Phase 1 Study of TLK286 (Telcyta) Administered Weekly in Advanced Malignancies

TLK286 {l-γ-glutamyl-3-[[2-[[bis[bis(2-chloroethyl)amino]phosphinyl]oxy]ethyl]sulfonyl]-l-alanyl-2-phenyl-(2R)glycine, hydrochloride salt}, a modified glutathione analog, is an investigational new prodrug in development for the treatment of cancer, with demonstrated single agent antitumor activity in Phase 2 studies in ovarian, nonsmall cell lung, breast, and colorectal cancers. TLK286 is designed to target the enzyme glutathione S-transferase (GST) P1–1 for activation (1, 2). The rationale for TLK286 is based on the observation that many common cancers have GST P1–1 levels two to four times those of matched normal tissues (3, 4, 5, 6, 7, 8, 9, 10, 11). The GSTs play a central role in drug detoxification and have been implicated in mediating cellular resistance to several classes of anticancer drugs (12). Levels of GST P1–1 have been shown to negatively correlate with prognosis in colorectal cancer, nonsmall cell lung cancer (NSCLC), ovarian cancer, gastric cancer, breast cancer, chronic lymphocytic leukemia, and gliomas. Studies have implicated GST P1–1 in the development of resistance to standard chemotherapy (13, 14, 15, 16, 17, 18).

TLK286 is metabolized by GST P1–1, releasing two electrophilic fragments, a tetrakis (chloroethyl) phosphorodiamidate and a glutathione vinyl sulfone (Fig. 1; Ref. 1, 2). Data suggest that the two electrophilic species cause a series of molecular events, including activation of the stress-response pathway that results in apoptosis (19). Cultured human cancer cells exposed to TLK286 undergo activation of the mitogen-activated protein (MAP) kinase, mitogen-activated kinase kinase 4, p38 MAP kinase, jun-NH₂-terminal kinase, and caspase 3, a process that leads to apoptosis, as evidenced by DNA fragmentation and the loss of membrane asymmetry. The proapoptotic and antitumor activities of TLK286 have been demonstrated in vitro against human cancer cell lines and in vivo in a number of xenograft tumor models, including those that have elevated levels of GST P1–1 as well as increased resistance to other anticancer agents (20, 21).

TLK286 was reported to be well tolerated when administered every 3 weeks (22, 23, 24). Principal toxicities included grade 1 or 2 nausea and vomiting, well controlled with standard antiemetics, fatigue and anemia. TLK286 administered every 3 weeks and every week has demonstrated significant single agent activity in six Phase 2 studies with objective tumor responses by RECIST criteria and improved survival over that expected historically in NSCLC, ovarian, breast and colorectal cancer (25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36).

The current study was intended to determine whether TLK286 could be administered using a weekly dose schedule, to permit flexibility of dose schedule in the design of subsequent single agent and combination studies. In addition, the more intensive weekly dose might be useful in treating earlier stage cancers in which dose density may be important. The principal objectives of this Phase 1 study in patients with advanced chemotherapy-resistant solid malignancies were to (a) determine the safety profile of TLK286 administered weekly, (b) determine the maximum-tolerated dose (MTD) of TLK286 administered weekly, (c) further characterize the pharmacokinetics of TLK286, and (d) assess indications of antitumor activity.

The study was approved by the Medical Human Subjects Committee at the University of California, Los Angeles. Written informed consent was obtained from all patients before entering the study. Patients with histologically confirmed advanced solid malignancies refractory to standard therapy, or for whom no effective therapy exists, were eligible. Other eligibility requirements were: age ≥18 years; Eastern Cooperative Oncology Group (ECOG) performance status ≤2; life expectancy ≥12 weeks; no chemotherapy, radiation therapy, immunotherapy, or investigational agents within 4 weeks before the study started (six weeks for nitrosoureas or mitomycin); acceptable hematopoietic function (absolute neutrophil count ≥1,500/mm³, platelet count ≥100,000/mm³, hemoglobin ≥9.0 g/dl); acceptable hepatic function (total bilirubin <2.0 mg/dl), acceptable renal function (creatinine <2.0 mg/dl), and aspartate transaminase/alanine aminotransferase concentrations ≤3.0 times the institutional upper limit of normal, unless caused by hepatic metastases, in which case elevations ≤5.0 times upper limit of normal were permitted. At least 2 weeks must have elapsed since major surgery, the administration of hematopoietic growth factors, or blood transfusion(s). Exclusion criteria included the following: pregnancy or lactation; symptomatic brain metastases; carcinomatous meningitis or hydrocephalus; uncontrolled infection; hematuria; concomitant malignancy within the last 5 years other than curatively treated carcinoma in situ of the uterine cervix or basal cell skin cancer; clinically significant hematological toxicity requiring a recovery period; and any coexisting medical problem of sufficient severity to limit full compliance with the study.

TLK286 for injection was supplied by Telik, Inc. (Palo Alto, CA) in vials containing 265 mg of sterile lyophilized drug product. The drug was reconstituted in water for injection, United States Pharmacopeia, to a concentration of 50 mg/ml, followed by dilution of the appropriate dose of TLK286 into 250 ml of 5% dextrose for injection, United States Pharmacopeia, and administered as a 30-minute constant-rate i.v. infusion in a peripheral vein weekly. A cycle of treatment was defined as 3 weekly treatments.

The starting dose of TLK286 was 60 mg/m² administered weekly, which was equivalent to one-third of the nontoxic single dose in the dog, the most sensitive species tested, and one-fortieth the nontoxic single dose in the rat. Subsequent dose escalations proceeded by a modified Fibonacci design at 120, 180, 300, 420, 540, 720, and 960 mg/m², with a minimum of three patients at each dose level. If a DLT was observed in one of the first three patients, three additional patients were entered at the same dose level. A DLT was defined as any TLK286-related grade 3 or greater nonhematological toxicity (excluding alopecia, nausea, vomiting, or diarrhea in the absence of optimal medical management) or any TLK286-related grade 4 hematological toxicity according to the WHO toxicity criteria scale during cycle 1. The MTD was defined as one dose below at which two patients of three to six patients treated at that dose level experienced a DLT. Intrapatient dose escalation was allowed, providing a higher dose was determined safe in a minimum of three patients.

TLK286 was administered weekly and a treatment cycle was defined as 3 weekly treatments or three weeks of therapy. Patients underwent tumor assessments on day 43 (i.e., after 2 cycles or 6 doses), and those patients who received clinical benefit (tumor regression or stabilization of disease) continued TLK286 treatment. Disease assessments were performed every 6 weeks. Patients remained on therapy until development of progressive disease, unacceptable toxicity, or withdrawal from the study. Dose modifications were specified in the protocol as follows: for TLK286-related grade 3 or 4 nonhematological toxicity, treatment was withheld until recovery to grade 1 or baseline, after which treatment could be reinstituted at the next lower dose level. For hematological toxicity, TLK286 treatment was withheld until recovery of absolute neutrophil count (ANC) to ≥1,500/mm³ and platelet count to ≥100,000/mm³. For nadir counts of ANC <1,000/mm³ or platelets <50,000/mm³, TLK286 treatment was resumed with a 25% dose reduction. ANC nadir counts <500/mm³ or platelets <25,000/mm³ required TLK286 to be resumed with a 50% dose reduction.

After obtaining written consent, and no more than 10 days before treatment initiation, all of the participants underwent a screening evaluation including complete medical history, physical examination with vital signs, and 12-lead electrocardiogram. Radiographic imaging studies required to assess the extent of disease were performed within 4 weeks of treatment initiation. Pretreatment laboratory evaluation included complete blood count with differential and platelets, serum chemistry profile, urinalysis, and pregnancy test (in women of childbearing potential). Laboratory and physical assessments with vital signs, recording of concomitant medications, and recording and grading of adverse events were assessed on day 1 (before infusion) and on days 2, 8, and 15 of each cycle.

Tumor responses were assessed by standard-response criteria as determined by the product of two perpendicular diameters of marker lesions applied at the widest portion of tumor. All of the measurable lesions were evaluated every 2 cycles (6 weeks). A complete response required the complete disappearance of all of the measurable and evaluable disease without the appearance of new lesions or disease-related symptoms for two measurements at least 4 weeks apart. A partial response (PR) required a ≥50% decrease from the baseline sum of the bi-dimensional products of the perpendicular diameters of all measurable lesions without the appearance of new lesions or the progression of evaluable disease documented by two measurements at least 4 weeks apart. Minor response was defined as the objective tumor reduction that did not meet the PR criteria. Progressive disease was defined as an increase of ≥25% in the sum of the products of the perpendicular diameters of all measured lesions, the appearance of any new lesion, or the reappearance of any lesion that had disappeared. Stable disease (SD) was defined as any measurements that did not meet the criteria for complete response, PR, or progressive disease.

Venous whole blood samples were obtained from an indwelling heparin lock catheter placed in the arm contralateral to the drug infusion site on day 1 and day 15 of cycle 1. Blood samples were processed, stored, and TLK286 blood levels were determined as described previously (23). Samples were obtained before the infusion, and 30 min, 1 h, and 4 h after the start of the infusion on day 1 and before the infusion on day 15 of cycle 1. Urine was collected over intervals of predose, 0–2, 2–4, and 4–6 h after the start of the infusion; urine volume was recorded.

Blood and urine concentrations of TLK286 were determined by liquid chromatography tandem mass spectrometry methods. The data presented uses parametric techniques (mixed-effects modeling), based on pooling data from a larger database using both dose schedules (weekly and every 3 weeks) and two sampling periods (day 1 and day 15 of cycle 1) in each dose schedule. Pharmacokinetics characteristics were determined using population methods (nonlinear mixed effects modeling or NONMEM) to determine the structural model (the number of compartments and to establish linearity) and the covariate model (whether morphometric, physiological, or other factors impacted on the pharmacokinetics). The analysis was performed using the NONMEM software package (version 5; Globomax, Baltimore, MD); graphics were prepared using S-PLUS (Insightful Corp., Seattle, WA). Estimates were obtained or derived for the volume of the central compartment, volume of the second compartment (for two-compartment model), total blood clearance, intercompartmental clearance, apparent volume of distribution at steady state, area under the blood concentration versus time curve, elimination half-life, and distribution half-life. Volume estimates for population mean parameters (typical values), standard errors of parameters, and interindividual variability were obtained.

Thirty-seven patients (23 males, 14 females, median age 66 years, range 23–79, median ECOG status 1, range 0–2) with a broad range of advanced solid malignancies (12 colorectal, 4 renal, 3 sarcoma, 2 NSCLC, 2 head and neck, 10 other) were enrolled between August 2000 and March 2002. Baseline characteristics are summarized in Table 1. Cancer diagnosis and history of prior cancer therapies are summarized in Table 2. All patients had disease metastatic to one or more sites, with the most common sites of metastatic disease being lung (28, 76%), liver (19, 51%), and lymph nodes (29, 78%). All but two patients had received multiple prior chemotherapy regimens (median, 5; range, 1 to >8), and 52% had received prior radiation therapy. All 37 patients were evaluable for safety; 31 patients were evaluable for tumor response.

All of the patients enrolled received TLK286 treatment. Dose intensity and dose modifications (reductions and intrapatient escalations) are summarized in Table 3. Thirty-seven patients received 111 cycles or 333 doses of TLK286 at eight dose levels ranging from 60 mg/m²-960 mg/m² administered weekly. The median number of cycles/doses administered per patient was 2 cycles/6 doses (range 1–16 cycles/3–48 doses). At least 3 patients were treated at each dose level. Three cohorts (180 mg/m², 420 mg/m², and 720 mg/m²) were expanded to 4, 5, and 4 patients, respectively, because of early withdrawal for reasons other than DLTs. Dose escalation to 960 mg/m² was achieved without observing any DLT, and this cohort was expanded to 12 patients. None of the 12 additional patients in the 960 mg/m² cohort experienced a DLT.

Two patients had TLK286 doses escalated, based on evidence of tumor response. A patient with renal cell carcinoma was treated at 60 mg/m² for the 6 cycles, and his dose was then escalated to 180 mg/m² for cycles 7–10 and escalated again to 300 mg/m² for cycles 11–14. At the end of cycle 14, he was found to have grade 2 thrombocytopenia, and his TLK286 dose was reduced to 180 mg/m². He continued at this dose for 2 more cycles until disease progression, for a total of 16 cycles (48 doses) of therapy. A second patient with renal cell carcinoma was treated at 300 mg/m² TLK286 for the first 9 cycles (27 doses), after which his dose was escalated to 420 mg/m² for cycles 10–13 (12 doses), until discontinued because of disease progression for a total of 13 cycles (39 doses).

The most frequently occurring treatment-related toxicities were nausea (57%), vomiting (41%), fatigue (54%), and anemia (30%). Antiemetic prophylaxis therapy was not initiated until ≥720 mg/m² dose level. Nausea and vomiting were well controlled with standard antiemetic prophylaxis therapy. Three patients had hematological toxicities (neutropenia, leucopenia, thrombocytopenia, and/or anemia) requiring dose reductions (1 patient from 540 mg/m² to 420 mg/m²; 2 patients from 960 mg/m² to 720 mg/m²).

TLK286 was well tolerated in this chemotherapy-resistant population with advanced malignancies. There were no treatment-related deaths, and no patient discontinued study treatment because of a TLK286 toxicity. The incidence of TLK286 treatment related toxicities is reported in Tables 4 and 5. Treatment-related grade 3 toxicities occurred in only six patients (16%). In the 960 mg/m² dose group, one patient experienced grade 3 anemia and thrombocytopenia, a second experienced grade 3 fatigue, a third had grade 3 leukopenia, neutropenia, and hypokalemia, and a fourth experienced hypersensitivity reaction characterized by hypotension and fatigue. The hypersensitivity reaction in this patient was prevented on subsequent infusions with appropriate premedication prophylaxis regimen of dexamethasone, H₁ and H₂ blockers. These patients resumed TLK286 therapy after appropriate treatment of their adverse events. None of the twelve patients treated at 960 mg/m² had a dose-limiting toxicity.

Table 4 summarizes TLK286-related hematological toxicities. Hematological toxicities were generally mild or moderate (grade 1 or 2) and did not generally lead to dose delays, nor were they cumulative in most patients, as follows: anemia (30%), thrombocytopenia (19%), leucopenia (14%), neutropenia (11%), and leukocytosis (3%). Grade 3 neutropenia, leukopenia, thrombocytopenia, or anemia was reported in one patient. There were no grade 4 toxicities. The reported hematological toxicities were not clinically significant and in general did not limit or delay TLK286 treatment. After 5 months of weekly therapy, two patients receiving TLK286 at 540 mg/m² developed intermittent thrombocytopenia (grade 1) that required treatment delays until recovery to baseline values.

The most frequent TLK286-related nonhematological toxicities were nausea and vomiting, well controlled with standard antiemetics, as well as fatigue, decreased appetite, and transient microscopic hematuria (Table 5). These toxicities were mild to moderate, with no grade 4 TLK286-related toxicities reported. Only four patients (11%) experienced any grade 3 nonhematological toxicity. Prophylactic antiemetics were not required until doses of 720 mg/m² and above. Nausea was well controlled when prophylactic prochlorperazine and dexamethasone were used. The symptom complex of hematuria, dysuria, and urinary frequency that comprised the DLT in the initial Phase 1 study of TLK286 given once every 3 weeks at 1280 mg/m² was not observed at any dose in this study, except for grade 1 dysuria and urinary frequency that was transient. Ten patients were granted protocol exceptions allowing enrollment with baseline microscopic hematuria because of a variety of underlying conditions (e.g., renal stones and tumor invasion); none of these patients developed clinically significant hematuria or any exacerbation of their underlying microscopic hematuria during the study.

Although this was primarily a safety and pharmacokinetic study, 31 of 37 patients completed at least 2 cycles of treatment and were evaluable for tumor response (Table 6). Nine (29%) of 31 evaluable patients had SD or significant tumor regressions that did not meet PR by standard criteria and continued on TLK286 treatment beyond tumor assessment on day 43 until disease progression (Table 7). Twenty-two (71%) had progressive disease. Eight patients had evidence of minor tumor regression (≤30% of reduction in tumor area) of one or more of their measurable disease sites. Two of these patients (NSCLC and renal cell carcinoma) achieved maximum reduction of tumor area after 5 and 6 cycles of TLK286 treatment, respectively. Patients with SD had the following cancer types: colorectal (2), renal cell (2), malignant chordoma (1), NSCLC (1), thyroid (1), ovarian (1), and sarcoma (1). All patients with SD were heavily pretreated with multiple chemotherapy regimens (median, 4; range, 1–7). The median number of cycles of TLK286 in these patients was five (range 3–16). The median duration of SD was 5.5 months (range, 3–14.5).

The pharmacokinetic characteristics of TLK286 are well described by an optimized two-compartment model in which the structural parameters are not normalized for size, but the data fit well when the structural parameters are normalized for body surface area and body weight. In this model the clearance and V₂ (volume of distribution at steady state) decrease with increasing dose. A model in which the structural parameters are not normalized for body size is presented in Tables 8,9,10,11, because the common practice of dosing chemotherapeutics is based on body surface area. The decrease in blood clearance with increasing dose causes the dose-normalized area under the blood concentration versus time curve to increase with dose. The combined effect of dose-related decreases in both clearance and volume of distribution is that increasing doses increase distribution half-life but decrease elimination half-life.

Complete and accurately timed urine collections were obtained on 56 occasions from 42 patients. The fraction of TLK286 eliminated in urine was approximately 5% with doses of 60–180 mg/m², increasing to ∼20% with a dose of 960-1280 mg/m². There is no relationship between the administered dose and the renal clearance of TLK286. It is remarkable that there are no systemic differences between blood concentrations of TLK286 on two occasions within individuals and interindividual variability in the plasma concentration versus time profile for TLK286. The ratio of the observed-to-predicted blood concentrations for the optimal models (Fig. 2) demonstrate that observed plasma concentrations rarely deviate >2- to 3-fold from the typical values predicted for a patient of that size who is given that dose (i.e., the population value, adjusted for size and dose). A dosing regimen based on the typical individual therefore is likely to lead to only small (<2- to 3-fold) differences from the expected plasma concentrations.

Figs. 3 and 4 present the data that describe the relationship between TLK286 clearance and laboratory data quantifying renal and hepatic function. Fig. 3 presents the relationship between TLK286 clearance and renal function (measured by serum creatinine and creatinine clearance), and Fig. 4 presents the relationship between TLK286 clearance and hepatic function (measured by serum bilirubin, aspartase transaminase, and alkaline phosphatase). Impaired renal or hepatic function, as assessed by serum creatinine, creatinine clearance, bilirubin, aspartate transaminase, and alkaline phosphatase do not appear to influence the elimination clearance for TLK286. Patients in the present study, however, have only mild to moderate abnormalities in renal and hepatic function.

TLK286 is a novel glutathione analog prodrug designed to undergo activation by GST P1–1, an enzyme that is over-expressed in many human malignancies. Results have shown that brief exposure to TLK286 induce apoptosis in cultured human cancer cells mediated by induction of the stress-response pathway resulting in sequential activation of the MAP kinase mitogen-activated kinase kinase 4, p38 MAP kinase, the MAP kinase jun-NH₂-terminal kinase, and caspase 3 (19).

The patients in this Phase 1 study had a variety of advanced chemotherapy-resistant cancers. Thirty-five patients (95%) had received prior chemotherapy, and 19 patients (52%) had also received prior radiation therapy. TLK286 treatment was generally well tolerated by these advanced-stage, heavily pretreated patients. Dose escalation of TLK286 to 960 mg/m² was achieved without any DLTs. TLK286 administered weekly was not escalated to 1280 mg/m², because 1280 mg/m² had been established as the MTD for single-dose TLK286 in the initial Phase 1 study. The most frequent treatment-related toxicities were nausea, vomiting, fatigue, and anemia. Most of the reported adverse events were mild or moderate (grade 1 or 2). There were no grade 4 drug-related adverse events. Grade 3 adverse events occurred in six patients and were not considered DLTs. The MTD was defined at 960 mg/m² weekly.

A comparison of the frequency and severity of adverse events in the two TLK286 studies showed comparable safety. In both studies, TLK286 was well tolerated, without treatment-related deaths, grade 4 toxicities, clinically significant myelosuppression, or thrombocytopenia. Both schedules had similar side-effect profiles including type, frequency, and severity of adverse events, indicating that TLK286 toxicity was not schedule dependent.

TLK286 treatment-related hematological toxicities reported as adverse events in this study were mild or moderate (grade 1 or 2 only) in the majority of cases with no treatment delays, although study patients were heavily pretreated and expected to have limited bone marrow reserve. The reported hematological toxicities were not clinically significant and did not limit or delay treatment. Anemia was the most frequently reported hematological toxicity. Thrombocytopenia, leukopenia, and neutropenia were reported in <20% of patients. There were no TLK286-related grade 4 toxicities, and grade 3 toxicities were infrequent (16%). These results in Phase 1 clinical trial are consistent with preclinical safety studies, which indicated that TLK286 does not cause bone marrow suppression.

No DLTs were observed in this study. In the first Phase 1 Study, where TLK286 was administered once every 3 weeks, the MTD was established at 1280 mg/m² (22, 23, 24). The two patients with grade 3 bladder symptoms that led to the determination of MTD exhibited a constellation of bladder-related toxicities characterized by urinary frequency, dysuria, and hematuria. We recommend that for patients who experience a grade 2–3 cystitis, TLK286 treatment be delayed until there is complete resolution of symptoms: i.e., urinalysis is normalized and subsequent dose modification of TLK286 dose by 20% with attention to oral and/or i.v. hydration. Because acrolein is not a metabolite of TLK286, mesna is not recommended to prevent cystitis. This DLT was not seen in the present study in doses of 60 mg/m²-960 mg/m² of TLK286 administered weekly.

The pharmacokinetic characteristics of TLK286 are well described by an optimized two-compartment model. The interindividual variability in the plasma concentration versus time profile for TLK286 is unremarkable. A dosing regimen based on the pharmacokinetic parameters for the typical individual would be expected to produce a plasma concentration profile and exposure not markedly different from that predicted.

Comparison of results of the two treatment schedules allows several conclusions. The dose below the MTD in the initial Phase 1 study was 1,000 mg/m² administered once every 3 weeks, resulting in a dose intensity of 333 mg/m²/week, whereas in the current study, 960 mg/m² was administered weekly without DLTs, resulting in a dose intensity of 960 mg/m²/week. Weekly dosing of TLK286 therefore allowed for 3-fold greater dose density. This dose schedule permits flexibility in planning future drug development studies with weekly dosing in earlier stage patients or combination studies with standard drugs that are administered on a weekly dose schedule.

Because this Phase 1 study was designed to evaluate the safety and pharmacokinetic characteristics of TLK286, a formal efficacy analysis was not planned. However, TLK286 did show single agent antitumor activity with 9 patients (29%) achieving a best response of SD or minor tumor regressions that did not meet PR criteria. Prolonged dosing did not lead to cumulative toxicity and was well tolerated, suggesting that TLK286 might be appropriate for advanced patients who may need prolonged therapy for disease control and palliation of symptoms.

The safety, predictable pharmacokinetics and antitumor activity observed across a broad range of malignancies support Phase 2 disease-specific investigations of TLK286 administered at 960 mg/m²/weekly.