CDK4 inhibitors (CDK4/6i), such as palbociclib, ribociclib, and abemaciclib, are approved in combination with hormonal therapy as a front-line treatment for metastatic HR+, HER2- breast cancer. Their targets, CDK4 and CDK6, are cell-cycle regulatory proteins governing the G1–S phase transition across many tissue types. A key challenge remains to uncover biomarkers to identify those patients that may benefit from this class of drugs. Although CDK4/6i addition to estrogen modulation therapy essentially doubles the median progression-free survival, overall survival is not significantly increased. However, in reality only a subset of treated patients respond. Many patients exhibit primary resistance to CDK4/6 inhibition and do not derive any benefit from these agents, often switching to chemotherapy within 6 months. Some patients initially benefit from treatment, but later develop secondary resistance. This highlights the need for complementary or companion diagnostics to pinpoint patients who would respond. In addition, because CDK4 is a bona fide target in other tumor types where CDK4/6i therapy is currently in clinical trials, the lack of target identification may obscure benefit to a subset of patients there as well. This review summarizes the current status of CDK4/6i biomarker test development, both in clinical trials and at the bench, with particular attention paid to those which have a strong biological basis as well as supportive clinical data.

This article is featured in Highlights of This Issue, p. 1

Breast cancer is the most common women's cancer worldwide, comprising 25% of total new cases diagnosed in 2018, and is the second leading cause of cancer-related deaths (1). Treatment depends on hormone receptor status, as well as the expression of the receptor tyrosine kinase HER2/neu. For patients with estrogen receptor (ER) and/or progesterone receptor (PR)–positive (HR+), HER2-negative metastatic disease, antiestrogen therapy is the backbone of treatment, which can be accomplished by ER downregulation (e.g., fulvestrant) or modulation (e.g., tamoxifen), or by aromatase inhibition (e.g., letrozole; ref. 2).

Palbociclib, ribociclib, and abemaciclib are first-generation cyclin-dependent kinase 4/6 (CKD4/6) inhibitors (CDK4/6i), which have demonstrated activity against HR+, HER2- breast cancer (3–5). Beginning with the FDA approval of palbociclib in 2015, the addition of these agents to antihormonal therapy has become a first-line option for patients with the HR+, HER2- phenotype in the metastatic setting (2). In the pivotal PALOMA-2 study, palbociclib, in combination with letrozole, was shown to increase median progression-free survival (PFS) from 14.5 to 24.8 months, compared with letrozole alone, in patients who had received no prior therapy for metastatic disease (6). The PALOMA-3 trial subsequently evaluated palbociclib in combination with fulvestrant in patients with metastatic disease, whom had progressed on prior endocrine therapy. The addition of palbociclib to fulvestrant increased median PFS from 4.6 to 9.5 months, leading to FDA approval of palbociclib in this second-line setting as well (7). Although these results are promising, the clinical use of CDK4/6 inhibitors is confounded by the high individual variability in clinical response. Many patients exhibit primary resistance to CDK4/6 inhibition and do not derive any benefit from treatment with these agents, often switching to chemotherapy within 6 months. Many other patients derive some benefit from treatment, but invariably become refractory, defined as secondary resistance.

Most targeted anticancer agents have a companion or complementary diagnostic biomarker to determine which patients have the actionable target, and are therefore likely to respond to treatment. A complementary diagnostic is a test which provides information relevant to determining the risk versus benefit of a particular treatment, but is not required by regulatory agencies in order to use the treatment (8). Complementary diagnostics often provide information relevant to a specific class of drugs, as opposed to a specific agent. A companion diagnostic, on the contrary, provides information that is required prior to using a drug, and is therefore frequently marketed with the drug. A classic example is HER2 testing by IHC or FISH, as only patients with documented HER2+ status are offered treatment with trastuzumab. At the current time, no clinically available biomarkers, other than ER/PR expression, are used to prescribe CDK4/6 inhibitors (9–13). Therefore, many patients receive treatment with these agents but do not benefit, and there are likely also many patients who could benefit, but are never offered treatment with one of these agents. One difficulty with the identification of a CDK4/6i biomarker is the fact that CDK4/6 activity is due to a phosphorylated multiprotein complex, and simple assessment of the complex's components has proven ineffective in predicting response.

Mechanism of action of CDK4/6 inhibitors

CDK4 and CDK6 have been sought-after drug targets since their discovery in the early 1990s. As cell-cycle regulatory proteins, they exist downstream of other oncogenic pathways, as a final common hub in cell-cycle progression (Fig. 1). Cyclin D complexes with CDK4 or CDK6 (DK4/6) to regulate the G1–S phase transition, allowing for progression through this checkpoint and the subsequent replication of DNA in S phase (ref. 14; Fig. 2). DK4/6 is actually a ternary complex and is only stable when associated with p27Kip1 or p21Cip1. p27 and possibly p21 must be phosphorylated on residue Y88/89 (p27) or Y76 (p21) to convert the closed ternary complex into an open, active complex (15). The cyclin D-CDK4/6-p27 complex is translocated to the nucleus by virtue of p27's Nuclear Localization Signal (NLS). The open complex now permits ATP access to the CDK4 active site, and also renders CDK4 competent to be phosphorylated on residue T172 by the Cyclin Activating Kinase (CAK), which fully activates CDK4 activity. Active DK4/6 can phosphorylate the retinoblastoma protein (Rb), causing the release of E2F transcription factors, which in turn lead to the transcription of genes required for DNA replication and cell-cycle progression, including Cyclin E. This leads to a positive feedback loop where E2F-mediated expression of Cyclin E drives activation of CDK2, which further phosphorylates Rb and causes the full release of E2F.

Figure 1.

All roads lead to cyclin D-CDK4/6, and as such has been a long sought-after therapeutic target. Cyclin D-CDK4/6 is downstream of most oncogenic signaling pathways, including tyrosine kinases, such as EGFR and Erb2/Her2, cytokine receptors, the Wnt and Hedgehog pathways, cadherin and integrin pathways, and the hormone receptors, estrogen (ER) and androgen (AR). Cyclin D is a direct transcriptional target of these pathways. Cyclin D-CDK4/6 phosphorylates RB, which permits cyclin E transcription, which then partners with CDK2 to fully phosphorylate RB, causing S phase–specific transcription. p16 is a member of the INK4A family and a specific CDK4/6 inhibitor. The active CDK4 complex is actually a trimer: cyclin D-CDK4/6-p27, as described in Fig. 2. The CDK4/6i, palbociclib, ribociclib, and abemaciclib are specific CDK4/6 inhibitors.

Figure 1.

All roads lead to cyclin D-CDK4/6, and as such has been a long sought-after therapeutic target. Cyclin D-CDK4/6 is downstream of most oncogenic signaling pathways, including tyrosine kinases, such as EGFR and Erb2/Her2, cytokine receptors, the Wnt and Hedgehog pathways, cadherin and integrin pathways, and the hormone receptors, estrogen (ER) and androgen (AR). Cyclin D is a direct transcriptional target of these pathways. Cyclin D-CDK4/6 phosphorylates RB, which permits cyclin E transcription, which then partners with CDK2 to fully phosphorylate RB, causing S phase–specific transcription. p16 is a member of the INK4A family and a specific CDK4/6 inhibitor. The active CDK4 complex is actually a trimer: cyclin D-CDK4/6-p27, as described in Fig. 2. The CDK4/6i, palbociclib, ribociclib, and abemaciclib are specific CDK4/6 inhibitors.

Close modal
Figure 2.

The activation of cyclin D-CDK4/6 is complex, explaining why the levels of cyclin D or CDK4 are not predictive of CDK4/6i response. Cyclin D is a direct transcriptional target of oncogenic signaling pathways (Fig. 1). It has a very short half-life, and upon assembly with CDK4, the complex rapidly dissociates unless it is stabilized by p27 or p21. p27 binds in two alternative conformations, a closed, inactive conformation or an open, active conformation, and this transition is mediated by Y phosphorylation of p27 itself by a nonreceptor bound Y kinase, such as Brk (Breast tumor Related Kinase; ref. 69). p27 causes the translocation of the complex to the nucleus via its NLS, where the open p27 Y phosphorylated-cyclin D–CDK4 complex is a substrate for additional phosphorylation by CAK. CAK phosphorylates CDK4 on reside T172, causing another conformational change and fully activating the complex. Only this modified complex can phosphorylate RB.

Figure 2.

The activation of cyclin D-CDK4/6 is complex, explaining why the levels of cyclin D or CDK4 are not predictive of CDK4/6i response. Cyclin D is a direct transcriptional target of oncogenic signaling pathways (Fig. 1). It has a very short half-life, and upon assembly with CDK4, the complex rapidly dissociates unless it is stabilized by p27 or p21. p27 binds in two alternative conformations, a closed, inactive conformation or an open, active conformation, and this transition is mediated by Y phosphorylation of p27 itself by a nonreceptor bound Y kinase, such as Brk (Breast tumor Related Kinase; ref. 69). p27 causes the translocation of the complex to the nucleus via its NLS, where the open p27 Y phosphorylated-cyclin D–CDK4 complex is a substrate for additional phosphorylation by CAK. CAK phosphorylates CDK4 on reside T172, causing another conformational change and fully activating the complex. Only this modified complex can phosphorylate RB.

Close modal

Initial attempts at developing CDK inhibitors yielded relatively nonspecific drugs which inhibited multiple CDKs and were toxic above low doses. Palbociclib was the first highly specific CDK4/6 inhibitor, with an affinity for CDK4 and CDK6 over 1,000 times greater than its affinity for other CDKs, such as CDK2 or CDK1 (14). Significant adverse events (AE) in early clinical trials included neutropenia and thrombocytopenia, due to the reliance of bone marrow progenitors on CDK6. Limited efficacy was seen in phase I studies in a variety of solid tumors and mantle cell lymphoma, a cyclin D1 overexpressing tumor type. Based on preclinical data suggesting increased sensitivity in HR+ breast cancer cell lines, which will be discussed later, subsequent clinical trials were initiated in the HR+, HER2- metastatic breast cancer population, eventually leading to approval of the drug in 2015. Two other CDK4/6 inhibitors, ribociclib and abemaciclib, have also been approved for this population. Although these three drugs are all approved for the same indications and all target CDK4/6, transcriptional and proteomic data now suggest that they have different target specificities. Ribociclib has a similar efficacy and toxicity profile to palbociclib, inhibiting both CDK4 and CDK6 (16). Abemaciclib on the other hand may have a greater affinity for CDK4 than CDK6, but also inhibits CDK9, DYRK/HIPK kinases, and GSK3 α/β, resulting in a different toxicity profile, with fatigue as the dose-limiting toxicity and diarrhea as a common AE (17, 18). Abemaciclib also has inhibitory activity against CDK2 and CDK1, suggesting that it may more closely resemble the previous pan-cdk inhibitors (18). In addition, abemaciclib is approved as monotherapy, as well as in combination with letrozole or aromatase inhibitors, as opposed to palbociclib and ribociclib, which are only approved in combination (19). It is also dosed continuously, which may offer kinetic advantages in some patients. At the current time, all three agents are under clinical investigation for a variety of other tumor types, including in combination with other agents.

Molecular subtypes of breast cancer

Molecular analysis has identified two major groups of breast cancer, luminal subtypes, and nonluminal subtypes (20, 21). Luminal subtype cancers, so named because their gene-expression profiles resemble those of luminal breast epithelial cells, are characterized by expression of ER-related genes. Within this group, differences in expression of ER-related and proliferation-related genes define the primary subtypes, Luminal A and Luminal B. The nonluminal subtypes are grouped into the HER2-enriched tumors and basal-like tumors. Molecular subtyping is used clinically to guide treatment in certain scenarios; however, IHC remains the standard for typing breast cancer into the major clinically relevant categories: HR+, HER2+, triple-positive, and triple-negative (TN). There is a general but imprecise correlation between molecular subtypes and receptor phenotype: Luminal A and B subtypes are typically HR+/HER2-, whereas the nonluminal HER2-enriched tumors are typically HR-/HER2+, and nonluminal basal-like tumors are frequently TN. HR+/HER2+ tumors remain the subject of some debate (22–24). Receptor status is important for guiding both initial treatment and metastatic disease. In the metastatic setting, HR+ tumors are treated with hormonal therapy, often now with the addition of anti-CDK4/6 therapy. HER2+ tumors are treated with an anti-HER2 monoclonal antibody, frequently with the addition of chemotherapy, whereas TN tumors are treated with chemotherapy alone and now in some settings with atezolizumab (anti–PD-LI; ref. 25). Furthermore, receptor status helps guide treatment options in the neoadjuvant and adjuvant settings.

ER expression

The initial rationale for using palbociclib in ER/PR+ patients was based on the preclinical work of Finn and colleagues, who discovered that luminal ER+ cell lines responded better to palbociclib treatment, in vitro, than nonluminal subtypes (26). This observation is consistent with the mechanism of action of palbociclib, as CCND1, the gene coding for Cyclin D1, is a transcriptional target of ER (27). Thus, overactivation of ER could lead to increased expression of Cyclin D1, causing DK4/6-mediated cellular proliferation.

Subsequent completed trials of palbociclib in breast cancer have only included patients with ER/PR+, HER2- tumors; thus, there is a sparsity of clinical data to directly support the hypothesis that palbociclib is only active in HR+ tumors without HER2 (6, 7, 28). The limited data come from a phase II study by DeMichele and colleagues, which evaluated palbociclib monotherapy in patients with advanced breast cancer, most of whom had been heavily pretreated (29). Thirty-three of 37 patients in that study were HR+/HER2-, and 4 were TN. Although 7 of the HR+ patients experienced a clinical benefit (21%), none of the 4 TN patients derived benefit. In another study using abemaciclib, 29 of 36 HR+ patients (81%) derived benefit (response or stable disease by RECIST v1.1 criteria), compared with 3 of 9 HR- patients (33%; ref. 17). It is worth noting, however, that the benefit experienced by HR- patients was only stable disease, not regression. The PALOMA-2 and PALOMA-3 trials, which included only HR+, HER2- patients, stratified patients by the level of ER positivity on IHC, but no statistically significant correlation between ER staining intensity and response to palbociclib was found (7, 30). Furthermore, despite only enrolling HR+ patients, 20% of patients exhibited primary resistance (<6 months PFS) in the PALOMA-3 study, and an additional 50% developed resistance within 2 years of starting therapy (7). Thus, hormone-receptor status by IHC may be an imprecise predictor of palbociclib responsiveness. Many HR+ patients do not benefit, whereas other non-HR+ patients, who could potentially derive benefit, do not receive treatment. Investigation into CDK4/6i biomarkers focuses on two interrelated issues: (1) the inability to identify the actionable, active CDK4 target and (2) preexisting or acquired resistance mechanisms that render the drugs ineffective, despite having the CDK4 target. The remainder of this review will summarize the status of the most promising potential biomarkers in both classes (Fig. 3).

Figure 3.

Potential biomarkers for CDK4/6i response. Left: Biomarkers to identify the CDK4 target would need to identify the fully active complex. CDK4T172 phosphorylation (orange P) or p27 pY phosphorylation (red P) delineates the active complex. Increased cyclin D levels should increase the cyclin D–CDK4 complex and additionally increase sequestration of p27 away from CDK2, activating that complex (14, 61). Alternatively, detection of DCAF could identify cells with genetic aberrations leading to increased cyclin D-CDK4 activity. Right: Biomarkers that identify resistance mechanisms would include loss of RB, which would render the presence of CDK4 unnecessary, or increased cyclin E-CDK2 activity due to increased cyclin E or reduced p27 inhibition. FAT, which increases cyclin D levels, might increase the CDK4/6 target directly or alternatively sequester p27 away from CDK2 complexes, increasing p27-free CDK2, so it might fall into both classes of biomarkers.

Figure 3.

Potential biomarkers for CDK4/6i response. Left: Biomarkers to identify the CDK4 target would need to identify the fully active complex. CDK4T172 phosphorylation (orange P) or p27 pY phosphorylation (red P) delineates the active complex. Increased cyclin D levels should increase the cyclin D–CDK4 complex and additionally increase sequestration of p27 away from CDK2, activating that complex (14, 61). Alternatively, detection of DCAF could identify cells with genetic aberrations leading to increased cyclin D-CDK4 activity. Right: Biomarkers that identify resistance mechanisms would include loss of RB, which would render the presence of CDK4 unnecessary, or increased cyclin E-CDK2 activity due to increased cyclin E or reduced p27 inhibition. FAT, which increases cyclin D levels, might increase the CDK4/6 target directly or alternatively sequester p27 away from CDK2 complexes, increasing p27-free CDK2, so it might fall into both classes of biomarkers.

Close modal

Cyclin D1, CCND1 amplification, and D-Cyclin activating features

The CCND1 gene encodes Cyclin D1, one of the three D-type cyclins. Cyclin D1 appears to be the most important of the three in the normal and malignant proliferation of breast tissue, though this may only be due to tissue-specific expression patterns, not a functional difference (31, 32). In addition to being a transcriptional target of ER, Cyclin D1 overexpression can also be driven by several upstream mitogenic signals, which have been proposed as potential mechanisms of endocrine therapy resistance (33). In Finn's preclinical data, CCND1 transcriptome expression was also found to correlate with different cell lines' sensitivity to palbociclib, in vitro, suggesting it may serve as a biomarker for response (26).

The PALOMA-1 trial evaluated CCND1 gene amplification by FISH, as well as Cyclin D1 protein expression by IHC (28, 30). Using either parameter, gene copy number or protein expression, there was no correlation with sensitivity to palbociclib; it was similar in all groups. The lack of predictive value of CCND1 gene amplification by FISH was also seen in the DeMichele study (29). The PALOMA-2 study, the largest clinical trial of palbociclib to date, also included an array of possible biomarkers as one of its secondary endpoints. Cyclin D1 IHC was included in this panel, as well as CCND1 mRNA expression, and neither showed any prognostic significance for palbociclib treatment (28, 34). Thus, although CCND1 levels were predictive biomarkers in cell lines, they were not in patients.

Further insight into the genetic aberrations responsible for Cyclin D–mediated proliferation is provided by Gong and colleagues' analysis of “D-Cyclin Activating Features (DCAF).” Gong and colleagues looked at 560 cell lines from a variety of cancer types to identify other tumor types that might be sensitive to CDK4/6 inhibition (35, 36). They found that cells with DCAF, including CCND1 translocation, CCND1-3 30 UTR loss (resulting in mRNAs lacking destabilizing elements), CCND2 or CCND3 amplification, CCNK (encoding Kaposi's sarcoma virus D-type cyclin), or FBX031 loss (encoding a ubiquitin ligase controlling cyclin D1 stability), tended to be sensitive to abemaciclib. Given the disparity between the prognostic significance of CCND1 mRNA levels in cell lines and in patients, it is unclear if looking at the DCAF themselves in patients will provide useful clinical information. If it does, it may still be a difficult biomarker to use, as multiple analysis tools would need to be involved to identify responsive patients.

p16 and CDKN2A amplification

CDKN2A, the gene encoding p16INK4A, was also assessed for amplification by FISH in the PALOMA-1 trial. p16INK4A is a critical tumor-suppressor protein which inhibits the activity of CDK4/6, and its expression has been shown to correlate with an improved prognosis in breast cancer (37). Low expression of CDKN2A, and thus p16INK4A, was hypothesized to correlate with increased CDK4/6 activity, and increased sensitivity to palbociclib. Finn's preclinical data supported this hypothesis; however, other preclinical data from Gong and colleagues suggested that CDK4/6 inhibition in CDKN2A-deficient cells only resulted in transient arrest, with resistance likely mediated by CDK2 (26, 35). Ultimately, the PALOMA-1 trial showed that CDKN2A copy number was not predictive of response to treatment (28, 30). p16INK4A expression was assessed by IHC in the PALOMA-2 and DeMichele trials, and did not predict response to therapy either (6, 29). CDKN2A mRNA was also looked at in the PALOMA-2 study, and was nonpredictive (34). Green and colleagues suggest that this result was not surprising as CDKN2A levels may be too low in HR+ cells to be predictive (38). They note that all of the high CDKN2A-expressing cell lines from Finn's 2009 panel were HR-, suggesting that selecting for HR+ tumors may eliminate the predictive value of CDKN2A. However, this biomarker may still have value for patient exclusion in other tumor types.

Phosphorylation status of CDK4

CDK4 activity is primarily regulated posttranslationally, by phosphorylation and association with other proteins, such as Cyclin D and p27. It has been shown that phosphorylation of the threonine 172 residue (pThr172) is the rate-limiting step in CDK4 activation (39). It is thus reasonable to hypothesize that tumor cells lacking Thr172 phosphorylation would progress through the cell cycle in a CDK4-independent manner, and thus be resistant to palbociclib. This was demonstrated by Raspé and colleagues, who found that pThr172 status correlated with sensitivity to palbociclib in 20 breast cancer cell lines (40). However, use of pThr172 as a clinical biomarker is limited by the difficulty in detecting it in formalin-fixed, paraffin-embedded (FFPE) tissue, so Raspé and colleagues identified a gene expression profile, which correlated with pThr172, to predict response to palbociclib. Using the profile, they were able to predict sensitivity to palbociclib in 49 of 52 breast cancer cell lines tested. Again, clinical studies will be needed to see if the profile is able to predict response to palbociclib in vivo.

Phosphorylation status of p27 at tyrosine-88

p27 is an important regulator of DK4/6, serving as either an inhibitor or activator of the complex depending on its phosphorylation status at the tyrosine-88 residue (pY88; ref. 41). Patel and colleagues hypothesized that p27 pY88 status could thus be indicative of DK4/6 axis activity, indicating potential sensitivity to CDK4/6 inhibitors. This hypothesis was supported in cell culture experiments, as the level of pY88 directly correlated with CDK4 activity and Palbociclib responsiveness: highly responsive MCF7 cells had lower levels of CDK4 activity and pY88, whereas Palbociclib-insensitive HCC1954 cells had much higher levels of pY88 and CDK4 activity (42). To further test this hypothesis, Gottesman and colleagues developed a dual IHC assay, staining for both total p27 and pY88 specifically. Using FFPE biopsy tissue, they demonstrated that pY88, which is negative in benign breast epithelium, was differentially expressed in otherwise pathologically identical HR+/Her2- tumors. This difference in pY88 stratified tumors in the Her2+ and TN subgroups as well (43). Lack of pY88 suggested that DK4 was inactive, and that these samples would not have the active CDK4 target.

To demonstrate the predictive value of pY88 staining for palbociclib response, Gottesman and colleagues obtained fresh tumor tissue samples from patients undergoing surgery and grew them in explant culture. The explants were stained for Ki-67 as a marker of proliferation at the onset of culture and following 60 hours of treatment with palbociclib or control. In parallel, they stained archival biopsy material from these patients with p27 and pY88 to stratify them into pY88+ or pY88- groups. Explants from the pY88+ group demonstrated a decrease in the percentage of Ki-67+ cells, indicating palbociclib-mediated growth arrest, whereas explants from the pY88- group had unchanged Ki-67 levels in the presence of palbociclib, indicating that they were resistant to Palbociclib-mediated arrest. Interestingly, pY88-positive tumors were seen in Her2+ and TN subtypes, supporting the idea that there may be subgroups of patients in those populations who could benefit from CDK4/6 inhibitors. Although this study demonstrated that the pY88 marker might be used for target identification and might predict primary resistance, additional work using patients treated clinically with CDK4/6i therapy will be required to assay the strength of this marker. Recently, this group reported stratification of patients who had previously been treated with palbociclib and letrozole into responsive and nonresponsive subgroups based on this biomarker (44).

Loss of Rb

Although the above-described biomarkers might help to identify patients with the active CDK4 target, resistance to CDK4/6i may exist a priori and be independent of the presence of the CDK4 target. For example, our knowledge about cell-cycle progression suggests that loss of Rb should confer resistance to Palbociclib. Without Rb, E2F transcription factors are free and active, regardless of CDK4/6 status (14). Preclinical data support the hypothesis that cell lines with lower baseline Rb expression are less sensitive to palbociclib, and that downregulation of Rb is a mechanism of secondary palbociclib resistance, in vitro (26, 45). Loss of Rb at the DNA level is rare in HR+ breast cancer, though it does occur in about one third of TN breast cancers, conferring a worse prognosis overall but increased sensitivity to chemotherapy (46). Consistent with this, less than 10% of patients in the PALOMA-1 and 2 trials, which only enrolled HR+ patients, did not express Rb by IHC (30, 47). Li and colleagues' study, which will be discussed later, found RB1 loss in 9 of 338 patients subsequently treated with CDK4/6i (48). Despite this small number of RB1 loss patients, they found a statistically significant difference in PFS on CDK4/6i therapy, of 3.6 months compared with 10.1 months, for patients with intact RB1.

Given that Rb expression was found to correlate with palbociclib response in preclinical models, Malorni and colleagues identified a panel of 87 genes whose expression corresponds to an Rb loss of function phenotype, despite an intact RB1 gene (49). They found that this signature (RBsig) was able to reliably discriminate between cell lines that are sensitive versus resistant to palbociclib treatment, with resistant cells exhibiting higher levels of RBsig (indicating decreased Rb function), a composite score of genetic changes in Rb-related genes. Similar approaches using different panels have previously identified cells with genetic alterations in the Rb pathway, and it is not clear if any of these offers benefits over each other (50, 51). The panel will need to be applied to tumor samples from palbociclib-treated patients to see if it can predict response clinically.

Cyclin E–CDK2

Given that the Cyclin E–CDK2 complex (EK2) is able to phosphorylate Rb and release E2F, some have speculated that upregulation of Cyclin E or CDK2 may compensate for CDK4/6 inhibition, allowing progression through the G1–S restriction point (42, 45, 52). Herrera-Abreu and colleagues generated palbociclib-resistant MCF-7 breast cancer cells via chronic exposure to the drug and showed that these cells harbored amplifications of the CCNE1 gene, which codes for Cyclin E1 (52). Although inhibition of CDK2 or Cyclin E1 expression using siRNA was ineffective at inhibiting proliferation on its own, the combination of CDK2 or CCNE1 siRNA knockdown with palbociclib caused cell-cycle arrest in palbociclib-resistant MCF-7 cells, suggesting that EK2 activation was the mechanism responsible for DK4/6 bypass.

Patel and colleagues also treated MCF-7 cells with palbociclib to explore induced resistance mechanisms (42). With ongoing palbociclib treatment, MCF-7 cells became resistant by day 6, with continued proliferation despite continually suppressed CDK4 activity. Using an in vitro kinase assay, Patel and colleagues showed that this was due to increased CDK2-mediated phosphorylation of Rb. Interestingly, CDK2 and Cyclin E protein levels were unchanged, suggesting a posttranslational mechanism responsible for increased CDK2 activity. Patel and colleagues found that p27 levels were also decreased in these samples and hypothesized that this may be responsible for increased CDK2 activity. In addition to forming the ternary complex with Cyclin D and CDK4/6, p27 can also serve as an inhibitor of the Cyclin E–CDK2 complex. Thus, ubiquitin-mediated degradation of p27 could allow for compensatory CDK2 activity in the setting of CDK4/6 inhibition.

Guarducci and colleagues generated palbociclib-resistant cell lines from 7 parental breast cancer cell lines sensitive to the drug, and performed gene expression profiling and Western blot to characterize changes associated with resistance (53). Although they observed a high degree of molecular heterogeneity in resistant cells, CCNE1 amplification and loss of RB1 were common events at time of resistance. Furthermore, these molecular changes were consistent with increased Cyclin E and decreased Rb protein expression by Western blot. Guarducci and colleagues then considered the CCNE1/RB1 ratio as a marker of palbociclib resistance, in order to take both into account as one biomarker. They found the ratio to be a better predictor of palbociclib resistance than either marker alone. In addition, they found that a high CCNE1/RB1 ratio predicted resistance to palbociclib in the publicly available NeoPalAna clinical trial dataset (described later), as well as poor overall survival in the METABRIC dataset, independent of molecular subtype.

Turner and colleagues recently reported retrospective gene expression profiling on 302 archival FFPE samples from the PALOMA-3 trial (54). Fifty-three percent of the samples were collected from primary tumors at the time of initial presentation, whereas 47% were collected from metastatic sites immediately prior to study entry, where the patients had already been exposed to at least one line of endocrine therapy. mRNA expressions of 2,534 cancer-related genes were analyzed. Looking specifically at the metastatic samples, patients with low CCNE1 expression had a median PFS of 14.1 months when treated with palbociclib plus fulvestrant, compared with 4.8 months for placebo plus fulvestrant, a statistically significant change. In the high CCNE1-expressing patients, the median PFS was 7.6 months for palbociclib plus fulvestrant-treated patients, compared with 4.0 months for placebo plus fulvestrant-treated patients, and did not reach statistical significance. Thus, the addition of Palbociclib provided a median PFS benefit, but the degree of benefit was decreased in the high CCNE1-expressing patients. A more modest benefit was seen in low CCNE1-expressing patients for the primary tumor samples, which was not statistically significant, consistent with this being a marker of evolved resistance. In the NeoPalAna trial, a study of combined hormonal therapy with palbociclib in the neoadjuvant setting for HR+ patients, increased expression of CCNE1, as well as CCND2 and CCND3 (encoding Cyclin D2 and D3, respectively) also correlated with treatment response (55). Interestingly, a recent in vitro study of 10 gastric cancer cell lines treated with palbociclib found Cyclin E protein expression by Western blot, but not CCNE1 amplification, to be predictive of resistance to palbociclib (56). Thus, there may be multiple mechanisms responsible for Cyclin E upregulation, and multiple assay methods may have to be employed for this to develop into a viable biomarker.

Vijayaraghavan and colleagues also found that CCNE1 gene expression in breast cancer cell lines correlated with palbociclib resistance (57). When they induced full-length Cyclin E, or low-molecular-weight Cyclin E (LMWE) overexpression, in previously sensitive cell lines, however, they found that only LMWE caused the cells to become resistant to palbociclib. LMWE is an oncogenic isoform whose expression leads to RB hyperphosphorylation, bypassing CDK4-dependent phosphorylation, and correlates with a worse prognosis (57, 58). It is created by posttranslational modification of intact Cyclin E; thus, there would not be differences in the CCNE1 mRNA. When they looked at protein expression in a set of TN cell lines, using Western blot, they found that palbociclib-resistant cell lines generally expressed LMWE, whereas sensitive cell lines expressed intact Cyclin E. They performed IHC staining on archival tumor tissue from 109 patients with advanced breast cancer who were subsequently treated with palbociclib, in combination with either letrozole or fulvestrant. Presumably because there are no antibodies specific for LMWE, they used cytoplasmic Cyclin E staining as a surrogate marker, because it had previously been shown to correlate with LMWE expression (59). Fifty-four of the 109 patients (49.5%) were found to have cytoplasmic Cyclin E staining, which was associated with a worse prognosis (P = 0.01). Excluding the 7 Rb- patients, patients with cytoplasmic Cyclin E staining had a median PFS of 13.4 months, compared with 36.5 months for patients with negative cytoplasmic Cyclin E staining. Given that cytoplasmic Cyclin E staining is a negative prognostic factor regardless of treatment, prospective randomized studies will be needed to determine whether it is indeed a useful prospective marker of palbociclib resistance.

CDK6 amplification and loss of the FAT1 tumor suppressor

Similar to the studies described above using palbociclib, Yang and colleagues generated abemaciclib-resistant MCF-7 breast cancer cell lines, and assayed for induced genetic changes associated with treatment. After 21 weeks of exposure, they obtained resistant clones with an IC50 nearly 10-fold higher than the parental clone (60). FISH demonstrated that this was due to gene amplification of the CDK6 locus, which resulted in a 7-fold increase in the levels of CDK6 mRNA and an associated increase in CDK6 protein levels. Knockdown of CDK6 using siRNA restored sensitivity to abemaciclib, and induced overexpression of CDK6 in abemaciclib-sensitive cells was sufficient to cause resistance. Surprisingly, this effect was not seen with induced overexpression of CDK4, suggesting that the increased CDK6 may not merely be increasing CDK4/6 kinase activity. Increased sequestration of p27 or p21 away from CDK2 may also account for this resistance, and thus results of this class of biomarkers may correlate with the increased Cyclin E-CDK2 category described above (14, 61). Yang and colleagues also generated abemaciclib-resistant T47D and CAMA-1 cell lines: The T47D cells exhibited a similar increase in CDK6 mRNA, but also had a prominent decrease in RB1 mRNA. The CAMA-1 cells had minimal changes in CDK6, with more significant changes appearing in RB1 and CCNE1, akin to Guarducci and colleagues and Dean and colleagues' data. Thus, different mechanisms of resistance may develop in different cell lines, under the same experimental conditions, but the same classes appear to persist.

Li and colleagues performed genetic sequencing, using the MSK-IMPACT platform, on pretreatment ER+/HER2- biopsy specimens from 348 patients subsequently treated with palbociclib, ribociclib, or abemaciclib, and correlated the results with PFS. Patients with a deleterious FAT1 mutation had a significantly decreased median PFS of 2.4 months, compared with 10.1 months for patients without FAT1 mutations (48). Loss of FAT1 in FAT1 knockout MCF7 and CAMA-1 cell lines did not promote accelerated growth in the absence of treatment; however, it did render cell lines resistant to all three CDK4/6i. In this setting, loss of FAT1 significantly increased CDK6 expression, and when they examined the MCF-7–resistant cell lines generated by Yang and colleagues, they found that the increased CDK6 in these lines correlated with reduced FAT1 levels.

FAT1 is a member of the cadherin superfamily, which has been found to interact with the Beta-catenin and Hippo signaling pathways. By assessing transcriptional targets of these pathways in parental and FAT1 knockout cell lines, Li and colleagues determined that although the Beta-catenin axis was unchanged, FAT1 knockout cells exhibited downregulation of Hippo-pathway signaling, which subsequently lead to increased CDK6 expression. Induced expression of the intracellular domain of FAT1 was sufficient to restore Hippo signaling and reduce CDK6 expression. Although this seems to be a viable mechanism of resistance, FAT1 mutations are only observed in approximately 6% of metastatic breast cancers prior to CDK4/6i treatment, so they are likely to represent just a small portion of patients with primary resistance, so its usefulness as a predictive biomarker will have to be determined.

Given the difficulty in developing a predictive biomarker for CDK4/6 inhibitors, alternative approaches to guide therapy have been explored. One such approach is to dose patients with a short course of CDK4/6i and to assess response.

Ki-67 expression

Ki-67 is a proliferative marker that has been shown to predict sensitivity to chemotherapy in HR+ breast cancer. Preclinical models have shown that palbociclib has activity against breast cancer cell lines with a wide range of proliferation rates, which was also seen in Dean and colleagues' ex vivo analysis of breast tumors treated with palbociclib in culture (62). Thus, it was not surprising that pretreatment Ki-67 expression did not correlate with response in the PALOMA-1 or PALOMA-2 trials (30, 47). However, changes in Ki-67 expression in response to palbociclib treatment have been used as a surrogate marker of drug sensitivity, and thus can be used to assess the prognostic value of other biomarkers without needing to wait for follow-up data to mature (55, 63). The POP trial found significant decreases in Ki-67 staining after just 14 days of palbociclib treatment before surgery, suggesting that it may be possible to determine primary resistance after just a short course of therapy, provided pre- and posttreatment tumor samples are available (63).

Early ctDNA changes with PI3KCA and ESR1 mutations

Examination of relative changes in the level of circulating cell-free tumor DNA in the plasma (ctDNA) has also shown promise in predicting CDK4/6i response as well. O'Leary and colleagues compared ctDNA levels from patients in the PALOMA-3 trial at baseline and on cycle 1 day 15 of treatment. They chose to look at two loci, PI3KCA and ESR1, which are frequently mutated in breast cancer (64). For patients with identifiable PI3KCA mutations at baseline, there was a significant reduction in PI3KCA ctDNA levels at C1D15 compared with baseline, for both palbociclib + fulvestrant and placebo + fulvestrant arms; however, the decrease was significantly greater in the palbociclib arm. Furthermore, the level of decrease predicted response to therapy (P = 0.0013). Though ESR1 ctDNA levels also decreased significantly at C1D15, they did not predict response to therapy. The authors found that this difference was because PI3KCA is usually a truncal, or originating mutation, present throughout the tumor, whereas ESR1 mutations are usually subclonal, arising in response to endocrine therapy (of note, patients in PALOMA-3 had already received prior endocrine therapy). The limitation of this approach is the need to identify a truncal mutation for each patient, and in the PALOMA-3 samples, only 22% of the patients had a PI3KCA mutation.

mRNA and proteomic signatures

Hafner and colleagues and others have reported that the different CDK4/6i drugs induce specific mRNA and proteomic signatures in treated cells, which in theory could be used to monitor drug response to assess treatment at the molecular level (18). These techniques would have to be adapted to work on circulating tumor cells in order to practically monitor initial treatment response. However, it is unclear how secondary resistance, which may overlay new signatures on top of old signatures, will be assessed. Green and colleagues have used an ADP acyl phosphate probe to measure CDK4 target engagement and demonstrate that target engagement correlates to palbociclib sensitivity (38). Again, how this would be clinically used and how secondary resistance, which may occur on top of original CDK4/6i target engagement, will be determined, is still unclear.

Will a biomarker that predicts response in HR+ breast cancer translate into the other breast cancer subgroups, as we hope use of the CDK4/6i therapies does? Foidart and colleagues showed that TN cell lines which expressed EGFR could be sensitized to palbociclib treatment after transfection with membrane-type-4 matrix metalloproteinase (MT4-MMP) DNA (65). MTR-MMP is a cell surface protein that has been shown to partner with EGFR and enhance its signaling upon binding ligand. They hypothesized that TN tumors which express EGFR and MT4-MMP might be more sensitive to CDK4/6 inhibition, as long as RB remained intact. They subsequently analyzed a series of 72 TN tissue samples, and 37 patient-derived xenografts (PDX), with IHC staining for MT4-MMP, EGFR, and RB. They found that approximately 50% of the TN samples and xenografts stained positive for all three biomarkers. They then treated select PDX-mice, representative of each biomarker combination (EGFR ±, MTR-MMP ±, Rb ±) with vehicle, palbociclib, erlotinib (EGFR inhibitor), or the combination, and assessed tumor growth. PDX positive for MT4-MMP, EGFR, and RB demonstrated significantly decreased proliferation when exposed to palbociclib treatment as compared with the other groups. Cotreatment with palbociclib and erlotinib yielded an additive effect. The data indicate that a subset of TN patients, potentially as high as 50%, may derive benefit from treatment with CDK4/6 inhibitors, and use of these markers might aid in identifying them.

Likewise, the PATRICIA trial is an ongoing study of palbociclib in combination with trastuzumab for patients with previously treated HR+, HER2+, advanced breast cancer (66). Previous work had shown that patients with the HR+, HER2+ phenotype by IHC can fall into one of several molecular subtypes, including Luminal A or B, or HER2-enriched (67). Given palbociclib's effectiveness in luminal subtypes, the authors hypothesized that patients with the Luminal A or B subtype might be more sensitive to this combination strategy than patients with other subtypes, such as HER2-enriched. This hypothesis was confirmed in the first stage of the trial, which showed a PFS of 10.37 months for patients with luminal disease, compared with 3.53 months for patients with other subtypes, with a P value of 0.023. Thus, molecular subtyping might be prudent to select which Her2+ patients are most likely to benefit from this combination strategy. This finding is especially relevant given the multiple clinical trials which are currently ongoing utilizing anti-CDK4/6 and anti-HER2 combination strategies for patients with the IHC HR+/HER2+ phenotype. We eagerly await the publication of their results, but it is likely that responses will be highly variable given the variety of molecular subtypes present in the trial cohorts.

The introduction of CDK4/6 inhibitors into the clinic has been a significant step forward in the treatment of HR+ metastatic breast cancer. The two main issues surrounding their use are (1) acquired drug resistance which develops in the majority of patients and (2) the lack of predictive biomarkers to determine who will benefit from treatment. These two issues are interrelated, as solving both requires an understanding of resistance mechanisms that develop before and during treatment. We have summarized the data from a number of biomarkers, which have been assessed clinically, as well as others which show promise in the laboratory. Unfortunately, to date there does not appear to be any one specific genetic aberration indicative of CDK4/6i sensitivity, as exists for other upstream, targetable enzyme mutations. Unlike many other kinases, which are therapeutic targets, the activation of DK4/6 is too complex for the levels of either of these proteins to be predictive. We have classified biomarker research into two categories: those that identify the actionable CDK4/6 target, such as changes in cyclin D or the phosphorylation status of CDK4 or p27, the lack of which may indicate primary resistance, and those that delineate resistance mechanisms, such as loss of RB or increased cyclin E-CDK2 activity, which can be defined as secondary resistance, even when they appear in the pretreatment condition. Most biomarkers for targeted therapies in the oncology field fall into the first category, but identification of biomarkers of resistance remains a problem universally. For example, although Her2 staining identifies tumors that may respond to trastuzumab, de novo or acquired resistance occurs clinically in 66% to 88% of Her2+ metastatic breast cancer (68). Thus, a focus on both classes of biomarkers should continue as it is likely that some combined use of a marker against the CDK4 target and against resistance mechanisms will produce the most accurate response data. Future clinical trials should include rational biomarker analyses, and additional experimental tissue collection, where possible, to facilitate application of these experimental biomarkers to clinical samples. Another factor will be ease of usage, and the confidence ratio, as a companion diagnostic must be highly predictive or we continue to run the risk of including and excluding patients, similar to the current scenario seen with the lack of any biomarker at all. Finally, many of the current trials for the expanded use of CDK4/6i therapy combines them with additional cytotoxic chemotherapies or radiotherapy, and this may necessitate a different set of biomarkers to predict response. However, identification of the actionable CDK4/6i target or the markers responsible for secondary resistance described in this review will be a good first start to increase our usage of these therapies.

S.W. Blain has ownership interest (including patents) in Concarlo Holdings, LLC, and has consultant/advisory board relationship with Princeton University. No potential conflicts of interest were disclosed by the other authors.

The authors thank Susan Gottesman, Vladislav Tsiperson, and Christopher Roman for critical reading of this article. This work was supported by NIH R01CA201536 (to S.W. Blain).

1.
U.S. Cancer Statistics Working Group
.
U.S. cancer statistics data visualizations tool, based on November 2017 submission data (1999-2015)
.
Bethesda, MD
:
U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute
; 
2018
.
2.
Gradishar
WJ
,
Forero
A
,
Pierce
LJ
. 
National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology. Breast Cancer (Version 1.2018)
. https://www.nccn.org/professionals/physician_gls/pdf/breast.pdf. Accessed October 19, 2018.
3.
Pernas
S
,
Tolaney
SM
,
Winer
EP
,
Goel
S
. 
CDK4/6 inhibition in breast cancer: current practice and future directions
.
Ther Adv Med Oncol
2018
;
10
:
1758835918786451
.
4.
Turner
NC
,
Neven
P
,
Loibl
S
,
Andre
F
. 
Advances in the treatment of advanced oestrogen-receptor-positive breast cancer
.
Lancet
2017
;
389
:
2403
14
.
5.
Ingham
M
,
Schwartz
GK
. 
Cell-cycle therapeutics come of age
.
J Clin Oncol
2017
;
35
:
2949
59
.
6.
Finn
RS
,
Martin
M
,
Rugo
HS
,
Jones
S
,
Im
SA
,
Gelmon
K
, et al
Palbociclib and letrozole in advanced breast cancer
.
N Engl J Med
2016
;
375
:
1925
36
.
7.
Cristofanilli
M
,
Turner
NC
,
Bondarenko
I
,
Ro
J
,
Im
SA
,
Masuda
N
, et al
Fulvestrant plus palbociclib versus fulvestrant plus placebo for treatment of hormone-receptor-positive, HER2-negative metastatic breast cancer that progressed on previous endocrine therapy (PALOMA-3): final analysis of the multicentre, double-blind, phase 3 randomised controlled trial
.
Lancet Oncol
2016
;
17
:
425
39
.
8.
Scheerens
H
,
Malong
A
,
Bassett
K
,
Boyd
Z
,
Gupta
V
,
Harris
J
, et al
Current status of companion and complementary diagnostics: strategic considerations for development and launch
.
Clin Transl Sci
2017
;
10
:
84
92
.
9.
Fang
H
,
Huang
D
,
Yang
F
,
Guan
X
. 
Potential biomarkers of CDK4/6 inhibitors in hormone receptor-positive advanced breast cancer
.
Breast Cancer Res Treat
2018
;
168
:
287
97
.
10.
Guarducci
C
,
Bonechi
M
,
Boccalini
G
,
Benelli
M
,
Risi
E
,
Di Leo
A
, et al
Mechanisms of resistance to CDK4/6 inhibitors in breast cancer and potential biomarkers of response
.
Breast Care (Basel)
2017
;
12
:
304
8
.
11.
Pandey
K
,
An
HJ
,
Kim
SK
,
Lee
SA
,
Kim
S
,
Lim
SM
, et al
Molecular mechanisms of resistance to CDK4/6 inhibitors in breast cancer: a review
.
Int J Cancer
2019
;
145
:
1179
88
.
12.
Portman
N
,
Alexandrou
S
,
Carson
E
,
Wang
S
,
Lim
E
,
Caldon
CE
. 
Overcoming CDK4/6 inhibitor resistance in ER-positive breast cancer
.
Endocr Relat Cancer
2019
;
26
:
R15
30
.
13.
Garrido-Castro
AC
,
Goel
S
. 
CDK4/6 inhibition in breast cancer: mechanisms of response and treatment failure
.
Curr Breast Cancer Rep
2017
;
9
:
26
33
.
14.
Sherr
CJ
,
Beach
D
,
Shapiro
GI
. 
Targeting CDK4 and CDK6: from discovery to therapy
.
Cancer Discov
2016
;
6
:
353
67
.
15.
Hukkelhoven
E
,
Liu
Y
,
Yeh
N
,
Ciznadija
D
,
Blain
SW
,
Koff
A
. 
Tyrosine phosphorylation of the p21 cyclin-dependent kinase inhibitor facilitates the development of proneural glioma
.
J Biol Chem
2012
;
287
:
38523
30
.
16.
O'Leary
B
,
Finn
RS
,
Turner
NC
. 
Treating cancer with selective CDK4/6 inhibitors
.
Nat Rev Clin Oncol
2016
;
13
:
417
.
17.
Patnaik
A
,
Rosen
LS
,
Tolaney
SM
,
Tolcher
AW
,
Goldman
JW
,
Gandhi
L
, et al
Efficacy and safety of abemaciclib, an inhibitor of CDK4 and CDK6, for patients with breast cancer, non-small cell lung cancer, and other solid tumors
.
Cancer Discov
2016
;
6
:
740
53
.
18.
Hafner
M
,
Mills
CE
,
Subramanian
K
,
Chen
C
,
Chung
M
,
Boswell
SA
, et al
Multiomics profiling establishes the polypharmacology of FDA-approved CDK4/6 inhibitors and the potential for differential clinical activity
.
Cell Chem Biol
2019
;
26
:
1067
80
.
e8
.
19.
Verzenio (abemaciclib) [package insert]
.
Indianapolis, IN
:
Eli Lilly and Company
; 
2018
.
20.
Perou
CM
,
Sørlie
T
,
Eisen
MB
,
van de Rijn
M
,
Jeffrey
SS
,
Rees
CA
, et al
Molecular portraits of human breast tumours
.
Nature
2000
;
406
:
747
52
.
21.
Sørlie
T
,
Perou
CM
,
Tibshirani
R
,
Aas
T
,
Geisler
S
,
Johnsen
H
, et al
Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications
.
Proc Natl Acad Sci
2001
;
98
:
10869
74
.
22.
The Cancer Genome Atlas Network
. 
Comprehensive molecular portraits of human breast tumours
.
Nature
2012
;
490
:
61
70
.
23.
Norum
JH
,
Andersen
K
,
Sørlie
T
. 
Lessons learned from the intrinsic subtypes of breast cancer in the quest for precision therapy
.
BJS
2014
;
101
:
925
38
.
24.
Bhargava
R
,
Dabbs
DJ
. 
Luminal B breast tumors are not HER2 positive
.
Breast Cancer Res
2008
;
10
:
404
.
25.
Marra
A
,
Viale
G
,
Curigliano
G
. 
Recent advances in triple negative breast cancer: the immunotherapy era
.
BMC Med
2019
;
17
:
90
.
26.
Finn
RS
,
Dering
J
,
Conklin
D
,
Kalous
O
,
Cohen
DJ
,
Desai
AJ
, et al
PD 0332991, a selective cyclin D kinase 4/6 inhibitor, preferentially inhibits proliferation of luminal estrogen receptor-positive human breast cancer cell lines in vitro
.
Breast Cancer Res
2009
;
11
:
R77
.
27.
Foster
JS
,
Henley
DC
,
Bukovsky
A
,
Seth
P
,
Wimalasena
J
. 
Multifaceted regulation of cell cycle progression by estrogen: regulation of Cdk inhibitors and Cdc25A independent of cyclin D1-Cdk4 function
.
Mol Cell Biol
2001
;
21
:
794
810
.
28.
Finn
RS
,
Crown
JP
,
Lang
I
,
Boer
K
,
Bondarenko
IM
,
Kulyk
SO
, et al
The cyclin-dependent kinase 4/6 inhibitor palbociclib in combination with letrozole versus letrozole alone as first-line treatment of oestrogen receptor-positive, HER2-negative, advanced breast cancer (PALOMA-1/TRIO-18): a randomised phase 2 study
.
Lancet Oncol
2015
;
16
:
25
35
.
29.
DeMichele
A
,
Clark
AS
,
Tan
KS
,
Heitjan
DF
,
Gramlich
K
,
Gallagher
M
, et al
CDK 4/6 inhibitor palbociclib (PD0332991) in Rb+ advanced breast cancer: phase II activity, safety, and predictive biomarker assessment
.
Clin Cancer Res
2015
;
21
:
995
1001
.
30.
Jiang
Y
,
Randolph
S
,
English
P
,
Kim
S
,
Huang
X
,
Press
MF
, et al
442ocell cycle biomarker analysis from the paloma-1/trio 18 palbociclib plus letrozole phase ii study in er-positive/her2-negative advanced breast cancer (abc)
.
Ann Oncol
2014
;
25
:
iv146
.
31.
Carthon
BC
,
Neumann
CA
,
Das
M
,
Pawlyk
B
,
Li
T
,
Geng
Y
, et al
Genetic replacement of cyclin D1 function in mouse development by cyclin D2
.
Mol Cell Biol
2005
;
25
:
1081
8
.
32.
Sicinski
P
,
Weinberg
RA
. 
A specific role for cyclin D1 in mammary gland development
.
J Mammary Gland Biol Neoplasia
1997
;
2
:
335
42
.
33.
Luqmani
YA
,
Alam-Eldin
N
. 
Overcoming resistance to endocrine therapy in breast cancer: new approaches to a nagging problem
.
Med Princ Pract
2016
;
25
:
28
40
.
34.
Finn
RS
,
Liu
Y
,
Martin
M
,
Rugo
HS
,
Dieras
V
,
Im
S-A
, et al
Comprehensive gene expression biomarker analysis of CDK 4/6 and endocrine pathways from the PALOMA-2 study [abstract]
. In:
Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5–9; San Antonio, TX
.
Philadelphia (PA)
:
AACR
;
Cancer Res 2018;78(4 Suppl):Abstract nr P2-09-10
.
35.
Gong
X
,
Litchfield
LM
,
Webster
Y
,
Chio
LC
,
Wong
SS
,
Stewart
TR
, et al
Genomic aberrations that activate D-type cyclins are associated with enhanced sensitivity to the CDK4 and CDK6 inhibitor abemaciclib
.
Cancer Cell
2017
;
32
:
761
76
.
e6
.
36.
Shapiro
GI
. 
Genomic biomarkers predicting response to selective CDK4/6 inhibition: progress in an elusive search
.
Cancer Cell
2017
;
32
:
721
3
.
37.
Peurala
E
,
Koivunen
P
,
Haapasaari
KM
,
Bloigu
R
,
Jukkola-Vuorinen
A
. 
The prognostic significance and value of cyclin D1, CDK4 and p16 in human breast cancer
.
Breast Cancer Res
2013
;
15
:
R5
.
38.
Green
JL
,
Okerberg
ES
,
Sejd
J
,
Palafox
M
,
Monserrat
L
,
Alemayehu
S
, et al
Direct CDKN2 modulation of CDK4 alters target engagement of CDK4 inhibitor drugs
.
Mol Cancer Ther
2019
;
18
:
771
9
.
39.
Bockstaele
L
,
Coulonval
K
,
Kooken
H
,
Paternot
S
,
Roger
PP
. 
Regulation of CDK4
.
Cell Div
2006
;
1
:
25
.
40.
Raspé
E
,
Coulonval
K
,
Pita
JM
,
Paternot
S
,
Rothé
F
,
Twyffels
L
, et al
CDK4 phosphorylation status and a linked gene expression profile predict sensitivity to palbociclib
.
EMBO Mol Med
2017
;
9
:
1052
66
.
41.
Blain
S
. 
Switching cyclin D-Cdk4 kinase activity on and off
.
Cell Cycle
2008
;
7
:
892
8
.
42.
Patel
P
,
Tsiperson
V
,
Gottesman
SRS
,
Somma
J
,
Blain
SW
. 
Dual inhibition of CDK4 and CDK2 via targeting p27 tyrosine phosphorylation induces a potent and durable response in breast cancer cells
.
Mol Cancer Res
2018
;
16
:
361
77
.
43.
Gottesman
SRS
,
Somma
J
,
Tsiperson
V
,
Dresner
L
,
Govindarajulu
U
,
Patel
P
, et al
Tyrosine phosphorylation of p27Kip1 correlates with palbociclib responsiveness in breast cancer tumor cells grown in explant culture
.
Mol Cancer Res
2019
;
17
:
669
75
.
44.
Blain
SW
,
Gottesman
SRS
,
Somma
J
,
Tsiperson
V
,
McGuinness
J
,
Ingham
M
, et al
Tyrosine phosphorylation of p27Kip1 associates with Palbociclib responsiveness in breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29–Apr 3; Atlanta, GA. Philadelphia (PA): AACR
;
Cancer Res
2019
;
79(13 Suppl)
:
Abstract nr LB-224
.
45.
Dean
JL
,
Thangavel
C
,
McClendon
AK
,
Reed
CA
,
Knudsen
ES
. 
Therapeutic CDK4/6 inhibition in breast cancer: key mechanisms of response and failure
.
Oncogene
2010
;
29
:
4018
32
.
46.
Witkiewicz
AK
,
Knudsen
ES
. 
Retinoblastoma tumor suppressor pathway in breast cancer: prognosis, precision medicine, and therapeutic interventions
.
Breast Cancer Res
2014
;
16
:
207
.
47.
Finn
R
,
Jiang
Y
,
Rugo
H
,
Moulder
SL
,
Im
SA
,
Gelmon
KA
, et al
Biomarker analyses from the phase 3 PALOMA-2 trial of palbociclib (P) with letrozole (L) compared with placebo (PLB) plus L in postmenopausal women with ER + /HER2– advanced breast cancer (ABC)
.
Ann Oncol
2016
;
27
Suppl 6
.
48.
Li
Z
,
Razavi
P
,
Li
Q
,
Toy
W
,
Liu
B
,
Ping
C
, et al
Loss of the FAT1 tumor suppressor promotes resistance to CDK4/6 inhibitors via the Hippo pathway
.
Cancer Cell
2018
;
34
:
893
905
.
e8
.
49.
Malorni
L
,
Piazza
S
,
Ciani
Y
,
Guarducci
C
,
Bonechi
M
,
Biagioni
C
, et al
A gene expression signature of retinoblastoma loss-of-function is a predictive biomarker of resistance to palbociclib in breast cancer cell lines and is prognostic in patients with ER positive early breast cancer
.
Oncotarget
2016
;
7
:
68012
22
.
50.
Herschkowitz
JI
,
He
X
,
Fan
C
,
Perou
CM
. 
The functional loss of the retinoblastoma tumour suppressor is a common event in basal-like and luminal B breast carcinomas
.
Breast Cancer Res
2008
;
10
:
R75
.
51.
Miller
TW
,
Balko
JM
,
Fox
EM
,
Ghazoui
Z
,
Dunbier
A
,
Anderson
H
, et al
ERα-dependent E2F transcription can mediate resistance to estrogen deprivation in human breast cancer
.
Cancer Discov
2011
;
1
:
338
51
.
52.
Herrera-Abreu
MT
,
Palafox
M
,
Asghar
U
,
Rivas
MA
,
Cutts
RJ
,
Garcia-Murillas
I
, et al
Early adaptation and acquired resistance to CDK4/6 inhibition in estrogen receptor-positive breast cancer
.
Cancer Res
2016
;
76
:
2301
13
.
53.
Guarducci
C
,
Bonechi
M
,
Benelli
M
,
Biagioni
C
,
Boccalini
G
,
Romagnoli
D
, et al
Cyclin E1 and Rb modulation as common events at time of resistance to palbociclib in hormone receptor-positive breast cancer
.
NPJ Breast Cancer
2018
;
4
:
38
.
54.
Turner
NC
,
Liu
Y
,
Zhu
Z
,
Loi
S
,
Colleoni
M
,
Loibl
S
, et al
Cyclin E1 expression and palbociclib efficacy in previously treated hormone receptor–positive metastatic breast cancer
.
J Clin Oncol
2019
;
37
:
1169
78
.
55.
Ma
CX
,
Gao
F
,
Luo
J
,
Northfelt
DW
,
Goetz
M
,
Forero
A
, et al
NeoPalAna: neoadjuvant palbociclib, a cyclin-dependent kinase 4/6 inhibitor, and anastrozole for clinical stage 2 or 3 estrogen receptor–positive breast cancer
.
Clin Cancer Res
2017
;
23
:
4055
65
.
56.
Min
A
,
Kim
JE
,
Kim
YJ
,
Lim
JM
,
Kim
S
,
Kim
JW
, et al
Cyclin E overexpression confers resistance to the CDK4/6 specific inhibitor palbociclib in gastric cancer cells
.
Cancer Lett
2018
;
430
:
123
32
.
57.
Vijayaraghavan
S
,
Karakas
C
,
Doostan
I
,
Chen
X
,
Bui
T
,
Yi
M
, et al
CDK4/6 and autophagy inhibitors synergistically induce senescence in Rb positive cytoplasmic cyclin E negative cancers
.
Nat Commun
2017
;
8
:
15916
.
58.
Keyomarsi
K
,
Conte
D
,
Toyofuku
W
,
Fox
MP
. 
Deregulation of cyclin E in breast cancer
.
Oncogene
1995
;
11
:
941
50
.
59.
Karakas
C
,
Biernacka
A
,
Bui
T
,
Sahin
AA
,
Yi
M
,
Akli
S
, et al
Cytoplasmic cyclin E and phospho-cyclin-dependent kinase 2 are biomarkers of aggressive breast cancer
.
Am J Pathol
2016
;
186
:
1900
12
.
60.
Yang
C
,
Li
Z
,
Bhatt
T
,
Dickler
M
,
Giri
D
,
Scaltriti
M
, et al
Acquired CDK6 amplification promotes breast cancer resistance to CDK4/6 inhibitors and loss of ER signaling and dependence
.
Oncogene
2017
;
36
:
2255
64
.
61.
Blain
SW
,
Scher
HI
,
Cordon-Cardo
C
,
Koff
A
. 
p27 as a target for cancer therapeutics
.
Cancer Cell
2003
;
3
:
111
5
.
62.
Dean
JL
,
McClendon
AK
,
Hickey
TE
,
Butler
LM
,
Tilley
WD
,
Witkiewicz
AK
, et al
Therapeutic response to CDK4/6 inhibition in breast cancer defined by ex vivo analyses of human tumors
.
Cell Cycle
2012
;
11
:
2756
61
.
63.
Arnedos
M
,
Bayar
MA
,
Cheaib
B
,
Scott
V
,
Bouakka
I
,
Valent
A
, et al
Modulation of Rb phosphorylation and antiproliferative response to palbociclib: the preoperative-palbociclib (POP) randomized clinical trial
.
Ann Oncol
2018
;
29
:
1755
62
.
64.
O'Leary
B
,
Hrebien
S
,
Morden
JP
,
Beaney
M
,
Fribbens
C
,
Huang
X
, et al
Early circulating tumor DNA dynamics and clonal selection with palbociclib and fulvestrant for breast cancer
.
Nat Commun
2018
;
9
:
896
.
65.
Foidart
P
,
Yip
C
,
Radermacher
J
,
Lienard
M
,
Blacher
S
,
Montero-Ruiz
L
, et al
Expression of MT4-MMP, EGFR and RB in triple negative breast cancer strongly sensitizes tumors to erlotinib and palbociclib combination therapy
.
Clin Cancer Res
2018
;
25
:
1838
50
.
66.
Ciruelos
E
,
Villagrasa
P
,
Paré
L
,
Oliveira
M
,
de la
Peña L
,
Pernas
S
, et al
PAM50 intrinsic subtype predicts survival outcome in HER2-positive/hormone receptor-positive metastatic breast cancer treated with palbociclib and trastuzumab: a correlative analysis of the PATRICIA (SOLTI 13-03) trial [abstract]
. In:
Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5–9; San Antonio, TX
.
Philadelphia (PA)
:
AACR
;
Cancer Res 2018;78(4 Suppl):Abstract nr P5-20-19
.
67.
Prat
A
,
Pascual
T
,
Adamo
B
. 
Intrinsic molecular subtypes of HER2+ breast cancer
.
Oncotarget
2017
;
8
:
73362
3
.
68.
Chung
A
,
Cui
X
,
Audeh
W
,
Giuliano
A
. 
Current status of anti-human epidermal growth factor receptor 2 therapies: predicting and overcoming herceptin resistance
.
Clin Breast Cancer
2013
;
13
:
223
32
.
69.
Patel
P
,
Asbach
B
,
Shteyn
E
,
Gomez
C
,
Coltoff
A
,
Bhuyan
S
, et al
Brk/Protein tyrosine kinase 6 phosphorylates p27KIP1, regulating the activity of cyclin D-cyclin-dependent kinase 4
.
Mol Cell Biol
2015
;
35
:
1506
22
.