Cell cycle regulation has been identified as an attractive target for targeted drug therapy. Given their kinase activity, the CDKs were pursued as drug targets. A large number of drug discovery programs have yielded potent small molecule CDK inhibitors, with several compounds successfully entering preclinical and early clinical trials. Until relatively recently, however, many CDK inhibitors have shown poor clinical activity accompanied by an undesirable adverse event profile. In general, CDK inhibitors can be broken down into two classes: first-generation inhibitors such as flavopiridol, R-roscovitine, and UCN-01, which tended to be less specific and broad in their ability to block a number of CDKs (pan-CDK inhibitors); and second-generation agents that are more specific to certain CDKs. The latter group of compounds has now shown more potent activity against their targets and a more favorable safety profile.
The first-generation CDK inhibitors
As mentioned, most of the first-generation compounds are not specific for any single CDK enzyme and act primarily as pan-CDK inhibitors. Despite initial enthusiasm generated by preclinical studies, however, many of these compounds suffered from low activity and/or toxicity in clinical studies.
Flavopiridol (National Cancer Institute) is the most studied of all first-generation CDK inhibitors, and is a classic pan-CDK inhibitor. In phase I and II studies, flavopiridol showed minimal single agent efficacy and was associated with several toxicities more typical of traditional cytotoxic agents, including infusion site irritation, gastrointestinal toxicity, and severe neutropenia [
61]. In metastatic breast cancers in particular, flavopiridol generated unacceptably high rates of neutropenia [
62]. At least a portion of this toxicity is attributable to the inhibition of transcription by the compounds effects on CDK9 and possibly CDK7 that lead to depletion of short-lived cell cycle and anti-apoptotic mRNA transcripts [
63]. Though this likely contributes to the in vitro efficacy of flavopiridol on tumors dependent on the expression of such transcripts, off-target effects in healthy tissues would contribute to the severe anti-proliferative toxicity observed in multiple clinical trials of this compound [
64].
Other examples of pan-CDK inhibitors include UCN-01 and R-Rescovitine (seliciclib; Cyclacel). UCN-01 is a staurosporine analog with broad activity against CDKs, AKT, Chk1, and protein kinase C. This drug showed good G1/S phase cell cycle arrest, induction of p21 and hypophosphorylation of pRb in preclinical models but phase I studies showed several dose-limiting toxicities, including hyperglycemia, arrhythmia, and pulmonary dysfunction [
65,
66]. Results of phase II studies in breast cancer were unimpressive [
67].
Specific CDK4/6 inhibitors
Recently, a number of inhibitors specific for CDK4 and CDK6 have entered clinical testing (Table
1). Palbociclib (PD 0332991; Pfizer) is furthest along in clinical development, having received US Food and Drug Administration (FDA) approval on 3 February 2015 for the first-line treatment of advanced post-menopausal ER+, HER2-negative breast cancer in combination with letrozole. It is an orally bioavailable, potent CDK4/6 inhibitor with an in vitro kinase IC50 of 0.01 μM and high selectivity when evaluating 36 other kinases including CDK2 (IC50 > 5 μM) [
69]. Preclinical studies have shown that palbociclib behaves very much like an agent specifically targeting CDK4/6. It exhibits potent inhibition of tumor cell proliferation accompanied by a pure G1 arrest, and dephosphorylation of pRb as well as a decrease in E2F-dependent gene expression [
70]. Further evidence of palbociclib's targeted design is the fact that it is completely inactive in pRb-negative tumor cell lines and xenografts [
9,
60,
70]. In phase I clinical studies palbociclib showed excellent bioavailability with a generally mild to moderate adverse events profile with the major dose limiting toxicities being related mainly to myelosuppression [
71].
Table 1
Current CDK4/6 inhibitors in clinical development
Palbociclib (PD-0332991) | Pfizer, Inc. | Approved | CDK4 (cyclinD1): 11 |
CDK4 (cyclinD3): 9 |
CDK6 (cyclinD2): 15 |
Ribociclib (LEE011) | Novartis | III | CDK4 (cyclinD1): 10 |
CDK6 (cyclinD2): 40 |
Abemaciclib (LY2835219) | Eli Lilly | III | CDK4 (cyclinD1): 2 |
CDK6 (cyclinD1): 9.9 |
Using an unbiased screening approach we performed preclinical work aimed at identifying breast cancers that might be growth inhibited by palbociclib and predictive markers of drug response. This was done by evaluating palbociclib’s growth inhibition effects in a large panel of molecularly characterized human breast cancer cell lines. This study identified that cell lines representing either the luminal, ER+ or HER2-amplified subtypes were most sensitive to palbociclib inhibition while those representing the non-luminal subtypes were most resistant [
9]. This work also demonstrated consistent synergistic growth inhibitory activity between palbociclib and tamoxifen or trastuzumab in ER+ and HER2-amplified cell models, respectively. Lastly, the drug showed activity in a model of acquired tamoxifen resistance leading to the concept that it may be clinically active in hormone-resistant, ER+ breast cancers.
These data were used to support the clinical development of palbociclib in a phase I/II study of frontline treatment of advanced ER+ post-menopausal breast cancer with a combination of palbociclib and letrozole. The phase I portion enrolled 12 patients and was designed to evaluate the safety of a dosing regimen consisting of 125 mg palbociclib orally given daily on a 3-week on/1-week off regimen in combination with daily letrozole [
72]. There were no treatment-related serious adverse events and the most common treatment emergent adverse events were leukopenia, neutropenia, and fatigue. However, there were no instances of neutropenic fever and there were no dose–dose interactions between palbociclib and letrozole.
The phase II study was developed as an open label trial in post-menopausal women with advanced ER+, frontline metastatic breast cancer. It was designed to compare progression-free survival (PFS) as its primary endpoint with safety and overall survival as secondary endpoints and randomized patients to receive either letrozole alone or the combination of letrozole and palbociclib. The study consisted of two parts that enrolled sequentially: part 1 required that patient tumors be ER+, the sole biomarker for study entry; part 2 enrolled the same population but patient tumors were also required to have either CCND1 (cyclin D1) amplification by fluorescence in situ hybridization (FISH) or CDKN2A (p16) loss by FISH as selection biomarkers in addition to the ER+ biomarker. While the preclinical data did not suggest that these genomic markers were required for augmented response, part 2 of the study was designed to determine whether the presence of these biomarkers might further enrich the responsive patient population.
Results from part 1 were presented at the IMPAKT meeting in 2012 [
73]. About half the women in each arm had not received any prior neoadjuvant or adjuvant systemic treatment for their diagnosis but about a third had received prior anti-estrogen therapy in early breast cancer settings. There was a significant improvement in PFS in part 1 with the median PFS increasing from 5.7 months with letrozole alone to over 18 months with the combination, resulting in a hazard ratio (HR) of 0.35 (95 % confidence interval (CI) 0.17–0.72,
P = 0.06). In addition, in patients with measurable disease the response rate increased from 32 to 52 % and the clinical benefit rate increased from 47 to 76 %. Dose reductions and delays were common in the palbociclib arm, but again, the most common treatment-related adverse events were leukopenia, neutropenia, and fatigue, although no instances of neutropenic fevers were reported. Retrospective biomarker analysis for CCND1 amplification and p16 loss was performed in the 66 patients from part 1. Though the groups were small, the HRs for each group demonstrated a consistent benefit regardless of the presence or absence of these biomarkers; biomarkers present (n = 21) HR = 0.37 (95 % CI 0.10–1.40,
P = 0.13), biomarkers absent (n = 25) HR = 0.19 (95 % CI 0.05–0.67,
P < 0.01), biomarker unknown (n = 20) HR = 0.59 (95 % CI 0.11–3.08,
P = 0.53). These data support the preclinical observation that ER positivity may be the best selection biomarker for patients likely to benefit from CDK4/6 inhibition.
An interim analysis combining parts 1 and 2, based on 50 % of events of the 114 needed for the final PFS analysis, was presented at the 2012 San Antonio Breast Cancer Symposium and the final results have been published [
73,
74]. These analyses included 165 patients and confirmed the benefit and safety profile observed initially in part 1. Specifically, the final results demonstrated that median PFS increased from 10.2 months with letrozole alone to 20.2 months with the combination (HR = 0.488 (95 % CI 0.319–0.748,
P < 0.001)). In cohort 1, median PFS was 5.7 months in the letrozole alone arm and was 26.1 months in the combination arm; in cohort 2 these numbers were 11.1 months and 18.1 months, respectively. HRs for both cohorts, 0.299 for cohort 1 and 0.508 for cohort 2, confirmed a benefit regardless of the presence of cyclin D1 amplification or p16, suggesting the most important determinant for benefit in this study is being ER+. The objective response rate for patients with measurable disease was increased from 39 to 54 % with the addition of palbociclib and the clinical benefit rate (complete response, partial response, and stable disease >6 months) for the intent-to-treat population improved from 58 to 81 %. The adverse event profile remained essentially the same. While the incidence of grade 3 and 4 neutropenia was 48 % and 6 %, respectively, there were no cases of neutropenic complications (that is, febrile neutropenia or serious infections). The lack of serious complications from the neutropenia may be explained by the cytostatic effect of CDK4/6 inhibition on the bone marrow which, compared with cytotoxic chemotherapy, results in a relatively short period of neutropenia. In addition, no mucositis or skin toxicity was associated with palbociclib, which are often considered sources of infection with chemotherapy-associated neutropenia. Preclinical studies suggest that CDK4/6 inhibition induces a reversible pharmacologic quiescence in hematopoietic stem/progenitor cells that differs significantly from cytotoxic effects and may explain the clinical observation [
75].
Together, the safety and efficacy data from this study resulted in palbociclib receiving a 'Breakthrough Therapy' designation from the US FDA and more recently accelerated approval for advanced ER+ breast cancer [
76,
77]. A phase III, double-blind, placebo-controlled study designed to confirm the phase II observations has completed accrual and results are awaited (PALOMA-2/TRIO-22, NCT01740427). Results of the PALOMA-3 study have recently been published and again demonstrate a significant improvement in PFS when palbociclib is used in combination with endocrine therapy [
78]. In this large phase III, placebo-controlled, double-blind study, palbociclib and fulvestrant was compared to fulvestrant and placebo. The study demonstrated a doubling of PFS. The PFS in the treatment arm was 9.2 months (95 % CI 7.5–not estimable) compared to 3.8 months (95 % CI 3.5–5.5) in the control arm. Unlike the PALOMA-1/TRIO18 and PALOMA-2/TRIO22 studies, this population of patients had a more endocrine-resistant disease, with the requirement to have progressed on or within 1 month of prior aromatase inhibitor for advanced disease, or within 12 months of completion or discontinuation of therapy for adjuvant therapy. This study also allowed pre-menopausal women that received goserelin as well. The safety profile looked very similar to what was seen in the PALOMA-1/TRIO18 study.
Single-agent activity of palbociclib has also been evaluated in a single arm phase II trial of palbociclib in advanced, heavily pre-treated breast cancer [
79]. Despite being tested in a heavily pre-treated cohort of patients (median lines of therapy = 3), single-agent activity was noted (clinical benefit 21 %, stable disease >6 months 14 %). Importantly, as the preclinical data suggested, this activity was seen in women with ER+ or HER2-amplified breast cancers. Myelosuppression again was the most frequently observed adverse event, with 46 % of patients requiring dose reductions and 25 % requiring dose interruptions.
In addition to palbociclib, two other small molecule CDK4/6 inhibitors are currently in early clinical development. Both have had their development programs expedited, going from phase I to phase III based on the palbociclib experience. The molecules and ongoing trials in breast cancer are highlighted in Tables
1 and
2, respectively. Phase I data with LY2835219 (abemaciclib; Eli Lilly) in patients with advanced malignancies was presented at the ASCO 2013 meeting [
80]. In this dose escalation study it was determined that the doses in the expansion phase were to be 150 mg and 200 mg twice a day continuously, without a dosing break like with palbociclib. They concluded that it had an acceptable safety profile and early signals of clinical efficacy were seen. Data on an expansion cohort of advanced breast cancer patients have been presented as well [
81,
82]. Two cohorts were examined, one with single-agent abemaciclib and one with abemaciclib and fulvestrant for ER+ disease. In the single-agent cohort, 47 patients with all subtypes of breast cancer were enrolled, but significant single-agent activity was seen only in women with ER+ breast cancer. The median lines of prior therapy in this group were 7 (2–16). The overall response rate in the 36 patients with ER+ disease was 33 % and the disease control rate was 80.6 %. Median PFS was 8.8 months for the ER+ cohort compared with 1.1 months in the ER-negative group. In the combination cohort, patients with ER+ metastatic breast cancer (n = 18) were treated with the combination abemaciclib plus fulvestrant. Patients received abemaciclib at 200 mg orally every 12 hours on a continuous schedule. Patients also received 500 mg fulvestrant intramuscularly every month. Patients in this cohort had a median of four lines of prior therapy. The disease control rate in the latter cohort was 72.2 %. Like palbociclib, neutropenia was seen in 40 % of all-grade cases, and 21 % of grade 3/4 cases. There was 66 % all-grades diarrhea reported, of which there were only 6 % grade 3 cases and no grade 4 cases. This side effect seems to indicate some differences between palbociclib and abemaciclib. The dose in phase III breast cancer studies is 150 mg daily every 12 hours, continuously.
Table 2
Currently registered clinical studies with CDK4/6 inhibitors in breast cancer
Palbociclib (PD-0332991) | First-line metastatic | PFS | Letrozole | 450 | III | NCT01740427 (PALOMA-2) |
| Metastatic | PFS | Fulvestrant | 417 | III | NCT01942135 (PALOMA-3) |
| High-risk adjuvant | iDFI | Anti-hormonal | 800 | III | NCT01864746 (PENELOPE-B) |
| Neo-adjuvant | pCR | Anastrozole | 29 | II | NCT01723774 |
| Pre-operative | ORR | Letrozole | 45 | II | NCT01709370 |
| Metastatic | MTD | Paclitaxel | 20 | I | NCT01320592 |
| Neo-adjuvant | Biomarker cCR | Letrozole | 306 | II | NCT02296801 (PALLET) |
| Metastatic | PFS | Exemestane versus capecitabine | 348 | III | NCT02028507 (PEARL) |
| Adjuvant | Treatment discontinuation | Letrozole | 160 | II | NCT02028507 |
| Metastatic | Dose/toxicity | TDM-1 | 17 | Ib | NCT01976169 |
| Neoadjuvant | RCB | Letrozole versus FEC-3 | 132 | II | NCT02400569 (NeoPAL) |
| Adjuvant | iDFS | SAT | 4600 | III | NCT02513394 (PALLAS) |
Ribociclib (LEE011) | First-line metastatic | PFS | Letrozole | 450 | III | NCT01958021 |
| Pre-surgical | PD | Letrozole | 120 | II | NCT01919229 (MONALEESA-1) |
| Metastatic | DLT, PFS | BYL719, letrozole | 300 | I/II | NCT01872260 |
| Metastatic | DLT, PFS | Exemestane, everolimus | 185 | Ib/II | NCT01857193 |
| Metastatic | PFS | Letrozole | 650 | III | NCT01958021 (MONALEESA-2) |
| Metastatic | DLT | Letrozole, buparlisib | 13 | I | NCT02154776 |
| Metastatic (pre-menopausal) | PFS | Tamoxifen, NSAI | 660 | III | NCT2278120 (MONALEESA-7) |
| Metastatic | DLT/PFS | BYL719a or BKM120 | 216 | I/II | NCT01872260 |
| Metastatic | PFS | Fulvestrant | 660 | III | NCT02422615 (MONALEESA-3) |
Abemaciclib (LY2835219) | Neoadjuvant | Biomarker | Anastrozole | 220 | II | NCT02441946 (NeoMONARCH) |
| Brain metastasis | Response | Single agent | 120 | II | NCT02308020 |
| Metastatic | PFS | NSAI | 450 | III | NCT02246621 (MONARCH-3) |
| Metastatic | Response | Single agent | 128 | II | NCT021024490 (MONARCH-1) |
| Metastatic | PFS | Fulvestrant | 630 | III | NCT02107703 (MONARCH-2) |
| Metastatic | PFS | NSAI, tamoxifen, exemestane, everolimus, trastuzumab | 102 | I | NCT02057133 |
Like palbociclib, LEE011 (ribociclib; Novartis) is being dosed at 600 mg daily, 3 weeks on and 1 week off. Limited data in breast cancer have been presented. In a large phase I study of advanced pRb + solid tumors, single-agent activity was seen in patients with breast cancer [
83]. The most common grade 3/4 toxicities at the recommended dose for expansion were neutropenia (26 %), leukopenia (16 %), and lymphonepnia (16 %). LEE011 is now moving ahead into more advanced studies in breast and other cancers. In addition, it is being evaluated in combination with the p110α-specific phosphoinositide 3-kinase inhibitor alpelisib (BYL719) and letrozole and in combination with everolimus plus exemestane. More mature data with both these compounds are eagerly awaited.