Abca7 deletion does not affect adult neurogenesis in the mouse

ATP-binding cassette transporter A7 (ABCA7) is a member of the “A” subfamily of ATP-binding cassette transporters that were initially characterized by their capacity to transport lipids and other lipophilic molecules across membranes [1,2]. ABCA7 is expressed in the brain where it appears to be predominantly localized in microglia and, to a lesser degree, in neurons and oligodendrocytes [3]. Several lines of evidence implicate a role for ABCA7 in the regulation of Alzheimer's disease (AD) risk and amyloid pathology. Data from human genome-wide association studies (GWAS) indicate ABCA7 is a risk factor for late-onset AD [4–9]. Additional genetic studies confirm the important association of ABCA7 single nucleotide polymorphisms (SNPs) and methylation changes with AD [7–9], and recent work highlights that loss-of-function ABCA7 variants confer increased AD risk [10]. This is in agreement with data showing that the AD-associated ABCA7 SNP (rs3764650 that does not result in an amino acid change) is associated with reduced ABCA7 expression levels in AD cases [11].

In vitro studies revealed that ABCA7 transfection into cell lines constitutively expressing human amyloid-β precursor protein (APP) results in a significant reduction in production of amyloid-β peptide (Aβ) peptides [12], whereas siRNA-mediated knockdown of ABCA7 increased Aβ production under similar experimental conditions [13]. In addition, phagocytic clearance of oligomeric Aβ was found to be impaired in ABCA7-deficient macrophages compared with wild-type (WT) macrophages [14]. In general agreement with these cell culture studies, work in AD mouse models has also provided evidence supporting a role for ABCA7 in amyloid homoeostasis. In the J20 AD transgenic mouse model, deletion of ABCA7 led to a doubling of insoluble Aβ levels and amyloid plaques in mice assessed at 17 months of age [14]. An independent study using the same Abca7 null (Abca7^−/−) mouse line but crossed with the TgCRND8 AD mouse model revealed an increase in the area of dense plaques at 18 weeks of age that the authors concluded was likely to be a reflection of the increased production of Aβ in the TgCRND8-Abca7^−/− mice [13]. It should be noted that even in the absence of the AD transgene expression, Abca7^−/− null mice were found to exhibit an age-dependent increase in the levels of the pathogenic (endogenous) Aβ42 peptide in the brain; assessed up to 6 months of age [13].

Although it is clear that ABCA7 has the capacity to regulate amyloidogenesis in mice, an entirely different question that has not been addressed centres on the potential impact that ABCA7 loss of function may have on adult neurogenesis. Our rationale for asking this question is based on previous studies showing that other ABC transporters including two that, like ABCA7, are also from the “Class-A” subfamily (i.e. ABCA2, ABCA3) appear to play a role in neurogenesis [15,16]. These studies along with the extensive literature highlighting the importance of adult neurogenesis in AD (see [17–20], for recent comprehensive reviews of this area) prompted our present study in which we have assessed cellular proliferation and neurogenesis in the dentate gyrus (DG) and subventricular zone (SVZ) of both WT and Abca7^−/− adult male mice.

Abca7-deficient mice were generated and genotyped as described previously [21]. All mice were on a C57BL/6j background and backcrossed >15 times. Test mice were males at ∼8.5 months of age (WT, n=7, 251±7 days old; Abca7^−/−, n=7, 256±9 days old). The animals we refer to as “WT” control mice were littermates derived from the same breeders that were used to generate the Abca7^−/− animals (i.e. both test cohorts were from Abca7^+/− breeders). Animal ethics approval was from the University of Wollongong Animal Ethics Committee.

5-Bromo-2-deoxyuridine (BrdU, Sigma, B5002) was administered via intraperitoneal injection (i.p.) at a dose of 50 mg/kg twice a day (8 a.m. and 8 p.m.) for three consecutive days prior to euthanasia of animals. The animals were killed by methoxyflurane inhalation 4 h after the final BrdU injection and perfused with 0.9% NaCl followed by 4% paraformaldehyde (PFA). Tissues were preparation as described previously [22]. In brief, brains were removed and post-fixed in phosphate-buffered 4% PFA at 4°C for 16 h then cryoprotected in 30% (w/v) sucrose in 0.1 M phosphate buffer for 48 h. One hemibrain was coronally sectioned at 40 μm with a cryostat (CM1950, Leica Microsystems), the sections were collected in six series and stored at −20°C in cryoprotectant (3:3:4, glycerol:ethylene glycol:0.1 M phosphate buffer, v/v/v) for histological or immunohistochemical analysis. Every seventh hemibrain section was used for neurogenesis detection as described previously [23,24]. Briefly, after cryoprotectant removal, the sections were post-fixed with Bouin-4 [4% (w/v) PFA, 1% (w/v) picric acid and 5% (v/v) acetic acid, to improve nuclear antigen unmasking] for 15 min, then processed in sodium borohydride (0.1 mg/ml in 0.01 M PBS, to remove excess reactive aldehyde groups), denaturation (2 M HCl, 22°C, 1 h, then neutralization with 0.1 M borate buffer, pH 8.5) and H₂O₂ quenching [1% (v/v) H₂O₂ in PBS, 22°C, 30 min to quench endogenous peroxidase activity]. After blocking in Vector MOM reagent (Vector), the sections were stained with rat monoclonal anti-BrdU (0.5 μg/ml; Abcam, AB6326, from Sapphire Bioscience Pty Ltd) and biotinylated with goat-anti-rat IgG (1:200, Vector Laboratories), followed by Vectastain Elite ABC reagent (Vector Laboratories) or Streptavidin-HRP (Sigma, S2438 1:2000) and staining was visualized with ImmPact diaminobenzidine (DAB) peroxidase substrate (Vector Laboratories). To confirm the specificity of primary antibodies, control experiments were performed where sections were incubated for 16 h in the absence of primary antibody. Doublecortin (DCX), a commonly used marker for neuroblast cells [25,26] in the adult brain, was detected using rabbit anti-DCX (1:2000, Abcam, Ab18723, from Sapphire Bioscience Pty Ltd) as described as above for the BrdU protocol except the denaturation and MOM blocking steps were omitted and the secondary antibody used was biotinylated goat-anti-rabbit IgG (Sigma, B7389, 1:2000). All the staining procedures were carried out within a single immunostaining session to minimize variability.

Immunohistochemically stained sections were captured using a Scanscope XT Image scanner at 20× (Scanscope Console V10.2.0352, Aperio Digital Pathology System, Aperio Technologies). The images were exported in tiff format from Imagescope software (Aperio Technologies), and analysed using ImageJ software (http://rsbweb.nih.gov/ij/) using an established quantitative method [27]. Representative images were assembled with Photoshop CS2 for presentation in the figures. The numbers of BrdU +ve and DCX +ve cells in the DG were directly counted in the immunostained sections (every seventh coronal section, five sections per mouse) in a blinded manner under 20× magnification. This was achieved using tiff files generated from the Scanscope XT Image scanned images and the data were expressed as positive cell number per mm². Due to the clustering of BrdU +ve cells in the SVZ and the overlap of neuronal structures in SVZ DCX staining, it was impractical to count absolute cell numbers and therefore the data were only acquired as positive stained occupied area (μm²) for this region of interest in each section (as described below).

To quantify BrdU +ve and DCX +ve staining, five coronal sections per brain were collected, 240 μm apart, grossly corresponding to sections 1.2–0.0 mm anterior to the Bregma (+1.2–0.0) according to the mouse brain atlas [28], were selected to measure the area occupied by positive staining in a defined region of interest area of the SVZ, i.e., the entire lateral wall of SVZ from corpus callosum to the anterior commissure (selected area: approximately 0.8–1.2 mm², mean ± S.E.M., 0.94 ±0.07 mm²); and for measurement of hippocampal neurogenesis, five coronal sections spanning from 1.7 to 2.7 mm posterior to the Bregma (−1.7 to −2.7) were chosen to analyse the area occupied by positive staining in the known neurogenic region, i.e. the DG. This region includes the molecular layer, outer granule cell layer, subgranular zone (the most neurogenic region of the hippocampus) and hilus in the hippocampus formation (selected area approximately 0.8–1.2 mm², mean ± S.E.M.: 0.93±0.11 mm²).

Quantitative data are presented as mean ± S.E.M. (represented by the error bars). Statistical differences were assessed using the Student's t test where P<0.05 was considered significant.

BrdU staining was used to assess cell proliferation in the DG of both WT and Abca7^−/− mice, whereas DCX staining was used as a marker of neurogenesis. Consistent with previous studies, we detected strong BrdU and DCX staining in the hippocampus (Figure 1). The immunohistochemical labelling was detected mainly in the subgranular zone of the DG. Evidence for neurogenesis was also clearly detected in the SVZ (Figure 2). Active cellular proliferation in the SVZ was detected by BrdU cluster-like staining within the wall of lateral ventricle (Figure 2E). Fine neuronal processes with bifurcations were evident in the SVZ when DCX immunostaining was applied (Figure 2F). We did not observe any gross (qualitative) differences to overall structure of proliferating cells when WT and Abca7^−/− mice were compared. Direct counting of BrdU +ve cells and DCX +ve cells in the DG indicated that Abca7^−/− loss did not significantly affect neurogenesis is this location (Figure 3). Furthermore, a quantitative evaluation of BrdU and DCX staining area in the DG and SVZ indicated there were no significant differences in neurogenesis comparing WT and Abca7^−/− mice (Figure 4).

Although we cannot rule out the possibility that glial progenitors are affected by Abca7 loss, this seems unlikely given the lack of significant change in BrdU incorporation in both the DG and SVZ. Also, our findings do not preclude the possibility that ABCA7 contributes to embryonic neurogenesis as this has not been assessed in the present study.

There is mounting evidence that ABCA7 plays a role in regulating the pathways leading to AD [29]. Studies in cell lines and mouse models suggest that a loss of ABCA7 function may have deleterious consequences for Aβ homoeostasis in the brain [12–14]. This is consistent with the previously reported significant association of ABCA7 SNPs with amyloid plaque load in human subjects [6]. Indeed, recent genetics studies have shown that loss-of-function ABCA7 variants confer increased AD risk [10]. Even though the available evidence points towards a role for ABCA7 regulating Aβ homoeostasis, the true function of ABCA7 in the brain is still an open question. It is therefore important to investigate additional pathways that may be influenced by ABCA7 in order to understand its true function(s) in the AD context. An additional rationale for the present study is the literature suggesting certain ABC transporters play a role in the regulation of neurogenesis. In this context, ABCA2, ABCA3, ABCB1 and ABCG2 have been suggested to play a role in neurogenesis [15]. Furthermore, ABCA2 is a marker of neural progenitors in the adult mouse brain [16]. Despite the structural similarities of ABCA7 with these proteins, and the fact that ABCA7 is expressed at a relatively high level in the brain [2], the results from our current study suggest that ABCA7 does not play a major role in the regulation of cell proliferation or neurogenesis in the DG or SVZ of adult mice.

It is also worth considering how loss of ABCA7 function may impact adult neurogenesis in the AD setting. Although the roles that APP proteolytic fragments may play in regulating neurogenesis are far from clear [17–20,30,31], there is evidence that Aβ42 may inhibit adult neurogenesis [32–34]. Indeed, as previously suggested [17], the majority of AD mouse models that are based on APP mutations exhibit reduced neurogenesis. It might therefore be predicted that in an environment where ABCA7 is required to remove a toxic insult (e.g. phagocytic clearance of Aβ peptides that would otherwise inhibit neurogenesis), an indirect effect due to the loss of ABCA7 function may become apparent. These types of hypotheses are difficult to test experimentally as it would be hard to separate the effects of the toxic insult from ABCA7 function per se. Nonetheless, it remains possible that ABCA7 may have an impact on neurogenesis under certain conditions that were not modelled in our current assessment. Based on our current data, however, it is reasonable to assume that ABCA7 is unlikely to play a major direct role in adult neurogenesis in humans. In conclusion, our current BrdU incorporation and DCX labelling experiments indicate that deletion of Abca7 does not have a significant impact on adult neurogenesis in mice.

Brett Garner, Tim Karl and Hongyun Li conceived and designed the experiments. Brett Garner wrote the first draft of the manuscript. Hongyun Li performed the experiments. All authors analysed the data and edited the manuscript for submission.

We thank Professor Mason W. Freeman (Massachusetts General Hospital, Boston) for providing the Abca7 null mice.

This work was supported by the National Health and Medical Research Council (NHMRC) of Australia Research Fellowships [grant numbers 1045643 (to T.K.) and 1109831 (to B.G.)].

Aβ

amyloid-β peptide

Abca7^−/−

Abca7 gene knockout mouse

ABCA7

ATP-binding cassette transporter, subfamily A, member 7

AD

Alzheimer's disease

APP

amyloid-β precursor protein

BrdU

bromodeoxyuridine

DCX

doublecortin

DG

dentate gyrus

PFA

paraformaldehyde

SNP

single nucleotide polymorphism

SVZ

subventricular zone

WT

wild-type-like