Veterinary Immunology and Immunopathology, 33 ( 1992 ) 129-143 Elsevier Science Publishers B.V., Amsterdam
129
The production of equine monoclonal immunoglobulins by horse-mouse heterohybridomas C.M. Richards a, H.A. Aucken b, E.M. Tucke#, D. Hannant c, J.A. Mumfor& and J.R. P o w e l l d aAFRC Institute of Animal Physiology and Genetics Research, Babraham, Cambridge CB2 4AT, UK bPublic Health Laboratory Service, Central Public Health Laboratory, 61 Colindale Avenue, London NW9 5HT, UK CAnimal Health Trust, P.O. Box 5, Newmarket, CB8 7DW, UK dUnilever Research, Sharnbrook, Bedford, UK (Accepted 18 September 1991 )
ABSTRACT Richards, C.M., Aucken, H.A., Tucker, E.M., Hannant, D., Mumford, J.A. and Powell, J.R., 1992. The production of equine monoclonal immunoglobulins by horse-mouse heterohybridomas. Vet. Immunol. Immunopathol., 33: 129-143. Studies were carried out to determine the optimum conditions for the production of equine monoclonal antibodies (MAbs). Lymphocytes from ponies immunised with influenza A equine 2 virus, isolate A/Equine/Newmarket/79 (H3N8) were fused with mouse myeloma (NSO) cells and with horse-mouse heterohybridomas made aminopterin-sensitive by selective growth in 8-azaguanine. Although all fusions initially resulted in heterohybridoma colonies that secreted equine immunoglobulin, many of these were unable to maintain secretion for longer than a few weeks. Increasing the time between immunisation and the booster injection of Newmarket/79 virus, the inclusion of Freund's incomplete adjuvant and the use of an aminopterin-sensitive primary heterohybridoma as the fusion partner, improved the production of HIg-secreting heterohybridomas. After two clonings eight cell lines were established which maintained anti-Newmarket/79 antibody secretion for over a year. FACS analysis of the cell lines provided a useful means of predicting breakdown of MAb secretion by the cell lines, thus enabling re-cloning to be carried out in time. ABBREVIATIONS ELISA, enzyme-linked immunosorbent assay; FCS, foetal calf serum; FIA, Freund's Incomplete Adjuvant; GAH, goat anti-horse; HAT, hypoxoanthine aminopterin thymidine; HIg, horse immunoglobulin; HS, horse serum; Ig, immunoglobin; MAbs, monoclonal antibodies; PBL, peripheral blood lymphocytes; RAH, rabbit anti-horse; SAH, sheep anti-horse; SDS---PAGE, sodium dodecyi sulphate polyacrylamide gel electrophoresis. Correspence to: C.M. Richards, AFRC Institute of Animal Physiology and Genetics Research, Babraham, Cambridge CB2 4AT, UK.
© 1992 Elsevier Science Publishers B.V. All fights reserved 0165-2427/92/$05.00
130
C.M. RICHARDS ET AL.
INTRODUCTION
Although rodent-derived monoclonal antibodies (MAbs) are suitable for a wide variety of different purposes in veterinary immunology, there are some instances where MAbs derived from other species are required. One such instance arises when the antibody is required for immunotherapy in which case it needs to be of the same species as the recipient to avoid triggering an immune response. Non-rodent MAbs are also required when studying host-parasite relationships in other animals as they are more likely to be raised against those parasite antigens seen by the host's own immune system. In the absence of suitable non-rodent myeloma cell lines for fusion, mouse myelomas have been used to produce human-mouse, sheep-mouse or cattle-mouse heterohybridomas and shown to be a successful means of producing cell lines secreting human (James and Bell, 1987), sheep and cattle (Groves and Tucker, 1989) MAbs respectively. Stable antibody-secreting heterohybridomas are, however, more difficult to produce than rodent hybridomas because of their inherent chromosome instability (Weiss and Green, 1967). This instability can sometimes be improved by creating aminopterin-sensitive heterohybridomas and using these as fusion partners in place of the rodent myeloma cell line (Tucker et al., 1984, 1987; Anderson et al., 1986, 1987; Groves et al., 1987, 1988; Kennedy et al., 1988; Flynn et al., 1989). This paper gives results of experiments to determine the feasibility of extending the above studies to produce horse-mouse heterohybridomas and compares their use as fusion partners with that of mouse myeloma cells. The antigen chosen was influenza A equine 2 because antibodies to this virus would be of potential use in diagnosis or treatment of this disease (review by Russel and Edington, 1985 ). MATERIALS AND METHODS
Immunisation
One isolate of the influenza A equine 2 virus A/Equine/Newmarket/79, used as the immunising agent, was prepared and administered at the Animal Health Trust (AHT), Newmarket. The virus was propagated by inoculation into the allantoic cavity of embryonated hens eggs. Welsh Mountain ponies of various ages and both sexes (AHT, Newmarket), were exposed to either an aerosol of Newmarket/79-infected allantoic fluid using a model 65 nebuliser (De Vilbiss, Somerset, PA, USA) as described by Mumford et al. (1990), or given an intramuscular injection of virus. The intramuscular injection of inactivated whole virus antigens or live virus was given as an equal volume of virus at 100/tg ml -~ and adjuvant (Anderson et al., 1987). Initially the adjuvant used was aluminium hydroxide, but for later immunisations this
EQUINE MONOCLONALIMMUNOGLOBULINS
131
was mixed, 1:1, with Freund's Incomplete Adjuvant (FIA). The emulsion was injected into the leg just below the popliteal lymph node to ensure local stimulation.
Cell lines Mouse myeloma NSO cells were used for primary fusions with equine lymphocytes. NSO, a derivative of NS1/Ag 4.1, does not synthesize mouse immunoglobin (Ig) chains. Several non-Ig-secreting heterohybridomas resulting from these fusions were selected for aminopterin-sensitivity by cloning on soft agar containing 20/zg ml-1 8-azaguanine (Tucker et al., 1984; Anderson et al., 1987). These aminopterin-sensitive lines (EQ17, EQ38, EQ42 and EQ43) were produced from fusions using peripheral blood lymphocytes (PBL) of different ponies, and then used as fusion partners in secondary fusions. All cells were maintained in RPMI-1640 medium (Flow Laboratories, Irvine, UK), containing penicillin-streptomycin ( 100 U ml-1 ) (Gibco, Paisley, UK) and foetal calf serum (FCS, Imperial Laboratories, Andover, UK) at 5-20%.
Lymphocytes Blood was collected from ponies into heparinised tubes (Vacutainers, Becton Dickinson UK, Oxford), mixed and allowed to stand for 20 min at 37 °C. The plasma was removed and layered onto Histopaque (Sigma Chemicals, Poole, UK) which contains ficoll 400, and the lymphocytes were separated by density gradient centrifugation at 150 × g for 40 rain in a Centra 7R (Damon/IEC, Dunstable, UK) centrifuge. The lymphocyte-rich layer was removed, washed and resuspended in RPMI-1640 with supplements plus HEPES at a final concentration of 0.01 M (Gibco, Paisley, UK). Lymph nodes and spleen were collected and prepared for fusion according to the method described by Anderson et al. ( 1987 ), using the medium described above.
Freezing and thawing of cells Cells for freezing were spun down at 400Xg for 5 min and resuspended at 2 X 106-107 cells ml-1 in 90% FCS, 10% dimethylsulphoxide (BDH Chemicals, Dagenham, UK). The cells were transferred to 2 ml cryotubes (Nunc, Kamstrap, Denmark), placed in polystyrene boxes and frozen at - 7 0 ° C for 24 h before being placed in liquid nitrogen. The cells were thawed quickly and transferred immediately to a universal tube containing fresh culture medium, spun at 400Xg for 5 min and resuspended in RPMI + 20% FCS.
132
C.M. RICHARDS ET AL.
Fusions Fusions were carried out by mixing equine lymphocytes with mouse NSO cells or equine heterohybridoma fusion partner at a 4: 1 ratio respectively, in the presence of 50% ( w / v ) polyethylene glycol (PEG 1500, Boehringer Mannheim, Germany). The fused cells were washed in RPMI containing 0.01 M HEPES buffer and resuspended in RPMI medium with 20% FCS at 10 6 cells m1-1 and plated out ( 1 ml per well) into 24-well plates (Linbro, Flow Laboratories, Irvine, U K ) . After 24 h in culture, 1 ml of selective medium, to give a final concentration of 10-4 M hypoxanthine, 4 × 10-7 M aminopterin and 1.6× 10 -5 M thymidine (HAT) (Anderson et al., 1987), was added. Control wells were also set up containing unfused myeloma cells and lymphocytes in medium with and without HAT.
Screening culture supernatants Culture supernatants were removed from wells containing fused cells and screened for both the presence of horse immunoglobulin (HIg) and specific antibody to Newmarket/79 virus, by solid-phase enzyme-linked immunosorbent assays (ELISA). The HIg ELISA plates (Falcon, Becton Dickinson UK, Oxford) were coated overnight at 4 °C with affinity-purified goat anti-horse (GAH) IgG (H + L), obtained from Jackson Immunoresearch Laboratories (Westgrave, PA), diluted to 2/~g m l - i in bicarbonate buffer pH 9.6. The antigen used to coat the H3N8 virus-specific assay plates was the third egg passage of Newmarket/79 strain containing 64 haemagglutination (HA) units diluted 1 : 250 in bicarbonate buffer, pH 9.6. HIg and H3N8 plates were then washed four times in phosphate-buffered saline containing 0.05% Tween 20 (Sigma Chemicals, Poole, U K ) , PBST. Culture supernatants were added for incubation at 37°C for 1 h, washed as above and incubated for 1 h in the presence of affinity-purified, alkaline phosphatase-conjugated GAH IgG (H + L) (Jackson Immunoresearch Laboratories, Westgrave, PA) diluted 1 : 5000 in PBST. After four more washes, p-nitrophenyl phosphate disodium (Sigma Chemicals, Poole, U K ) in substrate buffer (0.5 M NaHCO3 + 0.002 M MgCI2), pH 9.5, at 1 mg ml -l was added to each well, and the plates were incubated for a further hour at 37 ° C. Optical densities were read on a Titertek Multiskan automatic reader at wavelengths 405 and 510 nm absorbance. Positive controls included serum from a hyperimmune pony (Day 14 postH3N8 virus infection ), diluted 1 : 2000 in culture medium, and purified horse IgG (H + L) (Jackson Immunoresearch Laboratories, Westgrave, PA) diluted to 10 #g m l - 1. Culture medium was also used as the negative control in both H3N8 antibody and HIg assays.
EQUINE MONOCLONALIMMUNOGLOBULINS
133
Cloning Where possible, single distinct colonies were picked out of those 2-ml wells assayed by ELISA as positive for H3N8-specific antibody and transferred to 0.2 ml or 2 ml wells. Alternatively all cells from a positive well were maintained and cloned on soft agar according to the method described by KiShler (1979). All selected cell lines were cloned twice. Cloning efficiency was assessed as the percentage of anti-H3N8 antibody-secreting clones obtained after cloning.
Fluorescence-activated cell sorter To determine the cell surface expression of the HIg produced by the different heterohybridoma lines, cells were analysed using a FACScan (Becton Dickinson UK, Oxford). Cells ( 106) were incubated in 100 #1 of affinitypurified, fluorescein-conjugated GAH IgG ( H + L ) (Jackson Immunoresearch Laboratories, Westgrave, PA) diluted 1 : 5 in RPMI plus 5% FCS, for 45 min at 4 ° C. The cells were then washed twice with 100/tl culture medium and resuspended in 200/A PBS. Negative controls included non-Ig secreting lines and lines incubated in the absence of the fluorescent conjugate. The level of fluorescence above that of the negative controls was then measured for each line.
Chromosome analysis Cells grown in 50 ml tissue culture flasks were treated with Colcemid, (Gibco, Paisley, U K ) , at a final concentration of 0.06/~g m l - 1 and incubated for 4 h at 37°C, 5% COz. The cells were harvested, spun at 4 0 0 × g for 5 min and resuspended in 5 ml of 75 mM KC1, mixed and spun as above. The pellet was resuspended in 5 ml of fixative (3 parts methanol: 1 part glacial acetic acid) at room temperature for 15 min and then washed three times in fixative, reducing the volume successively to 0.5-1.0 ml. Two to three drops of the fixed cell suspension were dropped onto clean slides which had been washed and immersed briefly in 60% methanol. The slides were passed rapidly through a bunsen flame and stored at room temperature for 1 week prior to Giemsa staining (Lin et al., 1977).
Purification of monoclonal immunoglobulins Culture supernatants from the heterohybridoma cell lines and control 20% FCS in RPMI were concentrated ten times with Amicon Hollow Fibre concentrator with a molecular weight (MW) cut-off of 30 kDa and affinity-purified on a Protein G Sepharose 4 Fast Flow column (Pharmacia L.K.B. Bio-
134
C.M. RICHARDS ET A L
technology, Milton Keynes, UK) according to the method described by Bj6rck and Kronvall (1984).
Immunoglobulin analysis The immunoglobulins secreted by the heterohybridoma cell lines were radiolabelled over 16-18 h in 250/~1 lysine-free RPMI (Flow Laboratories, Irvine, UK) containing 50 /tCi m1-1 of [14C]-lysine monohydrochloride (Amersham, U K ) . The cells were spun at 400Xg for 5 min, and the supernatants collected and stored if necessary at - 2 0 ° C. These supernatants were then mixed 1 : 1 in sample buffer containing fl-2-mercaptoethanol and bromophenol blue, boiled for 3 min and examined by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE), as described by Laemmli (1970). Autoradiography was performed by exposure of the gels to Hyperfilm-B max (Amersham, U K ) and the size of the Ig chains estimated with respect to molecular weight markers (Sigma Chemicals, Dorset, U K ) also run on the gel. Monoclonal Ig samples, concentrated and purified on a Protein G column, were analysed by double immunodiffusion tests. Samples were added to wells in 1% barbitone agar (Hudson and Hay, 1976 ) for immunodiffusion against rabbit anti-horse (RAH) IgG (provided by Cambridge University Veterinary School) overnight at 37 ° C. RESULTS
Production of mouse-horse heterohybridomas Initial fusions were carried out between the mouse myeloma line NSO and equine lymphocytes (Table 1 ). For the first two fusions, PBL were taken from ponies 4 days after contact with ponies shedding the H3N8 virus (Table 1 ). One of these, Pony 37, was naive, i.e. it had not been exposed to the H3N8 virus previously, while the other, Pony 20, was injected with inactivated whole virus 4 and 2 weeks prior to in-contact infection. Both fusions resulted in hybrid colonies of which 25% and 10% respectively, were shown by ELISA to secrete HIg, but none was specific to H3N8 virus. With the aim of improving the immunisation procedure and source of lymphocytes, fusions were then carried out between NSO cells and lymphocytes taken from lymph nodes and spleen of a previously immunised pony (W5) 4 days after a secondary aerosol challenge and an intramuscular booster injection of live H3N8 virus in the region of the inguinal lymph node (Table 1 ). The growth of hybrid colonies varied considerably but fewer hybridomas were formed and fewer Ig-secretors were present than in the previous fusions car-
135
EQUINEMONOCLONALIMMUNOGLOBULINS TABLE 1 Production of horse-mouse heterohybridomas Pony
Horse lymphocytes
Wells with colonies (%)
Mean colonies per well
Secreting HIg (%)
371 202
PBL PBL
96 96
4.6 2.9
25 10
W53
Bronchial LN Mediastinal LN Inguinal LN Spleen
66 48 33 4
1.8 0.7 0.3 0.08
3 8 10 4
tPBL fused with NSO cells 4 days after in-contact infection with ponies secreting H3N8 virus. 2pBL fused with NSO cells 4 days after in-contact infection, 4 and 2 weeks after i.m. injection with H3N8 virus. 3Lymphocytes removed for fusion 4 days after an aerosol and i.m. injection of H3N8 virus, 8 weeks after a primary aerosol challenge. All HIGgG secretors were negative for H3N8-specific antibody when tested by ELISA.
ried out with PBL. The spleen fusion was particularly poor resulting in only 4% of the wells containing hybrid colonies. Aminopterin-sensitive horse-mouse heterohybridomas were prepared by further fusions of NSO cells with PBL followed by selection in 8-azaguanine. The ponies used to provide the PBL (W4, W2, 277 and 289) had been previously immunised with H3N8 virus but none was challenged immediately before fusion. Four non-HIg secreting hybrids were chosen for use themselves as fusion partners.
Comparison of fusion partners NSO cells and the four aminopterin-sensitive heterohybridoma fusion partners (EQ 17, EQ38, EQ42, EQ43 ) were fused with PBL from two immunized ponies (W2 and W4) 4 days after either an intravenous or intramuscular booster injection of Newmarket/79 virus respectively (Table 2). The fusion efficiencies for the different fusion partners were extremely variable, but for each pony, the overall best fusion partner in terms of percentage growth and mean colonies per well was the EQ 17 heterohybridoma. Both EQ 17 fusions produced some anti-Newmarket/79 secretors but these only maintained secretion for 2-3 weeks after fusion. Eight further fusions were carried out using PBL from four different immunised ponies for comparison of EQ 17 and NSO cells as fusion partners. Although the results (not shown) of the fusions varied slightly, in each case EQ 17 gave better fusion efficiencies and a higher number of Ig-secreting lines than NSO. The EQ 17 fusion with Pony W2 lymphocytes (Table 2 ) gave rise
136
C.M. R1CHARDS ET AL.
TABLE 2 Comparison of aminopterin-sensitive heterohybridomas and mouse myeloma NSO cells fused to PBL from immunised ponies Cell line used in fusion
Pony
Wells with colonies (%)
EQ 17 EQ38 EQ42 EQ43 EQI 7 NSO
W21 W2 W2 W2 W42 W4
100 100 25 10 100 100
Mean colonies per well 20.0 9.0 0.2 0.1 > 30 7.4
HIg secretors (%)
H3N8 antibody secretors (%)
17 6 2 0 47 27
2 0 0 0 17 0
ILymphocytes used 4 days after an i.v. injection, 13 and 11 months post-aerosol challenge. 2Lymphocytes used 4 days after an i.m. injection, 6 and 3 months post-aerosol challenge.
to more than 30 hybrid colonies per well, and 43% of wells with HIgG activity, compared with 7.4% colonies per well and 27% HIgG secretion for the relative NSO fusion. From this it was concluded that EQ 17 would be a better fusion partner than NSO for subsequent experiments.
Fusions using lymphocytes from different sources It was thought that failure to produce sufficient cells secreting antibody to H3N8 virus could be due to the fact that antibody levels remained high in the immunised pony long after the initial challenge, thus resulting in a poor secondary immune response (Hannant et al., 1989). A pony (277) was therefore rested for 27 months after secondary challenge (i.e. until antibody levels were low), and then given two intramuscular booster injections of Newmarket/79, close to the left sub-mandibular and popliteal lymph nodes. Both stimulated and unstimulated lymph nodes, PBL and spleen cells were fused with the EQ 17 fusion partner, 4 days post-boost. Table 3 summarises the resuits of these fusions. Good growth was observed in all fusions and therefore, because of the large number of samples for testing, only H3N8-specific assays were carried out. The PBL fusion produced no colonies secreting anti-H3N8 antibody, but all other fusions gave some H3N8-specific antibody secretors. As expected, the stimulated lymph nodes gave the highest number of anti-H3N8 secretors, of which two lines (A and B) were successfully expanded, cloned and maintained. Some of the excess lymphocytes from these lymph nodes were frozen in liquid nitrogen and used in later fusions (Table 3 ) which resulted in one more stable antibody-secreting cell line (line C ). A second previously vaccinated pony (83) was given an intramuscular boost 12 months after secondary challenge and the stimulated lymph nodes were
EQUINEMONOCLONALIMMUNOGLOBULINS
137
TABLE 3 Fusion efficiency of lymphocytes from different sources fused with aminopterin-sensitive EQ 17 cells Pony
Lymphocytes
Mean colonies per well
Secreting anti-H3N8 Week 3 (%)
Lines maintained
277 t
PBL S popliteai L.N. US popliteal L.N. S sub-mandibular L.N. US submandibular L.N. Spleen
12.2 17.3 13.5 15.4
0 38 2 20
0 2 0 3
S popliteal L.N. stored at -196oc S sub-mandibular L.N. stored at -196°C 832
S popliteal L.N. S mandibular L.N. S inguinal L.N.
Lines stable
A B
35.7
6.25
0
37.7
1.4
1
16.5
5.0
0
2.5
1.4
1
C
4.1 5.54
4.2 6.8
0 2
DE
38.5
3
FGH
12.5
~Lymphocytes removed for fusion 4 days post-i.m, injection, 30 and 27 months post-aerosol challenge. 2Lymphocytes removed for fusion 4 days post-i.m, injection, 30 and 12 months post-aerosol challenge. S, stimulated lymph node by local injection of virus; US, not injected locally; Lines maintained, secreted antibody for 2-3 months; lines stable, secreted antibody for over 1 year.
removed for fusion 4 days later (Table 3). A further five lines (D-H) were thus expanded, cloned and maintained in culture, giving a total of eight equine heterohybridoma cell lines secreting antibody specific for influenza A equine 2 virus.
Stability of horse-mouse heterohybridomas secreting specific antibody Cell lines A-H cloned twice on soft agar remained positive when assayed for H3N8 and HIgG specificity by ELISA for over 12 months in continuous culture. These lines were also tested by FACS analysis in order to calculate the percentage of cells in the population producing HIg. Table 4 gives the FACS results obtained immediately before the second cloning, i.e. approximately 6 months after fusion, for comparison with the cloning efficiencies (i.e. percentage of wells positive after cloning). In most cases the correlation between the percentage of cells with surface HIg (10-88%) and cloning efficiency (2-90%) was good, suggesting that FACS analysis is a useful test of
138
C.M. RICHARDS ET AL.
TABLE 4 Stability of heterohybridomas as judged by FACS analysis and second cloning of cell lines Hybridoma cell line
FACS 1% surface HIg
Cloning 2 efficiency (%)
A B C D E F G H
30 88 45 30 70 41 10 35
28 73 2 95 83 53 90 36
~FACS analysis of cell lines cloned once on soft agar, immediately before second cloning. 2Second cloning carried out 6-8 months after fusion.
HIg and MAb production. There were exceptions, however; line C had a very poor cloning efficiency, although 45% of its cells were HIg-positive by FACS. Conversely, lines D and G gave much higher cloning efficiencies than would be predicted by the FACS measurements; this may be due to the variability of the surface Ig expression. Some of the re-clones resulting from the above experiment were also analysed on the FACScan. All showed some improvement in the percentage of cells expressing HIg. In particular, line A increased from 30 to 50%; line C 45-96%; line G 10-33% and line H 35-76%. This indicated the importance of cloning for the maintenance of antibody-secreting heterohybridoma cell lines.
Immunoglobulin analysis 14C-lysine-labelled supernatants from lines A - H and EQ 17 were analysed by SDS-PAGE autoradiography (Fig. 1 ). The molecular weight of the heavy chains indicated that with the exception of line D, all samples were in the range of the IgG isotype, i.e. 48-52.5 kDa (Montgomery, 1973; McGuire et al., 1973 ). Even though line D was an antibody secretor it appeared to lack a heavy chain. This would indicate that the secretion of heavy chain in this sample was below detection by SDS-PAGE. The light chains present in all samples showed slight variation between the molecular weights of 24.5-27 kDa. Such variation has also been reported in both rodent (Howard et al., 1980 ) and cattle (Anderson et al., 1987 ) Ig light chains, and is thought to be due to variation in the carboxyl groups. Figure 2 illustrates precipitin reactions between seven of the eight monoclonal Ig samples with RAH IgG. Reactions of identity were observed be-
139
EQUINE MONOCLONAL IMMUNOGLOBULINS
H
G
F
E
D
C
B
A
EQ17
MW
-180 -116 84
Heavy
W
m
58 48.5
36.5 26.5 Light
~ --
Fig. 1. Autoradiograph of SDS-PAGE of 14C-labelled HIg chains in culture supernatants. Fusion partner EQ 17 showed no evidence of HIg secretion. The anti-H3N8 secreting lines A - H with the exception of D show heavy (48-52.5 kDa) and light chains (24.5-27 kDa). Line D did not give a visible heavy-chain band.
Fig. 2. Double immunodiffusion tests of the monoclonal Ig samples A-H, concentrated and purified on a Protein G column against RAH IgG. Control samples included concentrated, purified 20% FCS and normal horse serum (HS).
tween A and B, C and E as well as F, G and H, while a partial reaction was observed between B and C. This suggests the presence of at least two subtypes of HIgG. No precipitin lines were observed with the control FCS. Line D was again negative, but this was probably due once again to the low concentration of heavy chain. Immunodiffusion tests were also carried out with sheep anti-horse (SAH)
140
C.M. R1CHARDS ET AL.
IgG (Evaibios Laboratories, Petworth, U K ) and GAH IgG with similar results. Both FCS and normal mouse serum showed no reaction, indicating that the MAb Ig samples were of horse origin.
Chromosome analysis Previous experiments with cattle-mouse heterohybridomas had indicated that a possible explanation for the superiority of some heterohybridomas as fusion partners was that they retained more cattle chromosomes than others (Anderson et al., 1987 ). The chromosomal content of seven of the heterohybridoma cell lines including EQ 17 was therefore determined by examination of Giemsa-stained preparations of metaphase spreads. The equine karyotype consists of 64 chromosomes while NSO has approximately 65. The latter can be distinguished from horse chromosomes when stained in Giemsa by their dark centromeres compared with the clear centromeres of the horse chromosomes. Figure 3 shows a metaphase spread from a heterohybridoma in which approximately 25 out of the 90 chromosomes seen were horse. One non-HIg-secretor tested had a mean of 6% equine chroiii!i
¸3!3¸
~iiiiiii
M - Mouse H - Horse
Fig. 3. Giemsa-stained metaphase spread of a heterohybridoma cell line ( × 100). Twenty-five equine chromosomes, some of which have been indicated (H), are identified in this spread in a total of 90 chromosomes.
EQUINE MONOCLONALIMMUNOGLOBULINS
141
mosomes and the mean of each of seven HIg secreting lines tested ranged from 15 to 23%. EQ 17 fusion partner, another HIg non-secretor, had between 10 and 16 horse chromosomes out of a total of 116-129, giving an average of 11%. There was therefore no clear indication that the number of equine chromosomes retained was relevant to hybridoma efficiency. DISCUSSION Appleton et al. ( 1989 ) and Perryman et al. ( 1990 ) have described the production of cell lines secreting an equine MAb to influenza A equine l (H7N7) and equine infectious anaemia viruses respectively by fusing mouse myeloma cells with stimulated PBL and spleen cells from immunised ponies. The results reported in the present paper confirm and extend the feasibility of the heterohybridoma approach for the production of monoclonal equine Ig. In agreement with previous results using sheep-mouse (Flynn et al., 1989 ) and cattle-mouse (Tucker et al., 1984, 1987; Anderson et al., 1986, 1987 ) hybridomas, there was convincing evidence that a horse-mouse fusion partner (EQ 17 ), made HAT-sensitive by selective growth in 8-azaguanine, was more efficient than mouse myeloma cells alone in producing stable antibody-secreting lines. The reason for the particular efficiency of EQ 17 (M × H) cell line in this respect is unclear. Chromosome analysis of the EQ17 line has shown that it does not retain a particularly high percentage of horse chromosomes compared with other lines. Also, recent tests indicate that lines produced from secondary fusions in which EQ 17 was the fusion partner, do not retain more equine chromosomes than those from primary fusions. One clear observation from the present series of experiments was the importance of selecting the method and timing of immunisation in order to produce well-stimulated lymphocytes for fusion. A characteristic of the response to primary challenge with influenza A equine 2 virus is the sustained subsequent presence of antibody in the peripheral circulation. To achieve maximal stimulation from a booster injection it would appear to be necessary to wait until the primary antibody response has waned. In the case of the ponies used in the present study, this took many months. As has been found with the production of bovine MAbs to E. coli (Anderson et al., 1987 ) and alloantigens (Tucker et al., 1987 ), it was clear that best results were obtained with lymphocytes derived from locally stimulated lymph nodes. FACS analysis, although not specific for antibody to H3N8 virus, gave useful information about the Ig-secretion of each cell line. Although FACS results did not always correlate with cloning efficiencies, regular FACS analysis proved a useful means of determining the stability of a line and indicating when cloning may be necessary. Owing to the presence of residual Ig in the culture supernatants, the ELISA will not indicate loss of Ig-production immediately, and by the time they have done so, it may be too late to clone.
142
C.M. R I C H A R D S ET AL.
Analysis o f the monoclonal HIg samples by S D S - P A G E and double imm u n o d i f f u s i o n tests with R A H IgG suggested that at least seven o f the eight MAb lines are of the IgG isotype. The slight variation observed in the size of the heavy chain on the S D S - P A G E together with the reaction of partial identity observed by i m m u n o d i f f u s i o n m a y be due to the presence of different IgG subtypes. Previous work on HIg (Montgomery, 1973; McGuire et al., 1973 ) has indicated that at least four subtypes o f IgG exist, IgGa, IgGb, IgGc and possibly IgG (T). Further characterisation will therefore be required before the exact isotype o f each MAb-secreting line produced can be determined and it is hoped that these and other cell lines can be used for the i m m u n o chemical analysis of different HIg classes. As a result of experiments described in this paper, it has been possible to create not only cell lines producing HIg of undefined specificity but also lines producing MAb to influenza A equine 2 virus. We attribute this success to the immunisation schedule, the use of lymphocytes from locally stimulated lymph nodes, and the use of h e t e r o h y b r i d o m a fusion partner EQ 17, and plan to use these MAbs for further analysis o f the influenza A equine 2 virus. ACKNOWLEDGEMENT This work was supported in part by a grant from Centaur, R i c h m o n d , VA, USA, and Unilever Research, Sharnbrook, Bedford, UK.
REFERENCES Anderson, D.V., Clarke, S.W., Stein, J.M. and Tucker, E.M., 1986. Bovine and ovine monoclonal antibodies to erythrocyte membrane determinants, produced by interspecific (hetero-) myelomas.Biochem. Soc. Trans., 14: 72-73. Anderson, D.V., Tucker, E.M., Powell,J.R. and Porter, P., 1987. Bovine monoclonalantibodies to the F5 (K99) pilus antigen ofE. coli, produced by murine/bovine hybridomas. Vet. Immunol. Immunopathol., 15: 223-237. Appleton, J.A., Gagliardo,L.F., Antczak, D.F. and Poleman, J.C., 1989. Production of an equine monoclonal antibody specific for the H 7 haemagglutininof equine influenza virus. Vet. Immunol. Immunopathol., 23: 257-266. Bj6rck, L. and Kronvall, G., 1984. Purification and some properties of streptococcal Protein G, a novel IgG-bindingreagent. J. Immunol., 133: 969-974. Flynn, J.N., Harkiss, G.D. and Hopkins, J., 1989. Generation of a sheepX mouse heterohybridoma cell line (IC6.3a6T. l D7) and evaluation of its use in the production of ovine monoclonal antibodies. J. Immunol. Methods, 121: 237-246. Groves, D.J. and Tucker, E.M., 1989. The production and application of non-rodent monoclonal antibodies in veterinary science. Vet. Immunol. Immunopathol., 23: 1-14~ Groves, D.J., Morris, B.A. and Clayton, J., 1987. Preparation of a bovine monoclonal antibody to testosterone by interspecies fusion. Res. Vet. Sci., 43: 253-256. Groves, D.J., Clayton, J. and Morris, B.A., 1988. A bovine monoclonal antibody to oestrone/
EQUINE MONOCLONALIMMUNOGLOBULINS
143
oestradiol prepared by a (murine × bovine ) × bovine interspecies fusion. Vet. Immunol. Immunopathol., 18: 95-101. Hannant, D., Jesset, D.M., O'Neill, T. and Mumford, J.A., 1989. Antibody isotype responses with serum and respiratory tract to primary and secondary infection with equine influenza virus (H3N8). Vet. Microbiol., 19: 293-303. Howard, J.C., Butcher, G.W., Licence, D.R., Galfre, G., Weight, B. and Milstein, C., 1980. Isolation of six alloantibodies against rat histocompatibility antigens: cloned competition. Immunology, 41: 131-141. Hudson, L. and Hay, F.C., 1976. Practical Immunology. Blackwell, Oxford, pp. 88-93. James, K. and Bell, G.T., 1987. Human monoclonal antibody production. Current status and future prospects. J. Immunol. Methods, 100: 5-40. Kennedy, H.E., Jones, B.V., Tucker, E.M., Ford, N.J., Clarke, S.W., Furze, J., Thomas, L.H. and Stott, E.J., 1988. Production and characterization of bovine monoclonal antibodies to respiratory syncytial virus. J. Gen. Virol., 69: 3023-3032. KiShler, G., 1979. Soft agar cloning of lymphoid tumor lines: Detection of hybrid clones with anti-SRBC activity. In: I. Leftkovitz and B. Pernis (Editors), Immunological Methods. Academic Press, New York, pp. 397-401. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227: 680-685. Lin, C.C., Newton, D;R. and Church, R.B., 1977. Identification and nomenclature for G-banded chromosomes. Can. Genet. Cytol., 19: 271-282. McGuire, T.C., Crawford, T.B. and Henson, J.B., 1973. The isolation, characterisation and functional properties of equine immunoglobulin classes and subclasses. In: J.T. Bryans and H. Gerber (Editors), Proceedings of the 3rd International Conference on Equine Infectious Diseases. Karger, Basel, pp. 364-381. Montgomery, P.C., 1973. Molecular aspects of equine antibodies. In: J.T. Bryans and H. Gerber (Editors), Proceedings of the 3rd International Conference on Equine Infectious Diseases. Karger, Basel, pp. 343-363. Mumford, J.A., Hannant, D. and Jessett, D., 1990. Experimental infections of ponies with equine influenza (H3N8) virus by intranasal inoculation or exposure to aerosols. Equine Vet. J., 22: 93-98. Perryman, L.E., O'Rourke, K.I., Mason, P.H. and McGuire, T.C., 1990. Equine monoclonal antibodies recognise common epitopes on variants of equine infectious anaemia virus. Immunology, 71: 592-594. Russell, P.H. and Edington, N. (Editors), 1985. Orthomyxoviridae. In: Veterinary Viruses. Burlington Press, London, pp. 138-145. Tucker, E.M., Dain, A.R., Clarke, S.W. and Donker, R.A., 1984. Specific bovine monoclonal antibody produced by a re-fused mouse/calf hybridoma. Hybridoma, 3:171-176. Tucker, E.M., Clarke, S.W. and M6t6nier, L., 1987. Murine/bovine hybridomas producing monoclonal alloantibodies to bovine red cell antigens. Anim. Genet., 18: 29-39. Weiss, M.C. and Green, H., 1967. Human-mouse hybrid cell lines containing partial complements of human chromosomes and functioning human genes. Proc. Natl. Acad. Sci. USA, 58:1104-1111.
129
The production of equine monoclonal immunoglobulins by horse-mouse heterohybridomas C.M. Richards a, H.A. Aucken b, E.M. Tucke#, D. Hannant c, J.A. Mumfor& and J.R. P o w e l l d aAFRC Institute of Animal Physiology and Genetics Research, Babraham, Cambridge CB2 4AT, UK bPublic Health Laboratory Service, Central Public Health Laboratory, 61 Colindale Avenue, London NW9 5HT, UK CAnimal Health Trust, P.O. Box 5, Newmarket, CB8 7DW, UK dUnilever Research, Sharnbrook, Bedford, UK (Accepted 18 September 1991 )
ABSTRACT Richards, C.M., Aucken, H.A., Tucker, E.M., Hannant, D., Mumford, J.A. and Powell, J.R., 1992. The production of equine monoclonal immunoglobulins by horse-mouse heterohybridomas. Vet. Immunol. Immunopathol., 33: 129-143. Studies were carried out to determine the optimum conditions for the production of equine monoclonal antibodies (MAbs). Lymphocytes from ponies immunised with influenza A equine 2 virus, isolate A/Equine/Newmarket/79 (H3N8) were fused with mouse myeloma (NSO) cells and with horse-mouse heterohybridomas made aminopterin-sensitive by selective growth in 8-azaguanine. Although all fusions initially resulted in heterohybridoma colonies that secreted equine immunoglobulin, many of these were unable to maintain secretion for longer than a few weeks. Increasing the time between immunisation and the booster injection of Newmarket/79 virus, the inclusion of Freund's incomplete adjuvant and the use of an aminopterin-sensitive primary heterohybridoma as the fusion partner, improved the production of HIg-secreting heterohybridomas. After two clonings eight cell lines were established which maintained anti-Newmarket/79 antibody secretion for over a year. FACS analysis of the cell lines provided a useful means of predicting breakdown of MAb secretion by the cell lines, thus enabling re-cloning to be carried out in time. ABBREVIATIONS ELISA, enzyme-linked immunosorbent assay; FCS, foetal calf serum; FIA, Freund's Incomplete Adjuvant; GAH, goat anti-horse; HAT, hypoxoanthine aminopterin thymidine; HIg, horse immunoglobulin; HS, horse serum; Ig, immunoglobin; MAbs, monoclonal antibodies; PBL, peripheral blood lymphocytes; RAH, rabbit anti-horse; SAH, sheep anti-horse; SDS---PAGE, sodium dodecyi sulphate polyacrylamide gel electrophoresis. Correspence to: C.M. Richards, AFRC Institute of Animal Physiology and Genetics Research, Babraham, Cambridge CB2 4AT, UK.
© 1992 Elsevier Science Publishers B.V. All fights reserved 0165-2427/92/$05.00
130
C.M. RICHARDS ET AL.
INTRODUCTION
Although rodent-derived monoclonal antibodies (MAbs) are suitable for a wide variety of different purposes in veterinary immunology, there are some instances where MAbs derived from other species are required. One such instance arises when the antibody is required for immunotherapy in which case it needs to be of the same species as the recipient to avoid triggering an immune response. Non-rodent MAbs are also required when studying host-parasite relationships in other animals as they are more likely to be raised against those parasite antigens seen by the host's own immune system. In the absence of suitable non-rodent myeloma cell lines for fusion, mouse myelomas have been used to produce human-mouse, sheep-mouse or cattle-mouse heterohybridomas and shown to be a successful means of producing cell lines secreting human (James and Bell, 1987), sheep and cattle (Groves and Tucker, 1989) MAbs respectively. Stable antibody-secreting heterohybridomas are, however, more difficult to produce than rodent hybridomas because of their inherent chromosome instability (Weiss and Green, 1967). This instability can sometimes be improved by creating aminopterin-sensitive heterohybridomas and using these as fusion partners in place of the rodent myeloma cell line (Tucker et al., 1984, 1987; Anderson et al., 1986, 1987; Groves et al., 1987, 1988; Kennedy et al., 1988; Flynn et al., 1989). This paper gives results of experiments to determine the feasibility of extending the above studies to produce horse-mouse heterohybridomas and compares their use as fusion partners with that of mouse myeloma cells. The antigen chosen was influenza A equine 2 because antibodies to this virus would be of potential use in diagnosis or treatment of this disease (review by Russel and Edington, 1985 ). MATERIALS AND METHODS
Immunisation
One isolate of the influenza A equine 2 virus A/Equine/Newmarket/79, used as the immunising agent, was prepared and administered at the Animal Health Trust (AHT), Newmarket. The virus was propagated by inoculation into the allantoic cavity of embryonated hens eggs. Welsh Mountain ponies of various ages and both sexes (AHT, Newmarket), were exposed to either an aerosol of Newmarket/79-infected allantoic fluid using a model 65 nebuliser (De Vilbiss, Somerset, PA, USA) as described by Mumford et al. (1990), or given an intramuscular injection of virus. The intramuscular injection of inactivated whole virus antigens or live virus was given as an equal volume of virus at 100/tg ml -~ and adjuvant (Anderson et al., 1987). Initially the adjuvant used was aluminium hydroxide, but for later immunisations this
EQUINE MONOCLONALIMMUNOGLOBULINS
131
was mixed, 1:1, with Freund's Incomplete Adjuvant (FIA). The emulsion was injected into the leg just below the popliteal lymph node to ensure local stimulation.
Cell lines Mouse myeloma NSO cells were used for primary fusions with equine lymphocytes. NSO, a derivative of NS1/Ag 4.1, does not synthesize mouse immunoglobin (Ig) chains. Several non-Ig-secreting heterohybridomas resulting from these fusions were selected for aminopterin-sensitivity by cloning on soft agar containing 20/zg ml-1 8-azaguanine (Tucker et al., 1984; Anderson et al., 1987). These aminopterin-sensitive lines (EQ17, EQ38, EQ42 and EQ43) were produced from fusions using peripheral blood lymphocytes (PBL) of different ponies, and then used as fusion partners in secondary fusions. All cells were maintained in RPMI-1640 medium (Flow Laboratories, Irvine, UK), containing penicillin-streptomycin ( 100 U ml-1 ) (Gibco, Paisley, UK) and foetal calf serum (FCS, Imperial Laboratories, Andover, UK) at 5-20%.
Lymphocytes Blood was collected from ponies into heparinised tubes (Vacutainers, Becton Dickinson UK, Oxford), mixed and allowed to stand for 20 min at 37 °C. The plasma was removed and layered onto Histopaque (Sigma Chemicals, Poole, UK) which contains ficoll 400, and the lymphocytes were separated by density gradient centrifugation at 150 × g for 40 rain in a Centra 7R (Damon/IEC, Dunstable, UK) centrifuge. The lymphocyte-rich layer was removed, washed and resuspended in RPMI-1640 with supplements plus HEPES at a final concentration of 0.01 M (Gibco, Paisley, UK). Lymph nodes and spleen were collected and prepared for fusion according to the method described by Anderson et al. ( 1987 ), using the medium described above.
Freezing and thawing of cells Cells for freezing were spun down at 400Xg for 5 min and resuspended at 2 X 106-107 cells ml-1 in 90% FCS, 10% dimethylsulphoxide (BDH Chemicals, Dagenham, UK). The cells were transferred to 2 ml cryotubes (Nunc, Kamstrap, Denmark), placed in polystyrene boxes and frozen at - 7 0 ° C for 24 h before being placed in liquid nitrogen. The cells were thawed quickly and transferred immediately to a universal tube containing fresh culture medium, spun at 400Xg for 5 min and resuspended in RPMI + 20% FCS.
132
C.M. RICHARDS ET AL.
Fusions Fusions were carried out by mixing equine lymphocytes with mouse NSO cells or equine heterohybridoma fusion partner at a 4: 1 ratio respectively, in the presence of 50% ( w / v ) polyethylene glycol (PEG 1500, Boehringer Mannheim, Germany). The fused cells were washed in RPMI containing 0.01 M HEPES buffer and resuspended in RPMI medium with 20% FCS at 10 6 cells m1-1 and plated out ( 1 ml per well) into 24-well plates (Linbro, Flow Laboratories, Irvine, U K ) . After 24 h in culture, 1 ml of selective medium, to give a final concentration of 10-4 M hypoxanthine, 4 × 10-7 M aminopterin and 1.6× 10 -5 M thymidine (HAT) (Anderson et al., 1987), was added. Control wells were also set up containing unfused myeloma cells and lymphocytes in medium with and without HAT.
Screening culture supernatants Culture supernatants were removed from wells containing fused cells and screened for both the presence of horse immunoglobulin (HIg) and specific antibody to Newmarket/79 virus, by solid-phase enzyme-linked immunosorbent assays (ELISA). The HIg ELISA plates (Falcon, Becton Dickinson UK, Oxford) were coated overnight at 4 °C with affinity-purified goat anti-horse (GAH) IgG (H + L), obtained from Jackson Immunoresearch Laboratories (Westgrave, PA), diluted to 2/~g m l - i in bicarbonate buffer pH 9.6. The antigen used to coat the H3N8 virus-specific assay plates was the third egg passage of Newmarket/79 strain containing 64 haemagglutination (HA) units diluted 1 : 250 in bicarbonate buffer, pH 9.6. HIg and H3N8 plates were then washed four times in phosphate-buffered saline containing 0.05% Tween 20 (Sigma Chemicals, Poole, U K ) , PBST. Culture supernatants were added for incubation at 37°C for 1 h, washed as above and incubated for 1 h in the presence of affinity-purified, alkaline phosphatase-conjugated GAH IgG (H + L) (Jackson Immunoresearch Laboratories, Westgrave, PA) diluted 1 : 5000 in PBST. After four more washes, p-nitrophenyl phosphate disodium (Sigma Chemicals, Poole, U K ) in substrate buffer (0.5 M NaHCO3 + 0.002 M MgCI2), pH 9.5, at 1 mg ml -l was added to each well, and the plates were incubated for a further hour at 37 ° C. Optical densities were read on a Titertek Multiskan automatic reader at wavelengths 405 and 510 nm absorbance. Positive controls included serum from a hyperimmune pony (Day 14 postH3N8 virus infection ), diluted 1 : 2000 in culture medium, and purified horse IgG (H + L) (Jackson Immunoresearch Laboratories, Westgrave, PA) diluted to 10 #g m l - 1. Culture medium was also used as the negative control in both H3N8 antibody and HIg assays.
EQUINE MONOCLONALIMMUNOGLOBULINS
133
Cloning Where possible, single distinct colonies were picked out of those 2-ml wells assayed by ELISA as positive for H3N8-specific antibody and transferred to 0.2 ml or 2 ml wells. Alternatively all cells from a positive well were maintained and cloned on soft agar according to the method described by KiShler (1979). All selected cell lines were cloned twice. Cloning efficiency was assessed as the percentage of anti-H3N8 antibody-secreting clones obtained after cloning.
Fluorescence-activated cell sorter To determine the cell surface expression of the HIg produced by the different heterohybridoma lines, cells were analysed using a FACScan (Becton Dickinson UK, Oxford). Cells ( 106) were incubated in 100 #1 of affinitypurified, fluorescein-conjugated GAH IgG ( H + L ) (Jackson Immunoresearch Laboratories, Westgrave, PA) diluted 1 : 5 in RPMI plus 5% FCS, for 45 min at 4 ° C. The cells were then washed twice with 100/tl culture medium and resuspended in 200/A PBS. Negative controls included non-Ig secreting lines and lines incubated in the absence of the fluorescent conjugate. The level of fluorescence above that of the negative controls was then measured for each line.
Chromosome analysis Cells grown in 50 ml tissue culture flasks were treated with Colcemid, (Gibco, Paisley, U K ) , at a final concentration of 0.06/~g m l - 1 and incubated for 4 h at 37°C, 5% COz. The cells were harvested, spun at 4 0 0 × g for 5 min and resuspended in 5 ml of 75 mM KC1, mixed and spun as above. The pellet was resuspended in 5 ml of fixative (3 parts methanol: 1 part glacial acetic acid) at room temperature for 15 min and then washed three times in fixative, reducing the volume successively to 0.5-1.0 ml. Two to three drops of the fixed cell suspension were dropped onto clean slides which had been washed and immersed briefly in 60% methanol. The slides were passed rapidly through a bunsen flame and stored at room temperature for 1 week prior to Giemsa staining (Lin et al., 1977).
Purification of monoclonal immunoglobulins Culture supernatants from the heterohybridoma cell lines and control 20% FCS in RPMI were concentrated ten times with Amicon Hollow Fibre concentrator with a molecular weight (MW) cut-off of 30 kDa and affinity-purified on a Protein G Sepharose 4 Fast Flow column (Pharmacia L.K.B. Bio-
134
C.M. RICHARDS ET A L
technology, Milton Keynes, UK) according to the method described by Bj6rck and Kronvall (1984).
Immunoglobulin analysis The immunoglobulins secreted by the heterohybridoma cell lines were radiolabelled over 16-18 h in 250/~1 lysine-free RPMI (Flow Laboratories, Irvine, UK) containing 50 /tCi m1-1 of [14C]-lysine monohydrochloride (Amersham, U K ) . The cells were spun at 400Xg for 5 min, and the supernatants collected and stored if necessary at - 2 0 ° C. These supernatants were then mixed 1 : 1 in sample buffer containing fl-2-mercaptoethanol and bromophenol blue, boiled for 3 min and examined by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE), as described by Laemmli (1970). Autoradiography was performed by exposure of the gels to Hyperfilm-B max (Amersham, U K ) and the size of the Ig chains estimated with respect to molecular weight markers (Sigma Chemicals, Dorset, U K ) also run on the gel. Monoclonal Ig samples, concentrated and purified on a Protein G column, were analysed by double immunodiffusion tests. Samples were added to wells in 1% barbitone agar (Hudson and Hay, 1976 ) for immunodiffusion against rabbit anti-horse (RAH) IgG (provided by Cambridge University Veterinary School) overnight at 37 ° C. RESULTS
Production of mouse-horse heterohybridomas Initial fusions were carried out between the mouse myeloma line NSO and equine lymphocytes (Table 1 ). For the first two fusions, PBL were taken from ponies 4 days after contact with ponies shedding the H3N8 virus (Table 1 ). One of these, Pony 37, was naive, i.e. it had not been exposed to the H3N8 virus previously, while the other, Pony 20, was injected with inactivated whole virus 4 and 2 weeks prior to in-contact infection. Both fusions resulted in hybrid colonies of which 25% and 10% respectively, were shown by ELISA to secrete HIg, but none was specific to H3N8 virus. With the aim of improving the immunisation procedure and source of lymphocytes, fusions were then carried out between NSO cells and lymphocytes taken from lymph nodes and spleen of a previously immunised pony (W5) 4 days after a secondary aerosol challenge and an intramuscular booster injection of live H3N8 virus in the region of the inguinal lymph node (Table 1 ). The growth of hybrid colonies varied considerably but fewer hybridomas were formed and fewer Ig-secretors were present than in the previous fusions car-
135
EQUINEMONOCLONALIMMUNOGLOBULINS TABLE 1 Production of horse-mouse heterohybridomas Pony
Horse lymphocytes
Wells with colonies (%)
Mean colonies per well
Secreting HIg (%)
371 202
PBL PBL
96 96
4.6 2.9
25 10
W53
Bronchial LN Mediastinal LN Inguinal LN Spleen
66 48 33 4
1.8 0.7 0.3 0.08
3 8 10 4
tPBL fused with NSO cells 4 days after in-contact infection with ponies secreting H3N8 virus. 2pBL fused with NSO cells 4 days after in-contact infection, 4 and 2 weeks after i.m. injection with H3N8 virus. 3Lymphocytes removed for fusion 4 days after an aerosol and i.m. injection of H3N8 virus, 8 weeks after a primary aerosol challenge. All HIGgG secretors were negative for H3N8-specific antibody when tested by ELISA.
ried out with PBL. The spleen fusion was particularly poor resulting in only 4% of the wells containing hybrid colonies. Aminopterin-sensitive horse-mouse heterohybridomas were prepared by further fusions of NSO cells with PBL followed by selection in 8-azaguanine. The ponies used to provide the PBL (W4, W2, 277 and 289) had been previously immunised with H3N8 virus but none was challenged immediately before fusion. Four non-HIg secreting hybrids were chosen for use themselves as fusion partners.
Comparison of fusion partners NSO cells and the four aminopterin-sensitive heterohybridoma fusion partners (EQ 17, EQ38, EQ42, EQ43 ) were fused with PBL from two immunized ponies (W2 and W4) 4 days after either an intravenous or intramuscular booster injection of Newmarket/79 virus respectively (Table 2). The fusion efficiencies for the different fusion partners were extremely variable, but for each pony, the overall best fusion partner in terms of percentage growth and mean colonies per well was the EQ 17 heterohybridoma. Both EQ 17 fusions produced some anti-Newmarket/79 secretors but these only maintained secretion for 2-3 weeks after fusion. Eight further fusions were carried out using PBL from four different immunised ponies for comparison of EQ 17 and NSO cells as fusion partners. Although the results (not shown) of the fusions varied slightly, in each case EQ 17 gave better fusion efficiencies and a higher number of Ig-secreting lines than NSO. The EQ 17 fusion with Pony W2 lymphocytes (Table 2 ) gave rise
136
C.M. R1CHARDS ET AL.
TABLE 2 Comparison of aminopterin-sensitive heterohybridomas and mouse myeloma NSO cells fused to PBL from immunised ponies Cell line used in fusion
Pony
Wells with colonies (%)
EQ 17 EQ38 EQ42 EQ43 EQI 7 NSO
W21 W2 W2 W2 W42 W4
100 100 25 10 100 100
Mean colonies per well 20.0 9.0 0.2 0.1 > 30 7.4
HIg secretors (%)
H3N8 antibody secretors (%)
17 6 2 0 47 27
2 0 0 0 17 0
ILymphocytes used 4 days after an i.v. injection, 13 and 11 months post-aerosol challenge. 2Lymphocytes used 4 days after an i.m. injection, 6 and 3 months post-aerosol challenge.
to more than 30 hybrid colonies per well, and 43% of wells with HIgG activity, compared with 7.4% colonies per well and 27% HIgG secretion for the relative NSO fusion. From this it was concluded that EQ 17 would be a better fusion partner than NSO for subsequent experiments.
Fusions using lymphocytes from different sources It was thought that failure to produce sufficient cells secreting antibody to H3N8 virus could be due to the fact that antibody levels remained high in the immunised pony long after the initial challenge, thus resulting in a poor secondary immune response (Hannant et al., 1989). A pony (277) was therefore rested for 27 months after secondary challenge (i.e. until antibody levels were low), and then given two intramuscular booster injections of Newmarket/79, close to the left sub-mandibular and popliteal lymph nodes. Both stimulated and unstimulated lymph nodes, PBL and spleen cells were fused with the EQ 17 fusion partner, 4 days post-boost. Table 3 summarises the resuits of these fusions. Good growth was observed in all fusions and therefore, because of the large number of samples for testing, only H3N8-specific assays were carried out. The PBL fusion produced no colonies secreting anti-H3N8 antibody, but all other fusions gave some H3N8-specific antibody secretors. As expected, the stimulated lymph nodes gave the highest number of anti-H3N8 secretors, of which two lines (A and B) were successfully expanded, cloned and maintained. Some of the excess lymphocytes from these lymph nodes were frozen in liquid nitrogen and used in later fusions (Table 3 ) which resulted in one more stable antibody-secreting cell line (line C ). A second previously vaccinated pony (83) was given an intramuscular boost 12 months after secondary challenge and the stimulated lymph nodes were
EQUINEMONOCLONALIMMUNOGLOBULINS
137
TABLE 3 Fusion efficiency of lymphocytes from different sources fused with aminopterin-sensitive EQ 17 cells Pony
Lymphocytes
Mean colonies per well
Secreting anti-H3N8 Week 3 (%)
Lines maintained
277 t
PBL S popliteai L.N. US popliteal L.N. S sub-mandibular L.N. US submandibular L.N. Spleen
12.2 17.3 13.5 15.4
0 38 2 20
0 2 0 3
S popliteal L.N. stored at -196oc S sub-mandibular L.N. stored at -196°C 832
S popliteal L.N. S mandibular L.N. S inguinal L.N.
Lines stable
A B
35.7
6.25
0
37.7
1.4
1
16.5
5.0
0
2.5
1.4
1
C
4.1 5.54
4.2 6.8
0 2
DE
38.5
3
FGH
12.5
~Lymphocytes removed for fusion 4 days post-i.m, injection, 30 and 27 months post-aerosol challenge. 2Lymphocytes removed for fusion 4 days post-i.m, injection, 30 and 12 months post-aerosol challenge. S, stimulated lymph node by local injection of virus; US, not injected locally; Lines maintained, secreted antibody for 2-3 months; lines stable, secreted antibody for over 1 year.
removed for fusion 4 days later (Table 3). A further five lines (D-H) were thus expanded, cloned and maintained in culture, giving a total of eight equine heterohybridoma cell lines secreting antibody specific for influenza A equine 2 virus.
Stability of horse-mouse heterohybridomas secreting specific antibody Cell lines A-H cloned twice on soft agar remained positive when assayed for H3N8 and HIgG specificity by ELISA for over 12 months in continuous culture. These lines were also tested by FACS analysis in order to calculate the percentage of cells in the population producing HIg. Table 4 gives the FACS results obtained immediately before the second cloning, i.e. approximately 6 months after fusion, for comparison with the cloning efficiencies (i.e. percentage of wells positive after cloning). In most cases the correlation between the percentage of cells with surface HIg (10-88%) and cloning efficiency (2-90%) was good, suggesting that FACS analysis is a useful test of
138
C.M. RICHARDS ET AL.
TABLE 4 Stability of heterohybridomas as judged by FACS analysis and second cloning of cell lines Hybridoma cell line
FACS 1% surface HIg
Cloning 2 efficiency (%)
A B C D E F G H
30 88 45 30 70 41 10 35
28 73 2 95 83 53 90 36
~FACS analysis of cell lines cloned once on soft agar, immediately before second cloning. 2Second cloning carried out 6-8 months after fusion.
HIg and MAb production. There were exceptions, however; line C had a very poor cloning efficiency, although 45% of its cells were HIg-positive by FACS. Conversely, lines D and G gave much higher cloning efficiencies than would be predicted by the FACS measurements; this may be due to the variability of the surface Ig expression. Some of the re-clones resulting from the above experiment were also analysed on the FACScan. All showed some improvement in the percentage of cells expressing HIg. In particular, line A increased from 30 to 50%; line C 45-96%; line G 10-33% and line H 35-76%. This indicated the importance of cloning for the maintenance of antibody-secreting heterohybridoma cell lines.
Immunoglobulin analysis 14C-lysine-labelled supernatants from lines A - H and EQ 17 were analysed by SDS-PAGE autoradiography (Fig. 1 ). The molecular weight of the heavy chains indicated that with the exception of line D, all samples were in the range of the IgG isotype, i.e. 48-52.5 kDa (Montgomery, 1973; McGuire et al., 1973 ). Even though line D was an antibody secretor it appeared to lack a heavy chain. This would indicate that the secretion of heavy chain in this sample was below detection by SDS-PAGE. The light chains present in all samples showed slight variation between the molecular weights of 24.5-27 kDa. Such variation has also been reported in both rodent (Howard et al., 1980 ) and cattle (Anderson et al., 1987 ) Ig light chains, and is thought to be due to variation in the carboxyl groups. Figure 2 illustrates precipitin reactions between seven of the eight monoclonal Ig samples with RAH IgG. Reactions of identity were observed be-
139
EQUINE MONOCLONAL IMMUNOGLOBULINS
H
G
F
E
D
C
B
A
EQ17
MW
-180 -116 84
Heavy
W
m
58 48.5
36.5 26.5 Light
~ --
Fig. 1. Autoradiograph of SDS-PAGE of 14C-labelled HIg chains in culture supernatants. Fusion partner EQ 17 showed no evidence of HIg secretion. The anti-H3N8 secreting lines A - H with the exception of D show heavy (48-52.5 kDa) and light chains (24.5-27 kDa). Line D did not give a visible heavy-chain band.
Fig. 2. Double immunodiffusion tests of the monoclonal Ig samples A-H, concentrated and purified on a Protein G column against RAH IgG. Control samples included concentrated, purified 20% FCS and normal horse serum (HS).
tween A and B, C and E as well as F, G and H, while a partial reaction was observed between B and C. This suggests the presence of at least two subtypes of HIgG. No precipitin lines were observed with the control FCS. Line D was again negative, but this was probably due once again to the low concentration of heavy chain. Immunodiffusion tests were also carried out with sheep anti-horse (SAH)
140
C.M. R1CHARDS ET AL.
IgG (Evaibios Laboratories, Petworth, U K ) and GAH IgG with similar results. Both FCS and normal mouse serum showed no reaction, indicating that the MAb Ig samples were of horse origin.
Chromosome analysis Previous experiments with cattle-mouse heterohybridomas had indicated that a possible explanation for the superiority of some heterohybridomas as fusion partners was that they retained more cattle chromosomes than others (Anderson et al., 1987 ). The chromosomal content of seven of the heterohybridoma cell lines including EQ 17 was therefore determined by examination of Giemsa-stained preparations of metaphase spreads. The equine karyotype consists of 64 chromosomes while NSO has approximately 65. The latter can be distinguished from horse chromosomes when stained in Giemsa by their dark centromeres compared with the clear centromeres of the horse chromosomes. Figure 3 shows a metaphase spread from a heterohybridoma in which approximately 25 out of the 90 chromosomes seen were horse. One non-HIg-secretor tested had a mean of 6% equine chroiii!i
¸3!3¸
~iiiiiii
M - Mouse H - Horse
Fig. 3. Giemsa-stained metaphase spread of a heterohybridoma cell line ( × 100). Twenty-five equine chromosomes, some of which have been indicated (H), are identified in this spread in a total of 90 chromosomes.
EQUINE MONOCLONALIMMUNOGLOBULINS
141
mosomes and the mean of each of seven HIg secreting lines tested ranged from 15 to 23%. EQ 17 fusion partner, another HIg non-secretor, had between 10 and 16 horse chromosomes out of a total of 116-129, giving an average of 11%. There was therefore no clear indication that the number of equine chromosomes retained was relevant to hybridoma efficiency. DISCUSSION Appleton et al. ( 1989 ) and Perryman et al. ( 1990 ) have described the production of cell lines secreting an equine MAb to influenza A equine l (H7N7) and equine infectious anaemia viruses respectively by fusing mouse myeloma cells with stimulated PBL and spleen cells from immunised ponies. The results reported in the present paper confirm and extend the feasibility of the heterohybridoma approach for the production of monoclonal equine Ig. In agreement with previous results using sheep-mouse (Flynn et al., 1989 ) and cattle-mouse (Tucker et al., 1984, 1987; Anderson et al., 1986, 1987 ) hybridomas, there was convincing evidence that a horse-mouse fusion partner (EQ 17 ), made HAT-sensitive by selective growth in 8-azaguanine, was more efficient than mouse myeloma cells alone in producing stable antibody-secreting lines. The reason for the particular efficiency of EQ 17 (M × H) cell line in this respect is unclear. Chromosome analysis of the EQ17 line has shown that it does not retain a particularly high percentage of horse chromosomes compared with other lines. Also, recent tests indicate that lines produced from secondary fusions in which EQ 17 was the fusion partner, do not retain more equine chromosomes than those from primary fusions. One clear observation from the present series of experiments was the importance of selecting the method and timing of immunisation in order to produce well-stimulated lymphocytes for fusion. A characteristic of the response to primary challenge with influenza A equine 2 virus is the sustained subsequent presence of antibody in the peripheral circulation. To achieve maximal stimulation from a booster injection it would appear to be necessary to wait until the primary antibody response has waned. In the case of the ponies used in the present study, this took many months. As has been found with the production of bovine MAbs to E. coli (Anderson et al., 1987 ) and alloantigens (Tucker et al., 1987 ), it was clear that best results were obtained with lymphocytes derived from locally stimulated lymph nodes. FACS analysis, although not specific for antibody to H3N8 virus, gave useful information about the Ig-secretion of each cell line. Although FACS results did not always correlate with cloning efficiencies, regular FACS analysis proved a useful means of determining the stability of a line and indicating when cloning may be necessary. Owing to the presence of residual Ig in the culture supernatants, the ELISA will not indicate loss of Ig-production immediately, and by the time they have done so, it may be too late to clone.
142
C.M. R I C H A R D S ET AL.
Analysis o f the monoclonal HIg samples by S D S - P A G E and double imm u n o d i f f u s i o n tests with R A H IgG suggested that at least seven o f the eight MAb lines are of the IgG isotype. The slight variation observed in the size of the heavy chain on the S D S - P A G E together with the reaction of partial identity observed by i m m u n o d i f f u s i o n m a y be due to the presence of different IgG subtypes. Previous work on HIg (Montgomery, 1973; McGuire et al., 1973 ) has indicated that at least four subtypes o f IgG exist, IgGa, IgGb, IgGc and possibly IgG (T). Further characterisation will therefore be required before the exact isotype o f each MAb-secreting line produced can be determined and it is hoped that these and other cell lines can be used for the i m m u n o chemical analysis of different HIg classes. As a result of experiments described in this paper, it has been possible to create not only cell lines producing HIg of undefined specificity but also lines producing MAb to influenza A equine 2 virus. We attribute this success to the immunisation schedule, the use of lymphocytes from locally stimulated lymph nodes, and the use of h e t e r o h y b r i d o m a fusion partner EQ 17, and plan to use these MAbs for further analysis o f the influenza A equine 2 virus. ACKNOWLEDGEMENT This work was supported in part by a grant from Centaur, R i c h m o n d , VA, USA, and Unilever Research, Sharnbrook, Bedford, UK.
REFERENCES Anderson, D.V., Clarke, S.W., Stein, J.M. and Tucker, E.M., 1986. Bovine and ovine monoclonal antibodies to erythrocyte membrane determinants, produced by interspecific (hetero-) myelomas.Biochem. Soc. Trans., 14: 72-73. Anderson, D.V., Tucker, E.M., Powell,J.R. and Porter, P., 1987. Bovine monoclonalantibodies to the F5 (K99) pilus antigen ofE. coli, produced by murine/bovine hybridomas. Vet. Immunol. Immunopathol., 15: 223-237. Appleton, J.A., Gagliardo,L.F., Antczak, D.F. and Poleman, J.C., 1989. Production of an equine monoclonal antibody specific for the H 7 haemagglutininof equine influenza virus. Vet. Immunol. Immunopathol., 23: 257-266. Bj6rck, L. and Kronvall, G., 1984. Purification and some properties of streptococcal Protein G, a novel IgG-bindingreagent. J. Immunol., 133: 969-974. Flynn, J.N., Harkiss, G.D. and Hopkins, J., 1989. Generation of a sheepX mouse heterohybridoma cell line (IC6.3a6T. l D7) and evaluation of its use in the production of ovine monoclonal antibodies. J. Immunol. Methods, 121: 237-246. Groves, D.J. and Tucker, E.M., 1989. The production and application of non-rodent monoclonal antibodies in veterinary science. Vet. Immunol. Immunopathol., 23: 1-14~ Groves, D.J., Morris, B.A. and Clayton, J., 1987. Preparation of a bovine monoclonal antibody to testosterone by interspecies fusion. Res. Vet. Sci., 43: 253-256. Groves, D.J., Clayton, J. and Morris, B.A., 1988. A bovine monoclonal antibody to oestrone/
EQUINE MONOCLONALIMMUNOGLOBULINS
143
oestradiol prepared by a (murine × bovine ) × bovine interspecies fusion. Vet. Immunol. Immunopathol., 18: 95-101. Hannant, D., Jesset, D.M., O'Neill, T. and Mumford, J.A., 1989. Antibody isotype responses with serum and respiratory tract to primary and secondary infection with equine influenza virus (H3N8). Vet. Microbiol., 19: 293-303. Howard, J.C., Butcher, G.W., Licence, D.R., Galfre, G., Weight, B. and Milstein, C., 1980. Isolation of six alloantibodies against rat histocompatibility antigens: cloned competition. Immunology, 41: 131-141. Hudson, L. and Hay, F.C., 1976. Practical Immunology. Blackwell, Oxford, pp. 88-93. James, K. and Bell, G.T., 1987. Human monoclonal antibody production. Current status and future prospects. J. Immunol. Methods, 100: 5-40. Kennedy, H.E., Jones, B.V., Tucker, E.M., Ford, N.J., Clarke, S.W., Furze, J., Thomas, L.H. and Stott, E.J., 1988. Production and characterization of bovine monoclonal antibodies to respiratory syncytial virus. J. Gen. Virol., 69: 3023-3032. KiShler, G., 1979. Soft agar cloning of lymphoid tumor lines: Detection of hybrid clones with anti-SRBC activity. In: I. Leftkovitz and B. Pernis (Editors), Immunological Methods. Academic Press, New York, pp. 397-401. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227: 680-685. Lin, C.C., Newton, D;R. and Church, R.B., 1977. Identification and nomenclature for G-banded chromosomes. Can. Genet. Cytol., 19: 271-282. McGuire, T.C., Crawford, T.B. and Henson, J.B., 1973. The isolation, characterisation and functional properties of equine immunoglobulin classes and subclasses. In: J.T. Bryans and H. Gerber (Editors), Proceedings of the 3rd International Conference on Equine Infectious Diseases. Karger, Basel, pp. 364-381. Montgomery, P.C., 1973. Molecular aspects of equine antibodies. In: J.T. Bryans and H. Gerber (Editors), Proceedings of the 3rd International Conference on Equine Infectious Diseases. Karger, Basel, pp. 343-363. Mumford, J.A., Hannant, D. and Jessett, D., 1990. Experimental infections of ponies with equine influenza (H3N8) virus by intranasal inoculation or exposure to aerosols. Equine Vet. J., 22: 93-98. Perryman, L.E., O'Rourke, K.I., Mason, P.H. and McGuire, T.C., 1990. Equine monoclonal antibodies recognise common epitopes on variants of equine infectious anaemia virus. Immunology, 71: 592-594. Russell, P.H. and Edington, N. (Editors), 1985. Orthomyxoviridae. In: Veterinary Viruses. Burlington Press, London, pp. 138-145. Tucker, E.M., Dain, A.R., Clarke, S.W. and Donker, R.A., 1984. Specific bovine monoclonal antibody produced by a re-fused mouse/calf hybridoma. Hybridoma, 3:171-176. Tucker, E.M., Clarke, S.W. and M6t6nier, L., 1987. Murine/bovine hybridomas producing monoclonal alloantibodies to bovine red cell antigens. Anim. Genet., 18: 29-39. Weiss, M.C. and Green, H., 1967. Human-mouse hybrid cell lines containing partial complements of human chromosomes and functioning human genes. Proc. Natl. Acad. Sci. USA, 58:1104-1111.
