Send Orders for Reprints to [email protected] 478 Current Neurovascular Research, 2019, 16, 478-490 RESEARCH ARTICLE Combined Transcriptomic and Proteomic Analyses of Cerebral Frontal Lobe Tissue Identified RNA Metabolism Dysregulation as One Potential Pathogenic Mechanism in Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy (CADASIL)

Marie-Françoise Ritz1,*, Paul Jenoe2, Leo Bonati3, Stefan Engelter3,4, Philippe Lyrer3 and Nils Peters3,4

1Department of Biomedicine, Brain Tumor Biology Laboratory, University of Basel, and University Hospital of Basel, Hebelstrasse 20, 4031 Basel, Switzerland; 2Proteomics Core Facility, Biocenter, University of Basel, Klingelbergstrasse 50/70, 4056 Basel, Switzerland; 3Department of Neurology and Stroke Center, University Hospital Basel and University of Basel, Petersgraben 4, 4031 Basel, Switzerland; 4Neurorehabilitation Unit, University of Basel and University Center for Medicine of Aging, Felix Platter Hospital, Burgfelderstrasse 101, 4055 Basel, Switzerland

Abstract: Background: Cerebral small vessel disease (SVD) is an important cause of stroke and vascular cognitive impairment (VCI), leading to subcortical ischemic vascular dementia. As a he- reditary form of SVD with early onset, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) represents a pure form of SVD and may thus serve as a model disease for SVD. To date, underlying molecular mechanisms linking vascular pathol- ogy and subsequent neuronal damage in SVD are incompletely understood. A R T I C L E H I S T O R Y Objective: We performed comparative transcriptional profiling microarray and proteomic analyses Received: October 01, 2019 on post-mortem frontal lobe specimen from 2 CADASIL patients and 5 non neurologically dis- Revised: October 11, 2019 Accepted: October 15, 2019 eased controls in order to identify dysregulated pathways potentially involved in the development

DOI: of tissue damage in CADASIL. 10.2174/1567202616666191023111059 Methods: Transcriptional microarray analysis of material extracted from frontal grey and white matter (WM) identified subsets of up- or down-regulated genes enriched into biological pathways mostly in WM areas. Proteomic analysis of these regions also highlighted cellular processes identi- fied by dysregulated proteins. Results: Discrepancies between proteomic and transcriptomic data were observed, but a number of pathways were commonly associated with genes and corresponding proteins, such as: “ribosome” identified by upregulated genes and proteins in frontal cortex or “spliceosome” associated with down-regulated genes and proteins in frontal WM. Conclusion: This latter finding suggests that defective expression of spliceosomal components may alter widespread splicing profile, potentially inducing expression abnormalities that could contribute to cerebral WM damage in CADASIL.

Keywords: CADASIL, transcriptomic, proteomic, pathomechanisms, spliceosome, ribosome.

1. INTRODUCTION dominant arteriopathy with subcortical infarcts and leukoen- cephalopathy (CADASIL) is the most common hereditary Small vessel disease (SVD) of the brain is a common form of SVD [2], caused by mutations in NOTCH3 gene on condition responsible for approximately 25% of all strokes chromosome 19q12, which encodes a transmembrane recep- and 40% of vascular dementia [1]. Cerebral autosomal tor playing a crucial role within vascular smooth muscle and

pericyte signaling pathway [3]. Clinical manifestations ap- pearing in mid adulthood (between 30 and 40 years of age) *Address correspondence to this author at the University of Basel and comprise migraine with aura, lacunar stroke and vascular University Hospital of Basel, Department of Biomedicine, Brain Tumor cognitive impairment. Upon MR-imaging, the disease is Biology Laboratory, Hebelstrasse 20, 4031 Basel, Switzerland; E-mail: [email protected] characterized by typical SVD related lesions including dif-

1567-2026/19 $58.00+.00 © 2019 Bentham Science Publishers Transcriptomic and Proteomic Analyses of CADASIL Brains Current Neurovascular Research, 2019, Vol. 16, No. 5 479 fuse ischemic white matter lesions (WML), lacunar infarcts logical disease - in particular sporadic SVD, CADASIL or and microbleeds and eventually brain atrophy [4-6]. neurodegenerative diseases - were obtained from the Insti- tute of Pathology, University Hospital Basel and were used The extracellular domain of the mutated Notch3 receptor as healthy controls (72 ± 15 years, M/F: 4/1). The use of accumulates in the membrane of blood vessels, most likely postmortem material from deceased human subjects as con- leading to a progressive loss of vascular small muscle cells trols was approved by the local Ethics committee (EKBB, (VSMCs), thickening and stenosis of vessel walls by various Basel, Switzerland). Times to autopsy were <24 h post- types of collagens, laminins, and fibronectins, as well as de- mortem for all subjects. There was no significant difference posits of granular osmiophilic material (GOM) [7, 8] and in age between CADASIL and control groups (p=0.204, t- von Willebrand factor [9]. test). Tissue blocks dissected during autopsy were shock To date, the underlying molecular mechanisms linking frozen at -80°C until mRNA or protein extraction. vascular pathology in SVD and subsequent neuronal damage are incompletely understood. Various mechanisms have been 2.2. RNA Preparation and Microarray Processing discussed, including endothelial dysfunction, inflammatory The frozen brain samples were homogenized in Purezol processes, hemostasis, oxidative stress, disturbance of capil- solution (Bio-Rad Laboratories, Hercules, CA, USA) using a lary flow and altered function of the blood-brain-barrier [10, bead type-homogenizer (MM-301, Retsch, Haan, Germany). 11]. We recently studied the transcriptome profiles in post- Isolation of total RNA from the lysate including DNAse I mortem brains affected by sporadic SVD in comparison to treatment was carried out using the Aurum Total RNA Fatty normal brains in order to gain insight in the pathogenesis and Fibrous Tissue kit (BioRad Laboratories) according to [12]. We found regional alterations in gene expression in- the manufacturer’s protocol. The purity and integrity of the volving various pathways, such as inflammation, apoptosis, extracted total RNA were determined using the Agilent 2100 and alterations in lipid metabolism and coagulation. How- Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). ever, given the age association of sporadic SVD, there are All RNA samples had an integrity number > 7. Double potentially confounding alterations present, such as concomi- strand cDNA and labeled cRNA were prepared as previously tant neurodegeneration, having a possible influence on the published [14] and 12 µg fragmented cRNA was used for observed findings. As a hereditary condition with earlier hybridization to the Affymetrix GeneChip® Human Gene 1.0 onset, CADASIL can be regarded as a pure form of SVD, ST Arrays (Affymetrix, Santa Clara, CA, USA). Sample potentially allowing to study the underlying mechanisms of labeling, hybridization, scanning and data outputs were per- SVD in more detail. Therefore, in this study, a similar whole formed according to Affymetrix protocol. Data expression genome analysis was performed in CADASIL brains com- values were collected as individual CEL files. bined with a proteomic approach, in order to identify the correspondence of transcriptional responses to cellular pro- 2.3. Analysis of Gene Expression Data tein abundance. Indeed, proteins are effectors of biological ® function and their levels are not only dependent on mRNA Microarray data were normalized using Partek Genom- TM levels but also on host translational control and regulation ics Suite version 6.6 (Partek Inc., St. Louis, MO, USA).

[13]. Thus, the proteomics should be considered as the most Data analysis including background subtraction, normaliza- relevant data set to characterize a biological system. tion, and elimination of false positives was performed using the default parameters. The principal component analysis The outcome of such an analysis is thought to offer a (PCA) was performed to assess the variation in the expres- comprehensive view of the biological roles of the selected sion of genes among the different samples. The unsupervised genes/proteins through highlighting key pathways and cellu- gene hierarchical clustering containing all the genes differ- lar processes in which they are involved. ently expressed between the two groups was used to estimate similarities in expression patterns between samples. One- Focusing on the changes of expression of genes and pro- teins in the frontal lobe, this work identified deregulation of way analysis of variance (ANOVA) was used to identify the spliceosome, which could produce splicing abnormalities genes differentially expressed between CADASIL and con- leading to generalized protein disbalance. This finding repre- trol samples for each brain region. Criteria for significant sents a yet unsuspected molecular dysfunction in cerebral changes in signal levels were: False Discovery Rate (FDR) small vessel diseases that needs to be further investigated. value < 0.05, and a fold change F (log transformed base 2.0) ≥ 1.2 in both directions. 2. MATERIALS AND METHODS 2.4. Proteomic Analysis 2.1. Human Brain Samples 2.4.1. Protein and Peptide Preparation The use of material from human patients was approved Biopsies were minced into small pieces and homogenized by the local Ethics Committee (Ethikkomission Beider in 8 M urea in 50 mM Tris-HCl, pH 8.0, 75 mM NaCl with a Basel, EKBB, Basel, Switzerland). Kimble Pellet Pestle mixer (Sigma-Aldrich, Buchs, Switzer- Post-mortem brain material (frontal lobes containing gray land) fitted with a 1.5 mL pestle twice for 30 seconds. The and white matter) from two unrelated cases of CADASIL homogenate was pelleted at 12,000 rcf for 10 min and the (mean age ± standard deviation: 66 ± 1 y, M/F: 2/0) were supernatant was collected. Proteins were reduced with 10 obtained from the Neurobiobank München, Germany. Both mM dithiothreitol (DTT) at 55oC for 30 min and alkylated patients died from non-neurological causes. Post-mortem with 50 mM iodoacetamide for 15 min at room temperature brain tissue blocks from five subjects free of known neuro- in the dark. Individual homogenates were desalted on PD-10 480 Current Neurovascular Research, 2019, Vol. 16, No. 5 Ritz et al.

Fig. (1). Labeling strategy for multiplexed quantitative proteomic analysis of brain samples. Pooled control specimen from frontal cortex and frontal white matter (WM) were labeled with the TMT126 and 129 labels, while CADASIL 1 and 2 from frontal cortex and frontal white matter were labeled with TMT127/128 and TMT130/131, respectively. Equal amounts of peptide were pooled, separated by reverse-phase HPLC and analyzed by high-resolution mass spectrometry. (A higher resolution / colour version of this figure is available in the electronic copy of the article). columns (GE Healthcare, Glattbrugg, Switzerland) with 4 M 2.4.2. Mass Spectrometric Analysis urea in 50 mM Tris-HCl, pH 8.0, 75 mM NaCl. The protein- o The individual HPLC pools were analyzed by capillary containing fractions were pooled and stored at -20 C. Pro- liquid chromatography tandem mass spectrometry teins were digested with two rounds of endoproteinase LysC (1:50 final enzyme to substrate) at 37oC overnight, followed (LC/MS/MS) using a separating column (0.075 mm x 30 cm) o packed with Reprosil C18 reverse-phase material (2.4 m by two rounds of trypsin digestion at 37 C overnight (prior µ particle size, Dr. Maisch). The column was connected on line to trypsin digestion the concentration of urea was lowered to to an Orbitrap Lumos tribrid instrument (Thermo Scientific). 2 M). Digestion was stopped by adding TFA to 1% final The solvents used for peptide separation were 0.1% formic concentration and immediately desalted on SepPak car- acid in water (solvent A) and 80% acetonitrile containing tridges (Waters, Baden-Dättwil, Switzerland) according to the manufacturer’s recommendation. The absorbance at 280 0.1% formic acid in water (solvent B). 2 µl of peptide digest was injected with an Easy-nLC 1200 capillary pump nm was measured and the peptide concentration was calcu- (Thermo Scientific) set to 0.3 l/min. A linear gradient from lated [15]. The peptide amount was then adjusted to equal µ 0 to 35% solvent B in solvent A in 160 min was delivered levels between the pooled controls and the respective CA- with the nano pump at a flow rate of 250 nl/min. The eluting DASIL samples. The peptides were dried in a SpeedVac. peptides were ionized at 2.5 kV. The mass spectrometer was 50 µg peptide of each of the pooled controls’ and pa- operated in a data-dependent mode. The precursor scan was tients’ frontal cortex and white matter biopsies were derivat- done in the Orbitrap set to 120,000 resolution, while the ized with a six-plex tandem mass tag (TMT) kit (Thermo fragment ions were mass analyzed at 15,000 resolution. All Fisher Scientific, Reinach, Switzerland) according to the fractions were run in technical triplicates. manufacturer’s recommendations. The pooled controls, fron- 2.4.3. Protein Identification and Bioinformatics Analysis tal cortex and WM, were labelled with TMT 126/129 respec- tively, and frontal cortex and WM samples from patient 1 For protein identification, the MS/MS spectra were and 2 individually were labelled with TMT 127/128/130/131 searched against the H. sapiens databank from SwissProt respectively, as depicted in Fig. (1). After the labelling reac- with Proteome Discoverer 2.1 (Thermo Scientific) set to tion, the derivatized peptides were pooled and desalted as Mascot and Sequest HT search engines with 10 ppm precur- described above. The peptides were dried in a SpeedVac. To sor ion tolerance. The fragment ions were set to 20 ppm tol- lower the complexity, the TMT-labelled peptide pool of con- erance. The following modifications were used during the trols and patients was injected onto a Vydac 218TPN re- search: carbamidomethyl-cystein, lysine-TMT and peptide verse-phase column (Dr. Maisch, Ammerbuch-Entringen, N-terminal-TMT were set to fixed modifications, and oxi- Germany). Peptides were eluted at 30 µL/min with a linear dized methionine, and protein N-terminal acetylation were 110 min gradient from solvent A (20 mM ammonium for- set to variable modifications. The peptide search matches mate, pH 4.5) to 50% solvent B (80% acetonitrile containing were set to 1% false discovery rate. For relative protein 20 mM ammonium formate, pH 4.5) and fractions were col- quantification, the areas of the 126/127/128/129/130/131 lected every two min, concatenated into six pools, dried and reporter ions were quantitated in proteome discoverer. The stored at -20oC. results from Proteome Discoverer were imported into Scaf- Transcriptomic and Proteomic Analyses of CADASIL Brains Current Neurovascular Research, 2019, Vol. 16, No. 5 481

Fig. (2). Principal Component Analysis (PCA) of all samples from the frontal cortex (A) and WM (B) of CADASIL and control patients. The PCA shows the global analysis of the transcriptome of CADASIL and control samples in 3D illustrating the relationships between three vari- ables. Cortical CADASIL samples were less distinguished from control samples (red and blue spheres, respectively) than CADASIL WM samples. (A higher resolution / colour version of this figure is available in the electronic copy of the article). fold (Proteome Software Inc., Portland, OR, USA) for data clearly differentiated CADASIL samples from control sam- compilation and statistical analysis. ples. 2.4.4. Gene Ontology (GO) Term Enrichment 3.2. Number of Genes and Proteins Differentially Expressed in CADASIL Brains In order to determine overrepresented GO categories from the differently expressed gene and protein sets, DAVID Comparison of gene and protein expressions in CA- Bioinformatics Resource (http://david.abcc.ncifcrf.gov) was DASIL and control brains sorted up and down-regulated used to generate clusters of co-associated genes/proteins genes in each region (Table 1). A general observation is that within KEGG (Kyoto Encyclopedia of Genes and Genomes). frontal WM contains the highest number of genes and pro- EASE Score, a modified Fisher Exact test adopted by teins differently expressed between patients and healthy con- DAVID to measure enrichment in annotation terms, was set trols, as suggested by the sample clusters well separated in to 0.05 as a cutoff point. the PCA. Table 1. Number of genes and proteins found to be signifi- 3. RESULTS cantly differently expressed in the frontal lobe between CADASIL patients and controls. 3.1. Principal Component Analysis and Hierarchical Clustering Analysis - Frontal Cortex Frontal WM The Principal Component Analysis (PCA) and unad- justed hierarchical clustering were used to estimate how the Genes Up 569 1117 mRNA microarrays group together based on the similarity of expression features, and how expression of genes is similar Genes Down 850 1997 or different in different samples. Proteins Up 92 147 The PCA shows the global analysis of the transcriptome of CADASIL and control samples in 3D (Fig. 2A, B). As Proteins Down 57 220 illustrated in the figure, cortical CADASIL samples were 20 genes with the highest expression changes in both di- less distinguishable from control samples (Fig. 2A, red and rections for each brain region are shown in Table 2. The list blue spheres, respectively) than CADASIL WM samples of differently expressed genes also contains a high number (Fig. 2B). of small nucleolar RNAs (snoRNAs) overexpressed in CA- The unsupervised hierarchical clustering (Fig. 3A, B) DASIL frontal regions listed in Table 3, snoRNAs C/D box shows the expression pattern of individual genes in each SNORD115 and SNORD116. These non-coding genes are a brain regions of CADASIL and controls. This clustering class of small RNA molecules that primarily guide chemical 482 Current Neurovascular Research, 2019, Vol. 16, No. 5 Ritz et al.

Fig. (3). Expression profile of differential gene expression between control and CADASIL. All differentially expressed genes in frontal cortex (A) and WM (B) of CADASIL compared to control brains were hierarchically clustered, showing clear expression differences between both groups. (A higher resolution / colour version of this figure is available in the electronic copy of the article).

Table 2. "Top 20 genes" up- and down-regulated in the frontal cortex and white matter in CADASIL patients compared to controls.

Accession RefSeq//Gene symbol//Full Name P-value Fold Change

Frontal Cortex - -

AK172782 // GPAM // glycerol-3-phosphate acyltransferase, mitochondrial 3.0E-03 -4.5

NM_001004689 // OR2M3 // olfactory receptor, family 2, subfamily M, member 3 2.3E-04 -3.8

NR_002141 // OR2M1P // olfactory receptor, family 2, subfamily M, member 1 1.8E-03 -3.5 pseudogene

NM_022375 // OCLM // oculomedin 1.5E-05 -3.4

NM_001004491 // OR2AK2 // olfactory receptor, family 2, subfamily AK, member 2 6.3E-04 -3.3

NM_175911 // OR2L13 // olfactory receptor, family 2, subfamily L, member 13 1.9E-04 -3.2

NR_026965 // LOC121952 // hypothetical LOC121952 5.5E-04 -3.2

AK096229 // LOC100128364 // hypothetical LOC100128364 3.9E-04 -3.1

AF314543 // OCR1 // ovarian cancer-related protein 1 8.3E-05 -3.0

NR_029506 // MIR32 // microRNA 32 5.9E-05 -2.8

NR_002169 // OR1F2P // olfactory receptor, family 1, subfamily F, member 2 1.4E-04 -2.8

ENST00000383686 // LOC151760 // hypothetical LOC151760 1.4E-03 -2.8

NM_001168235 // FREM3 // FRAS1 related extracellular matrix 3 2.7E-03 -2.7 NM_018431 // DOK5 // docking protein 5 9.7E-04 -2.7 NM_017813 // IMPAD1 // inositol monophosphatase domain containing 1 5.3E-05 -2.6 NM_017425 // SPA17 // sperm autoantigenic protein 17 3.3E-04 -2.6 AK095315 // FBXO9 // F-box protein 9 1.4E-05 -2.5

Table 2. contd… Transcriptomic and Proteomic Analyses of CADASIL Brains Current Neurovascular Research, 2019, Vol. 16, No. 5 483

Accession RefSeq//Gene symbol//Full Name P-value Fold Change

NM_020202 // NIT2 // nitrilase family, member 2 2.5E-04 -2.4

NM_145049 // UBLCP1 // ubiquitin-like domain containing CTD phosphatase 1 8.2E-05 -2.4

NM_016940 // RWDD2B // RWD domain containing 2B 3.0E-04 -2.4

NM_002055 // GFAP // glial fibrillary acidic protein 1.5E-03 2.3

NM_004281 // BAG3 // BCL2-associated athanogene 3 1.3E-03 2.4

NM_032840 // SPRYD3 // SPRY domain containing 3 6.2E-04 2.5

NM_002300 // LDHB // lactate dehydrogenase B 2.2E-04 2.5

NM_007177 // FAM107A // family with sequence similarity 107, member A 2.5E-04 2.5

NR_003051 // RMRP // RNA component of mitochondrial RNA processing endoribonuclease 4.0E-04 2.5

NM_000661 // RPL9 // ribosomal protein L9 9.2E-08 2.5

NM_002954 // RPS27A // ribosomal protein S27a 1.1E-09 2.5

NM_003548 // HIST2H4A // histone cluster 2, H4a 4.1E-04 2.6

NM_001540 // HSPB1 // heat shock 27kDa protein 1 8.6E-06 2.6

NM_080671 // KCNE4 // potassium voltage-gated channel, Isk-related family, 1.1E-03 2.6 member 4

NM_001040058 // SPP1 // secreted phosphoprotein 1 4.7E-04 2.7

NM_003542 // HIST1H4C // histone cluster 1, H4c 5.6E-04 2.9

NM_001017 // RPS13 // ribosomal protein S13 1.8E-05 2.9

NM_002798 // PSMB6 // proteasome (prosome, macropain) subunit, beta type, 6 2.8E-05 3.1

NM_005345 // HSPA1A // heat shock 70kDa protein 1A 3.3E-03 3.2

NM_019058 // DDIT4 // DNA-damage-inducible transcript 4 2.7E-03 3.6

NM_139314 // ANGPTL4 // angiopoietin-like 4 1.1E-03 4.0

NM_005252 // FOS // FBJ murine osteosarcoma viral oncogene homolog 2.4E-03 4.1

NM_013332 // HILPDA // hypoxia-inducible lipid droplet-associated protein 2.4E-03 4.5

Frontal White Matter - -

NM_001004686 // OR2L2 // olfactory receptor, family 2, subfamily L, member 2 3.1E-04 -5.1

NM_001337 // CX3CR1 // chemokine (C-X3-C motif) receptor 1 4.0E-06 -4.5

NM_057749 // CCNE2 // cyclin E2 2.6E-04 -4.3

NM_033207 // OPALIN // oligodendrocytic myelin paranodal and inner loop protein 6.3E-03 -4.2

NR_029837 // MIR219-2 // microRNA 219-2 1.8E-03 -4.1

BC040288 // LOC100130428 // IGYY565 7.4E-08 -4.0

NM_024854 // PYROXD1 // pyridine nucleotide-disulphide oxidoreductase domain 1 1.3E-05 -3.8

AK097109 // LOC100131860 // hypothetical protein LOC100131860 4.9E-06 -3.7

NM_199136 // C7orf46 // chromosome 7 open reading frame 46 2.2E-07 -3.6

NM_181789 // GLDN // gliomedin 3.5E-03 -3.6

NM_001136002 // TMEM229A // transmembrane protein 229A 1.0E-04 -3.6

Table 2. contd… 484 Current Neurovascular Research, 2019, Vol. 16, No. 5 Ritz et al.

Frontal White Matter - -

NM_175911 // OR2L13 // olfactory receptor, family 2, subfamily L, member 13 7.6E-05 -3.6

NR_030741 // MIR9-2 // microRNA 9-2 4.0E-03 -3.4

NM_005019 // PDE1A // phosphodiesterase 1A, calmodulin-dependent 8.2E-04 -3.4

NM_001001963 // OR2L8 // olfactory receptor, family 2, subfamily L, member 8 3.9E-03 -3.4

NM_017512 // ENOSF1 // enolase superfamily member 1 9.1E-04 -3.4

NM_007078 // LDB3 // LIM domain binding 3 1.8E-03 -3.3

NM_173084 // TRIM59 // tripartite motif-containing 59 3.6E-03 -3.2

AK096606 // LOC100131170 // hypothetical LOC100131170 1.3E-04 -3.2

NM_001004491 // OR2AK2 // olfactory receptor, family 2, subfamily AK, member 2 1.1E-03 -3.1

NM_017594 // DIRAS2 // DIRAS family, GTP-binding RAS-like 2 3.8E-03 5.0

AF284753 // UIMC1 // ubiquitin interaction motif containing 1 2.4E-03 5.2

NM_001540 // HSPB1 // heat shock 27kDa protein 1 3.2E-09 5.2

NR_026703 // VTRNA1-1 // vault RNA 1-1 2.0E-03 5.3

NM_013230 // CD24 // CD24 molecule 4.8E-04 5.3

NM_000043 // FAS // Fas 8.7E-05 5.3

NM_006933 // SLC5A3 // solute carrier family 5 (sodium/myo-inositol cotransporter), member 3 2.9E-06 5.4

Frontal White Matter - -

NM_015675 // GADD45B // growth arrest and DNA-damage-inducible, beta 6.9E-07 5.4

NM_004281 // BAG3 // BCL2-associated athanogene 3 3.5E-07 5.6

NM_003956 // CH25H // cholesterol 25-hydroxylase 2.6E-04 5.7

NM_004120 // GBP2 // guanylate binding protein 2, interferon-inducible 3.6E-05 5.8

NM_001124 // ADM // adrenomedullin 3.3E-05 5.9

NM_001423 // EMP1 // epithelial membrane protein 1 1.0E-04 6.0

ENST00000361453 // ND2 // MTND2, mitochondrially encoded NADH dehydrogenase 2 2.8E-03 6.3

NM_001235 // SERPINH1 // serpin peptidase inhibitor, clade H (heat shock protein

47), member 1 9.0E-08 6.7

NM_001924 // GADD45A // growth arrest and DNA-damage-inducible, alpha 1.1E-06 6.9

NM_006981 // NR4A3 // nuclear receptor subfamily 4, group A, member 3 2.9E-09 7.1

NM_005252 // FOS // FBJ murine osteosarcoma viral oncogene homolog 8.3E-05 7.4

NR_022010 // PAR4 // Prader-Willi/Angelman region gene 4 4.6E-04 9.2

NM_013332 // C7orf68 // chromosome 7 open reading frame 68 9.6E-06 13.0

Transcriptomic and Proteomic Analyses of CADASIL Brains Current Neurovascular Research, 2019, Vol. 16, No. 5 485

Table 3. Small nucleolar RNAs (snoRNAs) found overexpressed in frontal cortex and white matter of CADASIL patients.

Accession RefSeq//Gene symbol//Full Name P-value Fold Change

Frontal Cortex - - NR_003328 // SNORD116-13 // small nucleolar RNA, C/D box 116-13 2.6E-06 3.31 NR_003340 // SNORD116-26 // small nucleolar RNA, C/D box 116-26 1.1E-03 3.36 NR_003339 // SNORD116-25 // small nucleolar RNA, C/D box 116-25 2.1E-03 3.36 NR_004398 // SNORD82 // small nucleolar RNA, C/D box 82 4.8E-07 3.66 NR_002736 // SNORD60 // small nucleolar RNA, C/D box 60 6.2E-07 4.02 NR_001295 // SNORD109A // small nucleolar RNA, C/D box 109A 3.7E-06 4.52 NR_000012 // SNORA68 // small nucleolar RNA, H/ACA box 68 7.4E-07 4.73 NR_003219 // SNORD114-26 // small nucleolar RNA, C/D box 114-26 1.8E-03 5.49 NR_002960 // SNORA20 // small nucleolar RNA, H/ACA box 20 9.2E-06 6.29 NR_004380 // SNORD104 // small nucleolar RNA, C/D box 104 1.5E-06 6.85 NR_002748 // SNORD45B // small nucleolar RNA, C/D box 45B 5.2E-05 7.42 Frontal White Matter - - NR_003340 // SNORD116-26 // small nucleolar RNA, C/D box 116-26 1.9E-05 5.93 NR_001291 // SNORD115-1 // small nucleolar RNA, C/D box 115-1 9.3E-04 6.13 NR_003298 // SNORD115-6 // small nucleolar RNA, C/D box 115-6 9.3E-04 6.24 NR_003357 // SNORD115-42 // small nucleolar RNA, C/D box 115-42 8.8E-04 6.38 NR_003304 // SNORD115-12 // small nucleolar RNA, C/D box 115-12 1.7E-03 6.47 NR_003343 // SNORD115-26 // small nucleolar RNA, C/D box 115-26 1.6E-03 6.53 NR_003342 // SNORD115-25 // small nucleolar RNA, C/D box 115-25 5.6E-03 6.93 NR_003303 // SNORD115-11 // small nucleolar RNA, C/D box 115-11 7.8E-04 7.14 NR_003359 // SNORD115-44 // small nucleolar RNA, C/D box 115-44 1.9E-03 7.40 NR_003195 // SNORD114-3 // small nucleolar RNA, C/D box 114-3 2.1E-04 9.92

Table 4. KEGG terms associated with down and up-regulated genes and proteins in CADASIL frontal cortex and WM, listed by decreasing statistical significance. Common terms between cortex and WM are underlined, and similar terms found asso- ciated with genes and proteins in the same region are in bold.

- RNA Up Protein Up RNA Down Protein Down

Frontal cortex Ribosome Ribosome GPI-anchor biosynthesis Histidine metabolism

- Spliceosome RNA processing Ubiquitin-mediated proteolysis Tight junction

Leukocyte transendothelial - Lysine degradation - Oxidative phosphorylation migration

- - RNA degradation Cell adhesion molecules

Frontal WM Ribosome Oxidative metabolism GPI-anchor biosynthesis Spliceosome

Complement and coagulation - Glycolysis Cardiac muscle contraction Ubiquitin-mediated proteolysis cascade

- Focal adhesion TCA cycle Spliceosome Purine metabolism

- P53 signaling pathway Calcium signaling Cell cycle Non-homologous end joining

- ECM-receptor interaction Long-term potentiation Pyrimidine metabolism Cell adhesion molecules

SNAR interaction in vesicular Leukocyte transendothelial - - Endocytosis transport migration 486 Current Neurovascular Research, 2019, Vol. 16, No. 5 Ritz et al. modifications of other RNAs, mainly ribosomal RNAs, tients. Damage to WM tracts and widespread WM lesions transfer RNAs and small nuclear RNAs. are characteristic of CADASIL as shown by imaging, and are of clinical relevance, correlated with cognitive impair- 3.3. Functional Categorization Of Differentially Ex- ment [4, 16]. By using direct processes found to be enriched pressed Genes and Proteins by both the transcriptomic and proteomic feature sets, we identified molecular pathways that may influence the normal To investigate which biological pathways and functions cerebral function and tissue integrity, and thus participate to are associated with genes and proteins showing statistically expression differences, we performed functional annotation WM lesions observed in CADASIL. When comparing the ® pathways associated with both, gene and protein sets, some analysis using DAVID Functional Annotation tools. For informative KEGG terms were highlighted. each region, we examined sets of over and under-expressed genes and proteins individually (Table 4). 4.1. Vascular and Extracellular Matrix Pathways Identi- Of interest, many terms related to RNA metabolism were fied found in both RNA and proteins expression analysis, such as Several cellular and functional processes associated with RNA processing, RNA degradation, basal transcription, change in gene and/or protein expression matched with the RNA polymerases, and spliceosome. Several terms high- subcortical localization of CADASIL pathology and with lighted the involvement of energy metabolism: glycolysis, already known dysfunctions, such as angiopathy of small TCA cycle, oxidative metabolism and oxidative phosphory- arteries, myelin rarefaction (leukoaraiosis), and hypoxia. lation. Other terms reflected impaired cell-cell interaction in CADASIL, such as ECM-receptor interaction, tight junction, Thickening of the arterial wall and subsequent loss of or cell adhesion molecules. vascular smooth muscle cells is observed in CADASIL. We Frontal cortex and WM showed common biological found overexpression of genes coding for various collagen terms, such as ribosome (associated with upregulated genes), types (collagen I, IV, V and VI), laminins, integrins, fi- GPI-ancored biosynthesis and ubiquitin-mediated proteolysis bronectin 1, syndecan 2/4 and tenascins, in WM of CA- (associated with downregulated genes) as well as leukocyte DASIL patients grouped under " extracellular matrix (ECM) trans-endothelial migration and cell adhesion molecules (as- receptor interaction pathway". The products of these genes sociated with downregulated proteins). may induce the accumulation of ECM and basal membrane molecules mostly in WM, and may result in defective vascu- Four of the processes were found to be enriched con- lar remodeling responses and finally stenosis as suggested comitantly by the transcriptomic and proteomic feature set. previously [17, 18]. Dysregulation of TGF-β signaling path- Other processes were enriched by either transcriptomic or way has been shown to participate in the Notch3-dependent proteomic features. toxicity due to co-aggregation of LTBP-1 (latent TGF-beta Pathways revealed by genes and proteins simultaneously binding protein 1) with Notch3 extracellular domain in were the terms “ribosome” and “RNA process- CADASIL [19]. In our analysis, TGF-β, LTBP1, 3 and 4 ing”/”spliceosome” associated with upregulated genes and were all overexpressed in WM, indicating their contribution proteins in frontal cortex, “glycolysis”/”oxidative metabo- to the arteriopathy. lism” associated with upregulated genes and proteins, or Smooth muscle cells (SMC)-related genes were also “spliceosome” associated with downregulated genes and identified in our study: LMOD1 coding for leiomodin 1, and proteins in frontal WM. TMOD3, coding for tropomodulin 3, two components of the 3.4. Genes and Proteins Commonly Identified by Both - actin cytoskeleton [20] were found overexpressed in CA- omics and Associated Pathways DASIL, suggesting that changes in the contractile apparatus of vascular SMC may contribute to CADASIL pathophysi- Various genes and corresponding proteins were com- ology. monly found in each brain region, again predominantly in WM (16 and 7 up and down-regulated in the cortex, 75 and Complement/coagulation pathways were associated with 166 up and down-regulated in the WM). The molecular func- changed genes in WM. Genes coding for TSP-1, TSP-2, tions associated with these genes/proteins (Table 5) revealed thrombomodulin, von Willebrand factor subtypes, and co- enhanced ribosomal function in the cortex, increased cell- agulation factors (Factor III, XIII, and Factor II receptor) cell and cell-ECM interactions, and hypoxia in WM, in addi- were upregulated. Although thrombomodulin is an antico- tion to deficiencies in several metabolic pathways and RNA- agulant [21], TSP-1, in conjunction with von Willebrand related pathways, such as RNA transport and splicing. The factors and fibrinogen, contributes to clot formation [22]. list of genes/proteins identified in a selection of these path- PAR4 (Prader Willi/Angelmann region gene 4 coding for a ways are given in Table 6. thrombin receptor inducing platelet activation and aggrega- tion [23] was also overexpressed in frontal WM. The corre- 4. DISCUSSION sponding proteins were however not found overexpressed in This microarray and proteomic study comparing gene WM. This raises the question if dysregulation at the gene and protein expression differences in post-mortem brains level influences protein expression and finally function of from CADASIL and controls implicates several pathways in the pathway. However, some of these genes code for soluble the pathology of CADASIL. White matter was the most af- factors that might have been lost or highly diluted during the fected by dysregulated genes and proteins in CADASIL pa- freezing/conservation/extraction procedures. Transcriptomic and Proteomic Analyses of CADASIL Brains Current Neurovascular Research, 2019, Vol. 16, No. 5 487

Table 5. Common cellular terms associated with genes and proteins up- or down-regulated in frontal lobe of CADASIL patients (n: number of genes/proteins).

- n UP n DOWN

Frontal cortex 16 Ribosome 7 - Frontal WM 75 Adherens junction 166 Metabolic pathways - - Focal adhesion - Purine metabolism Sulfur metabolism Biosynthesis of antibiotics - - HIF-1 signaling pathway - RNA transport Spliceosome Endocytosis

Table 6. List of genes/proteins from selected pathways common to the genomic and proteomic analyses.

Pathway Gene/Protein Spliceosome DEAD-box helicase 23 (DDX23) Catenin beta like 1 (CTNNBL1) Mago homolog B, exon junction complex core component (MAGOHB) Serine and arginine rich splicing factor 6 (SRSF6) Small nuclear ribonucleoprotein D3 polypeptide (SNRPD3) Small nuclear ribonucleoprotein polypeptide B2 (SNRPB2) RNA transport Eukaryotic translation initiation factor 2B subunit alpha (EIF2B1)

Mago homolog B, exon junction complex core component (MAGOHB) Nucleoporin 155 (NUP155)

Nucleoporin 54 (NUP54)

Nucleoporin 88 (NUP88) Nucleoporin 93 (NUP93)

Poly(A) binding protein interacting protein 1 (PAIP1) Endocytosis STAM binding protein (STAMBP) Epsin 2 (EPN2) Itchy E3 ubiquitin protein ligase (ITCH) Sorting nexin 1 (SNX1) Sorting nexin 6 (SNX6) Ubiquitin specific peptidase 8 (USP8) Vacuolar protein sorting 4 homolog B (VPS4B) Cell-cell interaction Brain-specific angiogenesis inhibitor 1-associated protein 2 (BAIAP2) Alpha-actinin-1 (ACTN1) Sorbin and SH3 domain-containing protein 1 (SORBS1) Collagen-α1(VI) chain (COL6A1) Filamin-A (FLNA) Laminin subunit alpha-5 (LAMA5) Ribosome RPL19 RPL21 RPL24 RPL3 RPL34 RPL37A RPL5 RPS13 RPS24 RPS8 HIF-1α signaling pathway Calcium/calmodulin-dependent protein kinase type II delta (CAMK2D) Hexokinase-1 (HK1) 3-phosphoinositide-dependent protein kinase 1 (PDK1) 488 Current Neurovascular Research, 2019, Vol. 16, No. 5 Ritz et al.

Microvessel thrombosis/fibrosis may lead to hypoxia in RNA transport pathway was also affected in CADASIL the surrounding parenchyma. Ischemic events are the most brains, in particular several nuclear pore complex proteins, frequent manifestations in CADASIL. Induction of genes called nucleoporins (Nups). Recently, a link between nu- involved in glycolysis and gluconeogenesis and reduced oxi- cleoporin defects and neurodegenerative disorders such as dative phosphorylation in the cortex and WM suggest that frontotemporal dementia, ALS or PD has been established brains affected by CADASIL are confronted to low oxygen. (for Nup358). Nup155, downregulated in frontal cortex of Astrocytes and oligodendrocytes are able to upregulate their CADASIL brain, has been found to play a potential role in glycolytic capacities during periods of hypoxia, with in- cardiovascular diseases [31], but no information is available creased activities of lactate dehydrogenases (LDHs) and py- so far concerning Nups in CADASIL. ruvate kinase [24, 25]. In our study, GFAP (the astrocytic A large number of genes coding for Small nucleolar marker) and LDHB were two of the most upregulated genes RNAs (snoRNAs), and particularly C/D box snoRNA in CADASIL frontal cortex with GFAP protein also found (SNORDs) were found changed in CADASIL brains. enhanced, suggestive of astrocytic reactivity. SNORDs regulate splice site selection and posttranscrip- tional modification of mRNAs and ribosomal RNAs. Another hypoxia-related gene, HILPDA, coding for the Through their ability to base pair with ribosomal RNA pre- hypoxia inducible lipid droplet-associated protein, was also cursors, they play important roles in the synthesis of ribo- overexpressed in CADASIL brains. This gene is a specific somes, and therefore affect the physiological condition of target of hypoxia-inducible factor 1 alpha (HIF-1α) [26]. cells and tissue, leading to various diseases. In contrast with Overexpression of HILPDA stimulates storage of fatty acids most snoRNAs, SNORD115 and SNORD116 instead appear [27] but also inflammatory responses through the induction to regulate the alternative splicing of gene transcripts in the of vascular endothelial growth factor A (VEGFA), interleu- brain [32]. SNORD116 changes expression levels of over kin-6 (IL-6) and macrophage migration inhibitory factor 200 genes and SNORD115 has been shown to influence the (MIF). VEGFA and MIF were overexpressed in frontal cor- ability of SNORD116 to change target gene expression [33]. tex, suggesting chronic and diffuse hypoxia and inflamma- Therefore, overexpression of these SNORDs leads to various tion in CADASIL brains. Moreover, transcripts for cellular psychic and behavioral aberrations as recently described stress induced by ischemia (HSPA1A and HSPA1B), DNA- [34]. damage-inducible transcript 4 (DDIT4) and angiopoietin-like Concerning genes belonging to ribosome biogenesis, 40 4 (ANGPTL4), both responding to hypoxia [28] and cellular genes coding for large ribosomal proteins (RPL) and small energy stress [29], were overrepresented in frontal cortex. ribosomal proteins (RPS) were found overexpressed in CA- Recently, MRI techniques based on the measurement of wa- DASIL frontal cortex in the present work, but only 15 RPL ter diffusion demonstrating that the extend of edema due to and RPS proteins were upregulated, 10 of them were com- hypoxia, glial alterations and neurodegeneration in CA- mon in both genomic and proteomic analyses. The disruption DASIL brains are associated with cognitive decline [4, 16], of ribosome biogenesis, with either increased or decreased support the implication of these pathways in the clinical expression of different ribosomal components, can promote manifestations of the disease. cell cycle arrest, senescence or apoptosis. Moreover, inter- ference with ribosomal biogenesis is often associated with 4.2. Newly Identified Pathways Dysregulated in CA- cancer and age-related degenerative diseases. Numerous DASIL studies suggest that ribosomal biogenesis is impaired in neu- rodegenerative diseases and may contribute to neuronal dys- Interestingly, spliceosome, RNA transport and ribosome function in two ways, loss of function and stress response were terms associated with genes and proteins significantly [35]. The role of an increased ribosome biogenesis in CA- downregulated in CADASIL. Defects in RNA metabolism DASIL is intriguing and needs further investigation. and splicing defects in particular, have been implicated in many diseases, such as cancer, neuro-muscular diseases Although only 2 patients with diagnosed CADASIL were (myotonic dystrophy, spinal muscular atrophy) and neurode- used in this study due to the paucity of freshly dissected generative diseases, e.g. amyotrophic lateral sclerosis (ALS), brains, we were able to confirm previously described patho- Frontotemporal Dementia (FTD), Parkinson’s disease (PD) physiologic features such vascular dysfunction, endothelial and Alzheimer’s disease (AD) [30]. As alternative splicing activation and rarefaction, fibrosis and thickening of ECM. affects numerous genes, it is not surprising that changes in Further comparison of the rather rare hereditary form of alternative splicing are frequently associated with human SVD with more easily available tissue from sporadic SVD diseases. Neurons are particularly vulnerable to disruption of may further elucidate the underlying mechanisms linking the RNA-binding protein dosage and dynamics; therefore, it is vasculopathy with consecutive brain tissue damage in SVD. possible that disruptions in spliceosomal proteins and vari- ous transcription factors have a central role also in CA- Superimposed pro-coagulant mechanisms, hypoxia- DASIL and SVD pathogenesis. It is often not clear whether a induced inflammation and cellular stress, as well as a defec- change in alternative splicing causes a disease or is an indi- tive ubiquitin proteasome system were highlighted, that may cator for an underlying defect. Mis-splicing has been previ- imply clinically relevant tissue damage and/or precipitate the ously implicated in both inherited and sporadic neurodegen- disease progression. Animal models may further help to un- erative diseases, such as ALS or FTD, suggesting that this derstand the underlying pathomechanisms and have in fact process is a promising target for future therapeutic develop- recently demonstrated various potential therapeutic ap- ments. proaches [36-38]. In human subjects, additionally to MR- Transcriptomic and Proteomic Analyses of CADASIL Brains Current Neurovascular Research, 2019, Vol. 16, No. 5 489 imaging approaches [4, 16], analyses of potentially interest- CONSENT FOR PUBLICATION ing blood-based markers, such as recently reported for neu- The informed consent forms were obtained from all of rofilament-light-chain in subjects with both hereditary as the patients legal proxies. well as sporadic SVD, may also further add to the under- standing of disease activity and progression [39]. AVAILABILITY OF DATA AND MATERIALS CONCLUSION The data supporting the findings of the article is available in the ArrayExpress Archive of Functional Genomics Data, In contrast to other neurological diseases in which mu- at https://www.ebi.ac.uk/arrayexpress/experiments/, refer- tated proteins are causing splicing dysfunctions, we showed ence number E-MTAB-3395 and the mass spectrometry pro- that in CADASIL, dysregulated expression of a set of genes teomics data have been deposited to the ProteomeXchange and proteins seems to affect correct RNA splicing or RNA Consortium (http://proteomecentral.proteomexchange.org) transcription. Consequently, multiple RNA/protein targets might be modified and could impact the pathology of CA- via the PRIDE partner repository with the dataset identifier PXD016066. DASIL, serving as a model for SVD. It is currently unknown whether and how these mis-splicing events are involved in the development of cerebral tissue damage in CADASIL. FUNDING This question could be answered by more sophisticated ap- This work was supported by grants from the Swiss Heart proaches, such as next-generation sequencing that generates Foundation to PL/NP, and Neurology Research Pool, Uni- aberrant alternative transcript profiles in a more sensitive versity Hospital Basel to NP. manner. CONFLICT OF INTEREST LIST OF ABBREVIATIONS The authors declare no conflict of interest, financial or AD = Alzheimer's Disease otherwise. ANOVA = Analysis Of Variance ACKNOWLEDGEMENTS CADASIL = Cerebral Autosomal Dominant Arte- We thank Dr. Thomas Arzberger, the Neurobiobank, In- riopathy with Subcortical Infarcts and stitute for Neuropathology, München, Germany, and Prof. Leukoencephalopathy Markus Tolnay, Institute of Pathology, University Hospital Basel, Switzerland for supplying the CADASIL and control DAVID = Database for Annotation, Visualization, brain specimens, respectively. We are thankful to Philippe and Integrated Discovery Demougin from the Life Sciences Training Facility, Pharma- zentrum, University of Basel, Switzerland for his technical FDR = False Discovery Rate expertise in microarray experiments and to Suzette Moes, Proteomics Core Facility, Biozentrum, University of Basel, FTD = Fronto-Temporal Dementia Switzerland, for her technical help in proteomic experiments. KEGG = Kyoto Encyclopedia of Genes and Ge- We warmly thank Dr. Caspar Grond-Ginsbach for very en- nomes thusiastic and helpful discussions. SVD = Small Vessel Disease REFERENCES [1] Pantoni L. Cerebral small vessel disease: From pathogenesis and TMT = Tandem Mass Tag clinical characteristics to therapeutic challenges. Lancet Neurol 2010; 9(7): 689-701. WM = White Matter [http://dx.doi.org/10.1016/S1474-4422(10)70104-6] [PMID: 20610345] WMLs = White Matter Lesions [2] Chabriat H, Joutel A, Dichgans M, Tournier-Lasserve E, Bousser MG. Cadasil. Lancet Neurol 2009; 8(7): 643-53. [http://dx.doi.org/10.1016/S1474-4422(09)70127-9] [PMID: ETHICS APPROVAL AND CONSENT TO PARTICI- 19539236] PATE [3] Joutel A, Corpechot C, Ducros A, et al. Notch3 mutations in CA- DASIL, a hereditary adult-onset condition causing stroke and de- The study protocol was approved by the local ethics mentia. Nature 1996; 383(6602): 707-10. committee Ethikkomission Beider Basel, EKBB, Basel, [http://dx.doi.org/10.1038/383707a0] [PMID: 8878478] [4] Holtmannspötter M, Peters N, Opherk C, et al. Diffusion magnetic Switzerland. resonance histograms as a surrogate marker and predictor of dis- ease progression in CADASIL: A two-year follow-up study. Stroke HUMAN AND ANIMAL RIGHTS 2005; 36(12): 2559-65. [http://dx.doi.org/10.1161/01.STR.0000189696.70989.a4] [PMID: No animals were used for studies that are the basis of this 16269644] [5] Peters N, Holtmannspötter M, Opherk C, et al. Brain volume research. All the human procedures were followed in accor- changes in CADASIL: A serial MRI study in pure subcortical dance with the ethical standards of the committee responsi- ischemic vascular disease. Neurology 2006; 66(10): 1517-22. ble for human experimentation (institutional and national), [http://dx.doi.org/10.1212/01.wnl.0000216271.96364.50] [PMID: 16717211] and with the Helsinki Declaration of 1975, as revised in 2013 [6] Erkinjuntti T, Gauthier S. The concept of vascular cognitive im- (http://ethics.iit.edu/ecodes/node/3931). pairment. Front Neurol Neurosci 2009; 24: 79-85. [http://dx.doi.org/10.1159/000197886] [PMID: 19182465] 490 Current Neurovascular Research, 2019, Vol. 16, No. 5 Ritz et al.

[7] Tikka S, Mykkänen K, Ruchoux MM, et al. Congruence between [23] Kahn ML, Zheng YW, Huang W, et al. A dual thrombin receptor NOTCH3 mutations and GOM in 131 CADASIL patients. Brain system for platelet activation. Nature 1998; 394(6694): 690-4. 2009; 132(Pt 4): 933-9. [http://dx.doi.org/10.1038/29325] [PMID: 9716134] [http://dx.doi.org/10.1093/brain/awn364] [PMID: 19174371] [24] Marrif H, Juurlink BH. Astrocytes respond to hypoxia by increas- [8] Ruchoux MM, Guerouaou D, Vandenhaute B, Pruvo JP, ing glycolytic capacity. J Neurosci Res 1999; 57(2): 255-60. Vermersch P, Leys D. Systemic vascular smooth muscle cell im- [http://dx.doi.org/10.1002/(SICI)1097- pairment in cerebral autosomal dominant arteriopathy with subcor- 4547(19990715)57:2<255::AID-JNR11>3.0.CO;2-6] [PMID: tical infarcts and leukoencephalopathy. Acta Neuropathol 1995; 10398303] 89(6): 500-12. [25] Fünfschilling U, Supplie LM, Mahad D, et al. Glycolytic oligoden- [http://dx.doi.org/10.1007/BF00571504] [PMID: 7676806] drocytes maintain myelin and long-term axonal integrity. Nature [9] Zhang X, Meng H, Blaivas M, et al. Von Willebrand Factor per- 2012; 485(7399): 517-21. meates small vessels in CADASIL and inhibits smooth muscle [http://dx.doi.org/10.1038/nature11007] [PMID: 22622581] gene expression. Transl Stroke Res 2012; 3(1): 138-45. [26] Denko N, Schindler C, Koong A, Laderoute K, Green C, Giaccia [http://dx.doi.org/10.1007/s12975-011-0112-2] [PMID: 22639698] A. Epigenetic regulation of gene expression in cervical cancer cells [10] Østergaard L, Engedal TS, Moreton F, et al. Cerebral small vessel by the tumor microenvironment. Clin Cancer Res 2000; 6(2): 480- disease: Capillary pathways to stroke and cognitive decline. J 7. Cereb Blood Flow Metab 2016; 36(2): 302-25. [27] Gimm T, Wiese M, Teschemacher B, et al. Hypoxia-inducible [http://dx.doi.org/10.1177/0271678X15606723] [PMID: 26661176] protein 2 is a novel lipid droplet protein and a specific target gene [11] Wiseman S, Marlborough F, Doubal F, Webb DJ, Wardlaw J. of hypoxia-inducible factor-1. FASEB J 2010; 24(11): 4443-58. Blood markers of coagulation, fibrinolysis, endothelial dysfunction [http://dx.doi.org/10.1096/fj.10-159806] [PMID: 20624928] and inflammation in lacunar stroke versus non-lacunar stroke and [28] Schwarzer R, Tondera D, Arnold W, Giese K, Klippel A, Kauf- non-stroke: Systematic review and meta-analysis. Cerebrovasc Dis mann J. REDD1 integrates hypoxia-mediated survival signaling 2014; 37(1): 64-75. downstream of phosphatidylinositol 3-kinase. Oncogene 2005; [http://dx.doi.org/10.1159/000356789] [PMID: 24401164] 24(7): 1138-49. [12] Ritz MF, Grond-Ginsbach C, Kloss M, et al. Identification of in- [http://dx.doi.org/10.1038/sj.onc.1208236] [PMID: 15592522] flammatory, metabolic, and cell survival pathways contributing to [29] Sofer A, Lei K, Johannessen CM, Ellisen LW. Regulation of cerebral small vessel disease by postmortem gene expression mi- mTOR and cell growth in response to energy stress by REDD1. croarray. Curr Neurovasc Res 2016; 13(1): 58-67. Mol Cell Biol 2005; 25(14): 5834-45. [http://dx.doi.org/10.2174/1567202612666151027151025] [PMID: [http://dx.doi.org/10.1128/MCB.25.14.5834-5845.2005] [PMID: 26503025] 15988001] [13] Vogel C, Marcotte EM. Insights into the regulation of protein [30] Faustino NA, Cooper TA. Pre-mRNA splicing and human disease. abundance from proteomic and transcriptomic analyses. Nat Rev Genes Dev 2003; 17(4): 419-37. Genet 2012; 13(4): 227-32. [http://dx.doi.org/10.1101/gad.1048803] [PMID: 12600935] [http://dx.doi.org/10.1038/nrg3185] [PMID: 22411467] [31] Zhang X, Chen S, Yoo S, et al. Mutation in nuclear pore compo- [14] Ritz MF, Grond-Ginsbach C, Fluri F, et al. Cerebral small vessel nent NUP155 leads to atrial fibrillation and early sudden cardiac disease is associated with dysregulation in the ubiquitin proteasome death. Cell 2008; 135(6): 1017-27. system and other major cellular pathways in specific brain regions. [http://dx.doi.org/10.1016/j.cell.2008.10.022] [PMID: 19070573] Neurodegener Dis 2017; 17(6): 261-75. [32] Kishore S, Khanna A, Zhang Z, et al. The snoRNA MBII-52 [http://dx.doi.org/10.1159/000478529] [PMID: 28810250] (SNORD 115) is processed into smaller RNAs and regulates alter- [15] Wiśniewski JR, Zougman A, Nagaraj N, Mann M. Universal sam- native splicing. Hum Mol Genet 2010; 19(7): 1153-64. ple preparation method for proteome analysis. Nat Methods 2009; [http://dx.doi.org/10.1093/hmg/ddp585] [PMID: 20053671] 6(5): 359-62. [33] Falaleeva M, Surface J, Shen M, de la Grange P, Stamm S. [http://dx.doi.org/10.1038/nmeth.1322] [PMID: 19377485] SNORD116 and SNORD115 change expression of multiple genes [16] Ban S, Wang H, Wang M, et al. Diffuse tract damage in CADASIL and modify each other’s activity. Gene 2015; 572(2): 266-73. is correlated with global cognitive impairment. Eur Neurol 2019; [http://dx.doi.org/10.1016/j.gene.2015.07.023] [PMID: 26220404] [Epub ahead of print]. [34] Gibbons A, Udawela M, Dean B. Non-coding RNA as novel play- [http://dx.doi.org/10.1159/000501612] [PMID: 31484188] ers in the pathophysiology of schizophrenia. Noncoding RNA [17] Chiang HY, Korshunov VA, Serour A, Shi F, Sottile J. Fibronectin 2018; 4(2): E11. is an important regulator of flow-induced vascular remodeling. Ar- [http://dx.doi.org/10.3390/ncrna4020011] [PMID: 29657307] terioscler Thromb Vasc Biol 2009; 29(7): 1074-9. [35] Turi Z, Lacey M, Mistrik M, Moudry P. Impaired ribosome bio- [http://dx.doi.org/10.1161/ATVBAHA.108.181081] [PMID: genesis: Mechanisms and relevance to cancer and aging. Aging 19407246] (Albany NY) 2019; 11(8): 2512-40. [18] Joutel A, Haddad I, Ratelade J, Nelson MT. Perturbations of the [http://dx.doi.org/10.18632/aging.101922] [PMID: 31026227] cerebrovascular matrisome: A convergent mechanism in small ves- [36] Ping S, Qiu X, Gonzalez-Toledo ME, Liu X, Zhao LR. Stem cell sel disease of the brain? J Cereb Blood Flow Metab (Nihongoban) factor in combination with granulocyte colony-stimulating factor 2015; [Epub ahead of print]. reduces cerebral capillary thrombosis in a mouse model of CA- [19] Kast J, Hanecker P, Beaufort N, et al. Sequestration of latent TGF- DASIL. Cell Transplant 2018; 27(4): 637-47. β binding protein 1 into CADASIL-related Notch3-ECD deposits. [http://dx.doi.org/10.1177/0963689718766460] [PMID: 29871518] Acta Neuropathol Commun 2014; 2(1): 96. [37] Ghezali L, Capone C, Baron-Menguy C, et al. Notch3ECD immu- [http://dx.doi.org/10.1186/s40478-014-0096-8] [PMID: 25190493] notherapy improves cerebrovascular responses in CADASIL mice. [20] Conley CA. Leiomodin and tropomodulin in smooth muscle. Am J Ann Neurol 2018; 84(2): 246-59. Physiol Cell Physiol 2001; 280(6): C1645-56. [http://dx.doi.org/10.1002/ana.25284] [PMID: 30014602] [http://dx.doi.org/10.1152/ajpcell.2001.280.6.C1645] [PMID: [38] Ping S, Qiu X, Kyle M, Hughes K, Longo J, Zhao LR. Stem cell 11350761] factor and granulocyte colony-stimulating factor promote brain re- [21] Van de Wouwer M, Collen D, Conway EM. Thrombomodulin- pair and improve cognitive function through VEGF-A in a mouse protein C-EPCR system: Integrated to regulate coagulation and in- model of CADASIL. Neurobiol Dis 2019; 132:104561. flammation. Arterioscler Thromb Vasc Biol 2004; 24(8): 1374-83. [http://dx.doi.org/10.1016/j.nbd.2019.104561] [PMID: 31376480] [http://dx.doi.org/10.1161/01.ATV.0000134298.25489.92] [PMID: [39] Duering M, Konieczny MJ, Tiedt S, et al. Serum neurofilament 15178554] light chain levels are related to small vessel disease burden. J [22] Bonnefoy A, Moura R, Hoylaerts MF. The evolving role of throm- Stroke 2018; 20(2): 228-38. bospondin-1 in hemostasis and vascular biology. Cell Mol Life Sci [http://dx.doi.org/10.5853/jos.2017.02565] [PMID: 29886723] 2008; 65(5): 713-27. [http://dx.doi.org/10.1007/s00018-007-7487-y] [PMID: 18193161]