Chronic neuron-specific tumor necrosis factor-alpha expression enhances the local inflammatory environment ultimately leading to neuronal death in 3xTg-AD mice

Abstract

Inflammatory mediators, such as tumor necrosis factor-alpha (TNF-alpha) and interleukin-1beta, appear integral in initiating and/or propagating Alzheimer's disease (AD)-associated pathogenesis. We have previously observed a significant increase in the number of mRNA transcripts encoding the pro-inflammatory cytokine TNF-alpha, which correlated to regionally enhanced microglial activation in the brains of triple transgenic mice (3xTg-AD) before the onset of overt amyloid pathology. In this study, we reveal that neurons serve as significant sources of TNF-alpha in 3xTg-AD mice. To further define the role of neuronally derived TNF-alpha during early AD-like pathology, a recombinant adeno-associated virus vector expressing TNF-alpha was stereotactically delivered to 2-month-old 3xTg-AD mice and non-transgenic control mice to produce sustained focal cytokine expression. At 6 months of age, 3xTg-AD mice exhibited evidence of enhanced intracellular levels of amyloid-beta and hyperphosphorylated tau, as well as microglial activation. At 12 months of age, both TNF receptor II and Jun-related mRNA levels were significantly enhanced, and peripheral cell infiltration and neuronal death were observed in 3xTg-AD mice, but not in non-transgenic mice. These data indicate that a pathological interaction exists between TNF-alpha and the AD-related transgene products in the brains of 3xTg-AD mice. Results presented here suggest that chronic neuronal TNF-alpha expression promotes inflammation and, ultimately, neuronal cell death in this AD mouse model, advocating the development of TNF-alpha-specific agents to subvert AD.

Figure 1

Neurons in the 3xTg-AD mouse brain endogenously express TNF-α. Combined immunohistochemistry for neuronal nuclear antigen (NeuN) or microglia/macrophages (F4/80) and in situ hybridization specific for mouse TNF-α was performed on brain sections from 6-month-old 3xTg-AD mice (n = 4) that were perfused and sectioned under RNase-free conditions. Radiolabeled ³⁵S-labeled riboprobes for an anti-sense TNF-α (A and C) and a sense TNF-α sequence (negative control, panel B) were used for in situ hybridization at a specific activity of 1 × 10⁸ cpm/μg Sections were subsequently incubated with an anti-NeuN or anti-F4/80 antibody to stain neurons and microglia/macrophages, respectively. Positive TNF-α signal via in situ hybridization was indicated by at least 15 clustered grains that colocalized to single immunohistochemically NeuN (A) or F4/80-expressing cells (C). Dotted boxes outline areas magnified in insets shown in (A) and (C). The numbers of F4/80⁺/TNF-α⁺ and NeuN⁺/TNF-α⁺ cells were enumerated using quantitative image analysis and are depicted graphically in (D). Error bars indicate SD. Scale bar =50 μm.

Figure 2

Vector construction and characterization of rAAV vectors expressing hTNFα and eGFP. Two rAAV vectors were constructed: one expressing human TNF-α (rAAV-TNFα) and a second expressing enhanced green fluorescent protein (rAAV-eGFP) as a negative control. The individual transgenes were placed under the transcriptional control of the human cytomegalovirus promoter (A). Baby hamster kidney cells were transfected with pAAV-TNFα or 293A cells transduced with 10⁹ particles of rAAV-TNFα. A human TNF-α enzyme-linked immunosorbent assay was performed on cell culture supernatants 72 hours after transduction according to manufacturer’s instructions to determine the concentration (pg/ml) of human TNF-α product by each vector tested (B). A cohort of 3xTg-AD mice subsequently injected with 3 × 10⁹ TU of rAAV-egfp (ipsilateral)/saline (contralateral) or rAAV-TNFα (ipsilateral)/saline (contralateral) were sacrificed 4 months postinjection, perfused, and sectioned coronally at 30 μm. Sections were co-incubated with a primary antibody specific for either eGFP (green signal; C, E, I, K) or human TNF-α (blue signal; F, H, L, N) and primary antibodies specific for neurons (NeuN; D, E, G, H; red signal) or astrocytes (GFAP; J, K, M, N; red signal). Following incubation with a designated set of secondary antibodies, photomicrographs of the transduced CA1 subregion of the hippocampal formation were captured by 2-color confocal microscopy (n = 3 per vector). Overlapping signals appear yellow in color for eGFP (E and K), or pink in color for human TNF-α (H and N). A separate set of 3xTg-AD mice (n = 4) were injected intrahippocampally with 3 × 10⁹ TU of rAAV-eGFP or rAAV-TNFα, sacrificed 4 months postinjection, hippocampal tissue microdissected, and mRNA isolated and processed for quantitative real-time RT-PCR using a human TNF-α-specific TaqMan fluorogenic primer/probe set (O). These mRNA samples were subsequently analyzed via real-time RT-PCR using a human APP^Swe, human PS1^M146V, or human Tau^P301L-specific TaqMan fluorogenic primer/probe set to determine whether rAAV-TNFα transduction resulted in transcriptional up-regulation of the resident transgenes (P, Q, or R, respectively). Error bars in B, O, P, Q, and R indicate SD. C-N: 40 × magnification.

Figure 3

rAAV-TNFα delivery results in selective localized enhancement of F4/80⁺ microglial cell staining and intracellular Aβ42 accumulation four months post-transduction. Two month-old 3xTg-AD mice were stereotactically injected with either rAAV-TNFα and saline (A–G) or rAAV-eGFP and saline (H–N) and were sacrificed at 6 months of age (n = 4 per group). Mice were perfused and brains were removed and sectioned at 30 μm for immunohistochemical staining for either microglia/macrophages using an F4/80 marker-specific antibody (A–C and H–J), for astrocytes using a GFAP-specific antibody (D, E, K, L), or for intracellular Aβ42 using an Aβ1–42 specific antibody (F, G, M, N). Sections were subsequently processed using DAB histochemistry and photomicrographs were obtained at original magnification ×1.25 (A and H), ×20 (B–E and I–L), or ×40 (F, G, M, N). Insets in (B), (D), and (F) represent digitally magnified images of the respective photomicrograph for better visualization of stained cell morphology. Scale bars in (L) and (N) = 50 μm. Quantitative image analyses for cells/pathologies staining positive for F4/80 (O), GFAP (P), and intracellular Aβ42 (Q) were performed on each brain and presented in histogram format. Error bars indicate SD. ***P < 0.001 as determined by one-way analysis of variance with Bonferroni posthoc analysis using the saline-injected hemisphere as the baseline comparator.

Figure 4

Four months of chronic TNF-α expression in 3xTg-AD hippocampus leads to subtle alterations in AD-related pathologies. 3xTg-AD mice injected with rAAV-TNFα (A, C, E, G) or rAAV-eGFP (B, D, F, H) at 2 months of age were sacrificed at 4 months postinjection and brains were sectioned at 30 μm and analyzed for the presence of human APP transgene product (A and B), extracellular human Aβ_1–42 (C and D), hAPP/Aβ (E and F), and a hyperphosphorylated epitope of hTau (G and H) via immunohistochemistry with DAB development. The inset in each panel represents a sixfold magnification of the region outlined by the dotted box for more optimal visualization of stained pathological hallmarks. Scale bar in (G) = 100 μm.

Figure 5

Chronic expression of TNFα reveals transcript level differences between 3xTg-AD and Non-Tg mice. Hippocampi were dissected from 3xTg-AD and Non-Tg mice injected with rAAV-TNFα or AAV-eGFP at 12 months of age (n = 6). RNA was extracted and cDNA was synthesized using Applied Biosystems High-Capacity cDNA Archive kit. Samples were analyzed by quantitative real-time RT-PCR for steady-state expression levels of the transcripts derived from the three resident transgenes in 3xTg-AD mice: human APP^Swe (A), human PS1^M146V (B), and human Tau^P301L (C). A separate 100-ng sample for each transduced 3xTg-AD and Non-Tg mouse was loaded onto Microfluidic Cards (Applied Biosystems) and qRT-PCR was performed. rAAV-TNFα values are compared to rAAV-eGFP control samples and data are expressed as fold-change for 3xTg-AD (D) and Non-Tg mice (E). *P < 0.05 as assessed by two-way analysis of variance and Bonferroni posthoc test for intragenotype comparisons. Additional intergenotype comparisons were performed between rAAV-TNFα injected Non-Tg and 3xTg-AD mice with one-way analysis of variance to decipher differences in transcript expression. ⁺P < 0.05; error bars represent the SD of the ΔC_T values.

Figure 6

Loss of NeuN-positive neurons specifically in 3xTg-AD mice injected with rAAV-TNFα indicates a role for TNF-α in AD-related neurotoxicity. Two month-old 3xTg-AD mice (A–D) and Non-Tg (E–H) control mice injected with rAAV-TNFα (A, C, E, G) or rAAV-eGFP (B, D, F, H) were sacrificed at 12 months of age, and brains were sectioned at 30 μm and analyzed for the expression of TNF-α (blue signal) or eGFP (green signal) by fluorescence immunocytochemistry and for the presence of NeuN-positive neurons via immunohistochemistry with DAB development. Photomicrographs of the injected hippocampal CA1 area were obtained (n = 6 per group). Scale bar in (H) = 50 μm.

Figure 7

Neuronally derived TNF-α induces heightened chronic microglial activation and contributes to neuronal loss in the CA1 and dentate gyrus of 3xTg-AD mice following 10 months of chronic cytokine expression. 3xTg-AD (A–F) and Non-Tg (G–L) mice stereotactically injected with either rAAV-TNFα (A–C and G–I) or rAAV-eGFP (D–F and J–L) were sacrificed at 10 months postinjection, and brains were sectioned at 30 μm and analyzed for the presence of F4/80-positive microglia via immunohistochemistry with DAB development and a subsequent Nuclear Fast Red counterstain. Photomicrographs of the injected hippocampi were obtained at original magnification ×10 (A, D, G, J). The CA1 (B, E, H, K) and DG regions (C, F, I, L) of the hippocampus were further magnified sixfold for more optimal visualization of stained microglia (brown/purple) and nuclei (red) in regions outlined with boxes. Double-headed arrows depict the boundaries of the pyramidal and granule layers of the CA1 and DG, respectively. Single-headed vertical arrows point to representative F4/80-positive microglia. Scale bars = 50 μm. Nuclear Fast Red-stained nuclei residing in the injected CA1 (M) and proximal DG layers (N) were enumerated (*P < 0.05 as determined by Student’s t-test when compared to rAAV-eGFP values). Error bars indicate SD.

Figure 8

Chronic TNF-α expression leads to a significant and specific enhancement in the numbers of CD45-positive cells at the vector infusion site in 3xTg-AD mice. 3xTg-AD (A–D) and Non-Tg (E–H) mice stereotactically injected with either rAAV-TNFα (A, B, E, F) or rAAV-eGFP (C, D, G, H) were sacrificed at 10 months post injection, and brains were sectioned at 30 μm and analyzed for the presence of CD45-positive cells at the vector-infusion sites via immunohistochemistry with DAB development. Arrows point to representative CD45-positive cells. CD45-positive cells residing in the injected CA1 (I) and proximal DG layers (J) were enumerated (*P < 0.002 and *P < 0.02 respectively, as determined by Student’s t-test when compared to rAAV-eGFP values). Error bars indicate SD.

Figure 9

Ten months after intrahippocampal administration of rAAV-TNFα, plaque load and tau pathology are appreciably reduced in 3xTg-AD mice. 3xTg-AD mice injected with rAAV-TNFα (A, C, E, G) or rAAV-eGFP (B, D, F, H) were sacrificed at 10 months post injection and brains were sectioned at 30 μm and analyzed for the presence of human APP transgene product (A and B), extracellular human Aβ_1–42 (C and D), hAPP/Aβ (E and F), and a hyperphosphorylated epitope of hTau (G and H) via immunohistochemistry with DAB development. The inset in each panel represents a sixfold magnification of the region outlined by the dotted box for more optimal visualization of stained pathological hallmarks. Scale bar in (G) = 100 μm.