【Review】Research progress of selective CDK inhibitors

CDK is a class of serine/threonine kinases that play a key role
in cell cycle regulation and cell transcription.
The dysregulation of CDK activity is closely related to the occurrence and development of tumors, and the design of anti-tumor drugs targeting CDK is a major research hotspot
.
In the previous article, we summarized the clinical research progress
of pan-CDK inhibitors.
Many related clinical trials of pan-CDK inhibitors have been terminated
due to narrow therapeutic windows.
Therefore, the development of selective CDK inhibitors is crucial, and the design strategy, structure-activity relationship, and therapeutic application
of selective CDK inhibitors are summarized here.

CDK1 inhibitors

CDK1 selective inhibitors can reversibly block the G2/M phase of the cell cycle and achieve cell synchronization in early mitosis
.

Figure: Avotaciclib's design strategy, chemical structure, and CDK inhibitory activity

Avotaciclib (BEY1107, 18) is an orally available CDK1 inhibitor currently in Phase I/II clinical trials for the treatment of locally advanced or metastatic pancreatic cancer (NCT03579836)
as a monotherapy agent or in combination with gemcitabine.
This compound not only controls the cell division of cancer cells, but also overcomes the tolerance of anti-cancer drugs by eliminating the targeting mechanism of cancer stem cells
.
When combined with the cytotoxic drug gemcitabine, avotaciclib is expected to overcome emerging resistance to gemcitabine
.
CDK2 inhibitors

CDK2/cyclin E complex plays an important role
in retinoblastoma (Rb) protein phosphorylation and G1/S phase transition.
CDK2/cyclin A complexes play an important role
in DNA synthesis in the S phase and activation of CDK1/cyclin B in the G2/M phase transition.
PF-07104091(19) is an orally available CDK2 inhibitor with potential antitumor activity
.
In the human ovarian cancer cell model (OVCAR3, OV5392), compound 19 (25, 75 and 175 mg/kg twice daily, po) can induce tumor reduction
in a dose-dependent manner.
In addition, the combined administration of 19 and the first-line drug pabucinil showed a synergistic effect
in breast cancer cell models.
Pfizer began a Phase I/II clinical trial of Compound 19 in 2020 to evaluate the safety and tolerability of Compound 19 and to test Compound 19 as the maximum tolerated dose of monotherapy (NCT04553133)
in patients with small cell lung, non-small cell lung, ovarian or breast cancer.

Figure: Structure and CDK inhibitory activity of PF-07104091 (19) and zotiraciclib (20).

Zotiraciclib (TG02,20) is a unique molecule
with a macrocyclic structure.
The conformation of the macrocyclic facilitates the presentation of key pharmacophores when binding to proteins
.
In addition, a large ring with basic nitrogen forms a salt bridge with the acidic residue Asp86 at the inlet of the CDK2 active site, which also explains the high activity
of compound 20.
As a highly effective oral CDK2 inhibitor, compound 20 can penetrate the blood-brain barrier and degrade MCL-1 and Myc oncoproteins
.
Overexpression of Myc protein is one of
the characteristics of human HGG and diffuse endogenous pontine glioma (DIPG).
In January 2021, the Phase Ib clinical trial of compound 20 (HHSN261200800001E)
was completed.
In addition, the combination of compound 20 and temozolomide in the treatment of patients with recurrent HGG has shown good tolerability and efficacy
.
The FDA and the European Medicines Agency (EMA) have approved compound 20 as orphan drugs for the treatment of gliomas
.
Compound 20 is currently undergoing two separate Phase I clinical trials for the treatment of glioblastoma multiforme (GBM) and DIPG (NCT03904628, NCT03224104).

CDK4/6 inhibitors

In recent years, the incidence and mortality rate of breast cancer has continued to rise
.
By determining the estrogen receptor (ER), progesterone receptor (PR), and HER2 status of the tumor, breast cancer can be divided into different subtypes
.
CDK4 and CDK6 are important drivers of the development and progression of HR+/HER2- breast cancer
.
CDK4/6 is a common downstream target for a variety of growth-promoting signaling pathways, including ER
.
In ER+ breast cancer, estrogen signaling leads to increased activity of the ER-cyclin D1-CDK4/6 pathway, driving cells from the G1 phase to the S phase, ultimately leading to uncontrolled cell proliferation
.
CDK4/6 inhibitors inhibit the formation of complexes between CDK4/6 and cyclin D, block ATP binding, thereby cutting off upstream growth signals and preventing the transition
from G1 to S.
The CDK inhibitors currently approved by the FDA are CDK4/6 inhibitors: Pfizer's pabucinil (PD-0332991, 21), Novartis' ribocicillib (LEE011, 22), Eli Lilly's abecsidlib (LY-2835219, 23), and G1 Therapeutics' triracicil (G1T28, 24).

The ATP-binding domains of CDK4 and CDK6 exhibit high structural similarity, and the differences between these domains are only Glu144 (CDK4) and Gln149 (CDK6).

All four inhibitors are ATP-competitive inhibitors that bind
to the ATP-binding domain of kinases.

Pyridine[2,3-d]pyrimidin-7-one derivative 25 is a CDK4/6 inhibitor that optimizes its structure to obtain compound 26 (PD0183812).

The compound has good CDK4/6 activity, but its selectivity for CDK isoforms is not as expected
.
Studies of amino substitution at parent nucleus C-2 have shown that 2-aminopyridine-substituted compounds are more selective than aniline-based substitutes
.
Subsequently, the parent nucleus was replaced with quinazoline to obtain compounds 27a, 27b, the pyridine ring was replaced by a benzene ring, and a methyl group was introduced at C-4 to obtain the highly selective compound 27c-f, but the activity of the new compound was still not high
.
The introduction of acetyl groups at C-3 of the parent nucleus gives 27 g-i, and these compounds exhibit very good activity and selectivity
.
Further optimization, different nitrogen-containing alkanes such as piperazine, homopiperazine, 4-hydroxypiperidine and morpholine were introduced on the C-2 side chain to obtain compound 21, which showed inhibitory activity
against drug-resistant CDK4/6.
21 demonstrated clinical activity in a variety of tumor types, including mantle cell lymphoma, breast cancer, and liposarcoma
.
21 Combination with tamoxifen has a synergistic effect
in the treatment of ER+/HER2- breast cancer.
In addition, 21 combinations with aromatase inhibitors (anastrozole and letrozole, NCT02942355, NCT05190094) or selective ER degraders (fulvestrant, NCT02738866, NCT04318223) have been evaluated in phase II/III clinical trials and have shown certain clinical advantages
.
Figure: Design strategy, chemical structure, structural modifications, and CDK inhibitory activity of Pabocilib (21) and SHR-6390 (28).

Replacing the piperazine ring with a piperidine ring yields SHR-6390(28).

Compared to 21, oral compound 28 eliminated resistance to endocrine therapy and HER2-targeted antibodies in ER+ and HER2+ breast
cancers.
In addition,28 in combination with endocrine therapy showed synergistic antitumor activity
against ER+ breast cancer.
The eutectic structure of 21 and CDK6 indicates that the pyrimidine ring interacts with amino acids in the hinge region and forms two hydrogen bonds
with Val101.
Cyclopentyl and multiple amino acid residues in hydrophobic bags form hydrophobic interactions
.
The 4-position methyl group occupies a narrow hydrophobic cavity
formed by Phe80.
The piperazine ring extends into the solvent region and forms a polar interaction
with the amino acid residues Asp102 and lle19 in the solvent-exposed region of the ATP binding site.
The crystal structure of CDK6 indicates that the aminopyrimidine on the left forms two hydrogen bonds and the amide carbonyl group forms one hydrogen bond
.
Cyclopentyl forms hydrophobic interactions
in hydrophobic pockets.
Piperazine extends into the solvent zone to form a salt bridge
.
Although the structure of 23 has changed significantly, the eutectic structure of 23 with CDK6 indicates that its key binding properties are essentially the same
as those of previously approved CDK4/6 inhibitors.
Aminopyrimidine still forms two hydrogen bonds, and pyridine remains
.
The protein loses its hydrogen bond with the carbonyl groups in 21 and 22, but forms a new hydrogen bond
with imidazole.
The isopropyl group in compound 23 can mimic the cyclopentyl group in compounds 21 and 22, producing a hydrophobic effect
.
The amide substituent of 22 is cyclized, and the pyrrolo[2, 3-d]pyrimidine parent nucleus is tricyclic to form a tricyclic and the cyclopentane substituent is changed to cyclohexane to obtain compound 24, and its binding pattern is presumed to be similar
to that of 22.
Figure: Interaction of palbocilib (21), ribocilib (22), abecilide (23) and CDK6

Compounds 22 and 23 are approved as first-line treatments for HR+/HER2-metastatic breast cancer, and their combination with nonsteroidal aromatase inhibitors (anastrozole and letrozole), tamoxifen or goserelin is under investigation
.
In addition, clinical studies of them as monotherapies and in combination with chemotherapy drugs for the treatment of other tumors, including melanoma, advanced solid tumors, lymphoma, and neuroblastoma, are also being further conducted
.
Compound 24 is the first "breakthrough therapy" approved by the FDA to improve outcomes for cancer patients undergoing chemotherapy
.
24 G1 phase of bone marrow cells can be preventively blocked before chemotherapy, which will significantly reduce the killing of bone marrow cells by chemotherapy drugs, thereby protecting bone marrow cells and the immune system
.

Betta Pharmaceuticals incorporated a saturated naphthenic in the benzimidazole parent nucleus of compound 23 to obtain the candidate compound BPI-16350(29), which is undergoing a Phase I clinical trial (NCT03791112) for the treatment of patients with
advanced solid tumors.
In addition, the researchers developed GLR-2007 (30) for the treatment of advanced relapsed or refractory solid tumors that do not respond well to standard clinical therapies, including non-small cell lung cancer (NSCLC) and GBM
.
Phase I/II clinical trials are currently underway for the treatment of advanced solid tumors (NCT04444427).

The drug has been approved as an orphan drug for the treatment of malignant glioma and has passed the fast-track review
for the treatment of glioblastoma.
30 was further modified to obtain lerociclib(G1T38, 31).

The study of compound 31 in combination with fulvestrant is in a phase I/II clinical trial for the treatment of patients with HR+/HER2- locally progressive or metastatic breast cancer following endocrine failure (NCT02983071).

ON-123300(32) is a dual inhibitor of CDK4/6 and AMPK-associated protein kinase 5 (ARK5) for the potential treatment
of advanced cancer.
Phase I clinical trials of compound 32 have been initiated for the treatment of patients with recurrent cancer or advanced cancer resistant to at least one prior treatment, including metastatic HR+/HER2-breast cancer resistant to approved second-generation CDK4/HER2-breast cancer (NCT04739293).

FLX-925(33) is a novel dual inhibitor targeting FMS-like tyrosine kinase-3 (FLT-3) and CDK4/6 in early-stage clinical studies for the treatment of relapsed or refractory AML
.
A Phase I clinical study evaluating the safety, pharmacokinetics, and pharmacodynamics of compound 33 in the treatment of patients with relapsed or refractory AML is further underway (NCT02335814).

Figure: Structure of compounds 29-33 and CDK inhibitory activity

In addition, the structure of compound 34-41, which is currently in clinical trials, has not been disclosed
.

CDK7 inhibitors

CDK7 is involved in both transcription and cell cycle regulation, is overexpressed in many types of cancer, and is also associated
with aggressive clinicopathological features and poor prognosis.
Several selective CDK7 inhibitors that have entered clinical trials have shown good anti-cancer efficacy in many preclinical models, including AML, prostate cancer, liver cancer, gastric cancer, cervical cancer, pancreatic cancer, colorectal cancer, neuroblastoma, NSCLC, and SOX-2-amplified lung squamous cell carcinoma
.

The pyrazolopyrimidine derivative samuraciclib (ICEC0942, 42) reversibly binds to the ATP-binding site of CDK7 (IC50=40nM) and exhibits antiproliferative activity in a range of cancers, including AML, small cell lung cancer (SCLC), and TNBC
.
There are two hydrogen bond interactions between the pyrazolo[1, 5-a]pyrimidine parent nucleus and the backbone of the hinge region Leu83, a third hydrogen bond is formed between the nitrogen atom of the piperidine ring and Asn132, and the preceding fragment also has electrostatic contact
with Asp145.
The good pharmacokinetic properties of compound 42 make this compound a promising clinical candidate
.
When combined with tamoxifen, 42 completely blocks the growth of ER+ xenograft tumors and is currently undergoing phase I/II clinical trials for the treatment of breast and prostate cancer (NCT03363893).

Figure: Structure and CDK inhibitory activity of selective CDK7 inhibitors 42-45

A number of covalent CDK7 inhibitors have also been developed due to the presence of cysteine and lysine residues
outside the ATP-binding pocket.
Covalently bound CDK inhibitors are generally more active and selective, have a longer duration of treatment, and have fewer
serious side effects.
THZ1(43) contains the N-phenylpyrimidine-2-amine parent nucleus, acrylamide irreversibly bound to nearby Cys312, and the aromatic portion is embedded inside CDK7, suggesting that the indole ring of 43 rotates so that it points rather than deep in
the open nucleotide-binding pocket.
Compound 43 can suppress many types of tumors in vitro and in vivo, such as osteosarcoma, esophageal squamous cell carcinoma, adult T-cell leukemia, MYCN-amplified neuroblastoma, melanoma, and SCLC, but the short half-life of 43 in vivo (45 min in mouse plasma) limits clinical development
.

Figure: Binding pattern of CAK-THZ1 (CAK: complex consisting of CDK7, cyclin H, and MAT1)

The CDK7 inhibitor SY-1365 (44), derived from compound 43, has improved
activity, selectivity, and metabolic stability.
The 1,3-cyclohexanediamine parent nucleus of 44 improves shape complementarity and hydrophobicity, and therefore exhibits higher activity and selectivity
than 43.
Compound 44, which limits the conformation of warheads by increasing steric hindrance around amide linkers, has entered phase I clinical trials for the treatment of ovarian and breast cancer (NCT03134638).

However, Syros Pharmaceuticals recently announced the discontinuation of clinical development of 44, prioritizing the advancement of a novel selective non-covalent CDK7 inhibitor, SY-5609 (45).

In addition to the aminopyrimidine hinge binding region, the basic nitrogen of the (S)-3 aminopiperidine ring can form a salt bridge with the ASP97 of CDK7, which is conserved in the CDK family
.
Substitution with trifluoromethyl increases activity but slightly reduces selectivity
.
Therefore, the introduction of a cyanide group at C-6 of the indole ring to protrude towards the P ring yields a more selective compound, which may be caused
by taking advantage of the difference between the P ring and other CDKs.
In addition, to improve the pharmacokinetic properties, a phosphine oxide fraction
was introduced due to its strong hydrogen bonding and high stability.
The introduction of dimethyl into compounds increases lipophilicity, regulates physicochemical properties and permeability, and improves oral bioavailability and selectivity
.
The binding mode suggests that the piperidine 6 position can provide a better carrier for non-conserved Val100, so the C-6 position introduces the dimethyl to give compound 45
.
The combination of compound 45 with fulvestrant showed antitumor activity
in preclinical models of ovarian cancer, TNBC and ER+ breast cancer.
A Phase I trial (NCT04247126)
for the treatment of patients with advanced solid tumors was initiated in early 2020.

In addition, XL-102 (AUR-102, 46, structure undisclosed) is an early clinical development of Exelixis' oral covalent CDK7 inhibitor for the treatment of solid tumors and is undergoing a Phase I clinical study (NCT04726332).

CDK8/19 inhibitors

CDK8 and its paralogous homologous CDK19 (97% homology) regulate the activity of RNA PolI I, and CDK8/CDK19 overexpression
has been observed in some cancers.
The latest research suggests that CDK8 and CDK19 may also play a role
in cell reprogramming.
Targeted inhibition of CDK8/CDK19 can activate the anticancer effect
of natural killer (NK) cells.
SEL120-34A(47) is a novel orally available inhibitor of benzimidazole tricyclic CDK8/19 that inhibits tumor growth, relieves splenomegaly, and promotes partial bone marrow recovery
in an orthotopic xenograft model of AML.
In 2020, the compound was granted orphan drug designation by the FDA for the treatment of AML
.
In addition, a Phase I/II study (NCT05052255) evaluating the safety, pharmacokinetics, and efficacy of 47 patients with metastatic or advanced solid tumors is underway, and the 47 trial for the treatment of AML or high-risk MDS patients is in Phase Ib (NCT04021368).

BCD-115(48), also a CDK8/19 inhibitor, has entered Phase I clinical trials (NCT03065010) for the treatment of ER+/HER2-locally progressive and metastatic breast cancer
.

Figure: Structure and CDK inhibitor activity of selective CDK8/19 inhibitors

CDK9 inhibitors

Researchers have developed a number of selective CDK9 inhibitors through modifications based on flavonoid base pan-CDK inhibitors aimed at increasing selectivity, reducing gastrointestinal toxicity, and improving oral bioavailability
.
TP-1287(49) is an orally bioavailable phosphate prodrug of compound 1 that effectively inhibits CDK9
.
Compound 49 targets and binds to CDK9, and this binding reduces the expression of CDK9 target genes, such as the anti-apoptotic gene MCL-1, and induces cell cycle arrest and apoptosis
in CDK9 overexpressed cancer cells in the G1 phase.
A multicenter phase I dose-escalation study showed that 49 resulted in a dose-dependent reduction in phosphorylated RNA PolI levels, which was consistent
with CDK9 inhibition.
After the administration of compound 49, a local response was observed in one sarcoma patient, and the condition stabilized in one patient with renal cell carcinoma (RCC) and two patients with bladder cancer, and based on these considerable results, 49 could be further developed clinically as monotherapy (NCT03604783).

Figure: Design strategy for selective CDK9 inhibitor 49-51

IIIM-290(50) is another novel orally available selective CDK9 inhibitor
.
Studies of the flavonoid base compound SAR have found that modification at the C-2 position of the parent nucleus can achieve higher selectivity and activity
.
Compound 50 is obtained by substituting 2-chlorophenyl on C-2 with 2,6-dichlorostyryl, which improves its selectivity for CDK9 and increases its inhibitory activity by a factor of 10
.

The type and location of halogen substituents on the C-2 styrene ring are closely related
to the selectivity and inhibitory activity of CDK9.
Chlorine substitution shows more potent activity
than fluorine substitution.
Chlorine has higher
ortho-substitution activity compared to para-substitution.
The binding pattern indicates that 50 occupies the ATP-binding cavity of CDK9, and the chromogen parent nucleus, hydroxypiperidine, and chlorine-substituted styrene are key fragments
that bind to the ATP-binding site of CDK9.
Compound 50 showed excellent pharmacokinetic properties, oral bioavailability and potent anticancer activity in multiple xenograft models, with the best
activity in pancreatic cancer models.
50 IND applications have been submitted and a Phase I/II clinical trial (TENDER no.
:12(318)/2021-P)
for the treatment of patients with locally progressive or metastatic pancreatic cancer is ongoing.

Trifluoromethyl was introduced into the counterpoint of compound 2 benzene ring to obtain voruciclib (P1446A-05, 51).

51 As monotherapy or in combination with KRAS inhibitors (sotolasib and adagrasib), it has shown initial activity
in KRAS mutated cancer cell lines.
In 2014, 51 clinical studies (NCT02117336)
for the treatment of patients with relapsed or refractory CLL were suspended.
51 is currently in another Phase I clinical trial to evaluate its safety and preliminary efficacy in patients with AML with relapsed or refractory B-cell malignancies and standard therapy (NCT03547115).

In drug design, sulfoneimide can enrich structural diversity, improve metabolic stability, optimize physicochemical properties, and introduce hydrogen bond receptors
/donors.
Atuveciclib (BAY-1143572, 52) with sulfonimide groups is the first orally selective CDK9 inhibitor
to enter clinical development.
Compound 52 has shown good efficacy
of monotherapy in multiple xenograft tumor models in mice and rats.
Two Phase I clinical trials have been completed to evaluate the safety, tolerability, pharmacokinetics and maximum tolerated dose
of 52 in patients with advanced malignancy (NCT01938638) and advanced acute leukemia (NCT02345382).
Replace triazine with pyridine to give BAY-1251152(53).

Compared to 52, 53 exhibited better activity (biochemical IC50=3nM, cell IC50=29nM), and a phase I dose-escalation study in patients with advanced cancer showed that 53 exhibited controllable safety, target pharmacodynamic activity, and antitumor activity (NCT02635672).

Figure: Structure and activity of selective CDK9 inhibitors 52-55

AZD4573 (54) is a highly selective (selective > 10-fold) CDK9 inhibitor that inhibits MCL-1 and induces apoptosis
in blood cancer cells.
The study found that when the terminal benzene ring was replaced by a heterocyclic ring, the physical and chemical properties were improved, but the activity was lost
.
By introducing a halogen into a pyridine ring to compensate for the lost activity, the resulting compound 54 achieves an overall balance
between activity and physicochemical properties.
54 can result in persistent tumor regression
in MM, AML, and non-Hodgkin lymphoma (NHL) xenograft model mice.
A phase I/II clinical trial (NCT04630756)
for the treatment of patients with advanced blood cancer is ongoing.

KB-0742(55) is an orally available selective CDK9 inhibitor with more than 50-fold
selectivity for CDK9 (IC50=6nM).
In castration-resistant prostate cancer (CRPC) cell models, 55 can reduce the expression
of the oncogene Myc by inhibiting CDK9 activity.
The compound has a long plasma half-life and exhibits dose-dependent target inhibitory activity, and has also shown good efficacy
in in vivo models.
A Phase I clinical trial is underway for the treatment of patients with relapsed or refractory solid tumors or NHL (NCT04718675).

GFH009 (56, structure undisclosed) is a potent CDK9 inhibitor with more than 100-fold selectivity and is undergoing a Phase I clinical trial evaluating the safety and tolerability of 56 patients with hematologic malignancies (AML, CLL, SLL and lymphoma) (NCT04588922).

Design of selective CDK inhibitors

The design of CDK-selective inhibitors needs to consider some key structural features
.
The development of selective inhibitors of CDK1 requires consideration of electrostatic interactions
between ligands and LYS89.
For CDK2 inhibitors, hydrogen bond interactions with GLU81 and LEU83 are key to
maintaining activity.
The presence of hydrophobic pockets consisting of PHE80, VAL64, HIS84 and GLN85 favors hydrophobic interactions and π-π accumulation
.
For selective inhibitors of CDK4/6, polar interactions with THR99/10 are required, and hydrogen bond interactions with conserved hinge residues HIS92 and VAL101 are also required
.
For the development of CDK9 inhibitors, strong hydrogen bond interactions with ASP167 and CYS106 are necessary
.
Several non-conserved regions in the CDK9 binding site, including hydrophobic regions and solvent-exposed regions, provide the possibility for the
design of selective CDK9 inhibitors.
Targeting the hinge region, DFG motif, and C-terminal extension region of CDK12 is a viable method
for discovering highly active and selective CDK12 inhibitors.

summary

Pan-CDK inhibitors have obvious disadvantages of narrow therapeutic window, and selective CDK inhibitors have a higher
safety profile by targeting specific CDK subtypes.
However, because the ATP-binding pockets of different CDK subtypes are very conserved, finding ATP competitive and highly selective CDK inhibitors presents significant challenges
.
The development of selective CDK inhibitors relies on innovations
in the backbone and structural characteristics of the maternal nucleus.
At present, four CDK4/6 inhibitors have been approved, and many other CDK inhibitors are also in the clinical research stage, with great application potential, and it is expected to make greater breakthroughs
in the development of selective CDK inhibitors in the future.