Investigational New Drugs 10:113-117, 1992. (cid:14)9 1992 Kluwer Academic Publishers. Printed in the Netherlands. Phase II clinical and pharmacological study of didemnin B in patients with metastatic breast cancer John A. Benvenuto 1, Robert A. Newman 1, Gary S. Bignami 2, T.J.G. Raybould 2, Martin N. Raber 1, Laura Esparza 1 and Ronald S. Walters 1 IDepartment of Medical Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, TX 77030, USA; 2Hawaii Biotechnology Group, Inc., Immunochemistry Division, Aiea, HI 96701 Key words: didemnin B, clinical trial, clinical pharmacology, enzyme immunoassay Summary Sixteen evaluable patients with metastatic breast cancer were entered into a phase II trial of didemnin B. They received the drug at an initial dose of 5.6 mg/m 2 every 21 to 28 days. Major toxicities noted were myalgia and nausea and vomiting while myelosuppression was mild. There were no complete responses; however, two minor responses were observed. The pharmacokinetics of didemnin B were studied in 10 patients who received the drug as 30 to 60 min i.v. infusions. A sensitive competitive inhibition enzyme immunoassay was used to quantitate didemnin B levels. Drug was observed to be rapidly cleared from plasma in a biphasic man- ner (tl(cid:1)89 = 0.12 hr, t(cid:1)89 = 4.8 hr). Although the assay could not identify the presence of specific metabo- lites, the increase of apparent didemnin B levels in plasma at later time points suggested the formation of unidentified metabolites which cross reacted with the antibody in the analytical procedure. In vitro experi- ments indicated that didemnin B was not bound to bovine serum albumin and only a minor portion (24%) of drug was found associated with red blood cells. Introduction Didemnin B is one of three structurally related cy- clic peptides isolated from a Caribbean tunicate [1]. The didemnins, of which didemnin B is the most ac- tive, have been shown to possess antiviral [1-3], antiproliferative [4], cytotoxic [1-3] and antitumor activities [2,3,5]. The mechanism of antitumor ac- tion of the didemnins is not known, but this family of compounds has been shown to inhibit protein synthesis and to a lesser extent DNA synthesis [6,7]. In addition, low doses of didemnin B can arrest progression of the cell cycle from G1 to S phase, although the drug can lead to death in all phases of the cell cycle. In previous clinical trials, toxicities of didemnin B included nausea and vomiting [8], elevations of SGOT [8] and neuromuscular toxicity [9] which was considered to be dose limiting. Because of a lack of a sensitive analytical assay, correlations of drug-mediated toxicities with pharmacokinetic parameters were not possible. To determine the clinical pharmacology of didemnin B, we developed a sensitive competitive enzyme immunoassay for this drug. The assay, which was capable of detect- ing didemnin B in plasma at a lower limit of quan- tification of 2 ng/ml, was used to investigate the pharmacokinetic parameters of this drug in patients with metastatic breast cancer. Materials and methods Patients considered eligible for this study signed in- formed consent forms and had failed no more than one regimen for recurrent disease. Patient charac-
`
`NPC02232466
`
`NOVARTIS EXHIBIT 2158
`Par v Novartis, IPR 2016-00084
`Page 1 of 5
`
`

`
`114 Table 1. Patient characteristics Number of patients entered 17 Evaluability status: Inevaluable 1 Response & toxicity 16 Age median (range) 56 (43-72) Performance status (Zubrod scale): 0 8 1 8 Histology: Carcinoma with sarcomatoid metaplasia 1 Duct carcinoma, invasive 14 Lobular carcinoma, invasive 1 Prior therapy: Chemotherapy 16 Hormone 8 Radiation 7 Surgery 16 Number of prior chemotherapy regimens: 1 14 2 2 Number of prior agents: 2 2 3 11 >3 3 teristics are listed in Table 1. Drug was adminis- tered in 30 to 60 min infusions and doses were either escalated or reduced based on incidence and/or severity of toxicity. Pretreatment antiemetic thera- py consisted of Decadron, Reglan and Benadryl. Reglan and Benadryl were also administered 3, 6, and 9 hr after administration of didemnin B. Plasma didemnin B pharmacokinetics were de- termined in 10 patients who received doses of 5.6 mg/m 2 (2 courses), 6.3 mg/m 2 (9 courses), and 7.5 mg/m 2 (2 courses). Blood was collected before in- fusion, at the end of infusion, and at 0.083, 0.25, 0.5, 1.0, 2.0, 4.0, 6.0, 8.0, 12.0, and 24.0 hr follow- ing infusion. Blood was centrifuged at 1000 x g to obtain plasma which was subsequently packed in dry ice and shipped to Hawaii Biotechnology Group for analysis. Didemnin B was analyzed by a competitive inhi- bition enzyme immunoassay; details of the assay have been published recently elsewhere [10]. This assay is capable of detecting didemnin B concentra- tions in plasma down to 2 ng/ml. BSA-coated microtiter plate wells treated with alkaline phos- phatase conjugate and substrate were included on each plate to measure background color develop- ment. The mean absorbance at 414 nm of these wells was subtracted from the mean absorbance at 414 of each set of controls and test replicates. Results were recorded as the means of triplicate analyses. Standard curves were constructed for each analy- sis using known concentrations of didemnin B. B/B 0 values were calculated by dividing the mean of a given set of replicates containing didemnin B inhibitor by the mean of all the wells containing no inhibitor. In vitro experiments were conducted by incubat- ing didemnin B (100 ng/ml) in whole human blood and in 4% BSA in PBS for 4 hr at 37 ~ The samples in whole blood were centrifuged to obtain plasma. The samples were frozen and shipped to Hawaii Bi- otechnology Group for analysis as described above. Plasma didemnin B concentrations were fitted to a two compartment model by nonlinear regression analysis using the RSTRIP Version 4.03 (Micro- Math, Inc., Salt Lake City, UT) microcomputer program. Pharmacokinetic parameters were calcu- lated using standard equations [11]; corrections were made for infusion times. Results There were no objective tumor responses observed in this study of the treatment of breast cancer with didemnin B although two minor responses were seen in soft tissue and lymph nodes. Grade 2-3 my- algias or motor deficits lasting from 5 to > 100 days were experienced by 5 of the 16 patients. All patients experienced nausea and vomiting and one had grade 2 diarrhea. Five courses of therapy were associated with grade 1 leukopenia of short dura- tion. Dose limiting toxicity included myalgias of grades 2 (2 patients) and grade 3 (1 patient) and grade 2 motor deficits in 2 patients. Electromyelo- graphic (EMG) studies were, however, normal in all symptomatic patients. A competitive inhibition enzyme immunoassay was used to quantitate didemnin B in plasma. The assay is based on didemnin A as hapten and is not specific for didemnin B. The plasma clearance of
`
`NPC02232467
`
`NOVARTIS EXHIBIT 2158
`Par v Novartis, IPR 2016-00084
`Page 2 of 5
`
`

`
`40011 i ,~176 ,0000ft 0 5 10 15 20 25 Time (hr) Fig. 1. Didemnin B plasma clearance in a patient administered 7.6 mg/m 2 of the drug. 14o I 12o I ~g ~ lOOb :> g so ~- i 60 40 g 20 0 0 2 4 6 8 I0 12 Time (hp) I i4 Fig. 2. Simulated didemnin B plasma clearance and possible metabolite formation based on feathering technique (see Con- clusions). Data are from a patient administered 5.6 mg/m 2. Plasma drug levels are indicated by ((cid:12)9 and the formation of possible metabolites are indicated by (*). didemnin B was observed to be biphasic (Figs. 1 and 2) with rapid distribution (t(cid:1)89 = 0.12 hr) and ter- minal (t(cid:1)89 = 4.8 hr) phases. The average phar- macokinetic parameters for the 13 courses were: AUC, 233 ng h/ml; Clt, 29.7 1/h/m2; and Vdss, 57.6 I/m 2 (Table 2). To obtain an estimate of the extent of protein binding and relative distribution in blood cells, drug was incubated for four hr with bovine serum albumin (prepared in phosphate buffered saline) at 37 ~ . Over 95~ of the added drug was detectable in the ultrafiltrate suggesting no appreciable protein binding. In addition, the four hr incubation of didemnin B in whole blood resulted in the detection of 76% of the drug, indicating that only 24~ of didemnin B was associated with blood cells. 115 Conclusions Figure 1 illustrates the extremely rapid decline of didemnin B during the initial phase of the plasma clearance curve. This precipitous decrease is fol- lowed by constant, or even increasing, low levels of drug. The rapid initial clearance may reflect seques- tration or metabolism of the drug, either of which may limit clinical effectiveness [5,8]. It has been conjectured, based on in vitro stem cell assays, that an AUC of 100-1000 ng h/mL must be achieved for clinical antitumor activity [12]. Our results indi- cate a mean AUC of 233 ng h/mL (range: 134- 462); higher than that expected for activity based on the in vitro studies. Since conjugates of didemnin A were used to produce immunogens to elicit anti- bodies to the ring that is common to the didemnins, it is possible that any didemnin B metabolites pos- sessing the ring structure may also contribute to this value. However, didemnin B is the most active didemnin known and any contribution from a ring containing metabolite would be expected to decrease the observed antitumor effect. Thus, for example, didemnin A, which has 1/10-1/20 [1,6] the cytotoxicity of didemnin B, is also detected by the assay and requires a hundred fold greater dose than didemnin B to achieve comparable antitumor activity against P388 leukemia in mice [1]. The metabolic inactivation of didemnin B is also suggested by animal studies in which didemnin B was only active by ip administration after ip inocu- lation of tumor [5]. The increase of apparent drug concentrations at later times observed in the majority of courses (9 of 13) we studied suggests possible metabolite formation; in the other in- stances collection times may not have been long enough to observe increases. Some of these curves could be dissected employing a feathering tech- nique that entailed using the last three concentra- tion-time points to generate a line from which the observed concentrations at earlier times were sub- tracted (the most outstanding example is shown in Fig. 2). This resulted in two components; the first of which may correspond to the elimination of didemnin B, while the second may indicate the ap- parent increase of a metabolite. These simulated drug and metabolite curves indi-
`
`NPC02232468
`
`NOVARTIS EXHIBIT 2158
`Par v Novartis, IPR 2016-00084
`Page 3 of 5
`
`

`
`116 Table 2. Pharmacokinetic parameters didemnin B Dose Ti t (cid:1)89 a t V2~3 AUCoo RT~ Clt Vdss (mg/m 2) (hr) (tl) (h) (ng h/mL) (h) (L/h/m 2) (L/m 2) 5.6 0.73 0.08 6.2 262.0 8.3 24.1 150.0 5.6 0.62 0.11 3.6 193.1 3.4 29.0 52.3 6.3 0.50 0.06 6.7 170.6 6.1 36.9 89.1 6.3 0.50 0.08 1.3 134.0 0.5 47.0 11.0 6.3 0.57 0.12 3.5 195.4 3.4 32.2 66.8 6.3 0.53 0.03 3.8 303.7 4.4 20.7 31.3 6.3 0.50 0.51 7.7 265.7 4.3 23.7 87.7 6.3 0.75 0.11 9.1 221.4 8.8 28.5 109.7 6.3 0.67 0.05 2.6 169.5 2.1 37.2 20.5 6.3 0.40 0.09 5.3 197.6 3.6 31.9 54.6 6.3 0.40 0.01 5.0 216.3 2.8 29.1 37.9 7.6 0.82 0.10 4.2 461.9 3.0 13.6 13.9 7.6 0.92 0.06 2.9 237.9 2.8 32.0 24.5 Median 0.12 4.8 233.0 4.1 29.7 57.6 _+ S.D. _+ 0.12 _+ 2.2 _+ 82.5 _+ 2.4 _+ 8.3 _+ 41.7 cate that plasma didemnin B concentrations de- crease at a much faster rate than apparent metabo- lite appears (Fig. 2). This disparity in rates suggests that formation of the putative metabolite is not the major clearance mechanism of didemnin B. The possibility exists that other highly degraded, ring opened metabolites are readily formed. These metabolites would not be detected by the assay and could not be expected to possess cytotoxicity simi- lar to didemnin B. The 'feathering' technique was not applicable to all the data, therefore the phar- macokinetic parameters presented in Table 2 are derived from the actual data points. The incubation of didemnin B with BSA in PBS revealed no ap- preciable protein binding. This result is in contrast to the report that diacetyldidemnin B is highly bound to human protein [13]. In addition, only 24% of didemnin B was found to be associated with human blood cells versus 45% of diacetyldidemnin B; the greatest uptake of diacetyldidemnin B was by lymphocytes [13]. Significantly, we observed no myelosuppression. These differences between didemnin B and its diacetyl derivative might be at- tributable to the greater lipophilicity of the diacetyl compound and the resulting increased protein bind- ing and ability to penetrate cell walls. Thus, the rapid plasma clearance of didemnin B associated with the possible formation of metabo- lites with less antitumor activity than didemnin B may limit the clinical effectiveness of the drug; there were no objective tumor responses achieved in this phase II clinical study of the treatment of breast cancer with didemnin B. Acknowledgements Supported in part by Contract No. N01-CM-07310 from the National Cancer Institute, National Insti- tutes of Health, Bethesda, MD. References 1. Rinehart KL Jr, Gloer JB, Hughes RG Jr, Renis HE, McGovern JP, Swynenberg EB, Stringfellow DA, Kuentzel SL, Li LH: Didemnins: Antiviral and antitumor depsipep- tides from a Caribbean tunicate. Science 212:933, 1981 2. Chun HG, Davies B, Hoth D, Suffness M, Plowman J, Flora K, Grieshaber C, Leyland-Jones B: Didemnin B The first marine compound entering clinical trials as an an- tineoplastic agent. Invest New Drugs 4:279, 1986 3. Rinehart KL Jr, Gloer JB, Wilson GR, Hughes RG Jr, Li LH, Renis HE, Mc Govern JP: Antiviral and antitumor compounds from tunicates. Fed Proc 42:86, 1983 4. LeGrue S J, Sheu T-L, Carson DD, Laidlaw JL, Sanduja SK: Inhibition of T-lymphocyte proliferation by the cyclic polypeptide didemnin B: No inhibition of lymphokine stimulation. Lymphokine Res 7:21, 1988
`
`NPC02232469
`
`NOVARTIS EXHIBIT 2158
`Par v Novartis, IPR 2016-00084
`Page 4 of 5
`
`

`
`117 5. Fimiani V: In vivo effect of didemnin B on two tumors of the rat. Oncology 44:42, 1987 6. Crampton SL, Adams EG, Kuentzel SL, Li LH, Badiner G, Bhuyan BK: Biochemical and cellular effects of didemnins A and B. Cancer Res 44:1796, 1984 7. Li LH, Timmins LG, Wallace TL, Krueger WC, Prairie MD, IM WB: Mechanism of action of didemnin B, a dep- sipeptide from the sea. Cancer Lett 23:279, 1984 8. Dorr FA, Kuhn JG, Phillips J, von Hoff DD: Phase I clini- cal and pharmacokinetic investigation of didemnin B, a cy- clic depsipeptide. Eur J Cancer Oncol 24:1699, 1988 9. Shin DM, Holoye PY, Murphy WK, Forman A, Papasozomenus S, Jones E, Raber MN, Hong WK: Phase I clinical trial of didemnin B (NSC-325319:DN-B) in non- small cell lung cancer (NSCLC); Neuromuscular toxicity is dose-limiting. Proc Amer Soc Clin Oncol 10:95, 1991 10. Raybould TJG, Grothaus PG, Simpson SB, Bignami GS, Lazo CB, Newman RA: An enzyme immunoassay for deter- mining plasma concentrations of didemnin B. J Clin Lab Analysis (in press) 11. Wagner JG: Linear pharmacokinetic equations allowing direct calculation of many needed pharmacokinetic parameters from the coefficients and exponents of polyex- ponential equations which have been fitted to the data. J Pharmacokin Biopharm 4:43, 1976 12. Jiang TL, Liu RH, Salmon SE: Antitumor activity of didemnin B in the human tumor stem cell assay. Cancer Chemother Pharmacol 11:1, 1983 13. Phillips JL, Schwartz R, yon Hoff DD: In vitro distribution of diacetyl didemnin B in human blood cells and plasma. Cancer Invest 7:123, 1989 Address for offprints: Department of Medical Oncology, Box 52, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030, USA
`
`NPC02232470
`
`NOVARTIS EXHIBIT 2158
`Par v Novartis, IPR 2016-00084
`Page 5 of 5