Isotype responses of infected, virus-vaccinated and peptide-vaccinated cattle to foot-and-mouth disease virus G. Mulcahy *~;, C. Gale*, P. Robertson*, S. Iyisan*, R.D. DiMarchi* and T.R. Doel* An ELISA to measure bovine serum immunoglobulin isotypes (IgG1, IgG2, IgM and lgA ) specific for f o o t - a n d - m o u t h disease virus (FMDV) or for synthetic F M D V peptides is described. Sera from cattle infected by FMD V, vaccinated with conventional inactivated virus vaccines or vaccinated with synthetic peptides were examined using this assay. Generally IgG subclasses dominated the antibody responses of all groups after an early IgM response had waned. An exception to this pattern was seen in the case of a group of immature calves given multiple or high doses of synthetic peptide and in which levels of l g M continued to rise until the end of the experimental period. Both infected animals a n d those vaccinated with inactivated virus mounted antibody responses in which IgG1 titres tended to predominate over those of IgG2. In some infected animals,, an early lgG2 response was evident but resolution of lesions and clinical recovery did not occur until IgG1 antibody appeared in the serum some days later. In synthetic-peptide immunized animals the response was more variable but IgGl : IgG2 ratios at 21 days postvaccination were significantly lower than those of virus-vaccinated animals. It is proposed that differences in the isotype profiles induced by conventional FMD vaccines and those resulting from vaccination of cattle with synthetic FMD V peptides may in part account for the lower protective index of peptide-induced antibodies.
Keywords:Foot-and-mouthdisease; cattle; isotypes Introduction Foot-and-mouth disease (FMD) is an economically important disease of ruminants for which vaccination is an important control strategy. Vaccines currently in use are based on inactivated virus and suffer from a number of disadvantages including short duration of protection, serotype specificity, risk of incomplete virus inactivation and risk of virus escape from vaccine production plants 1. Consequently, the use of alternative vaccines has received some attention in recent years, and protection of cattle against FMDV challenge has been achieved by immunization with synthetic peptide 2. However, protection with synthetic vaccine typically requires neutralizing antibody at concentrations about tenfold higher than those which would be expected to-confer protection on animals immunized with conventional vaccines of the same serotype 2. Since circulating antibody is essential for immunity to FMD 3 it seems that a large proportion of the antibodies induced by synthetic peptide, although capable of neutralizing FMDV in vitro, are not effective in preventing viral replication and dissemination following challenge of animals with virulent virus. In an attempt to understand the functional differences between virus and peptide induced antibodies we have examined a number of their characteristics, particularly that of isotype. The functional activities of antibodies are known to be dependent on their isotype4'5, and it is clear *AFRC Institute for Animal Health, Pirbright Laboratory, Pirbright, Woking, Surrey, GU24 0NF, UK. tEli Lilly Research Laboratories, Indianapolis, Indiana 46285, USA. tTo whom correspondence should be addressed. (Received 19 June 1989; revised 13 September 1989; accepted 26 September 1989) 0264-410X/90/030249q)8 © 1990Butterworth-Heinemann Ltd
that the immunoglobulin class or sub-class induced by a particular antigen is influenced by the protective mechanisms appropriate for clearance of that particular antigen. In order to examine possible differences in the isotype profiles of infected, conventionally vaccinated or peptide-vaccinated animals we have developed an ELISA to measure immunoglobulin isotypes specific for FMDV or synthetic FMDV antigens in bovine sera.
Materials and methods Sera Samples from conventionally vaccinated cattle were drawn from stocks collected at Pirbright in annual potency tests of the UK vaccine reserve. The vaccines used were aziridine-inactivated preparations of either O1 BFS 1860 or A24 Cruzeiro strains of FMDV adjuvanted with aluminium hydroxide/saponin. Typically, vaccines were diluted 1/2, 1/10 or 1/50, and were administered subcutaneously. The animals were bled 21 days postvaccination, immediately prior to viral challenge by intradermolingual inoculation of 10 ~ IDso homologous virus. Serial serum samples from animals vaccinated conventionally at various times before challenge, or challenged without prior vaccination were also examined. These animals were challenged by exposure to donor pigs as described by Donaldson and Kitching6. Sera from peptide immunized animals represented a variety of experimental regimes with respect to peptide dose, adjuvant and number of immunizations, and were usually collected at weekly intervals from vaccination until challenge. Unless otherwise indicated, the cattle were 6-9 month old Friesian heifers, and were immunized subcutaneously.
Vaccine, Vol. 8, June 1990 249
Isotope responses of cattle to FMDV: G. Mulcahy et al.
Peptides
Results
The immunizing peptides were synthesized by solidphase methodology as described previously 2. They consisted of 38 or 40 amino-acid residues, shown numerically, based on known immunogenic regions of the VP1 surface proteins of O1 Kaufbeuren and A24 Cruzeiro virus strains. Additional terminal and linking amino-acids were included as follows:
Both anti-peptide and anti-viral assays were found to give reproducible results and at the levels of antigen used competition between serum isotypes for antigen-binding sites was found not to occur. Antigen concentration in the assay did not affect the ratios of isotypes detected until it was lowered to one-tenth (peptide) or one-fifth (virus) of levels normally used. Sera from peptide-immunized animals were found to give similar isotype profiles whether titrated against synthetic peptide or virus; a representative analysis is shown in Figure I. Except where noted, experiments with antipeptide sera were carried out with virus as the antigen, while those from virus-immunized animals were titrated in all cases against virus as their anti-peptide titres were too low to be considered reliable. Sera from 57 animals used in trials of strategic reserve vaccine were examined. All were taken 21 days postimmunization, and of these 31 (54%) had IgG1 titres threefold or more above IgG2 titres while only 15 (21%) had IgG2 titres greater than or equal to IgG1 titres (Table I). In contrast, examination of 29 sera taken at a similar time point from peptide-immunized animals revealed that 23 (79%) had IgG2 titres greater than IgG1 titres and the remainder (21%) had similar titres of both isotypes (Table 2). Tables I and 2 also show the lack of correlation between protection and virus neutralizing antibody titre in the case of peptide vaccines. The mean antigen-specific IgG1 :IgG2 ratios of the 29 sera with the highest total anti-viral IgG titre from each group were calculated and found to be significantly different (p
- -
> > >
>
>
Peptide/ adjuvant
2800 4096 1400 512 2800 90 1400 32 1024 2800 32 1024 708 1024 1400 1400 1400 1400 1024 1400 32 2800 1400
0.2 mg/FCA 0.2 mg/FCA 0.2 mg/FIA 1 mg/FCA 1 mg/FIA 0.2 mg/FIA 5 mg/FCA 0.2 mg/FIA 1 mg/FCA 1 mg/FIA 0.2 mg/FIA 1 mg/FCA 0.2 mg/FCA 5 mg/FCA 5 mg/FCA 0.2 mg/FCA 0.2 mg/FCA 1 mg/FIA 1 mg/FCA + 0.2 mg/FIA 1 mg/FCA + 0.2 mg/FIA 0.2 mg/FIA 1 mg/FIA 0.2 mg/FIA
720 40 180 120 510 1400
1 mg/Oil 1 mg/Oil 1 rag/Oil 1 mg/Oil 1 mg/Oil 1 rag/Oil
aResistance to virus challenge; "The ELISA titres were obtained using sera taken 21-days postimmunization whereas the VN tests were performed on sera drawn immediately prior to challenge on days 28 to 53 (depending on protocol). In the case of animals PH74-PH76 and RA82-RA84 the second vaccination was given on day 21, and therefore had no bearing on ELISA titres. Animals RA82-RA87 received 40-residue peptide based on the sequence of A24 Cruzeiro virus; all others were vaccinated with the equivalent 01 Kaufbeuren peptide; CVirus neutralization titre
16 14 12 O t~
10
E
6
Z
4
0 0.03-O.06
O.5-1.0 2 . 0 - 4 . 0
>64
IgG I:lgG2 Figure 2 Frequency distributions of IgGl:lgG2 ratios of 21 day sera from virus-vaccinated and peptide-vaccinated animals. The lighter shading represents peptide-immunized animals, the medium shading virus-vaccinates and the dark shading the area of overlap between the two populations
28 days, after an initial dose of 0.2 mg peptide in oil. In addition to this initial immunization procedure, all animals received a final injection of 0.2 mg peptide in oil adjuvant. The ELISA results of anti-viral isotypes are shown for a representative animal from each group in Figure 6. Continuous infusion of peptide induced a strong IgG2 response and a somewhat weaker IgG1 response which peaked at days 29 and 36, respectively. The IgM
252
V a c c i n e , V o l . 8, J u n e 1990
response was marked and was still rising at the end of the experiment. Administration of three doses of peptide also induced this atypical IgM response and in addition produced relatively low IgG1 and IgG2 titres, suggesting inefficient class switching and an inability to mount an effective anamnestic response. The group receiving a single dose of peptide showed distinct peaks of IgGl, IgG2 and IgM, the dominant immunoglobulin being IgG2. It was not possible to challenge any of the cattle in this experiment. The influence of the route of immunization on isotype induction was examined in an experiment in which three groups of three six-month-old calves received 1 mg doses of O1K peptide in incomplete oil adjuvant by subcutaneous, intramuscular and intradermal routes, respectively. The results obtained with a representative animal from each group are shown in Figure 7. Animals immunized subcutaneously had a strong antibody response at 21 days. IgG1 was the predominant isotype but high titres of IgG2 and IgM and moderate titres of IgA were also observed. The titres comprising the secondary response were lower than those recorded after primary vaccination suggesting that a true anamnestic response had not occurred. In contrast, animals vaccinated by intramuscular and intradermal routes mounted strong anamnestic IgG1 responses. However, in terms of overall titres the group vaccinated intramuscularly responded more favourably than the group immunized intradermally. The cattle in the experiment were not challenged.
Isotope responses of cattle to FMDV: G. Mulcahy
et al.
5000 (~
b
4000
T
3000
T
2000 • ".• /
•/ •J
L %
1000 , _[ . . . . . . . . . .
0
_•s L .
.
.
.
.
I"-
4000
T
3000
2000
lOOO %.
~'"','"',"",'"L".,...',...,-., o
14
0
3
5
7
9
11
13
21
7
1
Time pre-and postimmunizotion
4
6
8
................. 10
12
14
28
(days)
Figure 3 Anti-FMDV isotype profiles of four cattle immunized with conventional inactivated O1 BFS 1860 vaccine. Animals were challenged 14 (a) and (c) 7 (b) or 4 (d) days after immunization. (The time of challenge is shown by the symbols on each panel.) None of the animals developed clinical signs following challenge. ( ) tgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . . ) IgA
Discussion Measurement of antigen-specific immunoglobulin classes and sub-classes by ELISA requires anti-isotype reagents which do not cross-react significantly but which are nonetheless sufficiently sensitive to facilitate workable assays. An ELISA system to measure anti-FMDV isotypes in bovine sera using polyclonal anti-immunoglobulin reagents has been described 9. However, in our hands similar reagents have proven cross-reactive to a degree which makes the assays unreliable and difficult to interpret. Therefore we developed a test system based on monoclonal antibodies of proven reactivity 7 and have found it to be satisfactory with respect to sensitivity and specificity. Our chosen test system uses antigen either bound directly to the plate (in the case of peptide) or via a capture antibody (in the case of virus), with which test serum is reacted and detected by means of isotype-specific monoclonal antibody and anti-mouse conjugate. A potential problem with this system is inter-isotype competition for antigen binding sites which can influence the results depending on the relative amounts and affinities of the various isotypes present in a given serum. An alternative system which avoids this problem is to coat the plates with the isotype-specific monoclonal antibody, and then add, sequentially, test serum, antigen
and antigen-specific conjugate. However, in this case all of the antibodies of a particular isotype will be bound, only a proportion of which will be specific for the relevant antigen and this limits considerably the sensitivity of the assay. In our experience the absorbance values obtained using this latter system were too low to provide reliable analysis of FMDV-specific isotype profiles and we therefore used the former type of assay, providing antigen in excess to prevent distortion of results by competition between isotypes. Although our system does not allow quantitative analysis of the levels of FMDV-specific immunoglobulin classes/sub-classes in a particular sample, it does permit valid comparisons to be made of the relative levels between one sample and another. We have shown that antibodies induced in cattle by synthetic peptides based on immunogenic sequences of VP1 and by inactivated viral vaccines have a comparable capacity to neutralize virus in vitro, but differ in their capability to confer protection to virus challenge in vivo (this paper and Ref. 2). This is so whether the peptides are based on sequences of Ox or A24 serotypes, the latter characteristically giving lower neutralizing titres than the former 3. We propose that this difference can be accounted for in part by the differential induction of antibody isotypes by soluble peptide or particulate viral antigens. Lack of correlation between neutralizing
Vaccine, Vol. 8, June 1990
253
Isotope responses of caWe to FMDV: G. Mulcahy et al. 10000
8000
6000
/
4000
2000
r
s*
....
J.
•
....
I-
d 8000
6000 /
•
4000 Sl f
~
i •
•/l
l
2000
I ... I -:.
0
3
5
• ___r
7
~
9
i " - - ~ . ~ , ~ ' :
11
....
13
21
0
3
5
7
9
11
13
21
Time postinfectlon Idays) Figure 4 Anti-FMDV isotypes of four cattle infected with O, BFS virus without prior immunization, all of which developed clinical signs and viraemia. In each case resolution of lesions and clearance of virus from the circulation coincided with the appearance of IgG1 in serum, even though IgG2 titres were apparent earlier than this in three of the four animals. Vesicles and circulating virus had disappeared by day 13 (a,c) 12 (b) and 11 (d). () IgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . . ) IgA
antibodies and protection has also been shown for anti-FMDV monoclonals *° which can protect suckling mice at concentrations too low to neutralize in vitro. Although other workers* ~.~2 have reported a correlation between neutralization titre and protection in animals immunized with FMDV peptides, the data refers in large measure to guinea-pigs-and pigs immunized with fusion proteins or peptides otherwise linked to a carrier, which are likely to elicit an entirely different response from that seen with our chemically defined, purely synthetic, peptide in cattle. The failure of high titres of neutralizing antibody to protect cattle immunized with synthetic peptides cannot be explained in terms of lack of homology among the SNT virus, cattle challenge virus and synthetic peptide, since typing with a panel of monoclonal antibodies has shown that both viruses react identically in the regions of VP1 corresponding to the peptide used for immunization (A. Samuels, personal communication). Furthermore, conventional O,BFS 1860 vaccines (from the MAFF strategic reserve) are tested with the same challenge virus and here the correlation is satisfactory. Development and maturation of a humoral immune response appropriate for a particular antigen involves processes of clonal selection and class switching that direct activity towards production of antibodies of high affinity and relevant isotype. These are progressive
254
Vaccine, Vol. 8, June 1990
processes whose effects are more obvious after second and subsequent exposure to antigen than after primary immunizations. Therefore, any differences apparent between the isotype profiles generated in response to FMDV and synthetic FMDV antigens should be most obvious in the sera of animals undergoing secondary or tertiary exposure to the respective antigens. It has been suggested that opsonization and enhanced phagocytosis of virus is more relevant to virus clearance in vivo than neutralization* 3. If this is the case, the protective capacity of different anti-FMDV antibody isotypes can be interpreted in terms of their functional activity and interaction with cells of the immune system such as macrophages. Knowledge of the functional activities of bovine immunoglobulin classes and sub-classes is relatively poor in comparison with the published information available for mouse and man. Complement activation is thought to be associated with bovine IgG1 but not IgG24, and there is evidence that IgG1 interacts with Fc receptors on bovine macrophages whereas IgG2 fulfils this role in the case of neutrophils .4. It has been shown that a predominantly IgG2 response occurs in mastitic infection of cows with Streptococcus agalactiae* 5, further suggesting that this isotype is important in defence against pathogens engulfed and killed primarily by granulocytes.
Isotope responses of cattle to FMDV: G. Mulcahy et al. 8"00
800
b
0 e
600
600
T
o/'/
400
400
ii # 200
Illl Ill
200
,," 0
U F-
6
14
7
21
28
;55
42
0
14oo
7
14
21
28
2560
d
C
12oo
2000 1ooo
sP o
1500
8oo
e
/ /
600
1000
a a s
400
/
500
200
/
e#
o
14
7
21
28
35 42 Time postvaccinatlon (days)
0
/ ........
~)
.. = ....
7
.14
.. • ..... ,...
.....
21
28
Figure 5 Anti-virus isotype profiles of four cattle given synthetic peptide based on the sequence of A=, Cruzeiro virus. Two of the animals (RA82 (a) and RA86 (b)) resisted virus challenge. These differ from the two unprotected animals (RA84 (c) and RA87 (d)) in having sharply rising IgG1 titres at the time of challenge. Second immunizations are indicated by the arrows in (a) and (c). ( ) IgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . --) IgA 20000
16000 O
15000
1
/"lt l i
/
.-'
•
6000
I-'
b
...."
5000 4000
....
'10000
8000
,t
3000
. IL...' /
5000
2000
11'
4000
1000
,,,. 0
8
c
12000
29
36
51
0
8
29
36
51
0
0 8 29 36 51 57 Time after initial vaccination (days] Rgure 6 Anti-peptide isotype profiles of three cattle immunized with a continuous infusion (a), double (b), or multiple (c) doses of synthetic FMDV peptide. Booster immunizations are indicated by the arrows, the period of continuous infusion in (a) by the horizontal line. ( ) IgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . . ) IgA 6000 I
14ooo
5000
12000
I
8000
3000
,
6000 2000
""
1000 0
'"
7
14
, ' ';""C " l,
c
2000
10000
4000 =
2500
b 1500 1000
4000 ,
21 2 8 3 5 4 2 4 9
500
2000 0
0
"7
14 21 2 8 3 5 4 2 4 9
0
/" 0
.
7
-~.~L'.'~ "~'
14 21 2 8 3 5 4 2 4 9
Time p o s t v a c c i n a t i o n (days) Rgure 7 Anti-peptide isotype profiles of three cattle given peptide in an oil adjuvant subcutaneously (a), intramuscularly (b) and intradermally (c). Second immunizations are indicated by the arrows. ( ) IgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . . . ) IgA
V a c c i n e , Vol. 8, J u n e 1990
255
Isotope r e s p o n s e s of cattle to FMDV: G. M u l c a h y et al.
If it is the case that macrophages play a critical role in the clearance of FMDV in cattle through the processes of opsonization, attachment to Fc and/or complement receptors and phagocytosis then the importance of antibody isotype to induction of protective immunity is clear. It has been shown that antibody isotype (i.e. constant region structure) can be influenced by the macromolecular form in which a B-cell epitope is presented 16, and isotype restriction of antibody responses to viruses in other species has been documented 17. The effect of antibodies ofisotypes unable to effect opsonization and macrophage phagocytosis on the immune status of a peptide-immunized animal clearly depends on whether those antibodies simply block access of 'useful' antibody molecules to the virus or can themselves stop virus replication by secondary, albeit less efficient, means. It seems likely that both effects may occur, the balance depending on the relative numbers and affinities of each isotype present. It is clear, also, that the direction of an immune response towards production of one isotype of antibody is a complex process which depends not only on the macromolecular form of the antigen presented, but also on the adjuvant in which it is presented and on the presence of T-cell epitopes on the antigen. Indeed, it has recently been_demonstrated is that individual B-cell epitopes within a peptide can vary with respect to immunoglobulin sub-class restriction, thus there are numerous parameters that can influence the immunoglobulin sub-classes generated during an immune response. In protection data obtained with peptide-immunized cattle we were not abte to discern patterns of isotype induction that are typical of protected or non-protected animals. However, there was an apparent tendency for peptide-immunized animals to be protected if they had rising IgG1 titres at the time of challenge, regardless of IgG2 levels. One animal (RA87) which had very high total antibody titres and which, surprisingly, remained susceptible to FMDV challenge was found on analysis of isotype profile to have mounted a response consisting almost entirely of IgG2 with negligible levels of other isotypes. This, together with the early non-protective IgG2 response in infected animals gives preliminary indications of the relevance of IgG1 to protection against FMDV in cattle. Peptide-immunized animals in which IgG2 was the predominant isotype were protected only if they had extremely high titres, indicating that virus clearance in the case of these animals may have been by secondary mechanisms.
Conclusions Cattle immunized with virulent or inactivated FMDV mount a humoral immune response in which the isotype profile differs significantly from that which occurs in cattle immunized with a synthetic peptide. The fact that an IgGl response is favoured in virus-immunized, and IgG2 in peptide-immunized animals may partially explain the lower protective capacity of these anti-peptide antibodies against FMDV challenge. Additional factors which have yet to be considered include individual animal variation in response to either type of antigen, reactivity of anti-isotypic antibodies with allotypic determinants and the effect of vaccine adjuvants or additives on class-switching and resultant antibody isotype profile. Elucidation of these parameters together with experiments to ascertain the mechanisms of FMDV clearance by
256
Vaccine, Vol. 8, June 1990
antibody interaction with cells of the immune and reticuloendothelial systems should in the future allow synthetic peptide vaccines against FMDV to be engineered towards the optimal protective response.
Acknowledgements The authors would like to thank Dr D. van Zaane, CDI, Lelystad, The Netherlands, for his gift of monoclonal anti-isotype antibodies, and Mrs Ann Boddy for typing the manuscript. The technical assistance of Mr Donald White is greatly appreciated.
References 1 Doel, T.R. Prospects for improved foot and mouth disease vaccines. Vaccine 1985, 3, 35-36 2 DiMarchi, R., Brooke, G., Gale, C., Cracknell, V., Doel, T.R. and Mowat, N. Protection of cattle against foot-and-mouth disease by a synthetic peptide. Science 1986, 232, 639-641 3 Van Bekkum, J.G. Correlation between serum antibody level and protection against challenge with FMD. Session of the Research Group of the Standing Technical Committee of the European Commission for the Control of Foot and Mouth Disease, Brescia, Italy, FAO, 1969 4 Sellei, J. Interaction of bovine immunoglobulins with complement. Comp. /mmunol. Microbio/. Infect. Dis. 1984, 1(](2), 93-98 5 McGuire, T.C., Musuoke, A.J. and Kurtti, T. Functional properties of bovine IgG1 and IgG2; Interaction with complement, macrophages, neutrophils and skin. Immunology 1979, 38, 249-256 6 Donaldson, A.I. and Kitching, R.P. Transmission of foot-and-mouth disease by vaccinated cattle following natural challenge. Res. Vet. Sci. 1989,40, 9-14 7 Van Zaane, D. and Ijzerman, J. Monoclonal antibodies against bovine immunoglobulins and their use in isotype-specific ELISAs for rotavirus antibody. J. Immunol. Methods 1984, 72, 427-441 8 Golding, S.M., Hedger, R.S., Talbot, P. and Watson, J. Radial immunodiffusion and serum neutralisation techniques for the assay of antibodies to swine vesicular disease. Res. Vet. Sci. 1976, 20, 142-147 9 Abu-Elzein, E.M. and Crowther, J.R. Detection and quantification of IgM, IgA, IgG1 and IgG2 antibodies against foot-and-mouth disease virus from bovine sera using an enzyme-linked immunosorbent assay. J. Hyg. (Camb.) 1981, 86, 79-85 10 McCullough, K.C., Crowther, J.R., Butcher, R.N., Carpenter, W.L., Brocchi, E., Capucci, L. and DeSimone, F. Immune protection against foot-and-mouth disease virus studied using virus neutralising and non-neutralising concentrations of monoclonal antibodies. Immunology 1986, 58, 421-428 11 Bittle, J.L., Houghton, R.A., Alexander, H., Shinnick, T.A., Sutcliffe, J.G. and Lerner, R.A. Protection against foot-and-mouth disease by immunization with a chemically synthesized peptide predicted from the viral nucleotide sequence. Nature 1982, ~ , 30-33 12 Broekhuijsen, M.P., Von Rijn, J.M.M., Blom, A.J.M, Pouwells, P.H., Enger-Valk, B.E., Brown, F. and Francis, M.J. Fusion proteins with multiple copies of the major antigenic determinant of foot-and-mouth disease virus protect both the natural host and laboratory animals. J. Gen. Virol. 1987, 68, 3137-3143 .13 McCullough, K.C., Parkinson, D. and Crowther, J.R. Opsonisationenhanced phagocytosis of foot-and-mouth disease virus. Immunology 1988, 65(2), 187-192 14 Howard, C.J., Taylor, G. and Brownlie, J. Surface receptors for immunoglobulin on bovine polymorphonuclear leucocytes and macrophages. Res. Vet. Sci. 1980, 20, 128-130 15 Mackie, D.P. and Logan, E.F. Changes in immunoglobulin levels in whey during experimental Streptococcus agalactiae mastitis. Res. VeL Sci. 1986, 40, 183-188 16 Fish, S. and Manser, T. Influence of the macromolecular form of a B cell epitope on the expression of antibody variable and constant region structure. J. Exp. Med. 1987, 166, 711-724 17 Courtelier, J.P., van der Logt, J.T.M., Hessen, F.W.A., Warner, G. and Van Snick, J. IgG2a restriction of murine antibodies elicited by viral infections. J. Exp. Med. 1987, 165, 64-68 18 Mathiesen, T., Brolinden, P.A., Rosen, J. and Wahren, B. Mapping of IgG subclass and T-cell epitopes on HIV proteins with synthetic peptides. /mmunology 1989, 67, 453-459
Keywords:Foot-and-mouthdisease; cattle; isotypes Introduction Foot-and-mouth disease (FMD) is an economically important disease of ruminants for which vaccination is an important control strategy. Vaccines currently in use are based on inactivated virus and suffer from a number of disadvantages including short duration of protection, serotype specificity, risk of incomplete virus inactivation and risk of virus escape from vaccine production plants 1. Consequently, the use of alternative vaccines has received some attention in recent years, and protection of cattle against FMDV challenge has been achieved by immunization with synthetic peptide 2. However, protection with synthetic vaccine typically requires neutralizing antibody at concentrations about tenfold higher than those which would be expected to-confer protection on animals immunized with conventional vaccines of the same serotype 2. Since circulating antibody is essential for immunity to FMD 3 it seems that a large proportion of the antibodies induced by synthetic peptide, although capable of neutralizing FMDV in vitro, are not effective in preventing viral replication and dissemination following challenge of animals with virulent virus. In an attempt to understand the functional differences between virus and peptide induced antibodies we have examined a number of their characteristics, particularly that of isotype. The functional activities of antibodies are known to be dependent on their isotype4'5, and it is clear *AFRC Institute for Animal Health, Pirbright Laboratory, Pirbright, Woking, Surrey, GU24 0NF, UK. tEli Lilly Research Laboratories, Indianapolis, Indiana 46285, USA. tTo whom correspondence should be addressed. (Received 19 June 1989; revised 13 September 1989; accepted 26 September 1989) 0264-410X/90/030249q)8 © 1990Butterworth-Heinemann Ltd
that the immunoglobulin class or sub-class induced by a particular antigen is influenced by the protective mechanisms appropriate for clearance of that particular antigen. In order to examine possible differences in the isotype profiles of infected, conventionally vaccinated or peptide-vaccinated animals we have developed an ELISA to measure immunoglobulin isotypes specific for FMDV or synthetic FMDV antigens in bovine sera.
Materials and methods Sera Samples from conventionally vaccinated cattle were drawn from stocks collected at Pirbright in annual potency tests of the UK vaccine reserve. The vaccines used were aziridine-inactivated preparations of either O1 BFS 1860 or A24 Cruzeiro strains of FMDV adjuvanted with aluminium hydroxide/saponin. Typically, vaccines were diluted 1/2, 1/10 or 1/50, and were administered subcutaneously. The animals were bled 21 days postvaccination, immediately prior to viral challenge by intradermolingual inoculation of 10 ~ IDso homologous virus. Serial serum samples from animals vaccinated conventionally at various times before challenge, or challenged without prior vaccination were also examined. These animals were challenged by exposure to donor pigs as described by Donaldson and Kitching6. Sera from peptide immunized animals represented a variety of experimental regimes with respect to peptide dose, adjuvant and number of immunizations, and were usually collected at weekly intervals from vaccination until challenge. Unless otherwise indicated, the cattle were 6-9 month old Friesian heifers, and were immunized subcutaneously.
Vaccine, Vol. 8, June 1990 249
Isotope responses of cattle to FMDV: G. Mulcahy et al.
Peptides
Results
The immunizing peptides were synthesized by solidphase methodology as described previously 2. They consisted of 38 or 40 amino-acid residues, shown numerically, based on known immunogenic regions of the VP1 surface proteins of O1 Kaufbeuren and A24 Cruzeiro virus strains. Additional terminal and linking amino-acids were included as follows:
Both anti-peptide and anti-viral assays were found to give reproducible results and at the levels of antigen used competition between serum isotypes for antigen-binding sites was found not to occur. Antigen concentration in the assay did not affect the ratios of isotypes detected until it was lowered to one-tenth (peptide) or one-fifth (virus) of levels normally used. Sera from peptide-immunized animals were found to give similar isotype profiles whether titrated against synthetic peptide or virus; a representative analysis is shown in Figure I. Except where noted, experiments with antipeptide sera were carried out with virus as the antigen, while those from virus-immunized animals were titrated in all cases against virus as their anti-peptide titres were too low to be considered reliable. Sera from 57 animals used in trials of strategic reserve vaccine were examined. All were taken 21 days postimmunization, and of these 31 (54%) had IgG1 titres threefold or more above IgG2 titres while only 15 (21%) had IgG2 titres greater than or equal to IgG1 titres (Table I). In contrast, examination of 29 sera taken at a similar time point from peptide-immunized animals revealed that 23 (79%) had IgG2 titres greater than IgG1 titres and the remainder (21%) had similar titres of both isotypes (Table 2). Tables I and 2 also show the lack of correlation between protection and virus neutralizing antibody titre in the case of peptide vaccines. The mean antigen-specific IgG1 :IgG2 ratios of the 29 sera with the highest total anti-viral IgG titre from each group were calculated and found to be significantly different (p
- -
> > >
>
>
Peptide/ adjuvant
2800 4096 1400 512 2800 90 1400 32 1024 2800 32 1024 708 1024 1400 1400 1400 1400 1024 1400 32 2800 1400
0.2 mg/FCA 0.2 mg/FCA 0.2 mg/FIA 1 mg/FCA 1 mg/FIA 0.2 mg/FIA 5 mg/FCA 0.2 mg/FIA 1 mg/FCA 1 mg/FIA 0.2 mg/FIA 1 mg/FCA 0.2 mg/FCA 5 mg/FCA 5 mg/FCA 0.2 mg/FCA 0.2 mg/FCA 1 mg/FIA 1 mg/FCA + 0.2 mg/FIA 1 mg/FCA + 0.2 mg/FIA 0.2 mg/FIA 1 mg/FIA 0.2 mg/FIA
720 40 180 120 510 1400
1 mg/Oil 1 mg/Oil 1 rag/Oil 1 mg/Oil 1 mg/Oil 1 rag/Oil
aResistance to virus challenge; "The ELISA titres were obtained using sera taken 21-days postimmunization whereas the VN tests were performed on sera drawn immediately prior to challenge on days 28 to 53 (depending on protocol). In the case of animals PH74-PH76 and RA82-RA84 the second vaccination was given on day 21, and therefore had no bearing on ELISA titres. Animals RA82-RA87 received 40-residue peptide based on the sequence of A24 Cruzeiro virus; all others were vaccinated with the equivalent 01 Kaufbeuren peptide; CVirus neutralization titre
16 14 12 O t~
10
E
6
Z
4
0 0.03-O.06
O.5-1.0 2 . 0 - 4 . 0
>64
IgG I:lgG2 Figure 2 Frequency distributions of IgGl:lgG2 ratios of 21 day sera from virus-vaccinated and peptide-vaccinated animals. The lighter shading represents peptide-immunized animals, the medium shading virus-vaccinates and the dark shading the area of overlap between the two populations
28 days, after an initial dose of 0.2 mg peptide in oil. In addition to this initial immunization procedure, all animals received a final injection of 0.2 mg peptide in oil adjuvant. The ELISA results of anti-viral isotypes are shown for a representative animal from each group in Figure 6. Continuous infusion of peptide induced a strong IgG2 response and a somewhat weaker IgG1 response which peaked at days 29 and 36, respectively. The IgM
252
V a c c i n e , V o l . 8, J u n e 1990
response was marked and was still rising at the end of the experiment. Administration of three doses of peptide also induced this atypical IgM response and in addition produced relatively low IgG1 and IgG2 titres, suggesting inefficient class switching and an inability to mount an effective anamnestic response. The group receiving a single dose of peptide showed distinct peaks of IgGl, IgG2 and IgM, the dominant immunoglobulin being IgG2. It was not possible to challenge any of the cattle in this experiment. The influence of the route of immunization on isotype induction was examined in an experiment in which three groups of three six-month-old calves received 1 mg doses of O1K peptide in incomplete oil adjuvant by subcutaneous, intramuscular and intradermal routes, respectively. The results obtained with a representative animal from each group are shown in Figure 7. Animals immunized subcutaneously had a strong antibody response at 21 days. IgG1 was the predominant isotype but high titres of IgG2 and IgM and moderate titres of IgA were also observed. The titres comprising the secondary response were lower than those recorded after primary vaccination suggesting that a true anamnestic response had not occurred. In contrast, animals vaccinated by intramuscular and intradermal routes mounted strong anamnestic IgG1 responses. However, in terms of overall titres the group vaccinated intramuscularly responded more favourably than the group immunized intradermally. The cattle in the experiment were not challenged.
Isotope responses of cattle to FMDV: G. Mulcahy
et al.
5000 (~
b
4000
T
3000
T
2000 • ".• /
•/ •J
L %
1000 , _[ . . . . . . . . . .
0
_•s L .
.
.
.
.
I"-
4000
T
3000
2000
lOOO %.
~'"','"',"",'"L".,...',...,-., o
14
0
3
5
7
9
11
13
21
7
1
Time pre-and postimmunizotion
4
6
8
................. 10
12
14
28
(days)
Figure 3 Anti-FMDV isotype profiles of four cattle immunized with conventional inactivated O1 BFS 1860 vaccine. Animals were challenged 14 (a) and (c) 7 (b) or 4 (d) days after immunization. (The time of challenge is shown by the symbols on each panel.) None of the animals developed clinical signs following challenge. ( ) tgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . . ) IgA
Discussion Measurement of antigen-specific immunoglobulin classes and sub-classes by ELISA requires anti-isotype reagents which do not cross-react significantly but which are nonetheless sufficiently sensitive to facilitate workable assays. An ELISA system to measure anti-FMDV isotypes in bovine sera using polyclonal anti-immunoglobulin reagents has been described 9. However, in our hands similar reagents have proven cross-reactive to a degree which makes the assays unreliable and difficult to interpret. Therefore we developed a test system based on monoclonal antibodies of proven reactivity 7 and have found it to be satisfactory with respect to sensitivity and specificity. Our chosen test system uses antigen either bound directly to the plate (in the case of peptide) or via a capture antibody (in the case of virus), with which test serum is reacted and detected by means of isotype-specific monoclonal antibody and anti-mouse conjugate. A potential problem with this system is inter-isotype competition for antigen binding sites which can influence the results depending on the relative amounts and affinities of the various isotypes present in a given serum. An alternative system which avoids this problem is to coat the plates with the isotype-specific monoclonal antibody, and then add, sequentially, test serum, antigen
and antigen-specific conjugate. However, in this case all of the antibodies of a particular isotype will be bound, only a proportion of which will be specific for the relevant antigen and this limits considerably the sensitivity of the assay. In our experience the absorbance values obtained using this latter system were too low to provide reliable analysis of FMDV-specific isotype profiles and we therefore used the former type of assay, providing antigen in excess to prevent distortion of results by competition between isotypes. Although our system does not allow quantitative analysis of the levels of FMDV-specific immunoglobulin classes/sub-classes in a particular sample, it does permit valid comparisons to be made of the relative levels between one sample and another. We have shown that antibodies induced in cattle by synthetic peptides based on immunogenic sequences of VP1 and by inactivated viral vaccines have a comparable capacity to neutralize virus in vitro, but differ in their capability to confer protection to virus challenge in vivo (this paper and Ref. 2). This is so whether the peptides are based on sequences of Ox or A24 serotypes, the latter characteristically giving lower neutralizing titres than the former 3. We propose that this difference can be accounted for in part by the differential induction of antibody isotypes by soluble peptide or particulate viral antigens. Lack of correlation between neutralizing
Vaccine, Vol. 8, June 1990
253
Isotope responses of caWe to FMDV: G. Mulcahy et al. 10000
8000
6000
/
4000
2000
r
s*
....
J.
•
....
I-
d 8000
6000 /
•
4000 Sl f
~
i •
•/l
l
2000
I ... I -:.
0
3
5
• ___r
7
~
9
i " - - ~ . ~ , ~ ' :
11
....
13
21
0
3
5
7
9
11
13
21
Time postinfectlon Idays) Figure 4 Anti-FMDV isotypes of four cattle infected with O, BFS virus without prior immunization, all of which developed clinical signs and viraemia. In each case resolution of lesions and clearance of virus from the circulation coincided with the appearance of IgG1 in serum, even though IgG2 titres were apparent earlier than this in three of the four animals. Vesicles and circulating virus had disappeared by day 13 (a,c) 12 (b) and 11 (d). () IgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . . ) IgA
antibodies and protection has also been shown for anti-FMDV monoclonals *° which can protect suckling mice at concentrations too low to neutralize in vitro. Although other workers* ~.~2 have reported a correlation between neutralization titre and protection in animals immunized with FMDV peptides, the data refers in large measure to guinea-pigs-and pigs immunized with fusion proteins or peptides otherwise linked to a carrier, which are likely to elicit an entirely different response from that seen with our chemically defined, purely synthetic, peptide in cattle. The failure of high titres of neutralizing antibody to protect cattle immunized with synthetic peptides cannot be explained in terms of lack of homology among the SNT virus, cattle challenge virus and synthetic peptide, since typing with a panel of monoclonal antibodies has shown that both viruses react identically in the regions of VP1 corresponding to the peptide used for immunization (A. Samuels, personal communication). Furthermore, conventional O,BFS 1860 vaccines (from the MAFF strategic reserve) are tested with the same challenge virus and here the correlation is satisfactory. Development and maturation of a humoral immune response appropriate for a particular antigen involves processes of clonal selection and class switching that direct activity towards production of antibodies of high affinity and relevant isotype. These are progressive
254
Vaccine, Vol. 8, June 1990
processes whose effects are more obvious after second and subsequent exposure to antigen than after primary immunizations. Therefore, any differences apparent between the isotype profiles generated in response to FMDV and synthetic FMDV antigens should be most obvious in the sera of animals undergoing secondary or tertiary exposure to the respective antigens. It has been suggested that opsonization and enhanced phagocytosis of virus is more relevant to virus clearance in vivo than neutralization* 3. If this is the case, the protective capacity of different anti-FMDV antibody isotypes can be interpreted in terms of their functional activity and interaction with cells of the immune system such as macrophages. Knowledge of the functional activities of bovine immunoglobulin classes and sub-classes is relatively poor in comparison with the published information available for mouse and man. Complement activation is thought to be associated with bovine IgG1 but not IgG24, and there is evidence that IgG1 interacts with Fc receptors on bovine macrophages whereas IgG2 fulfils this role in the case of neutrophils .4. It has been shown that a predominantly IgG2 response occurs in mastitic infection of cows with Streptococcus agalactiae* 5, further suggesting that this isotype is important in defence against pathogens engulfed and killed primarily by granulocytes.
Isotope responses of cattle to FMDV: G. Mulcahy et al. 8"00
800
b
0 e
600
600
T
o/'/
400
400
ii # 200
Illl Ill
200
,," 0
U F-
6
14
7
21
28
;55
42
0
14oo
7
14
21
28
2560
d
C
12oo
2000 1ooo
sP o
1500
8oo
e
/ /
600
1000
a a s
400
/
500
200
/
e#
o
14
7
21
28
35 42 Time postvaccinatlon (days)
0
/ ........
~)
.. = ....
7
.14
.. • ..... ,...
.....
21
28
Figure 5 Anti-virus isotype profiles of four cattle given synthetic peptide based on the sequence of A=, Cruzeiro virus. Two of the animals (RA82 (a) and RA86 (b)) resisted virus challenge. These differ from the two unprotected animals (RA84 (c) and RA87 (d)) in having sharply rising IgG1 titres at the time of challenge. Second immunizations are indicated by the arrows in (a) and (c). ( ) IgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . --) IgA 20000
16000 O
15000
1
/"lt l i
/
.-'
•
6000
I-'
b
...."
5000 4000
....
'10000
8000
,t
3000
. IL...' /
5000
2000
11'
4000
1000
,,,. 0
8
c
12000
29
36
51
0
8
29
36
51
0
0 8 29 36 51 57 Time after initial vaccination (days] Rgure 6 Anti-peptide isotype profiles of three cattle immunized with a continuous infusion (a), double (b), or multiple (c) doses of synthetic FMDV peptide. Booster immunizations are indicated by the arrows, the period of continuous infusion in (a) by the horizontal line. ( ) IgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . . ) IgA 6000 I
14ooo
5000
12000
I
8000
3000
,
6000 2000
""
1000 0
'"
7
14
, ' ';""C " l,
c
2000
10000
4000 =
2500
b 1500 1000
4000 ,
21 2 8 3 5 4 2 4 9
500
2000 0
0
"7
14 21 2 8 3 5 4 2 4 9
0
/" 0
.
7
-~.~L'.'~ "~'
14 21 2 8 3 5 4 2 4 9
Time p o s t v a c c i n a t i o n (days) Rgure 7 Anti-peptide isotype profiles of three cattle given peptide in an oil adjuvant subcutaneously (a), intramuscularly (b) and intradermally (c). Second immunizations are indicated by the arrows. ( ) IgG1, (. . . . ) IgG2, ( . . . . ) IgM, (. . . . . . ) IgA
V a c c i n e , Vol. 8, J u n e 1990
255
Isotope r e s p o n s e s of cattle to FMDV: G. M u l c a h y et al.
If it is the case that macrophages play a critical role in the clearance of FMDV in cattle through the processes of opsonization, attachment to Fc and/or complement receptors and phagocytosis then the importance of antibody isotype to induction of protective immunity is clear. It has been shown that antibody isotype (i.e. constant region structure) can be influenced by the macromolecular form in which a B-cell epitope is presented 16, and isotype restriction of antibody responses to viruses in other species has been documented 17. The effect of antibodies ofisotypes unable to effect opsonization and macrophage phagocytosis on the immune status of a peptide-immunized animal clearly depends on whether those antibodies simply block access of 'useful' antibody molecules to the virus or can themselves stop virus replication by secondary, albeit less efficient, means. It seems likely that both effects may occur, the balance depending on the relative numbers and affinities of each isotype present. It is clear, also, that the direction of an immune response towards production of one isotype of antibody is a complex process which depends not only on the macromolecular form of the antigen presented, but also on the adjuvant in which it is presented and on the presence of T-cell epitopes on the antigen. Indeed, it has recently been_demonstrated is that individual B-cell epitopes within a peptide can vary with respect to immunoglobulin sub-class restriction, thus there are numerous parameters that can influence the immunoglobulin sub-classes generated during an immune response. In protection data obtained with peptide-immunized cattle we were not abte to discern patterns of isotype induction that are typical of protected or non-protected animals. However, there was an apparent tendency for peptide-immunized animals to be protected if they had rising IgG1 titres at the time of challenge, regardless of IgG2 levels. One animal (RA87) which had very high total antibody titres and which, surprisingly, remained susceptible to FMDV challenge was found on analysis of isotype profile to have mounted a response consisting almost entirely of IgG2 with negligible levels of other isotypes. This, together with the early non-protective IgG2 response in infected animals gives preliminary indications of the relevance of IgG1 to protection against FMDV in cattle. Peptide-immunized animals in which IgG2 was the predominant isotype were protected only if they had extremely high titres, indicating that virus clearance in the case of these animals may have been by secondary mechanisms.
Conclusions Cattle immunized with virulent or inactivated FMDV mount a humoral immune response in which the isotype profile differs significantly from that which occurs in cattle immunized with a synthetic peptide. The fact that an IgGl response is favoured in virus-immunized, and IgG2 in peptide-immunized animals may partially explain the lower protective capacity of these anti-peptide antibodies against FMDV challenge. Additional factors which have yet to be considered include individual animal variation in response to either type of antigen, reactivity of anti-isotypic antibodies with allotypic determinants and the effect of vaccine adjuvants or additives on class-switching and resultant antibody isotype profile. Elucidation of these parameters together with experiments to ascertain the mechanisms of FMDV clearance by
256
Vaccine, Vol. 8, June 1990
antibody interaction with cells of the immune and reticuloendothelial systems should in the future allow synthetic peptide vaccines against FMDV to be engineered towards the optimal protective response.
Acknowledgements The authors would like to thank Dr D. van Zaane, CDI, Lelystad, The Netherlands, for his gift of monoclonal anti-isotype antibodies, and Mrs Ann Boddy for typing the manuscript. The technical assistance of Mr Donald White is greatly appreciated.
References 1 Doel, T.R. Prospects for improved foot and mouth disease vaccines. Vaccine 1985, 3, 35-36 2 DiMarchi, R., Brooke, G., Gale, C., Cracknell, V., Doel, T.R. and Mowat, N. Protection of cattle against foot-and-mouth disease by a synthetic peptide. Science 1986, 232, 639-641 3 Van Bekkum, J.G. Correlation between serum antibody level and protection against challenge with FMD. Session of the Research Group of the Standing Technical Committee of the European Commission for the Control of Foot and Mouth Disease, Brescia, Italy, FAO, 1969 4 Sellei, J. Interaction of bovine immunoglobulins with complement. Comp. /mmunol. Microbio/. Infect. Dis. 1984, 1(](2), 93-98 5 McGuire, T.C., Musuoke, A.J. and Kurtti, T. Functional properties of bovine IgG1 and IgG2; Interaction with complement, macrophages, neutrophils and skin. Immunology 1979, 38, 249-256 6 Donaldson, A.I. and Kitching, R.P. Transmission of foot-and-mouth disease by vaccinated cattle following natural challenge. Res. Vet. Sci. 1989,40, 9-14 7 Van Zaane, D. and Ijzerman, J. Monoclonal antibodies against bovine immunoglobulins and their use in isotype-specific ELISAs for rotavirus antibody. J. Immunol. Methods 1984, 72, 427-441 8 Golding, S.M., Hedger, R.S., Talbot, P. and Watson, J. Radial immunodiffusion and serum neutralisation techniques for the assay of antibodies to swine vesicular disease. Res. Vet. Sci. 1976, 20, 142-147 9 Abu-Elzein, E.M. and Crowther, J.R. Detection and quantification of IgM, IgA, IgG1 and IgG2 antibodies against foot-and-mouth disease virus from bovine sera using an enzyme-linked immunosorbent assay. J. Hyg. (Camb.) 1981, 86, 79-85 10 McCullough, K.C., Crowther, J.R., Butcher, R.N., Carpenter, W.L., Brocchi, E., Capucci, L. and DeSimone, F. Immune protection against foot-and-mouth disease virus studied using virus neutralising and non-neutralising concentrations of monoclonal antibodies. Immunology 1986, 58, 421-428 11 Bittle, J.L., Houghton, R.A., Alexander, H., Shinnick, T.A., Sutcliffe, J.G. and Lerner, R.A. Protection against foot-and-mouth disease by immunization with a chemically synthesized peptide predicted from the viral nucleotide sequence. Nature 1982, ~ , 30-33 12 Broekhuijsen, M.P., Von Rijn, J.M.M., Blom, A.J.M, Pouwells, P.H., Enger-Valk, B.E., Brown, F. and Francis, M.J. Fusion proteins with multiple copies of the major antigenic determinant of foot-and-mouth disease virus protect both the natural host and laboratory animals. J. Gen. Virol. 1987, 68, 3137-3143 .13 McCullough, K.C., Parkinson, D. and Crowther, J.R. Opsonisationenhanced phagocytosis of foot-and-mouth disease virus. Immunology 1988, 65(2), 187-192 14 Howard, C.J., Taylor, G. and Brownlie, J. Surface receptors for immunoglobulin on bovine polymorphonuclear leucocytes and macrophages. Res. Vet. Sci. 1980, 20, 128-130 15 Mackie, D.P. and Logan, E.F. Changes in immunoglobulin levels in whey during experimental Streptococcus agalactiae mastitis. Res. VeL Sci. 1986, 40, 183-188 16 Fish, S. and Manser, T. Influence of the macromolecular form of a B cell epitope on the expression of antibody variable and constant region structure. J. Exp. Med. 1987, 166, 711-724 17 Courtelier, J.P., van der Logt, J.T.M., Hessen, F.W.A., Warner, G. and Van Snick, J. IgG2a restriction of murine antibodies elicited by viral infections. J. Exp. Med. 1987, 165, 64-68 18 Mathiesen, T., Brolinden, P.A., Rosen, J. and Wahren, B. Mapping of IgG subclass and T-cell epitopes on HIV proteins with synthetic peptides. /mmunology 1989, 67, 453-459
