• HEPATOCELLULAR CARCINOMA
• CELL CYCLE
• PHARMACOTHERAPY

Hepatocytes are the foreign legionnaires in our body. They are exposed to various xenobiotics, which invigorates them and when transformed into malignant cells, renders them drug resistant. So far, treatment with the multikinase inhibitor sorafenib, which blocks Raf, vascular endothelial growth factor receptor and platelet-derived growth factor receptor signalling, is the only approved targeted therapy for advanced hepatocellular carcinoma (HCC). However, sorafenib extends the average life expectancy of patients only 2.8 months over the placebo,1 which offers room for improvement. Unfortunately, several other drugs and combinatorial therapies, such as sorafenib plus erlotinib failed in clinical trials.1 ,2 In this issue of Gut, Bollard et al3 tested the combination of sorafenib with palbociclib in preclinical HCC mouse models. Palbociclib (or PD0332991) is a reversible and orally bioavailable low-molecular-weight cyclin-dependent kinase (CDK)4/6 inhibitor without significant off-target effects.4 The history of palbociclib is tightly linked to the decipherment of G1 cell cycle checkpoint regulation that started in 1991 with the discovery of D-type cyclins. CDK4 and CDK6 were discovered soon thereafter. They bind physically to and are enzymatically activated by D-type cyclins, which leads to phosphorylation and inactivation of retinoblastoma (RB1), the key gatekeeper of G1/S transition. Given the importance RB1 in cancer development, pharmacists started to develop CDK4/6 inhibitors such as palbociclib (Pfizer), ribociclib (Novartis) and abemaciclib (Elli Lilly). However, early clinical trials with palbociclib monotherapy dampened the initial enthusiasm for this drug, although tumour regression was observed in some patients with liposarcoma. A breakthrough was achieved when palbociclib was combined with letrozole, an aromatase inhibitor, for the treatment of oestrogen receptor-positive breast cancer. The potent antitumour effects, observed in this combination therapy, led to Food and Drug Administration approval of palbociclib in 2015.5

The efficacy of cytostatic drugs such as CDK4/6 inhibitors in monotherapies might depend on their potential to induce an irreversible cell cycle exit called senescence. Senescent cancer cells readily acquire a senescence-associated secretory phenotype, which is characterised by cytokine production and attraction of cytotoxic immune cells that contribute to tumour cell killing.6 Bollard et al3 treated 15 HCC cell lines with palbociclib to assess cellular responses. Hep3B cells, which harbour a mutation in RB1, did not respond with cell cycle arrest as demonstrated previously.7 Twelve HCC cell lines displayed quiescence, which is a reversible cell cycle arrest, and two cell lines entered an irreversible senescence-like state. Treatment modifications and higher dosing improved the senescence-inducing capacity of palbociclib. Importantly, senescence was observed in vitro and in HCC xenografts. Previous studies with liposarcomas showed that palbociclib-induced senescence occurred preferentially in tumour cells with low expression of the p53 ubiquitin ligase Mdm2 (mouse double minute 2 homolog).5 However, the two HCC cell lines that entered senescence in the study of Bollard and coworkers were Huh7 and skHep1, which harboured mutated and wild-type p53, respectively.8 This challenges the value of the p53 status as a predictive marker for the senescence-inducing capacity of palbociclib in HCC.

The potential of CDK4/6 inhibitors to induce cancer cell death in monotherapies is limited.5 However, they might prove valuable when combined with drugs that affect mitogen-dependent signalling pathways or levels of D-type cyclins. The activity of D-type cyclin/CDK4/6 complexes is regulated by cytokines, hormone receptors and MAPK/PI3K (mitogen-activated protein kinase/phosphatidylinositol-3-kinase) downstream effector pathways of Ras. Antagonists of these pathways synergise with CDK4/6 inhibitors in G1 cell cycle arrest and might reprogramme their cytostatic capacity to induce senescence or even apoptosis of cancer cells. This mode of action explains the synergistic effects of CDK4/6 inhibitors with letrozole, MEK (mitogen-activated protein kinase kinase) and PI3K inhibitors5 and might also explain the synergistic effect of palbociclib and sorafenib in preclinical HCC mouse models as observed by Bollard et al.3 HCC formation was already reduced after palbociclib monotherapy in Myc;p53-sgRNA mice and in xenografted mice. Tumours showed signs of cellular senescence, but a fraction of tumour cells was negative for the senescence marker SAßGAL (senescence-associated beta-galactosidase) and tumours regrew after palbociclib withdrawal. Importantly, combination with sorafenib enhanced the antitumour effects of palbociclib significantly. Short combinatorial treatment led to tumour relapse, but extended therapy resulted in complete cures. Sorafenib decreased extracellular signal-regulated kinase phosphorylation and cyclin D1 protein levels in HCC cell lines, which might further potentiate its synergistic effects with palbociclib.3

Currently, palbociclib monotherapy is tested in a phase II clinical trial with patients with advanced HCC (NCT01356628). The preclinical data of Bollard and coworkers are promising and lack of efficacy of previous targeted therapies for HCC will facilitate the clinical approval of a palbociclib plus sorafenib combination. Three important issues should be addressed prior to application of this drug combination in patients with HCC: (1) adverse side effects; (2) stratification of patients with HCC into responders and non-responders; and (3) combination with drugs other than sorafenib. Adverse side effects of palbociclib include neutropenia, thrombocytopenia, fatigue, diarrhoea, anaemia and nausea. Despite negating reports about acute liver toxicity,9 there is also evidence for pseudocirrhosis and liver failure in palbociclib-treated patients with metastatic breast cancer.10 The side effects impose a week of drug holiday in current clinical regimens, which affects the senescence-inducing capacity of palbociclib. Therefore, palbociclib might be replaced by abemaciclib, which is less toxic and allows continuous treatment.5

Are there markers for the stratification of responders versus non-responders? Amplification of cyclin E, aberrant activation of CDK2 and loss of p21^Cip1 or p27^Kip1 in human tumours might bypass the requirement for CDK4/6 and confer de novo palbociclib resistance. Otherwise, certain tumours with loss of p16^INK4A or amplification of D-type cyclins showed a profound sensitivity to CDK4/6 inhibitors demonstrating their addiction to RB1 signalling.5 The same molecular changes are frequently found in HCC.1 The most intuitive stratification is certainly based on the RB1 status itself. At least for breast tumours, RB1 inactivation predicts responses to palbociclib.5 Initial preclinical experiments demonstrated that HCC tumour transplants responded equally to palbociclib irrespective of the RB1 status, which was attributed to a compensatory role of RB1-related pocket proteins and argues against RB1 as stratification marker.7 However, Bollard and coworkers extended on these analyses and showed that the RB1 status matters. They reduced RB1 activity in cell lines by CRISPR/Cas9 (clustered regularly interspaced short palindromic repeats/CRISPR-associated 9)-based and shRNA (short hairpin RNA)-based approaches, which blunted palbociclib responses. Although RB1 is infrequently mutated in HCC, analysis of human TCGA (The Cancer Genome Atlas) data revealed a RB1 loss of function signature in 30% of patient with HCC samples.3 Therefore, 70% of patients which retained RB1 signalling might respond to palbociclib.

Apart from sorafenib, combination of palbociclib with other targeted therapies, irradiation, chemotherapy or immunotherapy is conceivable. Here, it should be considered that the cell cycle arrest, opposed by palbociclib, can blunt responses to therapeutic modalities that affect mainly proliferating cancer cells such as cytotoxic drugs. Even immune checkpoint therapies rely on the ability of checkpoint inhibitors to restore proliferative expansion of T cells. In the next years, we will likely see combinations of CDK4/6 inhibitors with other drugs and refinement of dosing and therapy durations (figure 1). They will improve the capacity of CDK4/6 inhibitors to induce cellular senescence or apoptosis and minimise undesirable toxicity. There is hope that palbociclib plus sorafenib or another combination therapy with CDK4/6 inhibitors will lead to durable clinical responses in patients with HCC.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Proposed algorithm for the stratification of patients with advanced hepatocellular carcinoma (HCC) that might benefit from treatment with CDK4/6 inhibitors. The vast majority of human HCC (∼70%) exhibit an intact retinoblastoma 1 (RB1) axis and are presumably responsive to CDK4/6 inhibitors. Based on the data of Bollard and coworkers, these inhibitors should be combined with other therapeutic options (conventional therapies, immunotherapy or targeted therapies) for the effective treatment of patients with HCC. MAPK, mitogen-activated protein kinase; MEK, mitogen-activated protein kinase kinase; mTOR, mechanistic target of rapamycin; PI3K, phosphatidylinositol-3-kinase; TACE, transarterial chemoembolization.