Homocamptothecin, an E-Ring Modified Camptothecin with Enhanced Lactone Stability, Retains Topoisomerase I-targeted Activity andAntitumor Properties¹

Homocamptothecin (hCPT) is a semisynthetic analogue of camptothecin (CPT) with a seven-membered β-hydroxylactone resulting from the insertion of a methylene spacer between the alcohol moiety and the carboxyl function of the naturally occurring six-membered α-hydroxylactone of CPT. This E-ring modification provides a less reactive lactone with enhanced stability and decreased protein binding in human plasma. Biological testing against CPT revealed that, instead of being detrimental, the modified lactone of hCPT has a positive impact on topoisomerase I (Topo I) poisoning properties. In vitro tests showed hCPT to fully conserve the ability to stabilize Topo I-DNA cleavage complexes and to stimulate a higher level of DNA cleavage than CPT. A similar trend toward improvement was also observed in antiproliferative assays with human tumor cell lines (including cells overexpressing P-glycoprotein). In two distinct in vivo models, using L1210 murine leukemia or human colon carcinoma HT29, hCPT was found to be more efficacious than CPT. The slow, but irreversible, hydrolysis of hCPT, instead of the fast equilibrium of CPT, may account for its good in vivo activity. Overall, these results provide evidence that a highly reactive lactone is not a requisite for the Topo I-mediated antitumor activity of CPT analogues, and that hCPT is an interesting pharmacological tool with improved solution behavior as well as a promising new template for the preparation of more efficacious Topo I poisons.

The CPT³-derived Topo I inhibitors have undergone considerable development (1) in recent years, leading to the approval of topotecan and irinotecan as second-line chemotherapeutic agents against resistant cancers. In addition, several other molecules including CPT, 9-aminocamptothecin (9-AC), 9-nitrocamptothecin (9-NC), GG-211, and DX-8951f have been clinically evaluated (2), and many other potent CPT analogues are in research or preclinical stage of development (3). Structure-activity studies have shown the CPT skeleton to be amenable to substitution, providing analogues with improved pharmacological profiles. Most effort has targeted water-soluble and/or more active molecules, whereas compounds exhibiting dual activities involving DNA alkylation, intercalation, or minor groove binding in addition to Topo I inhibition have been the focus of recent studies (3).

It is a generally accepted rule for CPT-derived molecules that the α-hydroxylactone ring is an absolute requirement for in vitro and in vivo activities (1). Indeed, the discontinuation of early clinical trials with CPT-sodium was due to the lack of useful activity of the ring-opened, carboxylate form adopted at that time (4), and reported attempts to modify the α-hydroxylactone have failed to preserve Topo I-mediated activity (5, 6, 7). Recently proposed binding modes for CPT and Topo I-DNA cleavable complex further confirm the participation of the lactone ring structural features to the CPT pharmacophore (8, 9). Unfortunately, the α-hydroxylactone imparts high chemical reactivity to CPT and its analogues, which are thus intrinsically unstable molecules at physiological pH.

This instability arises from the rapid hydrolysis of the α-hydroxylactone in basic or neutral media to give the ring-opened, carboxylate form, which is essentially inactive. The reaction is reversible, and the lactone form predominates only at acidic pH. Pharmacokinetic studies have shown this pH-dependent hydrolytic equilibrium to be shifted toward the carboxylate form in plasma in a species-dependent manner that is less favorable in man than in rodents (10). This latter point has been invoked to explain the diminished efficacy of various CPT analogues in the clinic compared with the spectacular results often obtained with xenograft models (11). To circumvent the rapid opening of the lactone, the drugs may be administered by continuous infusion for long periods. Alternatively, prodrugs have been proposed to achieve sustained plasma levels of CPT analogues in their lactone form. No active principle combining improved stability with good in vitro and in vivo Topo I inhibitory activity has been previously reported.

We have recently described the semisynthetic preparation of a CPT analogue bearing a seven-membered β-hydroxylactone ring instead of the naturally occurring six-membered α-hydroxylactone (12). Because a one-carbon ring expansion is chemically termed a homologation, this new lactone ring-modified compound is a “homocamptothecin” (hCPT). Because of the reduced electrophilicity of a β-hydroxylactone, the racemic compound showed enhanced stability in buffer solutions, and in vitro testing showed unexpectedly good inhibition of the Topo I-mediated, supercoiled DNA relaxation, as well as antiproliferative activity on L1210 murine leukemia cells (12).

On the basis of these encouraging results, an analogue generation program has been undertaken to identify potential candidate molecules for clinical development (13) in parallel with further pharmacological evaluation of hCPT. The biologically active enantiomer has been isolated in optically pure form (13) and is referred to as hCPT (Fig. 1), whereas dl-hCPT designates the racemic mixture. Experimental comparison of hCPT with CPT should contribute to understanding of the role of lactone reactivity in biological activity.

CPT was purchased from Sigma (La Verpillière, France). Preparation and purification of racemic dl-hCPT and enantiomerically pure hCPT were performed according to previously published procedures (13). A three-dimensional structure obtained by X-ray diffraction on a synthetic precursor established that, as shown in Fig. 1, hCPT and CPT have identical spatial arrangements around their asymmetric carbons.

To 500-μl fractions of pooled human plasma (Transfusion Sud-Est Francilien, Rungis, France), distributed in polystyrene vials and preincubated at 37°C for 5 min, were added 5 μl of a solution of the drug—hCPT or CPT—in DMSO (10 mm) and the samples were incubated in capped vials (to prevent CO₂ loss), at 37°C. At defined times, three vials were opened, and the plasma proteins were precipitated by the addition of 2 ml of cold acetonitrile (instead of methanol, which might not be inert to CPT and analogues) at −50°C. The resulting mixture was centrifuged (at 2000 × g) at 4°C for 5 min, and the supernatant was analyzed immediately to avoid further chemical transformation. Samples were run on a 5-μm Nucleosil C18 column (4.6 × 150 mm) at 30°C with a flow rate of 1 ml/min of an isocratic eluent composed of 1 m tetrabutylammonium dihydrogen phosphate/acetonitrile/75 mm (pH 6.9) ammonium acetate buffer (5/125/870, v/v/v), and the eluted analytes were detected at 360 nm.

Supercoiled pKMp27 DNA (0.5 μg) was incubated with six units of human Topo I (TopoGen Inc., Colombus, OH) at 37°C for 45 min in relaxation buffer [50 mm Tris (pH 7.8), 50 mm KCl, 10 mm MgCl₂, 1 mm DTT, and 1 mm EDTA] in the presence of different concentrations of the drug under study. Reactions were terminated by adding SDS to 0.25% and proteinase K to 250 μg/ml. DNA samples were then added to the electrophoresis dye mixture (3 μl) and electrophoresed in a 1% agarose gel at room temperature for 3 h. Gels were stained with ethidium bromide (1 mg/ml), washed, and photographed under UV light. Similar experiments were performed using ethidium bromide-containing agarose gels.

The plasmid pKMp27 was linearized with EcoRI and labeled with [α-³²P]dATP in the presence of the Klenow fragment of DNA polymerase I. The labeled DNA was then digested to completion with AvaI. The cleavage reaction mixture contained 20 mm Tris-HCl (pH 7.4), 60 mm KCl, 0.5 mm EDTA, 0.5 mm DTT, 2 × 10⁴ dpm of α-³²P-DNA, and the indicated drug concentrations. The reaction was initiated by the addition of Topo I (10 units in 20 μl of reaction volume) and allowed to proceed for 45 min at 37°C. Reactions were stopped by adding SDS to a final concentration of 0.25% and proteinase K to 250 μg/ml, followed by incubation for 30 min at 50°C. Samples were denatured by the addition of 5 μl of buffer consisting of 0.45 m NaOH, 30 mm EDTA, 15% (w/v) sucrose, and 0.1% bromcresol green before loading onto 1% agarose gel in TBE buffer containing 0.1% SDS. Electrophoresis was conducted at 2V/cm for 18 h.

The 117-bp DNA fragment was prepared by 3′-³²P end labeling of the EcoRI-PvuII double digest of the plasmid pKMp27 using [α-³²P]dATP (6000 Ci/mmol) and avian myeloblastosis virus reverse transcriptase. The digestion products were separated on a 6% polyacrylamide gel under native conditions in TBE buffer [89 mm Tris-borate (pH 8.3) and 1 mm EDTA]. After autoradiography, the band of DNA was excised, crushed, and soaked in water overnight at 37°C. This suspension was filtered through a Millipore 0.22 μ filter, and the DNA was precipitated with ethanol. After washing with 70% ethanol and vacuum drying of the precipitate, the labeled DNA was resuspended in 10 mm Tris (adjusted to pH 7.0) containing 10 mm NaCl.

Each reaction mixture contained 2 μl of 3′-end ³²P-labeled DNA (∼1 μm), 5 μl of water, 2 μl of 10× Topo I buffer, and 10 μl of drug solution at the desired concentration (1–100 μm). After a 10-min incubation to ensure equilibration, the reaction was initiated by the addition of 2 μl (20 units) of calf thymus Topo I (Life Sciences, Cergy Pontoise, France). Samples were incubated for 45 min at 37°C before adding SDS to 0.25% and proteinase K to 250 μg/ml to dissociate the drug-DNA-Topo I cleavable complexes. The DNA was precipitated with ethanol, resuspended in 5 μl of formamide-TBE loading buffer, denatured at 90°C for 4 min, and then chilled in ice for 4 min before loading onto the sequencing gel. DNA cleavage products were resolved by PAGE under denaturing conditions (0.3-mm thick, 8% acrylamide containing 8 m urea). After electrophoresis (about 2.5 h at 60 W, 1600 V in TBE buffer, Life Technologies, Inc. sequencer model S2), gels were soaked in 10% acetic acid for 10 min, transferred to Whatman 3 MM paper, and dried under vacuum at 80°C. A Molecular Dynamics 425E PhosphorImager was used to collect data from the storage screens exposed to dried gels overnight at room temperature. Baseline-corrected scans were analyzed by integrating all of the densities between two selected boundaries using ImageQuant version 3.3 software. Each resolved band was assigned to a particular bond within the DNA fragment by comparison of its position relative to sequencing standards generated by the treatment of the DNA with dimethylsulfate followed by piperidine-induced cleavage at the modified guanine residues.

The A427 (lung carcinoma) and PC-3 and DU145 (prostatic adenocarcinoma) cell lines were obtained from American Type Culture Collection (Rockville, MD). The MCF7r (breast adenocarcinoma) and K562r (leukemia) cell lines were obtained from A-M. Faussat (Hôpital Hôtel-Dieu, Paris, France). They were derived from sensitive cell lines by prolonged exposure to Adriamycin and have been shown by flow cytometry to overexpress a functionally active P-glycoprotein. Cells (3000 for A427, MCF7r, and K562r, and 1500 for PC-3 and DU145, in 80 μl of DMEM at 4.5 g/l glucose supplemented by 10% heat-inactivated FCS, 50 units/ml penicillin-50 μg/ml streptomycin, and 2 mm glutamine—all from Life Technologies, Inc., Cergy-Pontoise, France) were seeded on a microtiter plate (tissue culture grade, 96 wells, flat-bottomed) 24 h before drug treatment. The cells were treated with the drugs, which had first been dissolved in DMSO (0.1% of final volume) and further diluted to 20 μl with culture medium, at final concentrations ranging from 5 × 10^-13 to 1 × 10^-6 m. At the end of a 72-h incubation period, the quantification of cell viability was evaluated by a colorimetric assay based on the reduction of the tetrazolium salt WST1 (Boehringer Mannheim, Meylan, France) by the mitochondrial dehydrogenases of living cells leading to the formation of soluble formazan. These experiments were performed twice with eight determinations per tested concentration. For each drug, the data points included in the linear part of the sigmoid were selected by linear regression analysis (linearity, deviation from linearity, and difference between experiments) to estimate the IC₅₀.

The mouse L1210 lymphoblastic leukemia cells were maintained by serial i.p. passage in DBA/2 mice (Iffa-Credo, l’Arbresle, France), and 10⁶ cells/0.2 ml were injected i.p. into B6D2F1 female mice (Iffa-Credo, l’Arbresle, France). The drugs were first dissolved in DMSO (2% of total volume) and further diluted with 0.9% NaCl solution. The injection volume was adjusted to 0.1 ml/10 g of body weight. The i.p. treatments began on day 1 after leukemia inoculation and were repeated daily until day 4, followed by 3 days without treatment. This 4-days-on/3-days-off cycle was repeated 1–5 times according to the experiment. The control mice were given the vehicle according to the schedule of the treated mice. Survival of the animals (5 mice/group) was monitored for up to 60 days, and the treatment was evaluated as the T/C index calculated from the median survival of the treated mice with respect to that of the control mice. Animal care was in accordance with institutional guidelines.

Tumors were established by s.c. injection of tissue culture-derived tumor cells (5 × 10⁶ cells/animal on the left dorsal surface) in 4–6-week-old NCr nu/nu female athymic nude mice (National Cancer Institute, Frederick, MD). Tumor volume (mm³) was calculated as

⁠, where w is the width and l is the length of the tumor as measured with calipers. The animals were monitored until day 10 postimplant, when the tumors reached an average size of approximately 175 mm³. The range of tumor volumes was then tallied, and the animals were randomized into appropriate groups (5 animals/treatment and 10 animals for the control group) so that the average tumor volumes were approximately the same for each group at the onset of dosing. The drugs were first dissolved in DMSO (3% total injection volume) and then diluted with Tween 20 (2% total injection volume); further dilutions were made with 0.9% NaCl solution. Animals were treated i.p. daily for 4 consecutive days, followed by 3 days without treatment. This 4-days-on/3-days-off cycle was repeated twice more for a total of 12 injections over the course of 3 weeks. The mice of the control group were given the vehicle according to the schedule of the treated mice. Three times weekly, tumor sizes and body weights were recorded for all of the animals. The T-C index, measured in days, is calculated as the difference in time required for the mean tumor volume to reach 250% of the tumor average volume at the onset of dosing for the treated group (T) and the control group (C). Animal care was in accordance with institutional guidelines.

The data (mean values of triplicate incubations) obtained by HPLC monitoring of the drugs in human plasma at 37°C and their best-fitting curves determined by least-squares linear regression are shown in Fig. 2. Clearly, homologation of CPT enhances plasma stability: by 30 min, CPT lactone levels fell to less than 10 μm, whereas hCPT lactone reached this low level only at the end of the 5-h time course of the experiment (Fig. 2,A). The nonequilibrating situation suggested by the linear decay of hCPT was confirmed by control experiments in which, upon incubation in human plasma at 37°C for 24 h, no ring-closure of the open form of hCPT was detected. Fig. 2 B corresponds to the measurement of the total drug fraction (lactone + carboxylate) remaining in solution after precipitation of the plasma proteins and shows hCPT to be fully recovered at each sampling time, in contrast to CPT for which the unbound fraction decreased rapidly from a 100 μm initial value to a steady level of 40 μm. The missing fraction of CPT presumably remained bound to plasma proteins, as demonstrated previously by others (14).

Negatively supercoiled plasmid pKMp27 was incubated with human Topo I and concentrations of CPT or hCPT ranging from 1 to 100 μm. The gel shown in Fig. 3A indicates that, even at low drug concentrations (1–5 μm), the intensity of the slowest-migrating band (corresponding to nicked + fully relaxed DNA) increases significantly for both hCPT and CPT while, in the absence of Topo I, neither drug unwinds DNA. To rule out a nonspecific effect due, for example, to direct interaction of hCPT with DNA, the relaxation assay was repeated on an ethidium bromide-containing agarose gel, as shown in Fig. 3B, and quantified in Fig. 3C. The level of nicked DNA molecules was considerably increased in the presence of hCPT, and was more potent than that observed with CPT under identical conditions.

A preliminary test on a 2.5-kbp EcoRI-AvaI restriction fragment showed that strong Topo I-mediated DNA cleavage sites were observed on agarose gel with both CPT and hCPT (Fig. 4,A). To quantify the respective cleavage efficiency of the drugs, a DNA fragment of 117 bp was used to map the induced Topo I cleavage sites. The EcoRI-PvuII double-digest of the pBS plasmid was uniquely end-labeled at the 3′-end at the EcoRI site and used as a substrate for the Topo I cleavage reactions. Cleavage products were analyzed on a sequencing polyacrylamide gel (Fig. 4,B), which fully confirmed that hCPT, like CPT, strongly promotes DNA cleavage in the presence of Topo I. Densitometric analysis of the gels (Fig. 4 C) revealed that cleavage at TG sites in the presence of low concentrations of hCPT was much higher than with CPT.

The antiproliferative activities of the drugs on the A427, PC-3, DU145, MCF7r, and K562r cell lines are presented in Table 1. On four of five cells lines, the IC₅₀s determined by WST1 colorimetric assay were lower for hCPT than for CPT, and both drugs had comparable IC₅₀s for MCF7r.

The data presented in Fig. 5,A show that when injected i.p. daily for 4 days, racemic dl-hCPT increased the life span of the mice (untreated animals died between day 9 and day 14 posttumor-cell injection) treated at doses ranging from 0.32 to 2.5 mg/kg, with maximal efficacy at 1.25 mg/kg. In contrast, CPT was active at only two doses, 1.25 and 2.5 mg/kg, and with lower efficacy. Fig. 5 B shows that when the four injections/week cycle was repeated for up to 5 weeks at subtoxic doses, the efficacy of CPT and dl-hCPT increased with the number of cycles. T/C index reached a maximum of 200% for the CPT groups, whereas many “long-term survivors,” with a T/C ≥ 300%, were seen in the dl-hCPT groups.

The doses used to treat the HT-29 grafted athymic mice were 0.625, 1.25, and 2.5 mg/kg, administered i.p., and the resulting T-C indexes were found to be dose-dependent. Animal weight loss, when observed, was less than 10% and recovered within 1 week after the end of treatment. The graph shown in Fig. 6 represents the evolution of the tumor size after treatments with the drugs at their MTD among the above-mentioned doses. Higher doses of hCPT than CPT (1.25 versus 0.625 mg/kg) could be administered in this model, and a T-C index of 12 days was obtained for hCPT instead of 4 days for CPT.

The insertion of a methylene spacer (—CH₂—) between the alcohol moiety and the carbonyl group of CPT lactone has two pronounced chemical consequences. One effect is due to the removal of the electronic induction of the hydroxyl oxygen atom on the carbonyl group, which results in a lower reactivity of the carboxyl function toward nucleophiles such as water, alcohols, or amines. The second effect stems from the thermodynamically disfavored closure of a seven-membered ring in comparison with a six-membered ring. Taken together, these effects afford a mechanistic explanation for the slow and irreversible hydrolytic ring-opening of hCPT, in contrast to the rapid hydrolysis of CPT to a pH-dependent equilibrium. These differences are exemplified by the distinct plasma hydrolytic profiles of the drugs (Fig. 2,A) corresponding to, respectively, a linear decay for hCPT and an exponential decay for CPT. An estimate of the relative reactivities is given by the initial slopes of the curves and shows hCPT to hydrolyze at a much lower rate, a highly desirable effect because the low in vivo stability of CPT analogues represents a hurdle to the exploitation of their full potential as anticancer agents (15). Another striking, and unexpected, difference was observed while monitoring the total drug (lactone + carboxylate) remaining in solution after the precipitation of the plasma proteins (Fig. 2 B). In contrast to CPT, which was found to be 60% protein-bound, hCPT was fully recovered over the time course of the experiment. This could advantageously reduce the species-dependencies observed for CPT analogues, due to specific albumin interactions (14).

A qualitative approach to the drugs mechanism of action reveals that, despite its strong effect on lactone chemistry, homologation of CPT fully conserves the ability to stabilize Topo I-DNA cleavable complexes. The results in Fig. 3 clearly indicate that hCPT stabilizes covalent Topo I-DNA species, as observed with CPT. Quantitatively, the sequencing polyacrylamide gel (Fig. 4,B) shows unambiguously that the Topo I inhibitory activity of hCPT is more potent than that of CPT. At the low concentrations of 1 and 5 μm, visual inspection of the gel suffices to demonstrate that the cleavage of DNA by the enzyme is weak in the presence of CPT and very pronounced with hCPT, and a densitometric analysis (Fig. 4 C) confirmed that DNA cleavage in the presence of hCPT was much higher than with CPT.

Quantitative improvement in terms of antiproliferative activity was also observed on a panel of cell lines representing various tissues and histological types. The IC₅₀s in Table 1 show that higher cytotoxicity against tumor cells is seen for hCPT versus CPT, except for the MCF7r line, where the IC₅₀s are comparable. In addition, the conserved, or improved, activity of hCPT on MCF7r and K562r suggests that CPT homologation does not lead to recognition by the multidrug-resistance phenotypes of these cells, a known advantage of several CPT analogues (16, 17).

The remaining point to be discussed concerns the in vivo results. The drugs were tested by i.p. administration in two different in vivo models, using racemic dl-hCPT on L1210 murine leukemia implanted by the i.p route, or enantiomerically pure hCPT on HT29 human colon adenocarcinoma xenografted s.c. In the L1210 model, dl-hCPT showed higher efficacy and was significantly active on a wider range of doses than CPT (Fig. 5A). For both drugs, the efficacy increased with the duration of the treatment at the doses of 0.64 and 1.25 mg/kg (Fig. 5,B), but the impressive number of long-term survivors (T/C ≥ 300%) obtained in six different groups demonstrates that dl-hCPT is more efficacious than CPT and well tolerated over a long period of administration. This trend is verified in the HT-29 xenograft model in which the dosing schedule of the 4-days-on/3-days-off cycle was repeated three times. The range of tested doses typically followed a geometrical progression with a factor of 2, until reaching an aproximate MTD. As shown in Fig. 6, the approximate MTDs were 1.25 mg/kg for hCPT and 0.625 mg/kg for CPT, again indicative of a better tolerance for hCPT. In addition, hCPT inhibited tumor growth to a considerably greater extent than did CPT.

In conclusion, the above results provide, for the first time, evidence that the highly reactive α-hydroxylactone ring of CPTs is not an absolute requirement for good Topo I-mediated activity. Indeed, homologation of the E-ring gives a more potent Topo I poison than CPT and exhibits high in vitro cytotoxicities and good in vivo efficacy. This novel E-ring modification also provides a more stable compound that hydrolyses less rapidly than the classical α-hydroxylactone of CPTs, and the resulting ring-opened compound is an independent, isolable, chemical entity that does not spontaneously recyclize. Although a potential ring-opening reaction with covalent binding to the enzyme is still possible, the lower reactivity of hCPT may have some mechanistic implications. Furthermore, it should influence the pharmacokinetics and pharmacodynamics of the drugs. Thus, hCPT represents not only an interesting pharmacological tool but also a promising template for the preparation of improved anticancer agents.

                  

We thank Jacques Pommier for the separation and purification of hCPT enantiomers; Nicole Muller, Nicole Baroggi, and Gilles Mario for expert assistance with HPLC analysis of plasma samples; Grégoire Prévost and colleagues for the cellular growth assays; and Mark Carlson and Jeffrey Lauer for participation in performing the xenograft experiments.