Phenotypical and functional heterogeneity of neural stem cells in the aged hippocampus

Abstract Adult neurogenesis persists in the hippocampus of most mammal species during postnatal and adult life, including humans, although it declines markedly with age. The mechanisms driving the age‐dependent decline of hippocampal neurogenesis are yet not fully understood. The progressive loss of neural stem cells (NSCs) is a main factor, but the true neurogenic output depends initially on the actual number of activated NSCs in each given time point. Because the fraction of activated NSCs remains constant relative to the total population, the real number of activated NSCs declines in parallel to the total NSC pool. We investigated aging‐associated changes in NSCs and found that there are at least two distinct populations of NSCs. An alpha type, which maintains the classic type‐1 radial morphology and accounts for most of the overall NSC mitotic activity; and an omega type characterized by increased reactive‐like morphological complexity and much lower probability of division even under a pro‐activation challenge. Finally, our results suggest that alpha‐type NSCs are able to transform into omega‐type cells overtime and that this phenotypic and functional change might be facilitated by the chronic inflammation associated with aging.


IFN-α administration
IFN-α (Mintenyi Biotec, REF: 130-093-131) was diluted in sterile saline and administered intraperitoneally at 4x10 5 IU/Kg. All mice received one single injection of IFN-α (at the same hour everyday), or saline, during 20 days and sacrificed 1day after the last injection in the short-term analysis or 1 month after the last injection of IFN in the long-term experiment.

Minocycline administration
Mynoclycline (Minocycline hydrochloride; Sigma, St Louis, MO, USA, REF: M9511) was diluted in the drinking drinking at a concentration of 0.533 mg/ml as mouse drank 6 ml within a day they would receive a dose of 100 mg/kg/day. Control mice received normal water without any addition.
Mice received minocycline in their drinking water for a total of 30 days. The amount of liquid intake was measured daily.

Immunohistochemistry
Experiments were performed essentially as described before following methods optimized for the use in transgenic mice (Encinas et al. 2006;Encinas and Enikolopov 2008;Encinas et al. 2011). Animals were deeply anesthetized and were subjected to transcardial perfusion with 30 ml of PBS followed by 30 ml of 4% (w/v) paraformaldehyde in PBS, pH 7.4. The brains were removed, cut longitudinally into two hemispheres and postfixed, with the same fixative, for 3 hr at room temperature, then transferred to PBS and kept at 4 o C. Serial 50 μm-thick sagittal sections were cut using a Leica VT 1200S vibrating blade microtome (Leica Microsystems GmbH, Wetzlar, Germany). Immunostaining was carried out following a standard procedure: the sections were incubated with blocking and permeabilization solution (PBS containing 0.25% Triton-100X and 3% BSA) for 3hr at room temperature, and then incubated overnight with the primary antibodies (diluted in the same solution) at 4 o C. After thorough washing with PBS, the sections were incubated with fluorochrome-conjugated secondary antibodies diluted in the blocking and permeabilization solution for 3 hr at room temperature. After washing with PBS, the sections were mounted on gelatin coated slides with DakoCytomation Fluorescent Mounting Medium (DakoCytomation, Carpinteria, CA). Those sections destined to the analysis of BrdU incorporation were treated, before the immunostaining procedure, with 2N HCl for 20 min at 37 o C, rinsed with PBS, incubated with 0.1M sodium tetraborate for 10 min at room temperature, and then rinsed with PBS. The GFP signal from the transgenic mice was detected with an antibody against GFP for enhancement and better visualization. The following antibodies were used: chicken anti-GFP (Aves Laboratories, Tigard, OR) at 1:1000 dilution; rabbit anti-Ki67 (Vector Laboratories, Burlingame,CA, USA) at 1:1000; rabbit anti-GFAP (Dako Cytomation) at 1:1000; rabbit anti-S100β (DakoCytomation) at 1:500; rat anti-BrdU (AbD Serotech, Kidlington, UK) at 1: 400; AlexaFluor 488 goat anti-chicken (Molecular Probes, Willow Creek Road, Eugene, OR) at 1:500; AlexaFluor 647 goat anti-rabbit (Molecular Probes) at 1:500; AlexaFluor 568 goat anti-rat (Molecular Probes) at 1:500; DAPI, at 1:1000 (Sigma) was used at counterstaining when required.

Image capture
All fluorescence immunostaining images were collected employing a Leica SP8 (Leica, Wetzlar, Germany) laser scanning microscopes and their corresponding manufacturer's software. The signal from each fluorochrome was collected sequentially, and controls with sections stained with single fluorochromes were performed to confirm the absence of signal leaking into different channels and antibody penetration. All images were imported into Adobe Photoshop 7.0 (Adobe Systems Incorporated, San Jose, CA) in tiff format. Brightness, contrast, and background were adjusted equally for the entire image using the "brightness and contrast" and "levels" controls from the "image/adjustment" set of options without any further modification. Al images shown are projections from z-stacks ranging from 10 (typically for individual cell images) to 20 microns of thickness.

Cell quantification
Quantitative analysis of cell populations (proliferation) in Nestin-GFP mice was performed by design-based (assumption free, unbiased) stereology using a modified optical fractionator sampling scheme as previously described (Encinas et al. 2004;Encinas and Enikolopov 2008;Encinas et al. 2011). Slices were collected using systematic-random sampling. The hemisphere was sliced sagittally in a lateral-to-medial direction, from the beginning of the lateral ventricle to the middle line, thus including the entire DG. The 50 μm slices were collected in 5 parallel sets, each set consisting of 14 slices, each slice 300 μm apart from the next. All BrdU cells per slice were counted, with a 63x oil immersion objective, to obtain absolute numbers of BrdU cells. α-NSCs were defined as radial glia-like cells positive for Nestin-GFP and GFAP with the soma located in the SGZ or the lower third of the GCL and with a process extending from the SGZ towards the molecular layer through the GCL. Ω-NSCs were Nestin-GFP and GFAP positive with a multibranches phenotype and the soma placed out of SGZ. The relative proportions of BrdU-positive NSCs types were referred to the total number of NSCs quantified per animal. For the absolute number of NSCs, the number of NSCs, excluding those in the uppermost focal plane, was counted in 100 μm-wide, 50 μm-tall 12 μm-deep z-stacks, in GFP and GFAP stained slices from Nestin-GFP mice, using a 63x oil immersion objective. At least 4 z-stacks were obtained from each slice. The values were normalized to the total volume of the SGZ+GCL for each animal. The total volume was obtained by measuring the area of the SGZ+GCL in each slice and measuring the thickness of each slice in at least 3 points. To measure the morphological changes in NSCs, at least 50 cells were randomly selected form 20 μm-thick z-stacks taken form Sal, LKA, IFN-α mice brain sections immunostained for GFP and GFAP. Primary processes were considered to be those emerging directly from the soma, and the secondary processes those emerging from the primary processes in the 30 μm closest to the soma.The area of DG in each z-stack was quantified using the Fiji Is Just ImageJ (Fiji) distribution of ImageJ, Using the total hippocampal volume, estimated in low magnification images of the whole series taken in a confocal microscope using Las AF lite.

Sholl analysis
Sholl analysis is an open-source plug-in for FIJI (Image J, Schindelin et al.2012) this plug-in performs the Sholl technique directly on 2D or 3D images of fluorescence labeled cells. It is based on an algorithm to retrieve data from pixel-based connectivity ( detailed in the user guide of http://fiji.sc/Sholl). For the morphological analysis of NSCs we obtained 3D reconstructions from confocal stack images(1024 pixels of resolution) . Single NSCs were analyzed using 3D Sholl analysis plugin (http://fiji.sc/Sholl_Analysis) as described in (Ferreira et al., 2014). Z-stack from Nestin-GFP/GFAP positive NSCs were collected in a random manner.Only entire NSCs were choosen to analize the complexity of the cells in the analisys. The thickness of each zstack depends on the NSCs, as the image must include all the cell body and the all the volumen ocupied by the branches and arborizations. Using the tool "polygon" of the Image J the outline of the celsl was delimited. Next, using the threshold we elaborate a mask to remove the background from the delimited image obtaining a single image which corresponds with our NSCs. Finally, using the tool "line" we draw a line from the middle of the some to the farthest arborization or segment of the NSC to stablish the lenght of the NSCs and the limit for the analysis.

Statistical analysis.
SigmaPlot (San Jose, CA, USA) was used for statistical analysis. 1-way ANOVA test was performed to determine the effect of the factor (Figure 1-3). In all cases, all pair-wise multiple comparisons (Holm-Sidak method or Dunn ś) were set as a post-hoc test to determine the significance between groups in each factor. For analysis of pairs of groups (Figure 4-6), a Student ś t test was performed. Hierarchical clustering was performed using Ward's method and squared Euclidean distances as linkage metric. For the Sholl analysis two-way repeated measures ANOVA followed by Bonferroni post-hoc test was performed. Only p<0.05 is reported to be significant. Data are shown as mean ± SEM (standard error of the mean). For the IFN-α experiment 2-way ANOVA was performed to analyzed the interaction between factors (age x treatment), no interactions were found. The fitting of linear and nonlinear regression models to data was compared in GraphPad Prism 5 (GraphPad Software, Incl, San Diego, CA) using Akaike's information criterion with correction for finite sample sizes (AICc) (Hurvich and Tsai, 1993). The exponential growth equation showed the best fit based on AICc with a normal distribution of the residuals, analyzed using the Shapiro Wilk normality test. In addition, correlation of these data was analyzed using the Pearson correlation coefficient.