CDK9: A Comprehensive Review of Its Biology, and Its Role as a Potential Target for Anti-Cancer Agents

Abstract

Cyclin-dependent kinases (CDKs) are proteins pivotal to a wide range of cellular functions, most importantly cell division and transcription, and their dysregulations have been implicated as prominent drivers of tumorigenesis. Besides the well-established role of cell cycle CDKs in cancer, the involvement of transcriptional CDKs has been confirmed more recently. Most cancers overtly employ CDKs that serve as key regulators of transcription (e.g., CDK9) for a continuous production of short-lived gene products that maintain their survival. As such, dysregulation of the CDK9 pathway has been observed in various hematological and solid malignancies, making it a valuable anticancer target. This therapeutic potential has been utilized for the discovery of CDK9 inhibitors, some of which have entered human clinical trials. This review provides a comprehensive discussion on the structure and biology of CDK9, its role in solid and hematological cancers, and an updated review of the available inhibitors currently being investigated in preclinical and clinical settings.

Keywords: CDK9 inhibitors; CDKs; P-TEFb; cancer; transcription.

Figure 1

Timeline for the discovery of P-TEFb and its biological roles.

Figure 2

The protein structure of monomeric CDK9 (Protein Data Bank: 3BLQ). The bilobal CDK9 structure is dominated by N-terminal β-sheets (1-4 are shown) and C-terminal α-helices (D-H are shown). The C-terminal also contains β-sheets (6-9, not shown). The two lobes are connected by a hinge region (green) that binds the adenine moiety of adenosine triphosphate (ATP). The N-terminus contains an αC helix and glycine-rich loop (G-loop, purple) which binds cyclin and ATP, respectively. The C-terminus comprises the catalytic loop (yellow), T-loop (brown), and DFG motif that binds Mg⁺². The threonine residue (Thr186) involved in CDK9 activation is found in the T-loop structure.

Figure 3

Sequence comparison between CDK9 and CDK2. The sequence identity between the two proteins is 31.9%. Green color indicates residues conserved between CDK9 and CDK2. Red underlined residues indicate the different functional subunits of the kinases. In the T-loop, the phosphorylation of a conserved threonine residue (labelled red) is vital for the activation of both CDK9 (Thr186) and CDK2 (Thr160). The sequence alignment was generated and % sequence similarity determined using UniProt (https://www.uniprot.org/align/) and sequence identifiers were P50750 for CDK9 and P24941 for CDK2.

Figure 4

Structure of CDK9-Cyclin T1 complex. Stereo ribbon plot of the 3D structure of CDK9-cyclin T1 in complex with ATP (Protein data bank: 3BLQ). CDK9 (blue) and cyclin T1 (dark red) make contact through the αC helix and β4 strand of CDK9 and H3, H4, and H5 helices of cyclin T1. The CDK9 T-loop and hinge regions are indicated.

Figure 5

Control of transcriptional elongation by P-TEFb. During active transcription, BRD4 recruits JMJD6 to 7SK snRNP anchored to anti-pause enhancers on chromatin. JMJD6 demethylates both H4R3me and the 5’ hairpin of 7SK RNA, breaking chromatin binding of the former and exposing the latter for degradation (Labelled as 1). Concurrently, acetylated histone (H3KAC)-bound BRD4 interacts with and extracts P-TEFb from 7SK snRNP (2). Protein phosphatases (PP2B and PP1α) also assist in the release of P-TEFb from 7SK snRNP by dephosphorylating CDK9 pThr186 (3). After release, CDK9 is re-phosphorylated on Thr186 by CDK7 and delivered by BRD4-JMJD6 to RNAP II that has been paused in the proximal promoter region. At this site, P-TEFb phosphorylates DSIF, NELF, and RNAP II CTD (4), allowing productive elongation (5).

Figure 6

Formation and composition of 7SK snRNP. Following the folding of 7SK RNA into a four-stem loop structure, MePCE and LARP7 bind and protect its 5’ and 3’ ends, respectively, from catalytic degradation. One mechanism of protection involves capping of the 5’ end of 7SK RNA by MePCE (depicted as a black dot). The stable 7SK snRNP core then binds dimers of HEXIM1 which exposes their P-TEFb binding domains. Subsequently, HEXIM1 binds activated P-TEFb (CDK9 phosphorylated on Thr186, green dot) and this inhibits its kinase activity. During transcriptional activation, P-TEFb is released and 7SK snRNP is stabilized by binding to heterogeneous nuclear ribonucleoproteins (hnRNPs).

Figure 7

Domains and post-translational modifications of HEXIM1 (A), cyclin T1 (B), and CDK9 (B). Phosphorylated (green), ubiquitinated (purple), acetylated (yellow), and PYNT (brown) motifs are shown. Numbers indicate the positions of amino acid residues. BR, basic region; AR, acidic region, encompassing AR1 and AR2; CR, coiled-coil region. Solid black lines with indicated spans of amino acid residues indicate the regions responsible for interactions with other binding partners. Modified from Cho, Schroeder and Ott, Cell Cycle, 9 (9), 1697-1705 (2010).

Figure 8

Recruitment of P-TEFb from 7SK snRNP. Various signaling pathways and stress conditions liberate P-TEFb from 7SK snRNP through PTMs of the components of 7SK snRNP (HEXIM1 and cyclin T1) or direct recruitment (BRD4, TAT, or super elongation complex, SEC). The major PTM involves phosphorylation (green dot) of HEXIM1 in various residues and acetylation (yellow dot) of cyclin T1. TAT recruits P-TEFb by collaborating with phosphatases (PPM1G and PP2γ) which dephosphorylate CDK9 on Thr186. Meanwhile, BRD4 interacts with JMJD6, a histone demethylase, which demethylates (black dot) the 5’ end of 7SK RNA and destabilizes the inhibitory core. AFF1/4, ALL-fused gene from chromosome 1/4 family member; ELL2, Eleven-nineteen lysine-rich in leukemia; ENL, Eleven nineteen leukemia; AF9, ALL-fused gene from chromosome 9.

Figure 9

P-TEFb is required for the MLL transcription program and leukemogenesis. MLL is a histone methyltransferase ubiquitously expressed in hematopoietic progenitor cells and plays a key role in their self-renewal. For unknown reasons, MLL gene (on chromosome 11q23) undergoes a trans-locational mutation where its 5’ end is fused with the 3’ end of numerous genes. The majority of these partner genes are nuclear transcription factors that recruit P-TEFb and DOTL-1 leading to upregulated expression of HOX, MEIS1, and FLT3. These proteins drive leukemogenesis by blocking differentiation and driving active proliferation.