Continuous treatment with abemaciclib leads to sustained and efficient inhibition of breast cancer cell proliferation

Abemaciclib is an oral, selective cyclin-dependent kinase 4 & 6 inhibitor (CDK4 & 6i), approved for hormone receptor-positive (HR+), human epidermal growth factor receptor 2-negative (HER2–) advanced breast cancer (ABC) as monotherapy for endocrine refractory disease, and with endocrine therapy (ET) for initial treatment and after progression on ET. Abemaciclib has also shown clinical activity in combination with ET in patients with high risk early BC (EBC). Here, we examined the preclinical attributes of abemaciclib and other CDK4 & 6i using biochemical and cell-based assays. In vitro, abemaciclib preferentially inhibited CDK4 kinase activity versus CDK6, resulting in inhibition of cell proliferation in a panel of BC cell lines with higher average potency than palbociclib or ribociclib. Abemaciclib showed activity regardless of HER2 amplification and phosphatidylinositol 3-kinase (PI3KCA) gene mutation status. In human bone marrow progenitor cells, abemaciclib showed lower impact on myeloid maturation than other CDK4 & 6i when tested at unbound concentrations similar to those observed in clinical trials. Continuous abemaciclib treatment provided profound inhibition of cell proliferation, and triggered senescence and apoptosis. These preclinical results support the unique efficacy and safety profile of abemaciclib observed in clinical trials.


Filter binding (FB) assays
Compounds were mixed with substrate mix (C-terminal retinoblastoma fragment [CTRF] peptide and ATP/33-P ATP) at a final concentration range of 2 µM-0.1 nM. The mix was incubated for 90 min at room temperature (RT) and the reaction was then stopped by adding 80 µl of 10% ortho-Phosphoric acid. The mix was next transferred to Multiscreen filter plates (Millipore) to retain phosphorylated peptide. The plates were read in a Microbeta Trilux instrument. Reaction mix with excess of EDTA was used as assay "bottom signal"; complete reaction mix without the inhibitor/compound was used as assay "top signal".
For IC 50 determination, results were fitted to a four-parameter dose-response curve using the following

Cell lines and culture condition
A panel of 32 BC cell lines (see Supplementary  Table 1 for details) was obtained from the American Type Culture Collection (ATCC, Manassas, Virginia, USA 30-4500 K) or the German Collection of Microorganisms and Cell Cultures GmbH (DSMZ). Cells were cultured according to ATCC or DSMZ recommendations for fewer than ten passages. Cells were seeded in 96 or 384-well plates and incubated overnight prior to treatment. were thawed and resuspended in IMDM, 10% hiFBS (Fisher), and 1% Pen/Strep (Invitrogen), supplemented with GMCSF 10 ng/ml, G-CSF 10 ng/ml, SCF 100 ng/ml, IL3 10 ng/ml and IL6 10 ng/ml.
The IC 50 of breast cancer cell lines was determine by curve fitting to a four-parameter logistic for each output using GENEDATA SCREENER ® tool: where Y = % inhibition (%Inh), X = concentration yielding y% inhibition, Bottom = minimum value of y attained by curve, Top = maximum value of y attained by curve, slope = steepness of curve at IC 50 . IC 50 : concentration of compound that reduces a given response (ligand binding, enzyme response) by 50%. %Inh = [(median Max-x/ median Max -median Min)]/100 Resulting IC 50 data for each treatment in each cell model were plotted as Geometric means in a waterfall plot using knime workflows. Biomarkers (ER or AR expression, HER2 amplification, or PIK3CA mutational) status was extracted from COSMIC db (COSMIC v79-Nov 2016).

In vitro drug treatment
Abemaciclib and its two main metabolites (internally synthesized) were added to the CD34+ progenitor cells at a concentration of 26 nM (Cmax, fu). After 13 days of maturation the cells per ml were measured using the flow cytometry technology. Data are normalized versus nontreated cells. The same procedure was also followed to investigate the effects of the combination of abemaciclib (100 nM) plus fulvestrant (30 nM) on the CD34 maturation process.
CD34+ bone marrow progenitor cells were treated with abemaciclib (250 nM), palbociclib (250 nM), or flavopiridol (250 nM) upon stimulation (IL3, GCSF, SCF, GM-CSF and IL6 cocktail), and using DMSO as nontreated controls. After 10 days of incubation cells were spun at 300 G, 5 mins and cell pellet was incubated with antibodies against CD13-FITC and CD11b-PE. After 10 minutes of incubation at RT protected from light, cells were washed with buffer (PBS, pH 7.2, 0.5% bovine serum albumin, 2 mM EDTA), spun at 300 G, 5 minutes, and supernatant was discarded. Cells were resuspended in buffer for acquisition in flow cytometer.
To measure the mitochondrial superoxide production after cell treatment, the Mitosox Red Mitochondrial Superoxide Indicator (Thermo, M36008) was used. For this, after cell detachment, the cells were allowed to recover in complete media at 37°C for 20-30 mins. Then, cells were washed in PBS 1x and cell pellet was with 5 μM MitoSOX in HBSS/Ca2+/Mg2+, 10 minutes, 37°C incubated following the vendors instructions. After the incubation time, cells were washed three times with HBSS/Ca2+/Mg2+ and prepared for acquisition in the cytometer (488-nm laser and 585/40 nm filter).

shRNA knockdown/phospho-Rb Western
MDA-MB-453 cells were seeded at a density of 400 000 cells per well in 6-well plates. CDK4 and CDK6 shRNA (Sigma, pLKO.1puro) lentivirus transduction was performed 24 h post-seeding with MOI of 10 in the presence of 10 µg/ml polybrene. Cell media was changed to fresh complete medium after 24 h transduction. Cells were collected 72 h post transduction to test CDK4 and CDK6 knock down and phospho-Rb and total Rb. Antibodies were anti-Rb antibody (Cell Signaling 9309), anti-phospho-Rb antibody (pRB-S780 BD Pharmingen 558385), anti-CDK4 antibody (Abcam ab75511) and anti-CDK6 antibody (Cell Signaling 3136) using ECL-HRP on Fujifilm LAS4000. Actin (Sigma A5441) was used as a loading control.

Analysis of cell viability
MDA-MB-453 CDK4 knockdown stable cell lines, generated with 1 µg/ml puromycin selection, were plated in 96-well plates at 4 000 cells/well. On Day 1, 3, and 7, cells were fixed with fixative Prefer ™ (Anatech, # 410) for 20 min and then stained with 20 µg/ml propidium iodide solution diluted in Phosphate Buffered Saline (PBS) containing 200 μg/ml Ribonuclease A (Sigma R6513). The plates were scanned with ACUMEN EXPLORER ™ to measure DNA content. To monitor senescence, cells were plated into 6-well plates at 150 000 cells/well. At Day 6, senescence was assessed using Cellular Senescence Assay Kit (Cell Biolabs, # CBA-230) per manufacturer's instructions.
To measure apoptosis, CD34+ cells were washed with FACS Flow, spun at 300 G for 5 min, and supernatant was removed. Cell pellet was incubated with Annexin V-FITC for 10 min in dark following vendor's instructions. Cells were then washed with Annexin buffer, centrifuged at 300 G for 5 min, supernatant was discarded, and cell pellet was resuspended in 100 µl of Annexin Buffer before being analyzed in the cytometer. Cells were analyzed using flow cytometry (events were gated for debris exclusion and singlets selection). A minimum of 5 000 cells were analyzed per sample. PI (1:200) was added automatically by the cytometer. The percentage of cells at each apoptotic phase was represented: Annexin V-/PI-(Alive). AnnexinV+/PI-(Early apoptosis), Annexin+/PI+ (Late apoptosis) and Annexin-/PI+ (Dead).
Anti-H2AX pS139-FITC expression was used to measure DNA damage. For this, after 13 days on treatment, cells were washed with PBS and cell pellet was incubated with Viobility 405/520 Fixable Dye for 15 min in the dark. Then, cells were washed and fixed using Inside Fix buffer for 10 min at RT following vendor's guidelines. Cells were then washed with PBS and cell pellet was permeabilized using Permeabilization Buffer A and incubated for 30 min in ice. Next, cells were washed, and incubated with anti-H2AX pS139-FITC for 10 min at RT. Finally, cells were washed and analyzed using the flow cytometer (gated on single and alive cells subpopulations). The mean fluorescence intensity, or median of fluorescence (MFI), was calculated, and compared for individual well and condition.
Cell cycle was analyzed using a Ki67/PI protocol. For this, after 13 days of treatment, cells were washed, fixed, and permeabilized following the same procedure as described above. Then, cells were labeled with Anti-Ki-67-Vio667 for 10min at RT. Next, cells were washed, and the cell pellet was incubated with PI/RNase A (Immunostep, PI/RNASE) for 15 min at RT in the dark. Cells were analyzed using the flow cytometer (gated for debris exclusion and single cells). A minimum of 5 000 cells were analyzed per sample. The different phases of cell cycle were gated as follows: G0 (Ki67-/PIlow), G1 (KI67+/PIlow), S (Ki67+/PImed), G2M (Ki67+/PIhigh). All data were analyzed using FlowJo 10.6 and Graph Pad prism 8.

Senescence assay. SA-b-Gal staining
Beta-Gal staining was performed using Senescencegalactosidase staining kit (Cell Biolabs) according to manufacturer's protocol. Cells were incubated at 37°C until beta-gal staining becomes visible. Development of color was detected under light microscope.
After treatment with compounds, the cells were incubated at 37°C and 5% CO 2 for 4 h before being washed once with PBS and adding 100 μl of fresh media. Cells were incubated once more for 1, 2, or 12 h at 37°C/ 5% CO 2 (washout [WO] step). After the WO, reference compound was re-added in minimal signal wells, to determine the minimal signal after WO (Supplementary Figure 1).
To monitor the phosphorylation of Rb (Ser780), a high content imaging cell-based assay was conducted, as previously used in Torres-Guzmán et al. [1]. After treatment, cells were fixed with 3.7% para-formaldehyde, permeabilized with cold Methanol and blocked with 1% BSA (Sigma) in PBS. Then, cells were treated with mouse anti-Rb (pRb Ser780) antibody (BD Pharmingen 558385) in 1% BSA in PBS overnight at 4°C. The next day, after several washing steps, cells were treated with goat antimouse IgG-Alexa Fluor ™ 488 (Thermo Fisher A11008) in PBS for 1 h at RT. After washing with PBS, 1:1 000 RNase A (Sigma R6513) and Propidium Iodide (PI) (Thermo Fisher P3566) dilution in PBS was added for 1 h at RT. Fluorescence plates were scanned with ACUMEN EXPLORER ™ monitoring Alexa Fluor 488 and PI signals using 488 nm wavelength.

Analysis and statistical considerations
Raw data were analyzed with FlowJo 10.6 software. Graph Pad v8.4.3 software was used for data analysis and representation of final readouts. JMP (Statistical Discovery from SAS) was used for the statistical treatment (ANOVA, pair-wise analysis) of the data.

Flow cytometry analysis
All analysis were carried out using the flow cytometry technology (Macsquant 10, Miltenyi).

Cell number
cells were gated on size (FSC) and internal complexity (SSC) for debris exclusion and the cells per ml (gated on singlets) were used as final readout. The percentage of cell proliferation inhibition normalized versus non-treated cells and staurosporine maximum inhibition was represented.

Apoptosis
Cells were washed with PBS 1X and the cell pellet was incubated with an antibody against Annexin V. After 10 minutes of incubation in dark, cells were washed twice using Annexin buffer V. Finally, cells were analyzed using the flow cytometer where PI (1:100) was automatically added. The percentage of cells at the different phases of apoptosis was represented (PI-/Annexin V-for Alive cells, PI-/Annexin V+ for Early apoptotic cells, PI+/Annexin V+ for Late apoptotic cells and PI+/Annexin V-for Dead cells).

Senescence
For fluorescent detection of β-galactosidase, cells were washed with PBS 1X and fixed using 2% PFA (Acros organics, 119690010) for 10 minutes. Then, cells were washed using PBS+1%BSA and incubated with the cell even green reagent (Thermo, C10841) following the vendors indications (2 h, 31°C, no CO 2 ). Finally, cells were washed and resuspended in 1% BSA in PBS for FACS analysis (488-nm laser and 525/50 nm filter). The results are represented as the percentage of green positive cells (gated on FSC, SSC, and singlets for debris exclusion and doublets respectively).

Potency of abemaciclib for CDK4: biochemicals and breast cancer cell-based assays
Abemaciclib is a more potent inhibitor of CDK4 than CDK6 in breast cancer cell lines.

Biochemical characterization
In biochemical assay, abemaciclib inhibits the kinase activity of CDK4/Cyclin D1 complexes with a Ki ATP = 0.6 nmol/l ± 0.3 nmol/l and of CDK6/Cyclin D3 complexes with a Ki ATP = 8.2 nmol/l ± 1.1 nmol/l, showing that in an in vitro cell-free assay, abemaciclib demonstrates specificity of approximately 14-fold for CDK4/Cyclin D1 over CDK6/Cyclin D3 complexes (Supplementary Table 1), regardless of the biochemical assay technic used (FB or TR-FRET). Under identical conditions, palbociclib did not exhibit such potency, and palbociclib could demonstrate a higher affinity to CDK6/ Cyclin D3 complexes as the IC 50 for CDK6/Cyclin D3 complexes was approximately 3-fold higher than the IC 50 for CDK4/Cyclin D1 complexes ( Figure 1A).

Cellular data
The effect of abemaciclib, palbociclib and ribociclib was further examined in two ER+ cell lines: one cell line CDK4-dependent (MDA-MB-453) and one cell line CDK6-dependent (NCI-H-1568). In the cellular assay, abemaciclib and ribociclib demonstrated a higher potency in the CDK4-dependent cell line, and abemaciclib was more potent than ribociclib in the CDK4dependent cell line (Supplementary Table 2). Consistent with the biochemical assay, palbociclib did not exhibit any selectivity towards CDK4/Cyclin D1 complexes and demonstrated a similar level of inhibition in both cell lines ( Figure 1C).

CDK4 plays a critical role in breast cancer
The role of CDK4 in cell proliferation was assessed in CDK4 knockdown stable cell lines generated in the AR+ breast cancer cell MDA-MB-453 (Supplementary Figure  2A) and cell proliferation and senescence were assessed. The kinase activity of CKD4/Cyclin D1 is involved in the cell cycle and triggers the progression from G1 to S1, and cell proliferation. In CDK4 knockdown cell lines CDK4-1 shRNA and CDK4-4 shRNA, the cells stopped growing and multiplying, thus indicating the interruption of cell proliferation (Supplementary Figure 2B). CDK4 knockdown also led to cell senescence (Supplementary Figure 2C). Taken together, these results demonstrate the dependency of the cell cycle on CDK4 in breast cancer cells expressing higher levels of CDK4. CDK4 is thus a key target in breast cancer therapy.

Breast cancer panel sensitivity to CDK4 & 6 inhibitors and biomarker analysis
Abemaciclib inhibits cell proliferation in a wide range of breast cancer cell lines, showing activity regardless of HER2 amplification, PI3KCA and BRCA gene mutation status.

Neutrophils maturation and neutropenia
CDK4 & 6 inhibitors share a common mechanism of action to induce neutropenia, which differs from the mechanism of action of chemotherapy agents and others pan-CDKs inhibitors (paclitaxel or flavopiridol) (Supplementary Figure 4). Abemaciclib metabolites, M2, and M20, show a similar effect on myeloid maturation of progenitor bone marrow cells. Cell surface markers (CD13 and CD11b) of myeloid maturation are less impacted by abemaciclib treatment than with palbociclib or a pan CDK inhibitor (Supplementary Figure 5). The combination with fulvestrant does not increase the impact on in vitro hematopoiesis (Supplementary Figure 6).
Flow cytometry analysis of stimulated healthy CD34+ cells (IL3, GCSF, SCF, GM-CSF and IL6 cocktail) for 10 days for myeloid cells to mature into neutrophils. Healthy CD34+ cells (Supplementary Figure  5A) were treated with abemaclib (b), palbociclib (c) or flavopiridol (d) upon stimulation and using DMSO as nottreated controls (a). CD13 and CD11b surface markers were monitored as a reference for mature neutrophil. Phase I corresponds to myeloblast population, Phase II corresponds to promyelocytes population, Phase III corresponds to myelocytes and Phase IV corresponds to metamyelocytes and band neutrophils.