Design of Antibody−Maytansinoid Conjugates Allows for Efficient

ARTICLE pubs.acs.org/bc

Design of Antibody-Maytansinoid Conjugates Allows for Efficient Detoxification via Liver Metabolism Xiuxia Sun, Wayne Widdison, Michele Mayo, Sharon Wilhelm, Barbara Leece, Ravi Chari, Rajeeva Singh, and Hans Erickson* ImmunoGen, Inc., 830 Winter Street, Waltham, Massachusetts 02451, United States

bS Supporting Information ABSTRACT: Antibody-maytansinoid conjugates (AMCs) are targeted chemotherapeutic agents consisting of a potent microtubule-depolymerizing maytansinoid (DM1 or DM4) attached to lysine residues of a monoclonal antibody (mAb) using an uncleavable thioether linker or a stable disulfide linker. Most of the administered dose of an antibody-based therapeutic is slowly catabolized by the liver and other tissues of the reticuloendothelial system. Maytansinoids released from an AMC during this catabolic process could potentially be a source of toxicity. To investigate this, we isolated and identified liver metabolites in mice for three different [3H]AMCs with structures similar to those currently undergoing evaluation in the clinic. We then synthesized each metabolite to confirm the identification and assessed their cytotoxic potencies when added extracellularly. We found that the uncleavable mAbSMCC-[3H]DM1 conjugate was degraded to a single major maytansinoid metabolite, lysine-SMCC-[3H]DM1, that was nearly 50-fold less cytotoxic than maytansine. The two disulfide-linked conjugates, mAb-SPP-[3H]DM1 and mAb-SPDB-[3H]DM4, were also found to be catabolized to the analogous lysine-linked maytansinoid metabolites. However, subsequent reduction, Smethylation, and NADPH-dependent oxidation steps in the liver yielded the corresponding S-methyl sulfoxide and S-methyl sulfone derivatives. The cytotoxic potencies of the oxidized maytansinoids toward several human carcinoma cell lines were found to be 5- to 50-fold less potent than maytansine. Our results suggest that liver plays an important role in the detoxification of both cleavable and uncleavable AMCs.

’ INTRODUCTION Antibody-drug conjugates (ADCs) are targeted chemotherapeutic agents that utilize the specificity of monoclonal antibodies (mAb) to deliver potent cell killing agents to cancer cells that express the target antigen.1-3 Several antibody maytansinoid conjugates (AMCs) that utilize0 the tubulin-targeting maytansi0 2 -(3-mercapto-1-oxopropyl)-maynoid thiols, N2 -Deacetyl-N 0 20 tansine (DM1) and N -Deacetyl-N2 -(4-mercapto-4-methyl-1oxopentyl)-maytansine (DM4), are currently being evaluated in clinical trials. The most advanced of these conjugates, trastuzumab-SMCC-DM1 (T-DM1), is in multiple advanced clinical trials for the treatment of HER-2 positive metastatic breast cancer.4 Two other AMCs, SAR3419 (huB4-SPDB-DM4) and lorvotuzumab mertansine (huN901-SPP-DM1), have both shown promising results in early clinical trials for the treatment of B-cell malignancies and for treating solid and liquid tumors that express CD19 and CD56, respectively.5,6 Three additional AMCs are in phase I testing.7-9 The AMCs currently in the clinic utilize three different linkers to attach the maytansinoid (DM1 or DM4) to a tumor-specific humanized antibody (Figure 1). T-DM1 utilizes an uncleavable thioether linker while the other candidates utilize disulfide linkers (Figure 1). During the preclinical evaluation of each AMC, conjugates with the different linkers were prepared and evaluated. The AMC that was found to yield the widest therapeutic r 2011 American Chemical Society

window—as defined by the difference between the minimally effective dose (MED) in mouse xenografts models and the maximum tolerated dose (MTD)—was selected. In the majority of cases, preclinical evaluation of the disulfide-linked AMCs were found to yield the widest therapeutic window,6-8,10 despite their having an MTD that was about 2-fold lower than that of the uncleavable versions.11 An exception to these findings was the uncleavable T-DM1 conjugate that was found to be slightly more active in mouse xenograft models than several disulfide-linked versions.11 Several studies point to enhanced bystander killing of the disulfide-linked conjugates as a possible explanation for their typically greater activity when compared with uncleavable conjugates in mouse xenograft models.10,12,13 Internalization and lysosomal processing of the uncleavable conjugate in targeted cancer cells yields the lysine-SMCC-DM1 metabolite. While this metabolite is highly cytotoxic to the targeted cell in which it is formed, it is unable to penetrate and kill neighboring “bystander” tumor cells. Internalization and lysosomal processing of the disulfide-linked conjugates yields the analogous lysine-linkermaytansinoid derivatives. However, additional intracellular Received: November 15, 2010 Revised: January 26, 2011 Published: March 10, 2011 728

dx.doi.org/10.1021/bc100498q | Bioconjugate Chem. 2011, 22, 728–735

Bioconjugate Chemistry

ARTICLE

Figure 1. Structures of antibody-maytansinoid conjugates prepared for the current study. Tritium was incorporated into the C-20 methoxy group to allow for the detection of maytansinoid metabolites. Antibody-maytansinoid conjugates currently under clinical evaluation with similar structures that lack this tritium label are listed.

huN901-SMCC-[3H]DM1 (3.6 D/A, 284 mCi/mmol DM1), and huN901-SPDB-[3H]DM4 (2.9 D/A, 300 mCi/mmol DM4). The maytansinoid metabolites—lysine-Nε-SPP-DM1, lysine-NεSMCC-DM1, DM1-SMCC, lysine-Nε-SPDB-DM4, S-methyl-DM1, S-methyl-DM4-sulfoxide, S-methyl-DM1 sulfoxide, and S-methylDM4 sulfone were synthesized at ImmunoGen (manuscript in preparation). The DM1, DM4, and S-methyl-DM4 were prepared as previously described.17 In Vitro Cell Proliferation Assays. The cytotoxic potencies of the maytansinoids toward human carcinoma cell lines were assessed using a WST-based cell viability assay. Cells were seeded in 96-well flat-bottom plates at a density of 1000 cells/well for A375, BJAB, COLO 205, and KB, and 5000 cells/well for MOLT-4. Dilutions of the AMC or free maytansinoid solutions were added to cells and incubated for 5 days at 37 °C with 6% CO2. The WST-8 agent was added and developed for 1 h for A375, COLO 205, and KB, and 2 h for BJAB and MOLT-4. Surviving fractions of cells were plotted versus conjugate or maytansinoid concentrations for determining IC50 values. Mouse Studies. Female CD-1 mice were obtained from Charles River Laboratory (Raleigh, North Carolina). Mice were intravenously injected with a bolus of 300 μg/kg (based on DM1 or DM4) of the conjugates or 1 mg/kg of free [3H]DM4 or [3H]DM1. Two mice were sacrificed at each of the designated time points following treatment. The gall bladder containing the bile was collected from the mice treated with [3H]DM4 and [3H]DM1 and frozen at -80 °C. After bile collection, the systemic circulation was flushed by injecting 5 mL PBS into the left ventricle and draining through an incision in the inferior vena cava. The whole liver tissues were collected in 50 mL preweighed polypropylene tubes and frozen at -80 °C. Extraction of [3H]Maytansinoid Metabolites from Liver Tissue. The liver tissues were thawed on ice, cut into small pieces, and homogenized with five volumes (v/w) 10 mM TrisHCl buffer, pH 7.5 containing 5 mM N-ethylmaleimide (NEM). The NEM was added to prevent thiol/disulfide exchange reactions during the workup by capping the thiol groups. The homogenates were placed on ice for 2 h to ensure complete alkylation, and then a portion of each homogenate (0.4 mL) was added to a 20 mL glass scintillation vial. The radioactivity associated with each homogenate was determined after solubilization with Solvable as described for tumor homogenates

reduction and S-methylation steps yield hydrophobic maytansinoid derivatives that are potentially able to diffuse throughout the tumor, reaching tumor cells that would otherwise not accumulate sufficient maytansinoid to kill them due to low expression of the target antigen or inaccessibility to the conjugate. These diffusible metabolites may also increase the toxicity of the disulfide-linked conjugates by exposing healthy tissues to free maytansinoids. Significant levels of the catabolites would be expected to form in the liver which is known to be the primary site of antibody catabolism.14,15 Other tissues of the reticuloendothelial system (RES) constitute additional potential catabolic sites for AMCs. Maytansinoid catabolites formed in these other tissues would ultimately be cleared through the liver since excretion studies have found liver to be the major route of AMC clearance (unpublished data). Given this central role of liver in antibody catabolism and clearance of AMCs, we set out to characterize the major maytansinoid catabolites and metabolites of the uncleavable mAb-SMCC-DM1 conjugate and the cleavable mAb-SPPDM1 and mAb-SPDB-DM4 conjugates. Our findings suggest that catabolism and liver metabolism plays an essential role in limiting the potential toxicity of AMCs.

’ EXPERIMENTAL PROCEDURES Reagents. RPMI 1640 and glutamine were from Cambrex Bioscience. Ultima Flo M scintillation fluid and Solvable were from PerkinElmer Life and Analytical Sciences. All cells were obtained from the American Type Culture Collection. All chemicals were obtained from Sigma Aldrich. The Gemini 5 μm C-18 column (0.46  25 cm) was obtained from Phenomenex and the 5 μm C8 column (150  2.1 mm) from Kromasil. The ULTRATURRAX T8 dispersing instrument with a S8N dispersing tool was obtained from IKA Works Inc. All antibody-maytansinoid conjugates were prepared at ImmunoGen from humanized anti-CD56 mAb, huN90116 using methods described elsewhere by Widdison et al.17 The linkers N-succinimidyl 4-(N-maleimidomethyl) cyclohexane-1 carboxylate (SMCC), N-succinimidyl 4-(2-pyridyldithio)butyrate (SPDB), and N-succinimidyl 4-(2-pyridyldithio) pentanoate (SPP), are all N-hydroxysuccinimide esters that react with amino groups of the protein. The ratio of linked maytansinoid molecules per antibody molecule (D/A) and the specific radioactivities for each conjugate was as follows: huN901-SPP-[3H]DM1 (3.5 D/A, 852 mCi/mmol DM1), 729

dx.doi.org/10.1021/bc100498q |Bioconjugate Chem. 2011, 22, 728–735

Bioconjugate Chemistry

ARTICLE

Figure 2. Maytansinoid metabolites of cleavable and uncleavable [3H]AMCs formed in liver over time. Mice (n = 10) were treated with a single 300 μg/ kg dose (based on DM1 or DM4) of mAb-SMCC-[3H]DM1, mAb-SPP-[3H]DM1, or mAb-SPDB-[3H]DM4, and two mice from each group were sacrificed after 2 h, 6 h, 1 d, 4 d, and 7 d and liver tissues samples were collected. Liver tissues were homogenized and portions of the homogenates were extracted and analyzed for 3H-maytansinoid metabolites by HPLC. The effluent from the samples was collected in 0.5 mL fractions, and the radioactivity corresponding to each fraction was determined by LSC. Alternatively, fractions in duplicate runs corresponding to the tritium peaks were analyzed by LC/MS for metabolite identification. The chromatograms show the fraction number on the abscissa and the counts per minute of tritium (CPM) on the ordinate. The sole metabolite (A) of the uncleavable conjugate is lysine-SMCC-DM1. The five metabolites for mAb-SPP-DM1 are lysine-SPP-DM1(A), DM1 (B), S-methyl-DM1 (C), S-methyl-DM1 sulfoxide (D), and S-methyl-DM1 sulfone (E). The corresponding five metabolites of mAb-SPDB-DM4 (lysine-SPDB-DM4, DM4, S-methyl-DM4, S-methyl-DM4 sulfoxide, and S-methyl-DM4 sulfone) are also denoted A-E.

previously.18 [3H]Maytansinoid metabolites were extracted by the procedure described previously with slight modifications.18 Aliquots of the homogenates (0.4 mL) were extracted three times with 3.4 volumes of ethyl acetate/methanol (7.5:1, v/v), and the pooled extracts were evaporated to dryness in an evacuated centrifuge. The dried extracts were dissolved in 0.12 mL of 20% aqueous acetonitrile containing 0.025% trifluoroacetic acid (TFA) and 25 mM catechol (to prevent oxidation), and the metabolites were separated by reversed-phase HPLC. In separate experiments, we found that 83% of the synthetic standard, S-methyl-[3H]DM1, could be recovered from the liver homogenate using this method. In addition, extracts from fresh liver homogenates mixed with mAb-SPDB-[3H]DM4 were found to contain none of the metabolites, indicating that the ex vivo workup released no artifacts (data not shown). The radioactivity associated with the metabolites in Figure 2 was converted to pmol/g of metabolite using the specific radioactivity of the DM1 and DM4 and the measured liver weights. The values were multiplied by 1.2 to account for the extraction efficiency (83%) of the synthetic S-methyl-[3H]DM1.

Extraction of [3H]Maytansinoid Metabolites from Bile. The

gall bladder containing the bile was mixed with 20% aqueous acetonitrile containing 5 mM NEM (0.12 mL), vortexed, and allowed to sit for 30 min to ensure complete alkylation of thiol groups. The debris was removed from the samples by sedimentation (13 000g, 5 min) and the supernatants were analyzed for metabolites by HPLC and liquid scintillation counting. Analytical Methods. All maytansinoids were separated on a Gemini 5 μm C-18 column (0.46  25 cm) equilibrated with 20% aqueous CH3CN containing 0.025% TFA. The chromatogram was developed with a 1 min hold at 20% aqueous CH3CN containing 0.025% TFA, followed by a linear gradient of 2% CH3CN/min for 15 min, and then 1% CH3CN/min for 20 min with a flow rate of 1 mL/min. The effluent was collected in 0.5 mL fractions. Tritium (cpm) associated with each fraction was determined by mixing each vial (0.5 mL) with 4 mL Ultima Gold liquid scintillation cocktail and counting for 5 min in a Tri-Carb 2900T liquid scintillation counter. Alternatively, fractions in duplicate runs corresponding to the peaks of tritium were evaporated to dryness, dissolved in 0.02 mL 50% aqueous CH3CN, and 730

dx.doi.org/10.1021/bc100498q |Bioconjugate Chem. 2011, 22, 728–735

Bioconjugate Chemistry

ARTICLE

analyzed for maytansinoids by liquid chromatography mass spectrometry (LC/MS). The [3H]-maytansinoid metabolites were separated on a 5 μm C-8 column (150  2.1 mm) equilibrated with 15% aqueous CH3CN containing 0.1% TFA. The chromatogram was developed with four linear gradients: 1.1% CH3CN/min for the first 15 min, 1.6% CH3CN/min for 10 min, 1% CH3CN/min for 10 min, and 1.6% CH3CN/min for 20 min with a flow rate of 0.22 mL/min.

’ RESULTS Administration of Antibody-[3H]Maytansinoid Conjugates to Mice. Three different antibody maytansinoid conjugates

(AMCs) with structures similar to those currently undergoing clinical evaluation were prepared for in vivo evaluation (Figure 1). The three conjugates were prepared from a humanized anti-CD56 antibody that does not cross-react with any antigen in mice, using either tritium-labeled DM1 or DM4 to enable the detection of maytansinoid metabolites.18 Mice were administered a single intravenous dose of 300 μg/kg (based on DM1 or DM4) of each 3 H-conjugate, and after 2 h, 6 h, 1 d, 4 d, or 7 d, mice were sacrificed and liver tissues were collected. The liver tissues were homogenized and alkylated with NEM to prevent thiol-disufide exchange reactions, and then analyzed for metabolites following extraction as described in Experimental Procedures. [3H]Maytansinoid Metabolites. Figure 2 shows the metabolites isolated from the livers of mice treated with the three 3Hconjugates. The metabolites were identified by comparing their HPLC retention times and mass spectral fragmentation patterns to synthetic standards as described previously18 (data not shown). The major metabolite of the uncleavable mAb-SMCC-[3H]DM1 in all of the samples was lysine-Nε-SMCC-DM1, suggesting complete proteolysis of the polypeptide backbone of the antibody component to its constituent amino acids. We also recently found this lysine-linker-maytansinoid to be the major metabolite in tumor tissue following administration of the anti-CanAg targeting huC242-SMCC-[3H]DM1 to SCID mice bearing CanAg positive COLO 205 tumors.18 We did not detect any additional liver metabolites, suggesting little if any metabolism of the maytansinoid macrocycle or the thioether linker. The additional minor peak in the 2 h sample was identified as N20 -[3-[[1[[4-(carboxy)cyclohexyl]methyl]-2,5-dioxo-3-pyrrolidinyl]thio]1-oxopropyl]-N20 -deacetyl-maytansine (MCC-DM1, 22.5 min, M þ Na = 997.4)—an unconjugated maytansinoid species present at low levels in the preadministration conjugate preparation (less than 5% of the total maytansinoid in the conjugate sample). As was observed for the uncleavable conjugate, significant levels of the lysine-linker-maytansinoid metabolites were detected in the liver extracts from the mice treated with the disulfide-linked conjugates (peak A, Figure 2) suggesting that the initial steps for AMC metabolism are linker independent. In addition to the lysine-linker-maytansinoids, DM1, DM4 (isolated as the stable NEM-trapped adducts DM1-NEM and DM4-NEM, formed during homogenization, Figure 2, peak B), and S-methyl DM4 (peak C) were also found to be significant metabolites of mAb-SPP-DM1 and mAb-SPDB-DM4, respectively. All of these metabolites had been identified in extracts from cancer cells exposed to the conjugates in previous studies.10,18 Presumably, these liver metabolites are formed by lysosomal degradation, disulfide reduction, and S-methylation of DM4 thiol—steps similar to those described previously for targeted cancer cells.10

Figure 3. Scheme for the in vivo metabolism of cleavable and uncleavable AMCs. Initial catabolism of both cleavable and uncleavable conjugates yields the corresponding lysine-linker-maytansinoid (May) metabolites. For the uncleavable mAb-SMCC-DM1, the lysine-SMCCDM1 is the only observed metabolite. For the cleavable disulfide-linked conjugates mAb-SPDB-DM4 and mAb-SPP-DM1, additional reduction yields DM4 and DM1, respectively. These free maytansinoid thiols are then methylated by an endogenous S-methyl transferase to form Smethyl-DM1 and S-methyl-DM4. Subsequent oxidation steps yield the corresponding sulfoxide and sulfone derivatives.

Importantly, several additional liver metabolites of the disulfide-linked conjugates were identified that were not observed in cancer cells from our prior studies. We observed S-methyl-DM1 from the livers of the mice treated with mAb-SPP-DM1 (peak C), which was unexpected because it was not identified as a tumor metabolite of mAb-SPP-DM1 in our previous study.18 These observations suggest that, while the efficiency of DM1 Smethylation by several carcinoma cells was found to be very poor, and much lower than for DM4,18 the liver has the capacity to more readily S-methylate both maytansinoid thiol compounds relative to tumor cells. The oxidized sulfoxide and sulfone derivatives of the S-methyl-DM1 or S-methyl-DM4 metabolites were also observed in the liver extracts of mice treated with the two disulfide-linked AMCs (peaks D and E in Figure 2). Ex vivo oxidation of both S-methyl-DM1 and S-methyl-DM4 to the corresponding sulfoxide and sulfone derivatives by human liver microsomes required NADPH, indicating that the oxidation was enzyme mediated (data not shown). The metabolites of the three conjugates isolated from the liver tissues suggest the scheme for the metabolism of AMCs displayed in Figure 3 where the initial step for the catabolism of both cleavable and uncleavable AMCs is degradation of the intact conjugate to the corresponding lysine-linker-maytansinoid. For the uncleavable mAb-SMCC-DM1, no further metabolism is observed. For the cleavable conjugates, additional reduction yields the maytansinoid thiols allowing for the additional Smethylation and oxidation steps that ultimately yield the Smethyl sulfoxide and S-methyl-sulfone metabolites. Exposure of Liver Tissue to Maytansinoid Metabolites. The concentration of the total metabolites of the conjugates in the liver tissue at each time (pmol/g) was determined from the 731

dx.doi.org/10.1021/bc100498q |Bioconjugate Chem. 2011, 22, 728–735

Bioconjugate Chemistry

ARTICLE

radioactivity associated with the metabolites (Figure 2) and plotted in Figure 4A. The AUCs for the metabolites of the three conjugates over 7 d were very similar, suggesting that the exposure of the liver tissue to the maytansinoid metabolites was linker independent. The fraction of the metabolite levels in Figure 4 as a percentage of total maytansinoids in the liver (intact conjugate þ metabolites) was over 50% indicating that the observed metabolites represent the majority of the maytansinoids in the liver (Supporting Information Figure S1). While the exposure levels were similar for the three conjugates, the chemical structures of the metabolites of the three conjugates were notably different with the charged lysine-SMCC-DM1 as the predominant metabolite of the uncleavable conjugate and mixture of the lysine-linker-metabolite and the additional four metabolites described in Figure 3 comprising the metabolites of the disulfide linked conjugates. Differences in the distribution of the metabolites of the two disulfide-linked conjugates are evident in the radiograms shown in Figure 2. To more easily compare the metabolite distributions of the two disulfide-linked conjugates, the percentages of the total recovered metabolites of each of the five observed metabolites (calculated from Figure 2) were plotted at each time point in Figure 4B. Clear differences in the percentages for the five metabolites of the two conjugates were evident at each time point. The metabolite representing the largest fraction of the mAb-SPP-DM1 conjugate at 2 h was DM1 (34%) followed by lysine-SPP-DM1 (33%), S-methyl-DM1 (12%), S-methyl-DM1 sulfoxide (12%), and S-methyl-DM1 sulfone (9%). In contrast, lysine-SPDB-DM4 constituted the largest fraction of the metabolites of the mAb-SPDB-DM4 conjugate at 2 h (56%) followed by DM4 (24%), S-methylDM4 (10%), S-methyl-DM4 sulfoxide (7%), and S-methyl-DM4 sulfone (3%). The percentages for these five metabolites all change markedly over the first 24 h with decreases of the lysineSPDB-DM4, DM4, and lysine-SPP-DM1 metabolites and increases for the sulfoxide and sulfone derivatives. From 24 h to 7 d, the metabolite percentages for all metabolites remain relatively constant even though the total metabolite levels are decreasing. The relative levels of lysine-SPP-DM1 and lysine-SPDB-DM4 versus DM1 and DM4, respectively, at the early time points (224 h) likely reflect the relative reactivity of the respective disulfide links to intracellular thiol-disulfide exchange reactions. Furthermore, DM1 levels may remain higher relative to DM4 levels throughout the time course (Figure 4B) if the rate of S-methylation of DM1 by liver cells is slower than for DM4, as described previously in human carcinoma cells.18 In Vitro Cytotoxic Potencies Associated with the Liver Metabolites. The metabolites reported here were all synthesized and the purified compounds were assayed for their cytotoxic potencies on a panel of tumor cell lines (Table 1). The S-methylDM1 and S-methyl-DM4 maytansinoids had similar potency to maytansine, and were about 1000-fold more cytotoxic than the lysine-linker-maytansinoids. Values for DM1 and DM4 are not presented because their IC50 values were highly variable when measured in vitro, likely due to the rapid formation of mixed disulfides via thiol/disulfide exchange reactions with cystine present at 0.2 mM in the culture medium (disulfides in serum proteins may also contribute to such reactions). The IC50 values for the S-methyl-DM1 sulfoxide were found to range from 7.1 nM to 23 nM—200- to 500-fold less potent than the S-methyl DM1 metabolite, whereas the S-methyl-DM4 sulfoxide was found to be 12- to 20-fold less cytotoxic than the S-methylDM4. The IC50 values for the S-methyl-DM1 sulfone and

Figure 4. Exposure of liver tissue to maytansinoid metabolites in mice. (A) The levels for the total maytansinoid metabolites in liver tissue at 2 h, 6 h, 1 d, 4 d, and 7 d following a single intravenous administration of 300 μg/kg (DM1 or DM4 dose) of mAb-SMCC-[3H]DM1 (9), mAbSPP-[3H]DM1 (b), and mAb-SPDB-[3H]DM4 (2). The radioactivities associated with the metabolites in Figure 2 were converted to pmol/ g of metabolite using the specific radioactivities of the DM1 and DM4 and the measured liver weights as described in Experimental Procedures. (B) Amounts of the five observed metabolites of the disulfide-linked conjugates at each time as percentages of total metabolite levels. The percentages for each of the five metabolites of the two disulfide-linked conjugates were calculated by dividing the amount of each metabolite (pmol/g) by the sum of the five metabolites (shown in A). The five metabolites for mAb-SPP-DM1 are lysine-SPP-DM1(A), DM1 (B), Smethyl-DM1 (C), S-methyl-DM1 sulfoxide (D), and S-methyl-DM1 sulfone (E). The corresponding five metabolites of mAb-SPDB-DM4 (lysine-SPDB-DM4, DM4, S-methyl-DM4, S-methyl-DM4 sulfoxide, and S-methyl-DM4 sulfone) are also denoted A-E.

S-methyl-DM4 sulfone metabolites were 2- to 7-fold more potent than their sulfoxide precursors. That the S-methyl-DM4 sulfoxide is more cytotoxic than the S-methyl-DM1 sulfoxide toward the carcinoma cell lines is perhaps a consequence of the three extra carbons of DM4 (compared to DM1) that increase its hydrophobicity and may thus increase cell permeability. Nonetheless, the oxidation of the maytansinoid metabolites in the liver clearly serves to reduce the cytotoxic potencies of the metabolites of both disulfide-linked conjugates. 732

dx.doi.org/10.1021/bc100498q |Bioconjugate Chem. 2011, 22, 728–735

Bioconjugate Chemistry

ARTICLE

Table 1. Cytotoxic Potencies of Maytansinoids Towards Human Carcinoma Cell Lines in Vitroa in vitro IC50 (nM) metabolites Maytansine

b

COLO205 (colon)

0.033

0.081

KB (cervix) 0.047

MOLT-4 (T-cell) 0.09

9.7

lysine-Nε-SPDB-DM4b

1.8

6.6

17

3.3

16

lysine-Nε-SMCC-DM1b

8.1

7.5

17

9.7

16

S-methyl-DM1b

0.020

0.012

0.033

0.022

S-methyl-DM4b

0.017

0.014

0.031

0.026

S-methyl-DM1 Sulfoxide

9.7

7.1

S-methyl-DM1 Sulfone S-methyl-DM4 Sulfoxide

0.55

1.7 0.55

DM1-MCC b

0.045

BJAB (B-cell)

lysine-Nε-SPP-DM1b

S-methyl-DM4 Sulfone a

A375 (melanoma)

0.075 22

50

61

17 5.9 1.3

0.17

0.63

21

50

10

23 3.5 0.12 0.80 29

>100

0.030 0.027 19 1.9 1.1 50

All metabolites were chemically synthesized. Cell-killing activities of the metabolites were measured after 5 days using a WST-based cell viability assay. IC50 values were established previously18 and shown here for comparison to liver-generated maytansinoid metabolites.

Metabolism of Intravenously Administered [3H]DM4 and [ H]DM1. Significant amounts of DM1, S-methyl-DM1, DM4,

SMCC-DM1 was previously shown to have low cytotoxic potency toward human carcinoma cells in cell-based viability assays;18 the low potency is most likely due to its inability to penetrate the cell due to the charged groups of the lysine residue. The lysine-SMCC-DM1 was found to be highly cytotoxic to the cancer cells in which it was formed following internalization and lysosomal degradation of the mAb-SMCC-DM1.10 The lysine-SMCC-DM1 metabolite released by intracellular processing in the liver tissue could also be potentially cytotoxic. However, preclinical studies have shown the uncleavable T-DM1 conjugate to be well-tolerated at high doses (50 mg/kg) in rats,11 and the uncleavable anti-CD79b conjugate, 10D10-SMCCDM1, was found to be well-tolerated when administered to cynomologus monkeys at ∼30 mg/kg for two doses three weeks apart, although mild and reversible elevation of liver enzymes and minimal degeneration of hepatocytes in the monkeys was noted.25 Furthermore, the dose-limiting toxicity observed in patients treated with T-DM1 was thrombocytopenia—no adverse liver toxicity was observed.4 These observations suggest that the exposure of liver tissue to the lysine-SMCC-DM1 metabolite of the mAb-SMCCDM1 conjugate reported here does not induce clinically significant liver toxicity at doses expected to be efficacious.4,11 One factor that may contribute to the ability of liver tissue to generate and eliminate lysine-SMCC-DM1 from mAb-SMCC-DM1 conjugates without undue toxicity is that maytansinoids, as tubulin-binding agents,26 are antimitotic and mature, fully differentiated, nondividing cells of organs such as liver may be able to tolerate significant exposure. Analysis of the metabolites in liver tissues from mice treated with the cleavable disulfide-linked conjugates suggested the pathway for their degradation and metabolism shown in Figure 3. As was observed for the uncleavable conjugate, the two disulfide-linked conjugates were first degraded to yield the lysine-linker-maytansinoid metabolites. The amounts of these lysine-linker-maytansinoid metabolites were highest at the earliest 2 h time points. Unlike the lysine-SMCC-DM1 metabolite of the uncleavable conjugate, the lysine-SPDB-DM4 and lysine-SPPDM1 metabolites of the disulfide-linked conjugates were found to undergo further intracellular transformations. The initial steps are similar to those described for the activation of the two conjugates in cancer cells.18 First, the lysine-SPDB-DM4 and lysine-SPP-DM1 metabolites are reduced to DM4 and DM1, respectively, likely due to thiol-disulfide exchange reactions. Next, S-methylation of the released maytansinoid thiols yields S-methyl-DM1 and S-methyl-

3

and S-methyl-DM4 metabolites were observed in the liver tissues up to 7 d after administration of the conjugates. If cleared from the liver without first being oxidized, these highly potent metabolites would likely be toxic to healthy tissue exposed to the metabolites during elimination. To test the efficiency of the liver oxidation, CD-1 mice were given a single dose of 1 mg/kg of unconjugated [3H]DM1 or [3H]DM4 (near the maximum tolerated dose), and after 1 h, mice were sacrificed and analyzed for maytansinoid metabolites in the liver tissue, bile, and plasma (Figure 5). Analysis of plasma indicated that the administered DM4 was rapidly S-methylated, possibly by the same S-methyltransferases in red blood cells shown to methylate the thiols of exogenous molecules such as penicillamine.19 Conversely, little S-methyl DM1 was observed in the plasma of the mice treated with DM1—consistent with its slow rate of S-methylation in cancer cells. For the mice treated with DM4, nearly equivalent amounts of S-methyl DM4, S-methyl-DM4 sulfoxide, and Smethyl-DM4 sulfone were observed in the liver tissue after 1 h. The corresponding S-methyl DM1, S-methyl-DM1 sulfoxide, and S-methyl-DM1 sulfone metabolites were observed in the liver tissues of mice treated with DM1. In addition, significant levels of DM1 were observed in the liver tissues and bile—again suggesting a lower rate for the S-methylation of DM1 as compared to DM4. Although significant levels of the highly cytotoxic S-methyl DM1 and S-methyl-DM4 metaboiltes were observed in the liver tissues of the mice, little of the S-methyl maytansinoids were observed in the bile samples with nearly all of the metabolites being the Smethyl sulfoxide and S-methyl-sulfone metabolites, suggesting efficient oxidation steps prior to billiary elimination.

’ DISCUSSION Our results suggest that liver plays an important role in the detoxification of both cleavable and uncleavable antibody-maytansinoid conjugates (AMCs). The lysine-SMCC-DM1 was the only metabolite observed for the noncleavable mAb-SMCC-DM1 conjugate. Significant levels were observed in the liver just 2 h following administration of the conjugates, consistent with other in vivo mouse studies that have reported efficient liver catabolism of radiolabeled antibodies20,21 and antibody-auristatin conjugates.22-24 The lysine-SMCC-DM1 metabolite of the uncleavable mAb733

dx.doi.org/10.1021/bc100498q |Bioconjugate Chem. 2011, 22, 728–735

Bioconjugate Chemistry

ARTICLE

Figure 5. Maytansinoid metabolites in plasma, liver, and bile 1 h following a single bolus intravenous administration of [3H]DM4 and [3H]DM1. Mice (n = 2) were given a single bolus intravenous administration of 1 mg/kg [3H]DM4 or [3H]DM4, and after 1 h, mice were sacrificed and the plasma, bile, and liver tissues were extracted for metabolites. Extracts were analyzed by HPLC as described in Figure 2 for the disulfide-linked conjugates. The chromatograms show the fraction number on the abscissa and the counts per minute of tritium (CPM) on the ordinate.

DM4. These diffusible and highly cytotoxic maytansinoids are then efficiently oxidized in the liver to polar S-methyl-maytansinoid sulfoxide and sulfone derivatives that have reduced cytotoxic potencies most likely due to low cell permeability. Interestingly, we observed no metabolism of the macrocyclic core of the maytansinoid molecule in any of the liver extracts. In a separate in vitro study, incubation of maytansine with purified human liver microsomes was reported to yield minor amounts of O-demethylation on the macrocyclic portion of maytansine.27 However, the authors noted that the observed O-demethylation may have arisen from the low purity of the maytansine used in the experiment. We repeated the in vitro experiments using purified maytansine and human liver microsomes and could detect no metabolism on the macrocycle of maytansine (unpublished data), suggesting that the observed demethylation in the prior study27 was presumably due to impurities in the maytansine. The resistance of maytansine to metabolism would explain why the most common and dose-limiting toxicity in patients treated with maytansine was gastrointestinal toxicity.28 In contrast, gastrointestinal toxicity has not been found to be dose-limiting in patients treated with AMCs. While the reason for this difference may stem from the slower clearance rate for the antibody-drug conjugates as compared to the small molecule maytansine, the detoxification of the conjugates via the metabolism described here may also reduce their gastrointestinal toxicities.

This material is available free of charge via the Internet at http:// pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Correspondence to Hans Erickson, ImmunoGen, Inc., 830 Winter Street, Waltham, MA 02451. Tel: 781-895-0734. Fax: 781-895-0611. E-mail: [email protected].

’ ACKNOWLEDGMENT We thank John Lambert for his many helpful suggestions and careful reading of the manuscript. ’ REFERENCES (1) Senter, P. D. (2009) Potent antibody drug conjugates for cancer therapy. Curr. Opin. Chem. Biol. 13, 235–44. (2) Polakis, P. (2005) Arming antibodies for cancer therapy. Curr. Opin. Pharmacol. 5, 382–7. (3) Lambert, J. M. (2005) Drug-conjugated monoclonal antibodies for the treatment of cancer. Curr. Opin. Pharmacol. 5, 543–9. (4) Krop, I. E., Beeram, M., Modi, S., Jones, S. F., Holden, S. N., Yu, W., Girish, S., Tibbitts, J., Yi, J. H., Sliwkowski, M. X., Jacobson, F., Lutzker, S. G., and Burris, H. A. (2010) Phase I study of trastuzumab-DM1, an HER2 antibody-drug conjugate, given every 3 weeks to patients with HER2-positive metastatic breast cancer. J. Clin. Oncol. 28, 2698-2704. (5) Smith, S. V. (2005) Technology evaluation: huN901-DM1, ImmunoGen. Curr. Opin. Mol. Ther. 7, 394–401.

’ ASSOCIATED CONTENT

bS

Supporting Information. An assessment of the radioactivity associated with liver tissues (mass balance) is provided. 734

dx.doi.org/10.1021/bc100498q |Bioconjugate Chem. 2011, 22, 728–735

Bioconjugate Chemistry

ARTICLE

(22) Alley, S. C., Benjamin, D. R., Jeffrey, S. C., Okeley, N. M., Meyer, D. L., Sanderson, R. J., and Senter, P. D. (2008) Contribution of linker stability to the activities of anticancer immunoconjugates. Bioconjugate Chem. 19, 759–65. (23) Alley, S. C., Zhang, X., Okeley, N. M., Anderson, M., Law, C. L., Senter, P. D., and Benjamin, D. R. (2009) The pharmacologic basis for antibody-auristatin conjugate activity. J. Pharmacol. Exp. Ther. 330, 932–8. (24) Kim, K. M., McDonagh, C. F., Westendorf, L., Brown, L. L., Sussman, D., Feist, T., Lyon, R., Alley, S. C., Okeley, N. M., Zhang, X., Thompson, M. C., Stone, I., Gerber, H. P., and Carter, P. J. (2008) AntiCD30 diabody-drug conjugates with potent antitumor activity. Mol. Cancer Ther. 7, 2486–97. (25) Zheng, B., Fuji, R. N., Elkins, K., Yu, S. F., Fuh, F. K., Chuh, J., Tan, C., Hongo, J. A., Raab, H., Kozak, K. R., Williams, M., McDorman, E., Eaton, D., Ebens, A., and Polson, A. G. (2009) In vivo effects of targeting CD79b with antibodies and antibody-drug conjugates. Mol. Cancer Ther. 8, 2937–46. (26) Gupta, S., and Bhattacharyya, B. (2003) Antimicrotubular drugs binding to vinca domain of tubulin. Mol. Cell. Biochem. 253, 41–7. (27) Liu, Z., Floss, H. G., Cassady, J. M., and Chan, K. K. (2005) Metabolism studies of the anti-tumor agent maytansine and its analog ansamitocin P-3 using liquid chromatography/tandem mass spectrometry. J. Mass Spectrom. 40, 389–99. (28) Issell, B. F., and Crooke, S. T. (1978) Maytansine. Cancer Treat. Rev. 5, 199–207.

(6) Al-Katib, A. M., Aboukameel, A., Mohammad, R., Bissery, M. C., and Zuany-Amorim, C. (2009) Superior antitumor activity of SAR3419 to rituximab in xenograft models for non-Hodgkin’s lymphoma. Clin. Cancer Res. 15, 4038–45. (7) Chen, Q., Millar, H. J., McCabe, F. L., Manning, C. D., Steeves, R., Lai, K., Kellogg, B., Lutz, R. J., Trikha, M., Nakada, M. T., and Anderson, G. M. (2007) Alphav integrin-targeted immunoconjugates regress established human tumors in xenograft models. Clin. Cancer Res. 13, 3689–95. (8) Ikeda, H., Hideshima, T., Fulciniti, M., Lutz, R. J., Yasui, H., Okawa, Y., Kiziltepe, T., Vallet, S., Pozzi, S., Santo, L., Perrone, G., Tai, Y. T., Cirstea, D., Raje, N. S., Uherek, C., Dalken, B., Aigner, S., Osterroth, F., Munshi, N., Richardson, P., and Anderson, K. C. (2009) The monoclonal antibody nBT062 conjugated to cytotoxic Maytansinoids has selective cytotoxicity against CD138-positive multiple myeloma cells in vitro and in vivo. Clin. Cancer Res. 15, 4028–37. (9) Lambert, J. M. (2010) Antibody-maytansinoid conjugates; a new strategy for the treatment of cancer. Drugs Future. (10) Erickson, H. K., Park, P. U., Widdison, W. C., Kovtun, Y. V., Garrett, L. M., Hoffman, K., Lutz, R. J., Goldmacher, V. S., and Blattler, W. A. (2006) Antibody-maytansinoid conjugates are activated in targeted cancer cells by lysosomal degradation and linker-dependent intracellular processing. Cancer Res. 66, 4426–33. (11) Lewis Phillips, G. D., Li, G., Dugger, D. L., Crocker, L. M., Parsons, K. L., Mai, E., Blattler, W. A., Lambert, J. M., Chari, R. V., Lutz, R. J., Wong, W. L., Jacobson, F. S., Koeppen, H., Schwall, R. H., KenkareMitra, S. R., Spencer, S. D., and Sliwkowski, M. X. (2008) Targeting HER2-positive breast cancer with trastuzumab-DM1, an antibodycytotoxic drug conjugate. Cancer Res. 68, 9280–90. (12) Kovtun, Y. V., Audette, C. A., Ye, Y., Xie, H., Ruberti, M. F., Phinney, S. J., Leece, B. A., Chittenden, T., Blattler, W. A., and Goldmacher, V. S. (2006) Antibody-drug conjugates designed to eradicate tumors with homogeneous and heterogeneous expression of the target antigen. Cancer Res 66, 3214–21. (13) Kovtun, Y. V., and Goldmacher, V. S. (2007) Cell killing by antibody-drug conjugates. Cancer Lett 255, 232–40. (14) Henderson, L. A., Baynes, J. W., and Thorpe, S. R. (1982) Identification of the sites of IgG catabolism in the rat. . Arch Biochem Biophys 215, 1–11. (15) Wright, A., Sato, Y., Okada, T., Chang, K., Endo, T., and Morrison, S. (2000) In vivo trafficking and catabolism of IgG1 antibodies with Fc associated carbohydrates of differing structure. Glycobiology 10, 1347–55. (16) Roguska, M. A., Pedersen, J. T., Henry, A. H., Searle, S. M., Roja, C. M., Avery, B., Hoffee, M., Cook, S., Lambert, J. M., Blattler, W. A., Rees, A. R., and Guild, B. C. (1996) A comparison of two murine monoclonal antibodies humanized by CDR-grafting and variable domain resurfacing. Protein Eng. 9, 895–904. (17) Widdison, W. C., Wilhelm, S. D., Cavanagh, E. E., Whiteman, K. R., Leece, B. A., Kovtun, Y., Goldmacher, V. S., Xie, H., Steeves, R. M., Lutz, R. J., Zhao, R., Wang, L., Blattler, W. A., and Chari, R. V. (2006) Semisynthetic maytansine analogues for the targeted treatment of cancer. J. Med. Chem. 49, 4392–408. (18) Erickson, H. K., Widdison, W. C., Mayo, M. F., Whiteman, K., Audette, C., Wilhelm, S. D., and Singh, R. (2010) Tumor delivery and in vivo processing of disulfide-linked and thioether-linked antibodymaytansinoid conjugates. Bioconjugate Chem. 21, 84-92. (19) Keith, R. A., Otterness, D. M., Kerremans, A. L., and Weinshilboum, R. M. (1985) S-Methylation of D- and L-penicillamine by human erythrocyte membrane thiol methyltransferase. Drug Metab. Dispos. 13, 669–76. (20) Franano, F. N., Edwards, W. B., Welch, M. J., and Duncan, J. R. (1994) Metabolism of receptor targeted 111In-DTPA-glycoproteins: identification of 111In-DTPA-epsilon-lysine as the primary metabolic and excretory product. Nucl. Med. Biol. 21, 1023–34. (21) Novak-Hofer, I., Zimmermann, K., Maecke, H. R., Amstutz, H. P., Carrel, F., and Schubiger, P. A. (1997) Tumor uptake and metabolism of copper-67-labeled monoclonal antibody chCE7 in nude mice bearing neuroblastoma xenografts. J. Nucl. Med. 38, 536–44. 735

dx.doi.org/10.1021/bc100498q |Bioconjugate Chem. 2011, 22, 728–735