Abstract
Heat shock protein 70, encoded by the HSPA1A gene in humans, is a key component of the machinery that protects neuronal cells from various stress conditions and whose production significantly declines during the course of aging and as a result of several neurodegenerative diseases. Herein, we investigated whether sub-chronic intranasal administration of exogenous Hsp70 (eHsp70) exerts a neuroprotective effect on the temporal cortex and areas of the hippocampus in transgenic 5XFAD mice, a model of Alzheimer’s disease. The quantitative analysis of neuronal pathologies in the compared groups, transgenic (Tg) versus non-transgenic (nTg), revealed high level of abnormalities in the brains of transgenic mice. Treatment with human recombinant Hsp70 had profound rejuvenation effect on both neuronal morphology and functional state in the temporal cortex and hippocampal regions in transgenic mice. Hsp70 administration had a smaller, but still significant, effect on the functional state of neurons in non-transgenic mice as well. Using deep sequencing, we identified multiple differentially expressed genes (DEGs) in the hippocampus of transgenic and non-transgenic mice. Furthermore, this analysis demonstrated that eHsp70 administration strongly modulates the spectrum of DEGs in transgenic animals, reverting to a pattern similar to that observed in non-transgenic age-matched mice, which included upregulation of genes responsible for amine transport, transmission of nerve impulses and other pathways that are impaired in 5XFAD mice. Overall, our data indicate that Hsp70 treatment may be an effective therapeutic against old age diseases of the Alzheimer’s type.
INTRODUCTION
Alzheimer’s disease (AD) is the most common form of dementia in the aging human population, accounting for 70–80% of total cases. AD is a progressive neurodegenerative disorder characterized by the formation of pathogenic soluble amyloid-β (Aβ) oligomers, Aβ plaques, neurofibrillary tangles, microglial activation, and neuronal loss [1–3]. AD progression is age-dependent and results in memory impairment and decline of other cognitive abilities [3]. While significant progress has been made uncovering the role of specific genes in AD [4, 5], little is known about global molecular changes that lead to the neurodegeneration and brain dysfunction.
Despite considerable efforts by many scientists around the world, no studies have demonstrated significant improvement in primary outcomes, nor have any successful therapeutic approaches been developed [4, 7]. Therefore, there is an urgent need for further elucidation of the molecular cascades involved in the progression of AD to reveal new pharmacological therapeutic targets. Although it is well known that various AD-associated neuropathologies have become evident in the brains of transgenic animal models and human patients, it is not clear what cascades may be responsible for the manifestation of this neurodegenerative disorder in the human population. However, there are particular consequences of the normal aging process that may reveal important targets for pharmacological interference to prevent or delay the progression of AD-related brain pathologies [8].
Disrupted protein homeostasis is one of the major hallmarks of AD-like neurodegeneration, which leads to the accumulation of toxic proteins with altered conformation, resulting in the apoptosis of neuronal cells. Notably, the temporal cortex and hippocampal regions, responsible for cognitive abilities, represent the brain regions most affected during the process of normal aging and in AD patients [3]. Moreover, in both cases, the level of neuronal death, but not the plaque number, correlates with the severity of cognitive impairment.
It has been previously demonstrated that endogenous heat shock proteins (Hsps) exert neuroprotective activity in rodent models of Huntington’s disease, amyotrophic lateral sclerosis, and several other pathologies [9–12]. Due to its chaperone properties, Hsp70 binds partially unfolded or misfolded proteins and either assists in their refolding, or directs them to a safe disposal site [13, 14]. However, the activity of endogenous Hsp70 appears to be insufficient in various neurodegenerative disorders [9, 15].
In our previous studies, we demonstrated a significant protective effect on various molecular and cognitive characteristics with intranasal administration of human recombinant Hsp70 in two transgenic mouse models of AD [16, 17]. It is worth emphasizing that the protective effects of Hsp70 are not only limited to its well-known chaperone properties but also include complex interactions with multiple cellular pathways. These interactions result in anti-apoptotic and anti-inflammatory effects, the ability to decrease the consequences of oxidative stress, and the preservation of normal function and homeostasis of proteasomal and lysosomal systems [18–20]. These beneficial properties allow Hsp70 to increase the survival of neuronal progenitors and, in cellular and mouse models of AD, to effectively protect neurons from the accumulation of Aβ [21, 22]. We have previously shown that, at the molecular level, exogenous Hsp70 dramatically diminished the accumulation of Aβ peptide in olfactory bulbectomized (OBX) animals and reduced Aβ-plaque formation in 5XFAD transgenic mice [17]. Therefore, in recent decades Hsps and, in particular, Hsp70 have emerged as critical regulators of proteins associated with neurodegenerative disease pathologies, including the accumulation of Aβ and tau in brain regions severely affected in AD patients [23–25].
Furthermore, we have shown that Hsp70 treatment extends the lifespan of wild-type mice and significantly improves life quality [26], i.e., influences the main risk factor for AD in the aging process. Therefore, recombinant Hsp70 definitely belongs to geroprotectors.
Unfortunately, in most AD models neuronal death manifested as a decrease in neuron density and/or increase in apoptosis is not observed [5]. It is of great importance to describe the entire spectrum of neuronal pathologies in AD models where such information is limited. Previously, we performed a detailed analysis of the major alterations in the state and morphology of neurons located in the hippocampus and cortex of OBX mice which represent a model of sporadic AD [5, 28]. Herein, we investigated the influence of sub-chronic eHsp70 administration on the neuronal morphology and density in transgenic 5XFAD mice that represent an early onset model of hereditary AD [5, 28]. Non-transgenic mice of the same age were used as controls. In the course of this analysis, we elucidated the similarities and differences in the pattern of neuronal pathologies occurring in the temporal cortex and areas of the hippocampus of transgenic 5XFAD mice.
It was a challenge to investigate the effect of subchronic Hsp70 treatment on the neuronal functional state and in parallel monitor resultant transcriptional changes in transgenic and control non-transgenic 5XFAD mice. Here, we reported the results of transcriptomic analysis in the hippocampus of 5xFAD mice, which depended on transgenicity per se and eHsp70 sub-chronic administration.
MATERIALS AND METHODS
Animals and treatment
Adult male transgenic 5XFAD mice (2.5 months old) were used in the experiments. Transgenic (Tg) and non-transgenic (nTg) 5XFAD mice treated with the same dose of Hsp70 (2 μg/mouse in 4 μl intranasal injections) daily for 1.5 months were divided into the following four groups: Tg+NaCl, Tg+ Hsp70, nTg+NaCl, and nTg+ Hsp70. It is of note that we successfully employed this dose in our previous studies [17, 26] and, hence, we decided to use the same dosage in the present study for comparison. Three mice of each group were used for histological analysis of brain regions.
Animals were maintained in their home cages in a climate-controlled room (21–23°C) with a 12:12 h light-dark cycle and had free access to water and food. 5XFAD mice (TG6799) have been described previously [27]. These mice co-express the Swedish (K670N/M671L), Florida (I716V), and London (V717I) mutations in human APP(695), and M146L and L286V mutations in PS1, with expression of both transgenes driven by the Thy1 promoter. Transgenic mice were acquired from Charles River laboratories and were maintained on a mixed SJL/C57Bl6 background. Mice were genotyped by PCR analysis of DNA extracted from ear biopsies. The transgenic cassette was detected using primers 5’-AGGACTGACCACTCGACCAG-3’ and 5’-CGGGGGTCTAGTTCTGCAT-3’, yielding a 377 bp product. All procedures involving animals were reviewed and approved by the Animal Care and Use Committee of Branch of Shemyakin & Ovchinnikov Institute of Bioorganic Chemistry. The procedures corresponded to principles and procedures approved by the NIH USA Committee for Animal Care and Experimental Use (NIH Publications N 8023, revised 1978).
Tissue preparation
Mice were anaesthetized with an overdose of Nembutal (60 mg/kg, i.p.) and transcardially perfused with 20 ml of 0.1 M PBS (pH 7.4). Brains were rapidly removed and divided by hemisphere. One hemisphere was frozen on dry ice and stored at –80°C for subsequent transcriptomic analysis. The other was fixed in 4% phosphate-buffered paraformaldehyde at 4°C for 48 h before being stored in 30% sucrose. Hemibrains were cut in the coronal plane into 10 μm sections on a cryostat and stored in a glycol-based cryoprotectant at –20°C until histological analysis.
Histology and morphology
General histological methods were performed as described previously [17]. A detailed description of all procedures is provided in the Supplementary Material.
Isolation and labelling of human recombinant Hsp70
In our research, we used human recombinant Hsp70 expressed in army worm (Spodoptera) cells or in E. coli. The original clone containing human Hsp70 cDNA (pBlueScriptSK(+)-hsp70) (a generous gift of Dr. R. Morimoto, North-Western Univ., Evanston, USA) was subcloned into the donor plasmid pFastBacHTb-hsp70 under control of the polyhedrin promoter and pET-14b expression vector. The plasmids were subsequently used in the Bac-to-Bac system (Invitrogen) and in E. coli to express human LPS-free Hsp70 as described in previous studies [29, 30]. The recombinant protein contained six His at its N-terminus, enabling its isolation from the cell extracts using an Ni-NTA resin column according to the manufacturer’s instructions (QUIAGEN, Ni-NTA Superflow BioRobot Hand Book).
The purity of Hsp70 preparations from E. coli and Spodoptera cells was confirmed by PAGE-electrophoresis followed by staining with Coomassie Blue and immunoblotting using monoclonal 3B5 anti-Hsp70 and N69 anti-Hsc70 antibodies. The protein concentration was measured according to Bradford’s protocol.
Transcriptome sequencing
Total RNA for the libraries was isolated from homogenized hippocampus samples of individual mice by QIAzol Lysis Reagent (Qiagen, Netherlands) and further isopropanol precipitation. The RNA concentration was measured with a Qubit®2.0 Fluorometer (Invitrogen, USA) and a NanoDrop® ND-1000 spectrophotometer (NanoDrop Technologies Inc., USA). The A260/A280 ratio of the RNA samples ranged from 1.8 to 2.0. The integrity of the isolated RNA (RNA integrity number, RIN) was determined using a Bioanalyzer Agilent 2100 (Agilent Technologies, USA). The RIN value for all samples was not less than 8.0. All samples were treated with DNase I (Fermentas, Lithuania).
To prepare samples for the mRNA sequencing libraries, we used the Illumina TruSeq™ RNA Sample Preparation Kit. Total RNA from each sample (2.5 μg) was used to purify the poly-A containing mRNA molecules by poly-T oligo-attached magnetic beads, with two rounds of purification. During the second elution of poly-A RNA, the RNA was also fragmented and primed for cDNA synthesis according to the manufacturer’s protocol. Then, the fragmented mRNA samples were subjected to cDNA synthesis using a SuperScript Double-Stranded cDNA Synthesis kit (Invitrogen, USA). The cDNA was converted into double-stranded (ds) cDNA and then Ampure XP beads were used to separate ds cDNA from the second-strand reaction mix.
The ds cDNA was blunt-ended through an end-repair reaction. Next, a single ‘A’ nucleotide was added to the 3′ ends of the blunt fragments to prevent them from ligating to one another during the adapter ligation reaction. The cDNA fragments were then ligated to specific RNA Adapter Indexes supplied in the kit. The In-Line Control DNA was added to each enzymatic reaction. To selectively enrich DNA fragments with adapter molecules on both ends and to amplify the amount of DNA in the library, the PCR process was used according to the manufacturer’s protocol (15 cycles).
The quantity of libraries was determined using a qPCR method on a Rotor-Gene 6000 PCR System (Qiagen, USA) according to the manufacturer’s protocol (Sequencing Library qPCR Quantification Guide). The primers matched sequences within adapters flanking an Illumina sequencing library. Before starting qPCR, a control template was selected to measure the libraries for quantification. The quality of libraries was determined on an Agilent 2100 Bioanalyzer (Agilent, USA) per the manufacturer’s instructions. The size and purity of the samples were checked. The final product produced a band at approximately 260 bp. Indexed cDNA libraries were normalized to 2 nM and pooled together in equal volumes. Sequencing was performed using the HiSeq™2000 platform (Illumina, USA), 50 bp single reads were generated. Approximately 50 billion reads were obtained for each sample.
Analysis of transcriptome sequencing results
The initial processing of the sequencing data was performed with PPLine toolkit [31], which including quality control (FastQC), read trimming and filtering (trimmomatic), alignment to mouse transcriptome GRCm38.78 (STAR), and gene expression quantification (HTSeq-tools). The following steps were performed in an R environment. Because our experiment has a complex design (transgenic/wt, and treat/control mice), we applied a generalized linear model (GLM) approximation approach to reveal genes with differential expression between groups using edgeR [32]. After read count normalization (TMM method), the derived gene expression profiles were analyzed to fit the following GLMs: Transgenic status + Treatment, which revealed genes being differentially expressed between the treat/control groups. The transcriptomic effect of HSP may be different in Tg and NTg mice, but we deliberately discarded the interaction member (e.g., we did not using linear model – Treatment + Transgenic status + Treatment: Transgenic status). This would abrogate the statistical significance of the derived results since it is equal to the comparison of small groups of samples (2 mice in nTg groups versus 3 mice in Tg groups) which did not pass permutation tests. Transgenic status, which revealed genes altered in transgenic mice in comparison to wt. Here we omitted the Treatment component in the last model since overall transcriptomic effect of HSP treatment was found to be much less than the changes induced by transgenicity.
We used these two different models instead of a single model to increase the reliability of derived results: first, the maximum variance is explained with the first GLM coefficient (e.g., with Transgenic status in Transgenic status + Treatment model); then the second coefficient (Treatment) is analyzed for the residual variance.
Permutation tests were performed to ensure the reliability of results. Comparison of large groups (e.g., all Tg versus all NTg; all HSP versus all NaCl) has passed this test, but the pairwise comparisons of elementary groups (e.g., Tg HSP versus NTg HSP, Tg HSP versus Tg NaCl, and other variants) have failed permutation tests because of the high variability of samples and insufficient group size (2 or 3 mice). To understand whether this result was due to biological or technical variability, we performed complete technical replicate (including RNA isolation, library preparation, and sequencing) and revealed a perfect match between the two technical replicates. Therefore, we concluded that biological variability plays a significant role.
Gene set enrichment analysis (Gene Ontology and KEGG) for the top 40, 80, 250, 500 and 1000 up/downregulated genes (p < 0.05) was performed using topGO and clusterProfiler [33, 34]. We used Pathview package [35] to visualize altered KEGG pathways.
RESULTS
Protective action of eHsp70 on temporal cortex and hippocampal neurons in 5XFAD transgenic mice
Memory loss in AD patients is linked with neuronal death and deterioration, with the temporal cortex and hippocampal C1 and C3 areas being the most affected during the course of this disease progression. Therefore, we investigated the morphological state and density of neurons in these brain regions in Tg and nTg 5XFAD mice that either were or were not administered Hsp70. The major observed neuronal pathologies included pyknosis, karyolysis, cytolysis, and vacuolization (Supplementary Figure 1). Additionally, neuronal density and the proportion of neurons with normal morphology were estimated.
ANOVA revealed significant effects of Hsp70 on the majority of the traits studied as its own main effect or in interaction with the genotype. Therefore, we performed post hoc analysis (test Bonferroni) and FDR test summarized in Table 1A and B.
The p values of pairwise differences obtained in Bonferroni post hoc tests (A) and FDR tests (B) for four experimental groups. A) Bonferroni post hoc tests results
p values for pairwise tests adjusted by multiplying with the number of all possible comparisons (Statistica 10.0), taking into account factors of genotype (2), medicinal effect (2), analyzed brain regions (3), and six histological indices, a total of 144 possible comparisons were used for estimation.
According to Benjamini and Hochberg [42] correction for the false discovery rate (FDR), the adjusted critical p-values at the first significance level for the variables Pyknosis, Karyolisis, Cytolis, Vacuolization, Normal neurons, and Density are: 0.021, 0.025, 0.044, 0.017, 0.033, 0.038, respectively, at the second significance level: 0.0042, 0.0038, 0.0075, 0.0017, 0.0067, 0.0075, respectively, and at the third significance level: 0.0003, 0.0003, 0.0005, 0.0001, 0.0004, 0.0007, respectively. (The significance levels are marked by one star, two stars and three stars, respectively).
Figure 1 and Supplementary Table 1 depict the results accumulated in the process of comparative analysis of neuron morphology and density in each group.
Proportion of normal neurons (A) and neuron density (B) in the brain regions of different groups depending on Hsp70 treatment. Square brackets above the columns indicate which group is compared with.
Notably, nTg mice did not display any significant changes in form, size, or structural organization of neurons. Most of the neurons in this group exhibited normal phenotypic characteristics that were evenly distributed, tigroid, and light colored, with centrally positioned nuclei and easily distinguishable nucleoli. Analysis also demonstrated that Hsp70 treatment results showed a significant protective effect on the morphological state of neurons in most brain regions examined (Fig. 1, Supplementary Table 1). In particular, Hsp70 treatment increased the proportion of normal neurons in Tg animals where this parameter was significantly diminished in comparison to nTg mice in both hippocampal regions and the temporal cortex.
Furthermore, the investigation of neuronal density, which represents one of the most important criteria of brain function, was significant decreased (p < 0.001) in CA3 region but not in the cortex of Tg mice compared to age-matched nTg animals (Fig. 1 and Supplementary Table 1). Importantly, Tg mice treated with Hsp70 displayed significantly higher neuronal density in the cortex compared with untreated nTg and Tg animals (p < 0.001). Therefore, Hsp70 treatment increased neuronal density in the temporal cortex but did not affect this parameter in the examined hippocampal regions (Fig. 1B). Similarly, although Hsp70 treatment exhibited beneficial effects for most pathologies assessed, the effect depended on the brain region and specific neuronal pathology being monitored. Table 1 and Supplementary Figure 2 depict statistical analysis of the major pathologies in hippocampal regions and the temporal cortex of the studied groups. It is evident that although eHsp70 improves most morphological and functional parameters, in a few cases, it exerts a detrimental effect. Thus, Hsp70 treatment increased the frequency of pyknosis in the both hippocampal regions of nTg mice (Supplementary Figure 2, Supplementary Table 1).
To better illustrate the effect of Hsp70 administration on the morphology of neurons, we compared the parameters indicated above in pairs of groups representing maximal interest in terms of dependence on eHsp70 administration. Such an approach enabled us to present all major data in a graphical form (Fig. 2) and monitor the events occurring in different brain regions of transgenic and non-transgenic animals.
Graphical illustration of the observed differences in neuronal density and morphology in the compared groups depending on transgenicity and Hsp70 treatment. The upper row in all panels depicts the state of temporal cortex; the middle row-CA1 area of hippocampus; the bottom row- CA3 area of the hippocampus. Black color of the rectangle denotes deterioration of a given parameter; a gray rectangle denotes the absence of any significant differences between the compared groups; white bars denote a significant decrease in pathology frequency or improvement of the studied parameter in comparison to the reference group. N, normal neuron proportion; p, pycnosis; c, cytolysis; k, karyolysis; den, density of neurons.
Figure 2 summarizes the effect of transgenicity and Hsp70 treatment on the compared morphological and functional neuronal state in each group. As expected, the brains of untreated Tg animals exhibit impairment in most histological, morphological, and functional characteristics when compared with nTg ones (Fig. 2).
Furthermore, Tg and nTg animals respond differently to Hsp70 treatment. Treatment improves a few parameters in nTg mice, including the proportion of normal neurons and neuron density in the cortex but leads to the deterioration of other parameters in CA1 and CA3 hippocampal regions (pyknosis and cytolysis). Hsp70 treatment exerted the most pronounced effect on neuronal morphology and density in Tg mice (Figs. 1 and 2, Table 1, and Supplementary Table 1). Thus, Tg mice treated with Hsp70 demonstrate significantly better outcomes in the cortex when compared with Tg untreated control animals (Fig. 2). The same is true concerning most pathologies studied when hippocampal regions from different groups are compared (Fig. 2, Table 1). This result is in agreement with a previous study on the effect of eHsp70 in OBX mice, another AD model, in which we demonstrated that the treatment may produce certain adverse effects in neuronal morphology in control sham operated mice. Furthermore, the treatment apparently results in neuronal protection predominantly in the case of neuropathology [17].
Interestingly, Tg 5XFAD mice treated with Hsp70 even exceeded nTg Hsp70-treated group in several important parameters, both in the temporal cortex and especially in hippocampal regions (Fig. 2, Table 1). Hence, Hsp70 treatment exhibited significantly stronger “normalizing” effects in terms of neuronal pathology when Tg animals were used. This analysis clearly demonstrates that Tg mice are more susceptible to the action of eHsp70.
Next, we investigated global transcriptomic changes in the brains of Tg versus nTg mice and the influence of eHsp70 treatment on the activity of differentially expressed genes (DEGs). Unfortunately, we were not able to obtain reproducible data when studying transcriptome of temporal cortex probably due to insufficient amount of the material and, hence, we concentrated on the investigation of hippocampal transcriptomes obtained for different experimental groups (Tg versus nTg and NaCl versus Hsp70).
RNA-Seq transcriptome analysis
The comparison of transgenic versus non-transgenic 5XFAD mice (model-Transgenic status) revealed 1897 DEGs (p < 0.05), 279 of which passed the FDR < 0.05 threshold. Gene set enrichment analysis (GSEA) of the upregulated genes demonstrated activation of pathways related to immune and inflammatory processes in transgenic mice, including cytokines and tumor necrosis factor production; leucocyte, lymphocyte, and macrophage migration and activation; and antigen processing and presentation (Fig. 3A, Supplementary Figure 3). Importantly, our data show perfect concordance with the results of Landel et al. [28], who studied transcriptomic changes in the hippocampus and cortex of 5XFAD mice of various ages (1, 4, 6, and 9 months). The spectra of Tg-associated overexpressed genes in young mice (1 month) was strikingly different from the other three ages. Landel found that 4 genes (Cst7, Clec7a, Itgax, and Ccl4) are among the top 10 upregulated in the hippocampus of 4-, 6-, and 9-month-old mice. According to our data, these genes are also among the top 10 overexpressed genes: they hold the 1st, 7th, 2nd, and 10th places, respectively, on the Tg-associated upregulated gene list.
Differentially expressed genes (DEGs) in the hippocampus of the compared groups (Tg versus nTg). The effect of transgenicity (left column) and Hsp70 administration in both groups (right column). Expression profiles of genes participating in various biological processes (Gene Ontology terms). For each term, genes are collected (irrespective of p-value) and sorted by decreasing LogFC. Overexpression is marked with red, while downregulation is marked in blue. Colour frames indicate GSEA p-values across the following statistically significantly enriched GO terms: top 40, 80, 250, 500, and 1000 upregulated (red frames) or downregulated (blue frames, see border color legend).
KEGG (Kyoto Encyclopedia of Genes and Genome) pathway of the hippocampi of Tg animals analysis also revealed activation of Toll-like and NOD-like receptor signaling pathways, Natural killer cell-mediated cytotoxicity, T cell and B cell receptor signaling pathways, TNF signaling pathway, activation of phagocytosis, and platelet activation. Additionally, possible dysregulation of MAPK, Rap1, cAMP, and calcium signaling was observed (Supplementary Figure 3).
Among the downregulated genes in transgenic 5XFAD mice, we found the following highly enriched GO terms related to neuronal function: nervous system development (GO:0007399), axon development (GO:0061564), neuronal differentiation (GO:0030182), neuronal development (GO:0048666), learning or memory (GO:0007611), cognition (GO:0050890), regulation of neuronal synaptic plasticity (GO:0048168), generation of neurons (GO:0048699), calcium ion transport (GO:0006816), and synaptic transmission (GO:00072681) (Fig. 3A). Notably, our data, derived for hippocampi, are not in complete agreement with the results of another group studying the same model [36] and who monitored profiling transcriptomic changes in the cortices of transgenic 5XFAD mice.
HSP-induced changes
Hsp70 treatment also introduced multiple changes in the transcriptome, but the extent of the changes was much smaller than when transgenic and naïve mice were compared. We found 994 DEGs (p < 0.05) in the Hsp70 treated group but only 19 of them passed Benjamini-Hochberg adjustment (FDR < 0.05). Furthermore, the average extent of gene upregulation was smaller than for the previous comparison (average LogFC for top 100 genes is 1.55 as opposed to 2.09 for the comparison of Tg versus NTg mice). However, GSEA for the top upregulated genes clearly demonstrated the over-representation of numerous genes participating in transmission of nerve impulses, including the dopamine biosynthetic process (GO:0042416), amine transport (GO:0015837), neurotransmitter loading into synaptic vesicles (GO:0098700), neuropeptide signalling pathway (GO:0007218), and G-protein coupled receptor signaling pathways (GO:0007186) (Fig. 3A,B and Supplementary Figure 3). In addition, upregulation of genes involved in the glutathione metabolic process (GO:0006749), detoxification (GO:0098754), and fatty acid metabolic process (GO:0006631) was observed. Glutathione is a well-known antioxidant and plays an essential role in the intracellular antioxidant defense against free radicals, especially OH. Decreased levels of glutathione in the aged brain are associated with increased oxidative stress and represent one of the major factors in AD development [37]. Surprisingly, exogenous Hsp induced overexpression of genes participating in MHC presentation, leucocyte differentiation, and antigen processing and presentation (Supplementary Figure 3). Notably, three of the five Hsp-treated mice showed the best transcriptomic response, including simultaneous Hsp-induced overexpression of the following genes involved in neuronal activity and neurogenesis: Slc6a3, Hcrt, Pmch, Foxa1, Th, Sim1, Chrna6, Avp, Lmx1b, and others.
DISCUSSION
In a previous investigation [17], we demonstrated the dramatic neuroprotective effect of recombinant human Hsp70 in the bilateral OBX and 5XFAD mouse models of AD-type neurodegeneration. We showed that intranasally administered Hsp70 rapidly enters the afflicted brain regions and mitigates multiple AD-like morphological and cognitive abnormalities observed in model animals [17, 38]. In particular, in OBX mice, Hsp70 diminished the level of neuronal pathology and normalizes the density of neurons in the hippocampus and temporal cortex, correlating with decreased accumulation of Aβ peptide and reduced Aβ-plaque formation in transgenic 5XFAD mice. Several behavioral tests also demonstrated that such sub-chronic Hsp70 treatment protects spatial memory in OBX and 5XFAD mice.
In another study exploring the same treatment, we demonstrated that in aging mice, eHsp70 extended lifespan and significantly improved exploratory and locomotor activities [26]. Moreover, eHsp70 administration resulted in larger synaptophysin immunopositivity in the brain and significantly decreased the accumulation of aging-related marker lipofuscin in the brains of old mice [26].
The analysis reported herein represents a logical extension of our previous studies on the therapeutic effects of eHsp70. We analyzed neuronal density and several other major parameters of neuron functional state in transgenic Tg and non-transgenic 5XFAD mice and evaluated the effect of eHsp70 administration.
In particular, significant protection was evident in the temporal cortex and two areas of the hippocampus (CA1 and CA3), the regions predominantly responsible for cognitive abilities, in most of the compared groups in this study. Notably, the extent of pathological changes observed in the brains of transgenic mice significantly exceeded that of nTg animals and included not only the hippocampus but the temporal cortex as well, where the density and quantity of normal neurons are also diminished (Figs. 1 and 2).
The effect of Hsp70 treatment differs characteristically in Tg and nTg groups; it is more pronounced for most parameters in the brains of Tg mice. In other words, in Tg mice, eHsp70 exerts highly significant beneficial effects on temporal cortical neurons and, to a lesser degree, on neurons in hippocampal areas (Figs. 1 and 2 and Supplmentary Figure 2). The molecular mechanisms underlying greater positive effect of Hsp70 administration in Tg mice warrants further investigation. One possibility is that basal Hsp70 levels in the neurons of Tg mice are lower than those in nTg.
To understand transcriptomic changes accompanying the early stages of AD-like degeneration in Tg 5XFAD mice, we analyzed DEGs from male mice at the age of four months. Previously, using the same model, a clear shift in gene expression patterns between mice at 1, 4, 6, and 9 months of age has been demonstrated [28]. All older mice share a large proportion of DEGs. Based on these data, we selected 4-month-old mice to represent an early onset model of AD-like degeneration for our analysis. Experimental mice (age 2.5 months) were sub-chronically treated intranasally with eHsp70 for 1.5 months. The Tg mice exhibited a very peculiar spectrum of DEGs in comparison with age-matched control nTg mice of the same strain (279 DEGs with FDR < 0.05). For technical reasons in the present study we concentrated on the transcriptomic changes only in the hippocampal regions that play pivotal roles in cognitive abilities, but we plan to investigate temporal cortex in this respect in our subsequent studies using different age groups for comparison.
All DEGs associated with transgenicity (Fig. 3) can be roughly divided into three major groups. The first group comprises genes associated with synaptic transmission at different levels. As expected, in Tg mice, we observed decreased expression of genes responsible for transmission of nerve impulses, regulation of neuronal synaptic plasticity, neurogenesis, generation of neurons, etc. (Fig. 3). Deterioration of all these processes should inevitably lead to the decrease of neuronal plasticity underlying learning and memory parameters significantly impaired in Tg 5XFAD mice. The observed deficit of dopamine-associated processes (biosynthesis and metabolism) in Tg mice is also significant in light of recent data suggesting an important role for this neurotransmitter in various neuropathologies, including AD and Parkinson’s disease [39]. The second group of DEGs specific for Tg mice include genes with increased expression compared to nTg animals. These upregulated genes are responsible for immunoreactivity and associated genes underlying inflammatory response and cytokine production, including TNF synthesis. The activation of these processes leads to the inhibition of genes responsible for cell division, growth and proliferation and, simultaneously, to the activation of genes underlying apoptosis. The death of neuronal cells in Tg 5XFAD mice is also associated with an observed decrease in expression of genes responsible for neurogenesis and the generation of neurons and, hence, results in a reduced quantity of normal neurons in hippocampus and diminished total neuronal density in the brains of Tg mice (Fig. 1).
The third group of DEGs specific for Tg mice include genes likely activated due to cellular compensatory mechanisms switched on due to transgenicity. This group includes genes responsible for ribosome biogenesis, translational kinetics and regulation of MAPK (mitogen-activated protein kinase). In mammals, this pathway is represented by several enzymes (e.g., ERK1, ERK2, etc.) that play an important role in neuroprotective mechanisms [40]. Activation of the MAPK cascade leads to increases in cell proliferation and growth, synaptic plasticity, and migration, i.e., processes significantly inhibited in the brains of Tg mice.
The comparison between treated and untreated animals demonstrated that eHsp70 significantly modulates the expression of the major above mentioned groups of DEGs. Thus, Hsp70 treatment increased the expression of the genes responsible for amine transport, transmission of nerve impulses, dopamine biosynthetic processes, and several other pathways compromised in Tg mice (Fig. 3). Furthermore, eHsp70 efficiently modulates oxidation-reduction processes and activates glutathione metabolism essential for ROS inhibition. Previously, we demonstrated with various in vitro systems that recombinant human Hsp70 efficiently decreases production of ROS and TNFα and is able to modulate apoptosis process [3, 38]. These results are also in accord with several recent studies that stressed the role of neuroinflammation in AD onset [6, 41].
Taken together, the results of the present study corroborate our previous results and the data accumulated by many other groups on the neuroprotective action of Hsp70. Transcriptomic analysis suggests that eHsp70 achieves its effects by inhibiting oxidative stress and suppressing neuroinflammation in Tg animals, which, hence, leads to the amendment of various pathologies in the brain of Tg mice associated with increased ROS production, synaptic dysfunction, and impaired transmission of nerve impulses observed in various AD models (reviewed in [3]). The suggested nasal delivery pathway has several advantages for treating neurodegenerative disorders, including convenience of administration and non-invasiveness.
