Borrelia burgdorferi-mediated induction of miR146a-5p fine tunes the inflammatory response in human dermal fibroblasts

Abstract

Colonization of a localized area of human skin by Borrelia burgdorferi after a bite from an infected tick is the first step in the development of Lyme disease. The initial interaction between the pathogen and the human host cells is suggested to impact later outcomes of the infection. MicroRNAs (miRNAs) are well known to be important regulators of host inflammatory and immune responses. While miRNAs have been shown to play a role in the inflammatory response to B. burgdorferi at late stages of infection in the joints, the contributions of miRNAs to early B. burgdorferi infection have yet to be explored. To address this knowledge gap, we used the published host transcriptional responses to B. burgdorferi in erythema migrans skin lesions of early Lyme disease patients and a human dermal fibroblasts (HDFs)/B. burgdorferi co-culture model to predict putative upstream regulator miRNAs. This analysis predicted a role for miR146a-5p in both, B. burgdorferi-infected skin and -stimulated HDFs. miR146a-5p was confirmed to be significantly upregulated in HDF stimulated with B. burgdorferi for 24 hours compared to uninfected control cells. Furthermore, manipulation of miR146a-5p expression (overexpression or inhibition) altered the B. burgdorferi driven inflammatory profile of HDF cells. Our results suggest that miR146a-5p is an important upstream regulator of the transcriptional and immune early response to early B. burgdorferi infection.

Fig 1. Upstream miRNA regulators predicted to influence the inflammatory response to early B. burgdorferi infection.

The lists of predicted miRNAs and the overrepresented DE gene targets were filtered; >1 interaction, P<0.05, FDR<10%. (A) Upstream miRNA regulation of differentially expressed (DE) genes in erythema migrans skin lesions of early Lyme disease patients. DE genes (circles) reported by Marques et al. [22] and predicted upstream miRNA regulators (squares) that had an overrepresented number of DE gene targets. Red and blue circles represent the reported upregulated and downregulated genes, respectively. Light red and light blue squares represent predicted upregulated and downregulated miRNAs, respectively. Grey squares represent miRNAs that have an overrepresented number of gene targets in both the upregulated and downregulated sets of DE genes. (B) Upstream miRNA regulation of DE genes from primary human dermal fibroblasts (HDFs) in response to stimulation with B. burgdorferi sensu lato for 24 hours. Network representing strong validated interactions for miRNAs that had an overrepresented number of targets within the upregulated genes (red circles) reported by Meddeb et al. [23]. Light blue squares represent presumably down-regulated miRNAs. (C) Venn Diagram outlines common and unique predicted upstream miRNA regulators for the two sets of DE genes reported for skin lesions of early Lyme disease patients (skin) and B. burgdorferi stimulated HDFs (fibroblasts).

Fig 2. B. burgdorferi stimulation of HDFs resulted in increased expression of miR146a-5p and its targets of regulation.

(A+C) RT-qPCR expression analysis of inflammatory genes and known targets of miR146a-5p regulation, normalized to RPL2 gene expression. (B) RT-qPCR expression analysis of miR146a-5p, normalized to miR191-5p. Data are presented as the relative log fold change in mRNA or miRNA levels in HDFs co-incubated with (+Bb) or without (-Bb) B. burgdorferi for 24 hours. Data represent the average of biological triplicates ± standard deviation. Statistical significance was determined by unpaired t test (GraphPad, Prism). ns, not significant; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Fig 3. The kinetics of expression miR146a-5p and miR146a-5p-related genes in B. burgdorferi stimulated HDFs are consistent with a model of a NF-ϰB-dependent hierarchical regulatory network.

(A) RT-qPCR expression analysis of inflammatory genes and miR146a-5p, normalized to RPL2 and to miR-191-5p gene expression, respectively. RT-qPCR data are presented as the relative log fold change in mRNA or miRNA levels in HDFs co-incubated with B. burgdorferi (+Bb) for 2 hours, (2 h), 6 hours (6 h) or 24 hours (24 h) compared with HDFs alone (-Bb) at each time point. (B) Immunoblot analysis of NF-ϰB and actin levels produced by HDFs co-incubated with B. burgdorferi (+Bb) compared with HDFs alone (-Bb) at each time point. Molecular weights are shown in kilodaltons (kDa). Representative data of three biological replicates are shown. Quantification of the NF-ϰB protein levels in each sample normalized to actin and relative to the NF-ϰB/actin signal intensity of HDFs alone (-Bb) at 2 hours. Data represent the average of biological triplicates ± standard deviation. (C) Quantification of the picograms per milliliter (pg/mL) of IL6 secreted by HDFs co-incubated with B. burgdorferi (+Bb) compared with HDFs alone (-Bb) at each time point. Data represent the average of biological triplicates ± standard deviation. (D) Immunoblot analysis of STAT1 and actin levels produced by HDFs co-incubated with B. burgdorferi (+Bb) compared with HDFs alone (-Bb) at each time point. Molecular weights are shown in kilodaltons (kDa). Representative data of three biological replicates are shown. Quantification of the STAT1 protein levels in each sample normalized to actin and relative to the STAT1/actin signal intensity of HDFs alone (-Bb) at 2 hours. Statistical significance was determined by Ordinary one-way ANOVA followed by the Welch and Brown-Forsythe ANOVA test for multiple comparisons (GraphPad, Prism). ns, not significant; *p< 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Fig 4. . Overexpression of miR146a-5p reduced the inflammatory response of B. burgdorferi stimulated HDFs.

HDF cells were transfected with the miR146a-5p mimic or the negative control (NC) mimic and stimulated with (+Bb) or without B. burgdorferi for 24 hours. (A) RT-qPCR expression analysis of miR146a-5p, normalized to miR191-5p. (B) RT-qPCR expression analysis of miR146a-6p target and related genes, normalized to RPL2 gene expression. RT-qPCR data are presented as the relative log fold change in miRNA or mRNA levels in cells transfected with the miR146a-5p mimic compared to cells transfected with the negative control mimic, in the absence or presence of B. burgdorferi. Data represent the average of biological triplicates ± standard deviation. Statistical significance was determined by Ordinary one-way ANOVA followed by the Welch and Brown-Forsythe ANOVA test for multiple comparisons (GraphPad, Prism). ns, not significant; *p< 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. (C) Immunoblot analysis of NF-ϰB and actin levels produced by the HDFs. Molecular weights are shown in kilodaltons (kDa). Representative data of three biological replicates are shown. Quantification of the NF-ϰB protein levels in each sample normalized to actin and relative to the NF-ϰB/actin signal intensity of HDFs alone treated with the negative control mimic (NC mimic). Data represent the average of biological triplicates ± standard deviation. (D) Quantification of the picograms per milliliter (pg/mL) of IL6 secreted by the HDFs. Data represent the average of biological triplicates ± standard deviation. Statistical significance was determined by unpaired t test (GraphPad, Prism). ****p < 0.0001. (E) Immunoblot analysis of STAT1 and actin levels produced by the HDFs. Molecular weights are shown in kilodaltons (kDa). Representative data of three biological replicates are shown. Quantification of the STAT1 protein levels in each sample normalized to actin and relative to the STAT1/actin signal intensity of HDFs alone treated with the negative control mimic (NC mimic). Data represent the average of biological triplicates ± standard deviation. (F) Immunoblot analysis of TRAF6 and actin levels produced by the HDFs. Molecular weights are shown in kilodaltons (kDa). Representative data of three biological replicates are shown. Quantification of the TRAF6 protein levels in each sample normalized to actin and relative to the TRAF6/actin signal intensity of HDFs alone treated with the negative control mimic (NC mimic). Data represent the average of biological triplicates ± standard deviation.

Fig 5. Inhibition of miR146a-5p resulted in increased expression of inflammation associated genes in B. burgdorferi stimulated HDFs.

HDF cells were transfected with the miR146a-5p specific inhibitor (miR146a-5p inh) or the negative control inhibitor (NC inh) and stimulated with (+Bb) or without B. burgdorferi for 24 hours. (A) RT-qPCR expression analysis of miR146a-6p target and related genes, normalized to RPL2 gene expression. RT-qPCR data are presented as the relative log fold change in mRNA levels in cells transfected with the miR146a-5p inhibitor compared to cells transfected with the negative control inhibitor, in the absence or presence of B. burgdorferi. Data represent the average of biological triplicates ± standard deviation. Statistical significance was determined by Ordinary one-way ANOVA followed by the Welch and Brown-Forsythe ANOVA test for multiple comparisons (GraphPad, Prism). ns, not significant; *p< 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. (B) Immunoblot analysis of NF-ϰB and actin levels produced by the HDFs. Molecular weights are shown in kilodaltons (kDa). Representative data of three biological replicates are shown. Quantification of the NF-ϰB protein levels in each sample normalized to actin and relative to the NF-ϰB/actin signal intensity of HDFs alone treated with the negative control inhibitor (NC inh). Data represent the average of biological triplicates ± standard deviation. (C) Quantification of the picograms per milliliter (pg/mL) of IL6 secreted by the HDFs. Data represent the average of biological triplicates ± standard deviation. Statistical significance was determined by unpaired t test (GraphPad, Prism). ****p < 0.0001, (D) Immunoblot analysis of STAT1 and actin levels produced by the HDFs. Molecular weights are shown in kilodaltons (kDa). Representative data of four biological replicates are shown. Quantification of the STAT1 protein levels in each sample normalized to actin and relative to the STAT1/actin signal intensity of HDFs alone treated with the negative control mimic (NC mimic). Data represent the average of four biological replicates ± standard deviation. (E) Immunoblot analysis of TRAF6 and actin levels produced by the HDFs. Molecular weights are shown in kilodaltons (kDa). Representative data of three biological replicates are shown. Quantification of the TRAF6 protein levels in each sample normalized to actin and relative to the TRAF6/actin signal intensity of HDFs alone treated with the negative control inhibitor (NC inh). Data represent the average of biological triplicates ± standard deviation.

Fig 6. Model of miR146a-5p-mediated modulation of inflammatory gene expression in B. burgdorferi-stimulated human dermal fibroblasts.

Stimulation of human dermal fibroblasts with B. burgdorferi results in increased expression of genes important for inflammation as well as the anti-inflammatory microRNA, miR146a-5p. Regulation by miR146a-5p reduces inflammatory gene expression levels perhaps by targeting mechanisms of NF-ϰB activation and/or the inflammatory gene transcripts directly.