Induction of Apoptosis Using Inhibitors of Lysophosphatidic Acid Acyltransferase-β and Anti-CD20 Monoclonal Antibodies for Treatment of Human Non-Hodgkin's Lymphomas

Purpose: Lysophosphatidic acid acyltransferase-β (LPAAT-β) is a transmembrane enzyme critical for the biosynthesis of phosphoglycerides whose product, phosphatidic acid, plays a key role in raf and AKT/mTor-mediated signal transduction.

Experimental Design: LPAAT-β may be a novel target for anticancer therapy, and, thus, we examined the effects of a series of inhibitors of LPAAT-β on multiple human non–Hodgkin's lymphoma cell lines in vitro and in vivo.

Results: We showed that five LPAAT-β inhibitors at doses of 500 nmol/L routinely inhibited growth in a panel of human lymphoma cell lines in vitro by >90%, as measured by [³H]thymidine incorporation. Apoptotic effects of the LPAAT-β inhibitors were evaluated either alone or in combination with the anti-CD20 antibody, Rituximab. The LPAAT-β inhibitors induced caspase-mediated apoptosis at 50 to 100 nmol/L in up to 90% of non–Hodgkin's lymphoma cells. The combination of Rituximab and an LPAAT-β inhibitor resulted in a 2-fold increase in apoptosis compared with either agent alone. To assess the combination of Rituximab and a LPAAT-β inhibitor in vivo, groups of athymic mice bearing s.c. human Ramos lymphoma xenografts were treated with the LPAAT-β inhibitor CT-32228 i.p. (75 mg/kg) daily for 5 d/wk × 4 weeks (total 20 doses), Rituximab i.p. (10 mg/kg) weekly × 4 weeks (4 doses total), or CT-32228 plus Rituximab combined. Treatment with either CT-32228 or Rituximab alone showed an approximate 50% xenograft growth delay; however, complete responses were only observed when the two agents were delivered together.

Conclusions: These data suggest that Rituximab, combined with a LPAAT-β inhibitor, may provide enhanced therapeutic effects through apoptotic mechanisms.

The safety and efficacy of monoclonal antibody therapy has made it a popular therapeutic modality for the treatment of patients with non–Hodgkin's lymphomas. The anti-CD20 antibody Rituximab is the most extensively used treatment for non–Hodgkin's lymphoma and has been shown to induce cytotoxicity by antibody-dependent cell-mediated cytotoxicity, complement activation, and the direct signaling of apoptosis (1–3). However, only ∼50% of patients respond to standard doses of Rituximab and even responding patients are not cured, suggesting that mechanisms of resistance exist that limit eradication of lymphoma cells after antibody binding to CD20 antigen (3). Therefore, alternative strategies are needed to augment Rituximab-mediated cytotoxicity for non–Hodgkin's lymphoma.

Inhibitors of lysophosphatidic acid acyltransferase-β (LPAAT-β) have recently been explored as novel anticancer therapeutic agents (4–7). The gene for LPAAT-β is encoded in a region of the class III human MHC and its product is an intrinsic transmembrane enzyme critical for the biosynthesis of phosphoglycerides (8). The specific role of LPAAT, also known as 1-acyl-sn-glycerol-3-phosphate-acyltransferase, is to catalyze the transfer of acyl groups from acyl-CoA to lysophosphatidic acid, to form phosphatidic acid (4, 9, 10). Phosphatidic acid has also been shown to be critical for signal transduction in the ras/raf/mitogen-activated protein kinase and PI3K/mTor oncogenic pathways (11–13). The most common isoforms are LPAAT-α and LPAAT-β, which show varied expression levels on different tissues (9, 10). Whereas LPAAT-α is expressed at a relatively constant level in virtually all human tissue tested, LPAAT-β is expressed at low levels in most normal tissues and is increased in epithelial tumor tissues and endothelial cells (4, 10). Enzymatic inhibition of the LPAAT-β enzyme seems to interrupt these oncogenic pathways, leading to apoptosis (4).

In this study, we investigated the effects of inhibitors of LPAAT-β activity in combination with Rituximab on human non–Hodgkin's lymphoma cells, in vitro and in vivo, in an effort to enhance the therapeutic efficacy of anti-CD20 antibody therapy by increasing the level of apoptosis that may be necessary to overcome antibody resistance. We have shown that exposure of human non–Hodgkin's lymphoma cells to the combination of LPAAT-β inhibitors and Rituximab enhances apoptosis in vitro and augments antitumor responses in vivo.

Non–Hodgkin's lymphoma tumor cell lines. The CD-20 expressing human Burkitt's lymphoma cell line Ramos was obtained from the American Type Culture Collection (Manassas, VA). The B lymphoma cell lines SU-DHL4 (DHL-4), TAB, Oci-Ly8, and FL-18 were gifts from Dr. David Maloney (Fred Hutchinson Cancer Research Center, Seattle, WA). Cell lines were maintained in RPMI 1640 with 2 mmol/L l-glutamine (Life Technologies, Grand Island, NY), supplemented with 10% fetal bovine serum (BioWhittaker, Walkersville, MD), 1 mmol/L sodium pyruvate, 100 units/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL Fungizone (BioWhittaker). Cells were incubated at 37°C with 5% CO₂ and maintained in log phase growth. Primary T and B cells were isolated and cultivated by Cell Therapeutics Incorporated (Seattle, WA). Peripheral blood mononuclear cells were isolated from whole blood using Histopaque (Sigma) gradient centrifugation. T and B lymphocytes were separated using magnetic beads from Biosource International (Camarillo, CA) coated with antibodies for CD3, CD3e, CD19, and CD8 from PharMingen (San Diego, CA). All primary populations were then cultivated in RPMI (Life Technologies, Carlsbad, CA) and 10% heat-inactivated fetal bovine serum. The “activated” population of B lymphocytes included 100 EU/mL lipopolysaccharide (Sigma, St. Louis, MO) and for T lymphocytes 2.5 μg/mL concanavalin A (Sigma).

Lysophosphatidic acid acyltransferase-β inhibitors and reagents. We studied three previously described triazine LPAAT-β inhibitors (CT-32228, CT-32176, and CT-32212) and two pyrimidine LPAAT-β inhibitors (CT-32615 and CT-32521; refs. 14, 15). The CT-32212 compound is structurally similar to CT-32176 and CT-32228, but is relatively inactive and was, therefore, used as a negative control (4). The molecular structures of the compounds are shown in Fig. 1.

Rituximab (Genentech, Inc., South San Francisco, CA) was used at a working concentration of 1.0 μg/mL for the in vitro studies and 200 μg/mouse/wk was injected i.p. for in vivo experiments. Sodium azide (Sigma) was used at a concentration of 2% as a positive apoptotic control. Annexin V conjugated to FITC (BD Biosciences Clontech, Palo Alto, CA) and propidium iodide (BD Biosciences) were stored at 4°C and used at 20 and 50 μg/mL, respectively, in apoptosis assays. FAM-peptide-fluoromethyl ketone (used in caspase inhibitor assays) was dissolved in 100% DMSO (American Type Culture Collection, Bethesda, MD) and stored at −20°C. [³H]thymidine (Perkin-Elmer, Boston, MA) was stored at 4°C and diluted to a working solution of 1 μCi/20 μL.

Total lysophosphatidic acid acyltransferase activity in non–Hodgkin's lymphoma cells. LPAAT activity was assayed in primary T and B lymphocytes and in hematologic cell lines as described by Hideshima et al. (7). This assay does not distinguish between LPAAT-α and LPAAT-β activities (data not shown) and, hence, will be called total endogenous LPAAT activity.

Growth inhibition assays. The effects of LPAAT-β inhibitors on the growth of lymphoma cell lines were tested using a [³H]thymidine incorporation assay (16). Cells were plated in 24-well flat-bottomed plates at a density of 1.25 × 10⁵ cells in 1.0 mL culture medium. The LPAAT-β inhibitors were freshly dissolved in 100% DMSO (American Type Culture Collection) and added to wells in concentrations ranging from 10 to 1,000 nmol/L. Cell suspensions were incubated with LPAAT-β inhibitors for 24 to 96 hours at 37°C. For the final 8 hours, 200 μL aliquots of each cell suspension were transferred in triplicate into 96-well flat-bottomed plates, and then 1 μCi of [³H]thymidine (Perkin-Elmer) was added to each of the 200 μL cell suspensions to bring the final volume to 220 μL. Untreated cells and cells that were incubated with DMSO diluent alone were used as negative controls. Following this final incubation at 37°C, cells were harvested onto glass fiber filters (Millipore, Inc., Billerica, MA) with a Packard Filtermate Harvester and [³H]thymidine incorporation was analyzed by liquid scintillation on a Packard Top Count Microplate Scintillation Counter (Packard Instruments, Meriden, CT). Counts per minute were analyzed and proliferation rates were calculated for each experimental condition. The mean and SE values for each time point were plotted to generate growth inhibition curves for each experimental group. All experiments were done in triplicate.

Apoptosis assays. Apoptosis assays were done using the ApoAlert apoptosis kit using Annexin V–conjugated FITC to detect phosphatidylserine translocated to the outer leaflet of the cell membrane and propidium iodide staining for detection of nonspecific cellular necrosis (BD Biosciences Clontech). Non–Hodgkin's lymphoma cells (1 × 10⁶/mL) were incubated at 37°C for 48 hours with LPAAT-β inhibitors at concentrations ranging from 50 to 100 nmol/L, either alone or in combination with 1.0 mg/mL Rituximab. Cell suspensions were subsequently washed and incubated with Annexin V-FITC (20 μg/mL), propidium iodide (50 μg/mL), and 1× binding buffer [0.1 mol/L HEPES (pH 7.4), 1.4 mol/L NaCl, 25 mmol/L CaCl₂; BD Biosciences Clontech]. Additional assays using the same concentrations of LPAAT-β inhibitor and Rituximab were done using the APO LOGIX Carboxyfluorescein Caspase (FAM-VAD-FMK) Detection kit (Cell Technology, Minneapolis, MN) for detection of activated caspases. Flow cytometry was done on a fluorescence-activated cell sorter and data analysis was done using CellQuest software (Becton Dickinson, San Jose, CA). Average and SE values were collected and are reported for each group of compounds studied. Sodium azide (2% final concentration) was used as a positive control. Negative controls included untreated cells and cell suspensions were treated with DMSO vehicle alone.

Mouse studies: Mice. Nonobese diabetic CB17 severe combined immunodeficient mice were obtained from the animal health facility at the Fred Hutchinson Cancer Research Center. Mice ages 6 to 8 weeks were kept in specific pathogen-free conditions and maintained under protocols approved by the Fred Hutchinson Cancer Research Center Institutional Animal Care and Use Committee.

Therapy experiments. Mice received s.c. injections of either 10 × 10⁶ Ramos lymphoma tumor cells, in experiments to establish palpable xenograft tumors, or 7 × 10⁶ cells for experiments designed to treat animals in minimal residual disease state. Mice with similar palpable tumor sizes (∼500 mm³) were selected for experimentation before receiving LPAAT-β inhibitor alone, Rituximab alone, or the combination of LPAAT-β inhibitor and Rituximab. In all murine experiments, CT-32228 was chosen as the LPAAT-β inhibitor because of its more favorable metabolic characteristics compared with CT-32615. Using human liver microsomes and a concentration of 5 mmol/L of each compound, 13% of CT-32228 was metabolized after 30 minutes, compared with 55% metabolism of CT-32615. Using mouse liver microsomes, at 20 mmol/L, CT-32228 was 25% metabolized after 30 minutes, compared with 41% metabolism of CT-32615. In minimal residual disease experiments, mice received the same therapeutic agents 72 hours following delivery of s.c. Ramos cells and before the appearance of palpable tumors. In each experiment, a group of five mice received DMSO alone as a control. Groups of five mice received either 72 mg/kg LPAAT-β inhibitor i.p. daily for 5 d/wk × 4 weeks (total 20 doses), 10 mg/kg Rituximab i.p. once weekly × 4 weeks (total 4 doses), or a combination of LPAAT-β inhibitor and Rituximab administered at identical doses to those used in single-agent animal groups. A dose of 72 mg/kg of LPAAT-β inhibitor was chosen for these studies based on the activity of these compounds in prior dose-escalation studies treating epithelial carcinomas using severe combined immunodeficient mice.⁶

In addition, doses of <72 mg/kg of LPAAT-β inhibitor alone had suboptimal efficacy in a dose-response study of murine lymphoma (data not shown). The dose of Rituximab (10 mg/kg) used was derived from previous xenograft experiments (17). Mice in this study were monitored every other day for general appearance, weight loss, and tumor volume measurements. Mice were euthanized if tumors grew large enough to cause obvious discomfort or impair ambulation.

Toxicity studies. In toxicity studies, groups of five mice received LPAAT-β inhibitors at doses of 18, 36, or 72 mg/kg for 5 d/wk with a maximum of 4 weeks (up to 20 total doses) of therapy. On the first day of each treatment week, mice were also bled through the retro-orbital venous plexus and blood was collected for analysis of serum alanine aminotransferase, aspartate aminotransferase, urea nitrogen, and creatinine levels. An untreated group and mice treated with only DMSO served as control mice, which were sacrificed at the end of the fourth treatment week.

In a separate series of experiments, groups of five mice assigned to treatment control groups in the manner described above were sacrificed and necropsied. Representative 5-μm sections stained by H&E were prepared from lungs, liver, spleen, kidneys, stomach, and small and large intestine. Slides were labeled with random numbers and each tissue analyzed in a blinded fashion for enumeration of apoptotic cells characterized by pyknosis, condensation of chromatin, and blebbing of apoptotic bodies. The degree of apoptosis in each tissue was graded semiquantitatively as absent, mild, moderate, or severe (18). This apoptosis scale was adapted from a routine grading system designed to evaluate apoptosis in patients with graft-versus-host disease following hematopoietic stem cell transplant. Grades were determined in the gastrointestinal tract by evaluation of the number of apoptotic cells in ∼50 mucosal crypts per glands for each therapeutic group analyzed. A mild apoptotic grade was assigned to specimens containing three to nine apoptotic cells per 50 epithelial crypt, a moderate grade was associated with 10 to 20 apoptotic cells per 50 crypts, and severe apoptosis was induced in samples with >20 apoptotic cells per 50 crypts.

Endogenous levels of lysophosphatidic acid acyltransferase activity in primary B and T lymphocytes and human non–Hodgkin's lymphoma cell lines. Total LPAAT activity was determined in a variety of human non–Hodgkin's lymphoma tumor cell lines and primary B and T lymphocytes cells using the cell-free enzymatic assay described in Materials and Methods. Activity levels in a panel of human lymphoma cells lines, including two Burkitt's lymphoma (Daudi and Ramos), two large B-cell lymphoma cell lines (Oci-Ly8 and TAB), and two transformed follicular non–Hodgkin's lymphoma cell lines (FL18 and DHL-4), displayed LPAAT activity levels ranging from 3.9 ± 0.2 to 8.0 ± 0.2 nmol/min/mg (Table 1). LPAAT activity levels in the majority of these human non–Hodgkin's lymphoma cell lines were greater than the levels seen in unstimulated and stimulated primary human B and T cells: unstimulated and stimulated primary B cells had an expression level of 2.6 ± 0.1 and 3.2 ± 0.4 nmol/min/mg, respectively, and stimulated and unstimulated T lymphocytes had endogenous LPAAT activity levels of 2.9 ± 0.1 and 4.5 ± 0.2 nmol/min/mg, respectively. A single murine B-cell line, A20, expressed significantly lower levels of LPAAT activity (0.9 ± 0.1 nmol/min/mg) compared with the human cell lines.

Effect of lysophosphatidic acid acyltransferase-β inhibitors on tumor cell growth. Because LPAAT activity was detected in human lymphoma cells lines, we determined the effects of a panel of LPAAT-β inhibitors on lymphoma cell proliferation in vitro. The LPAAT-β inhibitor CT-32228 and four structural analogues, CT-32176, CT-32521, CT-32615, and CT-32212, were studied using Ramos, FL-18, and DHL-4 malignant B lymphoma cell lines. The CT-32212 inhibitor did not exhibit an antiproliferative effect on any of the cell lines tested, consistent with previous results (4). Inhibitors CT-32228, CT-32176, CT-32521, and CT-32615 exhibited significant growth inhibition after 24 hours with an IC₅₀ of 50 to 100 nmol/L in all cell lines tested. Each of these four inhibitors exhibited maximal inhibition at concentrations of 500 nmol/L, with one of the most active compounds, CT-32615, inhibiting cell proliferation to <50% of control cells in DHL-4, FL-18, and Ramos cells by 24 hours (Fig. 2). Using 500 nmol/L of LPAAT-β inhibitor, Ramos cell proliferation decreased to 7.0%, 1.8%, and 0.8% at 24, 48, and 72 hours, respectively, compared with control cells treated with DMSO alone. In DHL-4 and FL-18 lymphoma lines using the same dose of LPAAT-β inhibitor, the levels of cell proliferation decreased to 40% and 8% at 24 hours, 30% at 48 hours, and <20% and 0.8% at 72 hours, respectively, compared with control DMSO-treated cells. These results suggest that inhibitors of the LPAAT-β isoform can induce cytotoxicity in transformed follicular, large B cell, and Burkitt's lymphoma cell lines in vitro.

Lysophosphatidic acid acyltransferase-β inhibitors induce apoptosis and augment the apoptotic activity of Rituximab in human non–Hodgkin's lymphoma cell lines: Induction of apoptosis by lysophosphatidic acid acyltransferase-β inhibitors. To determine the mechanism of the antiproliferative effect induced by these inhibitors, we assayed non–Hodgkin's lymphoma cell lines for apoptosis after incubation for 48 hours with 50 to 800 nmol/L of LPAAT-β inhibitors. In three separate experiments, compounds CT-32228, CT-32615, and CT-32176 induced significant apoptosis in all non–Hodgkin's lymphoma cell lines at concentrations that were also antiproliferative. After 48 hours, each compound induced Annexin V positivity in ∼50% of cells at inhibitor concentrations of 200 to 800 nmol/L (Fig. 3). The LPAAT-β inhibitor CT-32228 was less potent than the other inhibitors, producing minimal apoptosis at concentrations ≤100 nmol/L with >40% Annexin V–positive cells detected using 50 nmol/L of CT-32615. The structural analogue CT-32212 induced apoptosis in <5% of cells at the highest concentration tested.

Effect of the combination lysophosphatidic acid acyltransferase-β inhibitors and Rituximab on apoptosis. Rituximab treatment has been reported to induce direct apoptotic and antiproliferative effects through binding of the CD20 antigen on the surface of many malignant B-cell lymphoma lines in vitro (1) and in malignant human B cells in vivo (19). Because the in vitro data presented above suggests that at least part of the antiproliferative effect of LPAAT-β inhibition is due to apoptosis, we evaluated the effects of Rituximab combined with an LPAAT-β inhibitor, CT-32228. FITC-Annexin V staining was used to evaluate apoptosis in non–Hodgkin's lymphoma cell lines treated with 50 nmol/L CT-32228 and 1 μg/mL Rituximab, alone and in combination. CT-32615 was not studied in these experiments due to suboptimal metabolic properties as described in Materials and Methods. After 48 hours of incubation, all cell lines displayed increased percentages of cells with phosphatidylserine exposure on the outer leaflet of the cell following exposure to CT-32228 alone, ranging from 15.2% ± 2.1 to 23.5% ± 0.5, or Rituximab alone, ranging from 13.5% ± 3.6 to 24.8% ± 2.8 (Fig. 4). The addition of Rituximab to CT-32228 produced further increases in Annexin V positivity in the Ramos cell line after 48 hours. In Ramos cells, 30.0% ± 0.8 apoptotic cells were detected at 48 hours using the CT-32228 and Rituximab combination (Figs. 4 and 5).

Because Rituximab-induced apoptosis is believed to be mediated by activation of caspase cascades (1), we used the carboxyfluroescein (FAM)-labeled peptide fluoromethyl ketone to detect activated caspases in Ramos and DHL-4 cells after incubation with Rituximab, CT-32228, or combinations of the two agents (20). Cells that bind fluoromethyl ketone are suspended in the early phase of apoptosis, and can be detected by flow cytometry. Rituximab induced apoptosis in DHL-4 cells (14.0% ± 1.8) and in Ramos cells (7.3% ± 1.2) after 48 hours (Fig. 4). CT-32228 induced caspase activation in 9.0% ± 0.8 of DHL-4 cells and 10.5% ± 0.8 of Ramos cells after 48 hours. The combination of CT-32228 and Rituximab produced apoptosis in 26.2% ± 1.3 of DHL-4 cells and 21.2% ± 0.2 of Ramos cells. These results suggest that the combination of CT-32228 and Rituximab each act to induce apoptosis in B non–Hodgkin's lymphoma cells via a caspase-mediated mechanism of action.

In vivo treatment with lysophosphatidic acid acyltransferase-β inhibitors and Rituximab in a minimal tumor burden model. To assess the in vivo potential of LPAAT-β inhibitors, we initially tested them alone and in combination with Rituximab in a murine model of “minimal disease.” Mice were injected s.c. with 7 × 10⁶ Ramos Burkitt's lymphoma cells. Seventy-two hours after injection, mice were randomly allocated to one of four groups: (a) mice that received daily i.p. injections of 72 mg/kg of the LPAAT-β inhibitor CT-32228 alone for 5 d/wk for 4 weeks (total of 20 doses), (b) mice that received 10 mg/kg Rituximab once per week i.p. for 4 weeks (total of four doses), (c) the combination of CT-32228 and Rituximab at the doses and schedules above, or (d) DMSO diluent alone. Treatment doses were used based on results from previous studies, as noted in Materials and Methods. In addition, doses <72 mg/kg of the LPAAT-β inhibitor showed suboptimal efficacy in a prior dose-escalation study of murine lymphoma (data not shown). No tumor development was detected in any of the five mice treated with the combination of CT-32228 and Rituximab; however, four of five (80%) mice treated with Rituximab alone (P = 0.50) and five of five (100%) of mice treated with CT-32228 alone (P = 0.004) developed palpable tumors by day 16 after tumor implantation (Fig. 6). All control mice treated with the DMSO vehicle alone developed progressive tumor growth requiring euthanasia by day 31 after injection (P = 0.004 for comparison to combination of CT-32228 and Rituximab).

Although the LPAAT-β inhibitor alone at a dose of 72 mg/kg produced a tumor growth delay, mice treated with this agent displayed significant wasting, leading to euthanasia of all five mice by day 31 despite the absence of tumor progression (Figs. 6 and 8A). The five mice treated with CT-32228 and Rituximab in combination were also euthanized on day 42 due to progressive weight loss although none of them had evidence of tumor progression (Fig. 8A). By comparison, four of five mice treated with 10 mg/kg Rituximab did transiently develop a palpable tumor on day 16, which regressed by day 26 (Fig. 6). One mouse in the Rituximab group never developed a palpable tumor in this model. The mice in the single-agent Rituximab group were observed without any additional recurrences for at least 150 days after the beginning of the experiment without evidence of tumor development.

In vivo effects of lysophosphatidic acid acyltransferase-β inhibitor and Rituximab therapy in mice with established lymphoma xenografts. We recognize that the murine model described above using mice with a minimal human B-cell lymphoma tumor burden has limitations. This minimal residual disease model requires s.c. administration of the malignant cells and subsequent delivery of the therapeutic agent(s) before establishment of a distinct tumor mass. This model differs from the clinical setting experienced by most non–Hodgkin's lymphoma patients who have demonstrable lymph node and bone marrow–based disease at the time of treatment. Therefore, we investigated the use of the LPAAT-β inhibitor CT-32228 and Rituximab, used either alone or in combination, delivered after the establishment of a palpable xenograft tumor in mice. When Ramos tumor xenografts reached sufficient size (∼500 mm³), mice were treated with an i.p. injection of either 72 mg/kg of CT-32228, 10 mg/kg Rituximab, a combination of the two agents delivered together at the same doses, or with the DMSO vehicle alone. Complete tumor responses were seen in two of five mice receiving the combination of CT-32228 and Rituximab, with a maximal response seen 16 days after treatment (Fig. 7). Partial responses, as defined by at least a 50% reduction in the size of the tumor, were seen in the remaining three mice in this group (Fig. 7). In contrast, one of five mice receiving monotherapy with CT-32228 achieved a transient remission, with recurrence by day 23, and euthanasia was mandated in all mice in this group due to tumor progression by day 32. None of the five mice that received Rituximab alone achieved a complete remission and all of the mice receiving single-agent Rituximab were euthanized by day 26 due to excessive tumor growth.

Despite the ability of the LPAAT-β inhibitor CT-32228 to induce remissions when used in combination with Rituximab, the ability of this combined therapy to prolong survival was limited due to toxicities (described in detail below; Fig. 8B). In this palpable disease model, 60% (three of five) of mice receiving 72 mg/kg of LPAAT-β inhibitor alone and 100% (five of five) of mice that received 72 mg/kg CT-32228 in combination with Rituximab exhibited >15% weight loss, diarrhea, and huddling behavior. These toxicities were significant enough to prompt euthanasia between days 20 and 36, despite complete regressions of palpable tumor xenografts in all the mice receiving the combination of agents, which were durable up to the time of sacrifice. Conversely, mice treated with single-agent Rituximab and control mice showed no evidence of toxicities, although their survivals were limited by progressive tumor growth leading to death by days the 36 and 33, respectively.

Toxicity analysis in mice receiving CT-32228. To further elucidate systemic toxicities incurred as a result of treatment with CT-32228, hematologic, hepatic, and renal parameters were assessed in nontumor-bearing mice that received CT-32228 concentrations ranging from 18 to 72 mg/kg (Table 2). Blood was obtained weekly from the retro-orbital venous plexus to measure leukocyte counts and hemoglobin values. Minor decrements were observed in the leukocyte counts of all groups receiving DMSO regardless of the presence of an LPAAT-β inhibitor (Table 2), with nadirs occurring 7 days after therapy. Significant differences were seen only between control (1.1 ± 0.2 K/μL) and the highest concentration, 72 mg/kg (0.3 ± 0.0 K/μL). Nadirs were observed in hemoglobin levels after 21 days and in platelet levels after 21 days in mice receiving 36 mg/kg of the LPAAT-β inhibitor.

Assessments of hepatic and renal toxicities were done in a separate series of experiments by sampling of blood weekly for aspartate transaminase, alanine transaminase, and creatinine levels (Table 2). Approximately 2-fold increases in transaminase levels were observed by day 7 in groups of mice that received the LPAAT-β inhibitor at concentrations of 36 and 72 mg/kg; however, progressive toxicities limited the survival of mice in these groups, preventing further analysis. In contrast, no elevations in aspartate transaminase or alanine transaminase levels were detected in the group of mice receiving 18 mg/kg of CT-32228. Serum creatinine levels were unaffected during treatment.

Mouse weight loss was 0% for mice receiving DMSO vehicle alone, 0% for 18 mg/kg, 16% for 36 mg/kg, and 18% for 72 mg/kg. Thus, blinded examination of H&E–stained tissue sections from mice were analyzed for determination of apoptosis induction. Examination of sections from the gastrointestinal tract suggested that the gastrointestinal epithelium was the only tissue to develop measurable levels of cellular apoptosis among the tissues studied. The gastrointestinal tract showed minimal or absent apoptosis in epithelial cells of mice receiving 18 mg/kg of LPAAT-β inhibitor, indistinguishable from apoptosis scores observed in normal untreated animals or mice receiving only DMSO. Mild levels of apoptosis (three to nine apoptotic cells per 50 epithelial crypts) were seen in mice receiving 36 mg/kg. In contrast, severe generalized apoptosis (>20 apoptotic cells per 50 epithelial crypts) was observed in the gastrointestinal tract of in all animals receiving 72 mg/kg. Apoptosis was most marked in the colon, where almost all crypts in every section displayed multiple apoptotic cells, including many that had sloughed into the crypt lumens (Fig. 9). These alterations were milder in the small intestine and fore-stomach. The glandular stomach was least involved. The histology of other organs did not differ substantially between animals treated with CT-32228 and controls.

The production of lysophosphatidic acid and phosphatidic acid has been shown to be important for cell survival because these lipid mediators play key roles in cell cycle regulation due to their growth factor–like effects (9, 21, 22). The suppression of phosphatidic acid biosynthesis by small compounds that inhibit the interleukin 2 receptor–associated Jak/signal transducers and activators of transcription signaling pathway has been shown to result in complete suppression of malignant T-cell lymphoma growth (23). Phosphatidic acid has also been shown to be involved in the downstream activation of raf protein as part of the ras/raf/extracellular signal-regulated kinase signaling pathway. Experiments using LPAAT-β inhibitors on primary endothelial cells resulted in suppression of raf translocation to the cell membrane, which is necessary for raf activation, as well as inhibition of extracellular signal-regulated kinase phosphorylation (4). The phosphorylation/activation of another molecule important for cell survival, Akt, is diminished by the LPAAT inhibitors in vascular smooth muscle cells (4). It has been hypothesized that inhibition of these known oncogenic signaling pathways can be an effective tool to regulate tumor cell growth. To this end, investigators have targeted inhibitors to an enzyme required for phosphatidic acid synthesis, LPAAT-β, to diminish production of phosphatidic acid products and thereby interfere with the oncogenic pathways important for growth of human cancers, including non–Hodgkin's lymphoma. Preliminary studies have shown antiproliferative effects of the LPAAT-β inhibitors in numerous human tumor cell types, including those of prostate, breast, lung, pancreas, and colon (4, 5).

Inhibition of the LPAAT-β isoform has also been shown to have negative proliferative effects on hematopoietic cell lines. The inhibition of the LPAAT-β enzyme using nanomolar concentrations of aryldiaminotriazine compounds resulted in significant inhibition of cell proliferation in vitro in Epstein-Barr virus–transformed B non–Hodgkin's lymphoma, T-cell leukemia, chronic myeloid leukemia, promyelocitic leukemia, and Burkitt's lymphoma cell lines (4). Hideshima et al. (7) also showed that LPAAT-β inhibitors caused antiproliferative effects in multiple myeloma cells at concentrations of <200 nmol/L. Interestingly, the effects of LPAAT-β inhibitors were shown on the observed cytotoxicity toward myeloma cells in the microenvironment of the bone marrow, suggesting that inhibition of LPAAT-β can occur in a clinically relevant environment.

Consistent with these prior studies, we postulated that LPAAT-β may also be an effective novel therapeutic target for the inhibition of B non–Hodgkin's lymphoma cell growth. The biological effect of these compounds on non–Hodgkin's lymphoma cells, however, had not been previously characterized. In this study, we found that cell growth of human non–Hodgkin's lymphoma cells is inhibited up to 10-fold with the use of an LPAAT-β inhibitor compared with untreated malignant B cells. These results suggest that LPAAT-β can be a potential target for the inhibition of non–Hodgkin's lymphoma proliferation, and may serve as an important therapeutic intervention to treat non–Hodgkin's lymphoma patients. To further elucidate the mechanism of the antiproliferation effects mediated by LPAAT-β inhibition, we did a series of experiments examining the apoptotic potential of these novel agents on human non–Hodgkin's lymphoma cells. Whereas the protective role of phosphatidic acid against apoptosis is still a topic of debate, inhibiting enzymes involved in the production of phosphatidic acid has been shown to be associated with regulation of cell-signaling pathways involved in cell growth and viability (4, 5, 24). It has also been shown that new novel agents, such as the aryldiaminotriazines that inhibit LPAAT-β, can also induce apoptosis in a variety of tumor cell lines (4). We and others have hypothesized that apoptosis triggered by specific targeting may be preferred to apoptosis induced by untargeted therapies due to limited inflammatory changes and a potential reduced risk of mutation, helping to ensure that tumor cells are eliminated.

Caspases are cysteine proteins that mediate apoptosis after activation by specific signals originating from both outside and inside the cell (25). Other authors have suggested that apoptosis induced by LPAAT-β inhibitors in myeloma cells is mediated by activation of caspase-8 and caspase-7 (7) and involves ligation of death receptors (5), suggesting that the extrinsic pathway is utilized in apoptosis induced by these agents (25). The intrinsic pathway, on the other hand, involves the release of cytochrome c from mitochondria, which induces Apaf-1 and leads to the activation of caspase-3 and caspase-9 (25). This process is facilitated through an important group of Bcl-2 proteins, including Bax and Bak, that regulate the permeability of the mitochondrial membrane and consequently influence the induction of apoptosis (25, 26). Recent data show activated caspase activity and apoptosis in a majority of cell lines derived from a variety of human epithelial tumors when treated with LPAAT-β inhibitors (4, 5, 7). We found that the induction of apoptosis in multiple non–Hodgkin's lymphoma tumor cell lines mediated by LPAAT-β inhibitors is also mediated by caspase activation.

Anti-CD20 antibodies also induce cell death through apoptotic pathways, particularly when cross-linked on the surface of non–Hodgkin's lymphoma tumor cell lines (1, 27). Ligation of the CD20 antigen by Rituximab, a chimeric human IgG1 monoclonal antibody, has been shown to lead to intracellular increases in calcium levels and poly(ADP-ribose) polymerase cleavage in vitro, an important substrate for apoptosis (19). More recent data suggests that the mitochondrial intrinsic pathway plays a key role in Rituximab-mediated apoptosis by releasing cytochrome c, leading to the activation of caspase-9 and subsequent activation of caspase-3 (19). Additional antiapoptotic proteins, such as Bcl-2, also prevent the release of cytochrome c and act to inhibit apoptosis (28). Thus, down-regulation of Bcl-2 may lower the apoptosis threshold for activation and allow the apoptotic cascade to be triggered more easily in response to anticancer agents such as LPAAT-β inhibitors and anti-CD20 antibodies (1, 3, 19). Similar results have been achieved using antisense oligodeoxynucleotides to suppress Bcl-2 activity and create an environment that is more sensitive to the induction of apoptosis when used in the presence of Rituximab (29). Therefore, we utilized human non–Hodgkin's lymphoma cell lines with known elevated Bcl-2 levels to show that these lymphoma cell lines are also sensitive to treatment with LPAAT-β inhibitors. These results have led us to hypothesize that a combination treatment involving anti-CD20 antibody therapy and an LPAAT-β inhibitor may lead to an increased level of apoptosis induction in non–Hodgkin's lymphoma cells compared with the degree of apoptosis seen with the use of either agent alone. In a similar manner, inhibition of the PI3K/Akt survival pathway along with Gemcitabine has been found to effectively enhance apoptosis in innately drug-resistant human pancreatic cancer cells (26). The method of targeting multiple pathways of the apoptotic cascade may serve as a way to guard against possible mutations in a single apoptotic pathway, and effectively cast a wider tumoricidal net, preventing emergence of resistant tumor cells (25).

Despite the tolerability and widespread use of single-agent Rituximab, it is efficacious in only ∼50% to 60% of relapsed indolent non–Hodgkin's lymphoma patients and the median duration of response for relapsed non–Hodgkin's lymphoma patients is on the order of ∼12 months (1). Therefore, a combination treatment of Rituximab and LPAAT-β inhibitors may overcome cytotoxic resistance and lead to a more efficacious therapy for non–Hodgkin's lymphoma patients. Our results using athymic mice bearing human non–Hodgkin's lymphoma xenografts treated with both Rituximab and first-generation LPAAT-β inhibitors suggest that these agents may have additive effects when administered in combination. However, due to the toxicity of compounds, such as CT-322228 at doses required for adequate tumoricidal activity, additional classes of compounds with enhanced therapeutic effectiveness will need to be identified before clinical development can be undertaken. Additional screening of diversity libraries is under way to identify LPAAT-β inhibitors that induce less gastrointestinal apoptosis, while maintaining or augmenting tumoricidal activity.

We have documented that CT-32228, a first-generation LPAAT-β inhibitor used at high doses, induced apoptosis of cells in the intestinal epithelial lining, as previously seen with alterations in lysophosphatidic acid levels (28, 30). Given the described trophic effects of lysophosphatidic acid on intestinal epithelium (31), it is possible that altering the route of feeding or the complexity of the diet may influence the intestinal cell proliferative and barrier functions when exposed to LPAAT-β inhibitors (32). Thus, future studies using LPAAT-β inhibitors will explore the concomitant use of specific nutrients and diet-derived compounds, such as glutamate, glycine, and zinc oxide, as well as polypeptides such as teduglutide (glp-2), which has been shown to alter intracellular G protein signaling pathways and provide a cytoprotective effect in animal models of colon injury (33). Moreover, because the most severe apoptotic abnormalities were present in the large intestine, the potential role of luminal bacteria or bacterial products should be evaluated. Prior studies using elective microbial decontamination has been effective in protecting against mucositis associated with radiation therapy in patients with solid tumors (34, 35). We will also investigate the use of intraluminal antibiotics that change the gut flora as a tool to protect against colonic mucosal injury after LPAAT-β inhibitor administration. These potential improvements in administration of the LPAAT-β inhibitors, as well as the future identification of compounds with a superior therapeutic index, may lead to a novel therapeutic approach targeting the ras/raf pathway to provide enhanced therapeutic effects through apoptotic mechanisms.

We thank Alexey Ball for his expert technical assistance.