Purpose: Inhibitor of apoptosis proteins (IAP) promote cancer cell survival and confer resistance to therapy. We report on the ability of second mitochondria-derived activator of caspases mimetic, birinapant, which acts as antagonist to cIAP1 and cIAP2, to restore the sensitivity to apoptotic stimuli such as TNF-α in melanomas.
Experimental Design: Seventeen melanoma cell lines, representing five major genetic subgroups of cutaneous melanoma, were treated with birinapant as a single agent or in combination with TNF-α. Effects on cell viability, target inhibition, and initiation of apoptosis were assessed and findings were validated in 2-dimensional (2D), 3D spheroid, and in vivo xenograft models.
Results: When birinapant was combined with TNF-α, strong combination activity, that is, neither compound was effective individually but the combination was highly effective, was observed in 12 of 18 cell lines. This response was conserved in spheroid models, whereas in vivo birinapant inhibited tumor growth without adding TNF-α in in vitro resistant cell lines. Birinapant combined with TNF-α inhibited the growth of a melanoma cell line with acquired resistance to BRAF inhibition to the same extent as in the parental cell line.
Conclusions: Birinapant in combination with TNF-α exhibits a strong antimelanoma effect in vitro. Birinapant as a single agent shows in vivo antitumor activity, even if cells are resistant to single agent therapy in vitro. Birinapant in combination with TNF-α is effective in a melanoma cell line with acquired resistance to BRAF inhibitors. Clin Cancer Res; 19(7); 1784–94. ©2013 AACR.
Although major advances have recently been made in the treatment of malignant melanoma, this disease remains largely incurable once metastasized. The main reason is the development of resistance, even to new therapies recently introduced into the clinic or currently in late stage testing. In addition, the genetic heterogeneity of melanoma often limits the use of effective targeted therapies to specific subsets of patients. Novel small-molecule inhibitors that could be applicable for all patients with melanoma, regardless of genetic subgroup, would be highly beneficial to the current standard of care. We describe here a novel second mitochondria-derived activator of caspases mimetic small molecule, birinapant, which shows antitumor activity in major subgroups of cutaneous melanoma. Moreover, birinapant could be useful in overcoming acquired resistance to BRAF inhibitors, a patient population that is in urgent need of new treatment strategies.
Treatment options for metastasized melanoma, a disease with a low 5-year survival rate, have improved remarkably in the last 2 years. After a decades-long period, in which the gold standard remained dacarbazine chemotherapy with modest response rates (1), two new therapies have been U.S. Food and Drug Administration-approved recently (2–4) and several are currently in late-stage clinical development (5). Among small-molecule inhibitors, the majority targets 2 major pathways: the mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K) pathways. Therapy through modulating T-cell responses (4) is also highly promising, but unfortunately for all new therapies, most patients eventually progress.
Antiapoptotic mechanisms are also frequently at work in melanoma (6) and an attractive target for therapeutic intervention. For example, targeting the B-cell lymphoma 2 (Bcl-2) family of antiapoptotic proteins using Bcl-2 antisense (7) and Bcl-2 homology domain 3 (BH3) mimetics (8, 9) has been promising preclinically, but will require additional studies to fully realize their potential in clinical trials.
One of the factors responsible for the difficulties in engaging the apoptotic cascade efficiently in melanoma is the upregulation of conserved inhibitor of apoptosis proteins (IAP; refs. 10, 11). IAPs are a family of proteins defined by the presence of baculoviral IAP repeats. The best described members are XIAP, cIAP1, cIAP2, ML-IAP, and survivin. These proteins are associated with chemoresistance and poor outcome in many cancer types [reviewed in ref. (12)]. IAPs themselves are controlled by the second mitochondria-derived activator of caspases (SMAC). SMAC is released from mitochondria upon onset of apoptosis (13) and binds directly to IAPs leading to their degradation (14, 15). A number of IAP inhibitors, designed to function as SMAC mimetics, have shown preclinical activity in a variety of cancer types (16–19). In contrast to XIAP, cIAP1 and cIAP2 can bind to caspases but lack binding domains capable of inhibiting caspases (20). Previous studies have shown that synthetic IAP antagonists are able to induce apoptosis in tumor cells primarily through inhibition of cIAP1 and or cIAP2, which results in changes in TNF receptor (TNFR) complex signaling. These effects were dependent on TNF-α signaling and caspase-8 activation (21). Moreover, SMAC mimetics can perturb NF-κB signaling downstream of TNFR (22).
Birinapant is a novel dimeric SMAC mimetic designed to specifically target cIAP1 and cIAP2 for degradation, resulting in a switch in TNFR signaling; activation of TNFR upon binding of TNF-α in the presence of cIAP1 and cIAP2 leads to NF-κB activation and increased cell proliferation. In contrast, following cIAP1 and cIAP2 inhibition, TNF-α signaling leads to activation of caspase-8 and induction of apoptosis (12).
Immune cell infiltrates are often found in melanoma lesions, leading to chronic inflammation and (among other factors) elevated levels of TNF-α (23, 24). TNF-α can then be used by melanoma cells to promote cell growth, invasion, and metastasis (25). Binding of TNF-α to its receptor leads to the activation of TNFR complex-I, a membrane-localized activator of the canonical NF-κB pathway composed of TNFRSF1A associated via death domain (TRADD), the ubiquitin ligases TRAF2, TRAF5, cIAP1, and cIAP2, and the protein kinase RIPK1 (26). This leads to increased cell proliferation via an ubiquitination cascade and downstream activation of NF-κB. After SMAC mimetic-induced degradation of cIAP1 and cIAP2, binding of TNF-α to its receptor leads to the formation of a death complex and activation of caspases. This results in a reversal of the role of TNF-α from promoting proliferation to inducing cell death (12).
In the present study, birinapant in combination with TNF-α led to a reduction of viability in the majority of 17 melanoma cell lines tested. This effect was dependent on TNF-α.
Our laboratory has previously published on melanoma cell lines with acquired resistance to BRAF inhibitors (27). Notably, birinapant in combination with TNF-α elicited the same antimelanoma effect in a parental and BRAF inhibitor-resistant cell line. This suggests a potential use of birinapant for the treatment of patients with BRAF inhibitor-refractory melanoma.
Materials and Methods
Cell lines and reagents
Human melanoma cell lines were cultured in Dulbecco's modified Eagle's medium supplemented with 5% FBS and grown at 37°C in 5% CO2. All cell lines were periodically authenticated by DNA finger printing using Coriell's microsatellite kit and tested for mycoplasma by PCR.
Genomic DNA from samples were analyzed for mutations in BRAF, CDK4, CTNNB1, FBX04, KIT, MEK1, MEK2, MET, and NRAS by the nucleotide extension assay using the iPlex platform (Sequenom, Inc; ref. 28).
Birinapant was provided by TetraLogic. Birinapant is a novel bivalent, selective small-molecule peptidomimetic of the Smac tetrapeptide that binds with high affinity to multiple members of the IAP family including XIAP and cIAP1. The dissociation constant (Kd) for XIAP and cIAP-1 was determined to be 45 and <1 nmol/L, respectively (29). Birinapant was developed by TetraLogic Pharmaceuticals and designed on the basis of the 4 N-terminal amino acids of Smac/DIABLO (Ala- Val-Pro-Ile) and caspase-9 (Ala-Thr-Pro-Phe; ref. 30).
Human recombinant TNF-α was obtained from Invitrogen. TNF-α monoclonal antibody (mAb) MAB610 was obtained from R&D Systems.
cis-Diamineplatinum(II) dichloride (cisplatin) was obtained from Sigma-Aldrich.
For immunoblot analyses, adherent cells were washed with cold PBS containing 100 μmol/L Na3VO4, scraped, collected by centrifugation, and quick-frozen in dry ice before lysis. Xenograft tumors were snap frozen in liquid nitrogen immediately after harvesting. Tumor chunks were ground on liquid N2 using a MM2 mixer mill (Retsch). Cells and powder were lysed and equal amounts of protein (10–40 μg) was subjected to SDS-PAGE and proteins transferred onto PVDF membranes (Immobilon). Antibodies used were: cIAP1, AF8181; cIAP2, AF8171 (both R & D systems); PARP, 7150; GAPDH, 32233 (both Santa Cruz biotechnology); Caspase-8, 9746; NF-κBp65, 3031 (both Cell Signaling); XIAP, ADI-AAM-050 (ENZO); NFKB2 p100- p52, 4882 (Cell Signaling); RIP1, 51-6559GR, RAC1, 610651 (both BD Biosciences); ERK, sc-271270, pERK, sc 7976, (both Santa Cruz biotechnology). Membranes were probed with primary antibodies overnight at 4°C, then further incubated with Alexa Fluor-labeled secondary antibodies (IRDye 680LT goat-anti mouse, IRDye 800CW goat-anti rabbit, donkey-anti mouse, donkey-anti rabbit, or donkey-anti goat IRDye 800CW or 680LT (LI-COR) for 1 hour and scanned using the Odyssey system (LI-COR).
In vitro drug sensitivity assays
For monolayer cell culture assays, cells were allowed to attach for 24 hours and subsequently incubated with birinapant and/or TNF-α for 24 or 72 hours. CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay (MTS) assay (Promega) was conducted according to the manufacturer's description. For cell-cycle analysis, melanoma cells were fixed in 70% ethyl alcohol and stained with propidium iodide. Samples were subsequently analyzed with an EPICS XL (Beckman-Coulter) apparatus. Annexin V staining was conducted with an Annexin V allophycocyanin conjugate (Invitrogen) according to the manufacturer's description. Briefly, cells were treated with dimethyl sulfoxide control, birinapant, and/or TNF-α for 24 hours. Resuspended cells were washed and incubated with the conjugate for 15 minutes and Annexin-binding buffer was added. Samples were subsequently analyzed with an EPICS XL (Beckman-Coulter) apparatus.
451Lu and WM1366 melanoma cells were treated with birinapant 1 μmol/L in combination with TNF-α 1 ng/mL. Cells were then incubated for 72 hours in the presence or absence of Z-VAD-FMK (R & D systems) a pan-caspase inhibitor. Proliferation was assessed using the MTS assay.
451Lu and WM1366 melanoma cells were treated with birinapant 1 μmol/L in combination with TNF-α 1 ng/mL. Cells were then incubated for 72 hours in the presence or absence of Necrostatin-1 (N9037, Sigma) a RIP1 kinase inhibitor. Proliferation was assessed using the MTS assay.
RIPK1 gene expression was evaluated using commercially available gene expression assays from Life Technologies. Briefly, RNA was isolated using Qiagen RNeasy kit and RNA was quantified and normalized for addition to cDNA reaction. cDNA was created using cDNA kit from Life Technologies (4368814) and real-time PCR was carried out for RIPK1 (assay ID Hs00169407_m1) and GUSB (assay ID Hs99999908_m1) as a housekeeping gene. For analysis, the relative quantity of RIPK1 to GUSB was calculated.
For 3-dimensional (3D) melanoma spheroid assays, cells were prepared as described previously (31, 32). Collagen-embedded spheroids were incubated with reagents for 72 hours. Spheroids were imaged using a Leica TCS SP2 confocal microscope after addition of a live/dead viability assay kit using the fluorescent dyes calcein AM and ethidium homodimer-1 (Invitrogen). To assess viability, spheroids were grown in a 96-well format, and allowed to grow in the presence of birinapant and/or TNF-α for 72 hours, then AlamarBlue (Invitrogen) was added. Fluorescence was measured using an Envision Xcite plate reader (Perkin Elmer). All experiments were carried out in triplicate.
To measure TNF-α levels in cell culture supernatants, cells were grown for 3 days to 70% confluence. Supernatants were collected and 200 μL were tested in triplicate following the manufacturer's instructions using the BD OptEIA Human TNF-α ELISA Kit II (BD).
All animal experiments were carried out in accordance with Wistar IACUC protocol 111954 in NUDE mice. Ten animals each were inoculated s.c. with 1 × 106 451Lu or 1205Lu human melanoma cells in a suspension of Matrigel (BD Matrigel Basement Membrane Matrix, Growth Factor Reduced)/complete media at a ratio of 1:1. After formation of palpable tumors, mice from both tumor models were randomized into 2 groups. Both groups were treated intraperitoneal 3 times/week with either vehicle control or birinapant 30 mg/kg for 21 days. Birinapant was dissolved in 12.5% Captisol (Ligand Pharmaceuticals) in distilled water. Tumor size was assessed twice weekly by caliper measurement. Subsequently, satellite groups of ten and fifteen mice were inoculated in the same fashion with 451Lu and 1205Lu tumor cells, respectively. After tumors had reached a mean volume of 200 mm3 animals were dosed with either birinapant or vehicle control as described above. After 48 hours and 2 doses, animals were sacrificed and tumors were harvested at 4 time points after the last treatment. Tumor samples were snap frozen in liquid nitrogen for subsequent protein analysis and preserved as formalin-fixed, paraffin embedded blocks for immunohistochemistry.
Paraffin-embedded 5 μmsections were stained with activated caspase-3 mAb (5A1E, Cell Signaling). Briefly, after deparaffinization and endogen peroxidase blocking, antigen retrieval was conducted via heat mediation with citrate buffer. The antibody was diluted 1:100 in Superblock and incubated overnight at 4°C. After washing with Dulbecco's Phosphate Buffered Saline slides were incubated for 30 minutes in ABComplex/HRP conjugated Kit (DAKO). Following a further 3 washing steps, the DAB-chromogen solution Kit (DAKO) was incubated on slides until visible color reaction was observed. Slides were washed and counterstained in diluted (1:4 in aqua dest.) Harris' hematoxlin solution and mounted with Aquatex (Merck).
IC50 values were calculated with GraphPad Prism 5 (GraphPad Software Inc.). Statistical significance was determined using 2-sided Student t test. P < 0.05 was considered to be significant.
The majority of cell lines exhibited strong combination activity of birinapant and TNF-α
Reflecting the heterogeneity of melanoma, we used a panel of 17 human melanoma cell lines comprising major genetic subgroups of cutaneous melanoma, approximately reflecting the distribution seen in patients (ref. 33; Supplementary Table S1).
To test birinapant as a single agent and calculate the half maximal inhibitory concentration (IC50) values in this cell line panel, cells were treated with birinapant at 1, 10, 100, and 1,000 nmol/L for 72 hours and cell viability was assessed through the MTS assay. Only one cell line (WM9, BRAFV600E) out of 17 showed a response to birinapant (Fig. 1, gray bars, TL32711). However, when birinapant was combined with TNF-α (1 ng/mL), the majority of cell lines showed birinapant IC50 values in the low nanomolar range (Fig. 1, black bars). This response was independent of the cell line genetic background. All cell lines were resistant to TNF-α alone (Supplementary Fig. S1).
Three distinct response patterns following birinapant addition were observed. For subsequent experiments, 4 cell lines were therefore selected on the basis of their response to birinapant in combination with TNF-α: 1205Lu, resistant to the combination therapy; WM9, sensitive to birinapant as a single agent; and 2 cell lines sensitive to birinapant only in combination with TNF-α: 451Lu (BRAFV600E) and WM1366 (NRASQ61L; Fig. 2A).
cIAP1 and cIAP2 downregulation was observed in all response types, but apoptosis occurred only in sensitive cell lines
Immunoblot analysis showed inhibition of the birinapant target protein cIAP1 in all 4 cell lines. Although cIAP2 was upregulated in the 1205Lu and WM1366 cell lines upon treatment with TNF-α, and was again degraded in the combination with birinapant. No change in the levels of XIAP could be observed (Fig. 2B). After treatment with birinapant alone, apoptosis as assessed by cleavage of PARP could be observed in the single-agent sensitive cell line only. When birinapant was combined with TNF-α, PARP cleavage was now detectable in the combination-sensitive cell lines 451Lu and WM1366 as well, corresponding to the growth inhibition observed in the MTS assay. No cleavage of PARP could be observed in the resistant cell line 1205Lu, even after combination treatment. Notably, a decrease in NF-κB p65 protein levels could only be observed in the single-agent–sensitive cell line, WM9, after treatment with birinapant alone or in combination with TNF-α. An increase in NF-κB2 p100 levels could be observed with TNF-α stimulation, and this could be brought back to baseline by the combination of birinapant and TNF-α. The p52 band of NF-κB2 was not detectable. No change in the protein levels of RIP1 kinase, which was found at similar levels in all 4 cell lines, could be observed upon treatment with birinapant or TNF-α alone, but RIP1 was depleted in 3 of 4 cell lines when both agents where combined (Fig. 2B).
We then confirmed the induction of apoptosis seen at the protein level through cell-cycle analysis by propidium iodide staining. A significant increase in sub-G1 fractions, indicative of apoptosis, was seen in WM9 cells when treated with birinapant alone; in 451Lu and WM1366 cells only when birinapant was combined with TNF-α; but not in 1205Lu cells, even after combination treatment (Fig. 2C). To confirm the indication of apoptosis seen in the cell-cycle analysis, Annexin V staining was conducted. The relative increase in Annexin V–positive cells seen with birinapant alone only in WM9 and with birinapant in combination with TNF-α in WM9, 451Lu, and WM1366, but not in 1205Lu cells, could confirm the pattern of apoptosis seen by cell-cycle analysis (Fig. 2D).
To investigate a possible role of MAPK and RAC1 signaling in resistance to birinapant, phosphorylation status of ERK1/2 and protein levels of RAC1 upon treatment with birinapant at different time points in birinapant-sensitive and -resistant cell lines were assessed. We could observe a marked decrease in phosphorylation of ERK1/2 only at 24 hours and this decrease was seen in a birinapant-sensitive as well as a birinapant-resistant cohort of cell lines. RAC1 protein was found at similar levels in both sensitive and resistant cohorts and there was no significant change upon birinapant treatment (Supplementary Fig. S2).
An immunoblot conducted in the 451Lu cell line showed that birinapant could cause sustained cIAP1 protein degradation within one hour at a dose of 100 nmol/l, whereas XIAP levels were not affected (Fig. 3A).
To confirm that birinapant resistance was not caused by failure of target degradation, we conducted immunoblots for cIAP1 and XIAP in a cohort of birinapant-TNF-α combination therapy–resistant cell lines and found cIAP1 protein levels to be depressed to similar levels as seen in the birinapant-sensitive cell line 451Lu (Supplementary Fig. S3).
Next, we investigated whether the observed cell death with birinapant at a dose of complete cIAP1 inhibition in combination with TNF-α was caspase dependent. Two combination-sensitive cell lines (451Lu and WM1366) were treated with the combination of birinapant and TNF-α with or without addition of Z-VAD-FMK, a pan-caspase inhibitor. In both cell lines, the inhibition of caspases led to a complete restoration of proliferation (Fig. 3B). In a similar experiment, we then further explored the role of RIP1 kinase, an essential part of the caspase-initiating complex activated by TNF-α. By treating the 2 cell lines as above, now in the absence or presence of necrostatin-1, a RIP1 kinase inhibitor, we could again reverse the effect of birinapant and TNF-α and restore cell proliferation (Fig. 3C).
As loss of RIP1 expression could therefore be a possible mechanism of resistance to birinapant, we assessed RIP1 expression in the whole panel of cell lines. However, no loss or decrease in RIP1 expression in the 6 birinapant-resistant cell lines compared with the sensitive group could be observed. This was also true after stimulation with TNF-α 1 ng/mL (Fig. 3D).
Three-dimensional spheroid models confirmed responses seen in adherent cell cultures
Cells grown in 3D spheroid cultures were previously shown to have altered drug response profiles compared with adherent cell cultures (34), and are expected to more accurately predict in vivo efficacy due to similar architecture and microenvironmental signals (35, 36). The 4 previously selected cell lines were grown as 3D spheroids in a collagen matrix and treated with birinapant alone or in combination with TNF-α. A live/dead fluorescent cell stain was used to visually assess treatment effects using confocal microscopy (Fig. 4A): Spheroids of the birinapant single-agent–sensitive cell line, WM9, did indeed show an extensive reduction in live cells after addition of birinapant, but not after addition of TNF-α alone. The combination-sensitive cell lines, 451Lu and WM1366, retained the same response patterns in 3D cultures: both showed a marked decrease in live cells and increase in dead cells only after treatment with birinapant in combination with TNF-α. In addition, the cell line that was completely resistant to the combination treatment in adherent cell culture, 1205Lu, showed only slight growth retardation when grown as spheroids in the presence of birinapant in combination with TNF-α.
To objectively quantify viability in this model, we assessed metabolic activity of spheroids after treatment with birinapant in combination with TNF-α using Alamar Blue. The viability results mirrored the responses seen in the Live/Dead assay: a near total loss of viability in WM9, a dramatic decrease in viability in the combination-sensitive cell lines (451Lu, WM1366), and only a slight reduction of viability in the 1205Lu cell line (Fig. 4B).
Birinapant inhibits tumor growth in melanoma xenotransplantation models as a single agent
To investigate whether birinapant could inhibit melanoma tumor growth in an in vivo setting as a single agent, 2 cell lines were selected for xenotransplantation experiments: both were in vitro birinapant single-agent–resistant, but 451Lu did respond in vitro to the combination of birinapant with TNF-α, whereas 1205Lu did not respond to the combination treatment in vitro. Both cell lines were inoculated subcutaneously in NUDE mice and allowed to form palpable tumors before being randomized into vehicle control and birinapant treatment groups. During 3 weeks of dosing, birinapant showed an antitumor effect in both models, although the effect in the in vitro combination-sensitive cell line was more sustained with abrogation of tumor growth in the birinapant-treated animals. In contrast, 1205Lu tumors showed a marked slowing of tumor growth, but not abrogation of tumors (Fig. 5A).
In a subsequent in vivo experiment, we then went on to confirm birinapant target inhibition in both models by immunoblot of tumor lysates. Animals were again inoculated with both xenograft models and tumors allowed to form. Animals were then pretreated twice in an interval of 48 hours and tumors were harvested 3, 6, 12, and 24 hours after the second dosing. Compared with vehicle control, cIAP1 protein was reduced to low levels 3 hours later, and this effect was sustained for 24 hours in both models (Fig 5B). Staining for activated caspase-3 in biopsies of the same tumors showed a modest increase in apoptotic cells in the birinapant-treated animals compared with vehicle control, 24 hours after treatment (Fig. 5C).
To further investigate the combination activity between birinapant and TNF-α in vitro, the 4 cell lines previously selected on the basis of their response profiles were again used. All 4 cell lines had comparable levels of TNF-α in cell culture supernatants (Supplementary Fig. S4) and adding high levels of TNF-α alone to all 4 cell lines had no effect on cell proliferation (Supplementary Fig. S5). The one cell line sensitive to birinapant as a single agent, WM9, provided an opportunity to further investigate the role of TNF-α in this setting. We therefore added a TNF-α–blocking monoclonal antibody to WM9 cell cultures before treatment with birinapant. By gradually binding endogenous TNF-α in the supernatant with increasing doses of the antibody, it was possible to completely abrogate the effect of birinapant in a dose-dependent manner (Fig. 6A). This indicates that while no addition of exogenous TNF-α was necessary to observe a birinapant effect in this cell line, the observed antitumor effect was still dependent on endogenous TNF-α.
Additional evidence on the causative role of TNF-α on the effect of birinapant was provided by varying the schedule of combining birinapant with TNF-α: a combination-sensitive cell line treated with birinapant for 36 hours and subsequently incubated with TNF-α for 36 hours, showed significant growth inhibition. Conversely, when treated first with TNF-α for 36 hours and then with birinapant for 36 hours, cells were significantly less sensitive (Fig. 6B). Therefore, the combination activity between birinapant and TNF-α is schedule dependent. This corroborates the hypothesis that birinapant-mediated IAP inhibition has to occur before the activation of TNFR to induce an antitumor effect.
Previously, our group has generated human melanoma cell lines with acquired resistance to BRAF inhibitors (27). The mechanism of resistance in these cell lines was RAF isoform switching and increased IGF-1R/PI3K signaling. We compared the cell lines 451Lu (BRAFV600E, BRAF inhibitor sensitive) and 451Lu-BR (BRAF inhibitor resistant) in regard to birinapant response. Cell viability was not affected after treatment with birinapant alone in neither the parental nor the resistant cell line. When birinapant was combined with TNF-α, however, both cell lines showed a strong response that was identical between the sensitive and the resistant cell line (Fig. 6C).
Next, we investigated whether birinapant could sensitize melanoma cells to chemotherapy. For this experiment, 2 birinapant single-agent–resistant cell lines were treated with cisplatin, a cytotoxic drug, with or without addition of birinapant. Cisplatin as a single agent reduced viability in both melanoma cell lines (mean of both cell lines) at increasing doses up to 20 μmol/L. Birinapant-mediated cIAP1 degradation in combination with cisplatin significantly (P < 0.05) enhanced sensitivity to cisplatin in both cell lines (Fig. 6D).
A delicate balance between inducers and inhibitors of apoptosis exists in any given cell (37). In cancer cells, including melanoma, this equilibrium is often skewed in favor of survival, with IAPs being predominant (38, 39). Committing melanoma cells to apoptosis, which would be an ideal outcome for most therapies, therefore requires additional stimuli. Inhibition of IAPs for one can restore the balance between survival and death signals, thereby facilitating programmed cell death in melanoma.
Chronic inflammation in the tumor microenvironment of melanoma lesions often leads to elevated levels of TNF-α, at least in part, provided by the tumor-infiltrating immune cells such as macrophages (23–25). Levels of tumor-infiltrating macrophages have been shown to be associated with aggressive disease (40). All melanoma cell lines tested were resistant to treatment with exogenous TNF-α as a single agent. This was in line with previous clinical experiences, where minimal antitumor effect and significant toxicity was observed (41, 42). Birinapant-mediated downregulation of cIAP1 by itself did not induce any antitumor effect in vitro. As neither compound was effective alone, but both were highly effective in combination, this satisfies the definition of “coalism” according to the so-called Saariselkä agreement (43, 44). This was confirmed on a subset of cell lines to be apoptosis-mediated and dependent on caspases and RIPK1 activity.
Independent of mutational background (we assessed melanoma cell lines from major genetic subgroups), 2 patterns of responses emerged: cell lines that were either exquisitely sensitive or remarkably resistant to birinapant in combination with TNF-α at low or high doses of both compounds respectively. We confirmed downregulation of cIAP1 target protein in resistant and sensitive cell lines by Western blot analysis, but only the sensitive cell lines showed PARP cleavage, indicative of apoptosis. This observation could also be shown by increased sub-G1 cell fractions and increased Annexin V–positive cells, indicating an increase in apoptotic cells, in the combination-sensitive cell lines.
As a subgroup of melanoma cell lines of all genetic backgrounds tested was found to be resistant to birinapant in combination with TNF-α, we investigated whether levels of phosphorylated ERK 1/2 or RAC1 were increased after birinapant treatment in these cell lines. These were recently described to play a role in resistance to SMAC mimetics (45, 46). Although we observed a decrease in ERK1/2 activation after 24 hours of birinapant treatment in most cell lines tested, we could not find any differential regulation of these 2 proteins in the resistant cell lines as compared with the sensitive subset. Another possible mechanism of resistance to SMAC mimetics is loss of RIP1 expression. We observed the effect of birinapant to be dependent on RIP1 kinase in concordance with previously published reports on compounds from the same class (47). As there was no difference in RIP1 expression between birinapant-resistant and -sensitive cell lines in the panel used in this study, RIP1 expression is likely not a suitable biomarker to predict melanoma response to birinapant. Therefore, further investigation to define a biomarker for this subgroup is warranted, as this would be highly useful in the clinical application of birinapant.
Our observations remained consistent across increasingly complex models from monolayer culture, to 3D matrix cultures, to human melanoma xenograft models. While the 3D spheroid models closely mirrored the effects seen in adherent monolayer culture, the in vivo xenotransplantation experiment results were reflecting the complexity of the in vivo setting. While in vitro 451Lu cells responded only to the combination of birinapant and TNF-α, birinapant was highly active as a single agent in the in vivo model abrogating tumor growth. In addition, a cell line resistant to birinapant in vitro even in combination with TNF-α still showed slower tumor growth when treated with birinapant in vivo compared with vehicle-treated controls. This observation indicates the high complexity of melanoma growth in a tissue microenvironment providing a multitude of additional stimuli. Together, these results indicate a potential effectiveness of birinapant as a single agent in patients with melanoma.
Birinapant was effective as a single agent in vitro only in one of the 17 cell lines tested. We therefore investigated the role of TNF-α signaling on the birinapant effect in this cell line. Blocking endogenous TNF-α in the supernatant completely abrogated the effect of birinapant in a dose-dependent fashion, thereby showing the dependency of birinapant on concurrent TNF-α stimulation in vitro.
This dependency was furthermore schedule-specific, as we could show in a cell line sensitive to birinapant only in combination with TNF-α. The effect of the combination therapy was preserved when sequentially added in the following order: first birinapant, with downregulation of cIAP1 and cIAP2, and subsequent TNF-α, with activation of the TNFR complex 2 and the apoptotic cascade. Adding these 2 compounds in the reverse order diminished their effectiveness significantly.
With the prospect of increased numbers of patients with melanoma acquiring resistance to BRAF inhibitors, the question whether birinapant might be a feasible second-line therapy for such cases was explored. Indeed, a cell line with acquired resistance to BRAF inhibition did not alter its sensitivity profile to birinapant when compared with the parental cell line. This suggests birinapant as a possible second-line therapy in patients with acquired resistance to BRAF inhibitors.
Although targeted therapy with small molecules is showing impressive results in melanoma therapy, cytotoxic agents are still playing a major role in this disease when patients are either not eligible for kinase inhibitors and immunotherapy, or have relapsed on these drugs. We have therefore investigated the role of birinapant in sensitizing melanoma to cisplatin, a DNA-damaging agent. Although the results seen in this study are encouraging, more combinations and schedules will have to be explored as these studies are currently ongoing.
Disclosure of Potential Conflicts of Interest
E.M. Neiman is employed (other than primary affiliation; e.g., consulting) in Tetralogic Pharmaceuticals as a Research Scientist. M. McKinlay has ownership interest (including patents) in TetraLogic Common Stock. No potential conflicts of interest were disclosed by the other authors.
Conception and design: C. Krepler, S.K. Chunduru, R. Amaravadi, M. McKinlay, M. Herlyn
Development of methodology: C. Krepler, S.K. Chunduru, M. McKinlay, M. Herlyn
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Krepler, M.B. Halloran, X. He, M. Xiao, J. Villanueva, E.M. Neiman, K.L. Nathanson
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Krepler, X. He, M. Xiao, Y. Mitsuuchi, C. Benetatos, K.L. Nathanson, R. Amaravadi, M. McKinlay
Writing, review, and/or revision of the manuscript: C. Krepler, S.K. Chunduru, A.M. Vultur, Y. Mitsuuchi, C. Benetatos, K.L. Nathanson, R. Amaravadi, H. Pehamberger, M. Herlyn
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.M. Vultur, K.L. Nathanson
Study supervision: M. Herlyn
The authors thank the Microscopy Facility managed by James E. Hayden, the Flow Cytometry Facility managed by Jeffrey S. Faust, the Animal Facility managed by Denise DiFrancesco, and Protein Expression and Molecular Screening Facility managed by David C. Schultz, all at the Wistar Institute.
C. Krepler is a MAX KADE foundation postdoctoral research exchange grant recipient. The studies were funded by NIH grants CA25874, CA047159, CA 114046, and CA10815, and the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation.