A Phase I Study of DMS612, a Novel Bifunctional Alkylating Agent

Benzaldehyde dimethane sulfonate (DMS612, NSC 281612, 4-[bis[2-[(methylsulfonyl)oxy]ethyl]amino]-2-methyl-benzaldehyde, BEN) was derived from a parent compound identified in the NCI-60 cell line Anticancer Drug Screen based on enhanced cytotoxicity against renal cell carcinoma (RCC; http://www.dtp.nci.nih.gov; ref. 1). The structure of DMS612 and preclinical evidence (2) suggests that this agent functions as a bifunctional alkylator with structural similarities to chlorambucil, busulfan, and melphalan (3). In contrast with these classical alkylating agents, DMS612 and related analogs were found to have additional activity against human RCC cell lines in the NCI-60 cell line screen (4). In vitro treatment of the human RCC cell line 109 and the breast cancer cell line MCF-7 with DMS612 caused cell-cycle arrest at the G₂–M and S-phase (4) and resulted in increased p53 expression, indicating DNA damage. In the NCI Yeast Anticancer Drug Screen, DMS612 and related compounds were selectively toxic against yeast strains with mutations in rad18 (postreplication DNA repair), rad52 (recombination repair), and rad14 (nucleoside excision repair). Bioinformatic COMPARE analysis—using Pearson correlation of cell line GI50 values for a matrix of agents and cell lines found that DMS612 and related compounds reside in a unique cluster that is distinct from traditional alkylating agents, such as chlorambucil, carmustine, and busulfan (4). Given these characteristics, DMS612 was selected for further clinical evaluation.

Preclinical in vivo toxicology studies in rats and beagle dogs determined that dose-limiting toxicities (DLT) were mainly hematologic (leukopenia, thrombocytopenia, and reduced reticulocyte counts) and gastrointestinal (diarrhea and nausea/vomiting). The MTD of DMS612 dosed weekly three times was between 12 mg/m²/dose and 24 mg/m²/dose (2–4 mg/kg/dose) in Fischer 344 rats and greater than 30 mg/m² in beagle dogs (1.5 mg/kg/dose).

DMS612 has demonstrated antitumor activity in in vivo xenograft models: DMS612 treatment in SCID female mice bearing human RCC RXF-393 xenografts (DMS612 Investigator Brochure), (5, 6) was able to produce tumor regressions at all doses and schedules studied; this antitumor activity was confirmed with additional xenograft models using orthotopic implantation of RCC lines ACHN-luc and 786-0, with greater activity seen against the latter (DMS612 Investigator Brochure).

In this first-in-human phase I study, we determined the DLT and MTD of DMS612 administered by 10-minute infusion on days 1, 8, and 15 of a 28-day cycle. We also characterized the pharmacokinetics (PK) of DMS612 and its active metabolites, and demonstrated pharmacodynamic (PD) evidence of induction of DNA damage response by quantification of γ-H2AX evolution using immunofluorescence in peripheral blood mononuclear cells (PBMC) and hair follicles at several time points during the first treatment.

This multicenter study (ClinicalTrials.gov Identifier: NCT00923520) was conducted at the NCI Clinical Center (Bethesda, MD), University of Pittsburgh Cancer Institute (Pittsburgh, PA), and Penn State Hershey Cancer Center (Hershey, PA), in accordance with the Declaration of Helsinki. The Institutional Review Boards at the respective institutions approved the study.

Eligible patients were ≥ 18 years of age with advanced solid tumors or lymphoma for which effective therapy did not exist or was no longer effective. There was no limit on prior chemotherapy treatment, although prior radiation to more than 25% of bone marrow was prohibited. Patients had to be ≥4 weeks from prior chemotherapy, monoclonal antibody therapy, or experimental therapy; ≥2 weeks from prior sorafenib, sunitinib, or temsirolimus treatment; or ≥6 weeks from prior mitomycin C or nitrosoureas. Patients were required to have acceptable organ and marrow function: leukocytes ≥3,000/μL, absolute neutrophil count ≥1,500/μL, platelets ≥100,000/μL total bilirubin within normal institutional limits, AST (SGOT) and ALT (SGPT) ≤2.5 X institutional upper limit of normal, creatinine within normal institutional limits, or creatinine clearance >50 mL/min per 1.73 m² for patients with creatinine levels above institutional normal. The Eastern Cooperative Oncology Group performance status was required to be 0-2 and life expectancy ≥3 months. Toxicities from prior treatment must have resolved to ≤ grade 1 by Common Terminology Criteria for Adverse Events (CTCAE) v.4. Concomitant inhibitors and inducers of CYP3A4 were prohibited. Patients with central nervous system metastases were excluded, unless control had been achieved with either radiation or surgical resection at least 6 months before enrollment. Patients with uncontrolled medical illness including myocardial infarction within the past 6 months were excluded.

DMS612 was supplied by the Division of Cancer Treatment and Diagnosis of the NCI in sterile, single-use vials containing a lyophilized powder/cake of 10 mg DMS612 and 1,000 mg hydroxypropyl-beta-cyclodextrin for reconstitution in 9.3 mL water for injection. After dilution in 0.9% sodium chloride or 5% dextrose, DMS612 was administered intravenously by central or peripheral catheter over 10 minutes on days 1, 8, and 15 of a 28-day cycle. There was no set limit on the number of cycles of study treatment administered. Dose escalation followed a modified Fibonacci schema starting at a dose of 1.5 mg/m². A standard 3+3 design was used for dose escalation. Adverse events were assessed using CTCAE version 4.0. Vital signs and laboratory safety assessments were performed before each study treatment, and history and physical exam were performed at the start of each cycle.

DLT was defined as febrile neutropenia, or grade 4 neutropenia without fever for >5 days; grade 4 thrombocytopenia or anemia; grades ≥3 nonhematologic toxicity except for nausea, vomiting, diarrhea, or electrolyte abnormality corrected within 48 hours, or elevated creatinine corrected within 24 hours. The MTD was defined as the highest dose at which ≤1 of 6 patients experienced a DLT.

Patients were evaluated for disease status by clinical and radiographic examination every two cycles (8 weeks) using RECIST 1.0 (7).

Blood samples (4 mL) were collected in heparinized tubes on day 1 of the first cycle of therapy. Samples were collected before infusion, 5 minutes into infusion, and then at 2, 5, 15, 30, 45, 60, 120, 180, 240, 360, 480, and 1,440 minutes (24 hours) after the end of the 10-minute infusion. The samples were processed, and plasma concentrations of DMS612 and its metabolites were determined as previously described (2). PK parameters were extracted from the data by noncompartmental methods with PK Solutions 2.0 (Summit Research Services).

Blood was obtained before the first dose of DMS612 and at 120 and 240 minutes, and 24 hours following infusion of DMS612. Detection of γ-H2AX in PBMCs was performed as previously described (8). Briefly, PBMCs were isolated by Ficoll gradient, washed in 1× PBS, fixed in 2% paraformaldehyde (PFA), and spotted on slides by cytospin. PBMCs were permeabilized with prechilled ethanol 70%. Slides were stored at 4°C overnight then blocked for 30 minutes with 5% BSA before incubating 2 hours with a mouse monoclonal anti-γ-H2AX antibody (Millipore) and then 1 hour with a goat anti-mouse Alexa-488–conjugated IgG (Invitrogen). Finally, slides were incubated at 37°C for 5 minutes with a solution containing RNAse A (0.5 mg/mL) and propidium iodide (PI; 5 μg/mL). Slides were then mounted with mounting medium containing PI (Vectashield, Vector Laboratories, Inc.) and sealed with nail polish.

Hair samples were taken to measure γ-H2AX levels before infusion start and at approximately 120 and 240 minutes and 24 hours after the end of drug infusion on day 1 of the study.

Single hairs were plucked from the scalp with forceps (8). The objective was to acquire 12 hair that contained a fully intact follicle and sheath. Plucked hair bulbs were placed in 1× PBS on ice before fixation for 20 minutes at room temperature with 2% PFA. Following three washes in PBS, plucked hair was imaged and hair bulbs from anagen hair were selected for γ-H2AX detection. Samples were then permeabilized with prechilled ethanol 70%, washed, and attached to slides using a nail polish drop. γ-H2AX detection was performed as described above but incubation times with blocking and antibody solutions were increased by 50%. Washes were performed by slide immersion in PBS. Incubation with the RNAse (0.5 mg/mL) and PI (5 μg/mL) solution was performed at 37°C for 30 minutes and slides were mounted as described above. Hair shafts were then detached from the bulbs by using a razor blade and hair bulbs were mounted with mounting medium containing PI.

Samples were imaged by laser scanning confocal microscopy (Nikon PCM 2000, Nikon, Inc.). Optical sections through PBMCs and hair bulbs were combined in a maximum projection using the Simple 32 software (Compix Inc.). The foci were counted in PBMCs using the FociCounter software (9) by analyzing at least 200 cells, while foci were visually quantified by eye in the extremity of the hair bulbs.

Thirty-one patients with advanced solid tumors were enrolled in this phase I study of DMS612 administered on days 1, 8, and 15 of 28-day cycles. Patient characteristics are presented in Table 1. Patients with a range of solid tumors were enrolled, including 4 with RCC and 11 with colorectal carcinoma. The median number of prior chemotherapy regimens was 4 (range 0–13).

Patients were treated at six dose levels: 1.5 mg/m² (n = 3), 3 mg/m² (n = 3), 5 mg/m² (n = 6), 7 mg/m² (n = 3), 9 mg/m² (n = 12), and 12 mg/m² (n = 3). The number of cycles of study treatment ranged from 0 to 6 (median 2). The most common reason for discontinuation of study treatment was disease progression.

The DLTs observed in cycle 1 of study treatment are summarized in Table 2. Myelosuppression with thrombocytopenia and neutropenia was the principal DLT. Two of 3 patients treated at the 12 mg/m² dose level experienced DLT: one developed grade 4 neutropenia and one had prolonged grade 3 thrombocytopenia. At the lower dose level of 9 mg/m², only 1 of 12 patients experienced a DLT. As a result, the MTD was determined to be 9 mg/m². Delayed and prolonged thrombocytopenia and neutropenia were observed in patients treated at the 9 and 12 mg/m² dose levels, respectively. One of the patients with prolonged thrombocytopenia had metastatic RCC (mRCC) and had not received prior cytotoxic chemotherapy, suggesting that this toxicity may not be limited to patients with significant prior marrow exposure. Measured hematologic parameters eventually recovered in all of these patients after suspension of treatment. A bone marrow biopsy in one patient with prolonged thrombocytopenia showed normal trilineage hematopoiesis with decreased cellularity, consistent with the known effects on hematopoiesis seen more broadly with alkylating agents. Drug-related toxicities of all grades observed in cycle 1 of study treatment are presented in Table 3. Gastrointestinal toxicities were minimal and transient, and routine anti-emetic prophylaxis was not utilized. As shown in Table 3, grade 1 nausea was observed in 6 of 30 patients, and grade 2 nausea in only one patient.

DMS612 was detected in only 2 patients (14.5 ng/mL midinfusion after 9 mg/m²; 4.7 ng/mL midinfusion after 12 mg/m²), and the half-life could not be determined in either of these patients. The PK parameters of a benzoic acid analog (BA) of DMS612, the primary enzymatic metabolite and putative alkylating agent; and BA-(OH)₂, the major inactive, and ultimate metabolite in the detoxification pathway, are presented in the Supplementary Files, and a representative example is shown in Fig. 1A. Metabolic intermediates and glucuronide metabolites are shown in Supplementary Files (Supplementary Figs. S1A and S1B). The exposure (AUC_0-t and C_max) of BA (Fig. 1B) increased with dose. Exposure to BA appeared to increase more than linearly with dose as apparent clearance (Cl/F) of BA decreased with dose (Supplementary Fig. S1C), while the dose-normalized C_max of BA increased with dose (Supplementary Fig. S1D).

An antibody-based immunofluorescence assay to detect the phosphorylated form of the histone variant H2AX referred to as γ-H2AX, was used in this study to evaluate DNA damage response in surrogate tissues during cycle 1 of treatment. Two different tissues were evaluated in eligible patients: PBMCs and plucked hair bulbs. These were examined for DMS612-induced DNA damage (Fig. 2), as indicated by the induction of γ-H2AX foci. At the lowest dose, 1.5 mg/m², no γ-H2AX foci formed at any time up to 24 hours. At 3 and 5 mg/m², some induction of γ-H2AX foci is apparent in PBMCs at 24 hours, while at 7, 9 and 12 mg/m², induction is apparent at 4 hours. γ-H2AX immunofluorescence at 24 hours was dose dependent, with the 9 and 12 mg/m² values being about 3-fold higher than in the lower doses.

Increases in γ-H2AX immunofluorescence were more pronounced in plucked hair bulbs than in PBMCs, appearing earlier and with more numerous foci (Fig. 2). For example, in PBMCs, the signal at 3 mg/m² was about 25% of that at 9 mg/m² after 24 hours, whereas in the hair bulbs, the γ-H2AX apparent at 3 mg/m² was 65% of that at 9 mg/m². Interestingly, DNA damage, as reflected by expression of the γ-H2AX biomarker, appeared to correlate with the incidence and severity of myelosuppression, supporting the hypothesis that DNA damage is biologically relevant to the mechanisms of action of DMS612.

One patient with RCC and one patient with cervical cancer had a confirmed partial response. Both patients were treated at the MTD of 9 mg/m² (Fig. 3). Stable disease of at least 16-week duration was observed in 3 additional patients [colon (2) and RCC (1)].

This is a first-in-human phase I study to evaluate the safety, DLT, MTD, and PK and PD of DMS612 administered weekly for 3 of every 4 weeks. In patients with previously treated solid tumors, the MTD of DMS612 was 9 mg/m² when administered on days 1, 8, and 15 of a 28-day cycle. DLTs of neutropenia and thrombocytopenia were consistent with the proposed mechanism of DMS612 as an alkylating agent. In some patients, myelosuppression was prolonged, but recovery of measured hematologic parameters was documented in every subject. While reversible myelosuppression was observed, toxicities such as alopecia and peripheral neuropathy were not, and gastrointestinal toxicity was relatively modest. When combined with the in vitro observations regarding the activity of DMS612 in renal cell cancer cell lines, these findings suggest that DMS612 is an alkylating agent with unique tissue specificities that remain presently unexplained and, therefore, make the agent an interesting and unique compound for further study

PK analysis revealed that DMS612 was rapidly metabolized as parent drug was detected in only 2 patients at the earliest time point: 2 minutes after completion of infusion. This result is consistent with preclinical murine studies in which DMS612 was detected only within the first 10 minutes of an intravenous bolus administration of drug, confirming the extremely rapid metabolism of DMS612 to BA. The BA C_max in humans of approximately 4,000 ng/mL after 12 mg/m² DMS612 is in good agreement with the C_max of 18,000 ng/mL observed in mice after 60 mg/m² (20 mg/kg; ref. 3), arguing for the validity of the murine model for DMS612 pharmacology. PK analysis identified a large interpatient variability in the concentration and AUC of BA, the major metabolite of DMS612 (2). BA was present in all but 2 patients; however, these patients did have downstream BA metabolites detected. The BA plasma half-life was several-fold longer than the 5 minute in vitro plasma half-life of this highly reactive compound, suggesting continued metabolic generation of BA from DMS612 in tissues, followed by redistribution into the plasma compartment. This finding is indirect evidence of the ability of DMS612 to reach tissues with subsequent intracellular activation to BA, the likely effector of DNA damage, as previously suggested (2). These observations are substantiated by the findings from γ-H2AX assays, which show that DMS612 leads to DNA damage induction in both blood cells and hair follicles. The capacity for DMS612 to generate DNA breaks in complex tissue (i.e., hair follicles) suggests that the drug most likely reaches and induces damage in solid tumors. The selective cytotoxicity of the benzaldehyde dimethane sulfonate compounds to yeast strains with specific defects in DNA repair suggests that the status of DNA repair pathways in tumor cells is critical for determining DMS612 activity (4).

All of the DMS612 metabolites and respective glucuronide forms that were detected in preclinical studies were also observed in patients, supporting the translation of previous preclinical findings related to DMS612 pharmacology. DMS612 metabolism differs from that of conventional alkylating agents in a number of ways. DMS612 is a lipophilic prodrug of BA; intracellularly generated, BA is a much more reactive analog and the likely alkylator of cellular targets. We previously demonstrated that aldehyde dehydrogenase (ALDH) may be responsible for the intracellular metabolism of DMS612 to BA, and that DMS612 is more active in cell lines that express high levels of ALDH (ref. 2; and manuscript in preparation). Therefore, DMS612 appears to have unique characteristics that are quite different from other well-characterized alkylating agents. The importance may lie in the reported overexpression of ALDH in stem cells of many tumor types making these stem cells potential targets for DMS612 therapy (10). The prolonged myelosuppression observed with DMS612 is suggestive of enhanced toxicity to bone marrow stem cells. Intriguingly, hematopoietic stem cells express high levels of ALDH (11), potentially facilitating the rapid conversion of DMS612 to BA, resulting in higher levels of intracellular alkylation and cytotoxicity. The short half-life of DMS612 might be a favorable property for use in antibody–drug conjugates, or in nanoparticle drug delivery systems. Such an approach might also circumvent some of the bone marrow toxicity.

One patient with prolonged myelosuppression underwent evaluation with bone marrow aspiration and biopsy, which revealed marked hypocellularity but with normal trilineage hematopoiesis, a finding consistent with the known effects of other well-established alkylating agents. An alternative schedule of drug administration on days 1 and 2 on an every 21-day schedule is currently being investigated in a phase I cohort of patients with advanced solid tumors. We hypothesize that the prolonged (20-day) interval of this modified schedule would allow for greater hematopoietic recovery, and would potentially be associated with reduced myelosuppression at clinically effective doses. The patient population that received DMS612 includes patients with multiple prior therapies, median 4 (range 0 – 13). Cumulative myelosuppression is more likely in patients with extensive prior cytotoxic therapies. As a result, patients enrolled on this study may not predict tolerability in patients with RCC, who are principally treated with targeted agents and immunotherapy that is not generally associated with severe or prolonged myelosuppression.

In the study presented here, tumor responses were observed in one patient with mRCC and in one patient with advanced cervical carcinoma. The fact that these responses were observed in a heavily pretreated population with a first-in-human agent is certainly interesting, and warrants further investigation of antitumor activity. RCC is generally not responsive to traditional cytotoxic agents used to treat other solid tumors such as alkylating agents, plant alkaloids including the taxanes, and antimetabolites (12, 13). At present, the mechanisms of RCC chemoresistance are not well understood. However, contributing factors include overexpression of several important proteins, including multidrug resistance proteins, major vault protein, and glutathione-s-transferase (12, 14–16). The degree to which DMS612 is a substrate of these proteins is unknown. Notably, the sensitivity pattern for an ABCB1 substrate in the NCI-60 was not observed for DMS612 (1). The dimethane sulfonate compounds shared a unique profile of specificity against yeast strains with particular DNA repair mutations (4). This pattern of sensitivity could potentially be exploited for identification of a biomarker for predicting response among individual tumors with specific defects in DNA repair (17). Another potential biomarker is ATAD5, a DNA repair protein with polymorphic variants that were identified as highly correlated with DMS612 sensitivity in the NCI 60 cell line drug screen (18). Elucidating which DNA repair proteins are important for DMS612 sensitivity and resistance could allow patient selection or stratification in future clinical trials.

In conclusion, the predictable and well-tolerated safety profile and preliminary antitumor activity supported by pharmacodynamic evidence of DNA damage suggest that DMS612 is suitable for phase II studies in RCC and other solid tumors. The recent approvals of pemetrexed for non–small cell lung cancer and bendamustine for chronic lymphocytic leukemia and indolent B-cell lymphoma demonstrate that novel cytotoxic agents will continue to provide advances in cancer treatment even in this era of targeted therapy. DMS612 offers the interesting possibility of an alkylating agent with a unique toxicity profile. The lack of alopecia, nausea, or mucositis suggests the drug may not be activated in all tissues, and means improved tolerability over alkylating agents of the past. Strategies to avoid the cumulative myelotoxicity and to exploit DMS612 in renal cancer should be vigorously pursued.

No potential conflicts of interest were disclosed.

Conception and design: L.J. Appleman, S. Balasubramaniam, J.J. Wright, D.R. Kohler, C.P. Belani, J.H. Beumer, S.E. Bates

Development of methodology: S. Balasubramaniam, R.A. Parise, C.E. Redon, D.R. Kohler, C.P. Belani, J. Eiseman, J.H. Beumer, S.E. Bates

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L.J. Appleman, S. Balasubramaniam, R.A. Parise, C. Bryla, C.E. Redon, A.J. Nakamura, W.M. Bonner, D.R. Kohler, Y. Jiang, C.P. Belani, J. Eiseman, J.H. Beumer, S.E. Bates

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L.J. Appleman, S. Balasubramaniam, R.A. Parise, C. Bryla, C.E. Redon, W.M. Bonner, C.P. Belani, E. Chu, J.H. Beumer, S.E. Bates

Writing, review, and/or revision of the manuscript: L.J. Appleman, S. Balasubramaniam, R.A. Parise, C. Bryla, C.E. Redon, W.M. Bonner, J.J. Wright, R. Piekarz, D.R. Kohler, C.P. Belani, J. Eiseman, E. Chu, J.H. Beumer, S.E. Bates

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Balasubramaniam, C. Bryla, R. Piekarz, D.R. Kohler, J.H. Beumer

Study supervision: S. Balasubramaniam, C. Bryla, J.J. Wright, J.H. Beumer, S.E. Bates

The authors thank the Merrill Egorin Writing Group at UPCI for their constructive comments and input regarding the manuscript, the late Dr. Merrill Egorin for his input in the original design of this trial, the clinical trial group of Dr. Tito Fojo for participation in the conduct of this trial, and Rob Robey for editorial assistance.

This work was supported, in part, by the National Cancer Institute grants U01-CA099168 (to L.J. Appleman, R.A. Parise, Y. Jiang, C.P. Belani, J. Eiseman, E. Chu, and J.H. Beumer) and P30-CA47904 (to J.H. Beumer and R.A. Parise), and by the Intramural Research Program of the NIH, National Cancer Institute (to S. Balasubramaniam, C. Bryla, W.M. Bonner, C.E. Redon, A.J. Nakamura, J.J. Wright, R.A. Parise, and S.E. Bates).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.