Bioconjugate
ChemS@y JULY/AUGUST 1993 Volume 4, Number 4 0 Copyright 1993 by the American Chemical Society
ARTICLES Chemically Optimized Antimyosin Fab Conjugates with Chelating Polymers: Importance of the Nature of the Protein-Polymer Single Site Covalent Bond for Biodistribution and Infarction Localization Vladimir S. Trubetskoy,*v+Jagat Narula,' Ban An Khaw,' and Vladimir P. Torchilint Center for Imaging and Pharmaceutical Research, Massachusetts General Hospital, Bldg 149, 13th Street, Charlestown, Massachusetts 02129, Bouve College of Pharmacy and Health Sciences, Center for Drug Targeting and Analysis, Northeastern University, Boston, Massachusetts 02115. Received November 13, 1992
Murine antimyosin Fab fragment was conjugated with "'In-labeled N-terminal-modified DTPApolylysineusing three bifunctional reagents N-hydroxysuccinimide esters of 3-(2-pyridyldithio)propionic acid (SPDP conjugate), 4-(maleimidomethyl)cyclohexanecarboxylicacid (SMCC conjugate) and bromoacetic acid (BrAc conjugate) for potential localization of experimental myocardial infarction. Using various antibody preparations and a rabbit acute myocardial infarction model the following parameters were observed: (1)an in vitro antigen binding activity of SPDP conjugate = SMCC conjugate > BrAc conjugate, (2) a blood clearance rate of SPDP conjugate > BrAc conjugate > SMCC conjugate, (3) a liver and splenic accumulation of SPDP conjugate > BrAc conjugate > SMCC conjugate, and (4) the infarcted tissue activity showed an accumulation of SMCC conjugate > SPDP conjugate > BrAc conjugate This study exemplifies the importance of rational chemical design of antimyosin Fab-chelating polymer conjugate for improved target tissue localization in vivo.
INTRODUCTION The modification of monoclonal antibodies and their fragments with polylysine-based negatively charged chelating polymers has been recently shown to be effective in achieving improved immunoimaging of target tissues and reduced nontarget organ activities ( I ) . This approach has been demonstrated with monoclonal antimyosin Fab coupled with DTPA-PLL1 or deferroxamine-PLL ( 2 , 3 ) for radioimmunoscintigraphic localization of myocardial infarction. The use of chelating polymers relative to low molecular weight chelates (like DTPA cyclic anhydride or isothio-
* Correspondingauthor. TelephoneNo.:
(617)726-7834. FAX
NO.: (617) 726-7830.
+ Massachusetts General
t
Hospital. Northeastern University. 1Q43-18Q2/93/29Q4-Q25 1$04.00/0
cyanatobenzyl-DTPA) to modify proteins is preferred due to a dramatic increase in labeling efficiency and the possibility of linking antibody and the corresponding chelating polymer (carrying multiple radioactive metal atoms) via a single covalent bond which should prevent loss of immunoreactivity usually observed with modification with multiple low molecular weight chelators (4). Recently, a new coupling technique has been introduced which enabled a substantial increase in the isotope loading Abbreviations used DTPA, diethylenetriaminepentaacetic acid; PLL, poly-L-lysine;NPL, N-terminal modified polylysine; AMFab, antimyosin Fab fragment; SPDP, succinimidyl 3-(2pyridy1dithio)propionate; PDP, 3-(2-pyridyldithio)propionate; SMCC,succinimidyl4-(maleimidomethyl)cyclohexylcarboxylak BrAc-NHS, N-hydroxysuccinimide ester of bromoacetic acid; HBS, HEPES-buffered saline, pH 7.4. 0 1993 American Chemical Society
252
Trubetskoy et al.
Bloconjugate Chem., Vol. 4, No. 4, 1993
of the immunoconjugate with minimal loss in immunoreactivity. According to this procedure, a reactive group on the N-terminal of DTPA-polylysine is covalently conjugated to a monoclonal Fab fragment via a single covalent bond (5, 6). The polymer is linked to the antibody like a 'tail', thus avoiding multipoint binding and protein crosslinking. Since the nature of this single chemical bond between antibody and polymer may influence the biodistribution and target accumulation of the conjugate, we undertook this study to demonstrate the in vivo properties of AMFab modified with DTPA-NPL via different single point cross-linkers and to determine the optimal chemical design of the AMFab-DTPA-NPL conjugate for the best myocardial infarction localization. EXPERIMENTAL PROCEDURES
Monoclonal AMFab (R11DlO) was prepared as previously described (7). DTPA-NPL bearing a PDP group on its N-terminal was synthesized as described by Slinkin et al. (5). Briefly, e,N-carbobenzoxy-protectedPO~Y-D,Llysine (Sigma; Mw 10 kDa, 15mg), SPDP (Pharmacia, 2,5 mg), and triethylamine (0.5 pL) were mixed in 0.5 mL of dry dimethylformamide and incubated overnight at room temperature. The water-washed and lyophilized precipitate was deprotected with a 30 5% solution of HBr in glacial acetic acid (Aldrich, 2 mL). The deprotected polymer was washed with dry diethyl ether and lyophilized. Chelating residues were introduced into the polymer using DTPA bicyclic anhydride in 0.1 M bicarbonate, pH 8.0 (1.5 mg of anhydride/mg of NPL). Using trinitrobenzenesulfonic acid assay (8),DTPA residues were found to modify 95% of t-amino groups of the polymer. The unmodified lysine €-amino groups were blocked with succinic anhydride (15 mg/mg of dry polymer). Preparation of SPDP-, SMCC-, and BrAc-NHSModified AMFab. SPDP, SMCC (Sigma), and BrAcNHS (Sigma) were used as bifunctional modifiers to react with amino groups of AMFab. The reaction mixture typically consisted of 200 pg of AMFab and 10 pg of bifunctional reagent in 100 pL of HBS. The reaction was allowed to proceed for 1h at room temperature and then was terminated by dialysis against HBS at 4 OC. Conjugation of DTPA-NPL to AMFab. The homogeneity of AMFab and DTPA-NPL was confirmed by HPLC chromatography using Waters Protein-Pak 300SW column. Just prior to the reaction with modified AMFab, PDP-containing DTPA-NPL was activated by the treatment with 2.5 mM dithiotreithol in HBS for 20 min at 37 "C. Dithiotreithol was separated from the polymer using a Sephadex G-25 spin column equilibrated with HBS. Modified AMFab was mixed with reduced DTPA-NPL in a 1:l weight ratio and this reaction mixture was incubated overnight at 4 "C in HBS. The conjugate was separated from unmodified Fab and free polymer using a DEAE Sephadex A-25 ion-exchange process (2). The fraction eluted with 0.35 M NaC1,0.05 M phosphate, pH 6.0, was used for lllIn labeling. Immunoreactivity in Vitro. Immunoreactivity of resultant conjugates as well as AMFab modified with the cross-linkerswas determined by standard double-antibody ELISA using human heart myosin as the homologous antigen and peroxidase-labeled goat anti mouse IgG (Sigma) as the second antibody. All samples were run in duplicate and the highest average absorbance value on each plate was taken as 100% binding. Labeling and in Vivo Studies. 111In labeling was performed using the citrate transchelation method described in ref 2 except that a Sephadex G-100 column was
used instead of a G-25 column as in original protocol. Along with the conjugates, we used DTPA-PLL (Mw = 12 000) as a control polychelator for in vivo experiments. For biodistribution studies New Zealand white rabbits were used. The animals were anesthetized with ketaminel xylazine (40 and 5 mg/kg, respectively) and further maintained on mechanical ventilation. A left thoracotomy was performed and the mid left anterior descending coronary artery was occluded with suture. After 40 min the suture was removed. At 1h after reperfusion 100-300 pCi of l1lIn-labeled AMFab-DTPA-NPL conjugate or DTPA-PLL were injectedvia the ear vein. Blood samples were taken starting at 1min after conjugate administration up till 5 h to assess the clearance of the various preparations. After 5 h the animals were sacrificed by an overdose of pentobarbital. The hearts were excised and cut onto 0.5-cm slices. The slices were stained with 2% triphenyltetrazolium chloride solution for histochemical confirmation of myocardium necrosis. Then the heart slices were cut into 50-150-mg pieces (normal and infarcted heart tissue) and, along with samples of the major organs, were weighed and counted for lllIn radioactivity using a y-counter (Compugamma, LKB). The number of animals in each group ranged from three to five. RESULTS AND DISCUSSION
The aim of this study was to investigate the influence of the chemical bond between the AMFab and the chelating polymer on the biodistribution and infarct localization using ll'In-labeled DTPA-NPL conjugated to the antibody fragment via three different cross-linkers. The original procedure of conjugation using the disulfide bond was expanded to include SMCC- and BrAc-NHS-based conjugates. These latter conjugates utilized the stable C-S single thioether bond between Fab and the polymer (Figure 1). The choice of bifunctional reagents for the polymer coupling was determined by the following factors. Originally, the single-site Fab modification was developed for SPDP reagent only, so we needed to have a reagent introducing a stable C-S thioether bond for comparison. Among different maleimido-containing bifunctional reagents, SMCC has the most stable maleimido group (9). BrAc-NHS was chosen because it is not maleimidocontaining reagent and possesses a less bulky spacer than the cyclohexyl carboxylate moiety in SMCC (Figure 1). This reagent also has demonstrated elevated conjugation yields in earlier studies with conventional multipoint bound DTPA-PLL-AMFab conjugates (IO). Specific Activity of Conjugates. The specificactivity of DTPA-NPL itself was found to be approximately 70 pCi/pg of polymer by lllIn-labeling assessment. Average specific activities for conjugates (two or three preparations in each group) were found to be 145,144, and 166 mCi/lO mg of protein for SPDP-, SMCC-, and BrAc-based conjugates, respectively. Since the labeling procedure of any of the three conjugates involved usually resulted in average of 150pCi of lllIn/lOpg of protein, it was concluded that the Fab/polymer molar ratio in all conjugates is close to 1:l assuming that the molecular weights of the Fab and the polymer are 50 and 10 kDa, respectively. This finding was quite expectable since the DEAE Sephadex A-25 anion-exchange purification process used for isolation of the conjugates allows separation of the total conjugation mixture onto several fractions which differ in polymer/ protein ratio. Only one fraction of all three conjugates which has eluted at 0.35 M NaCl was used for labeling. Immunoreactivityin Vitro. After separation of the conjugates from unmodified AMFab and unreacted poly-
Biodistribution of Antibody Conjugates
Bloconlugete Chem., Voi. 4, No. 4, 1993 253
NH NH NH w N H C O - ( C H , ) , - S S -
reagent
+
80
a, b or c PDP-DTPA-NPL
1 t
-AT
BrAcconj.
+ wlvmer
~NOCO-(CH,l,-SS-~
p,
n n
pH 7.0
SPDP
"
DTPA-NPL-SH
SMCCconj.
+ SPDPconj.
-2
Fab
DTT,
-m-
'oak
bifunctional
0
QOCO-CH,-Br
e,
0
100
200
300
LNHS time, min
PDP-, maleimido-
or bromoacetyl- modified
Fab
NHCO-(CH,),-SS-NPL(DTPA)
Figure 1. Synthesis of AMFab-DTPA-NPL conjugates.
=I m
I
E
100
$?
80
Z
0
2 1 60
s 1
4'
1
A"
/ I
-2
1
0
log[AMFab
I
1
(1
i
--.--
native AMFab AMFab-SMCC AMFab-BrAc AMFab-SPDP BrAc-conj. --a-- SMCC-conj. SPDP-conj. -t
2
3
4
added], nglweil
Figure 2. Antigen binding activity assay (ELISA) for the native AMFab, bifunctional reagent-modified AMFabs, and different AMFab-DTPA-NPL conjugates.
mer using Sephadex A-25 ion-exchange chromatography, the immunoreactivity of the modified antibodies was tested in vitro using an ELISA (Figure 2). For comparison purposes the data on the immunoreactivity of bifunctional reagent-modified AMFabs are also included in the figure. As one can see, the introduction of the bifunctional reagent itself does not significantly alter the immunoreactivity of AMFab while having bulky polymer attached to the antibody fragment leads to more severe damage (Figure 2, dotted lines). BrAc-NHS-based conjugate demonstrated the most considerable loss of binding relative to both SPDP- and SMCC-based conjugates. Since the AMFab/ DTPA-NPL molar ratio was found to be the same in all three conjugates, the maximal loss of immunoreactivity in the BrAc-NHS-based conjugate might be explained by the fact that the BrAc modification site probablylies more closely to the antigen-binding region on the Fab protein globule which, in turn, leads to the impairment of antigen binding by the attached polymer.
Figure 3. Blood clearance of lllIn-DTPA-PLL and "'InAMFab-DTPA-NPL conjugates in rabbit. Mean values are plotted (three to five animals per group).
Blood Clearance. These immunoconjugates were then tested in the rabbit acute myocardial infarction model for target accumulation, clearance, and label biodistribution. Figure 3 shows the blood clearance patterns for "'Inlabeled DTPA-PLL (a polymer without any reactive group) and all three lllIn-labeled immunoconjugates. The disulfide-containing SPDP conjugate cleared from the circulation the most rapidly, while the SMCC-based conjugate remained in the circulation for the longest period. The behavior of the BrAc-based conjugate is more or less close to that for the SMCC-based conjugate. Quite expectable for a small polymer, lllIn-labeled DTPA-PLL itself rapidly cleared from the circulation. Thorpe et al. reported that the half-life of the disulfide bond containing immunotoxins in blood was approximately 6 h (11). Similarly, the immunoconjugate with a S-S bond used in our study had cleared maximally within the 5-h period, presumably due to the instability of the disulfides in vivo. Assuming that the radioactive polymer moiety of the conjugate is separated from the antibody, it should have a more rapid clearance from the blood, since it has a smaller molecular weight. Besides disulfidelinked immunotoxins, similar studies have also been performed where low molecular weight chelators have been conjugated to antibodies via different types of the covalent bond. Perala et al. has demonstrated the difference in blood clearance as well as in kidney and tumor activities of anti-prostate acid phosphatase IgG conjugated with DTPA directly, and via oxidized carbohydrate chains (12, 13). Deshpande et al. found that the clearance rate of conjugates with a reducible S-S bond of benzyl-EDTA and murine IgG was faster than conventional benzylEDTA-linked antibody. Interestingly, this group found the accumulation of the S-S bond containing conjugate in the liver to be lower than stable thioether bond I g G benzyl-EDTA conjugate (14). Biodistribution and Infarct Accumulation. Conjugates used in our investigation exhibited not only dissimilarities in blood clearance rates but also substantial differences in the biodistribution patterns (Figure 4). Disulfide-containingSPDP-linked conjugate a t 5-h postinjection time demonstrated high liver and splenic activities (up to 1% of injected dose/g of tissue in spleen). Stable thioether bond containing conjugates showed low radioactivity accumulation in the liver and spleen. Localization for S-S containing conjugate was remarkably lower in the infarct than thioether SMCC-based conjugate, perhaps due to the splitting of the polymer-Fab S-S bond in the blood or just due to increased blood clearance. Therefore, there is less intact conjugate to interact with the antigen.
254
Trubetskoy et al.
Bioconjugate Chem., Vol. 4, No. 4, 1993
Blood
Kidney
Spleen
Liver
Norm. myocard. SMCCconj. SPDPconj.
Infarct 0
1
2
YO injected dose/g of tissue Figure 4. Biodistribution of “‘In-DTPA-PLL and “‘InAMFab-DTPA-NPL conjugates in rabbits (5-h postinjection time, n = 3-5 animals per group).
BrAc-NHS-based conjugate demonstrated intermediate infarct accumulation. This may reflect the consequence of lower antigen binding activity found in the in vitro ELISA tests. For control polymer, DTPA-PLL, rapid clearance was followed by kidney accumulation without discrimination between infarcted and normal heart tissues (Figure 4). Unlike lllIn-DTPA-PLL, the DTPA-NPL accumulates mainly in liver and spleen as it can be concluded from the SPDP-based AMFab-DTPA-NPL biodistribution data. This phenomenon presumably takes place due to the following two-step process: (1)reduction of the SPDP-based conjugate in the blood with release of free sulfhydryl-containing lllIn-DTPA-NPL and (2) “secondary”binding of the polymer to the certain plasma proteins via sulfhydryl/disulfide exchange in a manner described in ref 15 for ar2-macroglobulin. These “secondary”-labeledproducts leave circulation relatively fast with subsequent accumulation of radioactivity in the liver and spleen. It is interesting to observe that differences in radioactivity accumulation for all three conjugates leveled off in normal myocardium and are statistically insignificant for kidney. Only “actively” absorbing organs (like target infarct area and RES organs) exhibited differences in the uptake of the conjugates. Target/nontarget accumulation ratio (which may be as high as 25 for the SMCC-based conjugate) is determined exclusively by high infarct tissue uptake. The SMCC-based conjugate showed the highest target area accumulation and the lowest liver and splenic activities. There is no polymer-Fab bond reduction and no drastic loss of immunoreactivity in this case. Unfortunately, the target/nontarget ratio of the SMCC-based conjugate was to some extent compromised due to the relatively high blood radiolabel content. Nevertheless, the infarcted tissue had maximal accumulation of this conjugate. CONCLUSION
Our data clearly demonstrated that the specific chemical design of single site linked AMFab-DTPA-NPL conjugates might be useful in the developmentof new myocardial infarct localization probes for radioimmunoscintigraphy. Together with the data of other researchers ( I I , I 2 ) these results indicate that the nature of the chemical bond of the linkage between the antibody fragment and the
chelating polymer is very important for the optimal performance of the immunoconjugates in vivo. Since the ideal properties of the lllIn-labeled immunoconjugate suitable for radioimmunoscintigraphy should be (1)long enough circulation time to provide sufficient contact with exposed antigen, (2) rapid removal from the circulation of nonbound conjugate after accumulation in the target area in order to increase target to blood activity ratios, and (3) insignificant accumulationof the conjugatein organs other than the target, we conclude that by varying the nature of the single covalent bond between antibody and chelating polymer one can influence and optimize the in vivo properties of the conjugate (such as circulation time and biodistribution) in order to make them closer to the properties of the ideal conjugate. Future studies should definitely focus on the Fab-polymer bonds of intermediate lability which allow rapid accumulation of the conjugate in the target area during the first phase after injection and rapid clearance of nonbound label during the second phase. ACKNOWLEDGMENT
The work was partially supported by funding from Sterling Winthrop Pharmaceutical Research Division, Inc. and NCI Grant #CA 50505. LITERATURE CITED (1) Torchilin, V. P., and Klibanov, A. L. (1991) The antibodylinked chelating polymers for nuclear therapy and diagnosis. CRC Crit. Rev. Ther. Drug Carrier Systems 7, 275-308. (2) Khaw, B. A., Klibanov, A. L., O’Donnel, S. M., Saito, T., Nossiff, N., Slinkin, M. A., Newell, J. B., Strauss, H. W., and Torchilin,V. P. (1991) Gamma imagingwith negatively chargemodified monoclonal antibody: modification with synthetic polymers. J. Nucl. Med. 32, 1742-1751. (3) Slinkin, M. A., Klibanov, A. L., Khaw, B. A., and Torchilin, V. P. (1990)Succinylated polylysineas a possible link between the antibody molecule and deferoxamine. Bioconjugate Chem. 1,291-295. (4) Yin, O., Narula, J., Nossif, N., and Khaw, B. A. (1991) Correlation of immunoreactivity and polymer formation to DTPA modificationof monoclonal antibody. Nucl. Med. Biol. 18,859-864. (5) Slinkin, M. A., Klibanov, A. L., and Torchilin, V. P. (1991) Terminal modified polylysine-based chelating polymers: highly efficient coupling to antibody with minimal loss in immunoreactivity. Bioconjugate Chem. 2, 342-348. (6) Trubetskoy, V. S.,Torchilin, V. P., Kennel, S. J., and Huang, L. (1992)Use of N-terminal modified poly(L-lysine)-antibody conjugate as a carrier for targeted gene delivery in mouse lung endothelial cells. Bioconjugate Chem. 3, 323-327. (7) Khaw, B. A., Mattis, J. A., Melincoff, G., Strauss, H. W., Gold, H. K., and Haber, E. (1984) Monoclonal antibody to cardiac myosin: Imaging of experimental infarction. Hybridoma 3, 11-23. (8) Habeeb, A. F. S. A. (1966) Determination of free aminogroups in proteins by trinitrobenzenesulfonic acid. Anal. Biochem. 14, 328-336. (9) Yoshitake, S., Imagawa, M., Ishikawa, E., Niitau, Y., Urushizaki, I., Nishiura, M., Kanazawa, R., Kurosaki, H., Tachibana, S., Nakazawa, N., and Ogawa, H. (1982) Mild and efficient conjugation of rabbit Fab’ and horseradish peroxidase using a maleimide compound and its use for enzyme immunoassay. J. Biochem. 92, 1413-1424. (10) Narula, J., Trubetskoy, V. S., O’Donnell,S. M., Strauss, H. W., Torchilin, V. P., and Khaw, B. A. (1992)Negativelycharge polymer-modified antimyosin for infarct imaging: effects of different cross-linkers. J. Nucl. Med. 33 (No 567), 959. (11) Thorpe, P. E., Wallace, P. M., Knowles, P. P., Relf, M. G., Brown, A. N. F., Watson, G. J., Knyba, R. E., Wawrzynczak,
Biodlstributlon of Antibody Conjugates
E. J., and Blakey, D. C. (1987) New coupling agents for the synthesis of immunotoxins containing a hindered disulfide bond with improved stability in vivo. Cancer Res. 47,59245931. (12) Perala, M., Vihko, P., Sodervall, M., Hekkila, J.,and Vihko, R. (1990)Biodistribution in normal mice lllIn-labeled prostatic phosphatase-specific antibody and its F(ab)g fragments derivatized site-specifically or via bicyclic diethylenetriaminepentaacetic acid anhydride. Eur. J. Nucl. Med. 16,621-626. (13) Perala-Heape, M., Vihko, P., Pelkonen, I., Laine, A., and Vihko, R. (1991)Effect of conjugation on the biodistributionof
Bioconjugate Chem., Vol. 4, No. 4, 1993 255
lllIn-labeled anti-PAP and anti-PSA monoclonal antibodies examined in nude mice with PC-82 human tumor xenografts. In uiuo 5, 159-166. (14) Deshpande, S. V., DeNardo, S. J., Meares, C. F., McCall, M. J., Adams, G. P., and DeNardo, G. L. (1989) Effect of different linkages between chelates and monoclonal antibodies on levels of radioactivity in liver. Nucl. Med. Biol. 16, 587597. (15) Ghetie, M.-A.,Uhr, J. W., and Vitetta, E. S. (1991)Covalent binding of human a2-macroglobulin to deglycosylated ricin A chain and its immunotoxins. Cancer Res. 51, 1482-1487.
