Journal of Immunological Methods, 131 (1990) 283-289 Elsevier

283

JIM05672

Highly sensitive enzyme immunoassays for antibodies to human tumor necrosis factor (TNF-ct) and lymphotoxin (TNF-fl) Herbert R. Lamche and Gi~nther R. Adolf Ernst-Boehringer-lnstitut fOr Arzneimittelforschung, Bender & Co GmbH, Department of Cell Biology, Vienna, Austria (Received 14 December 1989, revised received 2 February 1990, accepted 12 February 1990)

Two 'inverse sandwich' enzyme immunoassays (ELISAs) were developed for the detection and quantification of antibodies to human tumor necrosis factor (TNF-a) and lymphotoxin (TNF-/3), respectively. In these one-step assays, antibodies present in the sample linked antigen which had been covalently coupled to horseradish peroxidase to antigen bound to a solid phase (microtiter plates). The limits of detection of the assays were lower than those of neutralization bioassays; antibodies to TNF-a and TNF-fl being detected at concentrations as low as 2 ng/ml and 0.5 ng/ml, respectively, and no cross-reactivity was observed. The advantages of these ELISAs over other assay methods currently in use for the detection of antibodies include: (i) the convenience of the microtiter plate format and its suitability for testing large numbers of samples; (ii) the absence of radioactive tracers and precipitation steps; (iii) the high stability of the reagents; (iv) the avoidance of second antibodies and, thus, the possibility of testing samples from various species without modification of the assay and (v) the ability to detect low-affinity antibodies due to the absence of competitive reactions. The assays may be used without modification for the detection of antibodies in serum samples from both man and laboratory animals as well as in other samples such as hybridoma supernatants. Key words: Tumor necrosis factor; Lymphotoxin; Cytokine; ELISA, antibody; (Antibody)

Introduction Recombinant DNA technology has made possible the production of virtually unlimited quantities of highly purified human cytokines. The therapeutic potential of several cytokines is currently

Correspondence to: H.R. Lamche, Department of Cell Biology, Bender & Co GmbH, Dr. Boehringergasse 5-11, A-1121 Vienna, Austria. Abbreviations: BSA, bovine serum albumin; CV, coefficient of variation; ELISA, enzyme-linked immunosorbent assay; HRPO, horseradish peroxidase; MoAb, monoclonal antibody; PBS, phosphate-buffered saline; RIP, radioimmunoprecipitation assay; TNF, tumor necrosis factor.

under investigation in broad-based clinical trials, and a few products have already been registered as drugs in a number of countries. Although these proteins have identical, or very similar, structures to those present in the human body, several reports have described antibody development in patients after their administration (e.g., Panem 1984; Figlin and Itri 1988). In addition, the presence of autoantibodies to interferons has been observed in a few elderly patients before treatment, e.g., in cancer (Quesada and Gutterman, 1983) and in certain autoimmune diseases (Panem et al., 1982). In animal experiments, such as toxicological studies, repeated application of human proteins over prolonged periods of time is likely to result in

0022-17591907503.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

284 antibody induction. Antibodies may reduce ('neutralize') the biological activity of the applied protein and thus diminish or even abolish its therapeutic efficiency in patients (Von Wussow et al., 1987) and invalidate the results of an animal study. It is thus essential to monitor the development of antibodies with an appropriate assay system, which should be highly sensitive in order to detect even weak immunological responses. Methods frequently used to detect antibodies to cytokines include n e u t r a l i z a t i o n bioassays (Kawade, 1980), antibody competition immunoassays (Protzman et al., 1984), radioimmunoprecipitation assays (Chen et al., 1986; Palleroni and Trown, 1986) and indirect antibody ELISAs (Konrad et al., 1987). Each of these methods has its advantages, but also certain inherent disadvantages. Bioassays, for example, are often laborious and time-consuming procedures and show relatively poor precision. Concerns about safety and waste disposal as well as the limited stability of the labelled compound are drawbacks of immunoassays employing radioactive reagents and methods that depend on second antibodies require modification of the procedure for samples from different species. We have now developed two so-called 'inverse sandwich' ELISAs for the measurement of antibodies to two structurally related cytokines, tumor necrosis factor (TNF-a) and lymphotoxin (TNF-fl), which on the basis of in vitro data and animal studies may find applications in the therapy of malignant neoplasias and other diseases (for recent reviews, see Paul and Ruddle, 1988; Beutler and Cerami, 1989). These ' inverse sandwich' ELISAs use antigen coated onto the assay plate for the capture of antibodies in the sample, which at the same time bind soluble antigen covalently coupled to horseradish peroxidase. This assay system seems to overcome many of the disadvantages of conventional methods, allowing detection and quantitation of antibodies in a very sensitive, rapid, simple and versatile procedure.

1984) and purified to > 99% homogeneity were kindly provided by G. Bodo from the Department of Protein Chemistry at this institute. Rabbit antisera to recombinant human T N F - a and human TNF-fl were gifts by T. Chen, Genentech (San Francisco, CA) and F.J. Schneider of this department. Horseradish peroxidase ( H R P O ) was purchased from Boehringer Mannheim, bovine serum albumin (BSA) from Sigma, calf serum from Serva (Heidelberg, F.R.G.). All other reagents were of analytical grade.

Enzyme coupfing of antigens T N F - a and TNF-fl were coupled to horseradish peroxidase (HRPO) using a modification of the method described by Wilson and Nakane (1978). 16 mg H R P O dissolved in 1 ml water were activated by incubation with 200 /~1 of 100 mM sodium periodate for 20 min, followed by dialysis against 1 m M sodium acetate, pH 4.4. The antigens were dialyzed overnight against 10 mM sodium carbonate, p H 10.5. The p H of the H R P O solution was raised to 9.0-9.5 with 200 mM sodium carbonate buffer, p H 9.5. The antigen concentration was adjusted to 2 m g / m l and 1 ml of this solution was added to the activated peroxidase. After 2 h stirring at room temperature, 100 #1 sodium borohydride solution (4 m g / m l water) were added and the solution further incubated for 2 h at 4 o C. Finally, the solution of the HRPO-linked antigens was stabilized by addition of BSA to a concentration of 5 m g / m l . The conjugates were stored at 4 ° C. 'Inverse sandwich" ELISA The assay principle is illustrated in Fig. 1. Microtiter plates were coated with antigen (1 (>

,

O

Materials and methods

Reagents Recombinant human T N F - a and TNF-fl produced in E. coli (Gray et al., 1984; Pennica et al.,

antigen-coated solid p h a s e

antibody (sample)

HRPO-labeled antigen

Fig. 1. Schematic representation of the 'inverse sandwich' ELISA.

285

/~g/ml in 0.05 M sodium carbonate pH 9.6) for 1 h at room temperature (RT) or at 4 ° C overnight and washed once with deionized water. The plates were blocked with assay buffer (phosphatebuffered saline pH 7.4 (PBS) containing 5 mg BSA/ml and 0.05% Tween 20) for 1 h at RT followed by one wash step. Serial dilutions of prediluted serum samples (1/10 in PBS) were prepared in dilution buffer (PBS containing 10% fetal calf serum) directly on the plate (final volume 100 /~l/well). 50 /~1 of enzyme-labelled antigen appropriately diluted in assay buffer were added; the plates were incubated for 2 h at RT and washed three times. 100 /~1 of the substrate solution (sodium perborate, 1 mg/ml, and o-phenylenediamine, 3 mg/ml in 0.067 M potassium citrate buffer, pH 5.0) were added and the plates incubated for 20 min in the dark. The reaction was stopped by the addition of 100 /~1 2 M sulfuric acid. Finally, absorbance values were read at 492 nm (690 nm reference) in a microplate reader. Titration endpoints were determined at twice the absorbance value of a negative control (nonspecific binding; wells containing dilution buffer instead of a sample).

Antibody neutralization bioassay Biological assays measuring the cytotoxic activity of TNF-ct and TNF-fl on murine connective tissue (L-M) cells were performed essentially as described (Kramer and Carver, 1986; Kramer et al., 1986). A cytotoxic activity of 1 U / m l will by definition result in 50% lysis of the assay cells. In this assay system, we determined the specific cytotoxic activity of TNF-a as 5 × 10 7 U/mg, and the specific activity of TNF-fl as 5 x 108 U/mg; the interim reference preparation for recombinant human TNF-a, 86/659 (National Institute for Biological Standards and Control, U.K.), showed an activity of 56,000 U/ml, close to its assigned potency of 40,000 U/ml. To determine the neutralizing activity of antisera, serial two-fold dilutions of the samples were incubated with a constant amount of the respective cytokine for 1 h at 37°C; the mixture was then transferred to the assay plates containing preformed cell layers (final TNF concentration: 20 U/ml) and the residual cytotoxic activity of the samples was determined. The neutralization titer of a serum sample was

defined as the reciprocal of the dilution that reduced the cytotoxic activity to 1 U/ml.

Results

Assay optimization For optimization of plate coating the wells of the microtiter plates were coated with antigen at concentrations ranging between 0.1 and 30/xg/ml and all other assay parameters were kept constant. In the TNF-fl antibody assay, optimal responses were observed at concentrations between 0.3 and 1 /~g/ml; lower titers were obtained at lower as well as at higher concentrations (Fig. 2). Similar results were observed in the TNF-ot antibody assay. Coat antigen concentrations of 1/~g/ml were therefore chosen for further studies. For comparison of one-step, pseudo-one-step, and two-step (sequential) procedures the enzymelabelled antigen was added either immediately after the antibody-containing sample or 15 min or 1 h later. In a separate experiment, the conjugate was added after 2 h of incubation of the sample and washing of the plates (sequential incubation). As shown in Fig. 3 for the TNF-fl antibody ELISA, optimal results (highest absorbance, highest titer) were obtained when the conjugate was added immediately after the sample, whereas addition at later time points gave lower absorbances. Sequential incubation failed to generate a significant signal (data not shown). Variation of the incubation time after simultaneous addition 70000 60000 50000 40000 30000 20000 10000 0

I 1

0.1

coating

antibody

I 10 concentration

I 100 (~¢g/rnl)

Fig. 2. Optimization of the coat antigen concentration. TNF-fl at various concentrations was used to coat the assay plates; all other parameters were kept constant.

286 10"

101

E

100

1' (~

E

10-1

e~ O 10-2

'

.....

~'1

10-4

'

'

' '''"1

10-5

'

'

' .....

10-6

dilution

.1

1 10-7

2 x NSB

of a n t i s e r u m

Fig. 3. Standard curves obtained in the TNF-fl antibody ELISA. The conjugate was added 60 min (v), 15 rain (O), or immediately (11) after addition of antibodies.

of sample and conjugate showed that optimal resuits were obtained after 2 h; longer incubation did not result in further improvements (data not shown). For optimization of the conjugate the ratio of peroxidase to antigen in the conjugation reaction was varied from 1 : 1 to 32 : 1 (w/w). Fig. 4 shows that ratios of 8 : 1 or higher gave very similar results, whereas lower ratios resulted in lower absorbances. The final assay protocol selected was that described in the materials and methods section; Fig. 5 shows typical standard curves for both optimized ELISA systems. Characteristics of the "inverse sandwich' ELISAs Intra- and interassay variance. Artificial samples were generated by spiking normal rabbit

.01

........ 0.3

' ........ ' ........ ' ........ ' ........ ' 1 0.4 1 0.5 1 0.6 1 0 .7 1 0.a

dilution

of

antiserum

Fig. 5. Standard curves of the optimized ELISAs for antibodies to TNF-a (m) and TNF-fl (e). Bars indicate standard deviations from four and six parallel titrations, respectively; no bars are given when the deviations are smaller than symbol size. NSB, non-specific binding.

serum at three dose levels with rabbit antiserum to T N F - a and TNF-fl, respectively. Three replicates of each sample were run on the same plate or on three different plates in the same assay. The interassay variance was determined by running these samples in five different assays on different days; the logarithms of the observed titers, their standard deviations and coefficients of variation were calculated. The results are summarized in Table I.

TABLE I EVALUATION OF ASSAY PRECISION

101=

E

Artificial samples were generated by spiking rabbit anti-TNF-a or anti-TNF-fl at three dose levels into normal rabbit serum. Five independent assays were carried out; each sample was tested on three plates, three replicates per plate.

I00-

Sample

Mean log titer

10-1.

102

........ 10-2

I 10-3

'

' ''''"1

........ 10 -4

dilution

I 10-5

........

I 10-6

of a n t i s e r u m

Fig. 4. Optimization of the enzyme:antigen ratio in the conjugate. Symbols indicate the results obtained with HRPO:TNF-fl ratios of 4:1 (D), 8:1 (O), 16:1 (zx), or 32:1

(0).

Intra-assay variance CV (%)

Inter-assay variance CV (%)

TNF-a antibody ELISA Sample 1 2.44 Sample 2 1.90 Sample 3 1.40

1.4 2.1 3.1

4.1 4.1 4.6

TNF-fl antibody ELISA Sample 4 2.11 Sample 5 1.44 Sample 6 1.05

3.6 3.5 5.6

6.0 7.6 13.3

287 TABLE II COMPARISON OF ELISA AND BIOASSAY Antisera against recombinant human TNF-a derived from three different rabbits (samples A, B, C) were tested. Sample D: rabbit antiserum to recombinant human TNF-fl; sample E: antibodies purified from the serum of the same animal by ammonium sulphate precipitation; sample F: antiserum from a second rabbit. Sample

Titer bioassay TNF-ct antibody ELISA Sample A 71600 Sample B 16300 Sample C 45700 TNF-fl antibody ELISA Sample D 205000 Sample E 270000 Sample F 165000

Titer ELISA

Ratio of titers ELISA/bioassay

325 000 68000 190000

4.5 4.2 4.2

812000 1020000 715000

4.0 5.0 4.3

Limit of detection. Rabbit antibodies affinitypurified by chromatography on immobilized T N F - a or TNF-fl were titrated in the respective assays. The limits of detection, defined as the antibody concentrations at the titration end points, were determined as 2.1 n g / m l for antibodies to T N F - a and 0.44 n g / m l for antibodies to TNF-fl (means, two independent determinations). Comparison with the bioassay. Different antiserum samples were assayed in both a neutralization bioassay which measured the inhibition by the antibodies of the cytotoxic activity of the T N F s on murine L-M cells (Kramer et al., 1986) and in the appropriate ELISA. The results suggested that the ELISA procedure was more sensitive than the bioassay (Table II).

negative results were seen when antisera against TNF-fl were tested in the T N F - a antibody ELISA (data not shown). As the minimal detectable titer of a serum sample is 10, the assays discriminate by a factor of at least 10,000 between antibodies to the two related cytokines. Stability of antigen-HRPO conjugates. TNF-ct and TNF-fl coupled to H R P O , dissolved in PBS containing BSA (5 m g / m l ) and thimerosal (0.1 m g / m l ) were routinely stored at 4 ° C. N o loss of activity was observed after more than 1 year of storage.

Application for screening of hybridoma supernatants Screening of hybridoma supernatants usually requires the testing of a large number of supernatants within a relatively short time. To simulate such hybridoma supernatants, eight different purified murine monoclonal I g G antibodies (MoAbs) to h u m a n TNF-fl were diluted in cell culture medium. All MoAbs were originally identified by screening with a conventional antibody capture ELISA system, using rabbit antibodies to murine immunoglobulins in order to detect bound MoAbs. All of the antibodies were able to neutralize the cytotoxic activity of TNF-fl (Adolf, manuscript in preparation). As shown in Fig. 6, all antibodies were readily detected in the 'inverse sandwich' ELISA. Five of the antibodies gave positive reactions at concentrations of 10 n g / m l 101

10 0

Comparison with a radioimmunoprecipitation (RIP) assay. Affinity-purified rabbit antibodies to T N F - a were titrated in the ELISA as well as in a R I P assay performed essentially as described by Chen et al. (1986). The limit of detection of the ELISA (2.1 n g / m l ) was not significantly different from that of the R I P assay (1.9 n g / m l ; mean, two independent titrations). Specificity. Although T N F - a and T N F - f l share about 30% of their amino acid sequences, no specific signal was observed when high-titered rabbit antisera to T N F - a (Table II) were run in the TNF-fl antibody ELISA. Conversely, the same

10-1

10-2

i

10-2

i

flllll

i

10-1

i

i

llllll

I

10°

i

i

~lll~l

I

101

I

I

I~llll

I

10 2

antibody concentration (,ug/ml)

Fig. 6. Titration of monoclonal antibodies. Serial twofold dilutions of eight purified monoclonal IgG antibodies to human TNF-fl were tested in the ELISA: LTX-9 (O), LTX-19 (A), LTX-20 (O), LTX-21 (D), LTX-22(v), LTX-26(ll), LTX-27 (zx), LTX-28 (v).

288 or lower; no 'high-dose hook' effects (i.e., reduction of the signal at high antibody concentration due to limiting antigen concentration) were observed even at concentrations as high as 50 g g / m l . The ELISA was less sensitive for three of the antibodies and it was of interest that the relatively weak signals generated by MoAbs LTX-27 and LTX-28 at 5 g g / m l did not increase further at concentrations up to 50 g g / m l .

Discussion

The aim of this investigation was the development of rapid, convenient and versatile immunoassays for antibodies to T N F - a and TNF-fl, with at least the same sensitivity as the laborious and time-consuming neutralization bioassay. The format of an 'inverse sandwich' ELISA (illustrated in Fig. 1) was chosen since this type of assay has several advantages over other frequently used types of assays. (i) The ELISA format avoids the drawbacks of radiolabelled (usually iodinated) reagents, such as short half-life, radiation hazards, and expensive waste disposal. Moreover, the availability of automated equipment for handling of assays performed in microtiter plates permits the processing of large numbers of samples. (ii) Assays that detect antibodies captured by immobilized antigen using enzyme-labelled second antibodies often suffer from background problems. Furthermore serum samples from different species require different second antibodies and, eventually, different standards and controls. The latter argument also applies to precipitation assays and is of particular concern when homologous reference antisera are not available. (iii) Competitive assays may have disadvantages in the detection of low-affinity antibodies, which will be able to displace the high-affinity antibodies usually employed as assay reagents only at relatively high concentrations. An assay following a similar 'inverse sandwich' principle, but performed on beads, was used by Hermes et al. (1987) to detect antibodies to human interferon-a-2a. However, the development of this assay has not been described. Prerequisites for the development of an 'inverse sandwich' ELISA are

the availability of purified antigen and an antiserum from any species; a human reference serum containing specific antibodies, which will often be difficult or impossible to obtain, is not required. Optimization of the 'inverse sandwich' assays revealed two interesting, though not unexpected, features which result from the limiting antibody concentration employed and are not seen in conventional 'sandwich' ELISAs where antibodies are present in excess. First, variation of the antigen concentration used for coating resulted in maximal sensitivity between 0.3 and 1 g g / m l (Fig. 2); at higher solid-phase antigen concentrations, both antigen-binding sites of the antibodies preferentially bind to the immobilized antigen, resulting in reduced binding of the conjugate and, consequently, lower signals. Second, it is essential that antibodies and antigen conjugate are added simultaneously; preincubation of the coated plates with the antibodies before addition of the conjugate again results in preferential binding to the solid-phase antigen. A further interesting result was obtained in experiments using monoclonal antibodies to TNF-fl. Out of eight different MoAbs tested, two gave signals that at a concentration of approximately 5 g g / m l reached a plateau at a .relatively low absorption level; higher concentrations (up to 100 g g / m l ) did not give higher signals. It seems that the epitopes recognized by these monoclonals are inaccessible on a large proportion of antigen molecules, resulting in a condition of limiting antigen concentration. This phenomenon might arise from a preferred orientation of TNF-fl on the surface or from a preferred site of binding to the peroxidase during the coupling reaction. However, as antisera to protein antigens always contain multiple antibody species detecting multiple epitopes, this is not a concern with respect to the testing of serum samples. As the titration endpoint is defined as the titer resulting in an absorption twice as high as the assay blank, the minimal detectable dose as well as the interassay precision of the assays is dependent on the absolute value of the background absorption. Experience shows that due to variations in the quality of assay plates, substrate solution and antigen-peroxidase conjugate as well as in the washing of the plates, the background absorption is subject to some interassay variation. If

289 a s s a y p r e c i s i o n is o f p a r t i c u l a r c o n c e r n , a reference antiserum should be titrated on each assay p l a t e , a n d t h e s a m p l e titers c a n t h e n b e c o r r e c t e d a c c o r d i n g to t h e a s s i g n e d titer o f t h e r e f e r e n c e preparation. I n c o n c l u s i o n , w e feel t h a t t h e E L I S A s w e h a v e d e v e l o p e d r e p r e s e n t u s e f u l a l t e r n a t i v e s to c o n v e n t i o n a l assays for t h e d e t e r m i n a t i o n o f a n t i b o d i e s to h u m a n T N F - a a n d T N F - f l . T h e s a m e ' i n v e r s e sandwich' principle may well be applicable for the d e t e c t i o n o f a n t i b o d i e s to o t h e r a n t i g e n s , p r o v i d e d that sufficient quantities of highly purified antigen are available.

Acknowledgements W e t h a n k R. E r h a r t , C h . H a r r e r , I. S c h r i n z a n d I. S c h w e i g e r for e x c e l l e n t t e c h n i c a l a s s i s t a n c e , D r s . T. C h e n a n d F.J. S c h n e i d e r for t h e i r k i n d gifts o f a n t i s e r a , a n d Prof. G . B o d o f o r r e c o m b i n a n t TNFs. TNF antibody bioassays were kindly perf o r m e d in t h e l a b o r a t o r y o f D r . J. L i n d n e r .

References Beutler, B. and Cerami, A. (1989) The biology of cachectin/TNF a primary mediator of the host response. Ann. Rev. Immunol. 7, 625-655. Chen, A.B., Anicetti, V.R., Klassen, T.D., Berthold, W., Zahn, G., Wert, R.M., Geier, M.D. and Jones, A.J.S. (1986) A sensitive radioimmut~oprecipitation assay for the detection of antibody to recombinant human gamma-interferon: Comparison to a bioassay neutralization test. J. Interferon Res. 6, 313-320. Figlin, R.A. and Itri, L.M. (1988) Anti-interferon antibodies: A perspective.Sem. Hematol. 25, 9-15. Gray, P.W., Aggarwal, B.B., Benton, C.V., Bringman, T.S., Henzel, W.J., Lenng, D.W., Moffat, B., Ng, P., Svedersky, L.P., Palladino, M.A. and Nedwin, G.E. (1984) Cloning and expression of cDNA for human lymphotoxin, a lymphokine with tumour necrosis activity. Nature 312, 721-724. Hennes, U., Jucker, W., Fischer, E.A., Krummenacher, T., PaUeroni, A.V., Trown, P.W., Linder-Ciccolunghi, S. and

Rainisio, M. (1987) The detection of antibodies to recombinant interferon alfa-2a in human serum. J. Biol. Stand. 15, 231-244. Kawade, Y. (1980) An analysis of the neutralization reaction of interferon by antibody: A proposal on the expression of neutralization titer. J. Interferon Res. 1, 61-70. Konrad, M.W., Childs, A.L., Merigan, T.C. and Borden, E. (1987) Assessment of the antigenic response in humans to a recombinant mutant interferon beta. J. Clin. Immunol. 7, 365-375. Kramer, S. and Carver, M.E. (1986) Serum-free in vitro bioassay for the detection of tumor necrosis factor. J. Immunol. Methods 93, 201-206. Kramer, S.M., Carver, M.E. and Apperson, S.M. (1986) Comparison of TNF-a and TNF-fl cytolytic biological activities in a serum-free bioassay. Lymphokine Res. 5 (suppl. 1), S139-S143. Panem, S. (1984) Antibodies to interferon in man. In: J. Vilcek and E. DeMaeyer (Eds.), Interferon, Vol. 2. Elsevier, Amsterdam, pp. 175-183. Panem, S., Check, I.J., Henriksen, D. and Vilcek, J. (1982) Antibodies to a-interferon in a patient with systemic lupus erythematosus. J. Immunol. 129, 1-3. Palleroni, A.V. and Trown, P.W. (1986) A sensitive radioimmunoassay for detection of antibodies to recombinant human interferon-alpha-A. J. Interferon Res. 6, 705-712. Paul, N.L. and Ruddle, N. (1988) Lymphotoxin. Ann. Rev. Immunol. 6, 407-438. Pennica, D., Nedwin, G.E., Hayflick, J.S., Seeburg, P.H., Derynck, R., Palladino, M.A., Kohr, W.J., Aggarwal, B.B. and Goeddel, D.V. (1984) Human tumour necrosis factor: Precursor structure, expression and homology to lymphotoxin. Nature 312, 724-728. Protzman, W.P., Jacobs, S.L., Minnicozzi, M., Oden, E.M. and Kelsey, K. (1984). A radi0immunologic technique to screen for antibodies to alpha-2 interferon. J. Immunol. Methods 75, 317-323. Quesada, J. and Gutterman, J.U. (1983) Antibodies to human leukocyte interferon in cancer patients. Lancet i, 81-84. Von Wussow, P., Freund, M., Block, B., Diedrich, H., Poliwoda, H. and Deicher, H. (1987) Clinical significance of antiIFN-alpha antibody titres during interferon therapy. Lancet ii, 635-636. Wilson, M.B. and Nakane, P.K. (1978) Recent developments in the peroxidase method of conjugating horse radish peroxidase (HRPO) to antibodies. In: W. Knapp, K. Holubar and G. Wick (Eds.), Imrnunofluorescence and Related Staining Techniques. Elsevier-North Holland, Amsterdam p. 215.