Determination of the Proteomic Response to Lapatinib Treatment using a Comprehensive and Reproducible Ion-Current-Based Proteomics Strategy

The breast cancer cell lines SKBR3, BT474, EFM-192a, HCC1954, JIMT-1, UACC-732 and MDA-MB-453 were maintained in RPMI-1640 supplemented with 10% FBS. The SKBR3, BT474, and EFM-192a cell lines are lapatinib sensitive, with IC50 values below 1µM, while JIMT-1, UACC-732 and MDA-MB-453 are lapatinib-insensitive, with IC50 values >1µM ¹⁷.

Drug treatments were applied singly or in combinations as follows Laptinib 1µM (Sequoia Sciences, Saint Louis, MO, USA), 150 nM Herceptin (Roche IN, USA), 150 nM Afatinib (Sequoia Sciences) and 20 µM capecitabine (Sigma-Aldrich. St.Loius, MO, USA). Cell pellets were collected pre- and post-drug treatment and lysed in 50 mM Tris, pH 8, 150 mM NaCl, 2% NP-40, 0.5% sodium deoxycholate, and 0.1% SDS. For Western blot, cell lysates were centrifuged at 16,000 g for 20 minutes, after 12 hours of drug exposure. For mass spectrometry (MS), the protein lysate was clarified by ultracentrifugation (140,000g, 40 min, 4 °C). Protein concentration was determined by the BCA protein assay kit (Thermo Fischer Scientific, Rockford, IL USA).

IC50 assays were carried out on the SKBR3 cell line with UA62784 (Sigma-Aldrich) at concentrations of 25nM to 300nM and GSK2923295a (Cytokinetics Inc.) at concentrations of 6nM to 300nM. Combination assays were also carried out at these concentrations with 50nM Lapatinib. Drug treatments were carried out 24 hours after cells were seeded (4*10⁴ cells/96 well) and cell survival measured 5 days later by acid phosphatase assays ¹⁸.

Cell lysates from the SKBR3 cell line (+/- 12 hours 1µM lapatinib, n=6 biological replicates) were tryptically digested using an on-pellet-digestion procedure described previously ¹⁵. A customised nano-LC system ¹⁵, was used to separate peptides during a 5-hour LC gradient on a 50 cm 75 µm i.d, C18, 3 µm, 100A column. Mobile phase A was 0.1% formic acid in 2% acetonitrile and mobile phase B was 0.1% formic acid in 84% acetonitrile. The flow rate was 250 nL/min and the gradient profile was (i) a linear increase from 3% to 10% B over 5 min; (ii) an increase from 10 to 24% B over 115 min; (iii) an increase from 24 to 38% B over 70 min; (iv) an increase from 38 to 60% B over 50 min; (v) an increase from 60 to 97% B in 35 min, and finally (vi) isocratic at 97% B for 25 min. The optimal loading amount of peptide was identified experimentally and a loading mass of 6 µg per injection was employed per sample.

LTQ/ Orbitrap data was acquired over a period of 275 minutes, one scan cycle included an MS1 scan (m/z 300-2000) at a resolution of 60 000 followed by seven MS2 scans by LTQ, to fragment the seven most abundant precursors. The target value for MS1 by Orbitrap was 4 × 10⁶. The fragmentation type was CID with a normalized collision energy of 35%.

Sieve (Fiona build, v. 1.2, Thermo Scientific), was used for quantitative data analysis. All peptides differing significantly between the control and treated (Fisher’s combined probability test, p-value < 0.05) were selected for protein identification. Relative abundance of an individual protein was calculated as the mean AUC ratio for all peptides derived from that protein. Protein ratios were defined as the average abundance of a protein in lapatinib treated samples/control samples. This number was divided into 1 to be converted into +/- fold changes. Identifications were matched against a non-redundant human database derived from the Swissprot database (Feb 2010). The precursor mass tolerance was set to 25 ppm and a mass tolerance of 1.0 Da; fixed modification was carbamidomethyl and variable modifications methionine oxidation. Requirements for a successful identification was matching of at least 2 unique peptides, a peptide probability of >95%, a protein probability of >99% and Sequest restrictions of deltaCn scores of greater than 0.10 and XCorr scores that achieves a 0.5% peptide FDR were employed.

Raw LC/MS data was also analysed using Scaffold 3 software (Portland, OR) with protein identifications carried out through the Sequest server, as described above. GO annotations were retrieved from a human non-redundant Uniprot database and protein cellular localisation and cell processes represented as a percentage of the overall GO annotations retrieved.

Equal quantities of protein lysates pre- and post-drug treatment (n=3 biological replicates) were subjected to Western blotting ¹⁹. Antibody binding was visualised by incubating the blot for 5 minutes with ECL Plus Western Blotting Detection substrate (RPN2132, GE Healthcare, PA, USA) and florescence emission captured by scanning blots at 457 nm excitation, 520 nm emission (PMT 450) on a Typhoon Variable Trio 9400 scanner. Antibodies for TET2 (S-13, sc-136926), HER2 (^3B5, ab16901), CENPE (C-7488) and b-actin (A3854) were purchased from Santa Cruz Biotechnologies (CA, USA), Abcam (Cambridge, UK) and Sigma-Aldrich respectively.

qRT-PCR was carried out as previously described ¹⁰. RNA was isolated with Rneasy Mini kit (Qiagen, CA, USA, n=3 biological replicates). Primers, buffer and dNTPs were supplied by a High Capacity cDNA Reverse Transcription kit (Applied Biosystems). CENPE specific FAM-labelled primer (Hs0106824_m1, Applied Biosystems) was used to quantify CENPE cDNA by qPCR; GAPDH specific FAM label primer (Hs9999905_m1) was used to measure GAPDH cDNA, which acted as an endogenous control.

3x10⁴ SKBR3 cells/well were transfected over 24 hours, in a 24 well plate, using 1 µl siPORT Neofx transfection reagent (Applied Biosystems). siRNA knockdown was performed using 30nM scrambled control (Negative Control #2, Applied Biosystems), and 30nM CENPE siRNA (S2917, Applied Biosystems). Neofx was incubated with serum free Optimem media for 10 minutes, mixed with diluted siRNA, and incubated for another 10 mins. The Neofx-siRNA mix was then applied to cells. Post-24 hours fresh media +/- 100nM lapatinib was added to transfected cells; drug treatments lasted 5 days. Cells were then trypsinised and incubated with 1 part cell suspension: 3 parts Guava viacount reagent and incubated in the dark for 5 minutes before viable cells were counted on a Guava Easycyte (EMD Millipore, MA, USA).

A comprehensive, in-depth proteomic investigation is essential for a study of this nature. As the whole cell lysate is highly complex, a large number of tryptic peptides are retrieved by the on-pellet digestion procedure. To achieve sufficient chromatographic separation, high run-to-run reproducibility of retention times we employed a custom nano-LC/nanospray configuration.

In total, 1224 unique proteins were identified with high confidence and subsequently quantified. The full list of the proteins analyzed in this study, as well as the information on the peptides identified, is shown in SI Table 1. Among the 1224 protein analyzed, quantitative proteomic profiling of the SKBR3 cell line revealed an altered abundance of 67 proteins in lapatinib-treated compared to untreated cells. Of these, 21 demonstrated an increased abundance and 46 a decreased abundance in the treated cells. The cut-off for biomarker discovery was based on the calculation of the false-positive biomarker discovery rate, as described in our previous publications ^{14, 15}. Listed in Table 1, divided by function, are the protein names, number of peptides, p-values, and fold change of these proteins.

A Gene Ontology (GO) analysis of the identified proteins was performed using DAVID Bioinformatics Resources 6.7 (http://david.abcc.ncifcrf.gov/). A large number of proteins from plasma membrane and various organelles were identified indicating a comprehensive protein recovery by our gel-free sample preparation procedure 1A).

Figure 1A.Classifications of the 1224 unique proteins by Biological process, Cellular Compartment and Molecular function.

HER2 and TET2 were chosen for validation of LC-MS/MS data by Western blot. Both proteins were assessed in 7 cell lines (4 sensitive to lapatinib, 3 insensitive), Figure 1B. While TET2 was significantly changed in all 4 lapatinib sensitive cell lines, and not in the insensitive, HER2 only showed a significant increase in the SKBR3 cell line (p<0.05).

Figure 1B.TET2 and HER2 expression, in the absence (-) or presence (+) of 1µM lapatinib after 12 hours, in lapatinib sensitive cell lines and lapatinib insensitive cell lines (highlighted in bold) C) TET2 and D) HER2 densitometry. Fold change = Control vs. Lapatinib treated. * represents significance at p<0.05 by Students t-test.

CENPE, a protein with increased abundance (2.3 fold) in lapatinib-treated cells was also chosen for validation as it represented a possible target for therapeutic intervention as a specific drug-based inhibitor was available. Changes to CENPE abundance were assessed by Western blot (Figure 2A) with significant increases found in 4 lapatinib-sensitive cell lines (p<0.05). Additionally, qRT-PCR analysis, carried out to assess if any related changes were occurring in mRNA expression (Figure 2B), showed significant increases in CENPE mRNA in 4 lapatinib-sensitive cell lines (Figure 2B).

Figure 2.Expression of CENPE, in the absence (-) or presence (+) of 1µM lapatinib after 12 hours, in lapatinib sensitive cell lines and lapatinib insensitive cell lines (highlighted in bold). A) By western blot including densitometric measurement of protein fold change (control vs. Lapatinib treated). B) qRT-PCR measurement of expression changes of mRNA (control vs. Lapatinib treated) * represents significance at p<0.05 ** at p<0.01 by Students t-test

To determine if this trend of increased CENPE abundance was purely a lapatinib drug response, or if CENPE could alter the toxicological response to lapatinib in a clinically-relevant manner, CENPE protein levels were tested in response to a) other HER2 targeting agents and b) clinically relevant combinations of lapatinib with other drugs. In the SKBR3 and BT474 cell lines both afatinib (irreversible HER2 and EGFR inhibitory small molecule agent) and trastuzumab (HER2-inhibitory monoclonal antibody) treatment alone resulted in significantly decreased CENPE protein abundance, in contrast to lapatinib alone treatment (Figure 3).

Figure 3.A).CENPE protein expression in response to a 12 hour treatment with 150nM Afatinib, 150nM Trastuzumab (Her), 150nM Trastuzumab + 1µM Lapatinib (Lap) and 1µM Lapatinib + 20µM Capecitabine (Cap) with densitometric measurement of fold change (control vs. drug treated) in B) the SKBR3 cell line and C) the BT474 cell line. * represents significance at p<0.05 ** at p<0.01 by Students t-test

In contrast to the decreases shown when cells were treated by trastuzumab alone, the combination of lapatinib and trastuzumab resulted in 2-3 fold increases in CENPE protein in both the SKBR3 and BT474 cell lines (Figure 3). The combination of lapatinib and capecitabine also resulted in a similar increase in CENPE protein in both cell lines (Figure 3).

To evaluate if the combination of lapatinib and CENPE inhibition resulted in decreased cell growth, compared to either lapatinib alone or CENPE inhibition alone, siRNA knockdown of CENPE was carried out.

Knockdown of CENPE in SKBR3 cells resulted in approximately a 70% decrease in CENPE protein after 5 days (Figure 4A). This treatment had very little effect on cell survival (<1%). Lapatinib-treated cells showed a 63% decrease in cell survival. The combination of lapatinib and CENPE knockdown resulted in a even greater decrease in cell survival of 85% (Figure 4A).

Figure 4.A) The effect of CENPE knockdown by siRNA, with and without 100nM Lapatinib, on cell viability. Knockdown of CENPE expression confirmed by western blot. B) Effect on cell viability (after 5 days) by the CENPE inhibitor UA62784, alone and in combination with 50nM Lapatinib. C) Effect on cell viability (after 5 days) by the CENPE inhibitor GSK923295A, alone and in combination with 50nM Lapatinib. * represents significance at p<0.05 ** at p<0.01 by Students t-test

To determine if small molecule inhibitors would demonstrate a similar effect as siRNA, UA62784, a laboratory grade inhibitor of CENPE, was tested alone in the nM range (0-300nM) and in combination with 50nM lapatinib. The IC50 value of UA62784 in the SKBR3 cell line was 144nM, comparable but slightly higher than those published for other cancer cell lines ²⁰. At the lower nM ranges of UA63784, the addition of lapatinib resulted in significant decreases in cell survival compared to lapatinib or UA63784 alone (Figure 4B); the IC50 value of UA63784 when in combination with lapatinib dropped to 15nM.

The IC50 for another CENPE inhibitor, GSK23295A alone was lower at 16nM, than for UA62784 alone. Combination of this agent with 50 nM lapatinib resulted in an IC50 of 5nM for GSK23295A. Again at the lower end of the treatment range the combination of lapatinib and GSK23295A displayed synergy (Figure 4C).