A Calicheamicin Conjugate with a Fully Humanized Anti-MUC1

Feb 24, 2005 - Murine CTM01 is an internalizing murine IgG1 monoclonal antibody that recognizes the MUC1 antigen expressed on many solid tumors of ...
2 downloads 0 Views 280KB Size
Download PDF
Bioconjugate Chem. 2005, 16, 354−360

354

A Calicheamicin Conjugate with a Fully Humanized Anti-MUC1 Antibody Shows Potent Antitumor Effects in Breast and Ovarian Tumor Xenografts Philip R. Hamann,*,† Lois M. Hinman,† Carl F. Beyer,† Delores Lindh,† Janis Upeslacis,† Dan Shochat,† and Andrew Mountain‡ Wyeth Research, 401 N. Middletown Road, Pearl River, New York 10965 and Celltech R & D, 216 Bath Road, Slough, England, SL1-4EN. Received August 26, 2004; Revised Manuscript Received December 6, 2004

Murine CTM01 is an internalizing murine IgG1 monoclonal antibody that recognizes the MUC1 antigen expressed on many solid tumors of epithelial origin. Calicheamicin conjugates of this antibody have previously been shown to be potent, selective antitumor agents in preclinical models. A conjugate has now been made with a genetically engineered human version of this antibody, hCTM01. The hCTM01 is an IgG4 isotype, has an immunoaffinity approximately 30% higher than mCTM01 by competitive RIA, and is efficiently internalized into target cells. The hCTM01-NAc-gamma calicheamicin DM amide conjugate, referred to as CMB-401, shows targeted killing of MUC1-expressing cells in vitro and produces pronounced dose-related antitumor effects over an 8-fold dose range against a MUC1expressing, ovarian xenograft tumor, OvCar-3. The specificity of CMB-401 was confirmed by comparing its antitumor effects with those of an isotype-matched nonspecific conjugate against the MX-1 breast carcinoma. CMB-401, given either ip or iv, was highly active in these models in single and multiple dose regimens and gave complete regressions at the highest doses examined with good overall therapeutic ratios. CMB-401 also gave good antitumor effects at similar doses with a cisplatin-resistant MUC1-expressing cell line.

INTRODUCTION

The calicheamicins are highly potent antitumor antibiotics of the enediyne family. Originally isolated from a broth extract of the soil microorganism M. echinospora ssp. calichensis, the calicheamicins were detected in a screen for potent DNA damaging agents (1). The parent of this family of xenobiotics, calicheamicin γ1I (1, Figure 1), associates with the minor groove of DNA and, following reduction with glutathione (2), causes sequencespecific, double-stranded DNA cleavage (3, 4). Due to their mode of action and potency, several analogues of the calicheamicins have been tested in preclinical models for their usefulness as antitumor agents. Their development as single agent therapies has not been pursued because of delayed toxicities that limit the therapeutic dose range for treatment. However, their potency makes them appealing candidates for antibody-targeted chemotherapy, where limits in antigen expression make the potency of the cytotoxic agent crucial. The murine monoclonal antibody CTM01, also known as CT-M-01 or 7F11C7 (5), is an internalizing IgG1 that recognizes the MUC1 antigen present on a broad spectrum of solid tumors of epithelial origin. The MUC1 antigen is a glycoprotein that is expressed on the nonvascular surface of normal epithelial cells, where it is normally inaccessible from circulation. It is present in underglycosylated form on the nonvascular as well as the vascular surface of a number of epithelial cancer cells, such as breast, lung, ovarian, and colon carcinomas, thereby making it accessible for targeting. This antigen undergoes * To whom correspondence should be addressed. Phone: (845) 602-3423. Fax: (845) 602-5561. E-mail: [email protected]. † Wyeth Research. ‡ Celltech R & D.

cycling between the surface and the interior of the cell (5) and carries any bound antibody with it into the cell. Our results with CTM01 as well as numerous other antibodies suggest that they must be internalized to make effective targeting agents for potent drugs such as the calicheamicins (6, 7). CTM01 was selected for development after comparison of this antibody with a panel of anti-MUC1 monoclonal antibodies (8). Successful targeting of this antibody to tumors in humans has been demonstrated (9). A series of conjugates of the calicheamicins and mCTM01 have previously been described (10, 11). Initial examination of the structural requirement for the calicheamicin portion of the molecule led to the selection of the carbohydrate conjugate 4 which contains NAcgamma1 calicheamicin as the calicheamicin core, a disulfide stabilized by two adjacent methyl groups, and a hydrazone linkage to the oxidized carbohydrates of the antibody as a site of hydrolytic release (10). Further work showed that the inclusion of the hydrazone was necessary for the anti-CD33 antibody, P67.6, for potent, selective cytotoxicity (12). However, for CTM01, including this hydrolyzable bond was not necessary for activity, and the lysine-linked conjugate 7, referred to as the mCTM01 amide conjugate, was equivalent or even superior in all preclinical models that were examined, especially against resistant cell lines (13). CTM01 has now been fully humanized by genetic engineering as an IgG4 isotype (8) in order to decrease immunogenicity and permit repeated administration in 1 Abbreviations: cal equiv, calicheamicin equivalents; MTD, maximum tolerated dose; MED, minimum effective dose; ip, intraperitoneal; iv, intravenous, sc, subcutaneous; NAc, Nacetyl; DMA, dimethyl acid; gamma calicheamicin, calicheamicin γ1.

10.1021/bc049794n CCC: $30.25 © 2005 American Chemical Society Published on Web 02/24/2005

Humanized Anti-MUC1 Calicheamicin Conjugate

Bioconjugate Chem., Vol. 16, No. 2, 2005 355

have recently been reported (6, 7). Clinical experience with these two conjugates will also expand our understanding of the important issues that need to be addressed in order to make successful chemoimmunoconjugates. EXPERIMENTAL PROCEDURES

Figure 1. Structures of calicheamicin derivatives and conjugates.

humans. This fully humanized antibody, hCTM01, has been evaluated for its potential for targeted therapy by examining its ability to bind to and internalize into target cells. The potency and selectivity of the amide conjugate of hCTM01, referred to as CMB-401, in in vitro and in vivo tumor models is also described. CMB-401 (7) has been in clinical trials for ovarian cancer (14-16) as well as lung cancer (unpublished) with only limited evidence of activity. This is in contrast to gemtuzumab ozogamicin (Mylotarg) (17), which is an anti-CD33 calicheamicin conjugate approved for AML. The differences that may be responsible for the lesser activity include the presence of circulating antigen (18), possible detrimental effects of the predose of native antibody given in an attempt to clear the circulating antigen (19), differences in the linkers used (12), possible differences in resistance mechanisms (20), possible insufficient targeting of the conjugate to large solid tumors, etc. It is also possible that choosing a different tumor type, dosing patients after surgical removal of the primary tumor, or use in combination would show activity. Ongoing efforts may help to clarify these issues. Preclinical data on two more calicheamicin-antibody conjugates

Antibodies. The mCTM01 used in these studies was produced and purified from tissue culture supernatants by Celltech (Lot 122 88JC00-01). The fully humanized version of CTM01 was prepared by Celltech (8). hP67.6, an IgG4, anti-CD33 antibody that served as a nonbinding antibody control (21), and MOPC-21, also a nonbinding control (22), were also prepared by Celltech. Antibody Conjugates. The synthesis of the carbohydrate conjugates (4) of NAc-gamma calicheamicin DMH have been reported previously (10). Drug loadings averaged between 2 and 3 mol of calicheamicin per mole of antibody and free calicheamicin (predominantly structure 3) was normally 1-3%. The procedures for the synthesis of amide conjugates (7) have also been published (12). Drug loadings of 2 to 3 mol of calicheamicin per mole of antibody were seen and free calicheamicin (predominantly structure 5) was normally 0.5-1%. All conjugates were characterized as monomeric by analytical gel-exclusion chromatography (aggregate < 3%) and loading was determined by the ratio of UV absorption at 280 nm (protein) and 333 nm (calicheamicin) as reported (10, 12). Neither conjugation procedure significantly affected the immunoaffinity of the conjugates reported in this paper versus the starting native antibodies, as determined by competitive binding with antibody that had been radiolabeled with 125I-BoltonHunter reagent, as detailed below. Cell Lines. MX-1 is a human ductal breast carcinoma cell line obtained from the National Cancer Institute. MCF-7 is a human breast adenocarcinoma, pleural effusion. ZR-75-1 is a human breast carcinoma. Raji is a Burkitt’s Lymphoma cell line, ATCC # CCL 86. A2780 (23) is an ovarian carcinoma, and A2780DDP (24) was derived from A2780 by exposure to cisplatin and has enhanced DNA repair. The ovarian adenocarcinoma tumor OvCar-3 was obtained from the ATCC line HTB 161 and was passaged as an ascites tumor in nude mice. Culture Methods. The basic methods used are as published previously (10). Cells were propagated in RPMI 1640 medium containing 5% fetal calf serum, 5 µg/mL insulin and transferrin, 5 ng/mL selenium, and 50 µg/ mL gentamicin. Cultures were maintained in a humidified, 5% CO2 incubator at 37 °C and were subcultured once or twice a week as needed. For in vitro cytotoxicity tests, streptomycin (50 µg/mL) and penicillin (50 U/mL) were incorporated into the medium. Relative immunoaffinity of the antibodies and conjugates was measured by a competitive radioimmunoassay (RIA) in which several dilutions of the test sample were competed with iodinated CTM01 for binding to live cells. For each assay, 106 A2780DDP cells in 0.1 mL were incubated with 0.05 mL of 4 µg/mL 125I-CTM01 (specific activity about 0.3-0.5 µCi/µg) and serial 4-fold dilutions of the test samples, the highest concentration being 200 µg/mL. After a 1 h incubation at 0-4 °C, the cells were washed three times with cold Dulbecco’s phosphate-buffered saline (PBS) and transferred to fresh tubes and counted. Binding inhibition curves were plotted and calculated IC50 values for each conjugate were compared with that of the unmodified control antibody as a relative measure of retention of immunoaffinity. Determination of Antibody Internalization by Target Cells. The acid-dissociation method (25) has been

356 Bioconjugate Chem., Vol. 16, No. 2, 2005

adapted to quantitatively assess internalization of mCTM01 and hCTM01 by the human breast carcinoma cell line, MCF-7, and the cisplatin-resistant human ovarian carcinoma cell line, A2780DDP. Cells were collected by trypsinization and washed; viable cells were enumerated by Trypan Blue exclusion, and the suspension was adjusted to 5 × 106 viable cells/mL in medium (RPMI1640 containing 5% heat-inactivated fetal calf serum and 0.02 M HEPES, pH 7.4). This suspension was chilled on ice, and 125I-labeled antibody (labeled with Bolton-Hunter reagent; specific activity about 0.5 mCi/mg) was added to give a final concentration of 1-2 mg/106 cells. This suspension was distributed to two sets of duplicate tubes at 0.5 mL/tube. One set of tubes (control) was kept on ice throughout the incubation period; the other set was immediately transferred to a 37 °C water bath. At various times, the cell suspensions in duplicate tubes from each set were diluted with 2.5 mL of cold medium and centrifuged. The supernatants were decanted, and the cell pellets were washed with 3 mL of cold medium. The supernatants were decanted again, and the pellets were resuspended in 0.5 mL of 0.05 M glycine and 0.10 M NaCl, pH 2.8. These suspensions were incubated for 20 min at room temperature and then centrifuged. The acid supernatants were carefully removed and transferred to clean tubes; the pellets were resuspended in 0.5 mL of medium and also transferred to clean tubes. The radioactivity associated with cell pellets and acid supernatants was determined by gamma counting. This method is based on the finding that molecules of CTM01 that are bound to the outer plasma membrane of cells are released by exposure to isotonic pH 2.8 buffer, whereas internalized antibody molecules are not released. Thus, it is possible to distinguish between surface-bound and internalized CTM01 by this technique. Over time, there is a shift of radioactivity from the acid supernatants to the cell pellets for samples incubated at 37 °C but not for the controls incubated on ice. From the acid-stable, cellassociated cpm in the experimental samples (corrected for the small, constant, acid-stable cpm in the control samples) and the specific radioactivity of the antibody, the number of molecules of antibody internalized per cell (net uptake) is plotted vs time of incubation. In Vitro Cytotoxicity Assays. To evaluate the cytotoxicity of drug or conjugate samples, viable cells (105/ 0.2 mL) were aliquoted into 15 mL test tubes which contained 0.2 mL of the sample to be tested at the appropriate concentration. Concentrations were all normalized to ng/mL calicheamicin equivalents (cal equiv). Tubes were vortexed and incubated at 37 °C for 1 h, and the pellets were washed 3 times with 8 mL of medium. One milliliter of medium was added to each pellet, the cells were vortexed, and 0.2 mL was removed and placed in a well of a 96-well plate (2 × 104 cells). Alternately for continuous exposure, cells were plated without removing unbound test article by washing. These cells were incubated for 3 days, at which time 0.1 mL of supernatant was removed and replaced with 0.2 mCi of [3H] thymidine in 0.1 mL of fresh medium. Incubation was resumed for an additional 24 h at which point the cells were harvested and counted. Growth inhibition curves of each drug or conjugate were plotted, and the IC50 value (concentration of drug equivalents needed for 50% [3H]thymidine uptake inhibition) of each sample was determined. In Vivo Tests for Antitumor Activity. The cell lines were passaged in tissue culture or as ascites tumors in nude mice (OvCar-3). For each experiment, 8 × 106 cells were implanted subcutaneously into athymic mice. Tumors were staged for 4 to 10 days to achieve a size of

Hamann et al.

Figure 2. Competitive radioimmunoassay for mCTM01 and hCTM01. Affinity of the humanized antibody relative to the murine antibody is 1.3.

Figure 3. Internalization by MCF-7 (A) and A2780DDP (B) human cell lines. Frame A shows the effect of humanization on uptake, while frame B shows the effect of conjugation to make CMB-401 (7) on uptake of humanized antibody.

100-150 mg prior to treating ip at the indicated dose levels on days 1, 5, and 9 after staging. There were five or six mice per test group and 10 in the buffer control group. Tumor mass was determined by measuring two tumor dimensions once weekly for at least 28 days posttumor staging. Antitumor activity was defined as a sustained 58% minimum inhibition of mean tumor mass (NCI standard criteria) compared with untreated controls in groups with greater than 65% survivors. A therapeutic ratio was defined as the maximum tolerated dose/minimum effective dose (MTD/MED) and used as a measure of the therapeutic window for the conjugates or drugs tested. RESULTS

Relative Affinities of mCTM01 and hCTM01. The affinities of the murine and humanized versions of CTM01 were compared using 125I-labeled mCTM01 in a competition assay. As shown in Figure 2, the binding data follow a normal curve and show that hCTM01 has ∼30% increased affinity with respect to the murine antibody. Internalization of mCTM01, hCTM01, and CMB401. Internalization of the two CTM01 antibodies and CMB-401 was examined on two cell lines using a modification of the acid-dissociation method. The increased affinity of the genetically engineered antibody resulted in increased internalization into the MCF-7 cell line, as shown in Figure 3A, and these results confirmed the observation (5) that MUC1 is efficiently internalized. In

Bioconjugate Chem., Vol. 16, No. 2, 2005 357

Humanized Anti-MUC1 Calicheamicin Conjugate Table 1. In Vitro Cytotoxicity of CMB-401 and the mCTM01 Amide Conjugate IC50, 1 ha

IC50, 4 daysa

sample

MX-1

Raji

CMB-401 mCTM01 conjugatec NAc-gamma cal.

14.9 54.2 >20

>500 >500 0.53

indexb

specificity

>1000 >300 1

MX-1

Raji

indexb specificity

0.05 0.61 1.08

>100 >100 0.03

>70000 >6000 1

a IC s are given in ng/mL cal equiv, doses of antibody are approximately 40-fold higher. b The specificity index is defined as (IC 50 50 of conjugate of Raji/IC50 of conjugate on MX-1) ÷ (IC50 of NAc-gamma on Raji/IC50 of NAc-gamma on MX-1). c Same conjugate as CMB-401, but with murine antibody.

Table 2. In Vitro Activity of the CMB-401 and the MCTM01 Amide Conjugate IC50 (ng/mL) sample

A2780DDP

ZR-75-1

MCF-7

CMB-401 mCTM01 conjugated MOPC-21 conjugatee hP67.6 conjugatee NAc-gamma calicheamicin selectivity indexc

1.8 2.4 189 NDb 0.056 105

9.8 NDb NDb 331 0.038 34

17 20 305 NDb 1.8 18

a IC s are given in ng/mL cal equiv, doses of antibody are 50 approximately 40-fold higher. b ND ) not determined. c Selectivity index ) IC50(MOPC-21 or hP67.6 conjugate) ÷ IC50(CMB-401). d Same conjugate as CMB-401, but with murine antibody. e Same conjugate as CMB-401, but with indicated antibody.

this example at least 50 000 molecules per cell were internalized within the first hour of incubation at 37 °C. Figure 3B shows that the A2780DDP cell line had a slightly lower rate of internalization for hCTM01, but that CMB-401 was internalized at least as well, indicating that conjugating calicheamicin to the antibody does not interfere with internalization. Comparative Activity of CMB-401 and the Corresponding Murine Conjugate in Vitro. The mCTM01 amide conjugate and CMB-401 routinely had drug loadings of 2 to 3 mol of calicheamicin per mole of antibody and retained >80% of their immunoaffinity (data not shown). The in vitro activity of CMB-401 was compared with that of the mCTM01 amide conjugate. The MX-1 breast line was used as the antigen-positive cell line and Raji as an antigen-negative control cell line with either 1-h exposure or continuous (4-day) exposure to conjugate. Additionally, the activities of all the conjugates were compared with that of unconjugated NAc-gamma calicheamicin (2). As summarized in Table 1, CMB-401 had lower IC50s and higher specificities in both the 1-h incubation and continuous exposure protocols than seen with the mCTM01 amide conjugate. To evaluate the selectivity of these amide conjugates for other MUC1-expressing tumor cell lines, an in vitro comparison of CMB-401 with a nonspecific antibody conjugate, prepared with identical calicheamicin chemistries, was made. In this study, the mCTM01 amide conjugate and CMB-401 along with both NAc-gamma calicheamicin (2) and the nonbinding control conjugate of the murine IgG1 monoclonal antibody, MOPC-21, or the isotype-matched conjugate of hP67.6 were all tested against two human breast and one ovarian tumor cell lines. As the data in Table 2 indicate, both the mCTM01 amide conjugate and CMB-401 were equivalently active against the MUC1-expressing cell lines, with IC50s in the 10 ng/mL of cal equiv dose range. The specificity of CMB401 ranged from 20 to 100 with respect to a nonbinding conjugate. Table 2 also shows that both CTM01 amide conjugates were at least as potent on the cisplatin resistant cell line A2780DDP as on the other two cisplatin sensitive cell lines.

Figure 4. Response of OvCar-3. Xenograft tumors were treated with CMB-401 (7), the murine conjugate (7), or a representative calicheamicin derivative (6) alone. Athymic mice were implanted sc with OvCar-3 cells and then dosed when tumors were approximately 150 mg in size. The doses indicated were given on days 7, 11, and 15, as indicated by the arrows, and are given in calicheamicin equivalents. The corresponding concentration of antibody is 40-fold greater. Error bars are SEMs.

In Vivo Efficacy of CMB-401 versus OvCar-3. As shown in Figure 4, the mCTM01 amide conjugate and CMB-401 were both highly effective against the OvCar-3 ovarian carcinoma xenograft at doses ranging from 50 to 300 µg/kg cal equiv given three times each, with four out of five animals tumor-free in each of the 300 µg/kg treatment groups. This experiment was done with tumors staged to 150 mg prior to treatment. The conjugates were still effective at doses as low as 50 µg/kg, indicating a therapeutic ratio (MTD/MED) g 6. In contrast, neither the NAc-gamma calicheamicin (2, not shown) nor the NAc-gamma calicheamicin DMA (5, shown) used as controls, gave substantial antitumor effects at the MTD. The dose-response behavior of the conjugates in this experiment is described in Figure 5. With both conjugates, 13 of the 25 treated animals were tumor-free at day 28 of the test as indicated in the figure. Native antibody itself has no effect in this or any other xenograft experiment that has been performed. Evaluation of Specificity and Efficacy in Vivo versus MX-1. As shown in Figure 6, CMB-401 was compared with a conjugate prepared with the isotypematched, genetically engineered IgG4 antibody, hP67.6, that does not target to MX-1, conjugated to the same calicheamicin derivative. In this study, the two conjugates were administered ip on days 11, 15 and 19 post tumor implantation. At doses ranging from 150 to 450 µg/kg cal equiv per injection the specific hCTM01 amide conjugate was active while the nontargeting hP67.6 conjugate had comparatively modest activity (T/C ) 43%) at the top dose tested.

358 Bioconjugate Chem., Vol. 16, No. 2, 2005

Figure 5. Dose-response of the OvCar-3 ovarian xenograft tumors. Athymic mice were implanted sc with OvCar-3 cells and then dosed with the conjugates (7) when tumors were approximately 150 mg in size. The doses indicated were given on days 7, 11, and 15 and are given in calicheamicin equivalents. The corresponding concentration of antibody is 40-fold greater. Data shown is for day 28.

Figure 6. Response of the MX-1 breast xenograft tumors. Athymic mice were implanted sc with MX-1 cells and then dosed with the conjugates (7) when tumors were approximately 150 mg in size. The control conjugate is an identical conjugate made with the nonbinding, isotype-matched antibody, P67.6. The doses indicated were given on days 11, 15, and 19 and are given in calicheamicin equivalents. The corresponding concentration of antibody is 40-fold greater. Error bars are SEMs and are given in the negative direction for the control conjugate and in the positive direction for CMB-401.

Evaluation of Dose Schedule and Route of Administration. The MX-1 breast carcinoma was used to examine both route of administration and dose schedule for CMB-401. Figure 7 shows no difference in the activity of CMB-401 administered ip as a larger bolus injection versus the standard three doses separated by 4 days. There was also no significant difference in activity between the conjugates as a function of route of administration, ip versus iv, regardless of whether the conjugate was given as a bolus or as split doses (bolus iv not shown in graph for clarity). DISCUSSION

The feasibility of conjugating calicheamicin to oxidized carbohydrates on antibodies has previously been demonstrated (10). Using these carbohydrate conjugates, the

Hamann et al.

Figure 7. Effect of changing the fractionated dose IP to either a fractionated dose iv or to a bolus dose IP. Athymic mice were implanted sc with MX-1 cells and then dosed with CMB-401 (7) when tumors were approximately 150 mg in size. The doses indicated were given either on day 11, as indicated by the solid arrow (single dose), or on days 11, 15, and 19, as indicated by the open arrows, and are given in calicheamicin equivalents. The corresponding concentration of antibody is 40-fold greater. Error bars are SEMs.

correlation between the potency of calicheamicin analogues and their therapeutic potential as conjugates was investigated, and linker variations were also examined in an effort to optimize the therapeutic window for the conjugates. From these studies, the mCTM01 carbohydrate conjugate (4), which contains an NAc group, a disulfide that is sterically hindered by two methyl groups, and a hydrazone that allows for hydrolytic drug release, was chosen as the optimal drug for conjugation. Although incorporation of the NAc group lowered the potency of the conjugate somewhat, it significantly improved the therapeutic index. Increasing the stability of the disulfide is thought to decrease premature release and triggering of the calicheamicin in circulation. The hydrazone linkage was originally thought to be required for efficient drug release, as is the case for conjugates of other antibodies (12). The site of hydrolytic release was subsequently removed by attaching an appropriate calicheamicin derivative to the epsilon-amino groups of lysines (11). The mCTM01 amide conjugate (7) is not only as efficacious as the hydrazone conjugate in many in vitro and in vivo tumor models, but also shows enhanced activity over the hydrazone conjugate against tumors overexpressing Pglycoprotein-mediated multidrug resistance (MDR) (13). These studies suggest the hypothesis that attaching a cytotoxic agent, such as calicheamicin, by way of a serum stable linker to an internalizing carrier protein may help to bypass normal membrane-based mechanisms of drug resistance. There is precedence for this with other types of conjugates (26), although further research with various drugs and antibodies would be required to adequately define this possibility. The amide conjugate also shows good activity versus a cisplatin-resistant ovarian carcinoma line. That this conjugate has activity equal to or greater than that of the CTM01 carbohydrate conjugate indicates that the hindered disulfide offers ample means for release of the drug intracellularly, presumably through interaction with glutathione and other reduced thiols (2). Preliminary experiments indicate that the reaction rate

Humanized Anti-MUC1 Calicheamicin Conjugate

of a model calicheamicin disulfide with glutathione is consistent with reasonable stability in circulation where glutathione concentrations are lower, but sufficient reactivity to allow for release of an active calicheamicin derivative intracellularly where glutathione concentrations are 100-fold higher (27, 28). Specifically, about 2% reaction of the calicheamicin was seen at 0.02 mM glutathione in pH 7.4 phosphate buffer, which approximates the conditions in circulation, and 55% reaction was seen at 2 mM, which approximates the concentrations seen in cells. The next step in the development of the CTM01 calicheamicin conjugates has been the replacement of the murine antibody with a genetically engineered human antibody that should be much less immunogenic in patients. The data presented in this report demonstrate that the resultant hCTM01 amide conjugate is wellbehaved and is equivalent or superior to the corresponding murine version in all preclinical models. Good binding and internalization of this and all other calicheamicin conjugates is felt to be crucial for good activity. Binding and internalization has been shown to be unaffected by the change from mCTM01 to hCTM01 and by conjugation to calicheamicin. In fact, both the native hCTM01 antibody as well as its conjugate bind and internalize better than the murine counterparts. In vitro, CMB-401 is potently cytotoxic with IC50s in the 0.05 to 17 ng/mL cal equiv range and is highly selective with specificity indices of 20 to over 70 000, depending on the cell line and assay conditions. Specificity has been demonstrated both versus a nontargeted cell line and a nontargeting conjugate on various cell lines. The reason for modest selectivity in some cases is unknown; however, the activity in xenograft models reaffirms the potential of these conjugates. CMB-401 exhibits good stability in vitro as demonstrated by the retention of specificity seen after 4 days of the continuous exposure to conjugate at pH 7.4 and 37 °C. In vivo, CMB-401 is virtually indistinguishable from the murine counterpart. The MED for multiple experiments falls in the range of 25 to 50 µg/kg cal equiv while the MTD ranges from 300 to 400 µg/kg, in each case being given three times spaced by 4 days. The therapeutic window in multiple experiments is always > 6. CMB401 produced regressions of well established (150 mg) OvCar-3 ovarian xenograft tumors with long-term, tumorfree survivors being common not only at the MTD, but often at lower doses as well. Good activity is maintained over a 10-fold dose range in some experiments. The targeting of CMB-401 is demonstrated by its superior activity with respect to a conjugate made from the hP67.6 antibody which does not target the MX-1 breast xenograft. The nontargeting conjugate shows only modest in vivo activity at a dose that is almost 10 times the MED for CMB-401. Such modest activity is likely due to the tendency of xenograft tumors to accumulate macromolecules, but is significantly less than seen with the targeting antibody, where internalization takes the calicheamicin into the cell. Experiments which examine the dosing schedule show that similar activity is seen with a higher bolus dose instead of three fractionated doses spaced 4 days apart. This is somewhat surprising, because chemotherapeutics are often fractionated to maximize effect while minimizing toxicity. Administration iv was seen to give the same results as the more normally used ip administration. Data from previous experiments also show that the CTM01 conjugates are as potent, both in vitro as well as in vivo, on the MUC1-expressing, cisplatin-resistant cell

Bioconjugate Chem., Vol. 16, No. 2, 2005 359

line A2780DDP as against the cisplatin-sensitive target cell lines (11). These observations were important since clinical evaluation of CMB-401 activity was initially made in ovarian carcinoma patients who had relapsed following cisplatin-based chemotherapy (14, 15). CMB-401 is a very potent antitumor agent, being routinely effective and nontoxic in the dose range of 50300 µg/kg cal equiv. This potency is crucial considering that uptake of antibody conjugates by solid tumors in humans is low, typically in the range of 0.001-0.01% injected dose/g. The potential of the calicheamicins for targeted delivery with a fully humanized anti-MUC1 antibody has been demonstrated in preclinical models. The lack of sufficient activity in clinical trials (see the Introduction for discussion) was disappointing, but it gives us additional incentive to explore the many other properties a chemoimmunoconjugate and the tumors it targets must possess for ultimate success. LITERATURE CITED (1) Lee, M. D., Dunne, T. M., Siegel, M. M., Chang, C. C., Morton, G. O., and Borders, D. B. (1987) Calicheamicins, a novel family of antitumor antibiotics, 1. Chemistry and partial structure of calicheamicin γ1I. J. Am. Chem. Soc. 109, 34643466. (2) Myers, A. G., Cohen, S. B., and Kwon, B. M. (1994) A study of the reaction of calicheamicin γ1Ι with glutathione in the presence of double-stranded DNA. J. Am. Chem. Soc. 116, 1255-1271. (3) Ellestad, G. A., Ding, W.-D., Zein, N., and Townsend, C. A. (1995) DNA-cleaving properties of calicheamicin γ1. Enediyne Antibiotics as Antitumor Agents (Borders, D. B., and Doyle, T. W., Eds.) pp 137-160, Marcel Dekker, New York. (4) Lee, M. D., Ellestad, G. A., and Borders, D. B. (1991) Calicheamicins: Discovery, structure, chemistry, and interaction with DNA. Acc. Chem. Res. 24, 235-243. (5) Aboud-Pirak, E., Sergent, T., Otte-Slachmuylder, C., Abarca, J., Trouet, A., and Schneider, Y.-J. (1988) Binding and endocytosis of a monoclonal antibody to a high molecular weight human milk fat globule membrane-associated antigen by cultured MCF-7 breast cancer cells. Cancer Res. 48, 31883196. (6) DiJoseph, J. F., Armellino, D. C., Boghaert, E. R., Khandke, K., Dougher, M., Sridharan, L., Kunz, K., Hamann, P. R., Gorovits, B., Udata, C., Moran, J. K., Popplewell, A. G., Stephens, S., Frost, P., and Damle, N. K. (2004) Antibodytargeted chemotherapy with CMC-544: A CD22-targeted immunoconjugate of calicheamicin for the treatment of B lymphoid malignancies. Blood 103, 1807-1814. (7) Boghaert, E. R., Sridharan, L., Armellino, D. C., Khandke, K. M., DiJoseph, J. F., Kunz, A., Dougher, M. M., Jiang, F., Kalyandrug, Lyka B., Hamann, P. R., Frost, P., Damle, N. K. (2004) Antibody-targeted chemotherapy with the calicheamicin conjugate hu3S193-N-acetyl γ calicheamicin dimethyl hydrazide targets LewisY and eliminates LewisYpositive human carcinoma cells and xenografts. Clin. Cancer Res. 10, 4538-4549. (8) Baker, T. S., Bose, C. C., Caskey-Finney, H. M., King, D. J., Lawson, A. D., Lyons, A., Mountain, A., Owens, R. J., Rolfe, M. R., Sehdev, M., Yarranton, G. T., and Adair, J. R. (1994) Humanization of an anti-mucin antibody for breast and ovarian cancer therapy. Adv. Exp. Med. Biol. 1994, 3536135382. (9) Prinssen, H. M., Molthoff, C. F., Verheijen, R. H., Broadhead, T. J., Kenemans, P., Roos, J. C., Davies, Q., van Hof, A. C., Frier, M., den Hollander, W., Wilhelm, A. J., Baker, T. S., Sopwith, M., Symonds, E. M., and Perkins, A. C. (1998) Biodistribution of 111In-labeled engineered human antibody CTM01 (hCTM01) in ovarian cancer patients: Influence of prior administration of unlabeled hCTM01. Cancer Immunol. Immunother. 47, 39-46. (10) Hinman, L. M., Hamann, P. R., Wallace, R., Menendez, A. T., Durr, F. E., and Upeslacis, J. (1993) Preparation and characterization of monoclonal antibody conjugates of the

360 Bioconjugate Chem., Vol. 16, No. 2, 2005 calicheamicins: A novel and potent family of antitumor antibiotics. Cancer Res. 53, 3336-3342. (11) Hinman, L. M., Hamann, P. R., and Upeslacis, J. (1995) Preparation of conjugates to monoclonal antibodies. Enediyne Antibiotics as Antitumor Agents (Borders, D. B., and Doyle, T. W., Eds.) pp 87-105, Marcel Dekker, New York. (12) Hamann, P. R., Hinman, L. M, Beyer, C. F., Lindh, D., Upeslacis, J., Flowers, D. A., and Bernstein, I. (2002) An antiCD33 antibody-calicheamicin conjugate for treatment of acute myeloid leukemia. Choice of linker. Bioconjugate Chem. 13, 40-46. (13) Hamann, P. R., Hinman, L. M., Beyer, C. F., Greenberger, L., Lin, C., Lindh, D., Menendez, A. T., Wallace, R., Durr, F. E., and Upeslacis, J. (2004) An anti-MUC1 antibody-calicheamicin conjugate for treatment of solid tumors, choice of linker and overcoming drug resistance. Bioconjugate Chem. 16, 346-353. (14) Gillespie, A. M., Broadhead, T. J., Chan, S. Y., Owen, J., Farnsworth, A., Sopwith, M., and Coleman, R. E. (1998) Phase I study of ascending doses of the cytotoxic immunoconjugate CMB-401 (hCTM01-calicheamicin) in patients with epithelial ovarian cancer. Br. J. Cancer 78, 20. (15) Chan, S. Y., Gordon, A. Y., Coleman, R. Y., Hall, J. Y., Berger, M. S., Sherman, M. Y., Eten, C. Y., and Finkler, N. Y. (2003) A phase 2 study of the cytotoxic immunoconjugate CMB-401 (hCTM01-calicheamicin) in patients with platinumsensitive recurrent epithelial ovarian carcinoma. Cancer Immunol. Immunother. 52, 243-248. (16) Broadhead, T. J., Gillespie, A. M., Coleman, R. E., Owen, J., Skelton, L., Stephens, S., Farnsworth, A., Sopwith, M., and Chan, S. Y. (1998) Phase I study of the cytotoxic immunoconjugate CMB-401 (hCTM01-calicheamicin) in patients with epithelial ovarian cancer: Immune response and pharmacokinetics following repeated intravenous administration. Proceedings of the American Society of Clinical Oncology, May 16-19, 1998, Los Angeles, CA, American Society of Clinical Oncology: Alexandria, VA, 1998, abstract 1693. (17) Bross, P. F., Beitz, J., Chen, G., Chen, X. H., Duffy, E., Kieffer, L., Roy, S., Sridhara, R., Rahman, A. R., Williams, G., and Pazdur, R. (2001) Approval summary: Gemtuzumab ozogamicin in relapsed acute myeloid leukemia, Clin. Cancer Res. 7, 1490-1496. (18) Davies, Q., Perkins, A. C., Fries, M., Watson, S., Lalani, E.-N., and Symonds, E. M. (1997) The effect of circulating antigen on the biodistribution of the engineered human antibody hCTM01 in a nude mice model. Eur. J. Nuclear Med. 24, 206-209.

Hamann et al. (19) Broadhead, T. J., Davies, Q., Van Hof, A. C., Molthoff, C. F. M., Perkins, A. C., Roos, J. C., Prinssen, W. M., Verheijen, R. H. M., den Hollander, W., Sopwith, M., Frier, M., Kenemans, P., and Symonds, E. M. (1998) Clinical study to investigate the potential of an engineered human monoclonal antibody, hCTM01, for targeted chemotherapy of ovarian cancer. Tumor Target. 3, 55-57. (20) Walter, R. B., Raden, B. W., Cronk, M. R., Bernstein, I. D., Appelbaum, F. R., and Banker, D. E. (2004) The peripheral benzodiazepine receptor ligand PK11195 overcomes different resistance mechanisms to sensitize AML cells to gemtuzumab ozogamicin. Blood 103, 4276-4284. (21) Hamann, P. R., Hinman, L. M., Hollander, I., Beyer, C. F., Lindh, D., Holcomb, R., Hallett, W., Tsou. H.-R., Upeslacis, J., Shochat, D., Mountain, A., Flowers, D. A., and Bernstein, I. (2002) Gemtuzumab ozogamicin, a potent and selective anti-CD33 antibody-calicheamicin conjugate for treatment of acute myeloid leukemia. Bioconjugate Chem. 13, 47-58. (22) Melchers, F. (1970) Biosynthesis of the carbohydrate portion of immunoglobulins. Biochem. J. 119, 765-772. (23) Ozols, R. F. and Young, R. C. (1984) Chemotherapy of ovarian cancer. Semin. Oncol. 11, 251-263. (24) Newman, E. M., Lu, Y., Kashani-Sabet, M., Kesavan, V., and Scanlon, K. J. (1988) Mechanisms of cross-resistance to methotrexate and 5-fluorouracil in an A2780 human ovarian carcinoma cell subline resistant to cisplatin. Biochem. Pharmacol. 37, 443-7. (25) Matzku, S., Bro¨cker, E.-B., Bru¨ggen, J., Dippold, W. G., and Tilgen, W. (1986) Modes of binding and internalization of monoclonal antibodies to human melanoma cell lines. Cancer Res. 46, 3848-3854. (26) Goren, D., Horowitz, A. T., Tzemach, D., Tarshish, M., Zalipsky, S., and Gabizon, A. (2000) Nuclear delivery of doxorubicin via folate-targeted liposomes with bypass of multidrug-resistance efflux pump. Clin. Cancer Res. 6, 19491957. (27) Mamsoor, M. A., Svardal, A. M., and Ueland, P. M. (1992) Determination of the in vivo redox status of cysteine, cysteinylglycine, homocysteine, and glutathione in human plasma. Anal. Biochem. 200, 218-229. (28) Mistry, P. and Harrap, K. R. (1991) Historical aspects of glutathione and cancer chemotherapy. Pharm. Ther. 49, 125-132.

BC049794N