CDK9 inhibition as an effective therapy for small cell lung cancer

Abstract

Treatment-naïve small cell lung cancer (SCLC) is typically susceptible to standard-of-care chemotherapy consisting of cisplatin and etoposide recently combined with PD-L1 inhibitors. Yet, in most cases, SCLC patients develop resistance to first-line therapy and alternative therapies are urgently required to overcome this resistance. In this study, we tested the efficacy of dinaciclib, an FDA-orphan drug and inhibitor of the cyclin-dependent kinase (CDK) 9, among other CDKs, in SCLC. Furthermore, we report on a newly developed, highly specific CDK9 inhibitor, VC-1, with tumour-killing activity in SCLC. CDK9 inhibition displayed high killing potential in a panel of mouse and human SCLC cell lines. Mechanistically, CDK9 inhibition led to a reduction in MCL-1 and cFLIP anti-apoptotic proteins and killed cells, almost exclusively, by intrinsic apoptosis. While CDK9 inhibition did not synergise with chemotherapy, it displayed high efficacy in chemotherapy-resistant cells. In vivo, CDK9 inhibition effectively reduced tumour growth and improved survival in both autochthonous and syngeneic SCLC models. Together, this study shows that CDK9 inhibition is a promising therapeutic agent against SCLC and could be applied to chemo-refractory or resistant SCLC.

Fig. 1. Mouse and human SCLC and NSCLC cells show different sensitivities to dinaciclib.

A Viability of mouse SCLC cell lines as measured with Cell TiterGlow expressed as % of untreated control (100%) after 30-h treatment with different concentrations of dinaciclib: 1, 5, 10, 20, 25, 30, 50 and 100 nM. Mean + SD, n = 3. B Viability of human SCLC and NSCLC cell lines after 30-hour treatment 1, 5, 10, 30, 50 and 100 nM of dinaciclib. Mean + SD, n = 3. C Viability of mouse SCLC compared to NSCLC after 30-h treatment with 50 nM of dinaciclib. Mean + SD, n = 3. D Mouse SCLC cells were lysed with RIPA buffer after 18, 24 and 30 h treatment with dinaciclib (50 nM) or vehicle. Representative blots of 3 independent experiments. E Human SCLC cells were lysed with RIPA buffer after 18 h treatment with dinaciclib (50 nM) or vehicle. Representative blots of 3 independent experiments. p- phospho-, cl. cleaved. F Percentage of PI-positive cells after treatment with 50 nM dinaciclib and 5 µM emricasan (EM) as measured by Incucyte. Mean + SD, n = 3. G Dynamic BH3 profiling after 96 h incubation with 25 nM dinaciclib with the indicated peptides BIM, BAD (BCL-2, BCL-xL, BCL-W dependence), HRK (BCL-xL dependence) and MS1 (MCL-1 dependence). Results expressed as ∆% priming, representing the increase in priming compared to non-treated cells. Values indicate mean ± SEM from at least three independent experiments. Paired t- test of dinaciclib treated vs. ctrl in each condition, **p < 0.01 and *p < 0.05.

Fig. 2. CDK9 inhibition shows no synergy with standard chemotherapy in SCLC.

A Human cell lines were treated with increasing doses of dinaciclib (1, 5, 10, 25, 50 and 100 nM) for 30 h in the presence or absence of a combination of cisplatin and etoposide. H526: Cis 0.5 µM, Et 1 µM. H1694: Cis 1 µM, Et 0.5 µM. H2171: Cis 40 µM, Et 75 µM. Mean + SD, n = 3. Viability was measured by CTG and expressed as a percentage of the viability of control. B Human cell lines were treated with increasing doses of cisplatin and etoposide for 30 h in the presence or absence of 4 nM dinaciclib. Mean + SD, n = 3. C SCLC cells were lysed with RIPA buffer after 6 and 12 h of treatment with either 25 nM dinaciclib and/or cisplatin and etoposide. H526 & H1694: Cis 20 µM, Et 50 µM. H2171: Cis 40 µM, Et 100 µM. Representative blots of 3 independent experiments. Cis cisplatin, Et etoposide, cl. cleaved.

Fig. 3. Dinaciclib treatment efficiently kills SCLC cells with acquired resistance to chemotherapy.

A SCLC cells with acquired resistance to cisplatin (Cis R) and the naïve parental cell line (N) were treated with increasing doses of cisplatin (0.5, 1, 2.5,10, 25, 50 and 75 µM) for 30 h. Mean + SD, n = 3. B SCLC cells with acquired resistance to etoposide (Et R) and the naïve parental cell line (N) were treated with increasing doses of etoposide (1, 2, 5, 20, 50 and 100 µM) for 30 h. C Resistant and naïve cell lines were treated for 30 h with increasing doses of dinaciclib (1, 5, 10, 25, 50 and 100). Mean + SD, n = 3. Viability was measured by CTG and expressed as a percentage of the viability of control. Two-way ANOVA with Geisser-Greenhouse correction. *p-adj<0.05, **p-adj < 0.01.

Fig. 4. A new class of CDK9 inhibitor.

A Chemical structure of the compound VC-1. B Percentage of inhibition of CDK9 activity calculated from a fluorescence polarisation shift assay. C Radiometric selectivity profile assay measuring kinase activity of 16 CDK/Cyclin pairs. Mean + SD, n = 2. D, E Viability was measured by CTG and expressed as a percentage of the viability of Control after 30-hour treatment with different concentrations of VC-1 (100, 200, 400, 600, 800, 1000 and 1400 nM) Mean + SD, n = 3. F Percentage of PI-positive cells after treatment with 1 000 nM VC-1 and 5 µM emricasan (EM) as measured by Incucyte. Mean + SD, n = 3. G Human SCLC cells were lysed with RIPA buffer after 18 h of treatment with 1000 nM VC-1 or vehicle. Representative blots of 2 independent experiments. H IC₅₀ [μM] of twelve SCLC and eleven NSCLC human cell lines after 72 h of treatment with VC-1. Unpaired t-test NSCLC vs. SCLC **p = 0.0034.

Fig. 5. CDK9 inhibition significantly reduces tumour growth and prolongs survival.

A Survival curve for mice bearing tumours treated with dinaciclib. Log-rank (Mantel–Cox) test. *p = 0.0405. Treatment began upon tumour establishment with either dinaciclib (30 mg/kg) or vehicle (10% Hydroxypropyl Beta Cyclodextrin) twice per week, followed by a week of drug holiday until endpoint criteria were met (tumour volume >800 mm³). Tumours were measured every 14 days. Vehicle treated n = 6, dinaciclib treated n = 8. B Volume of autochthonous SCLC tumours as measured by MRI and the Horos software. C Representative images of MRI scans of vehicle and dinaciclib-treated mice since tumour volume reached >1 mm³ and 45 days after. D Weight of C57BL/6 mice treated with a single i.p. injection of 40 mg/kg of VC-1 (Acute) or three injections per week of 20 mg/kg (Chronic). E Percentage of liver weight with respect to the body weight of each mouse on day 14 of D. F 1380 cells were injected subcutaneously on the flank of C57BL/6 mice. Treatment consisted of either VC-1 (20 mg/kg), dinaciclib (30 mg/kg) or vehicle (10% Hydroxypropyl Beta Cyclodextrin) three times per week (VC-1) or twice per week, followed by a week of drug holiday (dinaciclib). Tumours were measured three times per week, n = 7 for each group. G Individual tumour volumes of D at day 21 expressed as fold change from day 0. One-way ANOVA, Dunnet’s multiple comparison test. *p-adj. < 0.05.