Bioconjugate Chem. 1093, 4, 121-126
121
Chemoimmunoconjugate Development for Ovarian Carcinoma Therapy: Preclinical Studies with Vinca Alkaloid-Monoclonal Antibody Constructs L. D. Apelgren,' D. L. Bailey, S. L. Briggs, R. L. Barton, D. Guttman-Carlisle, G. A. Koppel, C. L. Nichols, W. L. Scott, T. D. Lindstrom, A. L. Baker, and T. F. Bumol Lilly Research Laboratories, Indianapolis, Indiana 46285. Received September 4, 1992
Preclinical efficacy studies are presented in a human ovarian carcinoma model utilizing several novel conjugation strategies with the KS1/4 monoclonal antibody and derivatives of the vinca alkaloid acid desacetylvinblastine hydrazide. The chemoimmunoconjugatesKS1/4-/3-alanine-methylenemalonic ethyl ester-4-decacetylvinblastine 23-hydrazide (KS1/4-BAMME-DAVLB-HY), KS1/4-/3-alanine-5formylpyrrole-2-carboxylicacid-4-desacetylvinblastine 23-hydrazide (KS1/4-BAP-DAVLB-HY), and KS1/4-4-desacetylvinblastine 23-hydrazide were explored in the OVCAR-3 human ovarian carcinoma xenograft model. These conjugates, constructed with variable linker stability between the vinca alkaloid and the antibody, were studied by comparing the route of administration and the treatment schedule. Under these conditions a mean survival time from 28 to 35 days in untreated control animals was observed. Significant increases in survival (i.e. 3-9-fold over untreated control animals) were observed with all the immunoconjugates tested but with varying potency and efficacy dependent on linker strategy. Parallel therapy with equivalent doses of free DAVLB-HY or a non-antigen-binding immunoconjugate did not significantly increase the survival of the animals. These results suggest several chemoimmunoconjugate strategies for site-directed therapy of human ovarian cancer.
INTRODUCTION Human ovarian cancer is the fourth most frequent cause of cancer death in women and is the leading cause of death among patients in the United States with gynecologic malignancies (1,2). Surgery followed by chemotherapy and/or radiation therapy are the current treatments for this disease, which result in about 50 % of patients having recurrent disease with survival falling sharply with each advancing of the disease. Ovarian cancer is confined to the peritoneal cavity throughout most of its course, although extra-abdominal metastases can occur. This fact has led to various attempts of intracavitary therapy for ovarian cancer with chemotherapeutic agents (3) or radioactive colloids ( 4 ) which generally have been ineffective. Monoclonal antibodies have been used in the monitoring (5,6)and therapy (7)of gynecological carcinomas. Intraperitoneal therapy with monoclonal antibody-drug or monoclonal antibody-toxin immunoconjugates has been suggested by a number of investigators (8,9). Previously, our laboratory demonstrated that the monoclonal antibody-vinca alkaloid immunoconjugate LY203725 (KS1/ 4-4-desacetylvinblastine 23-hydrazide) significantly prolonged the survival of tumor-bearing animals in the OVCAR-3human ovarian carcinoma xenograft model (IO). In this immunoconjugate, the vinca alkaloid moiety was linked to oxidized carbohydrate residues on the monoclonal antibody through a hydrazone linkage. These linkages possess the ability to release active drug by virtue of their inherent hydrolytic instability under physiological conditions (11, 12). Extending previous studies in the OVCAR-3 xenograft model, this report describes two novel conjugation strategies in the preparation of monoclonal antibody-vinca alkaloid immunoconjugates which vary the chemical linker to potentially optimize delivery of the potent vinca alkaloid
* Author to whom correspondence should be addressed. 1043-1802/93/2904-0121$04.0010
moiety. In this study, we have chemically introduced aldehyde or methylenemalonate groups onto the lysine residues of the monoclonal antibody to provide the use of hydrazone- or enamide-based chemistry for conjugation. Thismodification theoretically achieves more control over the release properties of these linkages to antibodies through enhanced chemical stability. These studies with the OVCAR-3 human ovarian carcinoma model compare intravenous therapy and intraperitoneal therapy for these varied immunoconjugates and report efficacy results as survival advantages for each chemical linker between a common antibody and vinca alkaloid species. The data demonstrate the importance of linker chemistry strategy in structure-activity relationships in the preclinical development of chemoimmunoconjugates for ovarian cancer. EXPERIMENTAL PROCEDURES Cell Lines. The human ovarian cancer cell line OVCAR-3 (8,131was maintained in vivo by serial passage of ascites fluid in female athymic nude mice (outbred CD1; Charles Rivers). The animals were housed in isolator cages (Lab Products, Maywood, NJ) and all handling of the animals was done under laminar-flow hoods in an AALAC approved facility. The KS1/4 and 9.2.27 hybridoma cell lines were originally obtained from Dr. Ralph A. Reisfeld, Scripps Clinic and Research Foundation, and were maintained as previously described (IO). Preparation of Monoclonal Antibodies. The monoclonal antibodies KS1/4, which defines an adenocarcinoma-associated antigen highly expressed in ovarian carcinomas (14), and 9.2.27, which recognizesa chondroitan sulfate proteoglycan with high expression in human melanomas (151,were purified as described previously (IO). Preparation of KS 1/4-DAVLB-HY Immunoconjugate. KS1/4-4-desacetylvinblastine 23-hydrazide (KS1/ 4-DAVLB-HY, Lilly Research Labs serial number LY203725) is a covalent conjugate of the monoclonal antibody KS1/4 and 4-desacetylvinblastine 23-hydrazide, 0 1993 American Chemical Society
Apelgren et
122 Bioconjugate Chem., Vol. 4, No. 2, 1993
ai.
A KSl N-DAVLEHydruldo
OH
'*\
N-N=C--MoAb (via carbohydrate on Fc region)
B OH
KS114- BAP-D AVLB-Hydrarlde
C
N-LySine-MOAb
0
0
0
Figure 1. Structures of the monoclonal antibody-vinca alkaloid chemoimmunoconjugates, KS1/4-DAVLB-HY (A), KS1/4-BAPDAVLB-HY (B), and KSU4-BAMME-DAVLB-HY(C). a derivative of the vinca alkaloid vinblastine (11). This conjugate was prepared as previously described by the coupling of the vinca alkaloid to the monoclonal antibody through a hydrazone linkage to oxidized carbohydrate residues located in the Fc portion of the antibody (12) (Figure 1A). These linkages possess the ability to release active drug by virtue of their hydrolytic instability. In general, immunoconjugates with an average molar conjugation ratio (CR) of 4.5 mol of vinca alkaloid/mol of antibody were utilized. Preparation of KS 1/4-BAP-DAVLB-HY Immunoconjugate. KS1/4-j3-alanine-5-formylpyrrole-2-carboxylic acid-DAVLB-hydrazide (KSl/kBAP-DAVLB-HY) was prepared as follows: to a 33-mL solution of KS1/4 (12.3 mg/mL, 406 mg, 2.7 pmol) in 0.34 M borate buffer, pH 8.6, was added 5 mg (16 pmol) of the N-hydroxysuccinimide active ester of j3-alanine-5-formylpyrole-2carboxylic acid dissolved in 1 mL of acetonitrile. After stirring gently for 1 h at room temperature, the reaction mixture was chromatographedon a Sephadex G-25 column equilibrated wtih 0.1 M sodium acetate, pH 5.6. The derivatized antibody fraction was collected. To this was added 260 mg (0.3 mmol) of 4-desacetylvinblastine 23hydrazide sulfate and the reaction mixture stirred at room temperature overnight. The product was isolated by
chromatography over a Sephadex G-25 column equilibrated and eluted with PBS, pH 7.4. The protein fractions were pooled and concentrated in an Amicon stirred cell. Drug and protein content were determined by spectrophotometric analysis resulting, on average, in immunoconjugates with a CR of approximately 3.0. The reversible coupling of the vinca alkaloid to the monoclonal antibody occurs through aldehyde groups introduced onto lysine residues in the monoclonal antibody (Figure 1B). Preparation of KS 1/4-BAMME-DAVLB-HY and 9.2.27-BAMME-DAVLB-HY Immunoconjugates. KS1/4-i3-alanine-2-methylenemalonicacid ethyl ester4-desacetylvinblastine 23-hydrazide (KSUCBAMMEDAVLB-HY) was prepared as follows: the KS1/4 monoclonal antibody at a concentration of 15-20 mg/mL in 0.34 M sodium borate buffer, pH 8.6, was stirred with the addition over several minutes of NHS-BAMME-DAVLBHY in dry Nfl-dimethylformamide (DMF). The DMF solution is prepared at a concentration to deliver NHSBAMME-DAVLB-HY a t a molar input ratio (active ester: antibody) of 81-12:l and give a final conjugation reaction mixture of 7.5% DMF cosolvent. The reaction mixture is stirred slowly at 25 "C for 20 min after which it is centrifuged for 10 min to separate suspended precipitate. The supernate is then injected onto an appropriately sized
Bioconjugete Chem., Vol. 4, No. 2, 1993 12s
Chemoimmunoconjugate Therapy of Ovarian Cancer Table I. Properties of the Chemoimmunoconjugates Utilized abbrev KSlI4-DAVLB-HY KSl/CBAP-DAVLB-HY KSlI4-BAMME-DAVLB-HY DAVLB-HY (free drug)
ICW
ave
17 63 128 5
4.5 3 4.5
ave (ng/mL)a CRb % IR' 98 92 96
DAVLB-HY
releaserate (pg/h)d
0.122 0.082 0.004
a The ICWwas measured in a tissue culture microtiter proliferation assay as describedin ExperimentalProcedures. * Average CR is the molar conjugation ratiowhich is the moles of vinca alkaloid per mole of antibody. Average IR is the average immunoreactivity of the immunoconjugate measured by flow cytometry compared to unconjugated KS1/4 antibody. Average release of DAVLB-HY from 0.5 mg/mL conjugate in rat plasma as determined by HPLC analysis.
3 M W G (VINCA CONTENT)
20
0
100
200
DAYS
hydrophobic interaction phenyl-Sepharose column (PharmaciaFPLC) which is equilibrated with 155% acetonitrile, 85% 0.1 M phosphate, pH 6.10. The column is eluted with 10 column volumes of the same buffer followed by a step to 5 column volumes of 40% acetonitrile/60% phosphate. Fractions from the 15% CH&N eluent containing antibody, conjugate, and free BAMME-DAVLB-HY were pooled and injected onto a Sephadex G-25M desalting column (Pharmacia FPLC). Fractions containing immunoconjugate were pooled, concentrated, and filtered (Millex GV 0.22-pm filters, sterile). The coupling of the vinca alkaloid to the monoclonal antibody occurs in this strategy by attaching the hydrazide function of the DAVLB-HY to the linker via an enamide-like bond and attaching the linker to the monoclonal antibody by amide formation with the t-amino groups of the lysine residues (Figure 1C). 9.2.27-BAMME-DAVLB-HYwas prepared as described above for the KS1/4-BAMME-DAVLB-HY. In general, immunoconjugates with a CR of approximately 4.5 were utilized. Immunoreactivity of the Immunoconjugates. The antigen-binding capacity of the above-mentioned immunoconjugates was measured by flow cytometry with OVCAR-3 target cells as reported previously (10). The control immunoconjugate 9.2.27-BAMME-DAVLB-HY did not react with the OVCAR-3 target cells in this assay but retained good immunoreactivity toward antigenpositive human melanoma cells in separate experiments. In Vitro Cytotoxicity Assay. A tissue culture microtiter proliferation assay (16) was used to assess the in vitro potency of the immunoconjugates. For each immunoconjugate and the free vinca alkaloid, the percent growth inhibition values were calculated by comparison with the media control group (% inhibition = [(control treated)/controll X 100). A series of percent growth inhibitions was plotted against the corresponding drug concentrations, and the concentration inhibiting 50 % of the cell growth ( I C d was estimated. Stability Studies. Samples of the above conjugates (0.5 mg/mL) in rat plasma were incubated at 37 "C and aliquots were taken at various times. Samples were assayed for free vinca alkaloid by HPLC with electrochemical detection as an approximation of in vivo stability. Survival Experiments. In our previous study, we documented the optimal parameters for the use of vinca alkaloid chemoimmunoconjugates under stringent conditions for efficacy in the OVCAR-3xenograft model where free DAVLB-HY and an irrelevant monoclonal antibody conjugate were inactive (10). Dosing schedules using three of our four treatments with the immunoconjugates listed below, free drug, or a non-antigen-binding immunoconjugate were employed and are indicated in Results. Under these conditions, untreated control animals have a mean
0
100
200
DAYS
KSl/CBAMME-DAVLB-HYand 9.2.27-BAMME-DAVLB-HY against OVCAR-3ovarian carcinoma with a three dose treatment protocol. After OVCAR-3 Figure 2. Comparison of
tumor implantation on day 0, the animals (n = 7)were treated ip with a 3.0 [A]or 1.6 mg/kg [Bl (vinca content) dose on days 11, 14, and 17. This experiment is summarized in Table 11, experiment 1.
survival of 28-35 days after inoculation. Treatment was by intraperitoneal (ip) or intravenous (iv) injection of the immunoconjugate in 0.2 mL of PBS as indicated in the text. The date each experiment was terminated is indicated in Results. Statistical evaluation was performed on all in vivo experiments using the Kaplan-Meier survival curve analysis (17,18). RESULTS
I n Vitro Properties of Various Chemoimmunoconjugates. Various in vitro parameters of the three chemoimmunoconjugates utilized in this study are presented in Table I. The immunoreactivity of the conjugates to the OVCAR-3 target cell are comparable and display 292 % immunoreactivity as compared to unmodified KS1/ 4. In contrast, the varied conjugation strategies displayed variable in vitro cytotoxicity profiles toward human adenocarcinoma target cells. KS1/4-DAVLB-HY (ICw, 17 ng/mL) is approximately 3-fold less potent than the free drug while KSl/CBAP-DAVLB-HY (ICs0 = 63 ng/ mL) is approximately 12-fold less potent. KS1/4-BAMME-DAVLB-HY (ICs0 = 128ng/mL) is 25-fold less potent than the free drug and 2-6 times less potent than the other immunoconjugates tested. The rank order of in vitro potency in this assay is DAVLB-HY > KSl/CDAVLBHY > KS1/4-BAP-DAVLB-HY> KSU4-BAMME-DAVLB-HY. This ranking also correlates with preliminary immunoconjugate stability studies in rat plasma which indicate that KS1/4-DAVLB-HYis the least stable whereas
124
Apelgren et 81.
Bioconjugate Chem., Vol. 4, No. 2, 1993
Table 11. Inhibition of OVCAR-3Xenografts by Immunoconjugates dose exp. treatmentd route (mg/kg) 1 control iP 3.0 KS1/4-BAMME-DAVLB-HY !P 1.5 KSU4-BAMME-DAVLB-HY !P 3.0 9.2.27-BAMME-DAVLB-HY 1.p 1.5 9.2.27-BAMME-DAVLB-HY !P 2 control !P 1.5 KSlI4-BAP-DAVLB-HY !P 0.75 KSU4-BAP-DAVLB-HY !P 0.375 KS1/4-BAP-DAVLB-HY !P 1.5 DAVLB-HY !P 0.75 DAVLB-HY !P 3 control !P 0.75 KS1/4-DAVLB-HY KS1/4-DAVLB-HY 0.75 KS1/4-BAP-DAVLB-HY iP 0.75 KS1/4-BAP-DAVLB-HY 1v 0.75 KSl/CBAMME-DAVLB-HY iP 0.75 KSlICBAMME-DAVLB-HY 1v 0.75 DAVLB-HY iv 0.75
treatment schedule (day)a 11, 14, 17
mean survival (days)* 27 f 8 157 f 85* 143 52* 101 21* 60 24 28 f 7 228 129* 340 65* 219 127* 16 2 84 i 53 35 f 4 266 72* 172 81* 226 107* 254 58* 131 25* 72 f 46 39 f 4
day of terminationc 220
* * * * * * *
8, 10,13, 15
8, 11,15, 18
404
300
** * * *
;:
a Days after implant of OVCAR-3 cells. An asterisk indicates a p < 0.01 aa compared to control in a x2 paired analysis. On the day of termination, all surviving animals were sacrificed and survival was statistically evaluated. Each animal was given an ip injection of 6 X lo7 Ovar-3cells on day 0. The animals (n = 7) were treated on the days indicated in the treatment schedule. The dose received waa in mg/kg based on the vinca content of the immunoconjugate. Inhibition of xenograft growth was determined by increased survival time of the treated animals compared to the untreated control k & a l s .
KSl/CBAMME-DAVLB-HY is the most stable conjugate exhibiting minimal release of vinca alkaloid (Table I). The corresponding structures of the three chemoimmunoconjugates utilized in this study are shown in Figure 1. In Vivo Comparison of KSl/4-BAMME-DAVLBHY and 9.2.27-BAMME-DAVLB-HY. To evaluate the conjugate-mediated effect of KSl/CBAMME-DAVLBHY in the OVCAR-3ovarian carcinoma nude mouse model, a non-antigen-binding, subisotype-equivalent immunoconjugate, 9.2.27-BAMME-DAVLB-H,was tested. Since KSU4-BAMME-DAVLB-HY was the least potent chemoimmunoconjugate evaluated (Table I), dose selection was adjusted upward from previous studies (10)to 3.0 mg/kg vinca content for this immunoconjugate (based on milligrams of vinca alkaloid per kilogram of animal body weight). When the immunoconjugates 9.2.27-BAMME-DAVLB-HY and KSU4-BAMME-DAVLB-HYwere administered ip with a three dose treatment schedule, a significant increase in survival was obtained with KS1/ 4-bAMME-DAVLB-HY a t 1.5 and 3.0 mg/kg (Figure 2). A significant increase in survival was not observed at 1.5 mg/kg with the irrelevant conjugate 9.2.27-BAMMEDAVLB-HY,but improvement was seen at 3.0 mg/kg with this therapy schedule. These data indicate that a 2-3fold increase in survival time is obtained with KS1/4BAMME-DAVLB-HY therapy over that with the irrelevant conjugate 9.2.27-BAMME-DAVLB-HY at the 1.5 mg/kg dose level compared to untreated control animals. A statistical summary of this experiment is shown in Table 11. In Vivo Efficacy of KS1/4-BAP-DAVLB-HY. The in vivo efficacy of KS1/4-BAP-DAVLB-HYwas evaluated in the OVCAR-3 human ovarian carcinoma xenograft model. The in vitro cytotoxicity profile of KS1/4-BAPDALVB-HY was similar to that of KS1/4-DAVLB-HY (Table I), suggesting a lower dose strategy. The mice were given ip injections of the immunoconjugates and free DAVLB-HY at doses of 1.5,0.75, and 0.375 mg/kg on days 8,10,13, and 15 after inoculation with the OVCAR-3 cells (Figure 3). Significant survival was observed with this conjugate at the dose levels used compared to untreated control animals. The free drug, DAVLB-HY, was not effective in increasing the survival of the animals at the
-.
- -
CONTROL -0- KSIICBAP-HY 1.Smgkg -0KSl/+BAP.HY 0.15 m g k g KSl/CBAP-HY 0.375 mgmg --ADAVLB-HY 1.5mgRg
'Tm 7 80
20]1\
I
.
,
,
,
u ,
,
~
0 0
100
200
300
400
Days Figure 3. Evaluation of ip administration of KSlIQ-BAPDAVLB-HY and DAVLB-HY against OVCAR-3 ovarian carcinoma. After OVCAR-3tumor implantation on day0, the animals (n = 7) were treated ip on days 8,10,13, and 15 with a 1.5,0.75, or 0.375 mg/kg (vinca content) dose. This experiment is summarized in Table 11, experiment 2.
1.5 mg/kg dose level, indicating a therapeutic advantage with this immunoconjugate. A statistical summary of the results is shown in Table 11. The 0.75 mg/kg (vinca content) dose demonstrated optimal survival efficacy in this test, which suggests that this dose may reflect an optimal therapeutic index for this conjugation strategy. The results indicate that a 4-9-fold increase in survival time with KS1/4-BAP-DAVLB-HYwasobtained over that with free DAVLB-HY, compared to control animals. In Vivo Comparison of Intraperitoneal and Intravenous Therapy with KS1/4-DAVLB-HY, KS1/4BAP-DAVLB-HY, and KS 1/4-BAMME-DAVLB-HY. In order to compare the various linker strategies and the effect of the route of administration in the OVCAR-3 ovarian carcinoma nude mouse model, KSU4-DAVLBHY, KS1/4-BAP-DAVLB-HY, and KS1/4-BAMMEDAVLB-HY were administered at 0.75 mg/kg (vinca alkaloid content), either ip or iv, on days 8,11,15, and 18 after inoculation with OVCAR-3 cells (Figure 4). Both routes of administration produced significant survival of the test groups over the untreated control animals with the KSl/CDAVLB-HY and KS1/4-BAP-DAVLB-HY
Bioconju@te Chem., Vol. 4, No. 2, 1993 125
Chemoimmunoconjugate Therapy of Ovarian Cancer
-+-
A
CON-
-0- KSlICDAVLCHYIC --t K8llCDAVLCtW#V -t- D A V L B W N
DISCUSSION
-
?
a >
‘E
00.
at
44‘
3
KS1/4-BAMME conjugate administered iv, which did not demonstrate a survival advantage. A statistical summary of the results is shown in Table 11.
L
I
U
T
I .
100
0
200
300
IT CONTROL
KSll+BAP*HYfiP KSll&BAP*HYN DAVLMYrN
0 100
0
O f 0
I
100
-
300
200
300
A
A
L I .
200
.
-
1
,
DAYS Figure 4. Comparison of iv and ip administration of KS1/4DAVLB-HY (A), KSl/CBAP-DAVLB-HY (B) and KS1/4BAMME-DAVLB-HY (C) against OVCAR-3ovarian carcinoma. After OVCAR-3 tumor implantation on day 0, the animals (n = 7) were treated eith ip or iv, on days 11,15, and 18 at 0.75 mg/kg (vinca content). This experiment is summarized in Table 11, experiment 3.
conjugates, but only ip therapy was successful with the KSl/CBAMME-DAVLB-HY conjugate at this dose level. These results in total demonstrate a 3-5-fold increase in survival with either route of therapy, compared to untreated control animals, with the sole exception of the
This report examines the preclinical antitumor activity of several chemoimmunoconjugatesconstructed by varying the chemical linker between the monoclonal antibody KS1/4 to the vinca alkaloid DAVLB-HY. A critical factor in the antitumor activity of a conjugate will be its ability to release drug in an active form after tumor localization has been accomplished. The activity of KS1/4-DAVLBHY reported earlier (IO) may reflect this property. In this conjugate,hydrazone linkagesare thought to be formed when hydrazide-functionalizeddrugs are linked to oxidized carbohydrate residues. These linkages possess the ability to release active drug by virtue of their hydrolytic instability, in this case liberating DAVLB-HY, a highly cytotoxic vinca alkaloid (12). The heterogeneous nature of carbohydrate residues on the antibody glycoprotein inherently limits our ability to control the conjugate chemistry or the chemical properties of the linkages. Our assumption has been that the hydrolytic instability of hydrazone conjugates is an important factor in moAb targeted drug delivery and the control over this parameter is desirable for more effective chemoimmunoconjugatebased tumor therapy. In this report, we have employed alternative strategies to chemically introduce aldehyde or enamide groups onto lysine residues of the antibody to afford linking chemistries with differential stabilities. The @-alaninepyrrole(“BAP”)chemistry is an example of the successful application of the aldehyde/hydrazone-based approach. Alternatively, @-alanine-2-methylenemalonic acid ethyl ester (“BAMME”)was developed to release free DAVLB-HY by a hydrolytically cleavable enamide type linkage. These resulting two conjugates demonstrated excellent immunoreactivity and differential stability of the drug-monoclonal antibody linker, and their antitumor activity was compared to that established with KS114DAVLB-HY (IO). However, additional studies are warranted to establish potential toxicities associated with these alternative conjugation strategies to determine true therapeutic indices. This differential in vitro stability correlated with the differential in vitro cytotoxicity profiles of the KS1/4DAVLB-HY (IC50 = 17 ng/mL), KS1/4-BAP-DAVLB-HY (IC50= 63 ng/mL), and KSU4-BAMME-DAVLB-HY(IC50 = 128 ng/mL) chemoimmunoconjugates. All three chemoimmunoconjugates demonstrated significant survival advantages in vivo under varied conditions in the OVCAR-3human xenograft model. The KSl/CDAVLBHY and KSU4-BAP-DAVLB-HY conjugates had comparable in vivo efficacy. These data suggest that our observed 3.7-fold drop in in vitro potency between these two conjugates did not translate to a comparable decrease in in vivo antitumor activity. The increased stability of the chemicallinker of KSl/4-BAP-DAVLB-HY over KS1/ 4-DAVLB-HY could be an important factor when considering this agent for future studies, as clinical trials with KSl/CDAVLB-HY have resulted in some patients developing peripheral neuropathies, a possible consequence of inappropriate kinetics of release of vinca alkaloid (19). A more stable covalent bond between the drug moiety and the antibody and an increase in serum half-life may translate to more effective tumor localization and drug delivery. The least potent conjugate in vitro, KSl/CBAMMEDAVLB-HY, was also the least effective agent in vivo in
126 Bioconlugate Chem., Vol. 4, No. 2, 1993
increasing the overall survival of the animals. Route of administration did not alter the efficacy of the KS1/4DAVLB-HY and KSl/CBAP-DAVLB-HY conjugates. However, ip therapy with the KSl/CBAMME-DAVLBHY conjugate was significantly more effective than iv administration a t the 0.75 mg/kg vinca content dose. At this time, this differential activity between iv and ip administration of KSl/CBAMME-DAVLB-HY is not understood. Intraperitoneal therapy may provide a prolonged residence of the conjugate in the peritoneal cavity when administered as an intracavitary agent. In addition, both KSl/CBAMME-DAVLB-HY and the non-antigenbinding, subisotype-equivalent immunoconjugate 9.2.27BAMME-DAVLB-HY demonstrated efficacy at the 3 mg/ kg dose level after ip administration. This non-antigenmediated efficacy of 9.2.27-BAMME-DAVLB-HY may reflect a nonspecific drug depot effect in the peritoneal cavity with this immunoconjugate at this high dose. This study did demonstrate antigen-mediated tumor efficacy a t the 1.5 mg/kg (vinca content) dose level with the KS1/ 4-BAMME-DAVLB-HYconjugate where 9.2.27-BAMMEDAVLB-HY was inactive. Further studies of the biodistribution and/or catabolism of the conjugates after iv or ip administration are necessary to correlate efficacy,route of administration, and chemical linker employed. This result does suggest that the conjugate could have overall utility as an intracavitary agent where enhanced conjugate stability would be advantageous (20). The two novel conjugates introduced by this study, KSl/CBAP-DAVLBHY and KSl/kBAMME-DAVLB-HY, provide alternatives for further preclinical and clinical evaluation to determine their tumor saturation and drug-release characteristics. This report summarizes the initial characterization of two chemoimmunoconjugates of the KS1/4 monoclonal antibody with the potent vinca alkaloid DAVLB-HY. They were compared to the chemimmunoconjugate KS1/4DALVB-HY, which had been shown previously to have excellent in vivo activity (IO). The alternative chemical linkers reported here, i.e. BAMME and BAP, resulted in a more chemically stable chemoimmunoconjugate than KSU4-DAVLB-HY. All of the conjugates examined demonstrated significant survival advantages in the OVCAR-3human xenograft model under conditions where unmodified DAVLB-HY is inactive. These data suggest that the KS1/4 antigenic system represents a therapeutic target for human ovarian cancer when the KS1/4 monoclonal antibody is covalently conjugated to vinca alkaloid by several diverse linker chemistries. In addition, the BAMME and BAP linker chemistries may be utilized for the conjugation of drug moieties other thanvinca alkaloids to other antibodies or proteins where the hydrazone- or enamide-based chemistries are applicable, providing flexibility in future design of chemoimmunoconjugates for site directed therapy. LITERATURE CITED (1) Richardson, G. S., Scully, R. E., Nikrui, N., and Nelson, J. H., Jr. (1985) Medical progress: Common epithelial cancer of the ovary. N . Engl. J. Med. 312, 415-424. (2) Young,R. C., Knapp, R. C., Fuks, Z., and DiSaia, P. J. (1985) Cancer of the ovary. In Cancer Principles and Practice of Oncology (V. T. Devita, Jr., S. Hellman, and S. A. Rosenberg, Eds.) Vol. 1, p 1083, J B Lippincott, Philadelphia. (3) M a r k ” , M. (1984) Medical principles of intraperitoneal and intrapleural chemotherapy. In Intra-arterial and int racauitary cancer chemotherapy Howell, (S.B. Ed.) pp 1-69, Nijhoff, Boston.
Apelgren et al.
(4) Rosenshein, N. B., Leichner, P. K., and Vogelsang, G. (1979) Radiocolloids in the treatment of ovarian cancer. Obstet. Gynecol. Suru. 34, 708-720. ( 5 ) Bast, R. C., and Knapp, R. C. (1987) The Emerging Role of Monoclonal Antibodies in the Clinical Management of Epithelial Ovarian Carcinoma. In Important Aduances in Oncology (V. T. Devita, Jr. Ed.) pp 39-53, J. B. Lippincott Company, Philadelphia. (6) Simpson, J., and Schlom, J. (1988) The Use of Monoclonal Antibody B72.3 in the Management of Gynecologic Malignancies. Yale J. Biol. Med. 61, 351-366. (7) Riva, P., Marangolo, M., Lazzari, S., Agostini, M., Sarti, G., Moscatelli, G., Franceschi, G., Spinelli, A., and Vecchietti, G. (1989) Locoregional immunotherapy of human ovarian cancer: Preliminary results. Nucl. Med. Biol. 16, 659-666. (8) Pastan, I., Willingham, M. C., and Fitzgerald, D. J. (1986) Zmmunotoxins Cell 47, 641-648. (9) Pimm, M. V. (1988) Drug-Monoclonal antibody conjugates for Cancer Therapy: Potentials and Limitations. Crit. Rev. Ther. Drug Carrier Syst. 5, 189-226. (10) Apelgren, L. D., Zimmerman,D. L., Briggs, S. L., and Bumol, T. F. (1990) Antitumor Activity of the MonoclonalAntibodyVinca Alkaloid Immunoconjugate LY203725 (KS1/4-4-Desacetylvinblastine-3-carboxhydrazide)in a Nude Mouse Model of Human Ovarian Cancer. Cancer Res. 50, 3540-3544. (11) Laguzza, B. C., Nichols, C. L., Briggs, S. L., Cullinan, G. J., Johnson, D. A., Starling, J. J., Baker, A. L., and Bumol, T. F. (1989) New Antitumor Monoclonal Antibody-Vinca Conjugates LY203725 and Related Compounds: Design,Preparation, and Representative In Vivo Activity. J.Med. Chem. 32,548555. (12) Lindstrom, T. D., Althaus, W. A., Rutherbories, K. J., and Kau, D. (1989)Pharmacokineticsand disposition of the KS1/4 monoclonal antibody-deaacetylvinblastinehydrazideconjugate LY203725 in rats and monkeys. J. Pharmacol. Erp. Ther. 252, 1117-1124. (13) Hamilton, T. C., Young, R. C., Louie, K. G., Behrens, B. C., McCoy, W. M., Grotzinger, K. R., and Ozols, R. F. (1984) Characterization of a Xenograft Model of Human Ovarian Carcinoma Which Produces Ascites and Intraabdominal Carcinomatosis in Mice. Cancer Res. 44, 5286-5290. (14) Varki, N. M., Reisfeld, R. A., and Walker, L. E. (1984) Antigens associated with a human lung adenocarcinoma defined by monoclonal antibodies. Cancer Res. 44,681-686. (15) Bumol,T. F., and Reisfeld,R. A. (1982)Unique glycoproteinproteoglycan complex defined by monoclonal antibody on humanmelanomacells. Proc. Natl. Acad. Sci. U.S.A.79,12451249. (16) Mirabelli, C. K., Bartus, H., Bartus, J. 0. L., Johnson, R., Mong, S. M., Sung, C. P., and Crooke,S. T. (1985)Application of a tissue culture microtiter test for the detection of cytotoxic agents from natural products. J. Antibiot. 38, 758-766. (17) Thomas, D. G., Breslow, N., and Gart, J. J. (1977) Trend analysisand homogeneityanalysesof proportions and life table data. Comput. Biomed. Res. 10, 373-381. (18) Kaplan, E. L., and Meier, P. (1958) Non-parametric estimation from incomplete observations. J.Am. Stat. Assoc. 53, 457-481. (19) Schneck, D., Butler, F., Dugan, W., Littrell, D., Petersen, B., Bowsher, R., Delong, A., and Dorrbecker, S. (1990) Disposition of a murine monoclonal antibody vinca conjugate (KSl/CDAVLB) in patients with adenocarcinomas. Clin. Pharmacol. Ther. 47, 36-41. (20) Schlom, J. Horan Hand, P., Gerteiner, J. W., Colcher, D., Shrivastav, S., Carrasquillo, J. A., Reynolds, J. C., Larson, S. M., and Raubitschek, A. (1990) Innovations That Influence the pharmacology of Monoclonal Antibody Guided Tumor Targeting. Cancer Res. (Suppl.) 50, 8208-827s.
