Peripheral inflammation is associated with remote global gene expression changes in the brain

Abstract

Background: Although the central nervous system (CNS) was once considered an immunologically privileged site, in recent years it has become increasingly evident that cross talk between the immune system and the CNS does occur. As a result, patients with chronic inflammatory diseases, such as rheumatoid arthritis, inflammatory bowel disease or psoriasis, are often further burdened with neuropsychiatric symptoms, such as depression, anxiety and fatigue. Despite the recent advances in our understanding of neuroimmune communication pathways, the precise effect of peripheral immune activation on neural circuitry remains unclear. Utilizing transcriptomics in a well-characterized murine model of systemic inflammation, we have started to investigate the molecular mechanisms by which inflammation originating in the periphery can induce transcriptional modulation in the brain.

Methods: Several different systemic and tissue-specific models of peripheral toll-like-receptor-(TLR)-driven (lipopolysaccharide (LPS), lipoteichoic acid and Imiquimod) and sterile (tumour necrosis factor (TNF) and 12-O-tetradecanoylphorbol-13-acetate (TPA)) inflammation were induced in C57BL/6 mice. Whole brain transcriptional profiles were assessed and compared 48 hours after intraperitoneal injection of lipopolysaccharide or vehicle, using Affymetrix GeneChip microarrays. Target gene induction, identified by microarray analysis, was validated independently using qPCR. Expression of the same panel of target genes was then investigated in a number of sterile and other TLR-dependent models of peripheral inflammation.

Results: Microarray analysis of whole brains collected 48 hr after LPS challenge revealed increased transcription of a range of interferon-stimulated genes (ISGs) in the brain. In addition to acute LPS challenge, ISGs were induced in the brain following both chronic LPS-induced systemic inflammation and Imiquimod-induced skin inflammation. Unique to the brain, this transcriptional response is indicative of peripherally triggered, interferon-mediated CNS inflammation. Similar models of sterile inflammation and lipoteichoic-acid-induced systemic inflammation did not share the capacity to trigger ISG induction in the brain.

Conclusions: These data highlight ISG induction in the brain as being a consequence of a TLR-induced type I interferon response. As considerable evidence links type I interferons to psychiatric disorders, we hypothesize that interferon production in the brain could represent an important mechanism, linking peripheral TLR-induced inflammation with behavioural changes.

Figure 1

Systemic LPS injection results in an inflammatory response in the periphery and brain. (A) Plasma concentrations of (i) IL-1β, (ii) IL-6 and (iii) TNFα, 6, 12 and 48 hours following injection with LPS (100 μg i.p.) or vehicle. Data represent the mean ± the standard error of the mean (SEM). Significance was determined using two-way ANOVA: **P ≤ 0.01, ***P ≤ 0.001; n = 4/group. (B) Top enriched gene ontology classifications determined using DAVID Bioinformatics Resources. (C) Top significantly altered pathways determined using Ingenuity Pathway Analysis software. (B,C) Significance was calculated using Fisher’s exact test. (D) Comparison of entities identified using Partek Genomics Suite and Genespring GX software packages as being differentially expressed by ≥ 2-fold.

Figure 2

ISG induction in the brain and blood of LPS-treated mice. (A) Relative expression of ISGs in the brain and PBLs, 48 hours after LPS injection (100 μg i.p.) compared with vehicle injection. (B) Relative expression of ISGs in the brain, 6, 12 and 48 hours after LPS injection compared with vehicle injection. Gene expression normalized to TBP. Fold change calculated by comparing the normalized copy number in each sample with the mean of the vehicle-injected controls. Data represent mean ± SEM. Statistical significance was determined using two-way ANOVA: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001; n = 4/group.

Figure 3

Response in the brain to systemic administration of TNF. Mice were injected with two doses of TNFα (1 μg i.v.) or vehicle at 0 and 24 hours and relative expression of interferon-stimulated genes (ISGs) in the brain and peripheral blood leucocytes (PBLs) was determined 48 hours after initial injection. Gene expression was normalized to TBP. Fold change was calculated by comparing the normalized copy number in each sample with the mean of the vehicle-injected control group. Data represent mean ± SEM.

Figure 4

ISG induction in the brain and PBL of mice treated daily with LPS. Mice were injected daily with LPS (50 μg i.p.) or vehicle. (A,B) Relative gene expression in the brains and PBLs of mice, 2, 5 and 7 days following initial injection. (A) Relative expression of cytokines; (i) Il1b, (ii) Il6 and (iii) Tnfa. (B) Relative expression of target ISGs; (i) Cxcl10, (ii) Ifit1, (iii) Irf7, (iv) Oasl2 and (v) Gbp4. Gene expression normalized to TBP. Fold change was calculated by comparing the normalized copy number in each sample with the mean of the vehicle-injected controls. Data represent mean ± SEM. Statistical significance was determined using two-way ANOVA: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001; n = 4/group.

Figure 5

TLR-induced signalling pathways. When activated, toll-like receptors (TLRs) either heterodimerize, such as TLR2 with either TLR1 or TLR6, or homodimerize to induce signalling pathway activation. Signalling is regulated by adaptor molecules. MyD88 is an adaptor used by all TLRs aside from TLR3. TLR7 recruits MyD88 directly, whereas TLR4 and TLR2 recruit MyD88 through bridging adapter TIRAP. MyD88 recruitment ultimately leads to the activation of nuclear transcription factor NFκB by releasing the p50 and p65 subunits from the NFκB inhibitor IκB. Activated NFκB then translocates to the nucleus to induce the expression of inflammatory cytokines. When activated by TLR7, MyD88 can also induce the phosphorylation and dimerization of interferon regulatory factor 7 (IRF7). This activation of IRF7 triggers the induction of IFNα. An alternative signalling pathway downstream of TLR4 involves the recruitment of adaptor molecule TRIF through bridging protein TRAM. This ultimately leads to activation of IRF3 and subsequent IFNβ induction (For review see [31]). Note that for simplicity the full complement of extracellular proteins involved in TLR for function are not represented in this diagram. Specifically, TLR4 requires MD-2, CD14 and LPS binding protein to bind LPS efficiently.

Figure 6

Response in the brain to systemic administration of LTA. Mice were injected with two doses of LTA (500 μg i.v.) or vehicle at 0 and 24 hours and relative expression of interferon-stimulated genes (ISGs) in the brain and peripheral blood leucocytes (PBLs) was determined 48 hours after initial injection. Gene expression was normalized to TBP. Fold change was calculated by comparing the normalized copy number in each sample with the mean of the vehicle-injected control group. Data represent mean ± SEM; n = 4/group.

Figure 7

Response in the brain to IMQ-induced skin inflammation. Mice were treated daily with IMQ or control cream (Vaseline). Samples taken 24 hours after the fifth application. (A) H & E stains of (i) IMQ-treated and (ii) Vaseline-treated skin. Scale bar: 100 μm. Arrow points to epidermal hyperplasia. (B) Plasma concentrations of (i) IL-1β, (ii) IL-6 and (iii) TNFα. (C) Relative expression of ISGs, normalized to TBP, in brain and PBLs of IMQ-treated compared with Vaseline-treated mice. Fold change was calculated by comparing the normalized copy number in each sample to the mean of the vehicle-injected control group. Data represent mean ± SEM; n = 5/group.

Figure 8

Response in the brain to TPA-induced skin inflammation. Mice were treated daily with five topical applications of TPA or acetone. Samples were taken 24 hours after fifth application. (A) H & E stains of (i) TPA-treated and (ii) acetone-treated skin. Scale bar: 100 μm. Arrow points to epidermal hyperplasia. (B) Plasma concentrations of (i) IL-1β, (ii) IL-6 and (iii) TNFα. (C) Relative expression of ISGs, normalized to TBP, in brain and PBLs of TPA-treated compared to acetone-treated mice. Fold change was calculated by comparing the normalized copy number in each sample to the mean of the vehicle-injected control group. Data represent mean ± SEM; n = 5/group.