TLR4 as a negative regulator of keratinocyte proliferation

Abstract

TLR4 is an innate immune receptor with expression in human skin, keratinocytes as well as squamous cell carcinoma (SCC) of the skin. In the present study we investigate the role of TLR4 as a negative regulator of keratinocyte proliferation. We present here that the expression of TLR4 increased with the differentiation of cultured keratinocytes in a passage-dependent manner or under calcium-rich conditions. Moreover, the down-regulation of TLR4 by specific knockdown increased the proliferation of HaCaT keratinocytes in vitro. In addition, subcutaneously injected HaCaT keratinocytes with shTLR4 formed growing tumors in nude mice. In contrast, we observed lower proliferation and increased migration in vitro of the SCC13 cell line stably overexpressing TLR4 in comparison to SCC13 TLR4 negative cells. In vivo, SCC13 TLR4-overexpressing tumors showed delayed growth in comparison to TLR4 negative tumors. The overexpression of TLR4 in SCC13 tumor cells was followed by phosphorylation of ERK1/2 and JNK and increased expression of ATF3. In gene expression arrays, the overexpression of TLR4 in tumor cells correlated with gene expression of ATF-3, IL-6, CDH13, CXCL-1 and TFPI. In summary, TLR4 negatively regulates the proliferation of keratinocytes and its overexpression reduces tumor growth of SCC cells.

Fig 1. Expression of TLR4 in normal human and SCC skin.

The expression level of TLR4 was analyzed on formalin-fixed paraffin-embedded skin samples and tissue microarray samples by immunochistochemistry using specific anti human TLR4 antibody (HTA125, dilution 1:100). The expression was tested in 225 SCC patient samples (moderately or well differentiated) and in 23 samples from normal skin. (A) Positive staining for TLR4 in normal epidermis. (B) Positive staining for TLR4 in moderately differentiated SCC. (C) Positive staining for TLR4 in well differentiated SCC. TLR4 positive keratinocytes show red intensity staining. TLR4 expression is visible in the membrane regions and perinuclear (cytoplasmic) regions and pointed by arrows. (D) Positive staining for TLR4 found in normal skin and maintained in SCC. (E) Positive staining for TLR4 is found in SCC in different stages of differentiation (moderately and well differentiated SCC). (F) Positive correlation between TLR4 and IRF6 expression (staining intensity) in moderately versus well differentiated SCC epidermis. The intensity of staining was analyzed by arbitrary evaluation and presented as arbitrary units respectively as weak, middle and strong staining. The expression level (intensity) of TLR4, IRF6 and the IRF6/TLR4 correlation were evaluated using t-test and Anova. P-values below 0.0001 and 0.05 (p<0.05; p<0.0001) were considered as significant.

Fig 2. TLR4 expression increases with the differentiation of normal and SCC keratinocytes in vitro.

(A) Primary keratinocytes were grown in duplicates (n = 2) till full confluency at 96h and the expression of TLR4 and keratinocyte differentiation markers (involucrin and filaggrin) was compared that of low passage (growing) keratinocytes by western blot using specific antibodies. The results were reproducible in two independent experiments. (B) Primary and SCC-derived keratinocytes were grown in duplicates (n = 2) in the presence of Ca²⁺ for different time intervals: 24h, 48h and 96 h. The group “no calcium” was used as positive control for differentiation at 96 hours, the time point of maximal differentiation induced by cell-cell contact. The expression of TLR4 and involucrin was analyzed by western blot using specific antibodies. The results were reproducible in two independent experiments.

Fig 3. Knockdown of TLR4 induces proliferation of HaCaT keratinocytes.

(A) Growth of sh control and shTLR4 cells under puromycine selection visualized by light microscopy. (B) Knockdown of TLR4 induces cellular proliferation. The alteration of the proliferation was analyzed by BrDU proliferation assay. The significance was evaluated by t-test analysis on duplicates (n = 2) from three independent experiments (p*<0.05; p****<0.0001). As a control cells transfected with a control sh were used. (C) Expression of TLR4 on transcriptional level after TLR4 knockdown. The differential TLR4 expression in the control and shTLR4 cells was evaluated by qPCR using specific TLR4 primers. (D) Expression of TLR4 on protein level after TLR4 knockdown. The differential TLR4 expression was evaluated by conventional western blot analysis using a monoclonal anti TLR4 antibody (HTA 125).

Fig 4. Growth of shTLR4 cells in nude mice.

(A) Knockdown of TLR4 induces growth of HaCaT cells in nude mice (injected 4X10³ cells/mouse (mice, n = 6/group); p = 0.0002, t-test). The graphic represents the tumor growth (tumor volume) at different time intervals post injection. (B) Expression of TLR4 in the control and sh TLR4 tumors on end day. The differential TLR4 expression was evaluated on transcriptional level by qPCR using specific anti-human TLR4 primers. The tumor RNA and cDNA synthesis were performed using TRIzol reagent and RT respectively.

Fig 5. Overexpression of TLR4 in SCC13 tumor keratinocytes.

(A) Visualization of SCC13-TLR4-GFP cells by fluorescent microscopy. SCC13 cells were stably transfected with pUNO-TLR4-GFP and control pUNO-mcs-expressing vectors. TLR4-GFP positive cells revealed nice green fluorescent staining in contrast to control cells. (B) SCC13-TLR4 cells show lower proliferation and higher migratory activity. The proliferation and migration of TLR overexpressing cells were investigated by BrDU proliferation assay and scratch assay and compared to cells expressing control vector only. The alterations in proliferation and migration were statistically significant (n = 4, two independent experiments, p*** = 0.0006, p****<0.0001, t-test).

Fig 6. Delayed growth of SCC13-TLR4 overexpressing tumors.

(A) Delayed growth of SCC13-TLR4 tumors in nude mice. The graph represents the tumor volume of the TLR4 and control tumors at day 15 after injection (end day, n = 22; p = 0.0192, Wilcoxon matched-pairs signed rank test). (B) Expression of TLR4 in SCC13-TLR4 and SCC13-pUNO control tumors at end day. The expression level of TLR4 in control and TLR4+ tumors was analyzed by qPCR using specific anti human TLR4 primers. The TLR4 overexpressing tumors (n = 22) retained high levels of TLR4 expression. The graph represents an example of TLR4 expression in animal group 1^st (n = 5).

Fig 7. Immunochistochemical analysis of the SCC13 tumors.

(A) Hematoxilyn staining (HE) for determining the tumor architecture. The tumor region and the tumor surrounding tissue (immune infiltrate) are marked with black and red arrows respectively. The staining was performed in all tumors and using the standard procedure for HE staining. (B) Staining for ki67 proliferation marker in all SCC13-TLR4 and control tumors. The cells expressing ki67 showed a positive red staining and are pointed with arrows. The intensity of the staining varies depending on the ki67 expression level. (C) Percentage of Ki67 positive cells in SCC13-TLR4 tumors versus control tumors. The differences in the expression level based on the immunochistochemical staining were quantified as [%] positive cell/area stained tissue (p = 0.0116, t-test, n = 12 sample group).

Fig 8. Immunochistochemical analysis of the SCC13 tumor vascularization, immune infiltration and metastasis.

(A) Staining for CD31 and CD68 expression. The red positive signal corresponds to the expression of CD31 in the endothelial cells and the blood vessels respectively (black arrows). CD68 positive cells (macrophages) are stained in red and are present in the surrounding tumor infiltrate and the peripheral tumor area (black arrows). (B) Hematoxilyn staining for determining the organ-specific cellular architecture. The stainings were performed using specific antibodies against human Ki67, CD31 and CD68. All images are presented at low (5X) and highest (20X) magnification.

Fig 9. Functional relevance of TLR4 in SCC13-TLR4 cells.

(A) Altered expression profile of pERK, pJNK and ATF3 in SCC13-TLR4 cells. The functionality of TLR4 was investigated upon LPS treatment for different time points and the differential expression/ phosphorylation profile was analyzed by western blot using specific antibodies. The phosphorylation of ERK and JNK was induced at 15–30 min after LPS treatment in both of the cell lines. SCC13-TLR4 cells showed increased phosphorylation of ERK and JNK in LPS-free conditions in comparison to SCC13-pUNO cells. SCC13-TLR4 cells show an increased ATF3 expression in LPS-free conditions in comparison to SCC13-pUNO control cells. The ATF3 expression was induced at 15 min after LPS stimulation in both of the cell lines. The results represent an example of two independent LPS treatments. (B) Alteration of IL-6 secretion in SCC13-TLR4 cells. SCC13-TLR4 and control SCC13-pUNO- cells were seeded in duplicates (n = 2) and induced by LPS at time points 15, 30, 45 and 60 min. Six and 24h after LPS stimulation cell supernatants were collected and analysed for IL-6 expression using ELISA and compared to untreated cells. LPS untreated SCC13-TLR4 cells show lower levels of IL6 in the cellular supernatant in comparison to control SCC13-TLR4 cells.

Fig 10. Altered gene expression in SCC13-TLR4 versus SCC13-pUNO control cells revealed by microarray analysis.

Gene expression array: SCC13-TLR4 and control SCC13-pUNO- cells were seeded in tetraplicates (n = 4) and induced by LPS for 24 hours. The SurePrint G3 Human Gene Expression 8x60K (Agilent) with 50599 biological features was used to analyze the samples. Differentially expressed genes were selected as significant if the absolute log-fold change was more than 2 and FDR adjusted p-value was less than 0.05. (A) Clustering of significant probes and GO Analysis. Red GO categories are overrepresented among the significantly upregulated genes. Blue GO categories are overrepresented among the significantly downregulated genes. Black GO categories are overrepresented among all signifcantly regulated genes. The genes of interest with validated expression are presented in the table. (B) Positive correlation between TLR4 and ATF3 expression in moderately versus well differentiated SCC epidermis. The intensity of staining was analyzed by arbitrary evaluation and presented as arbitrary units respectively as weak, middle and strong staining. The expression level (intensity) of TLR4, ATF3, and the ATF3/TLR4 correlation were evaluated using t-test and Anova. P-values below 0.0001 and 0.05 (p<0.05; p<0.0001) were considered as significant.