Genomic deletion of TLR2 induces aggravated white matter damage and deteriorated neurobehavioral functions in mouse models of Alzheimer’s disease

Introduction

Alzheimer’s disease (AD) is a neurodegenerative disease typified by chronic inflammation and neuronal loss in the brain [1, 2]. In patients with familial Alzheimer’s disease (FAD), mutations in the APP gene, PSEN1 gene, and PSEN2 gene were found. In addition, other candidate genes associated with AD were also identified, of which the polymorphic apolipoprotein E (apoE) gene was reported to be the most related [3]. Due to these genetic mutations, the deposition of Aβ and the hyperphosphorylation of the tau-protein appear in the brain [4, 5], inducing the loss of neurons, the activation of astrocytes, and the hyperactivation of microglia cells [6, 7]. Activated microglia and astrocytes release pro-inflammatory cytokines, leading to inflammatory responses, which are involved in not only neuronal death and neurofibrillary tangle formation but also in Aβ clearance and neuroregeneration [8]. Substantial evidence has demonstrated that inflammation plays a key role in the pathological processes of AD [9]. Excessive inflammation associated with the deposition of Aβ and the hyperphosphorylation of the tau-protein results in neuronal loss and white matter damage (WMD) [10–12]. On the other hand, moderate inflammation is helpful for eliminating the deposition of Aβ and for neuroregeneration [13]. The mechanisms underlying the regulation and modulation of inflammation in AD brains are, however, unclear at present.

Toll-like receptors (TLRs) are a family of type-1 transmembrane receptors. TLRs, possessing the toll/ interleukin-1 receptor (TIR) domain and leucine-rich repeat (LRR) motifs, regulate host defensive response via the myeloid differentiation primary response 88 (MyD88)-dependent pathway and/or the MyD88-independent signaling pathway [14]. Activated by ligands, TLRs recruit serial downstream kinases, leading to the activation of nuclear factor kappa B (NF-kB) and/or interferon regulatory factor 3 (IRF3) and resulting in the release of pro-inflammatory factors and anti-inflammatory factors [15]. TLRs were found to be expressed in neural precursor cells, neurons, and glial cells, and are involved in the immune functional maturation of microglia, as well as in the differentiation and development of neurons [16].

Recently, the role of TLRs in the AD pathological process has attracted the attention of investigators. Previous studies demonstrated that modulating TLRs results in changes in pathology and neurobehavioral functions in AD rodent models [17–20]. For example, a deficiency of TLR4 in AD models up-regulated cytokines and glial cell activation [21]. The activation of TLR4 by agonists improved cognitive impairments in rat models of AD [22]. TLR2, a member of the TLR family but different from TLR4 and other TLRs, is mediated through MyD88 alone. Recent studies demonstrated that activated bone-marrow-derived microglia (BMDM) could uptake Aβ and help to clear Aβ deposition [23], while TLR2 deficit BMDM could not perform its Aβ clearance function; moreover, TLR2 deficiency aggregated cognitive dysfunction in APP/PS1 transgenic mice [24, 25]. In contrast, other studies have reported that long-term administration of the TLR2 inhibitor in AD mice could reduce Aβ aggregation and glial activation [26], and that TLR2 gene knockout and the blocking of the interaction between TLR2 and MyD88 could attenuate the neurotoxicity and pathological changes of AD [27, 28]. While these contradictory phenomena could be interpreted as the consequence of different experimental conditions, such as differences in animal models, observed time points, and cell types, it is nonetheless apparent that TLR2 does play a role in the process of AD, although the exact effect remains to be elucidated.

To clarify the role of TLR2 in the pathological process of AD, in the present study, TLR2 knockout (KO) plus APPswe/PSEN1dE9 transgenic mice (AD-TLR2KO) were generated. Cognitive and emotional behavioral tests were conducted on the mice at the age of 12 months. Cortical thickness and white matter integrity were evaluated using brain magnetic resonance imaging (MRI). Neuron loss was evaluated using NeuN staining. Aβ, GFAP, proteins related to synaptic function, endogenous ligands for TLR2, and the activation of downstream signaling of TLR2 in brain tissue were detected by Western blots and immunohistochemistry (IHC) staining.

Results

Discussion

Prior research has shown that AD is a heterogeneous disease involving multiple pathogenic factors, including β-amyloid deposition, tau-protein hyperphosphorylation, and inflammatory reaction, among other mechanisms [29], of which immune inflammatory response is the core pathological mechanism. The imbalance of immune inflammatory response and its regulatory mechanism promotes the occurrence and development of AD. Many immune inflammatory factors are involved in AD [30], including microglia, peripheral immune cells, the interleukin family, the tumor necrosis factor family, and TLR-mediated transduction pathways [7, 8, 13, 31, 32].

TLRs are a type I transmembrane receptor superfamily that plays an important role in the induction and regulation of immune inflammatory responses. At present, at least 13 TLRs have been found in mammals, including 10 TLRs (TLR1 to TLR10) in humans and 12 in mice (TLR1 to TLR9, TLR11 to TLR13) [33]. Activated by their ligands, TLRs activate downstream protein kinases through regulatory proteins, leading to the activation of nuclear transcription factors (NFκB) and/or the interferon regulatory factor (IRF), as well as to the release of immune inflammatory factors [33, 34]. TLR signaling is mainly mediated by two regulatory proteins, MyD88 (myeloid differentiation factor 88) and TRIF (TIR-domain-containing adapter-inducing interferon-β), which are, respectively, mediators of the MyD88-dependent signaling pathway and the TRIF-dependent (MyD88-independent) signaling pathway.

Recently, the role and mechanism of TLRs in the development of AD, and the role of specific drugs for TLRs in AD, have been reported [35]. TLR2, TLR4, TLR5, TLR7, and TLR9 mRNAs were found to be significantly up-regulated in the plaque-associated brain tissue of aged APP23 transgenic mice compared to the plaque-free tissue [36]. Expressions of mRNAs and proteins of TLR2 and TLR4 in peripheral blood mononuclear cells (PBMCs) were markedly elevated in late-onset AD (LOAD) patients [37]. Levels of TLR4-dependent cytokines, such as tumor necrosis factor (TNF)-alpha, interleukin (IL)-1beta, IL-10, and IL-17, in the brains of transgenic AD mice were significantly higher than those in non-transgenic WT mice [38]. Behavioral, molecular, and electrophysiological evidence indicates that early minor stimulation of microglial TLR2 and TLR4 receptors attenuates AD-related cognitive deficits in rats [39]. These data strongly suggest that TLR signaling is involved in AD progression and can be a new therapeutic target for AD.

TLR2, different from other TLRs, is only propagated through the MyD88-dependent pathway. Previous studies have indicated that mice lacking TLR2 receptors are largely protected against ethanol-induced cytokine and chemokine release, as well as against behavioral-associated effects during alcohol abstinence [40]. Activation of TLR2 by intracerebroventricular injection of its ligand, Pam3CSK4, induced sickness behavior symptoms, including anorexia, hypoactivity, and hyperthermia [41]. Recent studies on the role of TLR2 in AD have yielded conflicting results. For example, one study revealed that inhibiting TLR2 by the anti-TLR2 antibody improved performance in spatial learning by decreasing microglial and astroglial activation, and by reducing Aβ plaque burden [26]. However, another study reported that the intracerebroventricular administration of low-dose TLR2 ligands reduced disturbances in spatial and working memory and reversed the impaired long-term potentiation induced by Aβ [39].

To further clarify the effect of TLR2 on neurobehavioral and pathological changes in AD, in the present study, we used APPswe/PSEN1dE9 transgenic (AD, App⁺Psen1⁺) mice and TLR2 gene knockout (TLR2KO, Tlr2^-/-) mice to generate Tlr2^-/-App⁺Psen1⁺ (AD-TLR2KO) mice. Neurobehavior functions in the mice were evaluated with different gene types at the age of 12 months. The data from the Morris water maze (MWM) demonstrated that the genomic deletion of TLR2 alone prolonged the latencies in finding the hidden platform during the acquisition/learning phase, decreased the frequencies of crossing the platform area, reduced the time spent in the target quadrant, and increased the ratio of the edge-type strategy. Moreover, the genomic deletion of TLR2 prolonged the latencies in finding the hidden platform during the acquisition/learning phase and increased the ratio of the edge-type strategy. The results indicated that the TLR2 deficit alone impaired learning ability and reference memory function, and that this deficit also exacerbated learning impairment, but not reference memory function, in AD mice at the age of 12 months. Open field tests demonstrated that TLR2KO decreased exploration times in the central area among AD mice as well as the total traveled distance. In the tail suspension test, the rest time significantly increased in AD-TLR2KO mice when compared with WT mice. These data indicate that the genomic deletion of TLR2 significantly aggravated the increased anxiety and depression in AD mice at the age of 12 months.

Diffusion tensor imaging (DTI) is a vital, non-invasive method for characterizing microstructural changes or differences with neuropathology and treatment [42, 43]. To evaluate brain damage, in the present study, diffusion tensor images were obtained and analysis of selected ROIs from ECs was performed by using a 7.0-T BioSpec 70/20 MRI scanner and diffusion spectrum magnetic resonance imaging (DSI) studio post-analysis software. The FA value represents brain white matter integrity, which is highly sensitive to microstructural changes. Da and Dr are indicators of myelination and axonal damage. The mean diffusivity (MD) value may serve as a marker of tissue water in edema and proliferation in neoplasia [42, 44, 45]. Our results demonstrated that the FA value significantly decreased in TLR2KO, AD, and AD-TLR2KO mice. In addition, the values of Da decreased, while those of Dr increased, in AD-TLR2KO mice compared with WT mice. The cortical thickness measured from T2-weighted images decreased in TLR2KO and AD-TLR2KO mice compared with WT mice. These data indicate that the TLR2 deficit was involved in WMD, myelination and axonal impairment, and brain atrophy, and further aggravated WMD in AD mice. Impaired WMD might be a mechanism by which the TLR2 deficit induces neurobehavioral dysfunction.

Neuropathology in AD is characterized by the deposition of Aβ, senile plaques, and neurofibrillary tangles (NFTs). Gliosis and increased inflammation play important roles in the disease’s pathogenesis. A previous study demonstrated that low-dose TLR2 ligands decreased Aβ deposits [39]. Conversely, other studies reported that blocking TLR2 with the anti-TLR2 antibody reduced Aβ plaque burden in APP/PS1 mice [26], and that the disruption of TLR2–MyD88 interaction inhibited hippocampal glial activation and reduced Aβ burden in 5XFAD mice [28]. To understand the mechanisms underlying the effect of the TLR2 deficit on impaired neurobehavioral functions, we detected Aβ deposition in mouse brains. Our data showed that Aβ deposition significantly increased in AD, but not in TLR2KO, mice. The increased Aβ deposition in AD mice was not inhibited in AD-TLR2KO mice. The data indicated that the TLR2 deficit resulted in the impairment of neurobehavioral functions, possibly via a non-Aβ mechanism.

Glial activation and inflammatory reaction are important self-protection mechanisms [8]. However, excessive inflammatory response is detrimental and accelerates disease process [46]. Many studies have shown excessive astrocyte gliosis and inflammation in patients with AD and in animal models of AD [31, 47]. Our data demonstrated that GFAP, a marker for astrocytes, significantly increased in AD mice compared with WT mice. Moreover, the increased GFAP was enhanced in AD-TLR2KO mice. The data indicated that the TLR2 deficit increased GFAP activation in AD mice, which might be a mechanism by which the TLR2 deficit induces the impairment of neurobehavioral functions.

Previous studies have reported that synaptic function is impaired in AD [48–50]. To investigate the effect of TLR2 on synaptic plasticity, we detected the protein levels of postsynaptic density 95 (PSD-95) and synaptophysin (Syn) in brain tissue. The results demonstrated no difference in the levels of PSD-95 and Syn among the groups, which in turn indicates that TLR2 deficiency does not lead to the changes in proteins associated with synaptic plasticity in mice at the age of 12 months.

One of the most important pathological characteristics in AD brains is the loss of neurons. In the present study, neuronal numbers were evaluated in the brains of mice at the age of 12 months by IHC staining for a neuronal marker, NeuN. The results showed (Figure 8) that neuronal numbers significantly decreased in TLR2KO, AD, and AD-TLR2KO mice compared with WT mice (p<0.05). However, no further decreased neuronal numbers were observed in AD-TLR2KO mice compared with AD mice (p>0.05). These data indicate that TLR2 deficiency was not able to induce the loss of neurons in AD brains. In contrast, TLR2 deficiency alone induced the loss of neurons in brains without AD.

The classic hallmarks of AD include the formation of Aβ plaque deposits and NFTs. However, the mechanisms underlying the initiation and progression of AD remain unclear. Recent studies have demonstrated that endogenous danger signals stimulate innate immune cells via TLR2 and TLR4 and accelerate inflammatory responses [51]. Biglycan, one prototype extracellular matrix-derived damage-associated molecular pattern, functions as an endogenous agonist of TLR2/4 and mediates sterile inflammation through TLR2- and/or TLR4-dependent signaling pathways [52]. Another central endogenous inflammatory mediator, HMGB1, when interacting with TLR signaling pathways, contributes to the pathogenesis of several inflammatory disorders [53]. However, our data failed to show increased protein levels of biglycan and HMGB1 in AD mice compared with WT mice. This phenomenon could be interpreted as meaning that since the mice used in the present study were 12 months old, some biochemical reactions occurring in the early stage of the disease are not observable in the late stage. Interestingly, TLR2 deficiency significantly increased the protein level of biglycan, but not HMGB1, indicating that TLR2 might play an important role in the clearance of biglycan.

TLR2 signaling is mediated via the MyD88-dependent pathway, leading to the phosphorylation of NFκB and to inflammatory responses. Previous studies reported that the expression of TLRs was up-regulated and TLR-mediated inflammation was activated in AD brains [36–38]. Our data, showing increased protein levels of MyD88 and p-NFκB in AD mice at the age of 12 months, indicate that the TLR downstream MyD88-dependent pathway was activated in AD brains. In addition, increased MyD88 and p-NFκB were inhibited in AD-TLR2KO mice, indicating that, in the late stage of AD, the activation of MyD88-dependent signaling is mediated through TLR2.

In summary, the present study demonstrated that the genomic deletion of TLR2 induced impaired neurobehavioral function, WMD and loss of neurons, brain atrophy, and the activation of astrocytes. TLR2 deficiency exacerbated impaired neurobehavioral function and enhanced the activation of astrocytes in AD brains without affecting Aβ deposition. Further studies on molecular mechanisms, especially in the early stage of AD, must be conducted. In addition, since TLR signaling is involved in the pathology of AD, it has been considered as a therapeutic target for AD. Several studies have demonstrated that the activation or inhibition of TLR2 by an exogenous reagent could attenuate impaired neuronal function and pathological changes in AD models. However, the reported results have been controversial. Therefore, the exact effect of TLR2 agonists/antagonists on neurobehavioral functions and pathological changes in AD, their dose–effect relationship, and the time window for treatments require further investigation.

Materials and Methods