A Phase I/II Study of LY900003, an Antisense Inhibitor of Protein Kinase C-α, in Combination with Cisplatin and Gemcitabine in Patients with Advanced Non–Small Cell Lung Cancer

Of the approximately 175,000 patients in the United States newly diagnosed with lung cancer each year, roughly 85% will be classified as having non–small-cell lung cancer [NSCLC (1)]. The prognosis for patients with unresectable disease (most patients with stage IIIB tumors and all patients with stage IV disease) is particularly poor. Systemic chemotherapy with platinum-containing regimens has been shown to produce objective response rates that are higher than those achieved with single-agent chemotherapy and modest improvements in survival (2). The advent of a newer generation of chemotherapy drugs has resulted in an additional increase in the number of responses observed and in 1-year survival (3, 4, 5, 6, 7, 8). Unfortunately, antitumor responses are short-lasting, and only a minority of patients survive longer than a year (7). Therefore, new agents/strategies are needed for the treatment of this disease.

Antisense technology has been identified as a promising strategy for the design of drugs that are highly selective against a specific target (9, 10). Antisense oligonucleotides bind specifically to a region of their target mRNA, resulting in inhibition of the expression of a single member of a gene family without affecting others. This specificity provides a clear advantage in terms of minimizing toxicity.

Protein kinase C (PKC)-α, a member of a family of phospholipid-dependent, cytoplasmic serine-threonine kinases, has increased expression in tumor compared with normal tissue and has been implicated in malignant transformation and proliferation (11, 12, 13, 14). Moreover, inhibition of PKC-α has been shown to arrest the growth of prostate cancer, hepatoma, and medulloblastoma cell lines (15, 16, 17). LY900003 (Affinitak, ISIS 3521) is the sodium salt of a 20-mer phosphorothioate oligonucleotide that hybridizes to the 3′-untranslated region of human PKC-α mRNA and inhibits its expression through Rnase-mediated cleavage of hybridized PKC-α mRNA (18). The phosphorothioate structure of LY900003 (a sulfur substitution of a nonbridging oxygen on the backbone) confers increased resistance to exonuclease- and endonuclease-mediated degradation, resulting in a stable heteroduplex with its target mRNA (19). Selective sequence- and concentration-dependent inhibition of PKC-α expression by LY900003 has been demonstrated in vitro and in vivo (18, 20). LY900003 had growth-inhibitory activity as a single agent or in combination with cisplatin or paclitaxel (additive) in xenograft models.¹³

After administration, LY900003 undergoes a biphasic pattern of elimination with extensive and rapid uptake into tissues and a prolonged elimination half-life (21, 22). Toxicology studies in rodents and primates revealed complement activation and prolongation of activated partial thromboplastin time, effects that are correlated with higher plasma concentrations (23, 24). Because continuous intravenous infusions are associated with greater uptake into tissues and prolonged inhibition of PKC-α mRNA expression while minimizing plasma concentrations, protracted infusions have been selected as the best schedule for the development of this agent. Fatigue and thrombocytopenia are dose-limiting at a dose of 3.0 mg/kg/d over 21 days (25). Therefore, a dose of 2 mg/kg/d in this schedule has been recommended for combination studies with this agent.

Based on the in vivo superadditive interaction of LY900003 and cisplatin, the lack of antagonism with gemcitabine, and the encouraging antitumor activity observed for LY900003 given in combination with paclitaxel and carboplatin in a phase II study in NSCLC patients (26), we designed this phase I/II study of LY900003 in combination with cisplatin and gemcitabine. This study had as its main objectives the evaluation of the safety and pharmacokinetic profile of the combination in patients with advanced solid malignancies and the determination of the antitumor activity of the combination in patients with advanced NSCLC.

The study was conducted in two parts. Part 1 was an abbreviated phase I study to assess the safety of LY900003 in combination with commonly used doses of gemcitabine and cisplatin in patients with advanced solid malignancies. Part 2 was a “minimax-design” phase II evaluation of the antitumor activity of the triple drug combination in chemotherapy-naïve NSCLC patients. The starting doses were 2 mg/kg/d of LY900003 administered in a 14-day continuous intravenous infusion, gemcitabine (1,000 mg/m²) as a 30-minute intravenous infusion on days 1 and 8 of each cycle, and cisplatin (80 mg/m²) as a 60-minute intravenous infusion on day 1 of each cycle. Treatment was repeated every 3 weeks.

The first cohort of patients was composed of three to six patients receiving the initial dose of the combination. If none of the first three or one of six patients developed dose-limiting toxicity (DLT), an increment on the starting doses of gemcitabine and cisplatin to 1,250 and 100 mg/m², respectively, was to be performed in a subsequent cohort of three to six patients. Intrasubject dose escalation was not permitted. The recommended dose for part 2 of the study was the dose at which no more than one of six patients develops DLT. DLT was defined as one of the following: (a) an absolute neutrophil count of <500/μL for >7 days; (b) a platelet count of <10,000/μL; or (c) any grade 3 to 4 nonhematologic toxicity (other than nausea, vomiting, anorexia, stomatitis, esophagitis/dysphagia, or fever without infection) judged to be at least possibly related to the experimental drug combination.

Patients with histologically confirmed advanced solid malignancies for which treatment with cisplatin and gemcitabine was considered appropriate or those who had no effective conventional treatment option were eligible for the phase I portion of the study. The phase II portion required histologic or cytological diagnosis of NSCLC and either stage IV disease or stage IIIB disease due to pleural or pericardial effusion. Eligibility criteria also included the following: (a) age ≥ 18 years; (b) Eastern Cooperative Oncology Group performance status of 0 or 1; (c) adequate hematopoietic (absolute neutrophil count of ≥1,500/μL, platelets ≥ 140,000/μL, hemoglobin level of ≥9.0 g/dL), hepatic [total serum bilirubin level of ≤1.5 mg/dL; aspartate aminotransferase and alkaline phosphatase < 3× the upper limit of normal (<5× the upper limit of normal in the presence of liver metastases)], and renal function (serum creatinine < 1.5 mg/dL); (d) no brain metastases or leptomeningeal disease, unless the lesions were localized, had been previously resected or irradiated, and were stable and asymptomatic for at least 4 weeks after completion of treatment; (e) no known active infectious process requiring therapy; and (f) no ongoing therapy with coumadin, heparin, low molecular weight heparin, aspirin, or nonsteroidal anti-inflammatory drugs, although prophylactic low-dose coumadin to maintain central venous catheter patency or cyclooxygenase 2-specific inhibitors were permitted. Up to one previous chemotherapy regimen not containing mitomycin C or nitrosureas was permitted for the phase I portion of the trial, whereas no prior chemotherapy was permitted for the phase II portion of the trial. The presence of measurable or evaluable lesions that had not previously been irradiated was required for participation in the phase II portion of the trial. Patients gave written informed consent according to federal and institutional guidelines before treatment.

LY900003 was supplied by ISIS Pharmaceuticals, Inc. (Carlsbad, CA) as a lyophilized powder containing 175 mg per vial. The vials were reconstituted in normal saline (0.9% sodium chloride injection to a concentration of 10 mg/mL in a volume of no greater than 250 mL and administered as a continuous intravenous infusion in 7-day duration bags for a total of 14 days of a 21-day treatment cycle. A portable volumetric infusion pump through a 0.22 μm inline filter, via an indwelling venous catheter, was used for the infusions. Commercially available gemcitabine (Gemzar; Eli Lilly and Co., Indianapolis, IN) and cisplatin were obtained from the hospital pharmacy. Gemcitabine is supplied in vials of either 200 mg or 1 g, formulated with mannitol and sodium acetate as a sterile lyophilized powder. It was administered at the assigned dose by 30-minute intravenous infusions on days 1 and 8 of each treatment cycle, preceding cisplatin administration. Cisplatin, after dilution in 0.9% NaCl, USP, or 5% dextrose in 0.9% NaCl, was administered by 60-minute intravenous infusions on day 1 of each treatment cycle. Cisplatin administration was preceded and followed by intravenous hydration according to individual institution guidelines.

Histories, physical examinations, and routine laboratory studies were performed pretreatment and preceding each course of treatment. Routine laboratory studies included serum electrolytes, chemistries, and complete blood cell counts with differential white cell counts. Complete blood cell counts were also performed weekly, and blood clotting times, urinalysis, pregnancy tests (when indicated), chest radiographies, and electrocardiograms were performed before initiating treatment. If patients developed toxicity manifested by grade 3 to 4 abnormalities in hematologic or biochemical laboratory parameters, the tests were repeated immediately and then repeated daily until the toxicity resolved. Tumors were measured after every other course, and treatment was continued in the absence of progressive disease or intolerable toxicity. Separate criteria for evaluation of response were used for the two portions of the trial. A complete response (CR) was defined in both cases as the disappearance of all target and nontarget lesions. A partial response (PR) required at least a 50% reduction in the sum of the products of the maximum perpendicular diameters of target lesions compared with pretreatment measures, and progressive disease required an increase of at least 25% in the sum of the products of the maximum perpendicular diameters of target lesions for the phase I portion, whereas a 30% decrease in the sum of the maximum diameters of the indicator lesions defined a PR in the phase II portion [Response Evaluation Criteria in Solid Tumors (RECIST) criteria; ref. 27]. A 20% increase in the sum of the longest diameters defined progressive disease on the phase II. Objective responses required confirmation by a subsequent response evaluation separated by at least 4 weeks in both portions of the trial.

To study the pharmacokinetic behavior of LY900003, gemcitabine, and cisplatin when given in combination, blood samples were obtained from indwelling venous catheters placed in the arm contralateral to the infusion in patients participating in the phase I portion of the trial. To assess for a possible influence of LY900003 in the pharmacokinetics of gemcitabine and cisplatin, sampling was performed in the presence or absence of LY900003 administration. To accomplish this, gemcitabine and cisplatin on the first day (day 1) of the first cycle were administered 1 hour apart, and the cisplatin was completed 4 hours before the start of the LY900003 infusion, whereas on day 8 and in subsequent cycles, the LY900003 infusion began concurrently with gemcitabine.

On cycle 1/day 1, blood samples were collected before the gemcitabine infusion; immediately before the end of the 30-minute gemcitabine infusion; and at 10, 30, and 60 minutes after the gemcitabine infusion was stopped, the time at which the cisplatin infusion started. Additional samples were obtained immediately before cisplatin end of infusion (EOI) and at 15, 30 and 60 minutes after cisplatin was stopped, as well as at 2 and 3 hours after cisplatin EOI. On the following day (day 2), one sample was obtained 24 hours after the beginning of the gemcitabine infusion. On cycle 1/day 8, blood samples were collected before the gemcitabine infusion; immediately before the end of the gemcitabine infusion; and at 10, 30, and 60 minutes and 2, 3 and 4 h after the gemcitabine infusion was stopped. On cycle 1/day 15, blood samples were collected immediately before the discontinuation of the LY900003 14-day infusion; at 30 minutes after discontinuation of infusion; and at 1, 1.5, 2, and 3 hours after discontinuation of LY900003. On cycle 2/day 1, blood samples were collected before the gemcitabine and LY900003 infusion; immediately before the end of cisplatin infusion (90 minutes from the gemcitabine EOI); at 15 and 30 minutes after cisplatin EOI; and at 1, 2, and 3 hours after the cisplatin infusion was stopped. One additional sample was obtained the following day (cycle 2/day 2) 24 hours after the LY900003 infusion was started.

Separate tubes were required for the collection of blood samples, depending on the drug to be analyzed. For LY900003, 2.0 mL of blood using a lavender top tube containing EDTA as an anticoagulant were obtained at each time. For cisplatin, 2.0 mL of blood were collected in green top tubes containing heparin as an anticoagulant at each time point. After mixing and centrifuging both kinds of tubes, all plasma was removed and placed into two polypropylene tubes and frozen at −20°C until analysis. For gemcitabine, 4.0 mL of blood were collected in a previously prepared heparinized tube containing tetrahydrouridine. After collection, the tube was cooled in ice water and centrifuged for 10 minutes at 2,000 rpm. The plasma was transferred into two polypropylene tubes and frozen at −70°C until analysis. Gemcitabine concentrations were assayed by a high-performance liquid chromatography method with a lower limit of quantitation (LLOQ) of 0.15 μg/mL (28), cisplatin concentrations were assayed through atomic absorption spectrophotometry (LLOQ, 0.05 μg/mL; ref. 29), and concentrations of LY900003 and its chain-shortened metabolite by capillary gel electrophoresis (LLOQ, 0.206 μg/mL) were assayed as described previously (22).

Pharmacokinetic parameter estimates in each patient for gemcitabine, cisplatin, LY900003, and major metabolites (N-1, N-2, N-3, and N-4) were derived from individual concentration-time data sets using noncompartmental methods and the program WinNonlin Professional Version 3.3 (Pharsight Corp., Mountain View, CA). Values for the area under the concentration-time curve (AUC) were calculated using the linear trapezoidal method with extrapolation of the curve to infinity (AUC_0-∞), and from time 0 to the last sampling time at which the concentration could be measured (AUC_0-t). The following additional pharmacokinetic parameters were also determined: maximum plasma concentration (C_max), time to maximum plasma concentration (t_max), and apparent half-life (t_1/2). Total clearance was estimated from dose (in mg) over AUC_0-∞. For gemcitabine, the apparent volume of distribution at steady-state (V_ss) was calculated as follows: Dose × AUMC∞/AUC^2∞ − T × Dose/2 × AUC∞, where AUMC_∞ is the area under the first moment of the concentration-time curve from the start of the infusion to infinite time, and T is the infusion time. For LY900003, the steady-state plasma concentration (C_ss) was determined in cycle 1 from the mean of samples collected on day 2 (18.5 hours after start of LY900003 infusion), day 8 (168, 168.5, 168.67, 169, 169.5, 170.5, 171.5, and 172.5 hours after start of infusion), and day 15 (336 h after start of infusion). For cycle 2, the C_ss value for each patient was determined from the mean of samples collected on day 1 (3, 4, and 5 hours after start of infusion) and day 2 (18.5 hours after start of infusion).

Descriptive statistics were used to summarize all pharmacokinetic parameters. Mean, SD, coefficient of variation, and minimum, maximum, and median values were determined for all parameters. In addition, geometric mean and geometric coefficient of variation were determined for C_max, AUC_0-t, and AUC_0-∞. The possibility of a clinically relevant pharmacokinetic interaction was evaluated by mixed effects models (containing fixed and random/repeated effects) generated within the bioequivalence module of WinNonlin Professional Version 3.3 and determining the least squares mean ratios and associated 90% confidence intervals (CIs) of selected pharmacokinetic parameters calculated for LY900003, gemcitabine, and cisplatin, as well as the P values for any observed differences in the calculated least squares means. Geometric least squares means were determined for gemcitabine and cisplatin pharmacokinetic parameters, whereas untransformed least squares means were determined for LY900003 C_ss values. Samples obtained on day 1 served for evaluation of pharmacokinetics of gemcitabine and cisplatin alone; day 8 samples served for evaluation of gemcitabine and LY900003 at steady-state plasma concentrations, day 15 samples served for evaluation of LY900003 alone, and cycle 2, day 1 samples served for evaluation of concurrent administration of LY900003 with gemcitabine and cisplatin. The hypothesis of no relevant interaction between LY900003 and gemcitabine/cisplatin was to be accepted if the 90% confidence limits for the contrast were included within the reference regions.

Descriptive statistics (mean ± SD) were used to summarize change in the sum of the products of the maximum perpendicular diameters (part 1) or the sum of maximum diameters (part 2) of indicator lesions as a function of dose level, treatment duration, and cancer type. The number and percentage of patients achieving a CR or PR were tabulated. Survival time was defined as the number of days from first LY900003 dosing to the date of death or the last known date to be alive. Survival data were expressed by Kaplan-Meier survival curves. Response duration was defined as the time from first LY900003 dosing to first documented tumor progression and was expressed as a median with 95% CIs for patients with CRs or PRs. Time to tumor progression was defined as the time from first LY900003 dosing to first documentation of tumor progression, and it was to be measured by Kaplan-Meier curves.

The sample size for part 1 was variable, with three to six patients to be enrolled in each dose cohort. For part 2, the sample size was based on an assumed 30% objective response rate for the combination of gemcitabine and cisplatin, which was the approximate response rate in the two randomized trials described in the drug labeling of gemcitabine (4, 30). If at least 8 responses were observed among the first 28 evaluable patients, an additional 11 patients (total, 39 patients) were to be treated, with the goal of observing a total of at least 16 responses among those 39 patients. Based on the minimax design of Simon (31), an improvement in response rate from 30% to 50% with the addition of LY900003 would be excluded with 90% power and an α of 0.10 if fewer than 8 of 28 or 16 of 39 responses were observed.

The phase I portion of this study was limited to two institutions. Seven patients, all enrolled into dose cohort 1 (LY900003, 2.0 mg/kg/d; gemcitabine, 1,000 mg/m²; cisplatin, 80 mg/m²), received a total of 18 cycles of the combination (median, 2 cycles; range, 1–6 cycles). Patient characteristics are listed in Table 1. Primary tumors in these patients included a variety of advanced solid malignancies, and two of the patients had received one prior chemotherapy regimen. One patient was withdrawn before completing the first cycle due to an unrelated adverse event and was considered nonevaluable for toxicity purposes.

DLT, as prospectively defined, was only observed in one of the six evaluable patients (grade 3 asthenia/fatigue) during the first cycle of therapy. Grade 2 asthenia/fatigue was also experienced by two other patients during subsequent treatment cycles. Myelosuppresion, although common, was of short duration and was not associated with hospitalizations or other complications. Overall, four patients (11 cycles) experienced grade 3 to 4 neutropenia, and five patients experienced grade 3 to 4 thrombocytopenia. Dose reductions for either gemcitabine (5 cycles) or cisplatin (1 cycle) were required in three patients, and 6 cycles were delayed. Because of the relatively high frequency of moderate to severe thrombocytopenia observed, dose escalation was not performed, and cisplatin (80 mg/m²) and gemcitabine (1,000 mg/m²) were the recommended doses to combine with LY900003 (2 mg/kg/d) during the phase II portion of the trial. Two patients experienced discernable antitumor activity, including a PR in a patient with stage IIIB NSCLC and a minor response (48% decrease in index lesions) in a patient with metastatic pancreatic carcinoma. In addition, prolonged stabilization lasting 5 months was observed in two other patients, including a patient with metastatic malignant mesothelioma and a patient with NSCLC.

Mean noncompartmental pharmacokinetic parameters for gemcitabine and cisplatin during days 1, 8, and 22 are depicted in Table 2. These parameters were consistent with those reported previously in the literature (28, 29, 32). In general, plasma concentrations of gemcitabine were similar on day 1 (before the start of LY900003 infusion) and day 8 (during the LY900003 infusion) and declined in an apparent multiexponential fashion over time. Comparison of geometric least squares means between day 1 and day 8 revealed no statistically significant differences. In addition, the mean ratios of the respective geometric means on study days 1 and 8 for each parameter of interest (AUC, clearance, C_max, t_1/2, t_max, and V_ss) were near the ideal of 100% (range, 68.9–127.3%), suggesting little or no effect of LY900003 on the pharmacokinetics of gemcitabine. Similarly, plasma concentrations of total cisplatin were equivalent in cycle 1 (in the absence of LY900003) and cycle 2 (60 minutes after the start of LY900003). Comparison of geometric least squares means between cycle 1 and 2 revealed no statistically significant differences, and the mean ratios of the geometric means for each parameter of interest in cycles 1 and 2 were near the ideal of 100% (range, 102–109.6%).

The pharmacokinetic parameters for LY900003 were consistent with previously conducted clinical trials using similar long-term infusion regimens (25). A relatively short mean plasma t_1/2 value (1 ± 0.33 hour) was consistent with a rapid attainment of steady state for LY900003 and total oligonucleotide (within 3–5 hours of the start of infusion). Clearance was 113 ± 28 mL/h/kg, and the AUC_0-∞ was 251 ± 91 μg-h/mL. Table 3 depicts a summary of LY900003 and total oligonucleotide steady-state plasma concentrations in the absence and presence of gemcitabine, cisplatin, or the combination of these drugs. Overall, the results obtained suggest that gemcitabine, cisplatin, or the combination of both agents had no readily apparent effect on the plasma pharmacokinetics of LY900003, and the magnitude of any changes is not likely to be clinically significant.

Forty-four patients, 39 of whom were evaluable for response, were enrolled in the phase II portion of the trial as per the minimax design discussed in Patients and Methods. An additional cohort of 11 patients was added at the end of the trial with the objective of testing the tolerability of LY900003 and cisplatin (at the same doses tested) in combination with gemcitabine (1,250 mg/m²) on days 1 and 8 (expanded cohort). Table 4 summarizes the characteristics of both groups of patients. A total of 186 cycles (median, 4 cycles; range, 1–8 cycles) were administered to the first 44 patients, and 43 cycles (median, 4 cycles; range, 1–6 cycles) were administered to the expanded cohort.

Table 5 lists the percentage of patients developing neutropenia, anemia, and thrombocytopenia at each dose tested. Overall, moderate to severe neutropenia was frequent. Grade 3 and 4 neutropenia occurred in 23% and 11% of patients, respectively, among the group treated at the 1000 mg/m² dose of gemcitabine. Nine percent and 27% of patients in the expanded cohort developed grade 3 and 4 neutropenia, respectively. However, the neutropenia was associated with fever or sepsis in only three episodes. There were no episodes of grade 4 anemia, and 14% of patients treated on the phase II trial developed grade 4 thrombocytopenia. Neither grade 4 thrombocytopenia nor grade 4 anemia was observed in the expanded cohort. The episodes of anemia or thrombocytopenia were seldom clinically significant, with no serious bleeding episodes. However, thrombocytopenia was involved in 93% of the 28 dose reductions (16 patients) of LY900003 needed in the trial. Thirteen patients, including 1 patient in the expanded cohort, received platelet transfusions, and 21 patients (including 2 patients from the expanded cohort) were given red blood cell transfusions.

The principal nonhematologic toxicities of the combination are listed according to severity in Table 5. The most common nonhematologic effects were nausea, vomiting, asthenia (fatigue and malaise), and cathether-related events.

Nausea and/or vomiting occurred frequently. However, these effects were generally mild to moderate; 48% and 14% of courses were associated with grade 2 and 3 nausea, respectively, and 27% and 16% of courses were associated with grade 2 and 3 vomiting, respectively. All patients received prophylactic antiemetic treatment orally with a 5HT3 serotonin antagonist and dexamethasone. Grade 3/4 asthenia/fatigue/malaise was reported by 25% of patients. Although the contribution of therapy to this constellation of symptoms is hard to establish in patients with advanced malignancies, the transient nature of the highest grade of this symptomatology in some patients may indicate a possible relationship. Peripheral neuropathy was generally mild, with grade 2 symptoms occurring in 2% of patients. Overall, these effects appear no different than those expected from treatment with cisplatin. No significant differences in the frequency of nonhematologic toxicities were observed in the patients in the expanded cohort, who received slightly higher doses of gemcitabine.

There were 41 catheter-related events in 30 patients (48%) from 11 centers through part 1 and 2 of the study. These events included venous thrombosis (n = 19), pulmonary embolism (n = 7), bacteremia/sepsis (n = 4), localized infections (n = 9), and erythema/pain (n = 2). The median number of treatment cycles for patients experiencing a catheter-related event was 3.9 (range, 1–8 cycles); thus, central line complications did not necessarily limit the duration of treatment. Only one event was fatal.

The median number of cycles administered was 4 (maximum, 8 cycles). Overall, 16 patients (26%) required dose reductions of LY900003, with 28 instances of dose reductions among these patients. The most common reason was thrombocytopenia, which was involved in all but two dose reductions. A total of 78 cycles were delayed. Whereas many of these delays were for toxicity, there were other reasons for the delays, such as patient or site scheduling conflicts. Eight patients died within 30 days of LY900003 administration throughout both parts of the study; four of these patients died due to disease progression, whereas in the other four patients, the predisposing causes of death were cerebrovascular infarction (n = 2), sepsis, and pulmonary embolism.

Relevant details pertaining to the antitumor effects of the combination in all patients who participated in the study are depicted in Table 6. Patients who received LY900003 and either completed 2 cycles of therapy or discontinued therapy due to disease progression before completing 2 cycles were considered evaluable for response. Fourteen of 39 evaluable patients in the original part 2 cohort had a major response to treatment (1 CR and 13 PRs), for a response rate of 36%. Two of nine evaluable patients in the expanded cohort experienced PRs, for a combined part 2 response rate of 33%. The complete responder was a 61-year-old female with stage IIIB (mediastinal lymphadenopathy and pleural effusion) poorly differentiated adenocarcinoma. Disappearance of all indicator and nonindicator lesions was detected at the end of cycle 2 and was still evident at the patient’s therapy discontinuation (after cycle 6). The patient remained disease-free until the 20 month follow-up visit, when a new ovarian mass was detected. Pathology of this mass revealed serous adenocarcinoma, clearly different from the original primary. Twenty-five patients (52%) had stable disease as their best response, whereas progressive disease was the best response for seven patients (15%).

The median overall duration of response was 7 months (95% CI, 4.2–7.8 months), and the median duration of stable disease was 4 months (95% CI, 3–5.5 months). The median time to progression for the entire group of 55 patients was 3.9 months (95% CI, 3.3–4.7 months), whereas the median time to progression for the 48 response-evaluable patients was 4.4 months (95% CI, 3.5–5.5 months). The median survival time for the entire group of 55 patients was 8.9 months (95% CI, 7 months to not reached) and 10.5 months for the 45 patients receiving ≥2 full cycles of treatment.

This study addressed the safety, pharmacology, and efficacy of the combination of the PKC-α inhibitor LY900003 with cisplatin and gemcitabine in a sequential phase I/II study in NSCLC patients. The rationale for the development of this study included the known role of PKC isoforms in malignant transformation and proliferation (11, 12, 13, 14), as well as in the regulation of the phosphatidylinositol 3′-kinase/Akt and MEK/extracellular signal-regulated kinase pathways (20, 33). These pathways have been found to be constitutively active in the majority of NSCLC cell lines and can promote cellular survival and resistance to chemotherapy (33). The combination with cisplatin and gemcitabine was therefore expected to augment the antitumor activity of this chemotherapy regimen in NSCLC patients.

Results of the phase I portion of this study showed an acceptable safety profile for the combination. With the exception of cathether-related complications, the toxicity was very similar to what is generally expected for the cisplatin/gemcitabine regimen. In addition, the results of the pharmacokinetic studies indicate that no interactions between the agents occur. Clinical benefit was observed in patients with mesothelioma, NSCLC, and pancreatic adenocarcinoma.

The phase II portion accrued a total of 55 patients, 11 of whom were treated at slightly higher doses of gemcitabine. The number of patients treated should be considered adequate to judge whether a regimen’s activity in NSCLC is promising enough for further evaluation in the randomized setting. The response rate was 33% (16 of 48 evaluable patients); the time to tumor progression was 4.3 months, and the overall survival for the entire group of 55 patients was 8.9 months. Although the observed response rate for the LY900003/cisplatin/gemcitabine combination is slightly better than the activity reported for the cisplatin and gemcitabine combination in some but not all large multicenter studies (4, 7), this activity was lower than the prospectively set response rate of interest. Our results, together with the lack of benefit of LY900003 in combination with paclitaxel/carboplatin in a recently reported randomized phase III trial (34), raise important questions regarding the role of this treatment in unselected NSCLC patients.

At this juncture, further development of LY900003 in NSCLC depends on the answer to a number of questions. How relevant a target is PKC-α in NSCLC? Furthermore, was the failure to demonstrate clinical benefit in these trials related to either a less than optimal target inhibition at the dose tested, the “unselected” nature of the patients in terms of PKC-α expression, or the use of a suboptimal sequence of administration (i.e., concurrent versus sequential administration)?

Available data suggest that the doses recommended for human administration result in adequate tissue exposure (20). Therefore, current dosing and schedule should be sufficient to reach and inhibit the intended target. Nonetheless, the percentage of NSCLC in which this target is critical for tumor survival is still unknown. It is also entirely possible, as it appears to be for other biologically targeted agents, that combination with chemotherapy may not be the best strategy to exploit the antitumor activity of this agent. For example, inhibition of a different PKC isoenzyme (PKC-δ) has been shown to be more effective in potentiating chemotherapy-induced apoptosis (33).

In summary, the combination of LY900003, cisplatin, and gemcitabine is well tolerated and had moderate activity in unselected NSCLC patients. Better characterization of subsets of patients most likely to benefit from this therapy is imperative for further development of this combination.