Bacterial endotoxin enhances colorectal cancer cell adhesion and invasion through TLR-4 and NF-kappaB-dependent activation of the urokinase plasminogen activator system

Abstract

Perioperative exposure to lipopolysaccharide (LPS) is associated with accelerated metastatic colorectal tumour growth. LPS directly affects cells through Toll-like receptor 4 (TLR-4) and the transcription factor NF-kappaB. The urokinase plasminogen activator (u-PA) system is intimately implicated in tumour cell extracellular matrix (ECM) interactions fundamental to tumour progression. Thus we sought to determine if LPS directly induces accelerated tumour cell ECM adhesion and invasion through activation of the u-PA system and to elucidate the cellular pathways involved. Human colorectal tumour cell lines were stimulated with LPS. u-PA concentration, u-PA activity, active u-PA, surface urokinase plasminogen activator receptor (u-PAR) and TLR-4 expression were assessed by ELISA, colorimetric assay, western blot analysis and flow cytometry respectively. In vitro tumour cell vitronectin adhesion and ECM invasion were analysed by vitronectin adhesion assay and ECM invasion chambers. u-PA and u-PAR function was inhibited with anti u-PA antibodies or the selective u-PA inhibitors amiloride or WXC-340, TLR-4 by TLR-4-blocking antibodies and NF-kappaB by the selective NF-kappaB inhibitor SN-50. LPS upregulates u-PA and u-PAR in a dose-dependent manner, enhancing in vitro tumour cell vitronectin adhesion and ECM invasion by >40% (P<0.01). These effects were ameliorated by u-PA and u-PAR inhibition. LPS activates NF-kappaB through TLR-4. TLR-4 and NF-kappaB inhibition ameliorated LPS-enhanced u-PA and u-PAR expression, tumour cell vitronectin adhesion and ECM invasion. LPS promotes tumour cell ECM adhesion and invasion through activation of the u-PA system in a TLR-4- and NF-kappaB-dependent manner.

Figure 1

(A) LPS-stimulated tumour cells release u-PA in a dose-dependent manner. Following stimulation with various concentrations of LPS for 24 h, u-PA concentration in SW480 and SW620 cell culture supernatants was measured by ELISA as described in Materials and Methods. Data are expressed as the mean±s.d. (ng mg⁻¹ of protein) of three separate experiments conducted in triplicate. The statistical significance was compared with the cells incubated with culture medium (^* for 0.1 μg ml⁻¹, P<0.05) and for cells incubated with 0.1 μg LPS (^**P⩽0.05 for 1 μg LPS compared to 0.1 μg LPS). (B) LPS statistically increases u-PA release within 12 h of stimulation. Following stimulation with 0.1 μg ml⁻¹ LPS for various time periods, u-PA concentration in SW480 and SW620 cell culture supernatants was measured by ELISA as described in Materials and Methods. Data are expressed as the mean±s.d. (ng per mg of protein) of three separate experiments conducted in triplicate. The statistical significance was compared with the cells incubated with culture medium (^* for 0.1 μg ml⁻¹, ^** for 1 μg ml⁻¹, P<0.05). (C) LPS enhances u-PA activity in SW480 and SW620 cell supernatants. SW480 and SW620 cells were stimulated with various concentrations of LPS for 24 h in the presence or absence of the selective u-PA inhibitors amiloride (10 mg ml⁻¹) and WX-340 (1.0 mg ml⁻¹). The u-PA activity in culture supernatants was measured by colorimetric analysis as described in Materials and Methods. Data are expressed as the mean±s.d. (IU per mg of protein) from three separate experiments conducted in triplicate. The statistical significance was compared with the cells incubated with culture medium alone (^* for 0.1 μg ml⁻¹, ^** for 1 μg ml⁻¹, P<0.05). (D) LPS increases active u-PA expression in SW480 and SW620 cell supernatants. Following stimulation with LPS for 24 h, tumour cell culture supernatants were processed and concentrated as described in Materials and methods and western blot analysis was performed to detect active u-PA expression. Lane 1, u-PA positive control; lanes 2 and 3, CACO2 cells stimulated with 0 and 0.1 μg ml⁻¹ LPS respectively; lanes 4 and 5, SW480 cells stimulated with 0 and 0.1 μg ml⁻¹ LPS respectively; lanes 6 and 7, SW620 cells stimulated with 0 and 0.1 μg ml⁻¹ LPS respectively.

Figure 2

(A) SW480, SW620 and CACO2 cells constitutively express cell-surface u-PAR. (i) SW480, (ii) SW620 and (iii) CACO2 cells were analysed by FACScan analysis as described in Materials and Methods. Filled histograms represent isotype-matched mAb that served as a negative control; open histograms represent anti-u-PAR mAb. Shown are data from one representative experiment from three independent assays. (B) LPS enhances total cellular u-PAR expression. Following stimulation with various concentrations of LPS for 24 h, the u-PAR concentration in SW480 and SW620 cell lysates was measured by ELISA. Data are expressed as the mean±s.d. (ng per mg of protein) of six separate experiments conducted in triplicate. The statistical significance was compared with cells incubated with culture medium alone (^* for 0.1 μg ml⁻¹, P<0.05) and 0.1 μg LPS (^**P=<0.05 for 1 μg LPS compared to 0.1 μg LPS) and between SW480 and SW620 cells (^** for SW480 cells, P<0.05). (C) LPS enhances cell-surface u-PAR expression. Surface expression of u-PAR on SW480 (i) and SW620 (ii) cells was determined by FACScan analysis after stimulation with LPS for 24 h. Filled histograms represent 0 μg ml⁻¹ LPS+u-PAR mAb; open histograms represent 0.1 μg ml⁻¹ LPS+u-PAR mAb. Shown are data from one representative experiment with tabulated MFI (±s.e.m.) from three independent assays.

Figure 3

(A) LPS enhances SW480 and SW620 attachment to the extracellular matrix (ECM) protein vitronectin. After the cells were incubated with culture medium, 0.1 or 1 μg ml⁻¹ LPS for 24 h, the adherence of SW480 and SW620 cells to vitronectin was assessed as described in Materials and Methods. Data are expressed as the mean±s.d. from four separate experiments conducted in triplicate. Statistical significance was compared with the cells incubated with culture medium alone (^* for 0.1 μg ml⁻¹, P<0.05). (B) LPS enhances SW480 and SW620 ECM invasion. After the cells were incubated with culture medium, 0.1 or 1 μg ml⁻¹ LPS for 24 h, the ECM invasion of SW480 and SW620 cells was assessed as described in Materials and Methods. Data are expressed as the mean±s.d. from four separate experiments conducted in triplicate. Statistical significance was compared with the cells incubated with culture medium alone (^* for 0.1 μg ml⁻¹, P<0.05).

Figure 4

Effect of u-PA and u-PAR blockade on LPS-induced tumour cell adhesion and invasion. SW480 and SW620 cells were incubated with either isotype-matched control mAb (2.5 μg ml⁻¹) or u-PAR function-blocking mAb (2.5 μg ml⁻¹) or the selective u-PA inhibitors amiloride (10 μg ml⁻¹) and WXC-340 (1 μg ml⁻¹) for 1 h before 0.1 μg ml⁻¹ LPS stimulation. Tumour cell attachment to vitronectin (A) and ECM invasion (B and C) were assessed as described in Materials and Methods. Results are expressed as the mean±s.d. from three separate experiments conducted in triplicate. Statistical significance was compared with cells incubated with either culture medium plus control mAb (^*P<0.05) or LPS plus control mAb (^**P<0.05).

Figure 5

(A) SW480, SW620 and CACO2 constitutively express cell-surface TLR-4. (i) SW480, (ii) SW620, (iii) CACO2 and (iv) THP-1 (positive control) cells were analysed by flow cytometry using direct immunofluorescent staining as described in Materials and Methods. Filled histograms representing isotype-matched mAbs served as a negative control; open histograms represent anti-TLR-4 mAb. Shown are data from one representative experiment from three independent assays. (B, i–iii) Inhibition or deficiency of TLR-4 reduces LPS-enhanced u-PA activity. Following pre-incubation with 20 mg ml⁻¹ anti-TLR-4 function-blocking antibody or matched isotype control, cell supernatant u-PA activity were analysed by colorimetric analysis for SW480 (i), SW620 (ii) and CACO2 (iii) cell lines. Data are expressed as the mean±s.d. from six separate experiments conducted in triplicate. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (C, i–iii) Inhibition or deficiency of TLR-4 reduces LPS-enhanced surface u-PAR expression. Following pre-incubation with 20 mg ml⁻¹ anti-TLR-4 function-blocking antibody or matched isotype control, surface u-PAR expression was analysed by flow cytometry for SW480 (i), SW620 (ii) and CACO2 (iii) cell lines. Filled histograms represent 0 μg ml⁻¹ LPS+TLR-4 mAb; open histograms: red line represents 0.1 μg ml⁻¹ LPS+TLR-4 mAb, black line 0.1 μg ml⁻¹ LPS+cont mAb and blue line 10 mg ml⁻¹ butyrate. Shown are data from one representative experiment with tabulated MFI (±s.e.m.) from six independent assays. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (D) Inhibition or deficiency of TLR-4 reduces LPS-enhanced tumour cell vitronectin adhesion. Following pre-incubation with 20 μg ml⁻¹ anti-TLR-4 function-blocking antibody or matched isotype control, tumour cells were stimulated with 0.1 μg ml⁻¹ of LPS for 24 h and vitronectin adhesion assessed as described in Materials and Methods. Results are expressed as the mean±s.d. from four separate experiments, conducted in triplicate. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (E) Inhibition or deficiency of TLR-4 leads to reduced LPS-stimulated tumour cell extracellular matrix (ECM) invasion. Following pre-incubation with 20 μg ml⁻¹ anti-TLR-4 function-blocking antibody or matched isotype control, tumour cells were stimulated with 0.1 μg ml⁻¹ LPS for 24 h ECM invasion assessed as described in Materials and Methods. Results are expressed as the mean±s.d. from four separate experiments, conducted in triplicate. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05).

Figure 6

(A) LPS stimulation of SW480 and SW620 cells increased NF-κB activity. SW480 (i), SW620 (ii) and CACO2 (iii) cell lines were stimulated with 0.1 μg ml⁻¹ for 30 min cell after incubation with TLR-4 mAb, control mAb or medium alone. Cell and nuclear lysates were obtained and NF-κB activity assessed by ELISA as described in Materials and Methods. Data are expressed as the mean±s.d. of three separate experiments conducted in triplicate. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (B, i–iii) NF-κB inhibition impairs LPS-enhanced u-PA activity. Following pre-incubation with 100 mg ml⁻¹ SN-50 or SN-50M and subsequent stimulation with 0.1 μg ml⁻¹ LPS, cell supernatant u-PA activity for SW480 (i), SW620 (ii) and CACO2 (iii) cells was analysed by colorimetric analysis as described in Materials and Methods. Data are expressed as the mean±s.d. of six separate experiments conducted in triplicate. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (C) NF-κB inhibition impairs LPS-enhanced u-PAR surface expression. Following pre-incubation with 100 mg ml⁻¹ SN-50 or MSN-50 and subsequent stimulation with 0.1 μg ml⁻¹ LPS, surface u-PAR was analysed by flow cytometry as described in Materials and Methods for SW480 (i), SW620 (ii) and CACO2 (iii) cell lines. Filled histograms represent 0 μg ml⁻¹ LPS+SN-50M; open histograms: red lines represent 0.1 μg ml⁻¹ LPS+SN-50, black lines 0.1 μg ml⁻¹ LPS+SN-50M and green lines 10 mg ml⁻¹ butyrate. Shown are data from one representative experiment with tabulated MFI (±s.e.m.) from six independent assays. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (D) NF-κB inhibition impairs LPS-enhanced tumour cell vitronectin adhesion. Following pre-incubation with 100 mg ml⁻¹ SN-50 or SN-50M and subsequent stimulation with 0.1 μg ml⁻¹ LPS tumour cell attachment to vitronectin was assessed as described in Materials and Methods. Results are expressed as the mean±s.d. from four separate experiments, conducted in triplicate. Statistical significance was compared with cells incubated with culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (E) NF-κB inhibition impairs LPS-enhanced tumour cell ECM invasion. Following pre-incubation with 100 mg ml⁻¹ SN-50 or SN-50M and subsequent stimulation with 0.1 μg ml⁻¹ LPS, tumour cell ECM invasion was assessed as described in Materials and Methods. Results are expressed as the mean±s.d. from four separate experiments, conducted in triplicate. Statistical significance was compared with cells incubated with culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05).

Figure 7

(A) Protein synthesis inhibition impairs LPS-enhanced cell supernatant u-PA concentration. Following co-incubation with 10 μg ml⁻¹ cycloheximide and 0.1 μg ml⁻¹ LPS, cell supernatant u-PA concentration for SW480 (i), SW620 (ii) and CACO2 (iii) cells was analysed by ELISA as described in Materials and Methods. Data are expressed as the mean±s.d. of six separate experiments conducted in triplicate. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (B) Protein synthesis inhibition impairs LPS-enhanced u-PA activity. Following co-incubation with 10 μg ml⁻¹ cycloheximide and 0.1 μg ml⁻¹ LPS, cell supernatant u-PA activity for SW480 (i), SW620 (ii) and CACO2 (iii) cells was analysed by colorimetric analysis as described in Materials and Methods. Data are expressed as the mean±s.d. of six separate experiments conducted in triplicate. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (C) Protein synthesis inhibition impairs LPS-enhanced total u-PAR expression. Following co-incubation with 10 μg ml⁻¹ cycloheximide and 0.1 μg ml⁻¹ LPS, total cellular u-PAR expression was determined by ELISA as described in Materials and Methods. Data are expressed as the mean±s.d. of six separate experiments conducted in triplicate. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (D) Protein synthesis inhibition impairs LPS-enhanced surface u-PAR expression. Following co-incubation with 10 μg ml⁻¹ cycloheximide and 0.1 μg ml⁻¹ LPS, cell-surface u-PAR was analysed by flow cytometry as described in Materials and Methods for SW480 (i), SW620 (ii) and CACO2 (iii) cell lines. Filled histograms represent 0 μg ml⁻¹; open histograms: red lines represent 0 μg ml⁻¹ LPS+10 μg ml⁻¹ cycloheximide, black lines 0.1 μg ml⁻¹ LPS+10 μg ml⁻¹ cycloheximide and green lines 0.1 μg ml⁻¹ LPS. Shown are data from one representative experiment with tabulated MFI (±s.e.m.) from six independent assays. Statistical significance was compared with cells incubated in either culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (E) Protein synthesis inhibition impairs LPS-enhanced tumour cell vitronectin adhesion. Following co-incubation with 10 μg ml⁻¹ cycloheximide and 0.1 μg ml⁻¹ LPS, tumour cell attachment to vitronectin was assessed as described in Materials and Methods. Results are expressed as the mean±s.d. from four separate experiments, conducted in triplicate. Statistical significance was compared with cells incubated with culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05). (F) Protein synthesis inhibition impairs LPS-enhanced tumour cell ECM invasion. Following co-incubation with 10 μg ml⁻¹ cycloheximide and 0.1 μg ml⁻¹ LPS, tumour cell ECM invasion was assessed as described in Materials and Methods. Results are expressed as the mean±s.d. from four separate experiments, conducted in triplicate. Statistical significance was compared with cells incubated with culture medium alone (^*P<0.05) or 0.1 μg ml⁻¹ LPS (^**P<0.05).

Figure 8

Binding of LPS to TLR-4 leads to NF-κB activation, increased u-PA activity and u-PAR expression and ultimately enhanced tumour cell vitronectin adhesion and tumour cell extracellular matrix invasion.