Aurora kinase inhibition overcomes cetuximab resistance in squamous cell cancer of the head and neck

Abstract

Squamous cell cancer of the head and neck (SCCHN) is the sixth leading cause for cancer deaths worldwide. Despite extense knowledge of risk factors and pathogenesis about 50 percent of all patients and essentially every patient with metastatic SCCHN eventually die from this disease. We analyzed the clinical data and performed immunohistochemistry for Epidermal growth factor receptor (EGFR) and Aurora kinase A (Aurora-A) expression in 180 SCCHN patients. Patients characterized by elevated EGFR and elevated Aurora-A protein expression in tumor tissue represent a risk group with poor disease-free and overall survival (EGFR(low)Aurora-A(low) versus EGFR(high)Aurora-A(high), p = 0.024). Treating SCCHN cell lines with a pan-Aurora kinase inhibitor resulted in defective cytokinesis, polyploidy and apoptosis, which was effective irrespective of the EGFR status. Combined Aurora kinase and EGFR targeting using a monoclonal anti-EGFR antibody was more effective compared to single EGFR and Aurora kinase inhibition. Comparing pan-Aurora kinase and Aurora-A targeting hints towards a strong and clinically relevant biological effect mediated via Aurora kinase B. Taken together, our findings characterize a new poor risk group in SCCHN patients defined by elevated EGFR and Aurora-A protein expression. Our results demonstrate that combined targeting of EGFR and Aurora kinases represents a therapeutic means to activate cell cycle checkpoints and apoptosis in SCCHN.

Figure 1. EGFR and Aurora-A transcript levels in SCCHN and clinical outcome

A public database (www.oncomine.com) was searched for gene expression analyses studies that compare AURORA-A transcript levels in control tissue and SCCHN samples from patients who were alive or dead [22]. Shown is the log2 median-centered relative intensity of expression for AURORA-A (AURKA, upper panel, tumor versus control tissue: p = 0.002, [reporter: 34851_at]) and EGFR (lower panel, tumor versus control tissue: n.s., [reporter: 1537_at]).

Figure 2. EGFR and Aurora-A expression in tumor tissue and adjacent normal mucosa

(A) Histological assessment of EGFR and Aurora-A protein expression by immunohistochemistry. Shown are representative tumor samples that were graded as negative/low expression (middle panel), high expression (lower panel) and normal mucosa control tissue (upper panel). Bar equals 100 μm. (B) Within each patient sample the expression of Aurora-A and EGFR was assessed in normal adjacent tissue and tumor tissue. The differences are highly significant. Aurora-A: p<0.001; EGFR: p<0.001. The staining score is defined in the material and method section.

Figure 3. EGFR and Aurora-A expression assessed by IHC is an adverse prognostic factor in SCCHN

(A) EGFR: the difference in disease-free survival for patients with expression above median (green curve; n = 90) is not statistically different from the survival of patients with expression below median (blue curve; n = 90). p = 0.10. (B) Aurora-A: the difference in disease-free survival for patients with expression below median (blue curve; n = 90) is not statistically different from the survival of patients with expression above median (green curve; n = 90). p = 0.21. (C) The difference in disease-free survival of patients with EGFR^high and Aurora-A^high is statistically different from the survival of patients who are characterized by EGFR^low and Aurora-A^low. p = 0.024. The staining score is defined in the material and method section.

Figure 4. Expression and activity of Aurora kinases and EGFR in SCCHN cell lines

(A) Six SCCHN cell lines were assessed by immunoblotting for the expression of Aurora-A and Aurora-B, for Aurora kinase activity measured by Histone H3 phosphorylation at serine10 (S10-HH3), and for EGFR protein levels. (B) Upper panel: AURORA-A and AURORA-B transcript levels were assessed by real-time qRT-PCR. Shown is the relative expression normalized to the expression of Ubiquitin. Lower panel: Expression of EGFR analyzed by RT-PCR. None of the SCCHN cell lines express the EGFRvIII mutant. Transiently transfected NIH-3T3 cells expressing EGFRvIII (3T3-EGFRvIII) were included as a control. (C) Upper panel: CAL cells were treated with 200 nM Cetuximab for the indicated time and assessed by immunoblotting for suppression of EGFR downstream target phosphorylation. Lower panel: Treatment of FADU cells with 5 nM Pan-Aurora kinase inhibitor R763 for the indicated time. The activity of Aurora kinases was assessed by immunoblotting for S10-HH3. (D) SCCHN cell lines were treated for 24 hr with R763 at the indicated concentrations or carrier alone (0 nM). The representative histograms show the DNA content assessed by propidium iodide (PI) staining.

Figure 5. Combined exposure to EGFR antibody and Aurora kinase inhibitor results in fortified growth inhibition and apoptosis

(A) SCCHN cells were treated for a total of 14 days with cetuximab (200 nM), the Aurora kinase inhibitor R763, the combination of both (Cet+ R763), or carrier only (control). The cell number was counted at the indicated times and the -fold increase in cell number calculated. Note that the increase in cell number is given in a logarythmic scale. The combination of cetuximab and Aurora kinase inhibitor resulted in a significantly reduced -fold increase after 14 day treatment period in all cell lines investigated in comparison to all other conditions (Cet alone, R763 alone, control; p<0.05). (B) The indicated SCCHN cells were cultured for a 48 hr period with the indicated conditions (cetuximab 200 nM, R7635 nM) and assessed for DNA content by PI staining. The percentage of polyploid cells with a DNA content >4n is given. (C) Analysis of the cells shown in (B) for apoptosis (Annexin V-positive, PI-negative cell fraction) by flow cytometry. The bars represent the mean ± SD of 3 independently performed experiments. Statistically significant differences are marked (* indicates p<0.05).

Figure 6. Selective Aurora-A inhibition versus pan-Aurora kinase inhibition in combination with Cetuximab

(A) FADU cell were treated with 10 nM Mln for the indicated time. The effect of Aurora-A inhibition was assessed by immunoblotting for serine10-phosphorylated Histone H3 (S10-HH3). (B) Mln treatment (10 nM) for 48 hr resulted in a significant (p<0.05) but moderate increase of polyploid cells (>4n DNA content) as evaluated by PI flow cytometry. (C) Combined Aurora-A inhibiton with 10 nM Mln and EGFR inhibition with 200 nM cetuximab treatment results a significantly reduced cell number increase. (D) The indicated SCCHN cell lines were treated for 48 hr with carrier only or cetuximab plus R763 (Cet+ R763) or cetuximab plus Mln (Cet+Mln). The percentage of polyploid cells as defined by a DNA content >4n was measured by flow cytometry of PI stained cells. (E) Cells were treated as in (C). The percentage of apoptotic cells was assessed by Annexin V flow cytometry. The bars represent the mean ± SD of 3 independent experiments. The differences between Cet+ R763 versus Cet+Mln treatment are significant (p<0.05) for all cell lines tested with regard to polyploidy and with regard to apoptosis.