IL-6 and IL-8 secreted by tumour cells impair the function of NK cells via the STAT3 pathway in oesophageal squamous cell carcinoma

Abstract

Background: Recurrence and metastasis are the leading causes of tumour-related death in patients with oesophageal squamous cell carcinoma (ESCC). Tumour-infiltrating natural killer cells (NK cells) display powerful cytotoxicity to tumour cells and play a pivotal role in tumour therapy. However, the phenotype and functional regulation of NK cells in oesophageal squamous cell carcinoma (ESCC) remains largely unknown.

Methods: Single cell suspensions from blood and tissue samples were isolated by physical dissociation and filtering through a 70 μm cell strainer. Flow cytometry was applied to profile the activity and function of NK cells, and an antibody chip experiment was used to identify and quantitate cytokine levels. We studied IL-6 and IL-8 function in primary oesophageal squamous carcinoma and NK cell co-cultures in vitro and by a xenograft tumour model in vivo. Western blotting was used to quantitate STAT3 (signal transducer and activator of transcription 3) and p-STAT3 levels. Finally, we performed an IHC array to analyse IL-6/IL-8 (interleukin 6/interleukin 8) expression in 103 pairs of tumours and matched adjacent tissues of patients with ESCC to elucidate the correlation between IL-6 or IL-8 and clinical characteristics.

Results: The percentages of NK cells in both peripheral blood and tumour tissues from patients with ESCC were significantly increased in comparison with those in the controls and correlated with the clinical characteristics. Furthermore, the decrease in activating receptors and increase in inhibitory receptors on the surface of tumour-infiltrating NK cells was confirmed by flow cytometry. The level of granzyme B, the effector molecule of tumour-infiltrating NK cells, was also decreased. Mechanistically, primary ESCC cells activated the STAT3 signalling pathway on NK cells through IL-6 and IL-8 secretion, leading to the downregulation of activating receptors (NKp30 and NKG2D) on the surface of NK cells. An ex vivo study showed that blockade of STAT3 attenuated the IL-6/IL-8-mediated impairment of NK cell function. Moreover, the expression of IL-6 or IL-8 in tumour tissues was validated by immunohistochemistry to be positively correlated with tumour progression and poor survival, respectively.

Conclusions: Tumour cell-secreted IL-6 and IL-8 impair the activity and function of NK cells via STAT3 signalling and contribute to oesophageal squamous cell carcinoma malignancy.

Keywords: IL-6; IL-8; Nature killer cell; Oesophageal squamous cell carcinoma; STAT3 signalling.

Fig. 1

Percentages of NK cells in the peripheral blood of 35 healthy volunteers and in the peripheral blood, tumour and adjacent tissues of 52 ESCC patients. a Representative flow cytometry analysis of CD3-CD56+ NK cells in the peripheral blood, tumour and adjacent tissues. Whole peripheral blood-, tumour- and adjacent-tissue-derived cell suspensions were stained with an APC-conjugated antibody (Ab) to human CD56 and a FITC-conjugated Ab to human CD3. CD3–CD56+ NK cell percentages analysed from the lymphocyte gate as defined by the FSC and SSC dot plot. b Statistical analysis of CD3–CD56+ NK cell and CD3+ T cell percentages in the peripheral blood, tumour and adjacent tissue. Data analysed by Student’s t-test are shown as the mean ± SEM, *p < 0.05, ***p < 0.001, and ****p < 0.0001. c Representative flow cytometry analysis of CD3-CD56^bright and CD3-CD56^dim NK cells in the peripheral blood and tumour tissues of ESCC patients

Fig. 2

Phenotypic features of NK cells from the peripheral blood of normal volunteers and peripheral blood, tumour and tumour-adjacent tissues of ESCC patients. a Peripheral blood-, tumour- and adjacent tissue-derived cell suspensions were stained with Abs to CD3, CD56, CD16, NKp30, NKp44, NKp46, NKG2D, CD226, NKG2A, and CD158b. NK cells were gated as CD3-CD56+ events, and the expression levels of CD16, NKp30, NKp44, NKp46, NKG2D, CD226, NKG2A, and CD158b were then analysed. b Symbols represent individual values from 9 to 21 ESCC patients or volunteers analysed individually, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Functional characteristics of NK cells in ESCC patients. c Representative dot plots of KI67, granzyme B, and perforin expression levels in NK cells from the peripheral blood of normal volunteers and paired patient blood, tumour-adjacent and tumour tissues from each ESCC patient. d Statistical analysis of KI67+, granzyme B+, and perforin+ NK cell percentages. Symbols represent individual values from 15 to 16 ESCC patients or volunteers analysed individually. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Each experiment was repeated three times

Fig. 3

Primary oesophageal squamous carcinoma cell isolation, identification and co-culture. a ESCC#1 and ESCC#2 morphology under a light microscope. b Representative nude mice with subcutaneous tumours derived from 1 × 10⁶ ESCC#1 or ESCC#2 cells. c Representative images of H&E and IHC staining for CK5/6 and P63 in mouse xenograft tumours are shown. Scale bar of H&E staining images indicates 50 μM. Scale bar of IHC staining images indicates 20 μM d Isolated and pure NK cells from volunteer peripheral blood obtained by using an NK cell purification kit to achieve approximately 95% NK cell purity with less than 1% T cells. e RPMI1640 media (NC), the supernatant of KYSE150, EC9706, ESCC#1 or ESCC#2 cells co-cultured with NK cells. The CD56 + CD3^_ NK cells were stained with CD16, NKp30, NKp44, NKp46, NKG2D, CD226, NKG2A, and CD158b and analysed by flow cytometry. f Statistical analysis of NKp30+, NKG2D+, NKp44+, NKp46+, CD226+, NKG2A+, and CD158b + NK cell percentages

Fig. 4

IL-6 and IL-8 inhibit the function of NK cells in vitro. a Representative isobaric map of KI67, granzyme B and perforin in the NK cells of five different groups. b Statistical analysis of KI67+, granzyme B+ and perforin+ NK cell percentages. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. c Antibody chip assay showing that the levels of IL-6 and IL-8 are highest in the KYSE150, EC9706, ESCC#1 and ESCC#2 cells supernatants. d IL-6 and IL-8 expression was assessed in xenograft tumour specimens by IHC. Scale bar of IHC staining images indicates 50 μM. e Different conditional media were used to stimulate NK cells for 24 h, and NKp30, NKG2D, NKG2A, KI67, and granzyme B levels were detected by flow cytometry. f Statistical analysis of NKp30+, NKG2D+, NKG2A+, KI67+, and granzyme B+ NK cell percentages. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. g An LDH assay was performed to detect NK cell cytotoxicity in four groups. Each experiment was repeated three times

Fig. 5

IL-6 and IL-8 inhibit the function of NK cells in vivo. Xenograft model: a A total of 1 × 10⁶ ESCC#1 cells were injected into the left armpit of nude mice. After three rounds of NK cell injection, mice were sacrificed, and their tumours were measured. b The mean tumour volumes of the four groups are shown. c The number and malignancy degree of xenograft tumours was assessed in xenograft tumour specimens by H&E. Scale bar of H&E-stained images indicates 200 μM. d Statistical analysis of the number of lung neoplastic foci of four groups of nude mice. IL-6 and IL-8 inhibit NK cell function by increasing STAT3 phosphorylation. e Flow cytometry analysis of STAT1, STAT2, STAT3, and STAT6 expression levels in NK cells after three stimulations with different types of NK cell supernatant for 2 h. f Statistical analysis of STAT1+, STAT2+, STAT3+, and STAT6 + NK cell percentages. g Western blotting was used to detect STAT3 phosphorylation levels in NK cells after co-culture with ESCC#1 supernatant, co-culture with ESCC#1 supernatant+anti-IL-6 supernatant, and co-culture with ESCC#1 supernatant+anti-IL-8 supernatant for 2 h. Statistical analysis of p-STAT3 protein expression levels in NK cells. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Each experiment was repeated three times. h Western blotting was used to detect STAT3 phosphorylation levels in NK cells after co-culture ESCC#2 supernatant, co-culture ESCC#2 supernatant+anti- IL-6 supernatant, and co-culture ESCC#2 supernatant+anti- IL-8 supernatant stimulation for 2 h. Statistical analysis of p-STAT3 protein expression levels in NK cells. i Western blotting was used to detect STAT3 phosphorylation levels in NK cells after RPMI1640 (NC), 100 ng/mL rIL-6 or 100 ng/mL IL-8 stimulation for 2 h. Statistical analysis of p-STAT3 protein levels in NK cells. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Each experiment was repeated three times

Fig. 6

STAT3 feedback assay and correlation between IL-6 and multiple clinical parameters. a Flow cytometry was performed to detect the expression levels of NKP30, NKG2D, KI67 and granzyme B in co-culture models stimulated with 10 mM Stattic for 24 h in vitro. b Statistical analysis of NKp30+, NKG2D+, KI67+, and granzyme B+ NK cell percentages in three groups. Each experiment was repeated three times. c The expression level of IL-6 in paired tumour and adjacent tissues was detected by IHC. Scale bar of IHC staining images indicates 20 μM. d The percentage of IL-6-expressing cells was analysed for putative correlations with multiple clinical parameters. For the cumulative survival curves, patients were separated into two groups by the median value of tumour-infiltrating IL-6-positive cell percentages, and Kaplan-Meier plots were used to calculate cumulative survival differences. *p < 0.05, **p < 0.01, and ****p < 0.0001. Each dot represents 1 patient

Fig. 7

Correlation between IL-8 and multiple clinical parameters. The mechanistic action of IL-6 or IL-8 impairs NK cell function. a The expression level of IL-8 in paired tumour and adjacent tissues was detected by IHC. Scale bar of IHC staining images indicates 20 μM. b The percentage of tumour-infiltrating IL-8+ cells correlates with multiple clinical parameters of ESCC. The percentage of IL-8+ cells was analysed for putative correlations with multiple clinical parameters. For the cumulative survival curves, patients were separated into two groups by the median value of tumour-infiltrating IL-8+ cell percentages, and Kaplan-Meier plots were used to calculate cumulative survival differences. *p < 0.05, **p < 0.01, and ****p < 0.0001. Each dot represents 1 patient. c Schematic diagram. Tumour-derived IL-6 or IL-8 acts on NK cells in an autocrine manner. IL-6 and IL-8 reduce NKp30 and NKG2D expression on NK cell surfaces by enhancing STAT3 phosphorylation, which enhances tumour malignancy and impairs NK cell function