al., 1988; Talbot et al., 1988) but, all authors agree that this method is far more sensitive than conventional CDCA. In the flow cytometry technique, anti-HLA antibodies are considered to be present in the recipient when the mean fluorescence of the donor cells, after incubation with recipient serum and labeled anti-human globulin antibodies, differs significantly from donor cells which were exposed to a negative human serum. The aim of the cross-matching technique was to detect sensitization in candidates for kidney transplantation. We sought to establish whether a new method could be developed to measure alloantibodies in a sensitive and quantitative manner in sensitized rodents. Moreover, by using two-color fluorescence analysis, we reasoned that alloantibodies against class I and class II antigen-bearing cells could be independently demonstrated.
Materials and methods
A llosera B A L B / c anti-C57B1/6 and B A L B / c antiC 3 H / H e allosera were obtained by immunizing B A L B / c mice with a skin graft followed by three weekly i.p. injections of 1 x 108 allogeneic spleen cells in complete Freund's adjuvant. Ascites fluid was collected in the week following the third injection.
Target cells C57B1/6 (H-2 h) and C 3 H / H e (H-2 k) spleen cells were prepared by disrupting spleens and flushing the cells through a steel mesh. After ammonium chloride lysis, the cells were suspended in RPMI at a concentration of 5 x 106 or 1.67 x 106 cells/ml.
CDCA A modification of the assay previously described (Waer et al., 1982) was used. Spleen cells at a concentration of 5 × 106 cells/ml were suspended in RPMI medium supplemented with Hepes, glutamine and 2% fetal calf serum. A series of dilutions of alloserum (20/~1) was prepared and incubated with 20 #1 of the cell suspension for 30 min at room temperature in a microtiter plate. For an additional 45 min 20 ~1 of guinea pig complement (Behring Diagnostics, Brussels, Belgium), diluted 1 / 3 in RPMI was added. Only guinea pig complement having a control cytotoxicity with a negative serum of less than 20% was used. Viability was tested by adding 20 #1 eosin (2%). Results were expressed in terms of percent lysed cells. The cytotoxic titer was arbitrarily taken as the last dilution of antiserum causing more than 50 percent cell lysis.
Cell labeling for flow cytometry 0.5 x 106 or 0.167 x 106 target cells suspended in 0.1 ml of medium were incubated for 20 min with 0.1 ml alloserum in different dilutions. The negative control was obtained by incubating the cells with 0.1 ml normal mouse serum ( 1 / 1 0 diluted). The cells were washed twice with 2 ml ice cold PBS, followed by a second incubation at 4 ° C with 25 /xl ( 1 / 5 diluted) of a FITC-conjugated sheep anti-mouse IgG (Fc) (The Binding Site,
75
Birmingham, England). This second-step reagent was first ultracentrifuged to make sure that no globulin aggregates were present. The administered dose was judged to be saturating, since increasing the amount of the second-step reagent did not alter the fluorescence level whatever alloserum dilution was used. When two-color immunofluorescence was used to demonstrate alIoantibodies against CD4 ÷ and C D 4 - lymphoid cells, the cells were further incubated with normal mouse serum ( 1 / 1 0 diluted) to block open binding sites of remaining anti-mouse IgG (Fc) and were then washed twice with PBS. Finally, the cells were incubated with 25 #1 ( 1 / 5 diluted) of antiL3T4 PE m A b (Becton Dickinson, Mountain View, CA).
Flow cytometry A Facstar Plus flow cytometer (Becton Dickinson) was used. Acquisition and analysis of results were performed using the Consort 30 program. For each determination ten thousand cells were acquired in a lymphocyte gate. When the reproducibility of the method was studied, flow cytometer settings such as laser power, electronic and photomultiplier gains were carefully noted and the same working conditions were repeated in separate experiments. Fluorescence beads (Becton Dickinson) were used as controls.
Statistical analysis Linear regression, standard deviation and coefficient of variation were determined using classical methods.
Results
The sensitivity of the flow cytometry method was first compared to CDCA. Fig. 1 shows the cytotoxic potential of two allosera of strongly immunized mice. Both sera had a cytotoxic titer of 128. At the dilution of 1/256 the alloantibodies were no longer capable of inducing more than 50% cell lysis. Alloantibodies could thus be shown up to a dilution of 1/128. Fig. 2 shows two-color immunofluorescence of target C3H cells exposed to the same anti-C3H alloserum and stained with an anti-mouse I g G F I T C and anti-L3T4 PE. The
• ANTI-C57 BI
101
.
.
4
.
.
.
.
.
16
.
.
.
.
.
64
.
.
.
.
.
.
256
.
1024
-1 (P~LLOSERUM DILU'I'ION) Fig. 1. Cytotoxicity of strong BALB anti-C3H (o) or BALB anti-C57BI (e) sera in the CDCA method. In the doubling dilution series the cytotoxic titer for both sera was 128. The data points represent the mean values of three independent determinations. Cytotoxic titer is defined as the reciprocal of the last dilution of serum that causes more than 50% cell lysis.
latter staining divides the target cells into CD4 ÷ and C D 4 - populations. In mice CD4 ÷ cells bear class I antigens only whereas C D 4 - cells contain B cells which have both class I and class II antigens. Fig. 2A shows the negative control. The difference between the mean fluorescence of CD4 ÷ and C D 4 - cells was partly caused by weak nonspecific binding of immunoglobulins to Fc receptor bearing cells and partly by the fact that 1-2% of the B cells expressed IgG on their surface. Fig. 2B and 2C show the target cells after exposure to alloserum dilutions of 1/300 and 1 / 3 respectively. The C D 4 - population appeared to consist of two groups of cells as far as immunofluorescence intensity was concerned. This was due to the fact that class II positive cells were coated with anticlass II alloantibodies as well as anti-class I antibodies. Fig. 2 D demonstrates the histograms obtained with four different alioserum dilutions. The relationship between the reciprocal of the serum dilutions and the mean fluorescence is illustrated for both allosera by a log-log plot (Fig. 3). The curves can be divided into three parts. At dilutions of 1 / 3 and 1 / 1 0 the upper part tends to flatten, probably as a result of saturation of the cell surface antigens with alloantibody. This is
76
suggested by the fact that increasing the amount anti-IgG F I T C did not alter the mean fluorescence indicating that the second-step reagent did not play a limiting role. However, increasing or decreasing the number of target cells in the presence of the same amount of alloserum resulted in either an earlier or later inflexion of the curve,
respectively (see also Fig. 4). The second and more important part of the curves between dilutions of 1 / 3 0 and 1/10,000 is represented by a linear relationship. The correlation coefficients of the regression lines of the anti-C3H and the anti-C57B1 serum were r 1 -- - 0 . 9 9 6 and r2 = - 0 . 9 9 7 respectively. Finally, at the highest dilution, when few
,qr
o g.i
m
5.7
m
19
118.0 o
m
165.5
9.8
-a
m
,
O
--me
101
10 z
(A)
Iti ~j
tO 4
"1 On
GREEN
i
, ,,..|
| lwl,.|
i
| vw,H,j
IB 2
IB!
w • wl,,,,
IB ~l
I0 4
(B)
qf
m ~e
1QO
cq fll
966.1 2
t
1196.9
%
.Q!
we
'
~IOU
' ''''"l
I01
(C)
'
' ''''"l
10 2
,,iw.i
IB 2
GREEN
•
/ I~ ~
l
i !:
'i .
1
i
gs /
'~'
...... 101~j
........
• ,IH.
I0 4
al~/l~'
(D)
"~" . . .|.~. l J ' "
11~1~ '
........
GREEN
Fig. 2. Two-color immunofluorescence analysis of 0.5 × l0 s C3H target spleen cells. The cells were incubated first with BALB anti-C3H alloserum followed by sheep anti-mouse IgG (Fc) FITC (horizontal axis). In a third step the cells were incubated with anti-L3T4 PE (vertical axis). The numbers indicated near the squares represent the mean fluorescence units along the abcissa..4: normal BALB serum 1 / 1 0 diluted; B: anti-C3H serum 1/300 diluted; C: anti-C3H serum 1 / 3 diluted; D: histograms of all target cells. Histograms 1, 2, 3 and 4 were obtained with alloserum dilutions of 1/10, 1/30, 1 / 3 0 0 and 1/3000 respectively.
77
100
d o ~rr~5~
3E
I
~
1"I 10
0 ANTI-C.3H 0.5X10
'1 101
10 2
10 3
10 4
101
10 2
10 3
(/U.LO6ERUM DI.U/IO~
(ALL(~IERUM DILUTK~1
Fig. 3. Relationship between alloserum dilutions and mean fluorescence units. The correlation coefficients of the linear regression lines between the dilutions 1/30 and 1/10,000 are shown. The negative controls + 2 SD are also represented in the lower right comer (o and o). alloantibodies were present the curves flattened again. At a d i l u t i o n of 1 / 3 0 , 0 0 0 the m e a n fluorescences of the a n t i - C 3 H a n d anti-C57B1 sera were 11.4 a n d 12.4 U respectively. T h e c o r r e s p o n d i n g negative controls _+ 1 SD d e t e r m i n e d in four indep e n d e n t samples were 8.4 _+ 0.7 a n d 8.8 + 0.8 U ( P < 0 . 0 0 1 ) . T h u s the sensitivity of the m e t h o d was very high a n d alloantibodies could be detected at s e r u m d i l u t i o n s up to 1/30,000. C o m p a r e d to the C D C A m e t h o d detection of all o a n t i b o d y b y flow c y t o m e t r y was nearly 300 times m o r e sensitive. We further investigated whether the flow cyt o m e t r y results showed good reproducibility. The
10 4 -1
Fig. 4. Influence of the number of target cells. Both curves show the relationship between a dilution series of the anti-C3H serum and the mean fluorescence units when 0.5x 106 or 0.167 × 106 target cells were used. Data points represent the measured fluorescence minus the fluorescence of the negative control.
same sera were analysed at four i n d e p e n d e n t times. A t t e n t i o n was p a i d to ensure the same flow cyt o m e t e r settings, n u m b e r of target cells a n d serum dilutions. T h e results are shown in T a b l e I. F o r n e a r l y all d i l u t i o n s the coefficients of variation (CV) were less t h a n 10%. T h a t it is i m p o r t a n t to use a c o n s t a n t n u m b e r of target cells is d e m o n strated in Fig. 4. F r o m this it could be erroneously inferred that the curves were o b t a i n e d b y two d i l u t i o n series of which o n e c o n t a i n e d three times more a l l o a n t i b o d y t h a n the other whereas o n l y the n u m b e r of target cells differed.
TABLE I REPRODUCIBILITY OF THE FLOW CYTOMETRIC DETERMINATION OF ALLOANTIBODY (Dilution) -1
3 10 30 1130 300 1000 3000 10000 30000 Negative control
Anti-C3H alloserum
Anti-C57 BL
Mean fluorescence (U)
SD
CV
SD
CV
(~)
Mean fluorescence (U)
(U)
1058.0 922.0 628.0 340.0 136.0 53.0 23.0 13.6 11.4 8.4
62 83 37 24 12 4.6 3.2 1.3 0.6 0.7
(U)
(~)
5.9 9.1 6.2 7.3 8.9 8.7 13.9 9.6 5.3 8.3
1 721 1440 1024 464 168 60 27.1 14.8 12,2 8.8
148 58 39 32 18 8.6 4 0.8 1.0 0.8
8.6 4.0 3~8 6.9 10.7 14.3 14.8 5.5 8.2 8.9
78
Discussion According to Morris et al. (1982), assays for alloantibody should be highly sensitive, capable of expressing results in a quantitative manner, reproducible, rapid and simple to interpret. The flow cytometry method we have described in the present study fulfills all of these conditions. The method was able to detect surface bound IgG antibodies in alloserum dilutions up to 1/30,000. For the same target cell:alioserum ratio, alloantibodies could only be demonstrated up to a dilution of 1/128 in CDCA. The flow cytometry method is thus much more sensitive. This higher sensitivity may not directly be relevant for organ transplantation, as has been reported in crossmatching results (Lazda et al., 1988), but it can permit distinction between tolerance or unresponsiveness and diminished responses due to immunosuppressive therapy. Apart from its high sensitivity, the method is ideal for quantitating IgG alloantibodies. Indeed, under standardized flow cytometry conditions and using a constant number of target cells, a linear relationship between the log values of mean fluorescence and the reciprocals of serum dilutions was established. This relationship permits comparisons of serum from individual animals both to each other or to a 'standard' line of a known positive serum and to draw statistical conclusions from the measured mean fluorescence units. In CDCA, a series of dilutions has to be assayed to find the cytotoxicity titer and thereby the quantity of alloantibodies, in each new serum, whereas one determination is sufficient with the flow cytometry method. This, together with shorter incubation times, makes the flow cytometry determination two to three times faster than CDCA. In addition, by using two-color immunofluorescence analysis, the method can immediately demonstrate alloantibodies against target cells bearing class I antigens only and cells bearing class I and class II antigens. A minor disadvantage of the proposed method is that it measures only surface-bound IgG antibodies. In CDCA complement fixing antibodies, both lgG and IgM, are measured. On the other hand, non-complement fixing IgG antibodies are not detected by the latter method. We deliberately
focussed on IgG alloantibodies for two reasons. First, early IgM alloantibodies will rapidly disappear and will hardly be detectable after a few weeks (Jones et al., 1987). Second, a high percentage of mouse, as well as human B cells carry IgM and IgD on their surface and labeled second-step anti-mouse Ig antibodies will therefore non-specifically stain the B cells in the target cell population, thus interfering with the alloantibody determination. This is not the case when specific anti-IgG antibodies are used, since only a minor fraction of B cells express IgG at low density. However, when evaluating IgM alloantibodies two strategies can be considered. First, B cells can be removed from the target cell population by cell separation techniques. This will, of course, complicate and prolong the method markedly. A more elegant solution is to use twocolor immunofluorescence analysis as shown in Fig. 2. Only the T cell population is then taken into account. In the present experiments we focussed on alloantibodies, but it will be clear that xenoantibodies can also be measured in a similar way. In conclusion, we have described a method to measure alloantibodies which is both quantitative and sensitive. It is of potential value in studies of transplantation, tolerance and immunoresponsiveness.
References A1-Muzairai, I.A., Innes, A., Hillis, A., Stewart, K.N., Bone, J.M.., Catto, G.R.D. and MacLeod, A.M. (1989) Renal transplantation: Cyclosporin A and antibody development after donor-specific transfusion. Kidney Int. 35, 1057. Asfar, S.K., Power, D.A., Mason, R.J., MacLet~, A.M., Simpson, J.G., Whiting, P.H., Engeset, J. and Catto, G.R.D. (1985) Prolonged survival of rat renal allografts after multiple allogeneic pregnancies; strain specificity and role of erythrocyte antibody rosette inhibiting antibodies. Clin. Sci. 69, 41. Cunningham, C., Power, D.A., lnnes, A., Lind, T. and Catto, G.R.D. (1987) Maternal alloantibody responses during early pregnancy detected by a cellular enzyme-linked immunospecific assay. Hum. lmmunol. 19, 7. Dorval, G., Welsh, K.I. and Wigzell, H. (1975) A radio-immunoassay of cellular antigens on living cells using iodinated soluble protein A from Staphylococcus aureus. J. Immunol. Methods 7, 237. Garovoy, M.R., Reinschmidt, M.A., Bigos, M., Perkins, H., Colombe, B., Feduska, N. and Salvatierra, O. (1983) Flow
79 cytometry analysis: A high technology crossmatch technique facilitating transplantation. Transplant. Proc. 15, 1939. Garovoy, M.R., Colombe, B.W., Melzer, J., Feduska, N., Shields, C., Cross, D., Amend, W., Vincenti, F., Hopper, S., Duca, R. and Salvatierra, O. (1985) Flow cytometry crossmatching for donor-specific transfusion recipients and cadaveric transplantation. Transplant. Proc. 17, 693. Hudson, L. and Hay, F.C. (1980) Antibody interaction with antigen. In: Practical Immunology. Blackwell Scientific Publications, Oxford, p. 93. Iwaki, Y., Cook, D.J., Terasaki, P.I., Lau, M., Terashita, G.Y., Danowitch, G., Fine, R., Ettenger, R., Mendez, R., Kavalich, A., Martin, D., Soderblom, R., Ward, H., Berne, T., Lieberman, E. and Strauss, F. (1987) Flow cytometric crossmatching in human cadaver kidney transplantation. Transplant. Proc. 19, 764. Johnson, A.H., Rossen, R.D. and Butler, W.T. (1972) Detection of aUoantibodies using a sensitive antiglobulin microcytotoxicity test: Identification of low levels of preformed antibodies in accelerated allograft rejection. Tissue Antigens 2, 215. Jones, M.C., Power, D.A., Cunningham, C. and Catto, G.R.D. (1988) The influence of repeated transfusions and cyclosporine on secondary alloantibody responses in inbred rats. Transplantation 45, 1094. Lazda, V.A., Pollack, R., Mozes, M.F. and Jonasson, O. (1988) The relationship between flow cytometer crossmatch results and subsequent rejection episodes in cadaver renal allograft recipients. Transplantation 45, 562. Morris, R.J. and Williams, A.F. (1975) Antigens on mouse and
rat lymphocytes recognized by rabbit antiserum against rat brain: the quantitative analysis of a xenogeneic antiserum. Eur. J. lmmunol. 5, 274. Morris, R.E., Thomas, P.T. and Hong, R. (1982) Cellular enzyme linked immunospecific assay (CELISA). I. A new method that detects antibodies to cell-surface antigens. Hum. lmmunol. 5, 1. Power, D.A., Cunningham, C. and Catto, G.R.D. (1987) The role of RTI antigen differences in semi-allogeneic rat pregnancy. Clin. Sci. 72, 37. Sachs, D.A., Winn, H.J. and Russell, P.S. (1971) The immunologic response to xenografts: Recognition of mouse H-2 histocompatiblity antigens by the rat. J. lmmunol. 107, 481. Talbot, D., Givan, A.L., Shenton, B.K., Stratton. A., Proud, G. and Taylor, R.M.R. (1988) Rapid detection of low levels of donor specific IgG by flow cytometry with single and dual color fluorescence in renal transplantation. J. Immunol. Methods 112, 279. Terasaki, P.I., Bernoco. D., Parks, M.S.. Ozturk, G. and lwaki, Y. (1978) Microdroplet testing for HLA-A,-B and -D antigens. Am. J. Clin. Pathol. 69, 103. Thistlethwaite, J.R., Buckingham, M., Stuart, J.K., Gaber, A.O., Mayes, J.T. and Stuart, F.P. (1987) T cell immunofluorescence flow cytometry cross-match results in cadaver donor renal transplantation. Transplant. Proc. 17, 722. Waer, M., Ang, K.K., Vandeputte, M. and Van der Schueren, E. (1982) Influence of overall treatment time in a fractionated total lymphoid irradiation as an immunosuppressive therapy in allogeneic bone marrow transplantation in mice. Int. J. Radiat. Oncol. Biol. Phys. 8, 1915.
73
JIM05586
Measurement of alloantibody by flow cytometry Peter L. Leenaerts 1.2, Dirk De R u y s s c h e r 2, Michel V a n d e p u t t e 2 a n d M a r k W a e r 1,2 I Division of Nephrology, and " Rega Institute for Medical Research, University of Leuven, Louvain, Belgium (Received 14 September 1989, accepted 12 February 1990)
A new method based on flow cytometry has been developed to determine IgG alloantibodies in the serum of immunized rodents. The method utilizes target lymphoid cells, diluted serum and labeled anti-mouse or anti-rat IgG antibodies. In serum from highly immunized animals alloantibodies could be demonstrated up to a dilution of (3 x 104)- 1 which makes the test approximately 300 times more sensitive than a simultaneously performed complement-dependent cytotoxicity assay (CDCA). Moreover a linear relationship between the amount of alloantibody and the log value of the mean fluorescence of the target cells was found. This linearity permits the comparison of alloantibody production between individual samples by comparing the mean fluorescence values. The reproducibility of the assay was excellent since the coefficients of variation for all dilutions were less than 15%. By using two-color fluorescence the method can discriminate alloantibodies directed against class I and class II MHC antigens. Finally, in addition to its high sensitivity and good reproducibility, the method was found to be at least twice as fast as CDCA. Key words: AIIoantibody; Flow cytometry; Major histocompatibility complex antigen
Introduction Many techniques, ranging from precipitation and electrophoresis methods to ELISA assays, have been reported for the measurement of antibodies against soluble antigens (e.g., Hudson and Hay, 1980). However, the number of techniques to determine allo- or xenoantibodies against mem-
Correspondence to: P.L. Leenaerts, Stanford University Hospital, Division of Nephrology, Stanford, CA 94305, U.S.A. (Tel.: (415) 723-8322). * Supported by the National Fund for Scientific Medical Research of Belgium (N.F.W.G,O.). Abbreviations: CDCA, complement dependent cytotoxicity assay; EA, erythrocyte antibody coated; FC, flow cytometry; FITC, fluorescein isothiocyanate; PE, phycoerythrin; SD, standard deviation; CV, coefficient of variation.
brane-bound antigens, is much more limited. Indirect assays, such as EA (erythrocyte antibodycoated) rosette inhibition (Asfar et al., 1985; Jones et al., 1988) and indirect hemagglutination (Power et al., 1987), have been used for the detection of alloantibodies against red blood cells. Other indirect methods have utilized radiolabeled protein A molecules (Dorval et al., 1975) or iodinated antibodies against surface antigens (Morris et al., 1975). The latter methods suffer from the disadvantages associated with the use of radioactive materials. A cellular enzyme-linked immunospecific assay (CELISA) was developed by Morris et al. (1982) and was claimed to be more sensitive than a cytotoxicity assay. Despite its higher sensitivity and quantitative potential the method is not widely used (Cunningham et al., 1987; AI-Muzairai et al., 1989). The most frequently used technique for the detection of alloantibody is the complement-de-
0022-1759/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
74 pendent cytotoxicity assay. Whether a direct microcytotoxicity test (Terasaki et al., 1978) or an indirect antiglobulin test (Johnson et al., 1972) is utilized the basic principle remains the same. In the direct method, cells coated with complementfixing antibodies are directly lysed by complement whereas in the indirect method complement fixing antiglobulins are used to detect membrane-bound antibodies. Lysis is visualized using a vital dye exclusion test. CDCA is thus an on-off phenomenon and the amount of antibody bound to the cell surface is not directly measured. Quantification of alloantibodies can, however, be expressed as the serum dilution that still results in 80 or 50% cell lysis. Since complement activation, and thus cell lysis, requires the binding of a relatively large amount of antibody to the cell surface, it is not surprising that the sensitivity of this method is not very high. Sachs et al. (1971) showed that, even using strong xenogeneic antisera, the limit of detection of xenoantibodies was reached using a dilution of 1/128 or 1/256. In order to improve the sensitivity of assays for the detection of anti-HLA antibodies in kidney recipients, Garovoy et al. (1983) introduced a flow cytometry cross-matching method. Whether or not the increased sensitivity is relevant to clinical transplantation is a controversial issue (Garovoy et al., 1985; lwaki et al., 1987; Thistlewaite et al., 1987; Lazda et al., 1988; Talbot et al., 1988) but, all authors agree that this method is far more sensitive than conventional CDCA. In the flow cytometry technique, anti-HLA antibodies are considered to be present in the recipient when the mean fluorescence of the donor cells, after incubation with recipient serum and labeled anti-human globulin antibodies, differs significantly from donor cells which were exposed to a negative human serum. The aim of the cross-matching technique was to detect sensitization in candidates for kidney transplantation. We sought to establish whether a new method could be developed to measure alloantibodies in a sensitive and quantitative manner in sensitized rodents. Moreover, by using two-color fluorescence analysis, we reasoned that alloantibodies against class I and class II antigen-bearing cells could be independently demonstrated.
Materials and methods
A llosera B A L B / c anti-C57B1/6 and B A L B / c antiC 3 H / H e allosera were obtained by immunizing B A L B / c mice with a skin graft followed by three weekly i.p. injections of 1 x 108 allogeneic spleen cells in complete Freund's adjuvant. Ascites fluid was collected in the week following the third injection.
Target cells C57B1/6 (H-2 h) and C 3 H / H e (H-2 k) spleen cells were prepared by disrupting spleens and flushing the cells through a steel mesh. After ammonium chloride lysis, the cells were suspended in RPMI at a concentration of 5 x 106 or 1.67 x 106 cells/ml.
CDCA A modification of the assay previously described (Waer et al., 1982) was used. Spleen cells at a concentration of 5 × 106 cells/ml were suspended in RPMI medium supplemented with Hepes, glutamine and 2% fetal calf serum. A series of dilutions of alloserum (20/~1) was prepared and incubated with 20 #1 of the cell suspension for 30 min at room temperature in a microtiter plate. For an additional 45 min 20 ~1 of guinea pig complement (Behring Diagnostics, Brussels, Belgium), diluted 1 / 3 in RPMI was added. Only guinea pig complement having a control cytotoxicity with a negative serum of less than 20% was used. Viability was tested by adding 20 #1 eosin (2%). Results were expressed in terms of percent lysed cells. The cytotoxic titer was arbitrarily taken as the last dilution of antiserum causing more than 50 percent cell lysis.
Cell labeling for flow cytometry 0.5 x 106 or 0.167 x 106 target cells suspended in 0.1 ml of medium were incubated for 20 min with 0.1 ml alloserum in different dilutions. The negative control was obtained by incubating the cells with 0.1 ml normal mouse serum ( 1 / 1 0 diluted). The cells were washed twice with 2 ml ice cold PBS, followed by a second incubation at 4 ° C with 25 /xl ( 1 / 5 diluted) of a FITC-conjugated sheep anti-mouse IgG (Fc) (The Binding Site,
75
Birmingham, England). This second-step reagent was first ultracentrifuged to make sure that no globulin aggregates were present. The administered dose was judged to be saturating, since increasing the amount of the second-step reagent did not alter the fluorescence level whatever alloserum dilution was used. When two-color immunofluorescence was used to demonstrate alIoantibodies against CD4 ÷ and C D 4 - lymphoid cells, the cells were further incubated with normal mouse serum ( 1 / 1 0 diluted) to block open binding sites of remaining anti-mouse IgG (Fc) and were then washed twice with PBS. Finally, the cells were incubated with 25 #1 ( 1 / 5 diluted) of antiL3T4 PE m A b (Becton Dickinson, Mountain View, CA).
Flow cytometry A Facstar Plus flow cytometer (Becton Dickinson) was used. Acquisition and analysis of results were performed using the Consort 30 program. For each determination ten thousand cells were acquired in a lymphocyte gate. When the reproducibility of the method was studied, flow cytometer settings such as laser power, electronic and photomultiplier gains were carefully noted and the same working conditions were repeated in separate experiments. Fluorescence beads (Becton Dickinson) were used as controls.
Statistical analysis Linear regression, standard deviation and coefficient of variation were determined using classical methods.
Results
The sensitivity of the flow cytometry method was first compared to CDCA. Fig. 1 shows the cytotoxic potential of two allosera of strongly immunized mice. Both sera had a cytotoxic titer of 128. At the dilution of 1/256 the alloantibodies were no longer capable of inducing more than 50% cell lysis. Alloantibodies could thus be shown up to a dilution of 1/128. Fig. 2 shows two-color immunofluorescence of target C3H cells exposed to the same anti-C3H alloserum and stained with an anti-mouse I g G F I T C and anti-L3T4 PE. The
• ANTI-C57 BI
101
.
.
4
.
.
.
.
.
16
.
.
.
.
.
64
.
.
.
.
.
.
256
.
1024
-1 (P~LLOSERUM DILU'I'ION) Fig. 1. Cytotoxicity of strong BALB anti-C3H (o) or BALB anti-C57BI (e) sera in the CDCA method. In the doubling dilution series the cytotoxic titer for both sera was 128. The data points represent the mean values of three independent determinations. Cytotoxic titer is defined as the reciprocal of the last dilution of serum that causes more than 50% cell lysis.
latter staining divides the target cells into CD4 ÷ and C D 4 - populations. In mice CD4 ÷ cells bear class I antigens only whereas C D 4 - cells contain B cells which have both class I and class II antigens. Fig. 2A shows the negative control. The difference between the mean fluorescence of CD4 ÷ and C D 4 - cells was partly caused by weak nonspecific binding of immunoglobulins to Fc receptor bearing cells and partly by the fact that 1-2% of the B cells expressed IgG on their surface. Fig. 2B and 2C show the target cells after exposure to alloserum dilutions of 1/300 and 1 / 3 respectively. The C D 4 - population appeared to consist of two groups of cells as far as immunofluorescence intensity was concerned. This was due to the fact that class II positive cells were coated with anticlass II alloantibodies as well as anti-class I antibodies. Fig. 2 D demonstrates the histograms obtained with four different alioserum dilutions. The relationship between the reciprocal of the serum dilutions and the mean fluorescence is illustrated for both allosera by a log-log plot (Fig. 3). The curves can be divided into three parts. At dilutions of 1 / 3 and 1 / 1 0 the upper part tends to flatten, probably as a result of saturation of the cell surface antigens with alloantibody. This is
76
suggested by the fact that increasing the amount anti-IgG F I T C did not alter the mean fluorescence indicating that the second-step reagent did not play a limiting role. However, increasing or decreasing the number of target cells in the presence of the same amount of alloserum resulted in either an earlier or later inflexion of the curve,
respectively (see also Fig. 4). The second and more important part of the curves between dilutions of 1 / 3 0 and 1/10,000 is represented by a linear relationship. The correlation coefficients of the regression lines of the anti-C3H and the anti-C57B1 serum were r 1 -- - 0 . 9 9 6 and r2 = - 0 . 9 9 7 respectively. Finally, at the highest dilution, when few
,qr
o g.i
m
5.7
m
19
118.0 o
m
165.5
9.8
-a
m
,
O
--me
101
10 z
(A)
Iti ~j
tO 4
"1 On
GREEN
i
, ,,..|
| lwl,.|
i
| vw,H,j
IB 2
IB!
w • wl,,,,
IB ~l
I0 4
(B)
qf
m ~e
1QO
cq fll
966.1 2
t
1196.9
%
.Q!
we
'
~IOU
' ''''"l
I01
(C)
'
' ''''"l
10 2
,,iw.i
IB 2
GREEN
•
/ I~ ~
l
i !:
'i .
1
i
gs /
'~'
...... 101~j
........
• ,IH.
I0 4
al~/l~'
(D)
"~" . . .|.~. l J ' "
11~1~ '
........
GREEN
Fig. 2. Two-color immunofluorescence analysis of 0.5 × l0 s C3H target spleen cells. The cells were incubated first with BALB anti-C3H alloserum followed by sheep anti-mouse IgG (Fc) FITC (horizontal axis). In a third step the cells were incubated with anti-L3T4 PE (vertical axis). The numbers indicated near the squares represent the mean fluorescence units along the abcissa..4: normal BALB serum 1 / 1 0 diluted; B: anti-C3H serum 1/300 diluted; C: anti-C3H serum 1 / 3 diluted; D: histograms of all target cells. Histograms 1, 2, 3 and 4 were obtained with alloserum dilutions of 1/10, 1/30, 1 / 3 0 0 and 1/3000 respectively.
77
100
d o ~rr~5~
3E
I
~
1"I 10
0 ANTI-C.3H 0.5X10
'1 101
10 2
10 3
10 4
101
10 2
10 3
(/U.LO6ERUM DI.U/IO~
(ALL(~IERUM DILUTK~1
Fig. 3. Relationship between alloserum dilutions and mean fluorescence units. The correlation coefficients of the linear regression lines between the dilutions 1/30 and 1/10,000 are shown. The negative controls + 2 SD are also represented in the lower right comer (o and o). alloantibodies were present the curves flattened again. At a d i l u t i o n of 1 / 3 0 , 0 0 0 the m e a n fluorescences of the a n t i - C 3 H a n d anti-C57B1 sera were 11.4 a n d 12.4 U respectively. T h e c o r r e s p o n d i n g negative controls _+ 1 SD d e t e r m i n e d in four indep e n d e n t samples were 8.4 _+ 0.7 a n d 8.8 + 0.8 U ( P < 0 . 0 0 1 ) . T h u s the sensitivity of the m e t h o d was very high a n d alloantibodies could be detected at s e r u m d i l u t i o n s up to 1/30,000. C o m p a r e d to the C D C A m e t h o d detection of all o a n t i b o d y b y flow c y t o m e t r y was nearly 300 times m o r e sensitive. We further investigated whether the flow cyt o m e t r y results showed good reproducibility. The
10 4 -1
Fig. 4. Influence of the number of target cells. Both curves show the relationship between a dilution series of the anti-C3H serum and the mean fluorescence units when 0.5x 106 or 0.167 × 106 target cells were used. Data points represent the measured fluorescence minus the fluorescence of the negative control.
same sera were analysed at four i n d e p e n d e n t times. A t t e n t i o n was p a i d to ensure the same flow cyt o m e t e r settings, n u m b e r of target cells a n d serum dilutions. T h e results are shown in T a b l e I. F o r n e a r l y all d i l u t i o n s the coefficients of variation (CV) were less t h a n 10%. T h a t it is i m p o r t a n t to use a c o n s t a n t n u m b e r of target cells is d e m o n strated in Fig. 4. F r o m this it could be erroneously inferred that the curves were o b t a i n e d b y two d i l u t i o n series of which o n e c o n t a i n e d three times more a l l o a n t i b o d y t h a n the other whereas o n l y the n u m b e r of target cells differed.
TABLE I REPRODUCIBILITY OF THE FLOW CYTOMETRIC DETERMINATION OF ALLOANTIBODY (Dilution) -1
3 10 30 1130 300 1000 3000 10000 30000 Negative control
Anti-C3H alloserum
Anti-C57 BL
Mean fluorescence (U)
SD
CV
SD
CV
(~)
Mean fluorescence (U)
(U)
1058.0 922.0 628.0 340.0 136.0 53.0 23.0 13.6 11.4 8.4
62 83 37 24 12 4.6 3.2 1.3 0.6 0.7
(U)
(~)
5.9 9.1 6.2 7.3 8.9 8.7 13.9 9.6 5.3 8.3
1 721 1440 1024 464 168 60 27.1 14.8 12,2 8.8
148 58 39 32 18 8.6 4 0.8 1.0 0.8
8.6 4.0 3~8 6.9 10.7 14.3 14.8 5.5 8.2 8.9
78
Discussion According to Morris et al. (1982), assays for alloantibody should be highly sensitive, capable of expressing results in a quantitative manner, reproducible, rapid and simple to interpret. The flow cytometry method we have described in the present study fulfills all of these conditions. The method was able to detect surface bound IgG antibodies in alloserum dilutions up to 1/30,000. For the same target cell:alioserum ratio, alloantibodies could only be demonstrated up to a dilution of 1/128 in CDCA. The flow cytometry method is thus much more sensitive. This higher sensitivity may not directly be relevant for organ transplantation, as has been reported in crossmatching results (Lazda et al., 1988), but it can permit distinction between tolerance or unresponsiveness and diminished responses due to immunosuppressive therapy. Apart from its high sensitivity, the method is ideal for quantitating IgG alloantibodies. Indeed, under standardized flow cytometry conditions and using a constant number of target cells, a linear relationship between the log values of mean fluorescence and the reciprocals of serum dilutions was established. This relationship permits comparisons of serum from individual animals both to each other or to a 'standard' line of a known positive serum and to draw statistical conclusions from the measured mean fluorescence units. In CDCA, a series of dilutions has to be assayed to find the cytotoxicity titer and thereby the quantity of alloantibodies, in each new serum, whereas one determination is sufficient with the flow cytometry method. This, together with shorter incubation times, makes the flow cytometry determination two to three times faster than CDCA. In addition, by using two-color immunofluorescence analysis, the method can immediately demonstrate alloantibodies against target cells bearing class I antigens only and cells bearing class I and class II antigens. A minor disadvantage of the proposed method is that it measures only surface-bound IgG antibodies. In CDCA complement fixing antibodies, both lgG and IgM, are measured. On the other hand, non-complement fixing IgG antibodies are not detected by the latter method. We deliberately
focussed on IgG alloantibodies for two reasons. First, early IgM alloantibodies will rapidly disappear and will hardly be detectable after a few weeks (Jones et al., 1987). Second, a high percentage of mouse, as well as human B cells carry IgM and IgD on their surface and labeled second-step anti-mouse Ig antibodies will therefore non-specifically stain the B cells in the target cell population, thus interfering with the alloantibody determination. This is not the case when specific anti-IgG antibodies are used, since only a minor fraction of B cells express IgG at low density. However, when evaluating IgM alloantibodies two strategies can be considered. First, B cells can be removed from the target cell population by cell separation techniques. This will, of course, complicate and prolong the method markedly. A more elegant solution is to use twocolor immunofluorescence analysis as shown in Fig. 2. Only the T cell population is then taken into account. In the present experiments we focussed on alloantibodies, but it will be clear that xenoantibodies can also be measured in a similar way. In conclusion, we have described a method to measure alloantibodies which is both quantitative and sensitive. It is of potential value in studies of transplantation, tolerance and immunoresponsiveness.
References A1-Muzairai, I.A., Innes, A., Hillis, A., Stewart, K.N., Bone, J.M.., Catto, G.R.D. and MacLeod, A.M. (1989) Renal transplantation: Cyclosporin A and antibody development after donor-specific transfusion. Kidney Int. 35, 1057. Asfar, S.K., Power, D.A., Mason, R.J., MacLet~, A.M., Simpson, J.G., Whiting, P.H., Engeset, J. and Catto, G.R.D. (1985) Prolonged survival of rat renal allografts after multiple allogeneic pregnancies; strain specificity and role of erythrocyte antibody rosette inhibiting antibodies. Clin. Sci. 69, 41. Cunningham, C., Power, D.A., lnnes, A., Lind, T. and Catto, G.R.D. (1987) Maternal alloantibody responses during early pregnancy detected by a cellular enzyme-linked immunospecific assay. Hum. lmmunol. 19, 7. Dorval, G., Welsh, K.I. and Wigzell, H. (1975) A radio-immunoassay of cellular antigens on living cells using iodinated soluble protein A from Staphylococcus aureus. J. Immunol. Methods 7, 237. Garovoy, M.R., Reinschmidt, M.A., Bigos, M., Perkins, H., Colombe, B., Feduska, N. and Salvatierra, O. (1983) Flow
79 cytometry analysis: A high technology crossmatch technique facilitating transplantation. Transplant. Proc. 15, 1939. Garovoy, M.R., Colombe, B.W., Melzer, J., Feduska, N., Shields, C., Cross, D., Amend, W., Vincenti, F., Hopper, S., Duca, R. and Salvatierra, O. (1985) Flow cytometry crossmatching for donor-specific transfusion recipients and cadaveric transplantation. Transplant. Proc. 17, 693. Hudson, L. and Hay, F.C. (1980) Antibody interaction with antigen. In: Practical Immunology. Blackwell Scientific Publications, Oxford, p. 93. Iwaki, Y., Cook, D.J., Terasaki, P.I., Lau, M., Terashita, G.Y., Danowitch, G., Fine, R., Ettenger, R., Mendez, R., Kavalich, A., Martin, D., Soderblom, R., Ward, H., Berne, T., Lieberman, E. and Strauss, F. (1987) Flow cytometric crossmatching in human cadaver kidney transplantation. Transplant. Proc. 19, 764. Johnson, A.H., Rossen, R.D. and Butler, W.T. (1972) Detection of aUoantibodies using a sensitive antiglobulin microcytotoxicity test: Identification of low levels of preformed antibodies in accelerated allograft rejection. Tissue Antigens 2, 215. Jones, M.C., Power, D.A., Cunningham, C. and Catto, G.R.D. (1988) The influence of repeated transfusions and cyclosporine on secondary alloantibody responses in inbred rats. Transplantation 45, 1094. Lazda, V.A., Pollack, R., Mozes, M.F. and Jonasson, O. (1988) The relationship between flow cytometer crossmatch results and subsequent rejection episodes in cadaver renal allograft recipients. Transplantation 45, 562. Morris, R.J. and Williams, A.F. (1975) Antigens on mouse and
rat lymphocytes recognized by rabbit antiserum against rat brain: the quantitative analysis of a xenogeneic antiserum. Eur. J. lmmunol. 5, 274. Morris, R.E., Thomas, P.T. and Hong, R. (1982) Cellular enzyme linked immunospecific assay (CELISA). I. A new method that detects antibodies to cell-surface antigens. Hum. lmmunol. 5, 1. Power, D.A., Cunningham, C. and Catto, G.R.D. (1987) The role of RTI antigen differences in semi-allogeneic rat pregnancy. Clin. Sci. 72, 37. Sachs, D.A., Winn, H.J. and Russell, P.S. (1971) The immunologic response to xenografts: Recognition of mouse H-2 histocompatiblity antigens by the rat. J. lmmunol. 107, 481. Talbot, D., Givan, A.L., Shenton, B.K., Stratton. A., Proud, G. and Taylor, R.M.R. (1988) Rapid detection of low levels of donor specific IgG by flow cytometry with single and dual color fluorescence in renal transplantation. J. Immunol. Methods 112, 279. Terasaki, P.I., Bernoco. D., Parks, M.S.. Ozturk, G. and lwaki, Y. (1978) Microdroplet testing for HLA-A,-B and -D antigens. Am. J. Clin. Pathol. 69, 103. Thistlethwaite, J.R., Buckingham, M., Stuart, J.K., Gaber, A.O., Mayes, J.T. and Stuart, F.P. (1987) T cell immunofluorescence flow cytometry cross-match results in cadaver donor renal transplantation. Transplant. Proc. 17, 722. Waer, M., Ang, K.K., Vandeputte, M. and Van der Schueren, E. (1982) Influence of overall treatment time in a fractionated total lymphoid irradiation as an immunosuppressive therapy in allogeneic bone marrow transplantation in mice. Int. J. Radiat. Oncol. Biol. Phys. 8, 1915.
