Journal oflmrnunologtcal Methods, 81 (1985) 283-297 283 Elsewer JIM03554 Biochemical Aspects of Immunotoxin Preparation O Gros 1, p. Gros, F.K Jansen and H Vldal Centre de Recherche Chn- Mid), Groupe Sanofi, 34082 Montpelher Cedex, France (Recexved 8 January 1985, accepted 26 March 1985) The preparation of lmmunotoxms, hybrid proteins formed by dxsulfide bonding an antibody and the A-chain of ncm, has been studied m detail Optxmalcondxtlons, both for the modfflcatxonof the anubody and the couphng reaction between the modified antlbodxes and the toxin subumt, have been deterrmned Con&tlons of time. temperature and stolchlometrystu&ed suggested 2 protocols for each of the 2 steps of thxs preparation Purification and analysxs of the physlcochemlcal and blochermcal properties of the products yielded well-characterized agents, hkely to be of value m chnlcal studxes Key words tmmunotoxm - rwm - modtfwd anttbodws Introduction Immunotherapy of cancer, using 'magic bullets', formed by covalently hnkmg specific or selectwe anti-tumor antlbo&es to highly cytotoxlc agents, combines the respectwe advantages of the toxxclty of the agents with the specificity of the antibodies A prerequisite for such therapy is the preparatxon of well-characterized, reproducible and fully actwe products The numerous stu&es that have been carried out m this field by our group (Jansen et al, 1980, 1982, Blythman et al, 1981, Casellas et al, 1982, 1984) and several others (Nevalle and Youle, 1982, Thorpe et al, 1982, Vxtteta et al, 1982, Colombattl et al, 1983) suggest that the lmmunotoxms which are formed by covalent bonding, through a &sulfide bridge, monoclonal anttbo&es and toxins having no cell-binding potency are among the most promaslng developments m this field of treatment Several methods for couphng the 2 protein moieties have been suggested, aiming to prevent polymenzatton a n d / o r mactwatlon of the constituent proteins (Blair and i To whom repnnt requests should be addressed Abbrevlattons BGG, bovine gamma globuhn, DTNB, 5,5'-dxttuobls(2-mtrobenzolcacid) (Ellman's reagent), EDC, N-ethyl,N'(3-dlmethylarmnopropyl)carbodunude,IgG, lmmunoglobuhn G, 2-ME, 2mercaptoethanol, PDPA, 3(2-pyndyldlthlo)proplomc acid, SPDP, N-succmlmldyl,3(2-pyndyldltbao)proplonate, TCA, tnchloracetlc acid, 2-TP, 2-tbaopyndone 0022-1759/85/$03 30 © 1985 Elsevier Science Publishers B V (Blome&calDivision) 284 Ghose, 1983, Krohck et al, 1983) The most frequently adopted methods are represented by the following reactions Ab + RSSZ ~ Ab - SSZ (reaction 1) Ab - SSZ + T-SH --* Ab - SST + Z-SH (reaction 2) where Ab stands for the anUbody and T-SH for the toxin containing the required tinol radical Tins radical can either be naturally occurring in the protein or artxficmlly introduced by applying a reaction similar to reaction 1 to the native toxin, followed by reduction (or tinol-dlsulfide interchange) of the mixed disulfide obtained The reagent RSSZ is a heteroblfunctlonal agent winch enables the reactions involving each of its functions to be klnetlcally and stoichlometrlcally controlled, so that reactions 1 and 2 can occur sequentially and unlvocally The vast majority of researchers use the N-hydroxysuccinlmldyl ester (SPDP) of 3-(pyradyldithao)proplonic acid as a heteroblfunctlonal agent (Carlsson et al, 1978), whereas we generally use the same acid (PDPA), not in the form of an activated ester, but activated by previous reaction with a lamated amount of a watersoluble carbodiamlde, such as 1-ethyl-3-dlmethylarmnopropyl carbodxlmade (EDC) In this latter case. reaction 1 breaks down into 2 successive steps as shown below ~-SS-(CH 2 )2-COOH (PDPA) ' + ~SS-(CH 2 )2-COO C=NR 2 I NHR1 R 1 - N = C = N - R 2 (EDC) ( R - SSZ) R - S S Z + Ab ~ A b - S S Z (la) (lb) An excess of PDPA with respect to EDC prevents any side-reactions due to EDC cross-hnklng of proteins In most of our work, the A-chain of ricin has acted as the effectometer in the lmmunotoxln The separate functions of the 2 ricln subunlts A and B (Olsnes et al, 1974) originally hnked by a single &sulfide bridge suggests the possibility of coupling the A-chain using its free thlol group to the modified antibodies The restrictive property of the A-chain to recognize a strictly intracellular ribosomal target enables the hybrid molecule to dlSCrlrmnate by its antibody moiety between target cells and non-target cells We have studied the various stages In the synthesis of the lmmunotoxln, and exarmned the different parameters which may be important In establlsinng a standard production process In addition the physlcochemlcal and blologmal characteristics of the products obtained have been studmd 285 Materials and Methods The murlne monoclonal antl-DNP antibody (2F7) was obtained as previously described (Jansen et al, 1980) The murme monoclonal anh-T1 antibody (T101) (Royston et al, 1980) was purchased from Hybrltech (San Diego, CA) The hybridoma producing the rat monoclonal antx-Thy 1 2 antibody (AT15E) was obtained from Dr Colombatti (Lausanne) Rlcln A-chain was prepared and purified on a scale of 10 g per batch (Vidal, 1981) Pyndyl dlthloproplonlc acid (PDPA) was prepared by condensation of 3-mercaptoproplonlc acid and 2-pyndylsulfenyl chloride (Noel and Plchat, 1983) SPDP was purchased from Pharmacia BGG was purchased from Sigma All other reagents were of analytical grade Determmatwn of thtol and pyndyldtsulfide groups Thiol determinations were made according to the Ellman method using DTNB (Anderson and Wetlaufer, 1975) Pyndyldlsulfide groups were measured by spectrophotometric deterrmnatlon of the released 2-ttuopyrldone at 343 nm 2-Thaopyrldone is released from SSP groups by thlol disulfide exchange, an the presence of a 100-fold molar excess of a thlol contalmng reagent (2-mercaptoethanol, cysteme or glutathlone) Rwm A-cham methylatton Rlcin A-chain was 3H-dlmethylated according to a modification of the method of Means and Feeney (1968) Briefly, ricm A-chain was reacted for 1 h with formaldehyde and boro[3H]hydnde at 0°C, m a 0 2 M borate buffer, pH 10 Reagents in excess were removed by dialysis or gel filtration The thtol group of the modified A-chain was then restored by a 1 h reaction with 2-mercaptoethanol (2-ME) at 1% final concentration, at room temperature, followed by dialysis Molecular wetght evaluatton Preparations were analyzed by polyacrylamlde gradient gel electrophoresis (PAA2-16, Pharmacla) using by Coomassie blue G250 staining Preparations were reduced as reqmred with 2-ME before electrophoresls The electrophoretic profiles were quanutated by densltometry Computertzed analysts of the stattsttcal dtstrtbutton of hybrtd molecules m I T preparattons The areas corresponding to the 4 major components of IT preparations, obtained as percentages from densltometrlc recordings, were transformed into molar percentages Linear regression was then carned out for the pairs (n, log[p.(n v)]) where n was the number of coupled A-chams per lmmunoglobuhn molecule m the studxed component and p. the molar percentage of this protein in the immunotoxm solution Linear representations are consistent with Poisson's law Calculated percentages qn (according to Polsson's formula) were compared with the observed pn values using linear regression analysis 286 Enzymatw A-chain actlvttr Preparations were tested by incubation with a cell-free system from rat hver supplemented with poly-U and ~4C-Phe for 15 man at 28°C, in the presence of 2-ME as required Proteins were then precipitated by the ad&tlon of TCA, collected on filters and their radioactivity determined A-chain actlwty was then measured as the concentration glwng 50% m h i b m o n of protein synthesis (denoted IC50) Immunoglobuhn modtftcatton wtth PDPA A 21 m g / m l solution of PDPA m 2-methyl-2-propanol was maxed with a 64 m g / m l aqueous solution of EDC at a ratao (EDC PDPA) of 1 5 ( v / v ) After 5 nun at room temperature, the correct volume of this mixture, calculated according to the Stolchlometry chosen, was added to the buffered IgG solution (1-10 m g / m l m 125 m M phosphate buffer, p H 7 0) Fifteen nun later, the modified IgG was purafled by daalysls at 4°C against the same phosphate buffer Immunoglobuhn modzftcatwn with SPDP Thas modification was performed as suggested by the suppher (Pharmacla) Briefly SPDP solution in ethanol was mixed w~th I g G solution in phosphate buffer p H 7 0 and reacted for 30 nun at 30°C before purification by either dialysis or gel filtration Couphng reaction and trnmunotoxm punflcatton Mo&fied I g G and A-chain (both in phosphate buffer soluuon at 1-10 m g / m l ) were nuxed and reacted overnight before purification by usual gel filtration on Sephadex G-100 or by H P L C on a T S K SW 3000 column Results (A) Immunoglobuhn modtftcatlon I g G modification by PDPA previously activated by EDC was completed in 10 nun at 20°C whatever the stoIchlometric ratio (Fig 1) In the case of T101 antibody at 4 4 m g / m l , a 30 1 PDPA I g G molar ratio proxaded an IgG modified by about 1 0 SSP radical per mol, whale a 45 1 ratio provided an IgG mo&fied by about 3 7 SSP radicals per mol At 0°C the mo&ficatlon was slowed down and reduced by about 70% since only 1 2 SSP radicals were introduced per I g G after 30 nun of reaction with an excess of 45 molecules of PDPA SPDP modlficataon of I g G was not so rapad (Fig 2) At 30°C, modificataon rataos of 6 0 and 4 0 SSP radicals per I g G molecule were obtained after 30 nun of reaction with 12 and 8 equavalents of SPDP respectively At 0°C the modxficataon was almost antublted since the modification rauo obtained after 30 nun with 8 eqmvalents of SPDP was only 0 6 SSP radical per IgG molecule The modification ratio of the I g G by PDPA also varied accordang to the I g G concentration and the nature of the antibody (Tables I and II) With the same stoachaometry and the same concentration of IgG, the modification raUo reached 287 °c x. n tO to .... 1 / o C I // / P t Minutes Fig 1 K m e u c study of antibody mo&ficatlon with P D P A 1 5 m g of pure IgG T101 (4 m g / m l ) was allowed to react at 20°C wath P D P A prevaously activated by a hmated quantity of water-soluble carbodnrmde (EDC) (see Materials and Methods) At each tlmepomt the antibody soluuon was added to a 0 2 M glycme solution m order to stop IgG modlflcauon The extent of mod~flcauon was measured after dialysis by reduction of introduced activated disulfide groups, releasing 2 tbaopyndone winch absorbs at 343 n M Two stolctuometrlc ratios were stu&ed 30 ( I I ) or 45 (D D) P D P A per IgG, actwated by 20 and 30 E D C per IgG respectively Dashed curve ( n . . . . . . n) indicates the same modification c a m e d out w]th 45 P D P A at 0°C instead of 20°C 5 to to 4 3 :c 2 C 1 41 3 10 30 M,nutes Fig 2 Kanetic study of antibody mo&ficatlon with SPDP 1 mg of pure IgG T101 (4 m g / m l ) was allowed to react with SPDP At each tlmepoint the antibody solution was added to a 0 2 M glyone solution m order to stop IgG mo&hcatlon The extent of mo&ficatlon was measured after dialysis as indicated m Fig 1 Two stolchlometnc ratios were studied 8 ( I I ) or 12 (rn rn) SPDP per IgG Dashed curve ( I . . . . . ! ) indicates the same mo&ficatlon carried out with 8 SPDP at 0°C instead of 300C 288 TABLE I ANTIBODY M O D I F I C A T I O N I N F L U E N C E OF THE MONOCLONAL ANTIBODY Mo&ficatlon was carried out as described at a 45 1 PDPA lgG ratio using several monoclonal antlbo&es (IgG concentration 2 m g / m l ) Modification ratios obtained w~th TI01 IgG were determined w~th standard errors of the means (SEM) between 5 and 15% depending on the scale Ab Class -SSP/IgG Yield a b c d e f g IgG2a IgG2a IgG1 IgG1 IgG2a IgG2a IgG2a 72 43 41 38 36 32 22 91 100 70 100 99 100 100 a a With respect to antibody TABLE II A N T I B O D Y M O D I F I C A T I O N I N F L U E N C E OF THE IgG C O N C E N T R A T I O N RATIO Ab PDPA IgG ratio T101 45 2F7 -SSP/IgG Yield a (%) 44 81 88 24 45 31 99 94 90 60 50 89 29 57 92 79 AT 15E 45 72 11 2 28 34 59 28 BGG 40 3 6 9 10 20 18 95 92 92 50 3 6 9 15 27 22 96 98 92 60 3 6 9 21 36 28 94 98 94 a With respect to antibody IgG (mg/ml) 289 after 30 mln varied between 2 2 and 7 2 according to the antibody chosen and independently of the antibody class The modification rano of the annbodles with PDPA depended on the nature of the lmmunoglobuhn, the temperature, the amount of reagent and also on the IgG concentration (Table II) As the concentration of protein increased, the modification ratio rose to a maximum, before decreasing These results suggest a broad opumal protein concentration level of about 5 m g / m l At higher concentrations a relatwe increase xn the proportion of solvent (2-methyl-2-propanol or ethanol) added to the lmmunoglobuhn solunon was necessary If the IgG used was unstable, as an the case of AT15E, a high concentraUon led to a poor yield due to precipitation (28% at 11 2 m g / m l ) If the precipitated IgG were re&ssolved (for example m urea) the measured modification ratio was no higher than that obtained with the soluble modified IgG The denaturing action of the alcohol used to &ssolve the PDPA was therefore the mare cause of degradation Before couphng the modified antlbo&es to the rlcm A-chain, it ~s necessary to remove excess reagents, by e~ther gel filtration or dmlysls In order to prevent dilution, the latter is usually preferred, despite the length of the operation (B) Couphng of the modified tmmunoglobuhn to the rtcm A-chain A kinetic study of the couphng reaction of the modified T101 IgG to various quantities of A-chain at 25°C, showed that a slow reacUon took place, which was completed m about 20 h (Fig 3) The A-chain, the thlol of which had been preserved and was controlled by Ellman's reaction, was consumed in the couphng reaction with a yield of 80% However, about 90% of the A-chain molecules coupled after 24 h were already coupled after 6 h If this h n e n c study was performed at 0°C, only / ..... Ooc I£,'..- ......... 24 6 8 20 24 72 Hours Ftg 3 Kinetic study of coupling reaction 10 nag of modified IgG T101 (4 mg/ml) with about 5 SSP radicals per IgG molecule were reacted with ncln A-chain and the reaction monitored by measunng the released 2TP at 343 nm, at 20°C ( ) or at 0°C (. . . . . . ) Four A-cham/IgG StOlCtuometncratios were stu&ed [3, 1 0, o, 1 2, ×, 2 0, II, 3 0 290 60% of the A-chain molecules coupled after 72 h were coupled after 6 h On the basis of a 6 h reaction, about 30% of the A-chain molecules coupled at 25°C were coupled at 0°C at the same stolchiometrlc ratio W~th a large excess of A-chain, this problem proved to be of little importance compared to the risk of denaturation of the proteins, when reacted for more than 20 h at 25°C The lmmunotoxin thus produced can be purified by chromatography making use of the affimty of IgG for Protein A-Sepharose, or by gel filtration, either at low pressure with gels such as Sephadex (G-100 or G-200), Sephacryl ($200) or Sepharose (CL4B), or at high pressure (about 40 bars) with TSK columns (SW 3000 or SW 4000) (Fig 4) The advantage of the latter method is that the purified ~mmunotoxm is ready for use w~thm 30 mln, w~thout any loss of resolution However, it can only be applied to analytical size preparations (C) Analysts of the hvbrM molecules The capacity of rlcln A-chain to lnl'ubit the ribosomal protein synthesis in a cell-free system was observed at concentrations of 10"*-10 M (IC50) (Fig 5) If its accessible thlol group was modified, for example by N-ethylmalelmide, the activity of the A-chain was not slgmficantly affected This was also true if the A-chain was mono- or dlmethylated Coupled to an antibody by any kind of reagent, the A-chain potency was reduced to a fraction of the activity of the free subunlt On reducing the disulfide bond between the antibody and the A-chain by incubation with 0 2% 2-mercaptoethanol, the inhibitory activity of free A-chain was completely restored The same observation was made m the case of native rxcm if the reduction of the bond linking the 2 subunlts was carried out by incubation with 2% 2-mercaptoethanol However with rlcln, when the bond was not reduced, the A-chain moiety only displayed 0 5% of its full activity In the case of the lmmunotoxins, the amount of residual activity expressed by the A-chain, when it was not spht from the antibody, varied between 10 and 20% of its potential activity This residual activity increased according to the length of the arm hnkmg the 2 proteins (Fig 5) An analys~s of the amount of trltmm-labeled and enzymatlcally active A-chain in chromatography fractions of an immunotoxln obtained after coupling a monoclonal IgG and a 3H-dlmethylated A-chain indicated an equal amount of trltmm-labeled A-chain and of enzymatically active A-chain, confirming that the A-chain lost no potential activity when coupled to an antibody (Fig 6) In immunotoxins, the antibody moieties had binding properties on FACS analysis s~milar to those of the corresponding unmodified antlbo&es, with a few exceptions In such cases, for example with IgG 96 5 directed against the melanoma assocmted P97 antigen, the immunotoxm showed a loss of 25% binding activity (Casellas et al, 1982) Electrophores~s of the lmmunotoxin and its constituent proteins confirmed that the ~mmunotoxln consisted of a group of hybrid proteins, formed by IgG molecules coupled to A-chain molecules, m the corresponding ratios 1 0, 1 1, 1 2, 1 3 and so on Each band was exactly located as expected from the molecular weight of the corresponding hybrid (150,000, 180,000, 210,000, 240,000 ) Electrophoresls of the lmmunotoxln, after reduction with 2-ME, revealed 3 bands, 291 P1 a) 0 2 ~. O 0 01 P2 7'o 160 ml p1 b) Ol j P2 o Fig 4 Immunotoxm purification Two punhcaUon schemes for T101 lmmunotoran are illustrated a 4 mg of lmmunotoxan (0 9 mg/ml) were purified on a Sephadex G-100 colunm m 8 h P1 lmmunotoxm, P2 free A-chain, P3 low molecular weight products b 0 9 mg of lmmunotoxln (4 6 mg/ml) were purified by HPLC on a TSK SW 3000 column m 20 man P1 lmmunotoxln, resolved into lug,her and lower molecular weights, P2 free A-chain, fracUonatedinto A1 and A2 peaks, P3 and P4 low molecular weight products c o r r e s p o n d i n g respectively to the 3 pepUde subumts, which were h n k e d b y disulfide bridges, i e , hght a n d heavy I g G chains a n d rlcln A - c h a i n Electrophoresls of a l o n g - t e r m stored a n d n o n - r e d u c e d A - c h a i n sample revealed 2 b a n d s which corres p o n d e d to the A - c h m n m o n o m e r a n d & m e r , the m o n o m e r / d l m e r ratio (68 32%) was consistent with the free tluol tlter (0 70) of this sample (Table III) A m i x t u r e of free A - c h a i n a n d I g G m equal proportions, s u b m i t t e d to electrophoresls a n d d e n s l t o m e t n c analysis d e m o n s t r a t e d that the s t a l m n g response of the A - c h a i n was, o n a weight basis, equal to 75% of that of the l m r n u n o g l o b u h n This 292 ACtlwty (IC 50) Products 10 "~0 l q "g I0~"a [A-] (M) A-SH A - S1 ',r° I.N(~C 2H 5 I A-SS-CH 2 -CH2-CO- NH-2F7 A_ SS_(CH2)2_CO_ NH_(CH2)5_CO- NH- { 2F7 A-SS-B (RJcln) k~N\NNNNNNNNNNNNNNNNNNNNNNNN~NNN~NNN~N~NNN~ ~,~NNNNNNNNNNNN\~NNNNNNNNNN~NN~NNNNNNNNN~%.~NNN'~ ] ~,\\\\\\\\\\\\\\\~ activityexpressed by previously preparations act)v)ty expressed by preparat)ons w~thout reduction [----- reduced Fig 5 A-chain enzymatic activities in toxins and immunotoxms Preparations were tested m a cell-tree protein synthesis system from rat hver Concentrations of A-chain reducing 50% mhibmon were determined and compared with free A-chain as standard lnh~bmon was measured either w~th intact products or with previously reduced products diluted m 0 2% 2-ME perrmtted calculation of the relative response factors of the different hybrids and conversion of the areas under the scanning curve of an lmmunotoxm into the molar distribution of the different hybrids and thus allowed their comparison with the ..... 3H IPS OD 400 15 3OO olo ~3 0 200 i/ 601 70 ~ ,oo 90 100 110 120 130 Froctlons Fig 6 Sephadex G-100 chromatography of (3H-A) lmmunotoxm 12 mg of lmmunotoxJn resulting from the coupling reacUon between SSP modified IgG (2F7) and 3H-&rnethylated n o n A-cham were purified by gel filtration on a Sephadex G-100 column OpUcal density ( ), A-chain mhabmon of protein synthesis (IPS, - . . . . . ) and tntmm content ( ) were deterrmned in each fraction 293 150 KD 5 0 KD 25 KD A B C D E F Fig 7 Analysis of the molecular weights of lmmunotoxln Preparations were tested by electrophoresls on polyacrylamade gradient gel (PAA 2-16) Proteins were stained by Coomassle blue G250 A reduced IgG, B reduced l m m u n o t o x m , C lmmunotoxln, D equal weight rmxture of IgG and A-chum ( E + F) E IgG, F partially d~merlzed A-chain pattern calculated according to Po~sson's statistical d~stnbutlon law, using the formula Un Pn = __e-U n V T A B L E III D E N S I T O M E T R I C ANALYSIS OF A N I M M U N O T O X I N E L E C T R O P H O R E T I C P A T T E R N The optical density of each band m Fig 7 was measured and integrated in order to calculate the composition of the tested preparation The background due to impurities in the slab gel was approx 59~ MW % of total area in each path (mean of 2 determinations) Sample > 240000 240 000 210 000 180000 150000 60 000 50000 30 000 25 000 A B C 23 9 18 25 23 66 33 54 20 25 D E 54 14 98 26 F 32 68 294 TABLE IV OBSERVED AND CALCULATED PARAMETERS FOR TWO ELECTROPHORESIS EXPERIMENTS WITH IMMUNOTOXIN T101 lmmunotoxln TI01 IT ~ 1 TI01 IT ~ 2 Uo a 20 16 Uc b R ~" a d 1 96 1 11 0 987 0 950 1 08 1 01 re 0 990 0 999 N1 6 2 U0 observed average coupling ratio us average coupling ratio calculated from the slope of the first hnear regressmn log[(n ~) p.] = f(n ) R correlation coefficient of the first hnear regression d a average proportmnality factor between molar fractions observed and calculated using Polsson formula with u = uc e r correlation coefficient of the second linear regressmn (between observed and predicted molar fractions) f N number of records used for the analysis a b where p. is the fraction of molecular entitles c o n t a i n i n g n s u b s t i t u t i n g mo~enes m a m i x t u r e c o r r e s p o n d i n g to an average s u b s t i t u t i o n ratio u The results of the computerized analysis of the records of 6 electrophoresls experiments carried out w~th a T101 ~ m m u n o t o x l n having a n average coupling ratio close to 2 0 o n the one h a n d a n d of the records of 2 electrophoresls experiments carried out with a T101 l m m u n o t o x l n h a w n g an average coupling ratio close to 1 6 suggested that the observed d i s t r i b u t i o n can be predicted by Polsson's law (Table IV) The first correlation (see M a t e n a l s a n d Methods) c o n f i r m e d that the whole d i s t r i b u t i o n s were consistent with Polsson's f o r m u l a The second correlation d e m o n strated that the percentage of each c o m p o n e n t in the mixture could also be o b t a i n e d i n d e p e n d e n t l y according to tins f o r m u l a However. the & s c r e p a n c y between observed a n d calculated coupling ratios v a n e d between 2 a n d 30% a n d this p r o b a b l y reflected the &fference between biological a n d physical d e t e r m l n a U o n s of A - c h a i n content Discussion A biochemical study of the p r e p a r a t i o n of l m m u n o t o x i n s m a y be relevant in the synthesis of several h y b r i d proteins, designed to j o m together 2 proteins with different properties (Gxlhland et a l , 1978, Cawley et a l , 1981, Olsnes, 1981, T h o r p e a n d Ross, 1982) The widespread use of S P D P IS largely due to its commercial availability The c o m p a r i s o n winch has b e e n m a d e emphasizes the m a i n advantages of the P D P A a n d S P D P techniques, i e , the speed of the reaction even at 0 ° C a n d the high yield respectively The coupling reaction between the modified a n t i b o d y a n d the n c l n A - c h a i n needs 295 an excess of A-chain and a long incubation period at room temperature Tins reaction is slowed down at 0°C although it is possible to restore a good yield by increasing the excess of A-chain The following general protocols are suggested by our study They have enabled us to prepare more than 150 immunotoxln batches using more than 30 different antibodies Protocol P1 Protocol P'I 1 eq IgG + 6 eq SPDP 30°C 35 min 1 eq I g G + (45 eq PDPA + 30 eq EDC) room temperature 15 min I I dialysis dialysis $ IgG-SSP IgG-SSP Protocol P'2 Protocol P2 1 eq IgG-SSP + 2 5 A-SH 20 h at 25°C (overnight) 1 eq IgG-SSP + 7 eq A-SH 20 h at 0°C (overnight) I I T S K SW 3000 (HPLC) T S K SW 3000 (HPLC) or or Sephadex G-100 chromatography Sephadex G-100 chromatography Immunotoxln Immunotoxln The process (P1 plus P2) allows a coupling ratio between 1 and 2 ncin A-chain per I g G to be achieved We have attempted to increase the excess of reagents in order to couple a Ingher number of A-chains per I g G molecule but when the number of coupled A-chains per I g G molecule exceeded 4, both the yield and the stabdlty of the products were strongly reduced (data not shown) Unreacted SSP groups remain m the immunotoxlns, and could potentially react with tinol containing proteins in vivo, but that should be of less importance compared to glycosyl group-lectln interactions, for example, in the liver Pharmacokinetic studies addressing tins problem are under way Since the different molecular entitles corresponding to the formula IgG(A)~ (with x varying between 0 and n according to the coupling ratio) are distributed according to Polsson's law, the proportion of antibodies winch have not been coupled to the A-chain can be calculated The calculation shows that the average coupling ratio needs to be greater than 3 A-chains per I g G molecule In order to decrease the proportion of uncoupled I g G to less than 5% It was found that for 296 most preparations, where the couphng ratio varied between 1 and 2, 15-40% of uncoupled antibodies were present as predicted by the Polsson distribution This proportion can be reduced by the careful choice of chromatography fractions from a gel which has a fractionation capacity greater than Sephadex G-100, but this purification can be more efficiently achieved by a chromatography step based upon afflmty of the A-chain for an immobilized lectln (H Vldal, m preparation), which can remove all unconjugated antibodies According to the process described above involving PDPA, we have prepared lmmunotoxlns on scales ranging from 1 mg up to 18 g The largest batches have made it possible to study the stability of such hybrid proteins in sterile buffered solutions It has now been established that these drugs preserve both their chemical and biological properties for more than 2 years at 4°C Such storage conditions are much more favorable than freezmg which can induce degradation on thawing The analysis of enzyme activity of the coupled A-chain contained in these preparations shows that the A-chain thiol group is not directly involved m the activity mechanism, but that this group is probably close to the enzyme site since the coupling of macromolecules lnvolvmg this thlol group partially Inhibits the enzymatic activity Moreover~ residual activity is related to the length of the bridge hnklng the 2 proteins Separation of the A-chain from its carrier protein probably occurs m the target cell and the A-chain does not act until after it has been freed from its carrier, exther by reduction of the disulfide bond or by proteolysls If the release of the Q-chain is brought about by proteolysls a study of fragments of the proteolysed A-chain could confirm th~s hypothesis Acknowledgements The authors thank D Flageollet, R Bellamy and C Pages for their skllful technical assistance and P Casellas for reading the manuscript G Couzin is warmly acknowledged for her valuable help in preparing the manuscript References Anderson W L and D B Wetlaufer, 1975 Anal Blochem 67 493 Blair A H and I T Ghose, 1983, J Immunol Methods 59, 129 Blythman, H , P Casellas O Gros et al, 1981, Nature (London) 290, 145 Carlsson, J , H Drevm and R Axen, 1978, Blochem J 173, 723 Casellas P J P Brown, O Gros et al 1982, Int J Cancer 30, 437 Casellas, P, B Bourne P Gros and F K Jansen, 1984, J Blol Chem 259 9359 Cawley D B D L Simpson and H R Herschman 1981 Proc Natl Acid So U S A 78 3383 Colombattl, M , M Nabholz, O Gros and C Bron, 1983, J Immunol 131 3091 Gdhland, D G R J Colher, J M Moehrlng and T J Moehrmg, 1978, Proc Natl Acid So U S A 5319 Jansen F K H Blythman D Carnere et al, 1980, Immunol Lett 2 97 Jansen, F K H Blythman D Carnere et al, 1982, Immunol Rev 62, 185 Krollck, K A J W Uhr and E S Vltetta, 1983, Methods Enzymol 93, 333 75 297 Means, G E and R E Feeney 1968, Blochermstry 7, 2192 Neville D M and R S Youle 1982, Immunol Rev 62, 75 Noel, J P and L Plchat, 1983, J Label Compounds Radlopharm 20 1243 Olsnes S, 1981 Nature (London) 290, 84 Olsnes, S, K Refnes and A Plhl, 1974, Nature (London) 249, 627 Royston, 1, J A Majda, S M Baird, B L Meserve and J O Grlffiths, 1980, J Immunol 125, 725 Thorpe, P E and W C J Ross, 1982, Immunol Rev 62, 119 Thorpe, P E D C Edwards, A J S Davies and W C J Ross, 1982, in Monoclonal anubodles m chmcal medicine eds A McMlchael and J Fabre (Acadermc Press, London) p 167 Vidal, H 1981, Thesis, Umverslte des Scxences et Techmques du Languedoc Montpelher France Vltteta, E S, K A Krohck and J W Uhr, 1982, Immunol Rev 62, 159