Macrophage secretory IL-1β promotes docetaxel resistance in head and neck squamous carcinoma via SOD2/CAT-ICAM1 signaling

Abstract

Docetaxel (DTX) combined with cisplatin and 5-fluorouracil has been used as induction chemotherapy for head and neck squamous cell carcinoma (HNSCC). However, the development of acquired resistance remains a major obstacle to treatment response. Tumor-associated macrophages are associated with chemotherapeutic resistance. In the present study, increased infiltration of macrophages into the tumor microenvironment (TME) was significantly associated with shorter overall survival and increased resistance to chemotherapeutic drugs, particularly DTX, in patients with HNSCC. Macrophage coculture induced expression of intercellular adhesion molecule 1 (ICAM1), which promotes stemness and the formation of polyploid giant cancer cells, thereby reducing the efficacy of DTX. Both genetic silencing and pharmacological inhibition of ICAM1 sensitized HNSCC to DTX. Macrophage secretion of IL-1β was found to induce tumor expression of ICAM1. IL-1β neutralization and IL-1 receptor blockade reversed DTX resistance induced by macrophage coculture. IL-1β activated superoxide dismutase 2 and inhibited catalase, thereby modulating intracellular levels of ROS and inducing ICAM1 expression. Arsenic trioxide (ATO) reduced macrophage infiltration into the TME and impaired IL-1β secretion by macrophages. The combinatorial use of ATO enhanced the in vivo efficacy of DTX in a mouse model, which may provide a revolutionary approach to overcoming acquired therapeutic resistance in HNSCC.

Keywords: Cancer; Drug therapy; Macrophages; Oncology.

Figure 1. TAMs impair chemotherapeutic response in patients with HNSCC.

(A) CD163 expression in pathologic specimens of patients with HNSCC was examined by IHC. Representative images demonstrating high CD163 signal (CD163⁺) and low CD163 signal (CD163^–). Scale bar: 50 μm. (B) The correlation between CD163 levels in tissue specimens and therapeutic response to ICT. Positive and negative responses to ICT were defined as ≥70% or <70% decreases in the sum of the longest diameter of the target lesions compared with baseline sum diameters, respectively. Correlations between CD163 levels and (C) PFS or (D) OS were analyzed using the Kaplan-Meier method. FaDu cells were pretreated with different doses of CM for 24 hours and then treated with indicated doses of (E) DTX, (F) cisplatin, and (G) 5-fluorouracil (5-FU) for a further 48 hours. Cell viability was determined by MTT assay. Data were displayed as the means ± SD. For statistical analyses, a 2-tailed unpaired Student’s t test (B), log-rank test (C and D), or 1-way ANOVA with Tukey’s post hoc test (E–G) was used. *, P < 0.05; **, P < 0.01.

Figure 2. CM-induced ICAM1 increases tumor stemness in HNSCC.

(A) Pearson’s correlation between commonly identified proteins in both comparative proteomic analyses analyzed using log₂ ratio of abundance changes. (B) Heatmap clustering analysis of quantified proteins with significant abundance ratios (P < 0.05) in both comparative proteomes. (C) Scatterplots demonstrating the relationship between ratio weights and abundance ratios for each identified protein in the proteomic analysis. The color of protein dots represents the P value for corresponding abundance ratios. (D) ICAM1 expression in OECM-1 cells cocultured with THP-1–derived macrophages and PBMC M1-like or M2-like macrophages. (E) Spearman’s monotonic correlation between ICAM1 and CD163 expression in HNSCC analyzed using the TCGA RNA-Seq database on the GEPIA server. TPM, transcripts per million. ICAM1 in FaDu, OECM-1, and CE146 cells induced by (F) CM or (G) 20% CM at different periods. (H) Spheroid formation in FaDu and SAS cells in the presence or absence of 20% CM for 12 days (n = 3). (I) Number of spheroids formed in ICAM1-KD and control CE146T cells in the presence or absence of 20% CM for 12 days (n = 3). The effect of ICAM1 inhibition by shRNA was validated (right). Effects of (J) 20% CM treatment and (K) ICAM1-KD on CD44 expression in FaDu and OECM-1 cells. (L) mRNA levels of ZEB1 and ZEB2 in OECM-1 cells treated with CM determined by quantitative PCR and normalized to untreated control (set as 1, dashed line). GAPDH, β-actin, and 18S rRNA mRNA were internal controls for gene expression. (M) Levels of EMT-related proteins in OECM-1 cells induced by CM. β-actin, loading control. (N) Transwell migration and (O) invasion assays performed using OECM-1 cells in the presence or absence of 10% CM. Signal quantification with crystal violet extract measured by colorimetric analysis at 570 nm (n = 3). Means ± SD. Two-tailed unpaired Student’s t test (H, I, N, and O). **, P < 0.01.

Figure 3. CM-induced ICAM1 enhances HNSCC resistance to DTX.

(A) Representative IHC images demonstrating strong ICAM1 signal (ICAM1^S), moderate signal (ICAM1^M), and weak signal (ICAM1^W) in pathologic specimens of patients with HNSCC. Correlations between ICAM1 expression and (B) OS or (C) PFS were analyzed according to ICAM1 signals using the Kaplan-Meier method. (D) The expression of α-tubulin in ICAM1-KD and control CE146T cells with or without 10 nM DTX treatment was determined by fluorescence microscopic analysis. DAPI, nuclear staining. (E) Viability of CE146T cells with or without treatment with the ICAM1 inhibitor, A205804 (10 μM), with indicated doses of DTX was determined by MTT assay. The effect of A205804 on ICAM1 inhibition was determined by Western blotting (right panel). (F) Cell viability of ICAM1-KD and control CE146T cells cultured in indicated doses of DTX in the presence or absence of 20% CM was determined by MTT assay. Data were displayed as the means ± SD. For statistical analyses, a 2-tailed unpaired Student’s t test (E), log-rank test (B and C), or 1-way ANOVA with Tukey’s post hoc test (D and F) was used. *, P < 0.05; **, P < 0.01.

Figure 4. CM enhances ICAM1 expression through modulation of intracellular ROS levels.

(A) Mitochondrial superoxide levels and (B) intracellular ROS levels in FaDu cells with or without 20% CM treatment were determined using the tracer dyes, MitoSOX (Invitrogen) and CM-H2DCFDA (DCF, Invitrogen), respectively. (C) The expression of ICAM1 in FaDu and OECM-1 cells with or without 5 mM NAC pretreatment in the presence or absence of 20% CM for 24 hours was determined by Western blotting. (D) The expression levels of SOD1, SOD2, and CAT in FaDu and OECM-1 cells treated with indicated doses of CM were determined by Western blotting. (E) Intracellular ROS levels in FaDu and OECM-1 cells in the presence or absence of 20% CM or 30 μM PIO were determined using the tracer dye, DCF, by flow cytometry. (F) The protein levels of ICAM1 and CAT in FaDu and OECM-1 cells in the presence or absence of indicated doses of CM or 30 or 60 μM PIO were determined by Western blotting. β-actin, loading control. (G) Spheroid formation and (H) cell invasion of FaDu and OECM-1 cells with or without 30 μM PIO treatment. (I) Cell viability of FaDu and OECM-1 cells in the presence or absence of 20% CM or 30 μM PIO for 24 hours was determined under indicated doses of DTX by MTT assay. Data were displayed as the means ± SD. For statistical analyses, a 2-tailed unpaired Student’s t test (A, B, G, and H) or 1-way ANOVA with Tukey’s post hoc test (E and I) was used. *, P < 0.05; **, P < 0.01.

Figure 5. Macrophage secretory IL-1β induces ICAM1 in HNSCC.

(A) Cytokine levels in CM from monocultures of FaDu or THP-1, or FaDu–THP-1 coculture, were analyzed by Luminex Multi-Analyte Profiling (xMAP) system. (B) Expression levels of ICAM1 in OECM-1 cells induced by various kinds of cytokines were determined by Western blotting. (C) Expression levels of ICAM1 in different HNSCC cell lines induced by 3 ng/mL IL-1β were determined by Western blotting. (D) Expression levels of SOD and CAT in FaDu and OECM-1 cells induced by 3 ng/mL IL-1β were determined by Western blotting. Mitochondrial superoxide levels and (B) intracellular ROS levels in OECM-1 cells in the presence or absence of 3 ng/mL IL-1β were monitored using the tracer dyes, (E) MitoSOX and (F) DCF. (G) Expression levels of ICAM1 and CAT in OECM-1 cells in the presence or absence of 3 ng/mL IL-1β or 30 μM PIO for 24 hours were determined by Western blotting. (H) Expression levels of ICAM1, SOD2, and CAT in OECM-1 cells in the presence or absence of 20% CM or 50 nM anakinra for 24 hours were determined by Western blotting. β-actin, loading control. Viability of OECM-1 cells with or without 20% CM treatment in the presence or absence of (I) 50 nM anakinra or (J) 3 ng/mL IL-1β neutralizing antibody (4H5; InvivoGen) was determined under indicated doses of DTX by MTT assay. Data were displayed as the means ± SD. For statistical analyses, a 2-tailed unpaired Student’s t test (E and F) or 1-way ANOVA with Tukey’s post hoc test (I and J) was used. *, P < 0.05; **, P < 0.01.

Figure 6. ATO reduces IL-1β secretion from macrophages.

THP-1–differentiated macrophages cocultured with FaDu cells were treated with indicated dose of chemotherapeutic agents including Asadin (ATO), axitinib, atezolizumab, and cabozantinib. (A) Expression levels of pro–IL-1β in macrophages were determined by Western blotting. (B) Levels of mature IL-1β in CM were measured using ELISA kits (ARG80101; Arigo Biolaboratories). (C) Expression levels of ICAM1 in OECM-1 cells with or without 20% CM treatment in the presence or absence of 1 μM ATO for 24 hours were determined by Western blot assay. (D) Expression levels of ICAM1 in OECM-1 cells with or without coculture with THP-1–differentiated macrophages in the presence or absence of 1 μM ATO for 24 hours were determined by Western blotting. Expression levels of (E) caspase-1 and (F) autophagy-related proteins such as ULK1, LC3B, and p62 in THP-1–differentiated macrophages treated with indicated doses of ATO were determined by Western blotting. β-Actin, loading control. Data were displayed as the means ± SD. For statistical analysis, 1-way ANOVA with Tukey’s post hoc test (B) was used. **, P < 0.01.

Figure 7. ATO improves the efficacy of DTX in a nude mouse model of HNSCC.

(A) Schematic illustration of the animal experiment. (B) Representative images of the tongue before tumor cell inoculation (day 0) and a visible tumor mass (indicated by arrow) in the tongue (day 6). (C) Animals were sacrificed on day 18. Representative images demonstrating tumor masses in the tongues of mice treated with different chemotherapeutic drugs. (D) Tumor volume increases in individual mice after drug treatments (from day 6 to day 18) were calculated and plotted. Short bars indicate the average increase in tumor volume for each group. (E) Tumor masses were sectioned and embedded in paraffin. IHC analyses were performed with the indicated antibodies. Scale bar, 50 μm. (F) Mouse IL-1β concentrations in tumors were measured using ELISA kits (ARG80196; Arigo Biolaboratories). Data were displayed as the means. For statistical analyses, 1-way ANOVA with Tukey’s post hoc test (D and F) was used. *, P < 0.05; **, P < 0.01.

Figure 8. Loss of ATO and DTX synergy in a NOD/SCID mouse model.

(A) Schematic illustration of the animal experiment. (B) Representative images of the tongue before tumor cell inoculation (day 0) and a visible tumor mass (indicated by arrow) in the tongue (day 6). (C) Animals were sacrificed on day 18. Representative images demonstrating tumor masses in the tongues of mice treated with different chemotherapeutic drugs. (D) Tumor volume increases in individual mice after drug treatments (from day 6 to day 18) were calculated and plotted. Short bars indicate the average increase in tumor volume for each group. Data were displayed as the means. For statistical analysis, 1-way ANOVA with Tukey’s post hoc test (D) was used. **, P < 0.01.

Figure 9. Representative working model.

IL-1β secreted by TAMs activates SOD2 and inhibits CAT to modulate intracellular ROS levels, thereby inducing ICAM1 expression. ICAM1 increases tumor stemness and PGCC formation, thereby promoting DTX resistance in HNSCC. Pharmaceutical inhibitors or agents against the IL-1β-SOD2/CAT-ICAM1 pathway may sensitize HNSCC to DTX. The clinical drug, ATO, reduces macrophage infiltration and attenuates IL-1β secretion by targeting macrophages, thereby potentially improving the efficacy of DTX in treating HNSCC.