Three PDE4B-high cell lines–namely OCI-Ly1 (RRID: CVCL_1879), OCI-Ly10 (RRID: CVCL_8795), and Ramos (RRID: CVCL_0597)–were kindly provided by Dr. Ricardo Aguiar, University of Texas, Health Science Center at San Antonio (UTHSCSA). All cell lines were authenticated using short tandem repeat profiling (Cosmogen Tech, Seoul, Korea) within the last 3 years and all experiments were performed using mycoplasma-free cells. Cells were cultured in RPMI medium (GIBCO) supplemented with 10% fetal bovine serum (Hyclone, Radnor, PA), 1% N-2-hydroxyethylpiperanzine-N’-2-ethanesulfonic acid (HEPES), 1% L-glutamine, and 1% penicillin/streptomycin (PENSTREP) at 37°C in a 5.0% CO₂ incubator. The primary and secondary antibodies used for Western blot were as follows: anti-phospho-4EBP1 (1:1,000, Cat#9459, Cell Signaling Technology, Danvers, MA), myeloid cell leukemia 1 (MCL1; 1:1,000, sc-819, Santa Cruz Biotechnology, Santa Cruz, CA), β-actin (1:5,000, sc-47778, Santa Cruz Biotechnology), HRP-conjugated anti-rabbit (1:10,000, A120-101p, Bethyl, Montgomery, AL), and anti-mouse secondary antibodies (1:10,000, A90-116p-33, Bethyl). The reagents used in this study were cytarabine (JW-Life Science, Dangjin, Korea), forskolin (cat# BML-CN100-0010, Enzo Life Science, Farmingdale, NY), roflumilast (cat# 2675, BioVision, Milpitas, CA), and JC-1 dyes (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide, KA1-324, Abnova, Walnut, CA).

The three PDE4B-high cell lines (Ly1, Ly10, and Ramos) were plated in 96-well micro plates at a density of 3×10⁵ cells/well each and treated with cytarabine (0, 0.5, 1, and 2 μM), forskolin/roflumilast (0, 10, 20, and 40 μM/40 μM), or their combination as indicated for 48 hours. Cell viability was estimated using the Cell Titer 96 Aqueous Non-Radioactive Cell Proliferation Assay (MTS; Promega, Madison, WI). The MTS solution was added to the cells at 30 μL per well and the plate was incubated for 2 hours. The absorbance was recorded at 450 nm with a GloMax Microplate multi-reader (Promega).

Protein levels were analyzed using western blot. Briefly, cells were harvested and washed with 1× phosphate buffered saline and lysed in a lysis buffer (RIPA buffer), Na-vanadate (1 mM), β-glycerol phosphate (50 mM), protease inhibitor (cat#786-108, G-biosciences, St. Louis, MO), EDTA (5 mM), β-mercaptoethanol (cat#41300000-1, 142 mM, BioWORLD, Dublin, OH). The cell lysate was mixed with 5× sample buffer and the mixture was boiled at 100°C for 10 min. Subsequently, the samples were loaded on 12% and 15% polyacrylamide gels. Proteins were transferred to membranes using Mini Trans-Blot Cell and Criterion Blotter (Bio-Rad, Hercules, CA) and the membranes were blocked in 1% BSA (bovine serum albumin, cat#160069, MP Biomedicals, Santa Ana, CA) dissolved in Tris-buffered saline containing 0.1% Tween 20 (TBST). The membranes were probed by indicated primary antibodies and incubated on a rotator at 4°C overnight. Then, the membranes were washed for 5 minutes in TBST and treated with anti-mouse or -rabbit secondary antibodies for 1 hour at room temperature. After washing three times for 10 minutes each with TBST, proteins were detected using a chemiluminescent substrate (EZ west Lumi Plus, ATTO, Tokyo, Japan) and visualized on a chemiluminescence Imaging system (Luminograph II, ATTO).

The cell lines (Ly1, Ly10, and Ramos) at 1.0×10⁶ cells/well were treated with forskolin/roflumilast (20 μM/20 μM), cytarabine (0.5 μM), or their combination and incubated for 48 hours. Before staining, JC-1 dye was diluted 1:10 by RPMI medium. Then, 20 μL of JC-1 was added to each well and incubated for 20 minutes at 37°C. Using a fluorescence microscope (Olympus, Tokyo, Japan), JC-1-stained cells were detected under the 520 nm optical filter. All stained cells and red-stained cells were counted and the ratio of red-stained cells to all stained cells was calculated using Microsoft Excel.

For the clinical evaluation of our regimen, a pilot study was conducted. According to our center’s policy, relapsed/refractory NHL patients were treated with ESHAP (etoposide, cisplatin, methylprednisolone, and cytarabine), DHAP (dexamethasone, cytarabine, and cisplatin), or ICE (ifosfamide, carboplatin, and etoposide) chemotherapy regimens. Based on the findings our in vitro experiments, we selected ESHAP regimen in combination with roflumilast for clinical evaluation.

From September 2017 to May 2018, patients who received ESHAP chemotherapy combined with roflumilast were enrolled. Patients received roflumilast 500 mg once daily from the beginning of the first cycle of ESHAP chemotherapy to the end of the treatment. Dose modification including roflumilast and supportive care for toxicity were performed at the discretion of the investigator. For comparative analysis, medical records of patients who received original ESHAP chemotherapy from 2013 to 2018 were retrospectively reviewed. Criteria for eligibility included histologically confirmed DLBCL, which was relapsed/refractory after R-CHOP chemotherapy; age ≥ 18 years; and Eastern Cooperative Oncology Group (ECOG) performance score of 0–2. Patients with primary DLBCL of the central nervous system, human immunodeficiency virus infection, or post-transplant lymphoproliferative disorder were excluded.

Patient records of diagnosis, treatment outcomes, and safety were collected and analyzed. Treatment response was assessed using the National Cancer Institute (NCI)–sponsored Working Group guidelines. Safety was evaluated by reviewing the incidence, severity, relationship, and type of adverse events that occurred during the treatment and follow-up periods using the revised NCI Common Terminology Criteria for Adverse Events, ver. 4.03.

Combination index (CI) values were estimated using the CompuSyn software (ver. 1.0, ComboSyn, Inc., Paramus, NJ). We used the percent inhibition of cell viability resulting from drug combinations (fraction affected, Fa) versus individual agents. To identify statistically significant differences (p < 0.05), one-way ANOVA was performed using Excel 2007 (Microsoft) and Prism (GraphPad, San Diego, CA).

Chi-square test or Fisher exact test were used to compare categorical variables. Continuous variables were examined using t test or u test depending on data distribution. Progression-free survival (PFS) was calculated from the date of treatment start to the date of documented disease progression, the last follow-up visit, or death if the disease had not progressed until the time of investigation. Overall survival (OS) was calculated from the date of salvage chemotherapy start until either death or the date last known to be alive. PFS and OS curves were constructed using the Kaplan-Meier method. For all analyses, p < 0.05 was considered statistically significant. SPSS ver. 22.0 (IBM Corp., Armonk, NY) was used for all clinical data analyses.

To identify which of the drugs used for salvage chemotherapy of DLBCL had synergistic effect with roflumilast, the Ly1 cell line was treated with roflumilast (0 or 40 μM) in combination with forskolin (0, 10, 20, or 40 μM) and one of the six chemotherapy agents (cytarabine, cisplatin, mitoxantrone, etoposide, carboplatin, ifosfamide) commonly used for salvage chemotherapy of DLBCL. The treatment was conducted for 48 hours and the MTS assay was performed to estimate cell viability. The data were normalized to the respective values in vehicle-treated cells (no forskolin, roflumilast, or chemotherapy agents). Among these six drugs, cytarabine showed statistically significant synergistic effect when combined with roflumilast. In the case of ifosfamide and carboplatin, cell viability tended to increase when used alone, but it tended to decrease when used with roflumilast (Fig. 1).

To confirm the efficacy of the combinatorial effects of cytarabine and roflumilast on cell viability, three cell lines (Ly1, Ly10, and Ramos) were treated with forskolin/roflumilast, cytarabine, or their combination for 48 hours, and the MTS assay was performed. Treatment with forskolin/roflumilast or cytarabine alone suppressed cell viability, but their combinatorial treatment inhibited cell viability to a much greater extent when compared with each single treatment (Fig. 2A). To quantitatively determine the interactions between forskolin/roflumilast and cytarabine or the drug-induced cytotoxic synergy, CI values were calculated. These values indicated that combinatorial treatment with forskolin/roflumilast and cytarabine had strong and very strong synergy in Ly1 and Ly10, respectively, but had moderate synergy in Ramos (Fig. 2B). In conclusion, co-administration of cytarabine and roflumilast has synergistic effects on the inhibition of B-cell lymphoma proliferation than single administration of either drug.

Previous studies have shown that PDE4 inhibitors and cytarabine inhibit mammalian target of rapamycin (mTOR) signaling [11,12]. To investigate the combinatorial effects of forskolin/roflumilast and cytarabine on the mTOR signaling pathway and its downstream target MCL1, three B-cell lymphoma cell lines—Ly1, Ly10, and Ramos—were treated with forskolin/roflumilast, cytarabine, or a combination of the two. Western blot analysis of the downstream substrates of the mTOR signaling pathway was then performed. Combinatorial treatment using forskolin/roflumilast and cytarabine markedly downregulated phospho-4EBP1 and MCL1 compared to the individual treatments (Fig. 3). These results suggest that co-administration of forskolin/roflumilast and cytarabine is more effective for targeting the mTOR/MCL1 pathway in B-cell lymphoma.

Given that pro-survival MCL1 is synergistically repressed by forskolin/roflumilast and cytarabine, we assessed if this combination influences mitochondrial membrane potential (MMP). To measure MMP, the three cell lines—Ly1, Ly10, and Ramos—were treated with forskolin/roflumilast, cytarabine, or their combination for 48 hours. MMP was then assessed using JC-1 fluorescence staining. At high MMP, JC-1 forms an aggregation and emits red fluorescence, whereas at low MMP, it exists as a monomer and emits green fluorescence. As shown in Fig. 4A, forskolin/roflumilast or cytarabine alone had little impact on green fluorescence intensity compared with the control group in all three cell lines. However, combinatorial treatment decreased red fluorescence intensity and increased green fluorescence intensity in comparison with the individual treatments. To quantify the extent of apoptosis, all stained cells were counted and the ratio of the number of red-stained (live) cells to the number of all stained cells was determined. Quantification analysis confirmed that combinatorial treatment significantly reduced red-stained cells than individual treatments (Fig. 4B). These results indicate that co-administration of forskolin/roflumilast and cytarabine induces more intensive apoptosis in B-cell lymphoma cells than their single administration by disrupting MMP.

During the study period, 13 patients were enrolled in the roflumilast group and received ESHAP chemotherapy in combination with roflumilast as a salvage treatment. As a control group, medical records of 26 patients with relapsed/refractory DLBCL treated with ESHAP regimen were collected. Patient characteristics of the two groups at diagnosis and occurrence of relapsed/refractory disease are presented in Table 1. There were minor differences regarding patient characteristics between the two groups; however, these differences were not significant. The median age of the roflumilast group was 10 years greater than that of the control group (69 years vs. 59 years, p=0.093). In addition, the proportion of patients older than 65 years was higher in the roflumilast group (69.2% vs. 42.3%); however, this was not statistically significant. About 70% of the patients showed ECOG performance score of 0 or 1 before salvage chemotherapy. There was no significant difference in International Prognostic Index risk score and overall response rate (ORR) for R-CHOP chemotherapy between the two groups. Two of the 13 patients (15.4%) in the roflumilast group and eight of the 26 patients (30.8%) in the control group had refractory disease.

ORR of the roflumilast group (76.9%) was higher than that of the control group (53.8%). Six patients (46.2%) of the roflumilast group and nine patients (34.6%) of the control group achieved complete remission. More patients in the control group did not respond to ESHAP chemotherapy (23.1% vs. 42.3%). For median follow-up duration of 12.9 months, 46.2% of the patients from the roflumilast group and 76.9% of the control group experienced relapse or refractoriness to ESHAP chemotherapy (p=0.055). In the analysis of disease progression, median PFS showed no significant difference in the two groups (6.9 months vs. 5.7 months). For 1-year PFS rate, the roflumilast group showed better, though not statistically significant, results (50.0% vs. 25.9%) compared with the control group. In the survival analysis, there was no significant difference between the two groups regarding median OS (13.4 months vs. 12.9 months) and 1-year OS rate (53.8% vs. 53.8%). Further results of survival are shown in Table 2 and Figs. 5, 6.

We collected data on the adverse events associated with 50 cycles of chemotherapy for the roflumilast group and 54 cycles for the control group. We could collect adverse event records from 54 of 90 total cycles in the control group. Hematologic adverse events were not significantly different between the two groups. Hepatopathy (18.0% vs. 5.6%) and renal failure (16.0% vs. 1.9%) were more common in the roflumilast group, but there was no grade 3 or 4 event. The most common non-hematologic adverse events observed among the patients in the roflumilast group were asthenia, anorexia, nausea, and dizziness; most events were of grade 1 or 2. The most common non-hematologic adverse events of grade 3 or higher observed among the patients in the roflumilast group were anorexia and diarrhea. Adverse events leading to treatment-related mortality were observed in one case (7.7%) in the roflumilast group and in three cases (11.5%) in the control group. The occurrence of adverse events according to treatment group is summarized in Table 3.