Complete artificial saliva alters expression of proinflammatory cytokines in human dermal fibroblasts

Abstract

Complete artificial saliva (CAS) is a saliva substitute often used as a vehicle for test articles, including smokeless tobacco products. In the course of a study employing normal adult human dermal fibroblasts (HDFa) as a model in vitro, we discovered that CAS as a vehicle introduced a significant change in the expression of proinflammatory cytokines. To determine the effects of CAS on gene expression, real-time quantitative reverse-transcriptase PCR gene array analysis was used. Results indicate that robust changes in the expression of the proinflammatory cytokine interleukin 8 (IL8) and the vascular cell adhesion molecule 1 (VCAM1) occur within 5h of exposure to CAS. To determine whether CAS also alters cytokine release into the culture media, cytometric bead array assays for human inflammatory cytokines were performed. Analysis shows that CAS induced the release of IL8 and IL6. This study focused on determining which components in CAS were responsible for the proinflammatory response in HDFa. The following components were investigated: α-amylase, lysozyme, acid phosphatase, and urea. Results demonstrated that enzymatically active α-amylase induced gene expression for proinflammatory cytokines IL8, IL6, tumor necrosis factor-α, and IL1α and for VCAM1. Therefore, it is important to carefully evaluate the "vehicle effects" of CAS and its components in in vitro toxicology research.

Keywords: human dermal fibroblasts; interleukin 8; tumor necrosis factor-α; vascular adhesion molecule 1; α-amylase..

Fig. 1.

Changes in gene expression in HDFa in response to AS measured using quantitative real-time PCR (qRT-PCR). Data indicate fold change relative to vehicle, n = 3. Error bars indicate SEM. Asterisks indicate significant difference from cells exposed to CAS for 5h (*p < 0.05; ***p < 0.001). VEH = vehicle; -U = AS without urea; -E = AS without α-amylase, lysozyme, and acid phosphatase; -A = AS without α-amylase; -L = AS without lysozyme; -AP = AS without acid phosphatase.

Fig. 2.

Changes in proinflammatory cytokine release from HDFa into culture media in response to AS measured using CBA assays. Data indicate concentrations (pg/ml) of cytokines present in the culture media, n = 3. Error bars indicate SEM. Asterisks indicate significant difference from cells exposed to CAS for 5h (**p < 0.01). VEH = vehicle; -U = AS without urea; -E = AS without α-amylase, lysozyme, and acid phosphatase; -A = AS without α-amylase; -L = AS without lysozyme; -AP = AS without acid phosphatase.

Fig. 3.

Concentration of α-amylase and gene expression for IL8, TNF-α, VCAM1, and IL6. HDFa were treated with samples of α-amylase at final concentrations of 0.835, 2.505, 8.35, 25.05, and 83.5U/ml or vehicle (FBM without supplements). Cells were collected at 5h posttreatment. Effects of α-amylase on the expression of IL8, IL6, TNF-α, and VCAM1 were determined using qRT-PCR. A one-site hyperbola was generated to graphically represent gene expression relative to concentration of α-amylase (GraphPad). Nonparametric correlations of relative fold changes in gene expression and α-amylase enzymatic activity were performed using linear regression and Spearman’s r (GraphPad).

Fig. 4.

Enzymatic activity of heat-inactivated α-amylase.

Fig. 5.

Gene expression in HDFa in response to heat-inactivated α-amylase measured using qRT-PCR. HDFa were treated with samples of heat-inactivated α-amylase at final concentrations of 8.35U/ml. MilliQ-treated H₂O or FBM without supplements was used as a vehicle. Cells were collected at 5h posttreatment. Effects of heat-inactivated α-amylase on the expression of IL1α, IL6, IL8, TNF-α, and VCAM1 were determined using qRT-PCR. Nonparametric correlations of relative fold changes in gene expression and α-amylase enzymatic activity were performed using linear regression and Spearman’s r (GraphPad). Data indicate fold change relative to vehicle, n = 3–6. Error bars indicate SEM. Asterisks indicate significant difference from cells exposed to nonheat-inactivated α-amylase for 5h (*p < 0.05; **p < 0.01; ***p < 0.001). VEH = vehicle.

Fig. 6.

Correlation of gene expression in HDFa with α-amylase enzymatic activity. Nonparametric correlations are significant for IL8, TNF-α, VCAM1, IL1α, and IL6 (*p < 0.05; **p < 0.01).

Fig. 7.

Effects of α-amylase and supplemental growth factors on gene expression in HDFa for IL8, TNF-α, VCAM1, and IL6. At the time of treatment, cell culture media were replaced with serum-free defined media as follows: -G: FBM, GlutaMAX I, Pen-Strep, and phenol red; +G: the same with supplemental growth factors; -A: without α-amylase; +A: treated with α-amylase (8.35U/ml). Cells were collected at 5h posttreatment, and gene expression was determined by qRT-PCR. Error bars indicate SEM (n = 3 for all groups). Asterisks indicate a statistically significant difference between the groups indicated (*p < 0.05; ***p < 0.001). Statistical analyses not indicated graphically: For IL6 gene expression, significant differences were observed between −G/−A and +G/−A (***p < 0.001) and between −G/+A and +G/+A (*p < 0.05). For TNF-α, significant differences were observed between −G/−A and +G/−A and between −G/+A and +G/+A (*p < 0.05). No significant differences due to supplemental growth factors were observed for expression of IL8 or VCAM1. Between −G/−A and +G/−A, p values for IL8 and VCAM1 were 0.290 and 0.276, respectively.