Up-regulation of CDK9 kinase activity and Mcl-1 stability contributes to the acquired resistance to cyclin-dependent kinase inhibitors in leukemia

Abstract

Flavopiridol is a small molecule inhibitor of cyclin-dependent kinases (CDK) known to impair global transcription via inactivation of positive transcription elongation factor b. It has been demonstrated to have significant activity predominantly in chronic lymphocytic leukemia and acute myeloid leukemia in phase I/II clinical trials while other similar CDK inhibitors are vigorously being pursued in pre-clinical and clinical studies. Although flavopiridol is a potent therapeutic agent against blood diseases, some patients still have primary or acquired resistance throughout their clinical course. Considering the limited knowledge of resistance mechanisms of flavopiridol, we investigated the potential mechanisms of resistance to flavopiridol in a cell line system, which gradually acquired resistance to flavopiridol in vitro, and then confirmed the mechanism in patient samples. Herein, we present that this resistant cell line developed resistance through up-regulation of phosphorylation of RNA polymerase II C-terminal domain, activation of CDK9 kinase activity, and prolonged Mcl-1 stability to counter flavopiridol's drug actions. Further analyses suggest MAPK/ERK activation-mediated Mcl-1 stabilization contributes to the resistance and knockdown of Mcl-1 in part restores sensitivity to flavopiridol-induced cytotoxicity. Altogether, these findings demonstrate that CDK9 is the most relevant target of flavopiridol and provide avenues to improve the therapeutic strategies in blood malignancies.

Trial registration: ClinicalTrials.gov NCT00098371.

Conflict of interest statement

CONFLICT OF INTEREST

Drs. Byrd and Grever have a use patent on flavopiridol that has not been awarded and currently lacks financial value.

Figures

Figure 1. Cells acquired resistance to flavopiridol-induced cell death in vitro

697 parental and flavopiridol-resistant (Flavo-R) cells were treated with continuous flavopiridol (0.2μM or 0.3μM), or 1μM dinaciclib with 2-hour exposure and washout (w/o). Cell viability was measured post 24 hours by annexin V-FITC and PI-PE stains, followed by flow cytometry. Flavo-R also develops cross-resistance to dinaciclib with the significant survival advantage over parental cells at 24 hours post to dinaciclib treatment. Due to the similar effects of continuous 0.2μM, 0.3μM flavopiridol, p-values represent the average effect for both doses.

Figure 2. Phosphorylation of Ser2 of RNA Pol II CTD is more resistant against flavopiridol's drug action

697 parental and Flavo-R cells were treated with either 2μM flavopiridol with 4-hour exposure and washout (w/o), or 1μM dinaciclib with 2-hour exposure and washout (w/o) and collected at various time points as indicated in the figure. Protein lysates were prepared and subjected to immunoblotting for phosphorylation of Ser2 of RNA Pol II, total RNA Pol II and actin. Consistently, Flavo-R reveals more robust Ser2 phosphorylation with flavopiridol and dinaciclib, implicating higher activity of RNA Pol II. It also suggests that Flavo-R mechanistically establishes the resistance to dinaciclib in vitro in concert with observations in cell cytotoxicity described in Figure 1. Densitometry was applied to quantify the intensity of immunoreactive bands for phospho-Ser2 of RNA Pol II and, which was normalized to total RNA Pol II and the arbitrary numbers are shown at the bottom of the figure.

Figure 3. CDK9 kinase activity is upregulated to promote RNA Pol II activity, counter to the drug mechanism of flavopiridol

697 parental and resistant cells were treated with 4hr-exposure of 2μM flavopiridol and harvested for protein lysate at pre (0hr), 2, 4 and 6hr. Lysates were subjected to immunoblotting for phospho-Thr186 in the CDK9 activation loop as a surrogate for CDK9 kinase activity. Densitometry was applied to quantify the intensity of immunoreactive bands for phospho-Thr186 of both CDK9 isoforms, which were normalized to total CDK9 and the arbitrary numbers are shown at the bottom of the figure. CDK9 kinase activity of both isoforms was further increased with flavopiridol in Flavo-R.

Figure 4. Reduction of CDK9-55 protein expression after flavopiridol therapy is correlated with the response in OSU-0491 clinical trial

(A) Protein expression of both CDK9 isoforms is reduced in normal B-cells. CD19+ B-cells were selected from healthy donors or CLL patients and lysed for western blotting. Although the heterogeneity of protein expression is observed, overall CDK9 protein expression in CLL is less than in normal B-cells. (B) Protein lysates were prepared at pre and 1-hour post flavopiridol treatment collected on the OSU 0491 clinical trial. Comparison of CDK9-55 protein levels between post and pre showed significant decrease in partial responders but not in non-responders (p=0.018).

Figure 5. Mcl-1 protein levels are more stable to antagonize the flavopiridol-mediated depletion in Flavo-R

(A) Cells were treated with continues exposure of 0.3μM flavopiridol or 2-hour exposure of 2μM flavopiridol with washout (w/o) and harvested for lysates and RNA preparation at pre (0 hr), 1, 2 and 4hr post treatment. Quantitative real-time PCR with TaqMan probes for Mcl-1 was used to measure its transcript abundance after treatment. Flavo-R showed significantly more Mcl-1 transcripts with both doses of flavopiridol at 6-hr time point. (B) Immunoblotting was applied to detect Mcl-1 protein levels in protein lysates collected from cells treated with 2μM flavopiridol for 4 hours and washout (w/o). (C) Densitometry is utilized to quantify the intensity of immunoreactive bands for Mcl-1 that is normalized to GAPDH and the bar graph shows the average of densitometry measurement of three independent experiments. Mcl-1 protein expression is more stable with the flavopiridol treatment in these resistant cells. (D) Cells were treated with 100μg/ml cycloheximide (CHX) to inhibit overall protein translation and compared with Mcl-1 protein stability between parental cells and Flavo-R. Treated cells were collected at pre (0h), 2, 4, and 6hr to assay short-life Mcl-1 protein levels by immunoblotting. (E) The bar graph shows the average of densitometry measurement of the intensity of immunoreactive bands for Mcl-1, which was normalized to GAPDH in three independent experiments. Levels of Mcl-1 protein expression are more stable in Flavo-R.

Figure 6. MAPK/ERK-mediated Mcl-1 stabilization contributes to resistance of flavopiridol and shRNA knockdown of Mcl-1 restores partial sensitivity to flavopiridol

(A) Cells were treated with 4-hours 2μM flavopiridol with washout, harvested for protein lysates at pre (0h), 2, and 4hr, and yielded to immunoblotting for scrutinizing MAPK/ERK activation, which may contribute to preventing Mcl-1 from degradation. Phosphorylation of ERK was assayed by phospho-ERK1/2 antibodies to indicate the activation status of ERK kinase activity. (B) Densitometry was used to quantify the intensity of immunoreactive bands for phospho-ERK1/2, which was normalized to total ERK and GAPDH and the bar graph shows the average of densitometry measurement of three independent experiments. (C) Ubiquitinated proteins were pulled down with the ubiquitin-affinity agarose, subjected to SDS-PAGE separation and immunoblotted with the Mcl-1 antibody. The level of ubiquitinated Mcl-1 is reduced in Flavo-R with 4-hour exposure of 2μM flavopiridol, implicating that Mcl-1 in Flavo-R is less subjected to ubiquitin-conjugated degradation. (D) Stable clones for Mcl-1 and control knockdown in 697 and Flavo-R cells were treated with 0.3μM or 0.5μM flavopiridol continuously and assayed for the cellular apoptosis 24-hours post treatment with annexin V-FITC and PI-PE stains analyzed by flow cytometry.