Peptidylarginine deiminase 3 modulates response to neratinib in HER2 positive breast cancer

Abstract

Neratinib is a tyrosine kinase inhibitor that is used for the therapy of patients with HER2+ breast tumors. However, despite its clinical benefit, resistance to the drug may arise. Here we have created cellular models of neratinib resistance to investigate the mechanisms underlying such resistance. Chronic neratinib exposure of BT474 human HER2+ breast cancer cells resulted in the selection of several clones resistant to the antiproliferative action of the drug. The clones were characterized biochemically and biologically using a variety of techniques. These clones retained HER2 levels similar to parental cells. Knockdown experiments showed that the neratinib-resistant clones retained oncogenic dependence on HER2. Moreover, the tyrosine phosphorylation status of BT474 and the resistant clones was equally sensitive to neratinib. Transcriptomic and Western analyses showed that peptidylarginine deiminase 3 was overexpressed in the three neratinib-resistant clones studied but was undetectable in BT474 cells. Experiments performed in the neratinib-resistant clones showed that reduction of PADI3 or inhibition of its function restored sensitivity to the antiproliferative action of neratinib. Moreover, overexpression of FLAG-tagged PADI3 in BT474 cells provoked resistance to the antiproliferative action of neratinib. Together, these results uncover a role of PADI3 in the regulation of sensitivity to neratinib in breast cancer cells overexpressing HER2 and open the possibility of using PADI3 inhibitors to fight resistance to neratinib.

Fig. 1. Generation and characterization of neratinib-resistant cells.

A Schematic representation of the generation of neratinib resistant cells. BT474 cells were cultured for 3 months in the presence of different concentrations of neratinib. Twenty-three clones were able to grow in the presence of 10 nM neratinib and 1 in the presence of 25 nM neratinib. B Dose-response effect of neratinib on BT474 and neratinib-resistant clones. Cells were plated and 24 h later treated with the indicated doses of neratinib. After 3 days of treatment, cells were counted, and values normalized to untreated controls. Results are shown as mean ± SD of quadruplicates of an experiment that was repeated more than three times. C Plot of the effect of 5 nM neratinib on the indicated cells. **p < 0.05. D IC₅₀ values for the cell lines shown, calculated from data shown in A and using the GraphPad Prism 8 software. E, F Effect of other anti-HER2 therapies on the different cell models generated. Cells were plated and treated with the indicated doses. Proliferation was measured at 3 (E, F) or 5 days (G, H). Results are shown as mean ± SD of quadruplicates of an experiment that was repeated at least twice. I Comparison of the antiproliferative effect of the different anti-HER2 treatments used at 5 nM on the indicated parental or neratinib-resistant cells. **p < 0.05.

Fig. 2. Analysis of HER2 levels and localization in neratinib-resistant cells.

A HER2 levels and activation are maintained in neratinib-resistant clones. Fifteen micrograms of cell lysates were used to analyze HER2 levels by Western. In parallel, HER2 tyrosine phosphorylation was determined by immunoprecipitation of 50 μg of protein extracts with the anti-HER2 antibody, followed by Western with an antiphosphotyrosine antibody. Calnexin was used as a loading control. The position of the M_r markers is indicated. B Quantitation of HER2 (top) or pHER2 (bottom) levels shown in A. Results are shown as mean ± SD of 2 independent experiments. C Analysis of cell surface HER2 on neratinib-resistant cells. Cell monolayers were incubated with 10 nM trastuzumab for 2 h at 4 °C, and cell surface HER2 analyzed as described in materials and methods. D Cell-surface HER2 levels in parental and resistant cell lines analyzed by flow cytometry. Cells were incubated with 10 nM trastuzumab (green line) or PBS (black line) for 1 h at RT and analyzed as described in materials and methods. E Detection of HER2 by immunofluorescence. Cells were fixed and stained with trastuzumab to detect HER2 (red) or DAPI to localize DNA (blue). Scale bar: 50 μm.

Fig. 3. Effect of neratinib on HER2 and downstream targets.

A Action of neratinib on HER2, pHER2, S6 and pS6. Cells were treated with the indicated doses of neratinib for 6 h, lysed, and the proteins either directly analyzed by Western blot (HER2, S6 and pS6), or 50 μg of cell lysates immunoprecipitated with anti-HER2 followed by Western against phosphotyrosine. Calnexin was used as a loading control. B Quantitation of pHER2 (top) and pS6 (bottom) levels shown in (A). Each bar shows the mean ± SD of two independent experiments. C=untreated controls. C Knockdown of HER2 in parental and neratinib resistant cells. Cells were infected with viruses containing control vector (pLKO) or two different shRNA sequences targeting HER2 (shRNA #1 and shRNA #2). Cells were lysed and HER2 expression was measured by Western blot (top). Calnexin was also used as a loading control. HER2 knockdown effect on the proliferation rate was analyzed by cell counting after 3 days (bottom). Graph bars represent the mean ± SD of triplicates of an experiment that was repeated twice. **p < 0.01, calculated by the Mann–Whitney U test.

Fig. 4. Genomic characterization of neratinib-resistant cells.

A Principal component analysis (PCA) of data obtained from RNA microarrays from BT474 and neratinib-resistant clones. All cell lines were grouped according to its intrinsic variability in the 3 axes of the three-dimensional graph (PCA1, 2 or 3) that explains the original variation from among the data set. Replicates for each sample are shown in the same color. B Differentially expressed genes (DEGs) for each resistant clone compared to the parental cell line. C Venn diagram showing both individual and common differentially expressed genes between all resistant clones compared to the parental cell line. D DEGs commonly deregulated among the three neratinib-resistant clones and parental BT474. Cut-off for DEGs was ≥±2-fold change with a p value of 0.01. Genes that meet both criteria are colored green, if downregulated, or red, if upregulated. The y-axis represents the significance as −10 × log of p value. The x-axis represents the base 2 log of the mean differential expression (MDE). E Genes with a higher fold change underexpressed (COLEC12, LMO3, and UPK1A) or overexpressed (MGP, PADI3, and SERPINB5) in the resistant clones compared to the parental cell line BT474. F Analysis of the expression of MGP, PADI3 and SERPINB5 genes comparing normal and tumor tissues of patients with invasive breast carcinoma, using the TNMplot web tool. The p value of the statistical analysis (Mann–Whitney U test) provided by the database is shown.

Fig. 5. PADI3 overexpression in BT474 parental cell line.

A PADI3 and MGP expression levels in BT474 and the neratinib-resistant clones. PADI3 or MGP levels were analyzed by direct Western blot (top and middle panels) of 50 μg of protein extract with the appropriate antibodies. Loading controls used were calnexin (for PADI3) or GAPDH (for MGP). Quantitative analysis of PADI3 levels (bottom) for each cell line, represented in the graph as mean ± SD of four different experiments. B PADI3 or Flag-PADI3 expression in the different cell lines tested. Protein extracts from the BTRN-resistant clones isolated after transfection of Flag-PADI3 into BT474 cells were prepared and separated by SDS-PAGE. PADI3 levels (top) were determined as in A. The membrane was then stripped and re-probed with an anti-Flag antibody (middle). Calnexin was used as a loading control (bottom). C PADI3 expression levels of Flag-PADI3 clones and BT474 cell line. Data from (B) and another three independent experiments were used for the generation of the graphic, and represent the mean ± SD of the mentioned samples. D PADI3 implication in neratinib resistance. BT474, Flag-PADI3 clones and BTRN cells were plated and treated with 5 nM neratinib. After 3 days of treatment, cells were counted, and the data represented as the mean ± SD of triplicates (normalized to untreated controls) of a representative experiment that was repeated three times. **p < 0.01, calculated by the Mann-Whitney U test.

Fig. 6. Genetic and pharmacological inhibition of PADI3 in BT474 and neratinib-resistant clones.

A Knockdown of PADI3 in parental and neratinib-resistant cells. Cells were infected with lentivirus containing shRNA control (pLKO) or the indicated shRNA sequences targeting PADI3 (shRNA #36 and shRNA #37). After puromycin selection, cells were lysed and gene silencing was evaluated by Western blot by an anti-PADI3 antibody. Calnexin was used as a loading control. B Effect of PADI3 knockdown on cell proliferation rate. The proliferative capacity of transduced cells was evaluated after 3 days by cell counting. The results show the percentage of the mean ± SD of triplicates compared to the shRNA control (pLKO), from a representative experiment that was repeated twice. C Analysis of PADI3 knockdown on the response to neratinib. Transduced cells were treated with 5 nM neratinib for 3 days and counted. Results show the percentage of the mean ± SD of triplicates compared to the pLKO cells of a representative experiment that was repeated three times. **p < 0.01, calculated by the Mann–Whitney U test. D Pharmacological inhibition of PADI3 with Cl-amidine. Dose-response analysis of the effect of Cl-amidine on the proliferation of BT474 cells and its respective neratinib-resistant clones. Cells were counted after 3 days of treatment with the indicated doses and data are represented as the mean ± SD of triplicates. E Cells were plated and treated with 10 μM Cl-amidine, 5 nM neratinib or both, and the re-sensitization capacity of Cl-amidine to neratinib response was evaluated after 3 days by cell counting experiments. Graph bars represent the mean ± SD of triplicates normalized to untreated cells of an experiment that was repeated three times. **p < 0.01, calculated by the Mann–Whitney U test.