Identification and Activities of Human Carboxylesterases for the

Seattle Genetics, Inc. 21823 30th Drive SE, Bothell, Washington 98021. Received August 1, 2001;. Revised Manuscript Received September 26, 2001. CPT-1...
0 downloads 0 Views 109KB Size
Download PDF
1074

Bioconjugate Chem. 2001, 12, 1074−1080

Identification and Activities of Human Carboxylesterases for the Activation of CPT-11, a Clinically Approved Anticancer Drug Peter D. Senter,* Kevin S. Beam, Bruce Mixan, and Alan F. Wahl Seattle Genetics, Inc. 21823 30th Drive SE, Bothell, Washington 98021. Received August 1, 2001; Revised Manuscript Received September 26, 2001

CPT-11 is a clinically approved anticancer drug used for the treatment of advanced colorectal cancer. Upon administration, the carbamate side chain of the drug is hydrolyzed, resulting in the release of SN-38, an agent that has approximately 1000-fold increased cytotoxic activity. Since only a very small percentage of the injected dose of CPT-11 is converted to SN-38, there is a significant opportunity to improve its therapeutic efficacy and to diminish its systemic toxicity by selectively activating the drug within tumor sites. We envisioned that a mAb-human enzyme conjugate for CPT-11 activation would be of interest, particularly since the conjugate would likely be minimally immunogenic, and the prodrug is clinically approved. Toward this end, it was necessary to identify the most active human enzyme that could convert CPT-11 to SN-38. We isolated enzymes from human liver microsomes based on their abilities to effect the conversion and identified human carboxylesterase 2 (hCE-2) as having the greatest specific activity. hCE-2 was 26-fold more active than human carboxylesterase 1 and was 65% as active as rabbit liver carboxylesterase, the most active CPT-11 hydrolyzing enzyme known. The anti-p97 mAb 96.5 was linked to hCE-2, forming a conjugate that could bind to antigen-positive cancer cells and convert CPT-11 to SN-38. Cytotoxicity assays established that the conjugate led to the generation of active drug, but the kinetics of prodrug activation (48 pmol min-1 mg-1) was insufficient for immunologically specific prodrug activation. These results confirm the importance of hCE-2 for CPT-11 activation and underscore the importance of enzyme kinetics for selective prodrug activation.

INTRODUCTION

The anticancer drug CPT-11 (irinotecan) is a watersoluble derivative of camptothecin and is clinically approved for the treatment of colorectal carcinoma (reviewed in refs 1-5). As with other camptothecin family members, CPT-11 inhibits DNA topoisomerase 1, an enzyme involved in relaxation of supercoiled DNA. This leads to the generation of DNA single strand breaks and eventual cytotoxic activity. Although CPT-11 has demonstrated significant antitumor activities in the clinic, the drug has several serious side effects including gastrointestinal toxicity and neutropenia (1-5). There is a great deal of interest in understanding the biochemical basis for the dose-limiting toxicities of CPT-11 and to find improved forms and delivery mechanisms for the drug. One of the distinguishing features of CPT-11 is that it is metabolized to SN-38 (Figure 1), an agent that is more potent in assays for topoisomerase 1 inhibition and in cytotoxicity activity on transformed cell lines (6, 7). Early on, it was found that carboxylesterases effected the conversion of CPT-11 to SN-38 and that enzymes from several species carried out the reaction (8, 9). In these studies, the most active enzyme was from guinea pig livers and had very low specific activity (4.5 nmol min-1 mg-1). The human enzyme used for comparison had only 3% of this activity. It is noteworthy that carbamates in general and CPT-11 in particular can act as serine carboxylesterase inhibitors, due to the stability of the active site serine-carbamate intermediates formed (10). * To whom correspondence should be addressed. E-mail: [email protected]; phone: (425) 527-4710; FAX: (425) 5274109.

Consequently, even though carboxylesterases are widespread throughout the body (11-14), only 2-5% of the injected dose of CPT-11 in people is converted to SN-38. Most of the drug is excreted through the liver and kidneys without modification of the carbamate bond (15). If SN-38 could be intratumorally generated from CPT11 in an efficient manner, it might be possible to improve CPT-11 therapeutic efficacy and at the same time minimize the severe toxicities associated with this drug. Several methods can be envisioned to enhance intratumoral CPT-11 activation. One approach that has recently been explored involves gene therapy (15) in which the gene encoding an enzyme that can activate CPT-11 is transfected into the target cancer cell population. Candidate genes include those corresponding to rabbit liver carboxylesterase (16, 17), human and horse butyrylcarboxylesterase (18), and two forms of human carboxylesterase known as hCE-1 (16, 19, 20) and hCE-2 (21), all of which convert CPT-11 to SN-38. Of these enzymes, rabbit liver carboxylesterase (rCE) was regarded as being the most effective, since it catalyzed CPT11 carbamate hydrolysis 100-1000-fold more efficiently than hCE-1 (16, 17). Human rhabdomyosarcoma cells transfected with the gene for rCE were sensitized to CPT11 to a greater extent than were nontransfected cells or cells transfected with the gene encoding hCE-1. These results were confirmed in vivo in mice with subcutaneous tumor implants (16). Other studies have shown that some cell lines transfected with hCE-1 were modestly more sensitive to CPT-11 than control cells (19, 20). Another method to enhance the activation of CPT-11 within tumors would be to utilize a prodrug activation strategy known commonly as ADEPT, antibody directed

10.1021/bc0155420 CCC: $20.00 © 2001 American Chemical Society Published on Web 10/23/2001

Carboxylesterases for the Activation of CPT-11

Bioconjugate Chem., Vol. 12, No. 6, 2001 1075

Figure 1. Structure of the prodrug CPT-11 and its conversion to the active metabolite SN-38.

enzyme prodrug therapy (reviewed in refs 22-23). This is a two-step targeting strategy, involving the administration of a monoclonal antibody (mAb)-enzyme conjugate that binds to tumor associated antigens, followed at an appropriate time by a prodrug that is activated by the targeted enzyme. For the purposes of CPT-11, the enzyme would be selected for its ability to efficiently hydrolyze the carbamate bond present within the drug and to generate cytotoxic levels of SN-38 intratumorally. Such an enzyme/prodrug combination would be distinguished from others that have been reported, since the human enzyme would likely be minimally immunogenic, and the prodrug for activation is already clinically approved. Toward developing such a system, we noted that a detailed search for suitable human enzymes has not been published. Here, we report the purification and characterization of CPT-11 activating enzymes from human liver along with their applications for specific prodrug activation. EXPERIMENTAL METHODS

Materials and General Methods. Pharmaceutical grade CPT-11 was obtained as a pure preparation (by HPLC as indicated below) at 20 mg/mL. SN-38 was a gift from Dr. Ruiwen Zhang at the University of Alabama in Birmingham. The mAb 96.5 is of the IgG2a isotype and binds to the p97 antigen, which is expressed on most melanoma cells (24). The 3677 human melanoma (24) and L2987 human lung adenocarcinoma (25) cell lines were established from surgically resected tumors. A pathogenfree frozen human liver was obtained from the Northwest Tissue Center, Seattle, WA, and was stored at -80 °C. Protein assays were conducted using the Pierce Micro BCA assay. The purification of guinea pig liver microsome esterase has been previously described (26). Rabbit liver carboxylesterase was purchased from Sigma and purified by size exclusion chromatography to a high degree of homogeneity as evidenced by both SDS-PAGE and isoelectric focusing. Enzyme Assays. With p-Nitrophenylacetate as Substrate. Protein samples were added to 1 mL of a 0.1 mM solution of p-nitrophenylacetate in phosphate buffered saline pH 7.2 (PBS), and the change in absorbance at 405 nm was recorded for 1 min. The protein samples were diluted to ensure that the measurements were linear during this time period. A change in absorbance of 1 OD represents the generation of 90 nmol of p-nitrophenol. With CPT-11 as Substrate. Protein samples were added to a solution of CPT-11 in water so that the final CPT11 concentration was 8 µM. Incubation was carried out at 37 °C for 15-48 h, at which point HCl (final concentration 60 mM) was added to reform the lactone ring. Analysis by HPLC was performed as follows: injected 0.1 mL onto a C-18 column (4.6 × 75 mm) using a mobile phase of 60 mM ammonium carbonate, 4 mM tetrabutylammonium phosphate, 24% acetonitrile (pH 5.0) at 1 mL/min, and a detector set at 365 nm. The amount of

CPT-11 consumed and SN-38 generated was quantified using standard solutions of these agents that were treated identically to the test samples. The conversion of CPT-11 to SN-38 was linear over this time period. Enzyme Purification. Step 1. Microsome Purification. A 300 g portion of a human liver was thawed and blended in a homogenizer with 500 mL of 1.5% KCl. The mixture was centrifuged at 8700g for 4 h at 4 °C, and the clear supernatant was further centrifuged at 105000g for 1 h at 4 °C. Step 2. Ammonium Sulfate Precipitation. The microsome pellet was suspended in 100 mL of tris[hydroxymethyl]aminomethane (Tris) at pH 8.0, 1.5 g of saponin was added, and the mixture was gently rotated for 1 h at 4 °C. Saturated aqueous ammonium sulfate was slowly added to a final concentration of 30%, and the precipitated material was removed by centrifugation at 11300g for 30 min at 4 °C. To the supernatant was slowly added solid ammonium sulfate to a final concentration of 70%, and the precipitated material was removed by centrifugation at 11300g for 30 min at 4 °C. The pellet was resuspended in and dialyzed against 10 mM Tris pH 8.0 overnight. Step 3. Anion Exchange Chromatography. The dialysate from step 2 was applied to a 46 mL Q Sepharose Fast Flow column (Amersham Pharmacia), and the column was washed with 10 mM Tris pH 8.0 buffer until there was no absorbance at 280 nm. Bound proteins were eluted with a gradient of 10 mM Tris pH 8.0 to 10 mM Tris, 500 mM NaCl, pH 8.0, over 3 h at a flow rate of 5 mL/min. The fractions (10 mL each) were monitored for both p-nitrophenylacetate and CPT-11 hydrolysis activities. Significant activity was obtained in the unbound material, and in fractions 52-62. Step 4. Hydrophobic Interaction Chromatography. The unbound material and fractions 52-62 from step 3 were separately diluted with 1.5 vol of 2 M ammonium sulfate in PBS, applied to a 1.5 × 12 cm phenylsepharose column (Amersham Pharmacia), and the column was washed until there was no absorbance at 280 nm in the eluant. The bound material was eluted with a 400 mL linear gradient from 0.8-0 M ammonium sulfate in PBS. Fractions that contained enzymatic activity were pooled and concentrated to 2 mL by ultrafiltration. Step 5. Size Exclusion Chromatography. The material from step 4 was applied to a 2 × 65 cm Superdex 200 PS column (Amersham Pharmacia) and eluted with PBS. Fractions (3 mL each) were monitored for enzyme activity and purity by SDS-PAGE. Sequence Analysis. The purified proteins (0.15-0.3 mg) were dialyzed against 10 mM Tris pH 8.0, concentrated to dryness on a vacuum centrifuge, and were dissolved in 0.03-0.1 mL of cyanogen bromide (Aldrich) in 70% aqueous formic acid (Fisher, sequencing grade). After 24 h at room temperature, 1.5 mL of water was added, and volatile material was evaporated in a vacuum centrifuge. This was repeated twice with 0.3 mL of water

1076 Bioconjugate Chem., Vol. 12, No. 6, 2001

Senter et al.

Table 1. Purification Scheme for HCE-2 from Human Liver Microsomes Specific Activity specific activity step

total protein (mg)

p-nitrophenylacetate (µmol min-1 mg-1)

CPT-11 (pmol min-1 mg-1

total activitya (pmol/min)

fold purification

ammonium sulfate anion exchange hydrophobic interaction size exclusion

786 94 4 1.4

1.6 3.2 53 60

0.9 4.3 58 92

707 404 232 129

4.8 64 102

a

Based on CPT-11 as the substrate.

to remove traces of acid. The residual material was dissolved in PBS (0.1 mL), applied to a 10% NuPage gel (Novex), and the gel was run using Novex MES running buffer. The gels were blotted onto PVDF filters, which were then stained with Coomassie Blue. The bands were cut out and sequenced by Sandy Kielland at the Protein Microsequencing Laboratory, University of Victoria, Victoria, BC, Canada. Preparation of 96.5-hCE-2. A solution of the 96.5 mAb (0.29 mL at 7.0 mg/mL in PBS) was diluted with 0.05 mL of 0.5 M borate containing 0.5 M NaCl at pH 8.0, and 9 µL of N-(γ-maleimidobutyryloxy)succinimide ester (GMBS, Pierce) at 10 mM in DMSO was added. After 30 min at 37 °C, the solution was applied to a Sephadex G-25 PD 10 column (Amersham Pharmacia), and the modified protein was eluted in a 1.8 mL volume with 100 mM sodium phosphate containing 0.5 M NaCl at pH 7.2. A total of 1.77 mg of protein was obtained, based on the absorbance at 280 nm (E0.1% ) 1.4). The maleimide content was determined by diluting 0.3 mL of the protein with 0.6 mL of PBS. To this was added 10 µL of a 1 mM aqueous solution of 2-mercaptoethanol, followed 1 min later by 25 µL of 50 mM 5,5′-dithio-bis(2-nitrobenzoic acid) (DTNB). The difference in absorbance at 412 nm ( ) 14 150) as compared to a control containing no protein indicated that there were a total of 0.7 maleimides/mAb. The modified mAb was kept at 4 °C prior to being added to the enzyme. hCE-2 (0.8 mL at 0.83 mg/mL in PBS; 33 µmol min-1 mg-1 specific activity using p-nitrophenylacetate as substrate) was diluted with 0.2 mL of 0.5 M sodium borate containing 0.5 M NaCl at pH 8.0. The protein was cooled on ice and treated with 53 µL of 20 mM 2-iminothiolane in water. After 180 min at 4 °C, low molecular weight agents were separated on a Sephadex G-25 PD 10 column, and the protein was collected in 2 mL of 100 mM sodium phosphate containing 0.5 M NaCl at pH 7.2. The protein concentration was 0.49 mg/mL, based on absorbance at 280 nm (E0.1% ) 1.01). Sulfhydryl group analysis, determined by the addition of 25 µL of DTNB using the difference in absorbance at 412 nm ( ) 14 150), indicated that there were 2.5 SH groups/protein molecule. The proteins were combined and concentrated to 1 mL by ultrafiltration. After 15 min at room temperature, 10 µL of 10 mM N-ethylmaleimide in DMSO was added, followed 5 min later by 20 µL of 10 mM 2-mercaptoethanol in water. The solution was further concentrated to 0.2 mL and then applied to a 7.8 × 300 mm Bio-Sil size-exclusion chromatography column (Biorad), which was eluted with PBS. Fractions that contained mainly monomeric conjugate mixed with unconjugated 96.5 mAb were gel filtered into 25 mM Tris buffer containing 125 mM NaCl at pH 7, and then applied to a 1 mL Q Sepharose column equilibrated with the same Tris buffer. Under these conditions, unconjugated mAb bound very poorly to the resin. The column was washed until there was no absorbance at 280 nm, and bound material was

eluted with 25 mM Tris buffer containing 1 M NaCl at pH 7. The resulting conjugate was almost entirely monomeric, with very small amounts of unconjugated 96.5 and aggregated protein. The specific activity with p-nitrophenylacetate was 5.2 µmol min-1 mg-1 of total protein, or 17.1 µmol min-1 mg-1 hCE-2 component. Binding Studies. 3677 cells were suspended in polypropylene tubes (105 cells/tube in 0.6 mL of RPMI culture medium containing 10% fetal bovine serum), and 96.5-FITC (1 µg/mL final concentration) was added with or without 96.5-hCE-2 at 0, 10, 25, and 50 µg/mL final 96.5 concentration. After 1 h at 4 °C, the cells were washed, and fluorescence was determined on a fluorescence activated cell sorter. The competition of binding was compared to cells that were treated with 96.5-FITC (1 µg/mL final concentration) together with unconjugated 96.5 at 0, 10, 25, and 50 µg/mL. 96.5-hCE-2 and unconjugated 96.5 were indistinguishable in this competition binding assay. Cytotoxicity Assays. CPT-11 and SN-38 were made as 10 mM stocks in DMSO and diluted in media just prior to use. Cells (5 × 103 cells/well) were plated into 96-well tissue culture plates in 0.1 mL of RPMI medium containing 10% fetal bovine serum. To the wells were added 0.05 mL of the drugs at various concentrations, followed by 0.05 mL of the enzyme, 96.5-hCE-2, or PBS in media for the times indicated in the figure legends. The plates were centrifuged at 200g for 5 min, media was aspirated off, and the cells were washed three additional times by centrifugation and aspiration. Additional media (0.15 mL) was added, and incubation was continued for a total of 72 h, at which time the cells were treated with XTT (2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino)carbonyl]-2H-tetrazolium hydroxide, Sigma) and evaluated for toxicity according to previously described methods (27, 28). RESULTS

The purification protocol of human liver microsomal proteins that effected the conversion of CPT-11 to SN38 (Figure 1) is shown in Table 1. Throughout protein purification, an HPLC-based assay was used to monitor this hydrolysis reaction, and the data were supplemented with total esterase activity using p-nitrophenylacetate as a substrate along with SDS-PAGE analysis. In establishing the HPLC assay, it was found as reported earlier (10), that the conversion of CPT-11 to SN-38 was more facile at CPT-11 concentrations that minimally inhibited the enzymatic hydrolysis of p-nitrophenylacetate. This is most likely due to the stability of the active site serine carbamate formed during the course of CPT-11 hydrolysis (10). Anion-exchange chromatography of the ammonium sulfate precipitated proteins led to two separate populations of protein that hydrolyzed CPT-11, one that bound to the resin and was eluted with a salt gradient, and one that did not bind to Q Sepharose. The subsequent purification steps, illustrated in Table 1 for the material that bound to the anion-exchange resin, were used for

Carboxylesterases for the Activation of CPT-11

Bioconjugate Chem., Vol. 12, No. 6, 2001 1077

Figure 2. Analysis of purified proteins by gel electrophoresis. (A) The purification of carboxylesterases from human liver microsomes was monitored by SDS-PAGE (10% polyacrylamide). Lanes 1-4 illustrate the purification of hCE-2. Lane 1, ammonium sulfate precipitated protein; lane 2, after anionexchange chromatography; lane 3, after hydrophobic interaction chromatography; lane 4, after size exclusion chromatography. Lane 5, hCE-1 using the same purification steps. Molecular weight standards gave bands as indicated. (B) Isoelectric focusing gel (pH 3-7) of purified proteins. Lane 1, hCE-1; lane 2, hCE-2. The IEF protein standards were obtained from Amersham Pharmacia. Both gels were stained with Coomassie Blue.

Figure 3. Sequence data from purified proteins. The peptide sequences determined from cyanogen bromide fragments of the purified proteins are underlined and bold. The sequences were 100% homologous to the sequences within (A) hCE-1 (16) and (B) hCE-2 (29).

both populations. After hydrophobic interaction chromatography on phenyl Sepharose, followed by size exclusion chromatography, the resulting protein was homogeneous by SDS-PAGE (Figure 2A) and had a diffuse pI, most likely due to post-transcriptional protein modification, centered at pH 4.4 (Figure 2B). The active protein that did not bind to the Q Sepharose resin was similarly purified and had a pI value of approximately 6.0. The two purified proteins had apparent molecular masses of 65 kDa. Attempts to identify the proteins by N-terminal sequencing failed since both proteins apparently had blocked amino termini. For that reason, the proteins were fragmented with cyanogen bromide, and some of the fragments that were separated by SDS-PAGE were subjected to N-terminal sequence analysis using Edman degradation chemistry. The results (Figure 3) indicate that the primary amino acid sequence of the lower activity enzyme corresponds to human carboxylesterase 1 (hCE-1, Figure 3A), and the higher activity protein corresponds to human carboxylesterase 2 (hCE-2, Figure 3B). In both cases, there was 100% homology between the sequenced peptides and the published sequences for hCE-1 and hCE-2 (16, 29). The specific activities of the purified human carboxylesterases were compared to highly purified samples of rabbit liver carboxylesterase (rCE) and guinea pig liver carboxylesterase (gpCE). The latter two enzymes were chosen based on reports describing them as being proficient at catalyzing the hydrolysis of CPT-11 (9, 16). Unexpectedly, hCE-2 had 66% of the activity of rCE and gpCE on CPT-11. Consistent with previous reports, hCE-1 was significantly less active than hCE-2 and the other enzymes tested (9, 16, 17). In vitro studies were carried out to compare the enzymes for their abilities to enhance the cytotoxicity of CPT-11. Tumor cell lines were exposed to varying concentrations of SN-38, CPT-11, or CPT-11 in the presence of purified carboxylesterases for a total of 24 h. Cell viability was measured 2 days later using XTT, a tetrazolium dye that measures mitochondrial membrane potential (27, 28). As expected (6, 30), SN-38 was approximately 1000-fold more potent than CPT-11 (Figure

4). The degree to which CPT-11 was activated on both L2987 human lung adenocarcinoma (Figure 4A,B) and on 3677 human melanoma (Figure 4C) cells was greater for hCE-2, rCE, and gpCE than for hCE-1. In addition, it was found that the extent of prodrug activation for hCE-2, rCE, and gpCE was equivalent and dose-dependent on the L2987 human lung adenocarcinoma cell line. The cytotoxic activity was greater at 5 µg/mL carboxylesterase (Figure 4B) than at 1 µg/mL (Figure 4A). On 3677 cells, hCE-2 was slightly less effective than rCE and gpCE (Figure 4C). These data show that hCE-2 activates CPT-11 under the conditions of the cytotoxicity assays. Conjugates of hCE-2 were prepared by linking the enzyme to the 96.5 mAb, a murine IgG2a that binds to the p97 antigen on most melanomas and on several carcinomas (24). The initial approach used was to react hCE-2 with N-(γ-maleimidobutyryloxy)succinimide ester (GMBS), which reacts with reactive amines such as those present on lysine side chains. However, the conjugates formed by combining hCE-2-GMBS adducts with antibody sulfhydryl groups suffered from approximately 90% reductions in enzyme activity. An alternative approach was more successful, in which sulfhydryl groups were appended to hCE-2 using 2-iminothiolane, and then the modified protein was reacted with maleimide-substituted 96.5. The 96.5-hCE-2 conjugate was then subjected to a two-step purification procedure that involved gel filtration to remove aggregates and free enzyme, followed by anion-exchange chromatography to separate most of the unconjugated mAb. Analysis by SDS-PAGE indicated that the conjugate was >75% monomeric. As has been previously shown with 96.5 enzyme conjugates (24, 25), binding studies using flow cytometric analysis indicated that 96.5-hCE-2 bound selectively to antigen positive 3677 melanoma cells and had unimpaired binding activity as compared to unmodified 96.5 (see Experimental Methods). 96.5-hCE-2 retained 52 and 69% of the specific activity with p-nitrophenylacetate and CPT-11 as substrates, respectively, as compared to the hCE-2 that was used for the conjugation (Table 2). In vitro cytotoxicity experiments were conducted on 3677 melanoma cells (p97 antigen positive) that were

1078 Bioconjugate Chem., Vol. 12, No. 6, 2001

Senter et al.

Figure 4. In vitro cytotoxicity assays. (A) and (B) L2987 human lung adenocarcinoma and (C) and (D) 3677 human melanoma cells were exposed to drugs with or without purified carboxylesterases for 24 h, washed, and the cytotoxic effects were determined 48 h later using XTT as an indicator of cell viability. For (D), 96.5-hCE-2 was used at 5 µg/mL total protein, representing 1.5 µg/mL hCE-2. Conjugate treated cells were either exposed to conjugate throughout the period of CPT-11 treatment (bound + unbound), or were washed to removed unbound conjugate prior to CPT-11 treatment (bound). Table 2. Activities of Purified Carboxylesterases specific activity proteina

p-nitrophenylacetate (µmol min-1 mg-1)

CPT-11 (pmol min-1 mg-1)

hCE-1 hCE-2 rCE gpCE 96.5-hCE-2b

10 60 28 46 17

3.5 92 142 140 48

a The abbreviations used for the proteins are hCE-1, human carboxylesterase 1; hCE-2, human carboxylesterase 2; rCE, rabbit liver carboxylesterase; gpCE, guinea pig liver carboxylesterase; 96.5-hCE-2, the conjugate formed between the 96.5 IgG and hCE2. b The specific activity is based on the hCE-2 component of the conjugate, which had 33 µmol min-1 mg-1 and 70 pmol min-1 mg-1 activity on p-nitrophenylacetate and CPT-11, respectively.

exposed to 96.5-hCE-2 prior to treatment with CPT-11. If the cells were extensively washed prior to prodrug treatment, there was no enhancement in the cytotoxic effects of CPT-11 (Figure 4D). In contrast, cells that were treated with the combination of 5 µg of 96.5-hCE-2/mL (contains 1.5 µg/mL hCE component) and CPT-11 for 24 h were sensitized to CPT-11, albeit to a slightly lesser extent than corresponding cells treated with free hCE at 5 µg/mL. These results indicate that the enzyme within the conjugate is active and capable of activating CPT-11 under the conditions used in the cytotoxicity experiments, but that the amount of antigen-bound activity is too low to achieve detectable levels of prodrug activation. Several experiments were conducted using increased conjugate concentrations, exposing cells at high density with saturating concentrations of conjugate (31), using other mAb targeting vehicles for hCE-2, and using assays involving the binding of biotinylated hCE-2 to mAb-streptavidin conjugates. The results were invariable in that bound enzyme conjugates did not release cytotoxic concentrations of active drug. We conclude that the activity of hCE-2 is insufficient for targeting strategies utilizing

mAb-hCE-2 conjugates for antigen-specific CPT-11 activation. DISCUSSION

The use of mAb-enzyme conjugates for prodrug activation at tumor sites has received considerable attention, and several targeted enzyme/prodrug combinations have demonstrated significant levels of antitumor activity in preclinical cancer models (reviewed in refs 22-23). The most advanced systems include bacterial enzymes, such as carboxypeptidase G2 and β-lactamase for the activation of nonclinically approved anticancer prodrugs (reviewed in refs 22, 23). One of the shortcomings of using such proteins is their expected immunogenicity in people, as evidenced by the anti-CEA mAb conjugate A5B7-mAbcarboxypeptidase G2 where all patients treated mounted significant immune reactions against the foreign protein (32). An additional complication in such systems is that the prodrugs are not clinically approved, and have unknown toxicities, pharmacokinetics, and biological effects. A preferable combination would be to utilize mAbhuman enzyme conjugates for the activation of clinically approved anticancer prodrugs. No such examples exist, although there have been reports of activating clinically approved anticancer prodrugs with foreign proteins (33) and activating nonapproved anticancer prodrugs with human proteins (34, 35, reviewed in ref 26). In using a human enzyme for the activation of the clinically approved anticancer drug CPT-11, we were aware that the known enzymes that converted CPT-11 to SN-38 had very poor activities. With the intent of exploring which enzyme might be the most efficient, we began by purifying enzymes from human liver microsomes on the basis of their abilities to effect the hydrolysis reaction. This search led us to hCE-1, one of the most abundant of the microsomal carboxylesterases (16, 17), and to hCE-2, which was reported during the course of our work as being more active than hCE-1 with CPT-11 as a substrate (21). In our hands,

Carboxylesterases for the Activation of CPT-11

hCE-2 was 26-fold more active than hCE-1 with CPT-11 and was almost as active as rCE, the highest activity enzyme known for the hydrolysis of CPT-11 (16, 19). Previous studies have shown that cancer cell lines transfected with the genes for either rCE (16, 17) or with hCE-1 (19, 20) were rendered significantly more sensitive to CPT-11 treatment as compared to the parental cells. Our results would suggest that the gene for hCE-2 might have considerably greater potential for gene therapy for CPT-11 activation, since it is more active than hCE-1 and would not be expected to induce an immune reaction like rCE. It is significant that in vitro potentiation of CPT-11 activity could be obtained by gene therapy with hCE-1 and with rCE2 (16, 17, 19, 20) but that we were unable to demonstrate any such activity on cells that were exposed to 96.5-hCE-2. Another conjugate, 96.5-β-lactamase, has been used successfully for specific activation of many anticancer prodrugs under conditions involving subsaturating concentrations and prodrug exposure for very short periods of time (24, 25). Here, we were unable to effect prodrug activation even at saturating 96.5-hCE-2 concentrations and with prodrug exposures for as long as 24 h. We believe the reason for the discrepancy between these two systems lies in the kinetics of the enzymes used. While β-lactamase has specific activities in the range of 315 µmol min-1 mg-1 with a doxorubicin prodrug (25) to 1470 µmol min-1 mg-1 for a melphalan prodrug (36), hCE-2 has activity 106-107-fold lower than that. The fact that hCE-1 and rCE have proven useful in gene therapy studies might be attributed to very high level expression of the protein inside the target cells as compared to the amount that can be delivered through a mAb-enzyme conjugate, and to possible differences between intracellular and extracellular generation of SN-38. We conclude that the activity of hCE-2 must be significantly increased to be effective for ADEPT. One of the lowest activity enzymes successfully used to activate an anticancer prodrug in an immunologically specific manner was penicillin G-amidase for the activation of a prodrug of doxorubicin (31). The specific activity of the enzyme was 12.7 nmol min-1 mg-1, which is only 140fold greater than hCE-2. It is reasonable to expect that directed evolution of hCE-2 toward CPT-11 hydrolysis should achieve catalytic efficiencies in this range or higher. Such studies are warranted, given the clinical need to improve the therapeutic window for CPT-11 and the advantages in using a potentially nonimmunogenic conjugate for doing so. ACKNOWLEDGMENT

This work was supported by grant no. 1 R43 CA8510901 from the National Cancer Institute. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute. LITERATURE CITED (1) Vanhoefer, U., Harstrick, A., Achterrath, W., Cao, S., Seeber S., and Rustum, Y. M. (2001) Irintotecan in the treatment of colorectal cancer: clinical overview. J. Clin. Oncol. 19, 15011518. (2) Rothenberg, M. L. (2001) Irinotecan (CPT-11): recent developments and future directions-colorectal cancer and beyond. Oncologist 6, 66-80. (3) Masuda, N., Kudoh, S., and Fukuoka, M. (1996) Irinotecan CPT-11): pharmacology and clinical applications. Crit. Rev. Oncol./Hematol. 24, 3-26.

Bioconjugate Chem., Vol. 12, No. 6, 2001 1079 (4) Herben, V. M. M., ten Bokkel Huinink, W. W., Schellens, J. H. M., and Beijnen, J. H. (1998) Clinical pharmacokinetics of camptothecin topoisomerase inhibitors. Pharm. World Sci. 20, 161-172. (5) O’Reilly, S. O., and Rowinsky, E. K. (1996) The clinical status of irinotecan (CPT-11): a novel water soluble campothecin analogue: 1996. Crit. Rev. Oncol./Hematol. 24, 47-70. (6) Kaneda, N., Nagata, H., Furuta, T., and Yokokura, T. (1990) Metabolism and pharmacokinetics of the camptothecin analogue CPT-11 in the mouse. Cancer Res. 50, 1715-1720. (7) Kawato, Y., Aonuma, M., Hirota, Y., Kuga, H., and Sato, K. (1991) Intracellular roles of SN-38, a metabolite of the camptothecin derivative CPT-11, in the antitumor effect of CPT-11. Cancer Res. 51, 4187-4191. (8) Tsuji, T., Kaneda, N., Kado, K., Yokokura, T., Yoshimoto, T., and Tsuru, D. (1991) CPT-11 converting enzyme from rat serum: purification and some properties. J. Pharmacobiodyn. 14, 341-349. (9) Satoh, T., Hosokawa, M., Atsumi, R., Suzuki, W., Hakusui, H., and Nagai, E. (1994) Metabolic activation of CPT-11, 7-ethyl-10[4-(1piperidino)-1-piperidino]carbonyloxycamptothecin, a novel antitumor agent, by carboxylesterase. Biol. Pharm. Bull. 17, 662-664. (10) Rivory, L. P., Bowles, M. R., Robert, J., and Pond, S. M. (1996) Conversion of irinotecan (CPT-11) to its active metabolite, 7-ethyl-10-hydroxycamptothecin (SN-38), by human liver carboxylesterase. Biochem. Pharmacol. 52, 1103-1111. (11) Scott, D. F., Chacko, T. L., Maxwell, D. M., Schlager, J. J., and Lanclos, K. D. (1999) Expression and partial purification of a recombinant secretory form of human liver carboxylesterase. Protein Expr. Purif. 17, 16-25. (12) Khanna, R., Morton, C. L., Danks, M. K., and Potter, P. M. (2000) Proficient metabolism of irinotecan by a human intestinal carboxylesterase. Cancer Res. 60, 4725-4728. (13) Hennibelle, I., Terret, C., Chatelut, E., Bugat, R., Canal, P., and Guichard, S. (2000) Characterization of CPT-11 converting carboxylesterase activity in colon tumor and normal tissues: comparison with p-nitrophenylacetate converting carboxylesterase activity. Anticancer Drugs 11, 465470. (14) Guemei, A. A., Cottrell, J., Band, R., Hehman, H., Prudhomme, M., Pavlov, M. V., Grem, J. L., Ismail, A. S., Bowen, D., Taylor, R. E., and Takimoto, C. H. (2001) Human plasma carboxylesterase and butyrylcholinesterase enzyme activity: correlations with SN-38 pharmakoinetics during a prolonged infusion of irinotecan. Cancer Chemother. Pharmacol. 47, 283-290. (15) Niculescu-Duvaz, I., Spooner, R., Marais, R., and Springer, C. J. (1998) Gene-directed enzyme prodrug therapy. Bioconjugate Chem. 9, 4-22. (16) Danks, M. K., Morton, C. L., Krull, E. J., Cheshire, P. J., Richmond, L. B., Naeve, C. W., Pawlik, C. A., Houghton, P. J., and Potter, P. M. (1999) Comparison of activation of CPT11 by rabbit and human carboxylesterases for use in enzyme/ prodrug therapy. Clin. Cancer Res. 5, 917-924. (17) Danks, M. K., Morton, C. L., Pawlik, C. A., and Potter, P. M. (1998) Overexpression of a rabbit liver carboxylesterase sensitizes human tumor cells to CPT-11. Cancer Res. 58, 2022. (18) Morton, C. L., Wadkins, R. M., Danks, M. K., and Potter, P. M. (1999) The anticancer prodrug CPT-11 in a potent inhibitor of acetylcholinesterase but is rapidly catalyzed to SN-38 by butyrylcholinesterase. Cancer Res. 59, 14581463. (19) Kojima, A., Hackett, N. R., and Crystal, R. G. (1998) Reversal of CPT-11 resistance of lung cancer cells by adenovirus-mediated gene transfer of the human carboxylesterase cDNA. Cancer Res. 58, 4368-4374. (20) Kojima, A., Hackett, N. R., Ohwada, A., and Crystal, R. G. (1998) In vivo human carboxylesterase cDNA gene transfer to activate the prodrug CPT-11 for local treatment of solid tumors. J. Clin. Invest. 101, 1789-1796. (21) Humerickhouse, R., Lophrbach, K., Li, L., Bosron, W. F., and Dolan, M. E. (2000) Characterization of CPT-11 hydrolysis by human liver carboxylesterase isoforms hCE-1 and hCE2. Cancer Res. 60, 1189-1192.

1080 Bioconjugate Chem., Vol. 12, No. 6, 2001 (22) Dubowchik, G. M., and Walker, M. A. (1999) Receptormediated and enzyme-dependent targeting of cytotoxic anticancer drugs. Pharmacol. Ther. 83, 67-123. (23) Senter, P. D., and Springer, C. J. (2001) Selective activation of anticancer prodrugs by monoclonal antibody-enzyme conjugates. Adv. Drug. Del. Rev. In press. (24) Kerr, D. E., Schreiber, G. L., Vrudhula, V. M., Svensson, H. P., Hellstro¨m, I., Hellstro¨m, K. E., Senter P. D. (1995) Regressions and cures of melanoma xenografts following treatment with monoclonal antibody β-lactamase conjugates in combination with anticancer prodrugs. Cancer Res. 55, 3558-3563. (25) Svensson, H. P., Vrudhula, V. M., Emswiler, J. E., MacMaster, J. F., Cosand, W., J., Senter, P. D., and Wallace, P. M. (1995) In vitro and in vivo activities of a doxorubicin prodrug in combination with monoclonal antibody β-lactamase conjugates. Cancer Res. 55, 2357-2365. (26) Hosokawa, M., Maki, T., and Satoh, T. (1990) Characterization of molecular species of liver microsomal carboxylesterases of several animal species and humans. Arch. Biochem. Biophys. 277, 219-227. (27) Scudiero, D. A., Shoemaker, R. H., Paull, K. D., Monks, A., Tierney, S., Nofziger, T. H., Currins, M J., Seniff, D., and Byrd, M. R. (1988) Evaluation of a soluble tetrazolium/ formazan assay for cell growth and drug sensitivity in culture using human and other tumor cell lines. Cancer Res. 48, 4827-4833. (28) Donaldson, K. L., Goolsby, G. L., and Wahl, A. F. (1994) Cytotoxicity of the antitumor agents cisplatin and taxol during cell proliferation and the cell cycle. Int. J. Cancer 57, 847-855. (29) Pindel, E. V., Kedishvili, N. Y., Abraham, T. L., Brzezinski, M. R., Zhang, J., Dean, R. A., and Bosron, W. F. (1997) Purification and cloning of a broad substrate specificity

Senter et al. human liver carboxylesterase that catalyzes the hydrolysis of cocaine and heroin. J. Biol. Chem. 272, 14769-14775. (30) Senter, P. D., Marquardt, H., Thomas, B. A., Hammock, B. D., Frank, I. S., and Svensson, H. P. (1996) The role of rat serum carboxylesterase in the activation of paclitaxel and camptothecin prodrugs. Cancer Res. 56, 1471-1474. (31) Vrudhula, V. M., Senter, P. D., Fischer, K. J., and Wallace. P. M. (1993) Prodrugs of doxorubicin and melphalan and their activation by a monoclonal antibody-penicillin-G amidase conjugate. J. Med. Chem. 36, 919-923. (32) Bagshawe, K. D., and Begent, R. H. J. (1996) First clinical experience with ADEPT. Adv. Drug Del. Rev. 22, 365367. (33) Wallace, P. M., MacMaster, J. F., Smith, V. F., Kerr, D. E., Senter, P. D., and Cosand, W. L. (1994) Intratumoral generation of 5-flourouracil mediated by an antibody-cytosine deaminase conjugate in combination with 5-fluorocytosine. Cancer Res. 54, 2719-2723. (34) Wolfe, L. A., Mullin, R. J., Laethum, R., Blumenkopf, T. A., Cory, M., Miller, J. F., Keith, B. R., Humphreys, J., and Smith, G. K. (1999) Antibody-directed enzyme prodrug therapy with the T268G mutant of human carboxypeptidase A1: in vitro and in vivo studies with prodrugs of methotrexate and the thymidylate synthase inhibitors GW1031 and GW1843. Bioconjugate Chem. 10, 38-48. (35) Bosslet, J., Czech, J., and Hoffmann, D. (1994) Tumorselective prodrug activation by fusion protein-mediated catalysis. Cancer Res. 54, 2151-2159. (36) Kerr, D. E., Li, Z., Siemers, N. O., Senter, P. D., and Vrudhula, V. M. (1998) Development and activities of a new melphalan prodrug designed for tumor-selective activation. Bioconjugate Chem. 9, 255-259.

BC0155420