ATP-binding cassette protein ABCA7 deficiency impairs sphingomyelin synthesis, cognitive discrimination, and synaptic plasticity in the entorhinal cortex

Sphingomyelin (SM) is an abundant plasma membrane and plasma lipoprotein sphingolipid. We previously reported that ATP-binding cassette family A protein 1 (ABCA1) deficiency in humans and mice decreases plasma SM levels. However, overexpression, induction, downregulation, inhibition, and knockdown of ABCA1 in human hepatoma Huh7 cells did not decrease SM efflux. Using unbiased siRNA screening, here, we identified that ABCA7 plays a role in the biosynthesis and efflux of SM without affecting cellular uptake and metabolism. Since loss of function mutations in the ABCA7 gene exhibit strong associations with late-onset Alzheimer’s disease across racial groups, we also studied the effects of ABCA7 deficiency in the mouse brain. Brains of ABCA7-deficient (KO) mice, compared with WT, had significantly lower levels of several SM species with long chain fatty acids. In addition, we observed that older KO mice exhibited behavioral deficits in cognitive discrimination in the active place avoidance task. Next, we performed synaptic transmission studies in brain slices obtained from older mice. We found anomalies in synaptic plasticity at the intracortical synapse in layer II/III of the lateral entorhinal cortex but not in the hippocampal CA3-CA1 synapses in KO mice. These synaptic abnormalities in KO brain slices were rescued with extracellular SM supplementation but not by supplementation with phosphatidylcholine. Taken together, these studies identify a role of ABCA7 in brain SM metabolism and the importance of SM in synaptic plasticity and cognition, as well as provide a possible explanation for the association between ABCA7 and late-onset Alzheimer’s disease. Sphingomyelin (SM) is an abundant plasma membrane and plasma lipoprotein sphingolipid. We previously reported that ATP-binding cassette family A protein 1 (ABCA1) deficiency in humans and mice decreases plasma SM levels. However, overexpression, induction, downregulation, inhibition, and knockdown of ABCA1 in human hepatoma Huh7 cells did not decrease SM efflux. Using unbiased siRNA screening, here, we identified that ABCA7 plays a role in the biosynthesis and efflux of SM without affecting cellular uptake and metabolism. Since loss of function mutations in the ABCA7 gene exhibit strong associations with late-onset Alzheimer’s disease across racial groups, we also studied the effects of ABCA7 deficiency in the mouse brain. Brains of ABCA7-deficient (KO) mice, compared with WT, had significantly lower levels of several SM species with long chain fatty acids. In addition, we observed that older KO mice exhibited behavioral deficits in cognitive discrimination in the active place avoidance task. Next, we performed synaptic transmission studies in brain slices obtained from older mice. We found anomalies in synaptic plasticity at the intracortical synapse in layer II/III of the lateral entorhinal cortex but not in the hippocampal CA3-CA1 synapses in KO mice. These synaptic abnormalities in KO brain slices were rescued with extracellular SM supplementation but not by supplementation with phosphatidylcholine. Taken together, these studies identify a role of ABCA7 in brain SM metabolism and the importance of SM in synaptic plasticity and cognition, as well as provide a possible explanation for the association between ABCA7 and late-onset Alzheimer’s disease. Sphingolipids are biologically active lipids involved in several biological processes (1Hannun Y.A. Obeid L.M. Principles of bioactive lipid signalling: lessons from sphingolipids.Nat. Rev. Mol. Cell Biol. 2008; 9: 139-150Crossref PubMed Scopus (2545) Google Scholar, 2Bikman B.T. Summers S.A. Ceramides as modulators of cellular and whole-body metabolism.J. Clin. Invest. 2011; 121: 4222-4230Crossref PubMed Scopus (311) Google Scholar). Sphingomyelin (SM), a major sphingolipid, is abundant in the outer leaflet of cell membranes where it forms lipid rafts with cholesterol. Additionally, SM is the most abundant sphingolipid in plasma lipoproteins (3Hammad S.M. Blood sphingolipids in homeostasis and pathobiology.Adv. Exp. Med. Biol. 2011; 721: 57-66Crossref PubMed Scopus (40) Google Scholar). We previously showed that microsomal triglyceride transfer protein (MTP) and ATP-binding cassette transporter family A protein 1 (ABCA1) play important roles in determining plasma concentrations of SM (4Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Tarugi P. Hegele R.A. et al.Microsomal triglycerdie transfer protein transfers and determines plasma concentrations of ceramide and sphingomyelin but not glycosylceramide.J. Biol. Chem. 2015; 290: 25863-25875Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar, 5Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Rader D.J. Hussain M.M. ATP binding cassette family A protein 1 determines hexosylceramide and sphingomyelin levels in human and mouse plasma.J. Lipid Res. 2018; 59: 2084-2097Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar). Mechanistic studies showed that both of these proteins do not affect SM synthesis. It is known that MTP transfers lipids in vitro between vesicles and possibly to apoB in vivo to assist in lipoprotein assembly and secretion (6Hussain M.M. Shi J. Dreizen P. Microsomal triglyceride transfer protein and its role in apolipoprotein B-lipoprotein assembly.J. Lipid Res. 2003; 44: 22-32Abstract Full Text Full Text PDF PubMed Scopus (452) Google Scholar, 7Sirwi A. Hussain M.M. Lipid transfer proteins in the assembly of apoB-containing lipoproteins.J. Lipid Res. 2018; 59: 1094-1102Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar). We showed that MTP transfers SM in vitro and suggested that it might also do so in cells and facilitate secretion of SM with lipoproteins to explain significant reductions in plasma SM levels in MTP-deficient humans and mice (4Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Tarugi P. Hegele R.A. et al.Microsomal triglycerdie transfer protein transfers and determines plasma concentrations of ceramide and sphingomyelin but not glycosylceramide.J. Biol. Chem. 2015; 290: 25863-25875Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar). ABCA1 is involved in the efflux of phospholipids and cholesterol to high-density lipoproteins (HDLs) and its deficiency in humans and mice is associated with low levels of plasma HDL. Although chronic deficiency of ABCA1 significantly reduced plasma SM levels in humans and mice, acute knockdown or inhibition had no effect on SM efflux from human hepatoma cells (5Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Rader D.J. Hussain M.M. ATP binding cassette family A protein 1 determines hexosylceramide and sphingomyelin levels in human and mouse plasma.J. Lipid Res. 2018; 59: 2084-2097Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar). Therefore, we surmised that ABCA1 plays a role in the biogenesis of HDL, and HDL is involved in the efflux of SM from cells involving an unexplained mechanism. Here, using an unbiased siRNA-mediated knockdown (KD) approach, we identified that ABCA7 plays a role in the biosynthesis and efflux of SM. Since ABCA7 gene has been associated with late-onset Alzheimer’s disease (LOAD) (8Reitz C. Jun G. Naj A. Rajbhandary R. Vardarajan B.N. Wang L.S. et al.Variants in the ATP-binding cassette transporter (ABCA7), apolipoprotein E ε4,and the risk of late-onset Alzheimer disease in African Americans.JAMA. 2013; 309: 1483-1492Crossref PubMed Scopus (299) Google Scholar), we addressed the role of ABCA7 in the brain. We observed that ABCA7 knockout (Abca7−/−) mice have reduced levels of certain species of SM in the brain and that ABCA7 deficiency affects cognitive discrimination and synaptic plasticity in brain slices which can be mitigated by supplementation of SM. We previously reported that ATP-binding cassette family A protein 1 (ABCA1) deficiency in humans and mice decreases plasma SM levels (5Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Rader D.J. Hussain M.M. ATP binding cassette family A protein 1 determines hexosylceramide and sphingomyelin levels in human and mouse plasma.J. Lipid Res. 2018; 59: 2084-2097Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar). However, overexpression, induction, downregulation, inhibition, and KD of ABCA1, a well-characterized cholesterol and phospholipid transporter, in human hepatoma Huh7 cells does not decrease SM efflux to HDL (5Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Rader D.J. Hussain M.M. ATP binding cassette family A protein 1 determines hexosylceramide and sphingomyelin levels in human and mouse plasma.J. Lipid Res. 2018; 59: 2084-2097Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar). Therefore, we hypothesized that ABCA1 is not directly involved in the SM efflux to HDL but contributes to plasma SM levels by playing a role in the biogenesis of HDL. To identify the SM transporter, we individually knocked down 48 members of the ABC transporter family by using an siRNA library in Huh7 cells, which were treated with C6-NBD ceramide (Cer). Ceramide is taken up by cells and is converted to sphingomyelin (SM) and glucosylceramide (GluCer). After labeling, cells were incubated with bovine serum albumin (BSA) to quantify basal efflux (siCTRL-BSA) or efflux to HDL (siCTRL-HDL) (Fig. 1). Compared with BSA, HDL significantly increased SM efflux (Fig. 1, A and B), in agreement with findings from our earlier studies (5Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Rader D.J. Hussain M.M. ATP binding cassette family A protein 1 determines hexosylceramide and sphingomyelin levels in human and mouse plasma.J. Lipid Res. 2018; 59: 2084-2097Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar), thus confirming that HDL enhances SM efflux. KD of different genes either increased or decreased SM efflux to HDL (Fig. 1, A and B). Of interest was one protein, ABCA7, whose KD significantly decreased SM efflux to HDL (Fig. 1, A and B). We validated these results by performing a second screening. In agreement with the results of our first screening, incubation of siCTRL-treated cells with HDL resulted in significantly greater SM efflux than that with siCTRL-BSA, and this enhanced efflux significantly decreased after KD of ABCA7 (Fig. 1C). These results suggested that ABCA7 might be involved in the efflux of SM to HDL. In previous studies, we showed that MTP and ABCA1 deficiencies have no effect on SM synthesis; however, they were involved in secretion and efflux, respectively (4Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Tarugi P. Hegele R.A. et al.Microsomal triglycerdie transfer protein transfers and determines plasma concentrations of ceramide and sphingomyelin but not glycosylceramide.J. Biol. Chem. 2015; 290: 25863-25875Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar, 5Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Rader D.J. Hussain M.M. ATP binding cassette family A protein 1 determines hexosylceramide and sphingomyelin levels in human and mouse plasma.J. Lipid Res. 2018; 59: 2084-2097Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar). Therefore, we hypothesized that ABCA7 would not affect SM synthesis but would decrease efflux. To test this hypothesis, we transfected Huh7 cells with siABCA7 to knockdown (KD) ABCA7 by more than 80% compared to siCTRL-treated cells (Fig. 2, A–C). These cells were then incubated with C6-NBD Cer for 10 to 60 min. Measurement of intracellular sphingolipids revealed that cellular Cer levels increased with time, thus indicating time-dependent uptake (Fig. 2D). At the earliest time point studied, NBD-Cer had already been converted to GluCer, and the levels remained similar over 1 h in both siCTRL- and siABCA7-treated cells. In contrast, conversion of Cer to SM increased with time in both siCTRL- and siABCA7-treated cells. However, unexpectedly, these time-dependent increases were significantly lower in ABCA7 KD cells, thereby leading to lower intracellular SM levels than those in siCTRL-treated cells (Fig. 2D). These studies suggest that ABCA7 is involved in SM synthesis but not in the synthesis of GluCer. We considered the possibility that decreased synthesis of SM might have been an artifact due to the use of NBD-labeled C6-Cer as a precursor for SM synthesis. Therefore, we labeled siCTRL- or siABCA7-treated Huh7 cells with C16-[14C]-Cer and measured labeled sphingolipids in the media and cells. For an additional control, we performed KD of ABCA1, whose deficiency decreases plasma SM levels in mice and humans (5Iqbal J. Walsh M.T. Hammad S.M. Cuchel M. Rader D.J. Hussain M.M. ATP binding cassette family A protein 1 determines hexosylceramide and sphingomyelin levels in human and mouse plasma.J. Lipid Res. 2018; 59: 2084-2097Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar). First, we quantified the efflux of [14C]-sphingolipids to BSA and HDL in the media (Fig. 2E). We observed significantly greater levels of Cer, GluCer, and SM in the media of siCTRL-HDL– than siCTRL-BSA–treated cells. SiABCA1 had no significant effect on the efflux of Cer, GluCer, and SM to HDL. We did not observe any difference in the efflux of Cer to HDL between siABCA7-treated cells and siCTRL-HDL–treated cells. However, we observed significantly greater efflux of GluCer and significantly less efflux of SM to HDL in siABCA7-HDL than in siCTRL-HDL–treated cells. These studies indicated that siABCA7 significantly decreases [14C]-SM efflux to HDL. Next, we measured sphingolipids in cells (Fig. 2F). SiABCA1 and siABCA7, compared with siCTRL-HDL, had no significant effect on cellular Cer and GluCer levels. In contrast, siABCA7-treated cells had significantly lower levels of radiolabeled SM than siCTRL- and siABCA1-treated cells, thus indicating decreased synthesis. To determine whether ABCA7 might play a role in both synthesis and efflux or whether decreased efflux might be secondary to lower synthesis, we calculated the efflux as a percentage of the total SM level present in the cells and media (Table 1). SiABCA1 had no effect on Cer and SM efflux but increased GluCer efflux. In contrast, siABCA7 increased the efflux of Cer and GluCer but had no effect on SM efflux. Based on these calculations, we surmised that ABCA7 plays a role in SM synthesis, and decreased efflux is secondary to lower synthesis and lower cellular SM levels. Because of reduced SM synthesis in the absence of ABCA7, more Cer is likely to remain in cells, and conversion to GluCer is likely to be greater; therefore, these lipids should be available for efflux to HDL. Together, these studies indicate that KD of ABCA7 significantly decreases the conversion of [14C]-Cer to SM and the efflux of [14C]-SM to HDL.Table 1Efflux of different sphingolipids to HDL as a percentage of total lipidsCellular (AU)Media (AU)Total (AU)% Effluxp value[14C]-Ceramide CeramidesiCTRL-BSA7861 ± 8681447 ± 2149307 ± 78815.7 ± 2.9siCTRL-HDL8343 ± 9933791 ± 9812,134 ± 97831.4 ± 2.7siABCA1-HDL8312 ± 8423904 ± 12312,216 ± 90232.1 ± 2.1NSsiABCA7-HDL6504 ± 8074447 ± 35510,951 ± 73740.7 ± 4.2<0.05 GlucosylceramidesiCTRL-BSA5695 ± 297464 ± 1106123 ± 3477.6 ± 1.5siCTRL-HDL5601 ± 2881711 ± 3697312 ± 59923.2 ± 3.4siABCA1-HDL5553 ± 1313166 ± 8518719 ± 97835.9 ± 5.8<0.05siABCA7-HDL5779 ± 6003165 ± 1848945 ± 78435.4 ± 1.0<0.005 SphingomyelinsiCTRL-BSA5165 ± 3472008 ± 3337173 ± 67627.9 ± 2.1siCTRL-HDL4300 ± 2433674 ± 4097974 ± 35446.0 ± 3.6siABCA1-HDL4455 ± 3943878 ± 1108332 ± 48646.6 ± 1.7NSsiABCA7-HDL3326 ± 1842702 ± 2816028 ± 13444.8 ± 3.9NSData from Figure 2, B and C were used to calculate the percentage efflux with respect to total sphingolipid arbitrary units present in the media and cells at the end of the experiment. Statistical differences were calculated in comparison to siCTRL-HDL via Student’s t test.Abbreviations: AU, arbitrary units; NS, not significant. Open table in a new tab Data from Figure 2, B and C were used to calculate the percentage efflux with respect to total sphingolipid arbitrary units present in the media and cells at the end of the experiment. Statistical differences were calculated in comparison to siCTRL-HDL via Student’s t test. Abbreviations: AU, arbitrary units; NS, not significant. The above studies showed that ABCA7 is involved in SM synthesis. We also considered several other possible roles of ABCA7 in SM synthesis. It has been shown that ceramide transfer protein 1 (CERT1) is important for SM synthesis (9Kumagai K. Hanada K. Structure, functions and regulation of CERT, a lipid-transfer protein for the delivery of ceramide at the ER-Golgi membrane contact sites.FEBS Lett. 2019; 593: 2366-2377Crossref PubMed Scopus (43) Google Scholar). Therefore, we asked whether CERT1 expression is altered in Abca7 deficiency and measured ABCA7 and CERT1 mRNA levels in the livers of WT and Abca7−/−-deficient (KO) mice. Compared to WT, KO livers had barely detectable ABCA7 mRNA levels confirming ABCA7 deficiency (Fig. S1). In contrast, ABCA7 deficiency had no effect on CERT1 mRNA levels. These studies showed that ABCA7 deficiency has no effect on CERT1 expression. Next, we studied the uptake of NBD-labeled SM by these cells (Fig. 3, A–D). Both WT and KO hepatocytes took up similar amounts of NBD-SM fluorescence (Fig. 3B). Moreover, both cells showed similar uptake of NBD-SM (Fig. 3, A and C). We observed that there was an increase in NBD-Cer in cells with time (Fig. 3A). The production of ceramide was similar in WT and KO hepatocytes (Fig. 3, A and D). These studies showed that hepatocytes take up NBD-SM and a small portion is hydrolyzed to ceramide. Both uptake of SM and conversion to ceramide were unaffected by ABCA7 deficiency. To investigate further the role of ABCA7 in the hydrolysis of SM, we first labeled WT and KO hepatocytes with NBD-SM. Next, we studied time-dependent conversion of NBD-SM to NBD-Cer (Fig. 4). Total disappearance of NBD-SM was similar in both hepatocytes (Fig. 4B). We observed similar decline in cellular NBD-SM with time (Fig. 4, A and C). Moreover, the production of NBD-Cer was similar in WT and KO hepatocytes (Fig. 4, A and D). We did not see rapid increase in NBD-Cer levels corresponding to loss of cellular NBD-SM. These studies suggested that NBD-SM is metabolized to other unknown metabolites. Nevertheless, these studies showed that degradation of NBD-SM in hepatocytes was unaffected by the absence of ABCA7. In short, these studies showed that mouse hepatocyte take up SM. A portion of it is converted to NBD-Cer probably due to sphingomyelinase activity. More importantly, we did not find any differences between WT and KO hepatocytes with respect to SM uptake and metabolism. Therefore, it is likely that ABCA7 does not play a major role in SM uptake and hydrolysis. ABCA7 protein contains two each of extracellular, transmembrane, and intracellular ABC domains (10Aikawa T. Holm M.L. Kanekiyo T. ABCA7 and pathogenic pathways of Alzheimer's disease.Brain Sci. 2018; 8: 27Crossref PubMed Scopus (66) Google Scholar). The ABC domains (Fig. 5A) are important for protein function (10Aikawa T. Holm M.L. Kanekiyo T. ABCA7 and pathogenic pathways of Alzheimer's disease.Brain Sci. 2018; 8: 27Crossref PubMed Scopus (66) Google Scholar, 11Kaminski W.E. Piehler A. Wenzel J.J. ABC A-subfamily transporters: structure, function and disease.Biochim. Biophys. Acta. 2006; 1762: 510-524Crossref PubMed Scopus (156) Google Scholar). To address the roles of these domains in SM synthesis and efflux, we first transfected 293T cells with a control plasmid or a plasmid expressing FLAG-tagged mouse ABCA7 (Fig. 5B, top). Expression of ABCA7 resulted in significantly greater SM efflux than that observed in the control cells (Fig. 5B, bottom). Next, we expressed mouse WT ABCA7, R986H, and K2027Q missense mutations in cells (Fig. 5C, top). SM efflux was similar in cells expressing WT ABCA7 and the K2027Q missense mutant (Fig. 5C, bottom and right graph). However, SM efflux was significantly lower in cells expressing R986H than in WT cells (Fig. 5C, bottom and right graph). Next, we studied the effect of the R986H mutation on cellular SM levels and found that cells expressing R986H had lower levels of SM than WT cells (Fig. 5D). These studies showed that overexpression of ABCA7 increases SM synthesis and efflux. Expression of the R986H, but not the K2027Q, missense mutation decreased SM cellular levels and efflux, thus indicating that at least one of the ABC domain is important for cellular SM levels and efflux. We then studied the effect of overexpression of human ABCA7 and the R989H missense mutation orthologous to mouse R986H reported in a Belgian cohort (12Bossaerts L. Hendrickx Van de Craen E. Cacace R. Asselbergh B. Van Broeckhoven C. Rare missense mutations in ABCA7 might increase Alzheimer's disease risk by plasma membrane exclusion.Acta Neuropathol. Commun. 2022; 10: 43Crossref PubMed Scopus (4) Google Scholar) that might be associated with LOAD (10Aikawa T. Holm M.L. Kanekiyo T. ABCA7 and pathogenic pathways of Alzheimer's disease.Brain Sci. 2018; 8: 27Crossref PubMed Scopus (66) Google Scholar, 13De Roeck A. Van den Bossche T. van der Zee J. Verheijen J. De Coster W. Van Dongen J. et al.Deleterious ABCA7 mutations and transcript rescue mechanisms in early onset Alzheimer's disease.Acta Neuropathol. 2017; 134: 475-487Crossref PubMed Scopus (45) Google Scholar) on the synthesis and efflux of SM in mouse hippocampal neuronal HT22 cells. Overexpression of GFP-tagged ABCA7-R989H showed strong vesicular localization, similar to ABCA7-WT (Fig. 5E). In addition, similar cellular protein levels were detected with a mouse anti-GFP IgG antibody via immunoblotting (Fig. 5F). Next, we studied the effect of the human R989H mutant on intracellular and medium SM levels in cells incubated with HDL. R989H expressing cells, compared with WT ABCA7 expressing cells, had lower levels of SM in the media and in cells (Fig. 5G). These results indicated that the human ABCA7 missense mutant R989H is defective in SM synthesis and efflux in neuronal cells. KD, KO, overexpression, and mutagenesis studies in the cells described above identified ABCA7 as a protein involved in SM synthesis. Next, we extended these studies to mice and investigated the consequences of ABCA7 deficiency on the levels of different SM species in the plasma, liver, and brain. Several SM species were more abundant in the plasma of ABCA7-deficient (Abca7−/−, KO) mice than in WT controls (Fig. 6A). In contrast, we did not observe any significant change in the levels of these sphingolipids in the livers of KO mice (Fig. 6B). However, analysis of different species of SM revealed significantly lower levels of C22:1-, C22-, C24:1-, C24-, and C26:1-SM in the brains of KO mice than those in WT controls (Fig. 6C). These data suggest that ABCA7 may play a role in determining the levels of SM with longer fatty acid species in the brain. Given the association between ABCA7 mutations and LOAD (14Hollingworth P. Harold D. Sims R. Gerrish A. Lambert J.C. Carrasquillo M.M. et al.Common variants at ABCA7, MS4A6A/MS4A4E, EPHA1, CD33 and CD2AP are associated with Alzheimer's disease.Nat. Genet. 2011; 43: 429-435Crossref PubMed Scopus (1513) Google Scholar, 15Steinberg S. Stefansson H. Jonsson T. Johannsdottir H. Ingason A. Helgason H. et al.Loss-of-function variants in ABCA7 confer risk of Alzheimer's disease.Nat. Genet. 2015; 47: 445-447Crossref PubMed Scopus (229) Google Scholar), we used the active place avoidance (APA) task to evaluate hippocampal-dependent and hippocampal-independent behaviors (16Cimadevilla J.M. Kaminsky Y. Fenton A. Bures J. Passive and active place avoidance as a tool of spatial memory research in rats.J. Neurosci. Methods. 2000; 102: 155-164Crossref PubMed Scopus (78) Google Scholar, 17Cimadevilla J.M. Wesierska M. Fenton A.A. Bures J. Inactivating one hippocampus impairs avoidance of a stable room-defined place during dissociation of arena cues from room cues by rotation of the arena.Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 3531-3536Crossref PubMed Scopus (143) Google Scholar, 18Pavlowsky A. Wallace E. Fenton A.A. Alarcon J.M. Persistent modifications of hippocampal synaptic function during remote spatial memory.Neurobiol. Learn. Mem. 2017; 138: 182-197Crossref PubMed Scopus (20) Google Scholar) in WT and KO mice (Fig. 7A). Young-adult (4–6 months old) WT and KO mice quickly learned to actively avoid the shock zone, as evidenced by decreases in the number of shocks that they received in different trials (Fig. 7B). Older KO (14–18 months old) mice also learned to avoid the shock zone (Fig. 7D, T1–T8). The values of these parameters were not significantly different to those of WT mice of the same age. These studies suggested that young and old KO mice do not show any visuospatial memory disabilities as assessed by APA, in agreement with the studies of Logge et al. (19Logge W. Cheng D. Chesworth R. Bhatia S. Garner B. Kim W.S. et al.Role of Abca7 in mouse behaviours relevant to neurodegenerative diseases.PLoS One. 2012; 7e45959Crossref PubMed Scopus (39) Google Scholar), who did not observe a decline in cognitive function in young-adult mice. We then investigated cognitive discrimination by changing the location of the shock zone (APA conflict, Fig. 7A, TC1–TC4) (16Cimadevilla J.M. Kaminsky Y. Fenton A. Bures J. Passive and active place avoidance as a tool of spatial memory research in rats.J. Neurosci. Methods. 2000; 102: 155-164Crossref PubMed Scopus (78) Google Scholar, 17Cimadevilla J.M. Wesierska M. Fenton A.A. Bures J. Inactivating one hippocampus impairs avoidance of a stable room-defined place during dissociation of arena cues from room cues by rotation of the arena.Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 3531-3536Crossref PubMed Scopus (143) Google Scholar, 18Pavlowsky A. Wallace E. Fenton A.A. Alarcon J.M. Persistent modifications of hippocampal synaptic function during remote spatial memory.Neurobiol. Learn. Mem. 2017; 138: 182-197Crossref PubMed Scopus (20) Google Scholar). Representative recordings of mouse paths and shocks received by a KO and a WT mouse are shown in Figure 7C. These recordings show that older WT mice quickly learned to avoid the conflict area and displayed a regular pattern of movement within the test arena. In contrast, older KO mice received more shocks and showed a random pattern of movement in the arena. Quantification revealed a transient increase in shocks in both young-adult WT and KO mice (Fig. 7, B and D; TC1, TC4), thus indicating that both WT and KO young-adult mice learned to ignore the initial location of the shock and acquired the information on the new location of the shock zone. With continued training, both the WT and KO young-adult mice rapidly learned to discriminate between the old and new shock locations (Fig. 7B, TC1–TC4). However, older KO mice received significantly more shocks in the first two trials of the third day, in contrast to older WT mice (Fig. 7D, TC1–TC2), but eventually learned to avoid the new shock zone (Fig. 7D, TC3–TC4). These studies showed that KO mice required more shocks to learn the new location of the shock zone. We interpret these studies to suggest that older KO mice have cognitive discrimination deficits. Since animals use visual cues to locate the shock zone in the APA task, we performed a visual Cliff test (20Fox M.W. The visual cliff test for the study of visual depth perception in the mouse.Anim. Behav. 1965; 13: 232-233Crossref PubMed Scopus (59) Google Scholar) of visual depth perception on a group of KO (n = 10) and WT mice (n = 12), >12 months old, and observed no significant differences (ANOVA safe. 8.42 ± 0.36 trials versus 9.33 ± 0.35; F = 1.59, p = 0.11). We also did not find differences in distance traveled, the number of shocks per entrance, or between males and females (Fig. S2). These studies suggest that the differences between KO and WT mice seen in APA conflict were not due to major visual abnormalities, locomotor problems, or differences in the sensitivity to the shocks. Subsequently, we attempted to address reasons for cognitive decline in older ABCA7 KO mice. To test whether the KO mice had abnormalities in synaptic transmission, we evaluated the long-term potentiation (LTP) form of synaptic plasticity in the hippocampal CA3-CA1 synapse (Fig. 8A). No significant differences were observed 40 min after LTP induction in the CA3-CA1 synapse between WT and KO slices (Fig. 8B). Because CA3-CA1 LTP has been proposed to be required for hippocampus-dependent memory storage (18Pavlowsky A. Wallace E. Fenton A.A. Alarcon J.M. Persistent modifications of hippocampal synaptic function during remote spatial memory.Neurobiol. Learn. Mem. 2017; 138: 182-197Crossref PubMed Scopus (20) Google Scholar), these results are consistent with the behavioral data (Fig. 7B). Next, we evaluated the LTP in the intracortical synapse in layer II/III of the lateral entorhinal cortex (LEC). No significant differences were observed in the input-output relationships or basal paired-pulse ratio (PPR) (Fig. 8C). We found that both, post-tetanic potentiation (PTP) and LTP were significantly lower in KO mice (Fig. 8D; PTP = 141.8%; LTP = 117.9% 40 min after high frequency stimulation (HFS)) as compared with WT mice (PTP = 208.1% p ≤ 0.0001; LTP = 157.1% p ≤ 0.0001). These studies suggest that the electrophysiological differences observed in the entorhinal cortex maybe due to impaired synaptic plasticity in this region rather than differences in basal synaptic transmission. Next, we assessed whether the LTP deficiency in ABCA7 KO mice was different between female and male mice. We did not find significant differences in input/output relationships in male and female mice (Fig. 8E). Interestingly, we found that female KO mice have significantly lower LTP values (F1,10 = 9.765, p = 0.0108) (Fig. 8F); however, Sidak’s post hoc analysis