the New York University School of Medicine, New York, NY
    Dana-Farber Cancer Institute
    Massachusetts General Hospital, Boston, MA
    Cancer Therapy Evaluation Program, National Cancer Institute, National Institutes of Health, Bethesda, MD
    Millennium Pharmaceuticals Inc, Cambridge, MA

    ABSTRACT

    PURPOSE: We performed a phase I study of a day (D) 1 and D4 bortezomib administration once every 2 weeks to determine the recommended phase II dose and toxicity profile, and the extent of 20S proteasome inhibition obtained.

    PATIENTS AND METHODS: Patients with solid tumors or lymphomas were treated with bortezomib at 0.25 to 1.9 mg/m2 on D1 and D4, every 2 weeks. 20S proteasome levels in blood were assayed at baseline and at 1, 4, and 24 hours postdose in cycle 1.

    RESULTS: On this D1 and D4 every 2 weeks' schedule, dose-limiting toxicity (DLT) was evident at the 1.75 and 1.9 mg/m2 dose levels, most commonly in patients receiving individual total doses  3.0 mg. The main DLT was peripheral neuropathy evident at the higher doses and in patients previously exposed to neurotoxic agents. Other DLTs included diarrhea and fatigue; grade 3 thrombocytopenia was also noted. Reversible inhibition of 20S proteasome activity was dose dependent and best fit a total dose (mg) per fraction rather than mg/m2; 70% of baseline activity was inhibited by a dose of 3.0 to 3.5 mg given on D1 and on D4 every other week. Antitumor effects short of confirmed partial responses were observed in patients with melanoma, non–small-cell lung cancer, and renal cell carcinoma.

    CONCLUSION: Bortezomib (PS-341) is a novel antineoplastic agent that is well tolerated at doses not exceeding 3.0 mg (equivalent to 1.75 mg/m2), repeated on D1 and D4 every other week. This dose correlates with 70% inhibition of 20S proteasome activity. DLTs include neuropathy, fatigue, and diarrhea.

    INTRODUCTION

    The ubiquitin-mediated proteasome pathway regulates a group of intracellular proteins that govern cell cycle, tumor growth, and survival. This pathway is the principal mechanism of degradation for short-lived cellular regulatory proteins, including p53,1 cyclins and the cyclin-dependent kinase inhibitors p212 and p27,3 the estrogen receptor,4 and the inhibitor (IB) of nuclear transcription factor kappa B (NF-B).5 Central to the pathway is the 26S proteasome, an adenosine triphosphate–dependent, multicatalytic protease that selectively degrades polyubiquinated proteins. These proteins get marked for degradation by a multistep process that includes specific protein ligases. Inhibition of the 26S proteasome permits accumulation of substrate polyubiquinated proteins, while normal activity rapidly clears them from the cell. The key result of the 26S proteasome inhibition is the disruption of cell cycle checkpoints and apoptosis pathways, and thus offers a promising approach for the treament of malignancies.

    Malignant cells display greater susceptibility to proteasome inhibition than nonmalignant cells.6 Proteasome actions that are thought to be key to the malignant phenotype include: (1) degradation of IB leading to NF-B activation. NF-B mediates the transcription of proteins including vascular endothelial growth factor (VEGF) and the cell adhesion molecules (CAM) E-selectin, ICAM-1, and VCAM-1, implicated in angiogenesis and tumor metastasis in vivo, and regulates the inhibition of tumor necrosis factor–alpha (TNF) -mediated cell death.7 (2) Degradation of p53, p21, and p27 removing critical checkpoints from the cell cycle, especially those associated to DNA damage induced by chemotherapy or radiation; p53 induces G1 arrest through transcription of p21, which initiates DNA repair mechanisms and, when DNA repair is ineffective, initiates apoptosis.8

    Bortezomib, a dipeptidyl boronic acid analog, is a potent and specific (Ki  0.6 nmol/L) reversible proteasome inhibitor. Cells exposed to bortezomib accumulate in the G2-M phase of the cell cycle, and some undergo apoptosis.9,10 Among the multiple molecular mechanisms that may be implicated in bortezomib's interference with cell survival signals, only a small number of pathways have been studied and confirmed to date.6,11,12

    Animal toxicology demonstrated gastrointestinal and possibly cardiovascular dose-limiting toxicities (DLT), with a steep dose-response effect that correlated with the extent and duration of 20S proteasome in peripheral blood mononuclear cells (PBMCs). The 20S proteasome is a subunit of the 26S proteasome that is highly conserved across eukaryotes and forms a channel where proteolysis of proteins occurs following removal of their ubiquitin tag. Inhibition of the 20S subunit by bortezomib is dose-dependent and consistent across rat and cynomolgous monkey models, and thus rendering it the most promising laboratory marker of drug activity. DLT in animal species occurred at approximately 90% 20S proteasome inhibition.9

    Single-dose doses of 0.25 to 0.3 mg/kg were lethal to monkeys, as profound hypotension occurred within 4 to 8 hours of drug administration.13 Accordingly, phase I studies utilized a fractionated dose-schedule that would allow for the recovery of proteasome activity before any subsequent dosing, and initially closely monitored for acute cardiovascular changes.14-16 The assessment of proteasome inhibition has served as a guideline as to where one might expect the maximum-tolerated dose (MTD) but key tumor responses have been observed at doses lower that the MTD.

    Beginning in 1998, several phase I studies were launched by the pharmaceutical sponsor independently or together with the National Cancer Institute (Bethesda, MD). In humans, pharmacokinetics showed a short (< 15 minutes) initial distribution half-life (T1/2 ) and a large volume of distribution that was 15L/kg, indicative of rapid entry into the tissue compartment.17 Maximal inhibition of the 20S subunit of the 26S proteasome is seen in PBMCs 1 hour after administration of bortezomib, with full recovery within 48 to 72 hours.18 Given these observations, the current and other phase I studies used the extent of proteasome inhibition as a pharmacodynamic correlate to serve as a guideline where one might expect the MTD to occur. Data derived from the current study show clinical activity at doses lower than the MTD.

    The objectives of this study were to determine the DLT and recommended phase II dose (RPTD) of bortezomib given to patients with advanced solid tumors or lymphomas as an intravenous bolus twice weekly for 1 week followed by 1 week without therapy. At each dose level, bortezomib pharmacodynamics were evaluated by measuring the inhibition of 20S proteasome activity achieved in whole blood following dosing as compared with baseline. This pharmacodynamic end point was selected in lieu of pharmacokinetics of bortezomib that had been uninformative in animal models due to the plasma levels below 2 ng/mL, a large volume of distribution, and rapid clearance. The dose escalation steps were in part influenced by pharmacodynamic data predicting a gradual approach of up to 90% inhibition as the target maximum safe level of inhibition. Safety issues were a major consideration in the development of this pharmacodynamic approach, because this was the first compound of its class to be tested clinically, and information on the tolerance of even transient proteasome inhibition in humans was totally lacking.

    PATIENTS AND METHODS

    Patients 18 years old or older with histologically confirmed cancers were eligible after failure of standard therapy or if no other known effective treatment was available. Eastern Cooperative Oncolgoy Group (ECOG) performance status  2, hematologic parameters including absolute neutrophil count (ANC)  1,500/mm3 and platelets  100,000/mm3; adequate liver function defined as AST  2.5 x upper limit of normal, total bilirubin  1.5 mg/dL, and serum creatinine  1.5 mg/dL were required. Surgery within 2 weeks, chemotherapy or radiation therapy within 4 weeks, or chemotherapy with nitrosourea-containing regimens within 6 weeks of study entry was not allowed. Due to cardiovascular toxicities in animals, patients with electrocardiographic evidence of acute ischemia or significant conduction abnormality (bifascicular block, defined as left anterior hemiblock in the presence of right bunder branch block; second or third degree atrio-ventricular blocks) were excluded. The protocol was reviewed and approved by the Cancer Therapy Evaluation Program of the National Cancer Institute (Bethesda, MD), the Protocol Review and Monitoring Committees of the New York University Cancer Institute (New York, NY) and the Dana-Farber Cancer Institute (Boston, MA), and the institutional review boards of the respective institutions. All participating patients signed an informed consent form reviewed and approved by the Cancer Therapy Evaluation Program and the local institutional review boards.

    In this study, bortezomib was administered as an intravenous short infusion (5 minutes) on days 1 and 4 of a 14-day cycle using dosing on a body-surface area base. The starting dose level (0.25 mg/m2) and the per meter square dosing are part of the conventional Cancer Therapy Evaluation Program's initial approach to phase I studies with a new agent.19,20 This dosing schedule was selected on the basis of the pattern of 20S proteasome inhibition in animals; it was hypothesized that in humans, re-treatment on day 4 would allow full recovery of proteasome activity between doses, as seen in these models. Blood pressure assessments every 15 to 30 minutes for up to 6 hours after drug administration were carried out. Routine antiemetics were not administered, and routine use of hematologic growth factors was not permitted. Re-treatment required recovery from toxicity levels greater than grade 1 or dosing was omitted. Treatment was to be discontinued for unacceptable toxicity, failure to meet the eligibility criteria following a 2-week treatment delay, disease progression, or withdrawal of consent. Based on preclinical pharmacodynamic data and expected wide tissue distribution, dose levels to be assessed were to begin at 0.25, and be escalated to 0.5, 0.8, 1.0, 1.2, 1.45, 1.75, and 1.9 mg/m2 in cohorts of three patients per level. The 0.5 mg/m2 was omitted when parallel studies moved beyond this dose at the time that we completed the 0.25 mg/m2 dose level. Toxicity was assessed weekly using the National Cancer Institute Common Toxicity Criteria version 2.0. DLT was defined as grade 4 hematologic toxicity, febrile neutropenia, or grade 3 or 4 nonhematologic toxicity occurring during the first 2 cycles (28 days) of therapy. Complete blood counts, serum electrolytes levels, and liver function tests were performed weekly, and ECGs were performed before every 2 cycles (once every 4 weeks). Peripheral blood was collected for 20S proteasome activity on days 1, 2, 4, and 5 of cycle 1. Patients with cutaneous or subcutaneous lesions who gave consent had tumor biopsies performed for 20S proteasome activity, and immunohistochemical assessment of p27, p53, E2F-1, and cyclin E at baseline and on day 2 of cycle 1. Tumor response was assessed every 4 cycles (every 8 weeks) using Response Evaluation Criteria in Solid Tumors (RECIST) criteria. Three assessable patients were to be entered at each dose level, which was expanded to six patients if one of three patients experienced a DLT. The RPTD was defined as the highest dose level at which  one of six patients experienced a DLT. RPTD was eventually expanded to 13 patients to assess the risk of neuropathy in a cohort of six patients who had not received prior platinum or antitubulin agents, and who were free of neuropathy at baseline. A retrospective reanalysis of the toxicity and proteasome inhibition data was performed according to total dose administered in mg, and compared these data with those observed at dose levels in mg/m2. Using similar rules to define the RPTD in mg/m2, an RPTD in total mg per dose was suggested.

    Measurement of 20S Proteasome Inhibition in Blood and Tumor Tissue

    Heparinized blood was collected at baseline (two separate samples) and at 1, 4, and 24 hours following bortezomib administration on day 1 and day 4. Tumor samples were collected at baseline, 2 to 3 hours, and 24 hours after bortezomib administration in two patients with accessible tumors, with one patient sampled during two cycles of treatment. The tubes were centrifuged at 2,000 x g for 10 minutes, within 60 minutes of sample collection. The plasma was removed and the buffy-coat samples were resuspended in 1 mL of phosphate buffered saline, aliquoted into tubes and frozen at –20°C. Tumor samples were snap frozen in liquid nitrogen and stored at –80°C. The 20S proteasome inhibition was assessed by a fluorogenic substrate assay conducted at Millennium Pharmaceuticals (Cambridge, MA) that followed the validated ex vivo proteasome assay based on the chymotryptic:tryptic ratio method reported by Lightcap et al.18 Briefly, the % inhibition was calculated by the expression: 100 x (1–SpAI/SpAu) where SpAI is the chymotryptic specific activity of the 20S proteasome in the presence of an inhibitor such as bortezomib, and SpAu is the chymotryptic activity of the 20S proteasome in the absence of bortezomib. The two baseline blood samples were used for the SpAu calculation, whereas the single baseline sample was used for the tumor biopsies. The percent 20S proteasome inhibition for the 1-, 4-, and 24-hour sampling times were analyzed and averaged together for the day 1 and day 4 determinations for each patient at the three respective time points. Then the mean values were modeled as a function of the mg/m2 dose as well as the actual mg dose administered to each patient, using a nonlinear maximum time to effect (Emax) inhibitor model available in WINNonlin Professional version 4.1 (Pharsight Corp, Mountain View, CA).

    Methods for p27, p53, cyclin E, and E2F-1 Immunohistochemistry

    Immunohistochemical assessment of tumor levels of four proteasome-regulated proteins (p27, p53, E2F-1, and cyclin E) were prospectively planned. Serial tumor biopsies were planned in selected patients before bortezomib treatment, 2 to 3 hours after bortezomib treatment, and 24 hours after bortezomib treatment. The study was to be most informative in malignant lymphomas where loss of p27 correlates with prognosis.21 A hypothesis was advanced that reappearance of p27 would be associated with therapeutic response. It was uncertain how many patients with accessible tumors would agree to have surgery for the serial biopsies. Patients with accessible solid tumor lesions underwent incisional biopsy at baseline, at 2 to 3 hours, and at 24 hours after administration of bortezomib. Tumor specimens were preserved in 10% neutral buffered formalin, processed and paraffin-embedded according to standard histologic practice. Five-micrometer-thick tissue sections prepared onto charged glass slides were assessed immunohistochemically for the four proteasome-regulated markers using commercially available monoclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, CA; Pharmingen). The immunohistochemistry images for the pre- and postbortezomib treatment biopsies were captured at 60x magnification using an Olympus BX40 microscope and DP11 digital image system (Olympus; Melville, NY) as JPEG files. The intensity of immunohistochemical staining for each of the respective markers was quantified via a three-step procedure involving the following: (1) The analysis of the JPEG image with Adobe Photoshop software (San Jose, CA) to process the immunohistochemistry densities via the magic wand tool, extracting the stained tissue color,22 and converting the image to TIFF file format; (2) Kodak ID image analysis software (Kodak Imaging systems, Rochester, NY) was used for quantification of the pixel density employing an edge detection integration method; (3) The "summation of intensity values" from the Kodak ID density table were exported and calculated using Microsoft Excel software (Microsoft Corp, Redmond, WA).

    Methods for Assessment of Ubiquination of Proteins in PBMCs

    One sodium heparin cell preparation tube (CPT; Becton Dickinson, Franklin Lakes, NJ) was collected at each timepoint. The CPT was centrifuged within an hour at 1,650 x g for 20 minutes. The buffy coat was removed and washed twice with Dulbecco's Phosphate Buffered Saline (DPBS; HyClone, Logan, UT) with the PBMC cell pellet resuspended in 200 μL DPBS. Western Blot analysis was completed using 150 μL of cell suspension, as described previously.23 Membranes were exposed to ligand affinity-purified rabbit polyclonal antiubiquitin antibody; (generously provided by A. Haas, Medical College of Wisconsin, Milwaukee, WI) at 1 μg/mL. Secondary antibody exposure was with antirabbit IgG peroxidase-linked species-specific whole antibody from donkey (Amersham, Piscataway, NJ), at a dilution of 1:1,000 for 2 hours. Enhanced chemi-luminescence (ECL) exposure and image quantitation were performed as described previously,23 with the exception of the use of Kodak ID image analysis software (ver 4.0). The test of statistical significance used the Dunnett multiple comparison test employed in JMP statistical software (SAS institute Inc, Cary NC).

    Methods for Assessment of Bcl-2 Integrity in PBMCs

    Based on in vitro data that proteasome inhibition induces a phosphorylation of Bcl-2,11 and the cleavage of Bcl-2 can be detected in PBMCs, a limited monitoring of the Bcl-2 signal was conducted during the course of bortezomib treatment. Serial blood samples were probed by Western blot analyses for Bcl-2 at baseline and at 1, 4, and 24 hours following the dosing of bortezomib on day 1 and day 4 of the treatment cycle. Monoclonal anti–Bcl-2 antibody was obtained from Calbiochem (Cambridge, MA). The PBMC cells were treated as for the ubiquination assessments following methodology detailed in Methods for Assessment of Ubiquination of Proteins in PBMCs for the ubiquination assessments.

    RESULTS

    Forty patients were accrued between July 1999 and June 2001. The primary malignancies were colorectal carcinoma (nine patients), melanoma (seven patients), soft tissue sarcoma (six patients), renal cell carcinoma (four patients), adenocarcinoma of lung (three patients), and one each with tonsillar carcinoma, squamous cell esophageal carcinoma, gallbladder carcinoma, Ewing's sarcoma, hepatocellular carcinoma, rectal carcinoid, pancreatic carcinoma, prostate carcinoma, serous ovarian carcinoma, endometrial cancer, and cutaneous T-cell lymphoma. Patients' median age was 58.5 years (range, 25 to 78 years); 24 patients were men and 16 patients were women. Sixteen patients had a performance status (PS) of 0, 22 patients were PS 1, and two patients were PS 2. Twenty-two patients had received at least two lines of prior chemotherapy; only three patients were chemotherapy naive. Twenty-one patients had received prior radiation, and 31 patients had visceral disease.

    Drug administration is listed in Table 1. A total of 175 cycles (consisting of two doses) of therapy were administered, at a median of 4 cycles per patient (range < 1 to 15 cycles). The majority of patients had discontinued therapy because of progressive disease (29 patients); seven patients discontinued therapy because of adverse events (neuropathy,three patients; diarrhea, three patients; aspiration pneumonia, one patient), one patient died of a complication of malignancy while on study, and three patients elected to discontinue study therapy without experiencing either DLT or disease progression.

    All adverse events that were assessed as possibly, probably, or definitely related to bortezomib are listed in Table 2. No drug-related hypotensive events of any grade were observed, although one patient had a grade 3 hypotensive episode related to a non-neutropenic Staphylococcus epidermidis line infection (0.25 mg/m2 dose level). There was no clinically significant hematologic toxicity at any dose level. Fatigue was documented across all dose levels with the exception of dose level 3 (1.0 mg/m2). Four patients at 1.75 mg/m2 and 3 at the 1.9 mg/m2 dose levels showed grade 2 fatigue x 5, whereas one patient each at these dose levels show evidence of grade 3 fatigue.

    Sensory peripheral neuropathy (paresthesias, often of a burning quality, and mostly localized in the feet) was the principal DLT (Table 3). At the 1.9 mg/m2 dose level, dose-limiting neuropathy was observed in one of seven patients (baseline grade 1 toxicity, prior cisplatin, docetaxel, and vinorelbine), and another patient had grade 3 fatigue in the 28-day assessment period. A second patient (no baseline neuropathy, prior vincristine, and high-dose carboplatin) experienced grade 2 peripheral neuropathy in the first 28 days of therapy, progressing to grade 3 neuropathy during week 11; therapy had been discontinued on week 9 for disease progression. Although this second event did not meet the study definition of "dose-limiting" within the period of drug administration, when coupled with the pharmacodynamic data and subsequent course, it was so classified. The dose level of 1.75 mg/m2 was thus chosen as the clinical RPTD. Of 13 patients treated at the 1.75 mg/m2 dose level, only four had adverse events that were classified as dose limiting (nausea and diarrhea, one patient; diarrhea, one patient; fatigue one patient; neuropathy, one patient). The neuropathy was classified as grade 1 in the first 28 days of therapy, and progressed to grade 3 during week 5. The diarrhea was controlled by loperamide in the following cycle without dose reduction in one patient.

    On reanalysis of the toxicity data according to total mg per dose administered, 19 patients received doses of 3 mg or more (range, 3.0 to 4.2 mg). Among these patients are all three patients with dose-limiting neuropathy and seven patients with other DLTs. Only two of 21 patients at a dose lower than 3.0 mg experienced a dose-limiting event; one patient had an isolated grade 3 anemia at a dose of 2.6 mg, and one patient had grade 3 diarrhea at a dose of 2.0 mg. Using a total dose model, the recommended phase two dose is below 3.0 mg, and thus 2.6 mg with three patients having received this dose (none received doses from 2.7 to 2.9 mg).

    Further analysis of the neuropathy is provided in Table 3. Neuropathy of any grade was noted in a total of 11 patients; two of six patients at the 1.2 mg/m2 dose level, two of five patients at the 1.45 mg/m2 dose level, four of 13 patients at the 1.75 mg/m2 dose level, and three of seven patients at the 1.9 mg/m2 dose level. Onset of symptoms was generally rapid, occurring or progressing within 2 weeks of the first dose of bortezomib in four of the 11 patients (including two of the three patients with baseline grade 1 symptoms), and within 4 weeks in another two patients. The presence of neuropathy at baseline increased the risk of dose-limiting neurotoxicity, but there was no clear increase in risk from prior therapy with neurotoxic agents in the absence of baseline symptoms. At the RPTD, neuropathy was observed in two of six patients with no prior neurotoxic exposure, and two of seven patients who had received prior neurotoxic agents (cisplatin, carboplatin, oxaliplatin, vinblastine, and/or paclitaxel). The role of prior exposure to neurotoxic agents as a risk factor for neuropathy and its severity and reversibility should be further assessed in subsequent studies; only three patients treated at the RPTD received more than 8 weeks of therapy. The neuropathy symptomatically resembled a platinum or antitubulin sensory neuropathy; there was no suggestion of cold sensitivity as seen with oxaliplatin, and no motor symptoms were observed. Neuropathy persisted after discontinuation of bortezomib in several patients, and progressed after discontinuation of bortezomib in one patient.

    Diarrhea was the only other significant toxicity. Dose-limiting diarrhea was observed in one patient at the 1.2 mg/m2 dose level and in two of thirteen patients at the 1.75 mg/m2 dose level. In addition, four patients had grade 3 diarrhea occurring after the first 28 days of therapy (one patient at the 0.8 mg/m2 dose level, two at the 1.45 mg/m2 dose level, and one at the 1.75 mg/m2 dose level), and another nine patients, representing all dose levels, had grade 1 or 2 diarrhea. At the 1.75 mg/m2 dose level, six of 13 patients experienced diarrhea. The diarrhea resolved between cycles and dose reductions were avoided by the use of loperamide in subsequent cycles. Nevertheless, three patients discontinued therapy because of this toxicity. Aspiration pneumonias (fatal in one patient with brain metastases) occurred in two patients with declining performance status due to disease, which were probably unrelated to bortezomib.

    No complete or partial responses were observed according to RECIST criteria. Evidence of antitumor activity was seen in two patients with malignant melanoma. One patient with pulmonary metastases and previously radiated brain lesions had a partial response of 15 weeks' duration that was not confirmed because he died as a consequence of aspiration pneumonia before reassessment. The second patient had a partial response in pulmonary lesions but had stable disease in cutaneous lesions; the overall response was stable disease for 23 weeks. Time to treatment failure exceeded 10 weeks in eight additional patients; 14 to 20 weeks in three patients with soft tissue sarcoma, 16 to 21 weeks in two patients with renal cell carcinoma, and 10, 18, and 24 weeks in one patient each with rectal carcinoma, colorectal adenocarcinoma, and non–small-cell lung cancer, respectively. For three patients with soft tissue sarcoma, the time to treatment failure ranged from 14 to 20 weeks, and for two patients with renal cell carcinoma, 16 to 21 weeks. One patient each with rectal carcinoid, colorectal carcinoma, and non–small-cell lung cancer had time to treatment failures of 10, 18, and 24 weeks, respectively.

    Bortezomib inhibited 20S proteasome activity in a dose-dependent manner (in milligrams per meter squared, Fig 1A) for the assessments at 1, 4, and 24 hours on days 1 and 4 of cycle 1. The 1- and 4-hour inhibition data were best fit with a simple Emax inhibitory model (12.8% and 29.4% coefficient of variation [CV], respectively), whereas the 24-hour data showed a linear correlation (R2 = 0.64) with bortezomib dose. Bortezomib also inhibited 20S proteasome activity in a manner that were also fit to the actual dose administered to each patient (Figs 1B through 1D) for the assessments at 1, 4, and 24 hours. The 1- and 4-hour inhibition data showed better fits to Emax inhibitory model (12.9% and 18% CV, respectively), whereas the 24-hour data showed a similar moderate linear correlation (R2 = 0.61) with bortezomib dose as found with the dose/m2 24-hour correlation.24 The 20S proteasome inhibition was maximal at the first sampling point (1 hour) after drug administration; the doses of 1.2 mg/m2 (range, 1.8 to 2.7 mg), 1.45 mg/m2 (range, 2.2 to 3.22 mg), and 1.75 mg/m2 (range, 2.6 to 4.2 mg) produced proteasome inhibition of approximately 60, 65, and a range of 65% to 70%, respectively. The start of recovery of 20S proteasome activity was discernible at 4 hours, whereas at 24 hours, 50% recovery had occurred, and before redosing on day 4, proteasome activity had returned to baseline levels in all patients.

    Tumor samples were available for molecular correlates in only two patients; one with cutaneous melanoma and one patient with cutaneous T-cell lymphoma, both treated at the 1 mg/m2 dose level. Given the availability of data on only two patients, the results are not subject to any interpretation, although we did observe that the expression of p27 in the one patient with T-cell lymphoma was increased 27-fold from baseline at the 24-hour sample after bortezomib. The level of 20S proteasome inhibition in the three tumor samples was comparable with inhibition in whole blood (Table 4).

    Accumulation of Ubiquinated Proteins

    Based on the expected proteasome inhibition by bortezomib, it is expected that ubiquinated proteins would accumulate in cells. We conducted a limited assessment of protein ubiquination in four patients at the higher dose levels (1.75 mg/m2 and 1.9 mg/m2). Figure 2 contains a Western Blot showing the profile of ubiquinated proteins in PBMCs at 1-, 4-, and 24-hour sampling times. Peak accumulation of ubiquinated proteins was detected 4 hours after bortezomib dosing, with recovery to baseline levels before redosing on day 4. Figure 3 is a plot of the mean ± SEM determinations of summed ubiquination proteins at each time point (N = 4). Studies performed after day 4 dosing enhanced the day 1 findings. A statistically significant (P < .05) increase in ubiquination was found for the comparison of 4 hours at day 4 over the baseline values.

    Effects of Bortezomib on Bcl-2 Proteins

    Based on the preclinical studies,11 another marker of the effects of proteasome inhibition on the cell is an unusual cleavage pattern of Bcl-2 into 23kDa and 25kDa fragments; these cleavage products can be detected in peripheral blood mononuclear cells following in vitro exposure to bortezomib.11 The pattern of these cleavage products could be followed in peripheral blood mononuclear cells from four patients on dose level 6 treated with bortezomib (1.75 mg/m2; range, 2.6 to 4.2 mg). As shown in Figure 4, for a representative PBMC Western blot profile, the peak levels of the Bcl-2 and the Bcl-2 cleavage products occur 24 hours postdosing, both on days 1 and 4, and represent a significant change over baseline levels at 1, 4, and 24 hours (P < .05), both with respect to the 26kDa Bcl-2 and the 25kDa and 23kDa Bcl-2 cleavage products.

    DISCUSSION

    This was one of four dosing schedules of bortezomib assessed in parallel human phase I studies: (1) days 1, 8, 15, and 22 of a 35-day cycle14; (2) days 1, 4, 8, and 11 of a 21-day cycle15; (3) days 1, 4, 8, 11, 15, 18, 22, and 25 of a 42-day cycle16,25; and (4) days 1 and 4 of a 14-day cycle (our study's schedule). All of these studies had coordinated pharmacodynamic assessment of whole blood 20S proteasome activity and its inhibition using analogous timepoints. In our schedule, we reached a recommended phase II dose of 1.75 mg/m2/dose that achieves a 70% proteasome inhibition, and delivers 10.5 mg/m2 of bortezomib over a 6-week period. A similar dose-intensity was achieved over a 6-week period across two of the other schedules tested in phase I studies. The recently reported study by Papandreou et al,14 however, delivered half the mg/m2 over a 6-week period but also achieved doses of up to 2 mg/m2. All of the schedules achieved significant levels of proteasome inhibition (range, 65% to 70%) at their respective RPTDs. This schedule of day 1 and day 4 every 2 weeks, however, produced the highest RPTD (1.75 mg/m2/dose) of the different schedules tested. While the RPTD varied between schedules, 20S proteasome inhibition correlated well with dose per administration across studies. Although the specific major DLT varied from schedule to schedule, neuropathy, diarrhea, and fatigue are the predominant toxicities. The intermittent schedule in this trial may allow delivery of a greater overall dose per administration than more continuous schedules. In fact, this schedule may have allowed a better separation of acute from cumulative toxicities, and in this respect, our findings suggest that neuropathy may be related to peak levels rather than to cumulative dose.

    The current study points to a relationship between DLT and doses exceeding 2.9 mg. However, several questions need to be answered as one embarks in further development of this drug. Although an intermittent schedule is logical for an inhibitor of an essential function of all cells such as the proteasome and supported by preclinical data, one could carry forward several schedules for phase II clinical studies: weekly, twice weekly for 2 weeks, or twice weekly every other week (current schedule). In order to determine the schedule with the best therapeutic index, randomized phase II studies employing two diverse schedules of bortezomib should be considered. Randomized phase II designs may yield information on the relationships between peak levels and toxicities such as neuropathy, while also help to identify the preferable schedule in a pick-the-winner selection process.

    Peripheral neuropathy was not predicted as a significant toxicity by animal models, but was dose limiting in this study. The mechanism is unclear, but it has been suggested that proteasome inhibition causes accumulation of oxidized proteins, thus increasing neural cell susceptibility to oxidative injury. In vitro, heat shock proteins partially abrogate this effect by facilitating proteasome activity.26 Other authors have found that TNF-, matrix metalloproteinases 2 and 9 (MMP-2 and MMP-9), interleukin-1 (IL-1), and interleukin-6 (IL-6) play a role in the development of peripheral neuropathies27,28; their role in the development of bortezomib-induced peripheral neuropathy should be pursued in future studies. Susceptibility to develop such neuropathy appeared to be partly related to dose and to prior treatments with neurotoxic agents. However, it may also appear de novo, and one should consider the presence of pre-existing subclinical neurologic diseases, such as Charcot-Marie-Tooth disease in some patients experiencing neuropathy in future studies.29

    Bortezomib demonstrated some antitumor activity in our study (objective regression of lung metastases in two patients with melanoma). Its clinical activity on a day 1-, 4-, 8-, and 11-day schedule was subsequently established and led to an indication for previously treated multiple myeloma.30 The molecular mechanisms underlying antiproliferative and apoptotic actions are clearly multifactorial, with only a small number of mechanisms explored to date. We attempted to study the consequences of bortezomib inhibition in serial tumor samples, but only one patient with melanoma and one with lymphoma qualified for such sampling. We had expected to obtain serial biopsies from patients enrolled onto this study, with the assumption that a large majority of patients would have the diagnosis of lymphoma thereby having easily accessible lymph nodes for serial biopsies. The correlative studies were designed to document changes in expression of p27, p53, E2F-1, and cyclin E by immunohistochemical assays. The enrollment of the trial, however, was dominated by solid tumor types, of which easily accessible tissue was only attainable in the one patient with melanoma. Thus, the reportable data are limited to these two patients, and we have little further to contribute about molecular mechanisms underlying its antitumor effect. We did observe proteasome inhibition in these samples that ranged from 50% to 59% at 24 hours in two separate melanoma biopsies, and inhibition was even higher in the cutaneous T-cell lymphoma (87% inhibition). The pattern of PBMC Bcl-2 staining and the Bcl-2 cleavage products that started at 4 hours and continued to persist strongly at 24 hours appears to correlate well with the increased amount of apoptosis seen in in vitro studies 24 to 48 hours after exposure to the drug.10-12

    The better fit to an Emax inhibitory model of the proteasome in peripheral blood with the actual dose administered to the patients has led us to conclude that a recommended phase II dose should not exceed 2.9 mg. Although the proteasome inhibition starts to recover already by 50% at 4 hours, by 72 hours no proteasome inhibition was discernible. In the phase I study by Orlowski et al25 of patients with refractory hematologic malignancies, which employed a 4-week twice-weekly dosing schedule, the 20S proteasome inhibition was also determined to be time- and dose-dependent, however, an incomplete return to normal proteasome inhibition before the next dose was observed, as opposed to the recovery we had observed across all dose levels at 72 hours in our study. Despite the lack of baseline recovery in this study, the 1-hour 20S proteasome inhibition was shown to be consistent across the 8 dosing days of the 4-week cycle of dosing. A lower ED50 concentration of 1.01 mg was fit to their 1-hour 20S proteasome data, as opposed to the higher ED50 estimate of 1.25 mg that was determined from our more conservative dosing schedule. This is probably due to the lower maximum tolerated dose of "their" more dose-intense study of 1.38 mg/m2, which resulted in a mean inhibition 74% ± 2% as opposed to the maximal dose of a maximum tolerated dose of 1.75 mg/m2 in our study that resulted in a mean 1 hour 20S proteasome inhibition of 69% ± 7%. Also, the maximal individual bortezomib doses given in our study were up to 4.2 mg, as opposed to the 2.5 mg given in the study by Orlowski et al.25

    Clear effects on protein ubiquination in PBMCs were still seen at 4 and 24 hours. The 24-hour values at day 4 (Fig. 4) may reflect some residual effect of the earlier exposure to bortezomib on this second drug administration. The pattern of Bcl-2 cleavage product starting at 4 hours and persisting strongly at 24 hours correlates well with the increased amount of apoptosis seen in in vitro studies that is elevated at time periods from 24 to 48 hours postexposure to bortezomib.10-12

    In conclusion, bortezomib (PS-341) is an antineoplastic compound with a novel mechanism of action. The drug can be safely administered as an intravenous bolus without premedication, and is generally very well tolerated. Limited evidence of antitumor activity was observed, most evidently in patients with melanoma. The RPTD in this schedule of day 1 and day 4 every 2 weeks was defined to be 1.75 mg/m2. However, reanalysis of toxicities in relation to actual dose administered suggests that a total mg per dose of 2.6 is an attractive dose for phase II study. In fact, three patients at this dose and all patients below this dose received the drug without dose-limiting neuropathy. In contrast, sensory peripheral neuropathy occurred in two of seven patients at the 1.9 mg/m2 dose level, and in three of 14 patients who received 3.0 mg per dose; caution should be exercised with bortezomib in patients with pre-existing peripheral neuropathy. Fatigue and diarrhea are the only other clinically significant toxicities observed. Proteasome inhibition in the tumor samples in the two patients paralleled the proteasome inhibition seen in blood. At the RPTD, 20S proteasome activity is reversibly inhibited by 70% in whole blood samples. With this unique mechanism of action and the lack of significant myelosuppression, one can anticipate an eventual role in cancer treatment for bortezomib and newer proteasome inhibitors. The rationale for drug combinations with non-neurotoxic drugs has been advanced and such trials are ongoing in phase I and phase II studies. It is also possible that potentially neurotoxic drugs might be safely combined with bortezomib without undue problems if the dose of these agents is kept below a threshold associated with a high probability of inducing sensory neuropathy. Additional progress in the integration of bortezomib in cancer therapeutics will also require additional mechanistic studies.

    Authors' Disclosures of Potential Conflicts of Interest

    The authors indicated no potential conflicts of interest.

    NOTES

    Supported by research grants from the National Cancer Institute: U01 CA76642, M01 RR00096, and P30 CA16087 (New York University) and U01 62490 (Dana-Farber Cancer Institute/Massachusetts General Hospital).

    Authors' disclosures of potential conflicts of interest are found at the end of this article.

    REFERENCES

    Nagata Y, Anan T, Yoshida T, et al: The stabilization mechanism of mutant-type p53 by impaired ubiquitination: The loss of wild-type p53 function and the hsp90 association. Oncogene 18:6037-6049, 1999

    An WG, Hwang SG, Trepel JB, et al: Protease inhibitor-induced apoptosis: Accumulation of wt p53, p21WAF1/CIP1, and induction of apoptosis are independent markers of proteasome inhibition. Leukemia 14:1276-1283, 2000

    Alessandrini A, Chiaur DS, Pagano M: Regulation of the cyclin-dependent kinase inhibitor p27 by degradation and phosphorylation. Leukemia 11:342-345, 1997

    Lonard DM, Nawaz Z, Smith CL, et al: The 26S proteasome is required for estrogen receptor-alpha and coactivator turnover and for efficient estrogen receptor-alpha transactivation. Mol Cell 5:939-948, 2000

    Palombella VJ, Conner EM, Fuseler JW, et al: Role of the proteasome and NF-kappaB in streptococcal cell wall-induced polyarthritis. Proc Natl Acad Sci U S A 95:15671-15676, 1998

    Voorhees PM, Dees EC, O'Neil B, et al: The proteasome as a target for cancer therapy. Clin Cancer Res 9:6316-6325, 2003

    Read MA, Neish AS, Luscinskas FW, et al: The proteasome pathway is required for cytokine-induced endothelial-leukocyte adhesion molecule expression. Immunity 2:493-506, 1995

    Lane DP: The regulation of p53 function: Steiner Award Lecture. Int J Cancer 57:623-627, 1994

    Adams J, Palombella VJ, Sausville EA, et al: Proteasome inhibitors: A novel class of potent and effective antitumor agents. Cancer Res 59:2615-2622, 1999

    Ling YH, Liebes L, Jiang J-D, et al: Mechanisms of proteasome inhibitor PS-341-induced G2-M-phase arrest and apoptosis in human non-small cell lung cancer cell lines. Clin Cancer Res 9:1145-1154, 2003

    Ling YH, Liebes L, Ng B, et al: PS-341, a novel proteasome inhibitor, induces Bcl-2 phosphorylation and cleavage in association with G2-M phase arrest and apoptosis. Mol Cancer Ther 1:841-849, 2002

    Ling YH, Liebes L, Zou Y, et al: Reactive oxygen species generation and mitochondrial dysfunction in the apoptotic response to Bortezomib, a novel proteasome inhibitor, in human H460 non-small cell lung cancer cells. J Biol Chem 278:33714-33723, 2003

    Adams J: Development of the proteasome inhibitor PS-341. The Oncologist 7:9-16, 2002

    Papandreou C, Daliani DD, Nix D, et al: Phase I trial of the proteasome inhibitor bortezomib in patients with advanced solid tumors with observations in androgen-independent prostate cancer. J Clin Oncol 22:2108-2121, 2004

    Aghajanian C, Soignet S, Dizon DS, et al: A phase I trial of the novel proteasome inhibitor PS-341 in advanced solid tumor malignancies. Clin Cancer Res 8:2505-2511, 2002

    Erlichman C, Adjei AA, Thomas JP, et al: A phase I trial of the proteasome inhibitor PS-341 in patients with advanced cancer. Proc Am Soc Clin Oncol 20:85a, 2001 (abstr 337)

    Nix D, Pien C, Newman R, et al: Clinical development of a proteasome inhibitor, PS-341, for the treatment of cancer. Proc Am Soc Clin Oncol 20:86a, 2001 (abstr 339)

    Lightcap ES, McCormack TA, Pien CS, et al: Proteasome inhibition measurements: Clinical application. Clin Chem 46:673-683, 2000

    Guarino A, Rozencweig M, Kline I, et al: Adequacies and inadequacies in assessing murine toxicity data with antineoplastic agents. Cancer Res 39:2204-2210, 1979

    Muggia F, Carter S, Macdonald J: The cancer therapy evaluation program of the National Cancer Institute. Semin Oncol 8:394-402, 1981

    Chiarle R, Budel LM, Skolnik J, et al: Increased proteasome degradation of cyclin-dependent kinase inhibitor p27 is associated with a decreased overall survival in mantle cell lymphoma. Blood 95:619-626, 2000

    Matkowskyi K, Glover S, Benya R: Quantitative Immunohistochemistry: A new algorithm measuring cumulative signal strength and receptor number. Microscopy and Analysis 18:5-6, 2004

    Liebes L, Potmesil M, Kim T, et al: Pharmacodynamics of topoisomerase I inhibition; western blot determination of topoisomerase I and cleavable complexes in patients with upper gastrointestinal malignancies treated with topotecan. Clin Cancer Res 4:545-557, 1998

    Holford NHG, Sheiner LB: Kinetics of pharmacologic response. Pharmacol Ther 16:143-166, 1982

    Orlowski RZ, Stinchcombe TE, Mitchell BS, et al: Phase I trial of the proteasome inhibitor PS-341 in patients with refractory hematologic malignancies. J Clin Oncol 20:4420-4427, 2002

    Ding Q, Keller JN: Proteasome inhibition in oxidative stress nerotoxicity: Implication for heat shock proteins. J Neurochem 77:1010-1017, 2001

    Cui J, Holmin S, Mathiesen T, et al: Possible role of inflammatory mediators in tactile hypersensitivity in rat models of mononeuropathy. Pain 88:239-248, 2000

    Murwani R, Armati P: Peripheral nerve fibroblasts as a source of IL-6, TNFalpha and IL-1 and their modulation by IFNgamma. J Neurol Sci 161:99-109, 1998

    Chauvenet AR, Shashi V, Selsky C, et al: Vincristine-induced neuropathy as the initial presentation of Charcot-Marie-Tooth disease in acute lymphoblastic leukemia: A Pediatric Oncology Group study. J Pediatr Hematol Oncol 25:316-320, 2003

    Richardson PG, Barlogie B, Berenson J, et al: A phase 2 study of bortezomib in relapsed, refractory myeloma. N Engl J Med 348:2609-2617, 2003

查询更多Patients相关信息在本站>>

《临床肿瘤学医学期刊》2005年9月第23卷第9期