Molecular Psychiatry (2003) 8, 863–878 & 2003 Nature Publishing Group All rights reserved 1359-4184/03 $25.00 www.nature.com/mp
ORIGINAL RESEARCH ARTICLE
DNA microarray profiling of developing PS1-deficient mouse brain reveals complex and coregulated expression changes ZK Mirnics1,4, K Mirnics2,4, D Terrano3, DA Lewis2, SS Sisodia3 and NF Schor1 1
Pediatric Center for Neuroscience, Department of Pediatrics and Neurology, University of Pittsburgh, School of Medicine, Children’s Hospital of Pittsburgh, Pittsburgh, PA, USA; 2Departments of Psychiatry and Neurobiology, University of Pittsburgh, Pittsburgh, PA, USA; 3Center for Molecular Neurobiology, University of Chicago, Chicago, IL, USA Presenilin 1 (PS1) plays a critical role in the nervous system development and PS1 mutations have been associated with familial Alzheimer’s disease. PS1-deficient mice exhibit alterations in neural and vascular development and die in late embryogenesis. The present study was aimed at uncovering transcript networks that depend on intact PS1 function in the developing brain. To achieve this, we analyzed the brains of PS1-deficient and control animals at embryonic ages E12.5 and E14.5 using MG_U74Av2 oligonucleotide microarrays by Affymetrix. Based on the microarray data, overall molecular brain development appeared to be comparable between the E12.5 and E14.5 PS1-deficient and control embryos. However, in brains of PS1-deficient mice, we observed significant differences in the expression of genes encoding molecules that are associated with neural differentiation, extracellular matrix, vascular development, Notch-related signaling and lipid metabolism. Many of the expression differences between wild-type and PS1-deficient animals were present at both E12.5 and E14.5, whereas other transcript alterations were characteristic of only one developmental stage. The results suggest that the role of PS1 in development includes influences on a highly coregulated transcript network; some of the genes participating in this expression network may contribute to the pathophysiology of Alzheimer’s disease. Molecular Psychiatry (2003) 8, 863–878. doi:10.1038/sj.mp.4001389
Alzheimer’s disease; notch signaling; oligonucleotide GeneChips; presenilin; knockout; lipid metabolism; transcript network; in situ hybridization
Keywords:
Introduction Presenilins (PS1 and PS2) are highly homologous integral membrane proteins essential for the intramembranous ‘g-secretase’ cleavage of the b-amyloid precursor protein (APP), Notch, ErbB4, E-cadherin, the LDL receptor-related protein, CD44, Delta1, Jagged2 and nectin-1a.1–12 PS1 is either the catalytic subunit or a cofactor of a high molecular weight complex that has g-secretase activity.13,14 The vast majority of pedigrees with autosomal dominant familial Alzheimer’s disease (FAD) are linked to missense mutations in PS1 and expression of these PS1 variants leads to increased production of highly amyloidogenic Ab42 peptides.15–18 PS1-deficient mice die at birth or in utero with skeletal deformities, cerebral hemorrhage, impaired neurogenesis and cavitations throughout the brain.19–23 PS2-deficient mice are normal, viable and fertile.24 In Correspondence: Dr ZK Mirnics, PhD, Pediatric Center for Neuroscience, Departments of Pediatrics and Neurology, University of Pittsburgh, Pittsburgh, PA, 15213, USA. E-mail:
[email protected] 4 Both contributed equally to this work. Received 28 March 2003; revised 15 May 2003; accepted 16 May 2003
addition to being essential for early embryonic development and g-secretase cleavage, PS1 plays other physiological roles.25,26 Based on morphological and developmental analysis of PS1-deficient mice, it has been suggested that PS1 plays a role in cell proliferation, neuronal differentiation, cell adhesion and extracellular matrix formation.19,23,27–32 Within the developing nervous system, PS1 deficiency leads to thinning of the ventricular zone, loss of Cajal– Retzius neurons, cortical dysplasia, defective neuronal migration and leptomeningeal fibrosis.19,23 PS1-hypomorphic mice are viable and show markedly reduced g-secretase cleavage. Analyzing PS1-hypomorphic and wild-type (WT) mice with differential display, Liauw et al33 uncovered 19 differentially expressed genes between the two conditions, including robustly decreased expression of Hif1a, deltacatenin, and cell division cycle 10, and increased expression of nucleoside diphosphate kinase subunit A. The molecular events leading to the profound developmental alterations in PS1-deficient mice are not well understood. The present study was aimed at uncovering complex transcript networks that show alterations in the absence of intact PS1 function in the developing brain. The obtained results argue that PS1 plays a major role in the formation of extracellular
Expression profiling of PS1-deficient brain ZK Mirnics et al
864
matrix, in the differentiation of progenitor cells and in normal lipid metabolism. Furthermore, PS1 may regulate transcript levels of several genes that have been previously associated with the pathophysiology of Alzheimer’s disease (AD).
Materials and methods PS1-deficient animals The generation of PS1-deficient animals has been described.20 Timed matings were set up between heterozygous PS1 mice. The morning of the day when a vaginal plug is seen is designated as embryonic day 0.5 (E0.5). Embryos at E12.5 and E14.5 were dissected from pregnant females and the hindlimb was removed from each embryo for genotyping. Embryonic heads were removed, brains were rapidly dissected, frozen on dry ice and stored on 801C until use. Beyond E14.5, PS1-deficient embryos develop visible morphological defects in specific regions of the ventricular zone and show extensive brain hemorrhages. In selecting two earlier ages for our brain analysis (E12.5 and E14.5), we wanted to ensure that (1) the obtained microarray data were not a reflection of nonspecific brain deterioration and (2) we could assess developmental progression of the transcriptome in both PS1deficient and WT animals. This approach allowed us to separate age-specific PS1-dependent transcriptome changes from those persisting over a longer developmental timeline.
Microarrays A total of 12 embryonic brains were analyzed by microarrays in the current study: three WT and three PS1-deficient, each at ages E12.5 and E14.5. RNA isolation, cDNA synthesis, in vitro transcription and microarray hybridization were performed using standard protocols recommended by Affymetrix.34 Briefly, total RNA was isolated using Trizol reagent (Invitrogen). The total RNA was reverse transcribed using a T7-promoter coupled oligo(d)T primer (GeneChip T7Oligo(d)T Promoter Primer Kit, Affymetrix). After the second-strand cDNA synthesis, an in vitro transcription (IVT) reaction was performed using an Enzo BioArray High Yield RNA transcript labeling kit (Affymetrix). The quality of total RNA and cRNA was analyzed on Bioanalyzer (Agilent). The labeled IVT samples that passed the Test 3 arrays were hybridized to GeneChips MG_U74Av2 microarrays that contain B12 000 annotated genes and ESTs. 50 /30 RNA integrity in all samples used was better than 1:1.2. Data analysis Initial microarray images were analyzed by Affymetrix software Microarray Analysis Suite version 5 (MAS5).35 For each microarray, signal intensity was scaled to 200. Each PS1-deficient sample was compared to all three WT samples of appropriate age, creating a 3 3 comparison matrix (Figure 1). Standards and genes with marginal expression or absent expression were eliminated from the data set. The
Figure 1 Cross-comparison analysis of PS1-deficient and WT microarray data. Comparisons were independently performed for the E12.5 and E14.5 ages. The microarrays data were cross-compared between the PS1-deficient and WT embryos using MAS 5.0 for B12 000 genes represented on the MG_U74A oligonucleotide arrays. Microarrays from the WT embryos were used as baseline in each comparison, resulting in a cross-comparison matrix of nine calls. For each gene each ‘Increased’ call was assigned a value of þ 1 and ‘Decreased’ call was valued 1. ‘Not changed’ calls were assigned a value of 0. The values were summed up for each gene across the nine possible comparisons, thus defining a ‘net expression change (NEC)’ value that ranged from þ 9 (a gene increased in all comparisons) to 9 (a gene decreased in all comparisons). In the next step, the data set was reduced by eliminating spiked in standards and genes that reported ‘Absent’ and ‘Marginal’ calls in 450% of the microarrays. The reduced data set was sorted and the distribution of the genes was plotted based on the NEC value. Finally, genes with NEC value of o5 were eliminated from the data set and the remaining fully annotated genes are reported in Table 1. Molecular Psychiatry
Expression profiling of PS1-deficient brain ZK Mirnics et al
expressed genes were ranked by the frequency of Net Expression Change, which was defined as the number of decreased calls—number of increased calls across the nine comparisons between the PS1-deficient and WT animals for each age. Net expression changes of X5 occur at a probability of o 0.005 and only these observations were considered for further analysis. For the comparison of developmental progress, we performed hierarchical clustering by Genes@Work software36 using gene vector calculations, Euclidean distance with mean calculations, and center of mass calculations. Both genes and samples were clustered. The analysis was independently performed for E12.5 and E14.5. Cross-correlation of expression levels was performed on MAS 5.0 scaled intensity values that were exported into Microsoft Exel 2000. In situ hybridization Six additional PS1-deficient and six control E14.5 embryos were used to confirm microarray data by in situ hybridization. Briefly, fresh-frozen embryos were coronally sectioned under RNAse-free conditions at 20 mm thickness. Mouse probes were generated using gene-specific primers and normal brain cDNA as a template in a standard PCR reaction. The PCR product was cloned into a plasmid with T7 and SP6 polymerase sites. These plasmids were used to make a specific in situ RNA probe by an in vitro transcription reaction (Riboprobe combination system SP6/T7 RNA Polymerase, Promega). The 35S-labeled RNA probe was purified using an RNeasy column (Gibco). We determined the specific activity of the probe (B2 000 000 c.p.m. per slide) and denatured it before adding to the slides. Slides with brain sections were fixed with 4% paraformaldehyde followed by washes and dehydration through a series of graded alcohol solutions. Slides were dried and the radiolabeled probe added to them. Hybridization was carried out at 561C for 16–20 h. This was followed by posthybridization washes (4 SSC buffer with mercaptoethanol, 4 SSC, formamide in formamide buffer, 2 SSC, RNase A treatment, 2 SSC, 1 SSC, 0.5 SSC, 0.1 SSC) and exposure to X-ray film. After determining the optimal labeling time (24 h of film¼5 days of emulsion dip), slides were dipped in emulsion and kept in a lightproof box for varying lengths of time and then developed and counterstained with toluidine blue. All procedures have been described in detail previously.37–40
Results Molecular development of the brain is comparable in PS1-deficient and WT embryos As PS1-deficient animals died in late embryogenesis, we first wanted to establish if the normal molecular developmental program was preserved in PS1-deficient animals (Figure 2, left panel). This was achieved by comparing the molecular developmental progression of PS1-deficient animals from E12.5 and E14.5 to
that seen in WT animals. Using a cross-comparison paradigm between E12.5 and E14.5 WT animals, we defined 50 genes that showed prominent expression increases and 50 transcripts that reported significant decreases between these two developmental ages. This was followed by testing the expression levels of these 100 genes in the E12.5 vs E14.5 PS1-deficient animals (Figure 2, right panel). For these genes, we observed parallel progression of gene expression changes from E12.5 to E14.5 in WT and knockout (KO) animals. The results argue that basic molecular brain development, judged by 100 genes with agedependent transcription level, is comparable between the WT and KO animals. This suggests that in subsequent analyses, we would be able to identify changes that are due specifically to PS1 deficiency (and not simply due to the reflection of greatly impaired developmental progression).
865
PS1-deficient animals show complex gene expression changes in the brain At E12.5, the PS1-deficient embryos, when compared to control littermates, showed 35 upregulated and 35 dowregulated gene transcripts (Table 1, A1–2). Among the group of annotated genes with decreased expression, we identified several transcripts that were previously observed in independently generated PS1 KO animals, thus reproducing previously published observations.20,25 For example, at E12.5 we observed a decrease in protein tyrosine phosphatase Z (Ptprz; phosphacan; DSD1 proteoglycan), Notch1 and Hes5.20,23,25 At this age, genes with changed expression included members of the Notch (Notch1, Sfrp2) signaling pathway and many transcription factors (Lhx1, Uncx4.1, Sfrp2, Sall3, Pou3f4, Tbr1, Foxg1, Zfp312, Zac1, Eomes, Lhx8, Neurog2, Nfyb, Neurod6). Furthermore, genes contributing to normal neural/ glial differentiation were also affected (Notch1, Tbr1, Eomes, Neurog2, Neurod6). Changes in the expression of transcription factors in early-born neurons are likely to be related to a premature exit of progenitor cells from the cell division cycle.25 Finally, several extracellular matrix-related gene products were also downregulated (Tnc, Chl1, Ptprz, Ptprd), while cytoskeletal genes showed both expression increases (Actb) and decreases (Mapt, Nfl, Nfm, Ina, Sncg). At E14.5, the PS1-deficient embryos, when compared to control littermates, showed 17 upregulated and 56 downregulated gene transcripts (Table 1, B1– 2). The increased expression of Hba-x, Tpi, Hmox1, Hbb-b1, Alas2 and Hbb-y is most likely a response to hypoxic changes that may occur in the PS1-deficient brain. In contrast, decreased expression was reported for genes belonging to the same functional groups as observed in the E12.5 comparisons (transcription factors: Zfp36I2, Zhx1, Zfhx1a; Notch signaling: Notch1, En2, Dlx1, Hes5; lipid metabolism: Acas2, Pde1b, Scd1, Asml3a, Dhcr7; extracellular matrix: Ncam1, Chl1, Tnc, Ptprz, Ptprd). Interestingly, our analysis also found a reduction in PAF acetylhydrolase (phospholipase A2–Pla2g7) transcript (a gene Molecular Psychiatry
Expression profiling of PS1-deficient brain ZK Mirnics et al
866
Figure 2 Developmental progression of brain transcriptome from E12.5 to E14.5 in WT and PS1 / animals. Both samples and genes were clustered using Genes@Work. Genes are represented in rows, arrays are in columns. Shades of green are proportional to the increased expression in the PS1-deficient embryos, shades of red are proportional to the transcript decreases. Left panel: Hierarchical clustering of 100 genes that showed prominent developmental change between E12.5 and E14.5 WT embryos. Right panel: Hierarchical clustering of the same 100 genes in the E12.5 and E14.5 PS1-deficient embryos. Note that the same set of genes, defined in the WT samples, showed a comparable expression pattern in the PS1-deficient embryos, arguing that the overall molecular development form E12.5 to E14.5 was comparable in the WT and PS1-deficient embryos. Molecular Psychiatry
Expression profiling of PS1-deficient brain ZK Mirnics et al
Table 1 #
867
Gene expression changes in the brain of PS1-deficient embryos
Probe
Gene name
Symbol
A1. Increased expression in E12.5 PS1 / animals 1 100065_r_at Gap junction membrane channel protein alpha 1 2 101042_f_at Peptidase 4 3 101578_f_at Actin, beta, cytoplasmic 4 102001_at Ribonucleotide reductase M2 5 102277_at Zinc-finger protein 26 6 102632_at Calmodulin binding protein 1 7 102700_at T-box brain gene 1 8 102865_at MAD homolog 5 (Drosophila) 9 103654_at Nucleosome binding protein 1 10 104097_at Budding uninhibited by benzimidazoles 1 11 104400_at Prenylated SNARE protein 12 160159_at Cyclin B1, related sequence 1 13 160681_at Poly (A) polymerase alpha 14 160906_i_at Ectodermal-neural cortex 1 15 161049_at Forkhead box G1 (HNF3 family member) 16 92229_at Zinc-finger protein 312 17 92501_s_at ZF regulator of apoptosis and cell cycle arrest 18 92502_at ZF regulator of apoptosis and cell cycle arrest 19 93285_at Dual specificity phosphatase 6 20 93353_at Lumican 21 93401_g_at Baculoviral IAP repeat-containing 4 22 93630_at CUG triplet repeat, RNA binding protein 1 23 93880_at T-box 2; Eomesodermin (Eomes) 24 94197_at UDP-glucose ceramide glucosyltransferase 25 94754_at LIM homeobox protein 8 26 94853_at Guanine nucleotide binding protein, beta 1 27 95328_at Fucosyltransferase 9 28 95705_s_at Actin, beta, cytoplasmic; A-X actin 29 95791_s_at Splicing factor, arginine/serine-rich 10 30 97130_at Arginine-tRNA-protein transferase 1 31 97302_at Nd1 32 97421_at Fibroblast growth factor inducible 16 33 97792_at Neurogenin 2 34 98024_at Nuclear transcription factor-Y beta 35 98857_at Neurogenic differentiation 6; Atonal 2, Math2
Gja1 Pep4 Actb Rrm2 Zfp26 Calmbp1 Tbr1 Madh5 Nsbp1 Bub1 Ykt6 Ccnb1-rs1 Papola Enc1 Foxg1 Zfp312 Zac1 Zac1 Dusp6 Lum Birc4 Cugbp1 Tbr2 Ugcg Lhx8 Gnb1 Fut9 Actb Sfrs10 Ate1 Nd1 Fin16 Neurog2 Nfyb Neurod6
A2. Decreased expression in E12.5 PS1 / animals 1 100139_at Proprotein convertase subtilisin/kexin type 1 inhibitor 2 100548_at Phosphoprotein enriched in astrocytes 15 3 100566_at Insulin-like growth factor binding protein 5 4 100569_at Annexin A2 5 100972_s_at Chemokine (C-C motif) ligand 27 6 101993_at Tenascin C 7 102431_at Microtubule-associated protein tau 8 102732_at Talin 9 102923_at Prepronociceptin 10 103088_at Close homolog of L1 11 103575_at Neurofilament, light polypeptide 12 103581_at Cytosolic acyl-CoA thioesterase 1 13 104280_at Synuclein, gamma (persyn) 14 160172_at Maternally expressed gene 3 15 160899_at Purkinje cell protein 4 16 92346_at Neurofilament, medium polypeptide 17 92379_f_at Tyrosine phosphatase, receptor Z (DSD-1) 18 92499_at Unc4.1 homeobox (C. elegans) 19 92558_at Vascular cell adhesion molecule 1 20 92961_at LIM homeobox protein 1 21 93379_at Dihydropyrimidinase-like 4 22 93503_at Secreted frizzled-related protein 2 23 93896_at Tyrosine phosphatase, receptor type, D 24 94335_r_at Internexin alpha
Pcsk1n Pea15 Igfbp5 Anxa2 Ccl27 Tnc Mapt Tln Pnoc Chl1 Nfl Cte1 Sncg Meg3 Pcp4 Nfm Ptprz Uncx4.1 Vcam1 Lhx1 Dpysl4 Sfrp2 Ptprd Ina
Chromosome
10 29.0 cM 7 15.0 cM 5 80.0 cM 12 7.0 cM 1 E4 2 32.0 cM 13 35.0 cM XD 2 73.0 cM 11 A1 13 D1 12 F1 13 D1 12 21.0 cM 14 A1 10 15.0 cM 10 15.0 cM 10 C3 10 61.0 cM X A3-A5 2 47.5 9 67.0 cM 4 32.0 cM 3 H4 4 79.4 cM 4 A3 5 80.0 cM 11 72.0 cM 7 F3 1 G1 4 18.5 cM 3 H1 10 43.7 cM 6 29.0 cM X A1.1 1 93.8 cM 1 36.1 cM 9 37.0 cM 4 12.7 cM 4 32.2 cM 11 64.0 cM 4 B1 14 D1 14 28.7 cM 12 D1 14 12.5 cM 12 54.0 cM 16 69.9 cM 14 28.5 cM 6 A3.1 5 82.0 cM 3 50.8 cM 11 48.0 cM 7 F4 3 38.5 cM 4 38.0 cM 19 C3
Molecular Psychiatry
Expression profiling of PS1-deficient brain ZK Mirnics et al
868
Table 1 # 25 26 27 28 29 30 31 32 33 34 35
continued. Probe
Metallothionein 3 Mab-21-like 2 (C. elegans) Brain abundant, membrane attached signal 1 Latexin Microsomal glutathione S-transferase 3 Serine proteinase inhibitor E, member 2 Notch gene homolog 1, (Drosophila) T-cell acute lymphocytic leukemia 1 sal-like 3 (Drosophila) Fatty acid binding protein 7, brain POU domain, class 3, transcription factor 4
Symbol
Chromosome
Mt3 Mab21l2 Basp1 Lxn Mgst3 Serpine2 Notch1 Tal1 Sall3 Fabp7 Pou3f4
8 45.0 cM 3 15 B1 3 31.6 cM 1 H2 1 48.6 cM 2 15.0 cM 4 49.5 cM 18 E3 10 B4 X 48.4 cM
B1. Increased expression in E14.5 PS1 / animals 1 100446_r_at Small proline-rich protein 1B 2 100549_at Hemoglobin X, alpha-like embryonic chain 3 100574_f_at Glucose phosphate isomerase 1 - Neuroleukin 4 101069_g_at Makorin, ring finger protein, 1 5 102726_at Tachykinin 1 6 104432_at Ras homolog N (RhoN) 7 160101_at Heme oxygenase (decycling) 1 8 160568_at Enolase 1, alpha non-neuron 9 161763_r_at PI-4-phosphate 5-kinase, type II, gamma 10 162457_f_at Hemoglobin, beta adult major chain 11 92768_s_at Aminolevulinic acid synthase 2, erythroid 12 93563_s_at Nidogen 2 13 94524_at Death-associated protein 3 14 96072_at Lactate dehydrogenase 1, A chain 15 97180_f_at Hemoglobin Y, beta-like embryonic chain 16 99566_at Triosephosphate isomerase 17 99669_at Lectin, galactose binding, soluble 1
Sprr1b Hba-x Gpi1 Mkrn1 Tac1 Arhn Hmox1 Eno1 Pip5k2c Hbb-b1 Alas2 Nid2 Dap3 Ldh1 Hbb-y Tpi Lgals1
3 45.2 cM 11 16.0 cM 7 11.0 cM 6 B1 6 5.0 cM 11 C-D 8 35.0 cM 4 79.0 cM 10 D3 7 50.0 cM X 63.0 cM 14 A2 3 F2 7 23.5 cM 7 49.95 cM 6 60.2 cM 15 44.9 cM
B2. Decreased expression in E14.5 PS1 / animals 1 100047_at Synaptosomal-associated protein, 25 kDa 2 100144_at Nucleolin 3 100153_at Neural cell adhesion molecule 1 4 100548_at Phosphoprotein enriched in astrocytes 15 5 100566_at Insulin-like growth factor binding protein 5 6 101923_at Phospholipase A2 group 7, Lissencephaly 1 7 101929_at Expressed sequence C87222 8 101930_at Nuclear factor I/X 9 101991_at Flavin containing monooxygenase 1 10 101993_at Tenascin C 11 102786_at Chloride channel 3 12 103061_at Glutamic acid decarboxylase 1 13 103088_at Close homolog of L1 14 103581_at Cytosolic acyl-CoA thioesterase 1 15 104280_at Synuclein, gamma 16 104492_at Early B-cell factor 3 17 160273_at Zinc-finger protein 36, C3 H type-like 2 18 160546_at Aldolase 3, C isoform 19 160726_at Quaking 20 160848_at Zinc-fingers and homeoboxes protein 1 21 160921_at Acetyl-Coenzyme A synthetase 2 22 161051_at Hairy and enhancer of split 5 (Drosophila) 23 161059_at GABA transporter 1 24 92379_f_at Tyrosine phosphatase, receptor Z (DSD-1) 25 92380_r_at Tyrosine phosphatase, receptor Z (DSD-1) 26 92435_at Retinaldehyde binding protein 1 27 92545_f_at Prostaglandin D2 synthase (21kDa, brain) 28 92558_at Vascular cell adhesion molecule 1 29 92700_at Brevican 30 92731_at Pentaxin-related gene 31 92795_at Microtubule-associated protein 4
Snap25 Ncl Ncam1 Pea15 Igfbp5 Pla2g7 C87222 Nfix Fmo1 Tnc Clcn3 Gad1 Chl1 Cte1 Sncg Ebf3 Zfp36l2 Aldo3 qk Zhx1 Acas2 Hes5 Gabt1 Ptprz Ptprz Rlbp1 Ptgds Vcam1 Bcan Ptx3 Mtap4
2 78.2 cM 1 48.4 cM 9 28.0 cM 1 93.8 cM 1 36.1 cM 17C 11 E2 8 38.6 cM 1 H1 4 32.2 cM 8 32.2 cM 2 43.0 cM
Molecular Psychiatry
95340_at 95379_at 95674_r_at 96065_at 96258_at 97487_at 97497_at 97973_at 98852_at 98967_at 99386_at
Gene name
12 D1 14 12.5 cM 7 F3 17 E4 11 44.98 cM 17 5.9 cM 2 4 3 6 6 7 2 3 3 3 9
H1 81.5 cM 50.4 cM A3.1 A3.1 39.0 cM 12.9 cM 50.8 cM 48.0 cM 33.8 cM 58.0 cM
Expression profiling of PS1-deficient brain ZK Mirnics et al
Table 1
869
continued.
#
Probe
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
93026_at 93326_at 93382_at 93503_at 93896_at 94056_at 94057_g_at 94354_at 94872_at 95340_at 95356_at 95661_at 96065_at 96088_at 96295_at 97198_at 97248_at 97487_at 97497_at 98338_at 98394_at 98967_at 98989_at 99052_at 99586_at
Gene name
Symbol
Microsomal glutathione S-transferase 1 Transmembrane 4 superfamily member 2 Phosphodiesterase 1B, 63 kDa Secreted frizzled-related protein 2 Tyrosine phosphatase, receptor type, D Stearoyl-Coenzyme A desaturase 1 Stearoyl-Coenzyme A desaturase 1 ATP-binding cassette, sub-family A1 Sphingomyelin phosphodiesterase 3a Metallothionein 3 Apolipoprotein E CD9 antigen Latexin N-myc downstream regulated 2 Phosphoserine aminotransferase ATP-binding cassette, sub-family A1 Diazepam binding inhibitor Serine proteinase inhibitor E, member 2 Notch gene homolog 1 (Drosophila) Engrailed 2 Distal-less homeobox 1 Fatty acid binding protein 7, brain 7-Dehydrocholesterol reductase Zinc-finger homeobox 1a Cystatin C
that has been associated with lissencephaly), possibly explaining the characteristic brain morphology that PS1-deficient embryos develop.27 At both ages, some of the genes with decreased expression have been associated with the pathophysiology of AD. At E12.5, we found the expression of metallothionein 3 (Mt3), Tau (Mapt) and neurofilament (Nfl and Nfm) to be reduced, while at E14.5, Mt3 and apolipoprotein E (ApoE) showed lower expression levels.41–45 Selected gene expression changes were verified by in situ hybridization Of the genes that reported an expression difference in at least one developmental stage, five were randomly chosen for further verification by in situ hybridization using radioactive riboprobes (Figure 3). For these genes, both expression decreases (VCAM1, latexin) and increases (neurod6-math2, neuroleukin-Gpi1, lumican) were verified. PS1-dependent expression changes occurred across different brain regions and cell types, further demonstrating the complexity of the role of PS1 during early embryonic development. PS1-deficient animals show highly correlated expression decreases at E12.5 and E14.5 Gene expression decreases observed at both E12.5 and E14.5 are likely to have biological consequences. As they are present over a time window of several days in the developing brain, these gene expression decreases may play a more primary role in the PS1-dependent transcriptional network. We found that 15 genes
Mgst1 Tm4sf2 Pde1b Sfrp2 Ptprd Scd1 Scd1 Abca1 Asml3a Mt3 Apoe Cd9 Lxn Ndr2 Psat Abca1 Dbi Serpine2 Notch1 En2 Dlx1 Fabp7 Dhcr7 Zfhx1a Cst3
Chromosome 6 G1 X A1.3-A2 15 F3 3 38.5 cM 4 38.0 cM 19 43.0 cM 19 43.0 cM 4 23.1 cM 10 B4 8 45.0 cM 7 4.0 cM 6 58.0 cM 3 31.6 cM 14 C1 19 A 4 23.1 cM 1 E2 1 48.6 cM 2 15.0 cM 5 15.0 cM 2 44.0 cM 10 B4 7 F5 18 0.0 cM 2 84.0 cM
reported decreased expression at both E12.5 and E14.5 (Table 2), while no gene showed increased expression at both ages. The probability of 15 of 35 decreased genes at E12.5 recurring in 56 expression decreases at E14.5 is o0.0001. Genes with reported expression decreases at both ages included IGFbinding protein 5, tenascin C (hexabrachion), close homologue of L1, cytosolic acyl-CoA thioestherase 1, synuclein gamma (persyn), protein tyrosine phosphatase Z (DSD-1 proteoglycan, phosphacan) secreted frizzled-related sequence protein 2, protein tyrosine phosphatase D, Mt3, latexin, serine proteinase inhibitor E2, Notch1, phosphoprotein enriched in astrocytes 15 and fatty acid binding protein 7. Based on the mean of the average log ratios across the changed comparisons, the decreases ranged from 1.3fold (Pea15) to 9.2-fold (Sncg). Coregulation of transcripts across conditions and ages Since microarrays simultaneously analyze the expression of 412 000 genes in a single sample, the obtained results represent more than a set of independent observations; in particular, within each sample, the relationship between individual transcript levels carries important information. Within each individual sample, transcripts may increase or decrease in a comparable ratio, suggesting a coregulation. Alternatively, transcript level changes may occur independently within each sample. We hypothesized that the genes showing expression changes are highly dependent on each other (or on a common factor) regardless of the embryonic age or presence/absence Molecular Psychiatry
Expression profiling of PS1-deficient brain ZK Mirnics et al
870
Figure 3 In situ verification of selected data. Five genes with expression changes were randomly chosen for further verification using gene-specific riboprobes on a new set of E14.5 embryos. Drawing represents a coronal section from an E14.5 embryonic brain, dashed red boxes denote areas depicted in panels a–d micrographs. Panel e is located rostral to the plane shown here. VCAM-1 expression (a) was decreased in the neuroepithelial zone, while the most prominent decrease for Latexin mRNA was observed in the basal forebrain (b). In contrast, Neurod6 expression (c) was increased diffusely across the developing cortex, Lumican transcript (d) was elevated in the meninges, while Neuroleukin expression (e) was amplified in the most rostral part of the developing brain.
of the PS1 gene. To test the relationships between the transcripts of interest, we cross-correlated the expression levels of 27 genes (15 genes from Table 2 and 3 þ 3 gene expression decreases and increases at E12.5 and E14.5) across all 12 microarrays. The expression levels of many genes were highly correlated across all samples (Figure 4, Table 3) with correlations ranging as high as r¼0.97 (Notch1 vs Hes5).
is the result of a highly coregulated transcriptional network. In this study, we were able to identify a set of genes that exhibit PS1-dependent expression and that participate in differentiation, cytoskeletal and extracellular matrix formation, and lipid metabolism. Furthermore, we showed that these transcripts were coregulated with each other and with the molecules that are known to play a role in the pathophysiology of AD.
Discussion
Absence of PS1 accelerates neuronal and glial differentiation Anatomically, the PS1-deficient embryonic brain is characterized by an increased number of postmitotic
Our data indicate that the PS1-deficient embryonic brain shows a well-defined molecular phenotype that Molecular Psychiatry
Expression profiling of PS1-deficient brain ZK Mirnics et al
Table 2 #
871
Genes with decreased expression in both E12.5 and E14. 5 PS1 / brains
Probe
Gene name
Symbol
Chr
GO biological function
1 100566_at Insulin-like growth factor binding protein
Igfbp5
1
GO:1558; regulation of cell growth
2 101993_at Tenascin C
Tnc
4
GO:7155; cell adhesion
3 103088_at Close homolog of L1 4 103581_at Cytosolic acyl-CoA thioesterase 1
Chl1 Cte1
5 104280_at Synuclein, Sncg gamma (persyn) 6 92379_f_at Protein tyrosine Ptprz phosphatase, receptor type, Z (DSD-1 proteoglycan)
7 92558_at
Vascular cell Vcam1 adhesion molecule 1
8 93503_at
Secreted frizzled-related sequence protein 2 Protein tyrosine phosphatase, receptor type, D
9 93896_at
GO cell component
GO molecular function
GO:5576; GO:19838; extracellular growth factor binding GO:5520; insulin-like growth factor bind GO:5578; extracellular matrix
E12.5a E14.5a
0.81 0.67
1.85 1.66 1.27 0.47
12
GO:1676; long-chain fatty acid metabolism GO:6637; acyl-CoA metabolism
14 6
3
Sfrp2
3
Ptprd
4
10 95340_at
Metallothionein 3
Mt3
8
11 96065_at 12 97487_at
Latexin Serine (or cysteine) proteinase inhibitor, clade E, member 2
Lxn 3 Serpine2 1
GO:7185; transmembrane receptor protein tyrosine phosphatase signaling pathway GO:16337; cell–cell adhesion GO:7155; cell adhesion
GO:5737; cytoplasm
GO:16291; 0.83 0.68 acyl-CoA thioesterase GO:16292; acyl-CoA thioesterase I GO:4759; serine esterase GO:3824; enzyme GO:16787; hydrolase
GO:5737; cytoplasm
3.25 1.18
GO:5886; plasma membrane
GO:4725; protein 1.20 1.15 tyrosine phosphatase
GO:5886; plasma membrane GO:16021; integral membrane protein GO:16020; membrane
GO:5194; cell adhesion molecule
1.35 1.32
0.52 0.80
GO:7185; transmembrane receptor protein tyrosine phosphatase signaling pathway
GO:5886; plasma membrane GO:16021; integral membrane protein
GO:4725; protein 0.62 0.42 tyrosine phosphatase GO:16787; hydrolase
GO:5505; heavy metal binding GO:8201; heparin binding
0.92 0.70 1.27 0.73 1.66 0.83
GO:4867; serine protease inhibitor GO:4868; serpin
Molecular Psychiatry
Expression profiling of PS1-deficient brain ZK Mirnics et al
872
Table 2 #
continued.
Probe
13 97497_at
14 98967_at
Gene name
Notch gene homolog 1
Fatty acid binding protein 7, brain
15 100548_at Phosphoprotein enriched in astrocytes 15
a
Symbol
Notch1
Fabp7
Pea15
Chr
2
10
1
GO biological function
GO cell component
GO:30154; cell GO:16021; differentiation, integral DNA-dependent membrane protein GO:6355; GO:16020; regulation of membrane transcription GO:7275; development GO:6810; transport
GO:7242; intracellular signaling cascade GO:6915; apoptosis
GO:5624; membrane fraction
GO molecular function GO:5515; protein binding
E12.5a E14.5a
0.52 0.57
GO:5509; calcium ion binding
GO:8289; lipid binding GO:5488; binding GO:16329; apoptosis regulator
0.72 0.96
0.50 0.38
GO:5829; cytosol
Mean signal log ratio across changed comparisons; SLR of 1.0¼2-fold change.
neurons and reduced size of the ventricular zone.25 At the molecular level, we found that PS1-deficient mice have an elevated expression of Foxg1 (HNF3 family member) and transcription factors that are present in early-born neurons (Tbr1) and in postmitotic neurons (Neurod6-Math2 and Tbr2) (Figure 5A).46–48 Given that Tbr1 gene promoter has eight and Tbr2 has three HNF3-beta binding sites, we propose that the expression of Tbr1 and Tbr2 genes is most likely a consequence of Foxg1 upregulation. Furthermore, it is known that overexpression of HNF3 family members results in decreased Bfabp and engrailed 2 expression, both observed in the E12.5 PS1-deficient embryos.49 Tbr1 also regulates reelin expression (which has been marginally decreased in our data set), which binds to ApoE receptor.50,51 The reelin– ApoER interaction may at least partially account for the observed decrease in ApoE transcripts. In addition, we observed decreases of Hes5 (basic helix–loop–helix transcription factor and downstream effector of Notch1 signaling) and Igfbp5 transcripts. These transcript decreases are also consistent with premature differentiation of progenitors; absence of Hes5 in mice, as well as inhibition of Igfbp5, result in premature neuronal differentiation.52–54 PS1-deficient mice brains show downregulation of extracellular matrix transcripts Extracellular matrix molecules are essential for normal stratification of the cortex.47,55–57 PS1-deficient brains have disrupted laminar architecture and show a progressive loss of Cajal–Retzius neurons.19,25 Molecular Psychiatry
The defects in the cortical layer formation and neuronal migration, as well as an absence of Cajal– Retzius neurons are almost certainly related to the expression deficits of the extracellular matrix molecules we report in this study (Ptprz, Ptprd, Tnc, Sfrp2, Chl1, Lum, Bcan and Ncam).23 It is known that proteoglycans and extracellular matrix molecules influence neuronal migration and are essential for normal formation of radial glial processes,47,58 and we propose that the expression disturbances observed in our data set directly contribute to the thin and poorly defined cortical plate in the brain of PS1-deficient embryos. PS1-deficiency is associated with defects in lipid metabolism of the acyl-CoA pathway The concentration of cytosolic free long-chain acylCoA esters in the cells is regulated by the concentration of acyl-CoA binding protein (Dbi), fatty acid binding proteins (FABP) and acyl-CoA hydrolase (CteI) activity, which are all downregulated in PS1deficient mice.59,60.Dbi and Bfapb buffer large fluctuations in free long-chain acyl-CoA ester concentration, while CteI prevents accumulation of free unprotected long-chain acyl-CoA esters and ensures sufficient amounts of free CoA to support beta-oxidation and other CoA-dependent enzymes (Figure 5B).61–63 Changes in this pathway may also explain the observed change in retinaldehyde-binding protein 1 expression that is involved in the process of vitamin A esterification and storage, another acyl-CoA-dependent process.64,65
Expression profiling of PS1-deficient brain ZK Mirnics et al
873
Figure 4 Coregulation of changed gene expression in the developing mouse brain for four genes playing a role in AD and/or ageing pathophysiology (Notch1, Mapt, Mt3 and ApoE). For each microarray (sample), MAS 5.0 standardized signal intensities for individual genes were plotted on either the X- or Y-axis. Note the highly correlated expression levels between the pairs of genes within the brain of mice regardless of developmental stage or the presence/absence of the PS1-gene during development.
Furthermore, six more genes with altered expression in our study have been associated with lipid metabolism: acyl-CoA synthetase 2 (Acas2), stearoylCoA desaturase 1 (Scd1), acid sphingomyelinase-like phosphodiesterase 3a (Asml3a), reverse cholesterol transporter ATP-binding cassette A1 (Abca1), prostaglandin D2 synthetase (Ptgds) and ApoE.66–72 Interestingly, the expression of almost all of these listed genes involved in lipid metabolism is regulated by peroxisome proliferator activated receptors (PPARs), which suggests that the core of these expression alterations may be changes in PPAR expression or protein function.59,63,73–75 This hypothesis is also supported by the finding that ApoE-deficient mice have reduced expression of B-FABP and PPAR-alpha, further highlighting the central role of PPAR in the lipid transcript network.76 Furthermore, significant experimental evidence in non-neural tissue supports this hypothesis. For example, inhibition of Notch1 function decreased the expression of PPARdelta and PPARgamma in 3T3-L1 fibroblasts.77 Similarly,
Jagged-1, a soluble Notch1 ligand, induced maturation of human keratinocytes through PPARgamma.78 Although evidence for the interaction of Notch1 and PPAR in the nervous system is lacking, both literature findings and our data strongly argue that PS1 regulates lipid metabolism through the interaction of Notch1 and downstream activation of PPARs. Absence of PS1 affects expression of genes associated with aging and AD pathology Our microarray screen of PS1-deficient embryos identified decreases in Mt3, ApoE, tau, neurofilament and Notch transcripts, genes that have been associated with the pathophysiology of aging, dementia and AD.17,29,79–85 Furthermore, a microarray study of aging mice identified a subset of genes that show similar expression changes to those we observed in the PS1-deficient brains: stearoyl-CoA desaturase, lissencephaly-1 protein, ApoE and prostaglandin D2 synthetase.86 In addition, A-X actin is upregulated in PS1-deficient mice as well as in the brain of aged mice Molecular Psychiatry
874
Molecular Psychiatry
Table 3
Cross-correlation of PS1-dependent expression levels 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
1
1.00
0.42
0.16
0.65
0.64
0.50
0.14
0.49
0.16
0.51
0.60
0.70
0.66
0.38
-0.53
-0.34
-0.38
-0.46
-0.60
-0.35
-0.45
-0.10
0.57
0.12
0.34
0.39
0.56
2
0.42
1.00
-0.50
-0.03
-0.15
-0.36
-0.68
-0.29
-0.23
-0.21
0.03
0.10
-0.08
0.27
-0.33
-0.85
0.39
-0.36
-0.68
-0.87
-0.64
-0.80
0.52
-0.69
0.96
-0.40
-0.01
Serpine2
3
0.16
-0.50
1.00
0.49
0.29
0.23
0.76
0.63
0.46
0.81
0.55
0.59
0.75
0.17
-0.29
0.48
-0.63
0.12
-0.09
0.54
-0.17
0.78
0.10
0.74
-0.42
0.45
0.24
Cte1 Mt3 Ptprd Chl1 Tnc Vcam1 Ptprz Fabp7 Igfbp5 Lxn Sncg Pea15 Math2a Enc1a Hnf3a Alasb Tac1b Sprr1bb Maptc Sall3c Pcp4c Hes5d ApoEd Dbid
4
0.65
-0.03
0.49
1.00
0.80
0.59
0.64
0.84
0.46
0.84
0.65
0.87
0.75
-0.06
-0.13
0.20
-0.55
0.12
-0.23
0.06
-0.14
0.49
0.21
0.65
-0.08
0.72
0.86
5
0.64
-0.15
0.29
0.80
1.00
0.83
0.63
0.70
0.18
0.66
0.59
0.76
0.59
0.05
0.11
0.21
-0.65
0.01
-0.01
0.14
0.22
0.42
0.09
0.62
-0.27
0.83
0.86
6
0.50
-0.36
0.23
0.59
0.83
1.00
0.60
0.64
0.09
0.46
0.56
0.40
0.46
0.04
0.08
0.25
-0.53
-0.15
-0.01
0.27
0.24
0.42
0.05
0.56
-0.44
0.71
0.71
7
0.14
-0.68
0.76
0.64
0.63
0.60
1.00
0.76
0.40
0.74
0.46
0.58
0.54
-0.07
0.02
0.69
-0.71
0.27
0.19
0.72
0.23
0.94
-0.10
0.97
-0.65
0.73
0.51
8
0.49
-0.29
0.63
0.84
0.70
0.64
0.76
1.00
0.57
0.89
0.75
0.69
0.66
-0.07
-0.25
0.43
-0.50
0.24
-0.18
0.29
-0.12
0.62
0.08
0.81
-0.27
0.86
0.78
9
0.16
-0.23
0.46
0.46
0.18
0.09
0.40
0.57
1.00
0.55
0.10
0.34
0.33
-0.55
-0.45
0.44
0.11
0.38
-0.08
0.31
-0.12
0.44
-0.20
0.46
-0.14
0.48
0.16
10
0.51
-0.21
0.81
0.84
0.66
0.46
0.74
0.89
0.55
1.00
0.77
0.86
0.83
0.07
-0.30
0.36
-0.62
0.22
-0.26
0.26
-0.23
0.64
0.15
0.76
-0.20
0.74
0.70
11
0.60
0.03
0.55
0.65
0.59
0.56
0.46
0.75
0.10
0.77
1.00
0.67
0.77
0.45
-0.23
-0.10
-0.50
-0.22
-0.52
-0.12
-0.40
0.25
0.53
0.45
0.07
0.61
0.71
12
0.70
0.10
0.59
0.87
0.76
0.40
0.58
0.69
0.34
0.86
0.67
1.00
0.84
0.26
-0.24
0.05
-0.67
0.00
-0.30
-0.03
-0.22
0.40
0.37
0.54
0.06
0.64
0.70
13
0.66
-0.08
0.75
0.75
0.59
0.46
0.54
0.66
0.33
0.83
0.77
0.84
1.00
0.41
-0.34
0.01
-0.65
-0.26
-0.36
-0.01
-0.29
0.42
0.50
0.48
-0.05
0.56
0.53
14
0.38
0.27
0.17
-0.06
0.05
0.04
-0.07
-0.07
-0.55
0.07
0.45
0.26
0.41
1.00
-0.23
-0.53
-0.42
-0.69
-0.39
-0.37
-0.35
-0.23
0.77
-0.20
0.33
-0.08
-0.04
15
-0.53
-0.33
-0.29
-0.13
0.11
0.08
0.02
-0.25
-0.45
-0.30
-0.23
-0.24
-0.34
-0.23
1.00
0.20
-0.06
0.32
0.69
0.11
0.76
0.09
-0.44
0.06
-0.41
0.05
0.14
16
-0.34
-0.85
0.48
0.20
0.21
0.25
0.69
0.43
0.44
0.36
-0.10
0.05
0.01
-0.53
0.20
1.00
-0.37
0.73
0.64
0.89
0.54
0.84
-0.74
0.76
-0.86
0.43
0.17
17
-0.38
0.39
-0.63
-0.55
-0.65
-0.53
-0.71
-0.50
0.11
-0.62
-0.50
-0.67
-0.65
-0.42
-0.06
-0.37
1.00
0.00
-0.15
-0.39
-0.17
-0.64
-0.11
-0.66
0.47
-0.48
-0.52
18
-0.46
-0.36
0.12
0.12
0.01
-0.15
0.27
0.24
0.38
0.22
-0.22
0.00
-0.26
-0.69
0.32
0.73
0.00
1.00
0.52
0.44
0.40
0.43
-0.79
0.40
-0.40
0.24
0.20
19 20 21
-0.60
-0.68
-0.09
-0.23
-0.01
-0.01
0.19
-0.18
-0.08
-0.26
-0.52
-0.30
-0.36
-0.39
0.69
0.64
-0.15
0.52
1.00
0.50
0.93
0.36
-0.72
0.25
-0.73
0.12
-0.14
-0.35
-0.87
0.54
0.06
0.14
0.27
0.72
0.29
0.31
0.26
-0.12
-0.03
-0.01
-0.37
0.11
0.89
-0.39
0.44
0.50
1.00
0.43
0.85
-0.59
0.73
-0.87
0.26
-0.02
-0.45
-0.64
-0.17
-0.14
0.22
0.24
0.23
-0.12
-0.12
-0.23
-0.40
-0.22
-0.29
-0.35
0.76
0.54
-0.17
0.40
0.93
0.43
1.00
0.32
-0.64
0.25
-0.71
0.28
0.01
22 23 24
-0.10
-0.80
0.78
0.49
0.42
0.42
0.94
0.62
0.44
0.64
0.25
0.40
0.42
-0.23
0.09
0.84
-0.64
0.43
0.36
0.85
0.32
1.00
-0.32
0.93
-0.78
0.56
0.33
0.57
0.52
0.10
0.21
0.09
0.05
-0.10
0.08
-0.20
0.15
0.53
0.37
0.50
0.77
-0.44
-0.74
-0.11
-0.79
-0.72
-0.59
-0.64
-0.32
1.00
-0.22
0.63
-0.02
0.06
0.12
-0.69
0.74
0.65
0.62
0.56
0.97
0.81
0.46
0.76
0.45
0.54
0.48
-0.20
0.06
0.76
-0.66
0.40
0.25
0.73
0.25
0.93
-0.22
1.00
-0.68
0.75
0.56
25 26 27
0.34
0.96
-0.42
-0.08
-0.27
-0.44
-0.65
-0.27
-0.14
-0.20
0.07
0.06
-0.05
0.33
-0.41
-0.86
0.47
-0.40
-0.73
-0.87
-0.71
-0.78
0.63
-0.68
1.00
-0.38
-0.12
0.39
-0.40
0.45
0.72
0.83
0.71
0.73
0.86
0.48
0.74
0.61
0.64
0.56
-0.08
0.05
0.43
-0.48
0.24
0.12
0.26
0.28
0.56
-0.02
0.75
-0.38
1.00
0.73
0.56
-0.01
0.24
0.86
0.86
0.71
0.51
0.78
0.16
0.70
0.71
0.70
0.53
-0.04
0.14
0.17
-0.52
0.20
-0.14
-0.02
0.01
0.33
0.06
0.56
-0.12
0.73
1.00
a Increased in E12.5 only
c Decreased in E12.5 only
b Increased in E14.5 only
d Decreased in E14.5 only
correlation r= 0.70 to 0.80 correlation r> 0.80
correlation r= -0.70 to -0.80 correlation r< -0.80
Expression profiling of PS1-deficient brain ZK Mirnics et al
Gene #
Sfrp2 Notch1
Expression profiling of PS1-deficient brain ZK Mirnics et al
findings arising from our data set, animal models of aging and AD models, represents a novel, data-driven approach that is a promising starting point for forming novel hypotheses related to the role of PS1 in development and AD pathology.
875
Coregulation of PS1-dependent transcripts
Figure 5 Differentiation and lipid metabolism transcript networks are PS1-dependent. Symbols in box denote transcripts, solitary symbols denote proteins. Genes with decreased transcript levels are denoted with blue boxes; genes with increased expression are denoted in red boxes. White boxes represent gene products that could not be meaningfully assayed in our experiment (not expressed or absent probes). Full arrows denote already known regulatory relationship between genes (but not necessarily in the CNS), dashed arrows denote proposed interactions. Note that the absence of PS1 transcript triggers a complex cascade of inter-related transcript changes, what argues that these transcript changes have functional consequences at the protein level. (a) In PS1-deficient animals, differentiation is accelerated via a Notch–Hes-dependent mechanism. In addition, we propose that Foxg1 upregulation is resulting in upregulation of Tbr1 and Tbr2, as well as in a related downregulation of Bfap7 and En2 transcripts. (b) In the absence of PS1, increased differentiation and decreased proliferation are most likely regulated through the PS1– Notch1–Hes axis, while changes in lipid metabolism are most likely regulated via the PS1–Notch1–NFkB–PPAR system.
and in AD animal models.86 Finally, increases in Abca1 transporter expression and cholesterol efflux in brain cells and decreased Ab secretion may represent a mechanism to reduce amyloid burden in the brain in the future.87 In summary, the overlap in
In addition to participating in common cellular functions, many of the genes are coregulated within the same brain sample. On the other hand, expression of two genes between two conditions may be changed, but their expression may not be coregulated within each of the samples. In our study, we provide statistical evidence that many of the genes were coregulated within the samples regardless of the presence or absence of the PS1 gene or the age of the embryos. These coregulations are also consistent with literature reports (see also previous sections), including known coregulations between genes that belong to the same physiological cascade (Notch1 and Hes5; Dbi and Bfabp, Ptprz and Tnc, etc).88–90 Furthermore, using human DNA microarrays we found that some aspects of the coregulation of this PS1-dependent network appear to be preserved across the rodent brain and adult human prefrontal cortex. For example, both in the embryonic mice brain and the human prefrontal cortex, ApoE expression levels were coregulated with Mt3 (r¼0.83 and 0.81), Ptprz (r¼0.74 and 0.62) and Dbi (r¼0.73 and 0.65), and Dbi was also coregulated with Tnc (r¼0.78 and 0.60) and Mt3 (r¼0.83 and 0.81). These data suggest that at least some coregulations we uncovered are age- and species-independent, emphasizing the translational nature of combined genetic–genomic studies. However, it is important to note that many of the highly correlated expression decreases reported in Table 3 have not been described previously. These potential interactions between the presented genes can be tested at a protein level, where the direct interactions can be separated out from secondary and compensatory changes. These finding show the power of using transgenic animal models to tease out important coregulations of transcript networks. Perhaps more importantly, the transcript network we identified includes several genes that are known to participate in the pathophysiological process of AD—and perhaps some other important ones that we are not aware of to date.
Acknowledgements We thank Dr Pat Levitt for useful comments on the experimental design and manuscript, as well as Ms Deborah Hollingshead for the skillful microarray hybridizations. This work is supported by CHP Research Advisory Committee and Child Neurology Training Grant NS07495-01A1 and PIND Grant (ZKM), Carol Ann Craumer Endowment of CHP (NFS), 2002 NARSAD Young Investigator Award Molecular Psychiatry
Expression profiling of PS1-deficient brain ZK Mirnics et al
876
(KM) and Ellison Medical Foundation Senior Scholar Award (SSS) and The Fidelity Foundation (SSS).
21
References 1 De Strooper B, Saftig P, Craessaerts K, Vanderstichele H, Guhde G, Annaert W et al. Deficiency of presenilin-1 inhibits the normal cleavage of amyloid precursor protein. Nature 1998; 391: 387–390. 2 Song W, Nadeau P, Yuan M, Yang X, Shen J, Yankner BA. Proteolytic release and nuclear translocation of Notch-1 are induced by presenilin-1 and impaired by pathogenic presenilin-1 mutations. Proc Natl Acad Sci USA 1999; 96: 6959–6963. 3 Ni CY, Murphy MP, Golde TE, Carpenter G. gamma-Secretase cleavage and nuclear localization of ErbB-4 receptor tyrosine kinase. Science 2001; 294: 2179–2181. 4 Sisodia SS, Annaert W, Kim SH, De Strooper B. Gamma-secretase: never more enigmatic. Trends Neurosci 2001; 24: S2–S6. 5 Selkoe DJ. Presenilin, Notch, and the genesis and treatment of Alzheimer’s disease. Proc Natl Acad Sci USA 2001; 98: 11039– 11041. 6 Kim DY, Ingano LA, Kovacs DM. Nectin-1 alpha, an immunoglobulin-like receptor involved in the formation of synapses, is a substrate for presenilin/gamma-secretase-like cleavage. J Biol Chem 2002; 277: 49976–49981. 7 Kimberly WT, LaVoie MJ, Ostaszewski BL, Ye W, Wolfe MS, Selkoe DJ. Complex N-linked glycosylated nicastrin associates with active gamma-secretase and undergoes tight cellular regulation. J Biol Chem 2002; 277: 35113–35117. 8 Lammich S, Okochi M, Takeda M, Kaether C, Capell A, Zimmer AK et al. Presenilin-dependent intramembrane proteolysis of CD44 leads to the liberation of its intracellular domain and the secretion of an Abeta-like peptide. J Biol Chem 2002; 277: 44754– 44759. 9 Marambaud P, Shioi J, Serban G, Georgakopoulos A, Sarner S, Nagy V et al. A presenilin-1/gamma-secretase cleavage releases the E-cadherin intracellular domain and regulates disassembly of adherens junctions. EMBO J 2002; 21: 1948–1956. 10 May P, Reddy YK, Herz J. Proteolytic processing of low density lipoprotein receptor-related protein mediates regulated release of its intracellular domain. J Biol Chem 2002; 277: 18736–18743. 11 Sisodia SS, St George-Hyslop PH. Gamma-secretase, Notch, Abeta and Alzheimer’s disease: where do the presenilins fit in? Nat Rev Neurosci 2002; 3: 281–290. 12 Ikeuchi T, Sisodia SS. The Notch ligands, delta1 and Jagged2, are substrates for presenilin-dependent ‘gamma-secretase’ cleavage. J Biol Chem 2003; 278: 7751–7754. 13 Wolfe MS, De Los Angeles J, Miller DD, Xia W, Selkoe DJ. Are presenilins intramembrane-cleaving proteases? Implications for the molecular mechanism of Alzheimer’s disease. Biochemistry 1999; 38: 11223–11230. 14 Kimberly WT, Esler WP, Ye W, Ostaszewski BL, Gao J, Diehl T et al. Notch and the amyloid precursor protein are cleaved by similar gamma- secretase(s). Biochemistry 2003; 42: 137–144. 15 Scheuner D, Eckman C, Jensen M, Song X, Citron M, Suzuki N et al. Secreted amyloid beta-protein similar to that in the senile plaques of Alzheimer’s disease is increased in vivo by the presenilin 1 and 2 and APP mutations linked to familial Alzheimer’s disease. Nat Med 1996; 2: 864–870. 16 Borchelt DR, Thinakaran G, Eckman CB, Lee MK, Davenport F, Ratovitsky T et al. Familial Alzheimer’s disease-linked presenilin 1 variants elevate Abeta1-42/1-40 ratio in vitro and in vivo. Neuron 1996; 17: 1005–1013. 17 Price DL, Sisodia SS. Mutant genes in familial Alzheimer’s disease and transgenic models. Annu Rev Neurosci 1998; 21: 479–505. 18 Citron M, Eckman CB, Diehl TS, Corcoran C, Ostaszewski BL, Xia W et al. Additive effects of PS1 and APP mutations on secretion of the 42-residue amyloid beta-protein. Neurobiol Dis 1998; 5: 107–116. 19 Shen J, Bronson RT, Chen DF, Xia W, Selkoe DJ, Tonegawa S. Skeletal and CNS defects in presenilin-1-deficient mice. Cell 1997; 89: 629–639. 20 Wong PC, Zheng H, Chen H, Becher MW, Sirinathsinghji DJ, Trumbauer ME et al. Presenilin 1 is required for Notch1 and Molecular Psychiatry
22
23
24
25
26
27
28
29
30
31
32
33
34 35
36
37
38
39
40 41
DII1 expression in the paraxial mesoderm. Nature 1997; 387: 288–292. Wong PC, Borchelt DR, Lee MK, Pardo CA, Thinakaran G, Martin LJ et al. Familial amyotrophic lateral sclerosis and Alzheimer’s disease. Transgenic models. Adv Exp Med Biol 1998; 446: 145–159. Davis JA, Naruse S, Chen H, Eckman C, Younkin S, Price DL et al. An Alzheimer’s disease-linked PS1 variant rescues the developmental abnormalities of PS1-deficient embryos. Neuron 1998; 20: 603–609. Hartmann D, De Strooper B, Saftig P. Presenilin-1 deficiency leads to loss of Cajal–Retzius neurons and cortical dysplasia similar to human type 2 lissencephaly. Curr Biol 1999; 9: 719–727. Herreman A, Hartmann D, Annaert W, Saftig P, Craessaerts K, Serneels L et al. Presenilin 2 deficiency causes a mild pulmonary phenotype and no changes in amyloid precursor protein processing but enhances the embryonic lethal phenotype of presenilin 1 deficiency. Proc Natl Acad Sci USA 1999; 96: 11872–11877. Handler M, Yang X, Shen J. Presenilin-1 regulates neuronal differentiation during neurogenesis. Development 2000; 127: 2593–2606. Herreman A, Serneels L, Annaert W, Collen D, Schoonjans L, De Strooper B. Total inactivation of gamma-secretase activity in presenilin-deficient embryonic stem cells. Nat Cell Biol 2000; 2: 461–462. Saftig P, Hartmann D, De Strooper B. The function of presenilin-1 in amyloid beta-peptide generation and brain development. Eur Arch Psychiatry Clin Neurosci 1999; 249: 271–279. Schwarzman AL, Singh N, Tsiper M, Gregori L, Dranovsky A, Vitek MP et al. Endogenous presenilin 1 redistributes to the surface of lamellipodia upon adhesion of Jurkat cells to a collagen matrix. Proc Natl Acad Sci USA 1999; 96: 7932–7937. Johnsingh AA, Johnston JM, Merz G, Xu J, Kotula L, Jacobsen JS et al. Altered binding of mutated presenilin with cytoskeletoninteracting proteins. FEBS Lett 2000; 465: 53–58. Benussi L, Alberici A, Mayhaus M, Langer U, Ghidoni R, Mazzoli F et al. Detection of the presenilin 1 COOH-terminal fragment in the extracellular compartment: a release enhanced by apoptosis. Exp Cell Res 2001; 269: 256–265. Wirths O, Multhaup G, Czech C, Blanchard V, Tremp G, Pradier L et al. Reelin in plaques of beta-amyloid precursor protein and presenilin-1 double-transgenic mice. Neurosci Lett 2001; 316: 145–148. Song C, Perides G, Wang D, Liu YF. Beta-amyloid peptide induces formation of actin stress fibers through p38 mitogen-activated protein kinase. J Neurochem 2002; 83: 828–836. Liauw J, Nguyen V, Huang J, St George-Hyslop P, Rozmahel R. Differential display analysis of presenilin 1-deficient mouse brains. Brain Res Mol Brain Res 2002; 109: 56–62. Lipshutz RJ, Fodor SP, Gingeras TR, Lockhart DJ. High density synthetic oligonucleotide arrays. Nat Genet 1999; 21: 20–24. Liu G, Loraine AE, Shigeta R, Cline M, Cheng J, Valmeekam V et al. NetAffx: Affymetrix probesets and annotations. Nucleic Acids Res 2003; 31: 82–86. Califano A, Stolovitzky G, Tu Y. Analysis of gene expression microarrays for phenotype classification. Proc Int Conf Intell Syst Mol Biol 2000; 8: 75–85. Mirnics K, Middleton FA, Marquez A, Lewis DA, Levitt P. Molecular characterization of schizophrenia viewed by microarray analysis of gene expression in prefrontal cortex. Neuron 2000; 28: 53–67. Mirnics K, Middleton FA, Lewis DA, Levitt P. Analysis of complex brain disorders with gene expression microarrays: schizophrenia as a disease of the synapse. Trends Neurosci 2001; 24: 479–486. Mirnics K, Middleton FA, Lewis DA, Levitt P. The human genome: gene expression profiling and schizophrenia. Am J Psychiatry 2001; 158: 1384. Mirnics K, Lewis DA. Genes and subtypes of schizophrenia. Trends Mol Med 2001; 7: 281–283. Masliah E, Mallory M, Veinbergs I, Miller A, Samuel W. Alterations in apolipoprotein E expression during aging and neurodegeneration. Prog Neurobiol 1996; 50: 493–503.
Expression profiling of PS1-deficient brain ZK Mirnics et al 42 Price DL, Wong PC, Borchelt DR, Pardo CA, Thinakaran G, Doan AP et al. Amyotrophic lateral sclerosis and Alzheimer disease. Lessons from model systems. Rev Neurol (Paris) 1997; 153: 484–495. 43 Dowjat WK, Wisniewski H, Wisniewski T. Alzheimer’s disease presenilin-1 expression modulates the assembly of neurofilaments. Neuroscience 2001; 103: 1–8. 44 Colangelo V, Schurr J, Ball MJ, Pelaez RP, Bazan NG, Lukiw WJ. Gene expression profiling of 12633 genes in Alzheimer hippocampal CA1: transcription and neurotrophic factor down-regulation and up-regulation of apoptotic and pro-inflammatory signaling. J Neurosci Res 2002; 70: 462–473. 45 Morfini G, Pigino G, Beffert U, Busciglio J, Brady ST. Fast axonal transport misregulation and Alzheimer’s disease. Neuromolecular Med 2002; 2: 89–99. 46 Kageyama R, Ishibashi M, Takebayashi K, Tomita K. bHLH transcription factors and mammalian neuronal differentiation. Int J Biochem Cell Biol 1997; 29: 1389–1399. 47 Lambert de Rouvroit C, Goffinet AM. A new view of early cortical development. Biochem Pharmacol 1998; 56: 1403–1409. 48 Ueno M, Kimura N, Nakashima K, Saito-Ohara F, Inazawa J, Taga T. Genomic organization, sequence and chromosomal localization of the mouse Tbr2 gene and a comparative study with Tbr1. Gene 2000; 254: 29–35. 49 Zhou H, Hughes DE, Major ML, Yoo K, Pesold C, Costa RH. Atypical mouse cerebellar development is caused by ectopic expression of the forkhead box transcription factor HNF-3beta. Gene Express 2001; 9: 217–236. 50 Hsueh YP, Wang TF, Yang FC, Sheng M. Nuclear translocation and transcription regulation by the membrane-associated guanylate kinase CASK/LIN-2. Nature 2000; 404: 298–302. 51 Hiesberger T, Trommsdorff M, Howell BW, Goffinet A, Mumby MC, Cooper JA et al. Direct binding of reelin to VLDL receptor and ApoE receptor 2 induces tyrosine phosphorylation of disabled-1 and modulates tau phosphorylation. Neuron 1999; 24: 481–489. 52 Ohtsuka T, Ishibashi M, Gradwohl G, Nakanishi S, Guillemot F, Kageyama R. Hes1 and Hes5 as notch effectors in mammalian neuronal differentiation. EMBO J 1999; 18: 2196–2207. 53 Kageyama R, Ohtsuka T. The Notch-Hes pathway in mammalian neural development. Cell Res 1999; 9: 179–188. 54 Kabos P, Kabosova A, Neuman T. Blocking HES1 expression initiates GABAergic differentiation and induces the expression of p21(CIP1/WAF1) in human neural stem cells. J Biol Chem 2002; 277: 8763–8766. 55 Bothwell M, Giniger E. Alzheimer’s disease: neurodevelopment converges with neurodegeneration. Cell 2000; 102: 271–273. 56 Fairen A, Morante-Oria J, Frassoni C. The surface of the developing cerebral cortex: still special cells one century later. Prog Brain Res 2002; 136: 281–291. 57 Sobeih MM, Corfas G. Extracellular factors that regulate neuronal migration in the central nervous system. Int J Dev Neurosci 2002; 20: 349–357. 58 Halfter W, Dong S, Yip YP, Willem M, Mayer U. A critical function of the pial basement membrane in cortical histogenesis. J Neurosci 2002; 22: 6029–6040. 59 Helledie T, Antonius M, Sorensen RV, Hertzel AV, Bernlohr DA, Kolvraa S et al. Lipid-binding proteins modulate ligand-dependent trans-activation by peroxisome proliferator-activated receptors and localize to the nucleus as well as the cytoplasm. J Lipid Res 2000; 41: 1740–1751. 60 Knudsen J, Neergaard TB, Gaigg B, Jensen MV, Hansen JK. Role of acyl-CoA binding protein in acyl-CoA metabolism and acyl-CoAmediated cell signaling. J Nutr 2000; 130: 294S–298S. 61 Schjerling CK, Hummel R, Hansen JK, Borsting C, Mikkelsen JM, Kristiansen K et al. Disruption of the gene encoding the acyl-CoAbinding protein (ACB1) perturbs acyl-CoA metabolism in Saccharomyces cerevisiae. J Biol Chem 1996; 271: 22514–22521. 62 Hunt MC, Nousiainen SE, Huttunen MK, Orii KE, Svensson LT, Alexson SE. Peroxisome proliferator-induced long chain acyl-CoA thioesterases comprise a highly conserved novel multi-gene family involved in lipid metabolism. J Biol Chem 1999; 274: 34317–34326. 63 Wolfrum C, Borrmann CM, Borchers T, Spener F. Fatty acids and hypolipidemic drugs regulate peroxisome proliferator-activated receptors alpha- and gamma-mediated gene expression via liver
64
65
66
67
68
69
70 71
72
73
74
75 76
77
78
79
80
81
82 83
84
fatty acid binding protein: a signaling path to the nucleus. Proc Natl Acad Sci USA 2001; 98: 2323–2328. Ross AC. Cellular metabolism and activation of retinoids: roles of cellular retinoid-binding proteins. FASEB J 1993; 7: 317– 327. Hellemans K, Rombouts K, Quartier E, Dittie AS, Knorr A, Michalik L et al. PPAR{beta} regulates vitamin A metabolismrelated gene expression in hepatic stellate cells undergoing activation. J Lipid Res 2003; 44: 280–295. Nimpf J, Schneider WJ. From cholesterol transport to signal transduction: low density lipoprotein receptor, very low density lipoprotein receptor, and apolipoprotein E receptor-2. Biochim Biophys Acta 2000; 1529: 287–298. Fujino T, Kondo J, Ishikawa M, Morikawa K, Yamamoto TT. Acetyl–CoA synthetase 2, a mitochondrial matrix enzyme involved in the oxidation of acetate. J Biol Chem 2001; 276: 11420–11426. Santamarina-Fojo S, Remaley AT, Neufeld EB, Brewer Jr HB. Regulation and intracellular trafficking of the ABCA1 transporter. J Lipid Res 2001; 42: 1339–1345. Miyazaki M, Ntambi JM. Role of stearoyl-coenzyme A desaturase in lipid metabolism. Prostaglandins Leukot Essent Fatty Acids 2003; 68: 113–121. Brousseau ME. ATP-binding cassette transporter A1, fatty acids, and cholesterol absorption. Curr Opin Lipidol 2003; 14: 35–40. Joyce C, Freeman L, Brewer Jr HB, Santamarina-Fojo S. Study of ABCA1 function in transgenic mice. Arterioscler Thromb Vasc Biol 2003; 2: 2. Fujimori K, Fujitani Y, Kadoyama K, Kumanogoh H, Ishikawa K, Urade Y. Regulation of lipocalin-type prostaglandin D synthase gene expression by Hes-1 through E-box and interleukin-1 beta via two NF-kappa B elements in rat leptomeningeal cells. J Biol Chem 2003; 278: 6018–6026. Kainu T, Wikstrom AC, Gustafsson JA, Pelto-Huikko M. Localization of the peroxisome proliferator-activated receptor in the brain. Neuroreport 1994; 5: 2481–2485. Xing G, Zhang L, Heynen T, Yoshikawa T, Smith M, Weiss S et al. Rat PPAR delta contains a CGG triplet repeat and is prominently expressed in the thalamic nuclei. Biochem Biophys Res Commun 1995; 217: 1015–1025. Gonzalez FJ. Recent update on the PPAR alpha-null mouse. Biochimie 1997; 79: 139–144. Hung MC, Hayase K, Yoshida R, Sato M, Imaizumi K. Cerebral protein kinase C and its mRNA level in apolipoprotein E-deficient mice. Life Sci 2001; 69: 1419–1427. Garces C, Ruiz-Hidalgo MJ, de Mora JF, Park C, Miele L, Goldstein J et al. Notch-1 controls the expression of fatty acid-activated transcription factors and is required for adipogenesis. J Biol Chem 1997; 272: 29729–29734. Nickoloff BJ, Qin JZ, Chaturvedi V, Denning MF, Bonish B, Miele L. Jagged-1 mediated activation of notch signaling induces complete maturation of human keratinocytes through NF-kappaB and PPARgamma. Cell Death Differ 2002; 9: 842–855. Goedert M, Sisodia SS, Price DL. Neurofibrillary tangles and betaamyloid deposits in Alzheimer’s disease. Curr Opin Neurobiol 1991; 1: 441–447. Sisodia SS, Martin LJ, Walker LC, Borchelt DR, Price DL. Cellular and molecular biology of Alzheimer’s disease and animal models. Neuroimaging Clin N Am 1995; 5: 59–68. Ebneth A, Godemann R, Stamer K, Illenberger S, Trinczek B, Mandelkow E. Overexpression of tau protein inhibits kinesindependent trafficking of vesicles, mitochondria, and endoplasmic reticulum: implications for Alzheimer’s disease. J Cell Biol 1998; 143: 777–794. Neely MD, Montine TJ. Csf lipoproteins and Alzheimer s disease. J Nutr Health Aging 2002; 6: 383–391. Chung RS, Vickers JC, Chuah MI, Eckhardt BL, West AK. Metallothionein-III inhibits initial neurite formation in developing neurons as well as postinjury, regenerative neurite sprouting. Exp Neurol 2002; 178: 1–12. Miyazaki I, Asanuma M, Higashi Y, Sogawa CA, Tanaka K, Ogawa N. Age-related changes in expression of metallothionein-III in rat brain. Neurosci Res 2002; 43: 323–333.
877
Molecular Psychiatry
Expression profiling of PS1-deficient brain ZK Mirnics et al
878
85 Uchida Y, Gomi F, Masumizu T, Miura Y. Growth inhibitory factor prevents neurite extension and the death of cortical neurons caused by high oxygen exposure through hydroxyl radical scavenging. J Biol Chem 2002; 277: 32353–32359. 86 Jiang CH, Tsien JZ, Schultz PG, Hu Y. The effects of aging on gene expression in the hypothalamus and cortex of mice. Proc Natl Acad Sci USA 2001; 98: 1930–1934. 87 Koldamova RP, Lefterov IM, Ikonomovic MD, Skoko J, Lefterov PI, Isanski BA et al. 22R-Hydroxycholesterol and 9-cis-retinoic acid induce ABCA1 transporter expression and cholesterol efflux in brain cells and decrease Abeta secretion. J Biol Chem 2003; 22: 22.
Molecular Psychiatry
88 Milev P, Chiba A, Haring M, Rauvala H, Schachner M, Ranscht B et al. High affinity binding and overlapping localization of neurocan and phosphacan/protein-tyrosine phosphatase-zeta/beta with tenascin-R, amphoterin, and the heparin-binding growthassociated molecule. J Biol Chem 1998; 273: 6998–7005. 89 Yanase H, Shimizu H, Yamada K, Iwanaga T. Cellular localization of the diazepam binding inhibitor in glial cells with special reference to its coexistence with brain-type fatty acid binding protein. Arch Histol Cytol 2002; 65: 27–36. 90 Iso T, Kedes L, Hamamori Y. HES and HERP families: multiple effectors of the notch signaling pathway. J Cell Physiol 2003; 194: 237–255.
