Inhibition of soluble tumor necrosis factor ameliorates synaptic alterations and Ca2+ dysregulation in aged rats

Abstract

The role of tumor necrosis factor α (TNF) in neural function has been investigated extensively in several neurodegenerative conditions, but rarely in brain aging, where cognitive and physiologic changes are milder and more variable. Here, we show that protein levels for TNF receptor 1 (TNFR1) are significantly elevated in the hippocampus relative to TNF receptor 2 (TNFR2) in aged (22 months) but not young adult (6 months) Fischer 344 rats. To determine if altered TNF/TNFR1 interactions contribute to key brain aging biomarkers, aged rats received chronic (4-6 week) intracranial infusions of XPro1595: a soluble dominant negative TNF that preferentially inhibits TNFR1 signaling. Aged rats treated with XPro1595 showed improved Morris Water Maze performance, reduced microglial activation, reduced susceptibility to hippocampal long-term depression, increased protein levels for the GluR1 type glutamate receptor, and lower L-type voltage sensitive Ca(2+) channel (VSCC) activity in hippocampal CA1 neurons. The results suggest that diverse functional changes associated with brain aging may arise, in part, from selective alterations in TNF signaling.

Conflict of interest statement

Competing Interests: The authors have read the journal's policy and have the following conflicts: David E. Szymkowski is the Senior Director of Bio-therapeutics at Xencor: the company that manufactures the anti-TNF biologic XPro1595 used in our study. Dr. Szymkowski's role on this project was to provide council on dosing and to provide a background on XPro1595's unique mechanism of action. Dr. Szymkowski played no role in data collection or statistical comparisons. Because the neuroscience field is outside his area of expertise, he did not provide any council or shade any interpretation of XPro1595's effects on neurologic function. For these reasons, the authors that Dr. Szymkowski's status as an author for this study is appropriate and in no way compromised or biased the experimental results or interpretations. Dr. Hafiz Mohmmad Abdul is now employed by Proteostasis Therapeutics, Inc. Dr. Mohmmad Abdul took this position after the data in our article were collected and interpreted. Please note that Proteostasis Therapeutics Inc had no role whatsoever in any portion of this study. The corresponding author of this article, Dr. Chris Norris, currently serves as an Academic Editor at PLoS ONE. This does not alter the authors' adherence to all the PLoS ONE policies on sharing data and materials.

Figures

Figure 1. Increased TNFR1/TNFR2 ratio during aging.

A, Representative Western blots for TNFR1 and TNFR2 in hippocampal membrane fractions from young (6 mos; n = 6) and aged (22 mos; n = 10) male Fischer rats, with the Na/K-ATPase loading control shown below. B, There was no significant change in TNFR1 protein levels with age (p = 0.67); however, there was a significant decrease in TNFR2 with age (* p<0.05).

Figure 2. XPro1595 alters Iba-1 levels.

XPro1595 (0.08 mg/kg/day) was administered to the hippocampus over a six week period using Alzet osmotic pumps. Immunolabeling revealed a reduction in inflammatory phenotype in microglia (A, Iba-1) and, to a lesser extent, in astrocytes (D, GFAP) of XPro1595 treated aged rats. CA1 = CA1 stratum pyramidale. Representative Western blots and mean ± SEM protein levels are for Iba-1 (B and C) and GFAP (E and F) in hippocampal fractions of vehicle and XPro1595-treated rats. β-actin and GAPDH served as loading controls for Iba-1 and GFAP, respectively. Results revealed a marked, significant reduction in Iba-1 protein levels (* p<0.05) and a smaller, nonsignificant reduction in GFAP levels in the XPro1595 group (p = 0.45).

Figure 3. XPro1595 improves learning and synaptic measures. A,

For behavioral assays, XPro1595 (0.08 mg/kg/day) was administered to the hippocampus over a six week period. Performance (mean ± SEM) in the spatial version of the Morris Water Maze was measured by path length (cm), and normalized (%) to block one performance levels for each rat. Each group showed significant improvement by block 6 (p<0.01). However, XPro1595-, but not vehicle-treated, rats showed significant improvement as early as block 2 and 3, suggesting that TNF-blockade accelerates learning rates during aging. B, Mean EPSP slope amplitudes (vertical SEM bars) were plotted against mean FV amplitudes (horizontal SEM bars) to generate CA3-CA1 synaptic strength curves. Each average curve was fit with sigmoidal equations and compared across treatment groups using Z tests. Aged rats that received intraventricular infusions of XPro1595 over a four week period (0.08 mg/kg/day) exhibited a slight, but significant increase in maximal EPSP amplitude, relative to the vehicle group (z = 2.17). Other curve parameters (i.e. slope and half-maximal FV) were not affected by XPro1595. C, In a twin pulse paradigm (50 ms interpulse interval), the EPSP slope corresponding to pulse 2 (S2) was expressed as a percentage (mean ± SEM) of the pulse 1 (S1) EPSP slope to obtain measures of paired-pulse facilitation (PPF). No treatment effect was observed for PPF (p = 0.27). D, Time plot showing normalized mean ± SEM EPSP slope amplitudes collected before (1) and after (2) the delivery of prolonged 1 Hz stimulation (bar, 900 consecutive pulses). The inset shows representative CA1 EPSP waveforms averaged in the pre- (1) and post-1 Hz (2) periods for vehicle and XPro1595 groups. Scale bars indicate 0.5 mV vertical by 2 ms horizontal. E, Bar graph shows the amount of LTD, expressed as a percentage of the pre-1 Hz baseline (mean ± SEM) in each treatment condition. The results revealed significant LTD in vehicle, but not in XPro1595-treated rats (* p<0.05, repeated measures ANOVA).

Figure 4. Effects of XPro1595 on glutamate receptor protein levels. A,

Representative Western blots are shown for AMPA receptor (GluR1 and GluR2) and NMDA receptor subtypes (NR1, NR2A, and NR2B) in hippocampal membrane fractions from aged (22 month) rats treated for four weeks (intraventricular delivery) with vehicle or XPro1595 (0.08 mg/kg/day). The Na/K-ATPase loading control is shown below. B, XPro1595 treatment resulted in a selective increase in GluR1 levels (* p<0.05).

Figure 5. XPro1595 reduces L-VSCC activity in CA1 neurons of aged rats. A,

Left panel, Cartoon illustration of a partially dissociated hippocampal slice in which CA1 is “unzipped” along stratum pyramidale to expose CA1 neurons for patch clamp recording. Right panel, Photomicrograph showing a glass micropipette tip patched onto a CA1 pyramidal neuron in an “unzipped” slice. B, Three representative L-VSCC current traces and the average ensemble current (bottom trace) in cell attached patches from aged rats treated for six weeks (intrahippocampal delivery) with vehicle or XPro1595 (0.08 mg/kg/day). Traces were taken from 45–50 consecutive step depolarizations (150 ms duration) from −70 to +10 mV. C, Mean ± SEM current (I) densities (pA/μm2) for each treatment group are shown. XPro1595 significantly reduced VSCC I density nearly three-fold (* p<0.05). D–E, Representative Western blots (D) and mean+SEM protein levels (E) for the major pore-forming L-VSCC subunits (CaV1.2 and CaV1.3) from hippocampal membrane fractions. The Na/K-ATPase served as loading control. XPro1595 did not significantly alter either CaV1.2 (p = 0.17) or CaV1.3 (p = 0.63) levels.