Introduction

The immune system includes both innate immunity and adaptive immunity mechanisms, and their role is to identify and eliminate invading microbial pathogens [1, 2]. TLR signaling pathways are important for host defense, and aberrant TLR signaling will change immune homeostasis and result in immunopathological conditions such as endotoxin shock and autoimmune responses [3]. Therefore, TLR signaling pathways are supposed to be strictly and finely regulated [4, 5]. Type 1 diabetes (T1D) is generally considered to be caused by the interaction of the immune system with genetic susceptibility and environmental factors, such as viral infections [6,7,8]. Adaptive immunity has been confirmed to be involved in the pathogenesis of T1D. However, the role of innate immunity in T1D has only been investigated recently.

Epigenetic mechanisms can influence gene expression without changing the DNA sequence. DNA methylation, histone modification, and microRNAs are the three main epigenetic modifications. All of them are associated with transcriptional regulation and determination of the cellular transcriptome, thereby contributing to maintain normal cell function. Increasing evidence has shown that epigenetic mechanisms are involved in the regulation of TLRs. Since TLRs are involved in self vs. non-self-identification, they are thought to play an important role in the pathogenesis of immune disorders and autoimmune diseases [9]. Understanding the epigenetic mechanisms that regulate TLRs is thus important for the study of disease pathogenesis and the treatment of autoimmune diseases including T1D.

Toll-like receptors (TLRs) and TLR signaling pathways

Toll was first identified to play an important role in recognizing microorganisms and establishing the dorsal-ventral axis during the embryogenesis of Drosophila melanogaster in 1996 [10]. The homolog of Toll in humans is called the Toll-like receptor (TLR). There are 10 functional TLRs (TLR1–TLR10) that have been identified in humans. Since their discovery, TLRs have gained increasing attention for their role in the innate immune system as well as for their involvement in cross talk between the innate and adaptive systems (Fig. 1).

Fig. 1
figure 1

TLRs in innate and adaptive immunity. Pathogens, as well as some microRNAs, can be recognized by TLRs, and TLR signaling pathways will then be initiated, in turn upregulating the expression of co-stimulatory molecules (CD80/86), MHC-II, and inflammatory cytokines (such as IL-12). Recent studies have suggested that microRNAs serve as ligands of TLRs, such as TLR1, 7, and 8, which are shown in the figure. In addition, pathogens can be captured by dendritic cells and be processed and presented to naïve T cells. Together, these signals contribute to the development of antigen-specific adaptive immunity

Full size image

TLRs can be divided into five subfamilies based on their sequence, genomic structure, and function [11]. Although all TLRs are membrane proteins, TLR1, TLR2, TLR4, TLR5, TLR6, and TLR10 are localized on the plasma membrane. In contrast, TLR7, TLR8, and TLR9 are localized to intracellular vesicles, such as endosomes, lysosomes, and the ER, while TLR3 can be localized on both cell surfaces and endosomes [12]. TLRs are expressed in both innate and adaptive immune cells, such as dendritic cells, macrophages, neutrophils, B and T cells [13, 14], T regulatory cells [15], and on non-immune cells such as fibroblasts, epithelial cells, and islet beta cells [16]. However, the highest expression of TLRs is on tissues that are involved in innate immunity [17].

Different TLRs can recognize different pathogen-associated molecular patterns (PAMPs) or damage-associated molecular patterns (DAMPs) (Table 1). PAMPs are usually the pathogenic molecules produced by bacteria, fungi, parasites, and viruses, which can interact with cells of the innate immune system [1, 18]. DAMPs are the endogenous molecules released by necrotic cells as well as intracellular proteins, such as heat shock proteins [19]. TLRs are type 1 transmembrane glycoproteins that consist of an extracellular part of TLRs, with a leucine-rich repeat domain (LRR) that specifically recognizes PAMPs, and the intracellular domain of TLRs with a Toll/interleukin-1 receptor (TIR) domain that recruits downstream adaptor proteins [20, 21]. The activation of TLR signaling transduction pathways will result in a series of downstream signaling events that lead to the activation of transcription factors and ultimately cause the upregulation of various pro-inflammatory cytokines, chemokines, and interferons (IFNs) and activate the adaptive immune system [22, 23].

Table 1 Toll-like receptors and their ligands
Full size table

There are two TLR-mediated signaling pathways after TIR adaptor binding, namely the myeloid differentiation primary response 88 (MyD88)-dependent pathway and the TIR domain-containing adaptor-inducing IFN-β (TRIF)-dependent pathway [24]. The MyD88-dependent pathway is common to most TLRs. Upon ligand recognition, TLRs, except for TLR3, will recruit adaptor protein MyD88 and launch the pathway. This will lead to the expression of inflammatory cytokines. In contrast, TLR3 does not use the MyD88-dependent pathway but instead uses the TRIF-dependent pathway [25]. The TRIF-dependent pathway will cause activation of TRAF3 and subsequent activation of interferon regulatory factor 3 (IRF3) and mediates type I interferon (such as IFN-β) expression [26].

Generally, the activation of TLR signaling pathways will cause subsequent immune responses to eliminate pathogens. However, aberrant activation of TLR signaling pathways may disturb immune homeostasis. Thus, its aberrant activation can be involved in the pathogenesis of autoimmune diseases, such as systemic lupus erythematosus, multiple sclerosis, T1D, and cancer [27]. Given the importance of TLRs in host defense and the pathogenesis of disease, there must be a complex regulation mechanism to strictly modulate TLR signaling pathways.

Epigenetic regulation of TLRs

Various mechanisms have been proposed to maintain the immunological balance that evolved to regulate TLR signaling pathways [4, 9, 28]. Epigenetics has recently emerged as a new method involved in the regulation of TLRs and has received increasing attention in the field. There are three main epigenetic modifications that take part in the regulation of TLRs: DNA methylation, histone modification, and non-coding RNA. We will discuss here the direct role of epigenetic modification in regulating TLRs (Fig. 2).

Fig. 2
figure 2

Epigenetic regulation of TLRs. TLRs are expressed in both innate and adaptive immune cells. TLRs can be regulated by DNA methylation, histone modifications, and microRNAs, respectively, which will eventually lead to the altered expression of TLRs on the membranes or endosomes. Firstly, DNA methylation that occurs on the promoter region of TLR genes, such as TLR1, 2, 3, 4, 5, 6, and 8, can reduce the expression of TLR on the membranes or endosomes. Secondly, depending on the type of modification, the histone modifications that occurs on the nucleosome near the promoter of TLR genes, such as TLR2, 3, 4, and 5, can also positively or negatively regulate the expression of TLRs on the membranes or endosomes. Finally, TLRs, including TLR2, 3, 4, 7, and 9 can be silenced directly by microRNAs (miRNAs), such as miR-105, 143, 26a, 223, 3148, and let-7i, at the post-transcriptional level. MiRNAs are transcribed from genomic DNA and processed into precursor miRNAs (pre-miRNAs). Pre-miRNAs are then exported to the cytoplasm to generate mature miRNAs. The mature miRNAs are assembled as part of the miRNA-RISC complex, which can silence TLR transcripts through post-transcriptional regulation, and lead to the decreased expression of TLR on the membranes or endosomes

Full size image

DNA methylation and histone modification in TLR regulation

DNA methylation usually refers to the biochemical process in which a methyl group is added to the cytosine DNA nucleotides. DNA methylation is an important epigenetic control mechanism associated with gene regulation. Histones are the primary protein components of eukaryotic chromatin found in cell nuclei and play a role in gene regulation. H1, H2A, H2B, H3, and H4 are the major families of histones. The possible modification types of histones include acetylation, phosphorylation, methylation, ubiquitination, SUMOylation, and so on [29]. These modifications will alter the histone’s interactions with DNA and lead to epigenetic changes, which may regulate disease processes .

Recently, the role of DNA methylation and histone modification in TLR regulation has been studied (Table 2). For example, Haehnel et al. [5] analyzed the expression of TLR2 in human monocytes and macrophages. They found that DNA methylation of the proximal human TLR2 promoter is associated with TLR2 repression. Zampetaki et al. [35] reported that TLR4 expression underwent DNA methylation and histone modification in mouse embryonic stem cells but not in embryonic stem cell-derived differentiated smooth muscle cells. Takahashi et al. [36] suggested that TLR4 gene transcription is downregulated by epigenetic modification, including histone deacetylation and DNA methylation, in human intestinal epithelial cells.

Table 2 DNA methylation and histone modification of TLRs
Full size table

The above studies have provided us with evidence that the expression of TLRs can be directly regulated by DNA methylation and histone modification. We note that the studies regarding DNA methylation and histone modification in the direct regulation of TLRs are still limited. However, considering the important role of DNA methylation and histone modification in the regulation of gene expression, we believe that there will be increasing research on the regulation of TLRs by DNA methylation and histone modification.

MicroRNAs and TLR signaling pathways

Non-coding RNAs (ncRNAs) are RNA molecules that are transcribed from DNA but have low protein coding potential [39]. MicroRNAs are the most studied ncRNAs to date. MicroRNAs usually induce gene silencing by binding to the 3′UTR of targeted mRNAs, and they play key roles in the post-transcriptional regulation of gene expression. MicroRNAs have recently been found to play a role in linking the innate and adaptive immune systems [40]. It has been shown that microRNAs are regulated by TLR signaling; however, microRNAs can also regulate TLR signaling pathways by targeting TLRs, TLR associated signaling proteins, transcription factors, cytokine mRNAs, and TLR signaling regulators [41]. We will now discuss the microRNAs that are involved in regulating the expression of TLRs, as well as the unconventional roles of microRNAs (Table 3).

Table 3 Relations between microRNAs and TLRs
Full size table

MicroRNAs serve as ligands of TLRs

The conventional role of microRNAs is regulation of gene expression. However, recent studies have suggested that microRNAs act as ligands of TLRs [42, 43]. Since microRNAs are short single-strand RNA molecules, they may bind to the TLRs. MicroRNAs have been suggested to be ligands of TLRs. The activation of nuclear factor κB (NF-κB) by miR-122, miR-15b, miR-21, and miR-155 stimulation can be blocked with TLR1 blocking antibody, while knockdown of TLR1 by shRNA resulted in NK-92 cells losing their capability for miRNA-mediated activation [44]. Another study found that the miR-21 and miR-29a released from cancer cell exosomes can bind to TLR7 in mice and TLR8 in human macrophages [45]. MiR-21 released from tumors can also induce myoblast apoptosis in cancer cachexia via binding to TLR7 and initiating activation of the downstream c-Jun N-terminal kinase pathway [46]. These findings are important for new drug development in cancer treatment. Let-7, a microRNA with a high expression level in microglias and neurons in the central nervous system, was reported to activate TLR7, and the expression level of Let-7b was increased in the cerebrospinal fluid (CSF) of Alzheimer’s disease patients [47]. The unconventional role of Let-7b on TLR7 was confirmed by another study [48]. These studies provide evidence that microRNAs can directly bind to TLRs and initiate stimulation of signaling pathways.

MicroRNAs serve as regulators of TLR mRNA expression

Another direct role of microRNA on TLRs is shown by the fact that the TLR expression level is manipulated by microRNAs. MicroRNAs may control cell differentiation and cell-specific functions through TLRs. A recent study analyzed the expression of TLR2 in rheumatoid arthritis (RA) fibroblast-like synoviocytes (FLS) following LPS and BLP treatments. Decreased expression of miR-19b was then demonstrated by microRNA microarray, and miR-19a/b was found to directly target TLR2 mRNA and regulate the expression of IL-6 and matrix metalloproteinase 3 release [49]. MiR-105 and miR-143 were also reported to modulate TLR2 protein expression by directly targeting TLR2 mRNA [50, 51]. The relationship between miR-26a and TLR3 was studied in rat macrophages. MiR-26a mimic treatment was shown to inhibit TLR3 expression and improve pristine-induced arthritis in rats [52]. The Let-7 microRNA family was recently reported to regulate TLR4 expression. They showed that overexpression of Let-7e downregulated TLR4 expression, while inhibition of Let-7e could upregulate TLR4 expression [53]. Recently, a group reported that the expression of TLR7 and TLR9 was regulated by miR-126 [54].

The recent discovery of this new role of miRNAs raises interesting questions. For example, why do some miRNAs have dual functions, that is, some miRNAs can silence TLRs through post-transcriptional regulation, while on the other hand, they can serve as ligands and interact directly with TLRs. How is the interaction between the two pathways regulated? What is the role of miRNAs as a TLR ligand in the pathogenesis of disease? Further studies are needed to assess the biological significance of these miRNAs’ dual roles and their mechanisms in disease states. Although many questions remain to be solved, the discovery of this new miRNA function opens up a new and exciting field of miRNA research on TLRs. The unconventional role of microRNA on TLRs is important in that it is a new mechanism of disease pathogenesis as well as a new strategy for disease treatment.

Taken together, epigenetic mechanisms, such as DNA methylation, histone modification, and microRNAs, have provided many new strategies in regulating TLR signaling pathways. There are additional studies that address the epigenetic regulation of Toll-like receptor pathways including TLRs, adaptor proteins, and transcription factors. However, we discussed here the direct regulation of TLRs by these epigenetic mechanisms. DNA methylation, histone modification, and microRNAs may be involved in the pathogenesis of a series of diseases through their interaction with TLRs either directly or indirectly. The study of epigenetic regulation of TLRs may provide new clues to clarify the pathogenesis of autoimmune diseases, including T1D.

T1D and TLRs

T1D is a T cell-mediated autoimmune disease that is characterized by selective destruction of pancreatic β cells in genetically susceptible individuals [8]. Humoral immunity is involved in the pathogenesis of T1D, and several autoantibodies against β islets have been identified so far. Studies from our group suggest the diagnostic value of autoantibodies in T1D [55,56,57]. In contrast, cellular immunity has been thought to play a more pivotal role in T1D pathogenesis [58].

Devaraj et al. [59, 60] reported that the surface expression levels of TLR2 and TLR4 were increased in T1D monocytes compared with controls, while the ligands of TLR2 and TLR4, such as endotoxin, Hsp60, and HMGB1, were also increased in T1D patients. These data indicate that the inflammatory response triggered by TLR2 and TLR4 is important in the development of T1D. Our group analyzed the expression of TLR2 and TLR4 in monocytes of T1D patients. However, the results showed that the expression of both TLR2 and TLR4 was not changed in T1D patients compared with controls [61].

Fulminant type 1 diabetes (FT1D) is a subtype of T1D characterized by the abrupt onset of insulin-deficient hyperglycemia [62]. Shibasaki et al. [63] detected the expression of TLR3, TLR7, and TLR9 in the pancreas of FT1D patients by immunohistochemistry and in situ hybridization. Studies from our group found a significant reduction of Foxp3 expression in PBMCs of FT1D patients, which indicated that there may exist a Treg development/function defect. A series of experiments were conducted, and the data suggested that DNA methylation may impair TLR9-induced Foxp3 expression by preventing IRF-7 from binding to the Foxp3 promoter and thus impairing Foxp3 expression [64].

TLRs induced diabetes in a mouse model

Studies performed on mouse models of type 1 diabetes, especially the non-obese diabetic (NOD) mouse, strongly support the hypothesis that TLRs are involved in the pathogenesis of T1D in NOD mouse.

Vallois et al. [65] found that, compared with the diabetes prone NOD mouse, the expression level of TLR1 in splenocytes and thymocytes from diabetes-resistant Idd6 NOD.C3H-congenic mice was decreased. They suggested that TLR1 pathways are involved in the induction of diabetes in the NOD mouse. Alyanakian et al. [66] found that the oral administration of a bacterial extract (OM-85) can delay the onset of diabetes in NOD mice through TLR2-, TLR4-, and MyD88-dependent signaling pathways. Wen et al. [16] found that upregulation of TLR3 by Poly I:C can lead to diabetes in the B6/RIP-B7.1 mouse, and the underlying mechanism may be related to the upregulated APCs and islet Ag-reactive T cells, which can then destroy islet β cells. In addition, Wen et al. [67] found that MyD88, the key adaptor protein that recognizes microbial stimuli in TLR signaling pathways, was important in the pathogenesis of diabetes because the specific pathogen-free MyD88−/− NOD mice do not develop T1D. Because MyD88 is common to multiple TLRs, further studies were conducted in NOD mice lacking individual TLRs (TLRKO). They found that TLR2, TLR3, and TLR4 were dispensable for the development of T1D when deleted individually. They also found that MyD88 deficiency can change the composition of the gut microbiota [67]. The role of microbiota in the pathogenesis of T1D was also reported to be related to TLRs in several other studies [68, 69].

It has been hypothesized that the TLRs are involved in the balance between CD4+CD25+ T regulatory cells (Tregs) and T effector cells. TLRs may have an effect on autoimmune responses in several ways, such as activation of APCs and T cells [16, 70] and an effect on modulating Tregs [71, 72]. Filippi et al. [73] found that diabetes can be prevented in prediabetic NOD mice after treatment with the TLR2 agonist Pam3CSK4 (P3C), and they also found an increased number and function of Tregs. An enhanced function of Treg cells through TLR2 was also reported by Karumuthil-Melethil et al. [74]. However, there are several contradictory studies [75, 76]. TLRs are important regulators of Tregs, and it seems that the number and function of Tregs can either be increased or decreased; the concentration of TLR ligands may have a role in influencing the above different outcomes.

TLRs induced diabetes in a rat model

Bio-breeding diabetes-prone (BBDP) and bio-breeding diabetes-resistant (BBDR) rats are two inbred strains of BB rats, and they are prone to develop diabetes similar to that of human T1D.

Ewel et al. [77] found that a high dose of Poly I:C (a TLR3 agonist) can accelerate the development of diabetes in BBDP rats. However, Sobel et al. [78] found that treatment of BBDP by low doses of Poly I:C (a TLR3 agonist) can prevent diabetes. Sobel et al. [79] also investigated the role of Poly I:C in BBDR rats. They found that both 5 and 10 mg/g body weight can significantly induce the development of diabetes in BBDR rats. It seems that contrary effects of Poly I:C administration are dose-related.

Guberski et al. [80] found that the Kilham rat virus (KRV) can induce diabetes in naive BBDR/Wor rats. However, they did not identify which TLR pathways were involved in the pathogenesis of the disease. Zipris et al. [81, 82] found that the activation of TLRs (TLR2, TLR3, TLR4, TLR7, TLR8, and TLR9) by Kilham rat virus (KRV) infection can induce T1D in BBDR rats. Recently, Tirabassi et al. [83] used KRV rat cytomegalovirus (RCMV), H-1, vaccinia, Coxsackie B4, and Poly I:C, and they found that KRV and RCMV can induce diabetes in up to 60% of LEW.1WR1 rats, whereas other viruses cannot. Taken together, data obtained from both the mouse and rat models imply that TLRs are involved in the pathogenesis of T1D. Virus infection, TLR ligand activation, and their related pathways may play a pivotal role in mediating TLR-induced diabetes.

Genetic association of TLRs with T1D

Accumulating genetic association studies also support the hypothesis that TLRs are associated with T1D in humans. Considering the role of TLRs in mediating the link between the environment factors and the adaptive immune system, Pirie et al. [84] assessed the role of TLR3 in South African T1D patients. They found a significant association between TLR3 and T1D. However, the results were no longer significant after correction of the P value [84]. Assmann et al. [85] found that rs3775291 and rs13126816 polymorphisms in TLR3 were associated with risk for T1D, while rs5743313 and rs11721827 polymorphisms were associated with age at T1D diagnosis and poor glycemic control. They also found that the number of risk alleles seemed to influence the risk for T1D, which suggests that these polymorphisms might interact in the susceptibility to T1D [85]. Park et al. [86] investigated the association of TLR2 in Korean T1D patients, and they found that TLR2 polymorphisms were associated with T1D and the differences were not influenced by HLA genes. The authors suggested a close relationship between innate and adaptive immunity. The results were confirmed by Bjørnvold et al. [87] in a Norwegian population. However, there are also studies that have demonstrated no association of TLR2 and TLR4 polymorphisms with T1D [88, 89]. Sun et al. [90] genotyped 28 SNPs in TLR1–6 and TLR8–9 in 429 T1D patients and 300 controls, and they found that SNPs and haplotypes in TLR1 and TLR6 were associated with T1D in a Chinese population. However, it should be noted that the sample size in some of the studies was rather small.

Given that the SNPs in TLRs may change the expression, cell surface trafficking, and functional responses level of TLRs [91, 92], these genetic association studies are important in support of the view that some TLRs may be involved in the pathogenesis of T1D. However, all of these association studies failed to genotype all SNPs in the reported TLR genes. Thus, the TLR allele or the haplotype may not have been accurately reconstructed. Further studies performed in different populations are needed to explore the possible association between TLRs and T1D.

Conclusions/summary

T1D is an autoimmune disease [93]. As a component of innate immunity, TLRs are believed to be involved in the pathogenesis of autoimmune diseases including T1D. Indeed, results from NOD mouse model and rat model studies, genetic association studies, and TLR expression studies in T1D patients all support this view. Epigenetics has emerged as a new mechanism in regulating TLR expression. DNA methylation, histone modification, and microRNAs can directly regulate the expression of TLRs and thus influence TLR signaling pathways. The reversible nature of epigenetic alterations allows epigenetic regulators of TLRs as potential options for T1D therapy. Some therapeutic strategies have been proposed [94,95,96,97]. However, the conflicting findings in animal models suggest that TLRs may act as a double-edged sword. Their safety and efficacy should be well considered before applying them to the treatment of T1D. Many more studies of TLRs should be performed to clarify the underlying epigenetic mechanisms of TLR regulation and the role of TLRs in the pathogenesis, prevention, and treatment of T1D. Further studies are still needed to demonstrate the role of epigenetic regulation of TLRs in the pathogenesis of T1D.