Aging Does Not Affect Axon Initial Segment Structure and Somatic Localization of Tau Protein in Hippocampal Neurons of Fischer 344 Rats

Abstract Little is known about the specific contributions of aging to the neuron dysfunction and death in Alzheimer’s disease (AD). AD is characterized by the pathological accumulation of abnormal tau (a microtubule-associated protein), and the mislocalization of tau from the axon to the somatodendritic compartment is thought to play an important role in disease pathogenesis. The axon initial segment (AIS) is thought to play a role in the selective localization of tau in the axonal compartment. Thus, disruption in the AIS barrier may allow tau to diffuse freely back into the somatodendritic compartment and potentially lead to neurotoxicity. Here, we analyzed AISs using stereological methods and protein immunoblotting, and the localization of tau was assessed with immunofluorescence optical density measurements and protein immunoblotting. None of the outcome measurements assessed, including AIS structure, AIS protein levels, the distribution of tau in neurons of the hippocampus (HP), and total tau or phospho-tau protein levels were different in young, middle-, and old-age Fischer 344 rats. The outcome measurements assessed, including AIS structure, AIS protein levels, the distribution of tau in neurons of the HP, and total tau or phospho-tau protein levels were not different in young, middle-, and old-age Fischer 344 rats, with the exception of a small reduction in AIS volume and diameter in the CA2 region of aged animals. These data suggest that aging largely has no effect on these properties of the AIS or tau distribution, and thus, may not contribute directly to tau mislocalization.


Introduction
Alzheimer's disease (AD) is the most prevalent neurodegenerative disease of individuals over the age of 65, and affects over 6.4 million adults in the United States (James et al., 2014). The leading risk factor for developing the disease is aging, and while the cause of AD is still unclear, the accumulation of inclusions comprised of tau protein is a hallmark of the disease (Santuccione et al., 2013). It remains unclear whether changes in the distribution of tau that are reminiscent of AD-related changes, such as accumulation in the somatodendritic compartment, occur during normal aging. Aged Fischer 344 rats are commonly used for aging research and exhibit agerelated behavioral deficits of learning and memory as well as motor impairment (Spangler et al., 1994;van der Staay and Blokland, 1996;Cochran et al., 2014). Additionally, aging causes altered neuronal protein expression, axonal atrophy, and some gliosis in the F344 rat model (Parhad et al., 1995). For example, glial fibrillary acidic protein (GFAP) is shown to increase significantly in the hippocampus (HP) of aged rats both with and without cognitive impairment (VanGuilder et al., 2011). Thus, Fischer 344 rats are a good model to recapitulate factors of human aging (van der Staay and Blokland, 1996) and determining whether aging effects the AIS and tau distribution. The underlying hypothesis is that aging causes alterations in AIS integrity and tau mislocalization, which may represent events that contribute to aging as a strong risk factor for AD.
Tau, a microtubule-associated protein, is thought to contribute to the development and progression of AD pathology, but the mechanisms behind tau toxicity remain largely unknown. Age-related accumulations of tau were reported in some nonhuman primate brains (Oikawa et al., 2010;Perez et al., 2013), yet systematic analysis of the cellular localization of tau accumulation across aging is not well studied. Unfortunately, human tissue studies are often limited in scope across age groups making it hard to appreciate changes from young to old age. In healthy neurons, tau is enriched in the axonal compartment, where it may stabilize microtubules and play a role in regulating axonal functions such as transport (Tytell et al., 1984;Binder et al., 1986;Kanaan et al., 2013). The redistribution of tau from the axonal compartment to the somatodendritic compartment is thought of as an important event in tau-mediated neurotoxicity (Buee et al., 2000). The preferential localization of tau in axons is mediated, at least in part, by the axon initial segment (AIS), which acts as a retrograde barrier for freely diffusing tau (Li et al., 2011;Sun et al., 2014;Zempel and Mandelkow, 2014;Sohn et al., 2016). Importantly, very little is known about the connections between aging, the AIS, and tau distribution. The AIS is a selective diffusion barrier separating the axonal compartment from the somatodendritic compartment (Winckler et al., 1999). The establishment of the AIS creates a filter that regulates intracellular traffic based on protein size or the type of motor proteins carrying cargo along microtubules (Song et al., 2009;Leterrier and Dargent, 2014). Ankyrin G (AnkG) is well established as a necessary protein for the development and maintenance of the AIS as a selective filter, as well as a necessary protein for developing and maintaining axonal polarity (Zhou et al., 1998;Hedstrom et al., 2008). AnkG also has an integral role in recruiting and binding structural proteins to the AIS, such as ␤IV-spectrin and neurofascin (Hedstrom et al., 2007;Freal et al., 2016). Interestingly, a single previous report showed that some structural proteins in the AIS (specifically ankyrin) are reduced with age in wild-type mice (Bahr et al., 1994). Thus, we set out to determine if the structure of the AIS and/or tau localization is altered during normal aging. Our data show that AIS structure, levels of multiple AIS proteins, and the distribution of tau in HP neurons are not altered with advancing age in Fischer 344 rats.

Animals
Young adult (4 months), middle-aged (14 months) and aged (24 months) male Fischer 344 rats were used for all experiments. Six animals per age group (n ϭ 6) were used to obtain fixed tissue for histologic analysis while five or six animals (n ϭ 5-6) per age group were used for fresh brain homogenate. The animals were provided rat chow and water ad libitum and housed in a reverse light-dark cycle room. All animal studies were performed in accordance with standard regulations and were approved by the Michigan State University Institutional Animal Care and Use and Committee.

Tissue processing
Animals used for collection of fresh brain tissue were transcardially perfused with 200 ml of 0.9% saline containing heparin (10,000 U/l). The brains were extracted and the HP was dissected and frozen on dry ice. For collection of fresh tissue for AnkG immunoblotting, animals were perfused for 5 min with saline containing heparin (ϳ60 mL). The extracted HPs were then immediately frozen in liquid nitrogen and stored in liquid nitrogen until processing for AnkG immunoblotting. To collect fixed tissue, the saline perfusion was followed by 200 ml phosphate buffered 4% paraformaldehyde. The brains were postfixed in 4% paraformaldehyde for 24 h. After postfixation, the brains were embedded into gelatin blocks for sectioning (Smiley and Bleiwas, 2012). The gelatin block was equilibrated in 20% glycerol. The gelatin block was cut into 40-m-thick coronal sections on a freezing, sliding stage microtome. Sections were stored in cyroprotectant until processed for immunohistochemistry or immunofluorescence.

Immunoblotting
Tissue from the HP was homogenized in 300 l of 10 mM Tris, 1 mM EDTA, 0.8 mM NaCl, and 10% sucrose buffer containing protease and phosphatase inhibitors (10 g/ml pepstatin, 10 g/ml leupeptin, 10 g/ml bestatin, 10 g/ml aprotinin, 1 mM PMSF, 10 mM ␤-glycerophosphate, 1 mM sodium orthovanadate, 10 mM sodium fluoride, and 1 mM tetra-sodium pyrophosphate decahydrate), using a sonicator (XL-2000, Misonix, 10 ϫ 1 s bursts at power level 1). Lysates were cleared of cellular debris by centrifugation at 22,000 ϫ g for 20 min at 4°C. The resulting supernatants were collected for analysis and the total protein content was assessed using the Bradford protein assay (B6916, Sigma). The samples were diluted in Laemmli buffer and heated to 95°C for 10 min. Lysate samples separated using SDS-PAGE (50 g total protein/ lane for AIS proteins; 20 g/lane for tau protein) on 4-20% Criterion TGX (Bio-Rad) gradient gels at 250 V and transferred to nitrocellulose membranes for 50 min (66458; Pall Life Sciences) to quantify the amount of AIS proteins and tau between age groups (n ϭ 5/group).
Because of the large size of the AnkG protein, a blotting protocol optimized for large molecular weight proteins was used (Fairbanks et al., 1971;Davis and Bennett, 1982;Bolt and Mahoney, 1997). The flash frozen HPs were dropped into nine volumes of urea buffer (8 M urea, 5% SDS, 5 mM N-ethyl maleimide, 10 mM HEPES, 10 g/ml leupeptin, and 10 g/ml pepstatin) at 65°C and immediately pulverized with an Eppendorf tube pestle (#12-141-363, Thermo Fisher) for 10 s and then sonicated, as described above. The samples were then diluted in Laemmli buffer and heated to 95°C for 10 min. The protein concentrations of the resulting lysate samples were measured using the SDS Lowry protein assay as described (Cox et al., 2016). Lysate samples (100 g/lane; n ϭ 6/group) were separated using SDS-PAGE on 3-8% Criterion XT Tris-Acetate in XT buffer (Bio-Rad) gradient gels at 150 V and transferred to nitrocellulose membranes in transfer buffer [40 mM Tris, 20 mM sodium acetate, 2 mM EDTA, pH 7.4, 20% (v/v) methanol, 0.05% (w/v) SDS] for 120 min at 10 V. After transferring, the blots were stained with Ponceau S [0.04% (w/v) Ponceau S (#P3504, Sigma), 0.1% (v/v) acetic acid] for 5 min followed by 2 ϫ 5 min washes in 5% acetic acid (v/v) and then two washes in water. The blots were imaged before proceeding to blocking and Ponceau bands were used to normalize the quantified AnkG signals.

Immunohistochemistry and immunofluorescence
A one in six series of tissue sections were processed for each age group (n ϭ 6) for immunohistochemical detection of AnkG. The tissue was rinsed in 0.1 M TBS, pH 7.4, containing 0.5% Triton X-100 (TBS-Tx) six times for 10 min each. The tissue was then incubated in 3% H 2 O 2 in TBS-Tx for 1 h at room temperature to quench endogenous peroxidase activity, and then rinsed again. An avidin/ biotin blocking kit (Vector Labs SP-2001) was used to block endogenous avidin in the gelatin matrix. Nonspecific antibody binding was inhibited by incubating the tissue in blocking buffer [10% goat serum (GS)/2% bovine serum albumin/0.5% TBS-Tx] for 1 h at room temperature. The tissue was then incubated over night at 4°C with rabbit anti-AnkG primary antibody (H-215, Santa Cruz sc-28561) 1:1000 in dilution buffer (2%GS and 0.5% TBS-Tx). The tissue was rinsed (6 ϫ 10 min in 0.5% TBS-Tx) and incubated for 2 h in biotinylated goat anti-rabbit secondary antibody (BA-1000, Vector) at a concentration of 1:500 in dilution buffer. The tissue was then rinsed again and incubated in avidin-biotin complex (ABC) solution (PK-6100, Vector) for 1hour. The ABC solution contained 50 l of solution A and 50 l of solution B in 10 ml of TBS made according to the manufacturer's instructions. After rinsing the tissue, the staining was developed using 3,3'-diaminobenzidine (D5637, Sigma) solution (50 mg/ml, 0.5% TBS-Tx, and 0.003% H 2 O 2 ) for 12 min. The tissue was then rinsed, mounted on microscope slides, and coverslipped with CYTOSEAL 60 (#8310-16, Thermo Scientific).
A one in six series of tissue sections (n ϭ 6/group) was used for the immunofluorescent labeling of tau. All tissue was stained simultaneously in staining dishes using the same reagents. Nonspecific antibody binding was inhibited by incubating the tissue in blocking buffer (10% GS, 2% bovine serum albumin, and 0.5% TBS-Tx) for 1 h at room temperature. The tissue was then incubated over night at 4°C with mouse anti-tau primary antibody (Tau7; 1:3500) in dilution buffer (2%GS and 0.5% TBS-Tx). The tissue was rinsed (6 ϫ 10 min in 0.5% TBS-Tx) and incubated for 2 h in Alexa Fluor 488-conjugated goat anti-mouse IgG (H ϩ L; #A-11001, Thermo Fisher) at a concentration of 1:500 in dilution buffer. The tissue was then incubated in DAPI (0.5 g/ml in 0.5% TBS-Tx for 10 min), rinsed (5 ϫ 10 min in 0.5% TBS-Tx), and the sec-tions were mounted on microscope slides. To block endogenous autofluorescence, the tissue was treated with Sudan black. The slides were incubated in 70% ethanol for 2 min, then in a saturated solution of Sudan Black B (Fisher, #AC419830100) for 5 min. The slides were then differentiated in 70% ethanol until background (gray matter) was pale gray, and rinsed in dH 2 O (2 ϫ 3 min). The slides were then coverslipped using Vectashield hardset mountant (H-1400, Vector; Kanaan et al., 2008).

Stereology
The spaceballs stereological probe is an unbiased and systematic stereological sampling method to estimate the total length of a population of fibers in 3D space, and this probe was used to quantify AIS length in the HP of rats. This stereological method was performed using serial sections (one in six series). Sampling grids (i.e., CA: 500 ϫ 500 m; DG: 200 ϫ 500 m) were chosen for each region to allow for ϳ200 fiber intersections to be counted in the and 270-kDa (B) isoforms of AnkG do not change between young, middle, and old rats. Note that the other isoforms of AnkG not analyzed are isoforms that are not AIS-specific. C, AnkG Western blottings were normalized to Ponceau S (C) staining for loading control. D, E, The amount of ␤IV-spectrin (D) and neurofascin (E) also remain unchanged during aging in rats. F, The level of axonal neurofilaments, as indicated by the SMI-312 antibody. The SMI-312 antibody is an axonal marker that labels phospho-neurofilaments in axons. SMI-312 levels are not changed with age, confirming that total axon content is similar across age groups. G, H, The total level of tau detected with Tau7 (G) and polyclonal R1 (H) do not change with age in the rat HP. I, The levels of PHF1, a phospho-epitope of tau, are not changed with age. J, The total levels of microtubules, labeled with ␤III-tubulin, remained unchanged with advancing age. K, GFAP, an astrocytic marker known to increase with age, is significantly increased in the HP of old age rats compared to young and middle aged rats; one-way ANOVA, Tukey's post hoc, ‫ء‬p Ͻ 0.05. L, Western blot band intensities are normalized to GAPDH, which did not change with age. All data are displayed as mean Ϯ SEM.
Negative Results HP for each hemisphere and to yield a Gunderson coefficient of error Ͻ0.1 for all samples. The mean coefficient of error was 0.058 Ϯ 0.001 (SEM). A hemisphere probe with a radius of 11 m was used to sample sites throughout the HP. The mean measured tissue thickness was ϳ13-14 m. A 4ϫ objective was used to outline each region and a 60ϫ oil objective lens (1.35 numerical aperture) was used for all stereological counts.
The AutoNeuron probe in Neurolucida (MBF Bioscience) was used to acquire measurements of individual AIS length, volume and diameter in the regions of the HP (CA1, CA2, CA3 and dentate gyrus) in young, middle, and old aged animals. In each region, an image stack was acquired with a 60ϫ oil objective lens through the complete depth of each section at a step-size of 0.5 m. A 3D representation of each individual AIS in the HP was created in the software using the following settings for Au-toNeuron: image type, brightfield; max process diameter, 2.00 m; XY region, all; Z range, all; trace somas, no; soma sensitivity, 50; min soma diameter, 2.00 m; seed detection sensitivity, 70; seed response filter, 4; tracing sensitivity, 60; tolerance to gaps, low; and connect branches, no. These settings were chosen to best detect the AIS of each neuron, while not falsely detecting background tissue stain or failing to differentiate individual AISs.

Cellular protein quantification
The tau immunofluorescence-stained tissue was imaged on a Nikon AI confocal system and a Nikon Eclipse Ti microscope with a 40ϫ oil objective (1.30 numerical aperture). Image stacks (0.5 m step size) were acquired of neurons in each HP region (CA1, CA2, CA3, and the dentate gyrus) using acquisition settings in the linear range of fluorescent intensity without saturation of tau signal and the same acquisition settings were used for all animals. Four image stacks were acquired per animal. The neurons were analyzed with NIS Elements (v4.30, Advanced Research, Nikon). The average intensity of Tau7 immunofluorescent signal was measured in the soma of individual pyramidal neurons in all CA regions (CA1, CA2, and CA3). Cells were chosen as randomly as possible, but only cells with an entire cross-section through the middle of the cell (i.e., the z-slice with the largest nuclear width) within the acquired z-stack and without overlapping cell bodies were used for analysis. To increase the rigor of this analysis, the experimenter was blinded to the condition of the samples. A box (ϳ8-12 m 2 ) was drawn as large as possible inside the somatic compartment of each neuron, excluding the DAPI-positive nuclei. The mean value of fluorescence intensity was measured for each box within a total of 180 cells/animal/group. The density and small cytoplasm of dentate granule cells precludes the ability to reliably measure signal intensity within individual cells. Instead, the average fluorescence intensity within a rectangular box (ϳ5000 m 2 ; six boxes/animal/group) was used to perform a regional analysis of immunofluorescence signal intensity within the dentate gyrus ( Fig. 3F-I).
For axonal tau intensity measurements, a 4x objective (0.13 numerical aperture) was used to capture two to three images of the dorsal HP in three serial tissue sections per animal. The entire visible regions of the fimbria, alveus, stratum lacunosum, and stratum moleculare were outlined in NIS Elements software. The average fluorescent intensity per area was measured for each axonal region.

Statistical analyses
All data were analyzed using Prism software (v6.0) and all data are presented as mean Ϯ SEM. Statistical significance between age groups was determined using oneway ANOVA. Significance was set at p Ͻ 0.05. Tukey's post hoc test was used for post hoc comparisons when significance of p Ͻ 0.05 was reached. If no overall significance was achieved, no post hoc analyses were used.
There was no change across age groups in the levels of total tau protein as indicated by Tau7, a C-terminal pantau antibody (F (2,12) ϭ 0.7686, p ϭ 0.4852; Fig. 1G). As a secondary method to confirm there were no changes in total tau levels we also probed blots with the R1 tau antibody, a pan-tau rabbit polyclonal antibody (F (2,12) ϭ 0.6824, p ϭ 0.5240; Fig. 1H) and found no aging-related changes (Berry et al., 2004;Horowitz et al., 2006). To determine if specific tau phosphor epitopes were changed with aging in the HP we probed blots with PHF1 antibody and found no differences (F (2,12) ϭ 2.361, p ϭ 0.1366; Fig.1I). We also probed with the AT8 antibody and the phospho-Ser422 antibody, both of which are diseaserelated modifications of tau (Guillozet-Bongaarts et al., 2006;Jeganathan et al., 2008) and found no signal for either epitope in young, middle or old age animals (data not shown). Tubulin was analyzed to measure the amount of microtubules as a control since tau is a microtubuleassociated protein. There was no change in ␤III-tubulin across age groups (F (2,12) ϭ 0.8679, p ϭ 0.4446; Fig. 1J).
As a positive control for detection of age-related changes of protein, blots were probed for GFAP, a protein previously shown to change in HP lysates of aging rats (VanGuilder et al., 2011). A significant increase in GFAP was found in the old rats compared with the young rats (F (2,16) ϭ 4.529, p ϭ 0.0276; Fig. 1K).

Stereological assessment of AIS length and morphology
Stereological analysis of the AIS length (using the spaceballs probe in AnkG-stained tissue) showed no differences in the total length of all AISs in the HP across age groups (F (2,15) ϭ 0.07323, p ϭ 0.9297; Fig. 2A). We used the AutoNeuron module as a second method to analyze individual AISs in the CA1, CA2, CA3, and dentate gyrus regions of the HP and found that length did not change across age groups in any of these regions (length: CA1:  Fig. 2C,D). However, a small but significant change was detected in the volume and diameter of CA2 neurons when comparing young and old animals (volume: CA2: F (2,15) ϭ 4.460, p ϭ 0.0302; and diameter: CA2: F (2,15) ϭ 4.062, p ϭ 0.0389; Fig. 2C,D). The measured length of the AIS in HP neurons was 32.86 Ϯ 1.99 m in young, 32.33 Ϯ 1.379 m in middle, and 31.53 Ϯ 1.361 m in old aged animals (Fig. 2B), consistent with previous findings in cultured HP neurons (Grubb and Burrone, 2010;Leterrier et al., 2015).

Optical density measurements of tau levels in somata and axonal layers of the HP
Optical density measurements of Tau7 immunofluorescence in the somatic compartment of individual HP neurons in tissue sections showed no change across age Figure 2. Structural analysis of the AIS with AnkG reveals regional changes in the HP during aging. A, The total length of AISs in the HP estimated with the spaceballs stereology probe is unchanged across age groups. B-D, The Neurolucida AutoNeuron analysis reveals that the length (B), volume (C), and diameter (D) of the AISs across young, middle, and old aged rats is not changed in CA1, CA3, or the dentate gyrus of the HP. The volume and diameter of the AISs in the CA2 region were significantly reduced in the old age compared with the young rats (one-way ANOVA, Tukey's post hoc, ‫ء‬p Ͻ 0.05). All data are displayed as mean Ϯ SEM. E-G, Representative images of the AISs (AnkG-positive immunostain, brown) in the dentate gyrus of the HP at ages 4 (E), 14 (G), and 24 (I) months, and the corresponding Neurolucida AutoNeuron tracings (F, H, and J; multicolor overlay to visually differentiate individual AISs). Scale bar, 20 m (E-I). groups (Fig. 3A). Individual analysis of pyramidal neurons in the CA regions did not show an age-related difference (CA1: F (2,15) ϭ 1.775, p ϭ 0.2033; CA2: F (2,15) ϭ 0.7810, p ϭ 0.4757; CA3: F (2,15) ϭ 3.513, p ϭ 0.0561; Fig. 3A), nor did the dentate gyrus show a change in regional intensity with advancing age in rats (F (2,15) ϭ 1.924, p ϭ 0.1804; Fig. 3A). Analysis of tau intensity in HP strata enriched in axonal projections did not detect any change with age (fimbria: F (2,15) ϭ 0.3902, p ϭ 0.6836; alveus: F (2,15) ϭ 0.6694, p ϭ 0.5267; stratum lacunosum: F (2,15) ϭ 0.08560, p ϭ 0.9184; stratum moleculare: F (2,15) ϭ 0.1433, p ϭ 0.8676; Fig. 3B).

Discussion
With the etiology of AD still unknown, it is important to investigate known risk factors and pathologic changes associated with the disease. The axonal enrichment of tau may deteriorate during the pathogenesis of AD as tau appears to accumulate in the somatodendritic compartment. Considering that aging is the leading risk factor for AD, investigation into the normal aging process may lead to the discovery of previously unappreciated anomalous features that contribute to disease vulnerability. Thus, we investigated the integrity of the AIS, the barrier involved in maintaining axonal localization of tau, over the span of aging in rats to establish whether normal aging might affect the AIS structure and/or tau distribution.
We evaluated age-related changes in tau and the AIS using multiple markers and complementary approaches. However, we found no evidence that the levels of AIS proteins or tau proteins change with age in the HP, nor levels of somatic or axonal tau change in HP neurons with advancing age in rats. The length of the AIS was not found to change in any region of the HP, but a discrete and specific effect was detected in the CA2 region (ϳ8% decrease from young to old). The reduction of AIS volume and diameter without a change in length indicates possible atrophy the axonal projections specifically in CA2 neurons, but biological significance of these changes in Figure 3. Optical density measurement of tau immunofluorescence in the somata and axons of HP neurons shows no changes across age. A, The intensity of somatic tau (using Tau7, a total tau antibody) in neurons in the CA1, CA2, CA3, and dentate regions of the HP shows no change with advancing age. B, No change in the regional intensity of tau is detected in the axon-enriched strata of the HP (i.e., the fimbria, alveus, stratum lacunosum, and stratum moleculare). All data are displayed as mean Ϯ SEM. C-E, Representative images of CA1 HP neurons positive for Tau7 at 4 (C), 14 (D), and 24 (E) months. The red rectangles are examples of the areas that fluorescence intensities were measured within individual neurons. F, Image illustrating the regions (in red) used for analysis of somatic tau intensity in the dentate gyrus. G-I, Enlargements represent the analyzed regions of 4-month-old (G), 14-month-old (H), and 24-month-old (I) rats. Scale bar, 20 m. the context of tau distribution and AIS functionality remain unclear. Overall, the findings presented here provide a strong case against aging-related changes in the total tau levels, somatic tau levels, some phosphorylated forms of tau (i.e., PHF1, AT8, pS422), levels of AIS structural proteins (i.e., AnkG, neurofascin, ␤IV-spectrin), or AIS morphology within the HP, thus, these factors are unlikely to contribute to the risk of developing AD. Importantly, these findings do not rule out that other variables related to tau and/or the AIS (e.g., other forms of tau, other characteristics/functions of the AIS, etc.) are changed during normal aging and might contribute to susceptibility for AD.
Before this work, a single study assessed the aging changes in the AIS and showed changes in ankyrin and spectrin proteins using Western blottings of telencephalic tissue from aged mice (Bahr et al., 1994). The discordant findings reported here could be due to the use of rats instead of mice, the methods used to immunoblot ankyrin or the increased specificity of antibodies currently available to the isoforms of proteins localized to the AIS (i.e., ankyrin vs AnkG 480/270 kDa or spectrin vs ␤IV-spectrin).
Multiple complimentary experimental approaches were used in the evaluation of the effects of aging on both tau and the AIS. Examination of aging-related changes in total tau protein levels was conducted through Western blotting using multiple tau markers, however, these methods do not allow cell-specific measurements. To assess aging-related changes specifically within levels of somatic tau we used fluorescence intensity measurements within the soma of individual HP neurons. Since tau is a microtubule-associated protein, we wanted to establish whether changes in tubulin content was altered with advancing age, but there were no age-related changes in total tubulin levels and the amount of tau per tubulin remained constant within the HP, which aligns well with previous studies showing HP neurons are not lost in normal aging (both in humans and in rats; West et al., 2004;Stanley et al., 2012). These findings demonstrate that the total amount of tau and the amount of tau in the soma remain unchanged in the HP of aging rats.
The same level of rigor was applied to the investigation of the AIS by using a multifaceted approach to evaluate changes in aging rats. Multiple key AIS proteins (AnkG, ␤IV-spectrin, and neurofascin) were examined in Western blottings, and we measured SMI312 levels to determine if total axonal content was changed. The lack of aging changes in all AIS proteins assessed and SMI312 suggest that axonal content and key structural components of the AIS are not altered with age. This conclusion is further supported by the lack of changes in tubulin described above. Stereological analysis of the AIS was conducted using two complementary approaches. The space balls probe is an established methodology for detecting agingrelated changes in fiber structures within the brain (Ypsilanti et al., 2008), and here we used it to measure AIS length. The AutoNeuron method in Neurolucida was used to perform 3D measurements of individual AISs and showed a lack of age-related changes in AIS length and only a mild, region-specific decrease in width and volume. This probe has been used previously to measure the diameter of dendrites in pyramidal neurons and compare them across brain regions (Amatrudo et al., 2012).The unbiased stereological measurement of AIS length in the entire HP is a significant advantage of this approach to analyze AIS morphology. The length of the AIS was previously measured with ␤IV-spectrin using a nonstereological method in selected z-stacks of the HP (Baalman et al., 2013). Interestingly, they report that a shortening of AISs correlates to cognitive impairment following explosioninduced brain trauma in Sprague Dawley rats. This work suggests that a shortening of the AIS may correlate to cognitive impairment, but we did not detect a shortening of AISs with age in the current work. Collectively, these data strongly demonstrate that the gross structure of the AIS (i.e., length, diameter, volume) do not change during the normal aging process in most cells of the HP.
Although we are reporting mostly negative data, it is unlikely that the Fischer 344 rat was not an appropriate model of aging to use for this study. We included age groups that span the spectrum from young (4 months), middle (14 months), and old (24 months), which is near the end of the normal lifespan and a time at which agingrelated impairments in memory and cognition occur. For example, an age-related decline in learning ability at 24 months of age was observed using a 14-unit T-maze and shock-motivated one-way active avoidance test in Fischer 344 rats (Spangler et al., 1994). Additionally, 24month-old Fischer 344 rats exhibit pathologies in multiple organ systems associated with advanced aging, further supporting their utility as a model of aging. We confirmed that our techniques and methods are capable of detecting other changes in normal aging that were previously reported (i.e., GFAP increase; VanGuilder et al., 2011).
Several studies showed that there is heterogeneity among aged animals on a number of learning and memory functions (e.g., water maze), as well as other behavioral tasks, indicating that individual animals respond to the aging process differently (Gage et al., 1984;Collier and Coleman, 1991;Freeman et al., 2009). These studies, made it clear that aged animals can be separated into unimpaired or impaired groups based on performance and that these changes correspond to some neurochemical and neuroanatomical changes. However, the data reported here demonstrate that two distinct populations of aged animals do not exist in regard to the specific AIS and tau parameters studied, suggesting these variables do not underlie the behavioral and cognitive decline seen in aging Fischer 344 rats. Analysis of these parameters in young, middle, and old humans may be necessary to definitively establish whether the same holds true for humans.
The negative findings presented here provide important insight into the aging-related changes in tau and AIS in the HP. This information is important considering the contention in the field that tau mislocalization is important in AD and the recent focus on the role of the AIS in tau distribution (Zempel et al., 2010;Li et al., 2011;Sohn et al., 2016). Thus, future investigations should focus on alternative aspects of tau and the AIS to further identify whether aging-related changes may contribute to the risk of AD.