A Novel Role for the Interleukin-1 Receptor Axis in Resistance to Anti-EGFR Therapy

Abstract

Cetuximab (CTX) is a monoclonal antibody targeting the epidermal growth factor receptor (EGFR), commonly used to treat patients with metastatic colorectal cancer (mCRC). Unfortunately, objective remissions occur only in a minority of patients and are of short duration, with a population of cells surviving the treatment and eventually enabling CTX resistance. Our previous study on CRC xenopatients associated poor response to CTX with increased abundance of a set of pro-inflammatory cytokines, including the interleukins IL-1A, IL-1B and IL-8. Stemming from these observations, our current work aimed to assess the role of IL-1 pathway activity in CTX resistance. We employed a recombinant decoy TRAP IL-1, a soluble protein combining the human immunoglobulin Fc portion linked to the extracellular region of the IL-1-receptor (IL-1R1), able to sequester IL-1 directly from the medium. We generated stable clones expressing and secreting a functional TRAP IL-1 into the culture medium. Our results show that IL-1R1 inhibition leads to a decreased cell proliferation and a dampened MAPK and AKT axes. Moreover, CRC patients not responding to CTX blockage displayed higher levels of IL-1R1 than responsive subjects, and abundant IL-1R1 is predictive of survival in patient datasets specifically for the consensus molecular subtype 1 (CMS1). We conclude that IL-1R1 abundance may represent a therapeutic marker for patients who become refractory to monoclonal antibody therapy, while inhibition of IL-1R1 by TRAP IL-1 may offer a novel therapeutic strategy.

Keywords: MAPK; cetuximab; colon cancer; colonspheres; consensus molecular subtype; inflammation; resistance.

Figure 1

IL-1R1 abundance predicts response to CTX in patients. Expression analysis of IL-1 receptor in 67 CRC patients treated with CTX. Responsive patients (OR) as well as stable disease (SD) display low amounts of IL-1R1 compared to the non-responsive patients (PD). The differences in IL-1R1 expression between individual groups (OR/SD/PD) were determined by one-way analysis of variance (ANOVA), followed by Fisher LSD test.

Figure 2

A recombinant decoy containing IL-1R1 inhibits Caco-2 growth. (A) Western blot analysis of Caco-2 TRAP IL-1 and Fc. Actin served as loading control. 500,000 cells were plated in medium with 10% FBS, then cells were serum starved overnight. The day after cells were harvested and total proteins extracted for TRAP IL-1 detection. (B) Western blot analysis of three replicates of Caco-2 Fc and Caco-2 TRAP IL-1 soup. 500,000 cells were plated in DMEM supplemented with 10% FBS and after 5 days the soup was harvested and 20 μL used for the analysis. (C) Colony-forming assay of Caco-2 Fc and Caco-2 TRAP IL-1. 4000 cells/well were plated and grown in the absence or presence of CTX (5 μg/mL) for 10 days in medium containing 10% of serum. Cells were then fixed, stained with crystal violet and photographed. Representative figures (left) and quantification (right) of the covered areas by ImageJ are provided. The statistic was calculated by 2-way ANOVA, *** p < 0.0005. These experiments were repeated at least three times. (D) Western blot analysis of Caco-2 TRAP IL-1 clones soup. 1.35–1.44–1.47–1.11 and 1.51 are clones derived from a single cell. Each clone soup was collected 5 days after seeding. TRAP IL-1 (purified protein) and Fc are intended as positive and negative controls respectively. (E) Clones from D were seeded and both living and death cells were counted. Statistical analysis was performed by one-way ANOVA, comparing the mean of proliferation of each clone to the control cells. Dunnet correction for multiple comparisons was applied. **** p < 0.0001. (F) Cell count of Caco-2 Fc and Caco-2 TRAP IL-1 (clone 1.35). 100,000 cells/Petri were seeded with 10% of serum. After 24 h medium was changed with 10% of serum in the presence or absence of CTX (5 μg/mL) and cells were counted after 24, 48 and 72 h. A 2-way ANOVA was performed, by comparing the matched values for each time point (24, 48 and 72 h) to the Fc control cells. **** p < 0.0001.

Figure 3

Caco-2 TRAP IL-1 display decreased growth in suspension as colonspheres. (A) 4× magnification of Caco-2 Fc and Caco-2 TRAP IL-1 producing “spheroid-like” structures, under the indicated treatments. Scale bar 100 μm. (B) Number of filled spheroids presented as average ± S.E.M. 2-way ANOVA with Bonferroni Test, ** p < 0.01; **** p <0.0001; (C) Quantification of spheroid size measurements under the indicated treatments in 10% FBS medium supplemented with EGF 10 ng/mL and CTX 5 μg/mL. Columns represent volume averages ± S.E.M, 1-way ANOVA, ** p < 0.01; **** p < 0.0001. (D) 3D confocal microscopy of colonspheres Fc and TRAP IL-1. In the left panel is reported the DAPI and phalloidin staining. In the middle panel are featured single optical sections collected at 15 μm intervals. In the panel on the right, the xz optical section passing through the maximum diameter of the spheroids (upper part) and 3D rendering focused on the nuclear density and z-depth of the spheroids (yellow square), reported as a scale of colors, red indicates a 5 μm depth and blue 35 μm depth (bottom part).

Figure 4

Bimodal activation of ERK and AKT after stimulation with IL-1A. (A) Western blot analysis of phospho-AKT s473 (pAKT) and phospho-ERK (pERK) levels in Caco-2 Fc and Caco-2 TRAP IL-1 cells (time course). At day one, 500 k cells were plated for each condition in medium with 10% FBS. At day 2, cells were serum starved overnight and the day after IL-1A (10 ng/mL) was added to the medium of growing cells for 1, 3, 6, 8, 12 and 24 h. After treatment cells were harvested, total proteins extracted and quantified. Monoclonal antibody against total AKT and ERK1/2 served as loading control. (B) Quantification of pAKT and pERK by Image Lab is provided. (C) Western blot analysis of phospho-EGFR Tyr1068, EGFR and IL-1R1 in Fc and TRAP IL-1 cells treated as in A. B-actin served as loading control.

Figure 5

IL-1 receptor expression predicts survival in CRC patients. (A) A cohort of 1211 patients was divided into two groups according to IL-1 receptor abundance. In the Kaplan-Meier plot, the black line represents patients with an overall low IL-1R1 expression, while the red line represents subjects with high expression of IL-1R1. For each patient, the relapse free survival (RFS) is reported over time and expressed in months. (B) mRNA abundance of IL-1R1 in the four consensus molecular subtypes (CMSs). (C) Stratification of patients using the CMS criterion. Patients are divided into four subtypes. CMS1 (Immune): hypermutated, microsatellite unstable and strong immune activation. CMS2 (Canonical): epithelial, marked WNT and MYC signaling activation. CMS3 (Metabolic): epithelial and evident metabolic dysregulation. CMS4 (Mesenchymal): prominent transforming growth factor-b activation, stromal invasion and angiogenesis. (D) KM plot of IL-1 receptor expression using the de Sousa Classification of Colon Cancer subtypes.