The details of these procedures have been described in our previous studies [Yamamoto et al (4) and Yoshida et al (5)].

Three strains of rat, SHR/Izm, SHRSP/Izm and WKY/Izm, were provided by the Disease Model Cooperative Research Association, Kyoto, Japan. Three-week-old rats were purchased and maintained for 2 days in our animal facility and were used as 3-week-old rats. Five-week-old rats were purchased and, after being maintained for 1 week in our animal facility, were used as 6-week-old rats.

Briefly, total RNA was purified using a miRNeasy kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.

Expression profiling was generated using a 4x44K whole rat genome oligo microarray version 3.0 G2519F (Agilent Technologies Inc., Santa Clara, CA, USA). Eighteen microarray analyses as 1 color experiment were performed using the WKY rats, SHRs, and SHRSP at 3 and 6 weeks old as biological triplicates. Each gene expression profile was compared between the SHRs and WKY rats and also between the SHRSP and SHRs.

Total RNA (200 ng) was reverse transcribed into double-stranded cDNA by the AffinityScript Multiple Temperature Reverse Transcriptase (Agilent Technologies Inc.) and amplified. The resulting cDNA was used for in vitro transcription by T7-polymerase and labeled with cyanine-3-labeled cytosine triphosphate (Perkin-Elmer, Wellesley, MA, USA) using a Low Input Quick Amp Labeling kit (Agilent Technologies Inc.). After being labeled and fragmented, each cRNA sample was hybridized on Agilent 4×44K whole rat genome arrays (Agilent Design #028282). After washing, the slides were scanned using an Agilent Microarray Scanner (G2505C; Agilent Technologies Inc.). Feature Extraction software (version 10.5.1.1) was used to convert the images into gene expression data.

Raw data were imported into Subio Platform version 1.12 (Subio Inc., Kagoshima, Japan) and raw intensity data were normalized to the 75th percentile intensity of probes above background levels (gIsWellAbove=1). SHR- and SHRSP-specific genes were defined as those with signal ratios with more than a 4.0-fold increase or less than a 4.0-fold decrease in expression. Raw data have been accepted in the Gene Expression Omnibus (GEO, accession no. GSE41453).

To validate the results obtained by the microarray analysis, 11 enriched genes were randomly selected from 39 enriched unique genes, and RT-qPCR was performed under 15 different experimental conditions. Statistical comparisons between the microarray and RT-qPCR data were performed using Spearman's rank correlation test.

Annotation enrichment analysis was performed using the DAVID (http://david.abcc.ncifcrf.gov/) web tool (version 6.7, 2010) (7,8) with GenBank IDs bearing Entrez Gene ID (Table I, unique genes identified). This web-based resource provides a set of functional annotation tools for the statistical enrichment of genes categorized into GO terms. We used the GO FAT category, which filtered out very broad GO terms to identify statistically enriched functional groups. The annotated gene and protein symbols were written in italic and regular fonts, respectively.

IPA software (IPA^®; Qiagen Redwood City, CA, USA, http://www.qiagen.com/ingenuity) was applied to microarray analyses that were conducted to provide functionality for the interpretation of the gene expression data. IPA was performed with Agilent probe IDs bearing Entrez Gene ID as an input for data (Table I, mapped probes). This web tool was used to overlay functions and diseases, and to categorize SHR- and SHRSP-specific genes according to disease-related or functional annotations. It identified the biological functions and/or diseases in the Ingenuity Knowledge Base (Spring 2014 version) that were the most significant to each of the category sets. The probability of the assignment was expressed by a P-value calculated using the right-tailed Fisher's exact test. The path explorer tool was also used to identify relevant interactions among SHR- and SHRSP-specific genes and to identify the shortest literature-supported paths between genes.

IPA was performed using the IPA database (Spring 2014 release of IPA) and the probe IDs of each gene. The data obtained with DAVID were based on the database (version 6.7, 2010) and GenBank IDs of each gene. Since the renewal dates of these two databases were different, small differences were observed between these two annotation results.

We compared gene expression profiles between the SHRs and WKY rats and also between the SHRSP and SHRs, at 3 and 6 weeks of age, and isolated SHR- and SHRSP-specific genes using genome-wide microarray technology. Since we expected the expression of candidate genes to be regulated before elevations in blood pressure (BP), i.e., in the pre-hypertensive period, we examined the expression profiles of each probe using RNA samples prepared from the kidneys, and isolated a total of 353 SHR- and SHRSP-specific probes showing more than a 4-fold increase or less than a 4-fold decrease in expression (Table I).

We classified the 353 probes into 4 groups, from G-1 to G-4 (Table I). G-1 probes were isolated from the rats at 3 weeks of age and contained 87 SHR-specific probes. Their expression profiles were displayed as a heatmap using the Subio Platform (Fig. 1). These 87 probes corresponded to 69 unique genes, 44 of which showed more than a 4-fold increase and 25 showed less than a 4-fold decrease in expression (Table I). G-2 contained 96 SHR-specific genes isolated from the rats at 6 weeks of age, G-3 contained 35 SHRSP-specific genes isolated from the rats at 3 weeks of age, and G-4 contained 32 SHRSP-specific genes isolated from the rats at 6 weeks of age (Table I).

Using the DAVID web tools, the candidate genes causing hypertension, stroke and ADHD were selected from each group as significantly enriched genes. We isolated a total of 61 enriched genes consisting of 39 unique genes (Table II).

In order to verify the results obtained from microarray analysis, we randomly selected 11 out of the 39 genes (Table III-A), performed 15 real-time RT-qPCR experiments (Table III-B), and compared the results obtained with those of the micro-array experiments by applying Spearman's rank correlation test. The results supported a correlation between the reults of these two different experiments as rs=0.814 with a two-tailed P-value <0.001.

A total of 69 G-1 genes included 26 enriched genes categorized with 3 GO terms: i) GO:0005576 (extracellular region); ii) GO:0008289 (lipid binding); and iii) GO:0055114 (oxidation reduction) (Table II, G-1). A total of 96 G-2 genes included 24 enriched genes categorized with 4 GO terms: i) GO:0003013 (circulatory system process); ii) GO:0055114 (oxidation reduction); iii) GO:0010817 (regulation of hormone levels); and iv) GO:0006775 (fat-soluble vitamin metabolic process) (Table II, G-2).

A total of 35 G-3 genes included 6 enriched genes categorized with 2 GO terms: i) GO:0003013 (circulatory system process); and ii) GO:0051918 (negative regulation of fibrinolysis) (Table II, G-3). A total of 32 G-4 genes included 5 enriched genes categorized with 2 GO terms: i) GO:0051918 (negative regulation of fibrinolysis) and ii) GO:0030097 (hemopoiesis) (Table II, G-4).

Although 26 enriched G-1 genes and 5 G-4 genes did not include genes categorized with circulatory system process, 24 enriched G-2 genes included 7 genes, and 6 enriched G-3 genes included 4 genes categorized with circulatory system process, respectively (Table II).

As described above, the SHR- and SHRSP-specific genes were classified into 4 groups (Table I), and then categorized based on disease-related or functional annotations using IPA. The results obtained are summarized in Table IV, and identified among other significantly enriched functional categories, such as ‘endocrine system disorders', ‘cardiovascular disease', ‘cardiovascular system development and function' and ‘hereditary disorder' (Table IV).

G-1 genes included 2 genes, cystic fibrosis transmembrane conductance regulator (Cftr) and serine peptidase inhibitor, Kazal type 3 (Spink3) categorized as ‘endocrine system disorders (idiopathic pancreatitis)' (Table IV, G-1) (20,21). G-2 genes included 8 genes: angiotensin I converting enzyme (Ace), deiodinase, iodothyronine, type II (Dio2), acyl-Coenzyme A oxidase 2 (Acox2), fin bud initiation factor homolog (Fibin), flavin-containing monooxygenase 2 (Fmo2), indolethylamine N-methyltransferase (Inmt), myosin XVI (Myo16) and zinc finger and BTB domain containing 16 (Zbtb16) categorized as ‘cardiovascular disease (hypertension)' (Table IV, G-2) (22–24). G-3 genes included 6 genes: Agt, Apoh, epoxide hydrolase 2 (Ephx2), histidine-rich glycoprotein (Hrg), ryanodine receptor 1 (Ryr1) and vascular endothelial growth factor B (Vegfb) categorized as ‘cardiovascular system development and function (development of cardiovascular system)' (Table IV, G-3) (25–30). G-4 genes included 6 genes: Btg3 associated nuclear protein (Banp), Ephx2, retinoid X receptor gamma (Rxrg), Ryr1, RNA-binding protein fox-1 homolog 1 (Rbfox1) and Zbtb16 categorized as ‘hereditary disorder (Huntington's disease)' (Table IV, G-4) (31–33).

Since our working hypothesis is that G-1 genes include genes that regulate the expression of G-2 genes, we examined the interactions between 69 G-1 and 96 G-2 genes using IPA, and found 5 direct and 3 indirect interactions (Table I and Fig. 2): Rxrg and group-specific component (Gc) interacted with cytochrome P450 subfamily 24 (Cyp24a1) (34,35); Bcl6 interacted with the following 3 genes: Zbtb16 (9,10), protocadherin 9 (Pcdh9) (11) and Spi-B transcription factor (Spib) (12); Cftr interacted with Ephx2 (36); tumor protein p73 (Tp73) interacted with tetraspanin 1 (Tspan1) (37); and Sox2 interacted with Tp73 (14).

Other than the 8 interactions between the G-1 and G-2 genes, we identified 3 interactions among the G-1 genes: Tp73 interacted with Tspan1; Sox2 interacted with Tp73; Sox2 interacted with connective tissue growth factor (Ctgf) (38); and among the G-2 genes: Gc interacted with Cyp24a1; Cftr interacted with Ephx2; Tp73 interacted with Tspan1, respectively. We also found 12 and 16 self-control genes among the SHR-specific G-1 and G-2 genes, respectively (Fig. 2).

However, we did not detect any interactions between the G-1 genes and the majority of BP-controlling G-2 genes, such as Ace, Agtrap, Cftr, glucagon-like peptide 1 receptor (Glp1r), kininogen 2 (Kng2), myosin light chain, phosphorylatable, fast skeletal muscle (Mylpf), Acox2, Dio2, Fibin, Fmo2, Inmt and Myo16 (Table II, G-2; GO:0003013, circulatory system process and Table IV, G-2; cardiovascular disease: hypertension).

Since the enriched G-3 genes were expected to regulate the expression of the G-4 genes, we examined the interactions between 35 G-3 and 32 G-4 genes using IPA, and found that Agt interacted not only with Agtrap (16,17) expressed in the rats at 3 and 6 weeks of age, but also indirectly interacted with Zbtb16 (39) expressed in the rats at 6 weeks of age (Fig. 3). In addiiton, a total of 5 self-control genes, such as Agtrap, Ephx2, Apoh, Ryr1 and zinc finger protein 597 (Zfp597) were found to be expressed in the SHRSP at 3 and 6 weeks of age (Fig. 3).

The description and reference of each gene are summarized in Table V.

The first aim of the present study was to identify candidate genes that triggered hypertension in SHRs, the second was to identify genes related to stroke-prone symptoms, and the third was to identify genes related to ADHD. We compared gene expression profiles between SHRs and WKY rats and also between SHRSP and SHRs at 3 and 6 weeks of age, and isolated a total of 232 unique genes showing more than a 4-fold increase or less than a 4-fold decrease in expression as SHR- or SHRSP-specific genes (Table I). We expected a number of these genes to be related to hypertension, susceptibility to stroke and ADHD.

The IPA path explorer tool suggested the presence of 5 direct interactions between 69 G-1 and 96 G-2 genes (Fig. 2): i) Rxrg interacted with Cyp24a1 (34); Bcl6 interacted with the following 3 genes: ii) Zbtb16 (9,10), iii) Pcdh9 (11) and iv) Spib (12); and v) Sox2 interacted with Tp73 (14).

Rxrg and Cyp24a1: Rxrg encodes a member of the retinoid X receptor (Rxr) family of nuclear receptors, which are involved in mediating the antiproliferative effects of retinoic acid. This receptor forms dimers with retinoic acid, thyroid hormone and vitamin D receptors, increasing both DNA binding and transcriptional function on their respective response elements. Cyp24a1 encodes a member of the cytochrome P450 superfamily of enzymes. Cytochrome P450 proteins are monooxygenases that catalyze a number of reactions involved in drug metabolism and the synthesis of cholesterol, steroids and other lipids. By regulating vitamin D3 levels, this enzyme plays a role in calcium homeostasis and the vitamin D endocrine system.

Other than these 5 direct interactions between G-1 and G-2 genes, we identified one direct interaction between G-1 genes; Sox2 interacted with Ctgf (38), which encodes a mitogen that is secreted by vascular endothelial cells. This encoded protein plays a role in chondrocyte proliferation and differentiation, cell adhesion in many cell types, and is related to platelet-derived growth factor. Ctgf has been linked to the development and progression of diabetic vascular and renal disease. Low-density lipoproteins (LDL) have previously been shown to induce the expression of Ctgf in aortic endothelial cells (40) (Fig. 2).

Even based on the interactions, described above, we were unable to pinpoint the candidate gene(s) causing hypertension. Although these predicted interactions included the hypertension-related G-2 genes, Ephx2 and Zbtb16, they did not include other hypertension-related genes, such as Ace, Agtrap, Cftr, Glp1r, Kng2, Mylpf, Acox2, Dio2, Fibin, Fmo2, Inmt and Myo16 (Table II, G-2; GO:0003013, circulatory system process and Table IV, G-2; cardiovascular disease: hypertension). In order to identify further interactions between SHR-specific G-1 and G-2 genes, we applied the IPA path explorer tool, suggested the presence of one gene that assisted in these interactions, and found such a condition when we proposed the Jun proto-oncogene (Jun) (41), which interacts directly with specific target DNA sequences to regulate gene expression. The presence of Jun has been shown to facilitate interactions between 3 G-1 genes: Bcl6 (13), Sox2 (15) and ankyrin repeat domain 35 (Ankrd35) (42), and 8 G-2 genes: Spib (43), Ephx2 (44), Tp73 (45), Dio2 (46), cytochrome P450, family 8, subfamily b, polypeptide 1 (Cyp8b1) (47), Mylpf (48), glial cell derived neurotrophic factor (Gdnf) (49) and neurofilament, heavy polypeptide (Nefh) (50) (Fig. 2).

These findings suggested that Bcl6 and Sox2 were the candidate genes responsible for causing hypertension in SHRs.

Since the candidate gene(s) found to cause stroke in SHRSP were expected to be included in the G-3 genes, we focused on the interaction between G-3 and G-4 genes (Fig. 3). Our results revealed that G-3 genes included 3 typical blood pressure-related genes, Ephx2, Kng2 and Agtrap (Table II, G-3; GO:0003013, circulatory system process). These 3 genes isolated from the SHRSP at 3 weeks of age were not isolated from the SHRs at 3 weeks of age, but were isolated from the SHRs at 6 weeks of age (Table II). These results indicated that the expression of genes related to BP control proceeds more rapidly in SHRSP than in SHRs during their development.

The IPA path explorer tool revealed one interaction among G-3 genes and 8 self-controlling genes (Fig. 3). One of the G-3 genes, Agt, interacted with another G-3 gene, Agtrap, and Agt also interacted with 2 G-4 genes, Agtrap and Zbtb16 (Fig. 3). These results suggest that Agt, Agtrap and Zbtb16 play pivotal roles in causing stroke-prone symptoms. Moreover, G-4 genes including 9 self-controlling genes (Fig. 3), and self-control genes, such as Agtrap, Ephx2, Apoh, Ryr1 and Zfp597, were expressed in the 3- and 6-week-old SHRSP.

In order to detect further interactions between SHRSP-specific G-3 and G-4 genes, we applied the IPA path explorer tool, suggested the presence of one gene that assisted these interactions, and found such a condition when we proposed tumor protein p53 (Tp53) (51), which interacts directly with specific target DNA sequences to regulate gene expression. The presence of Tp53 facilitated interactions between 2 G-3 genes, Agt (18) and Apoh (19), and 3 G-4 genes, Apoh, Vegfb (52) and Banp (51) (Fig. 3).

Four enriched G-3 genes were categorized as GO:0003013 (circulatory system process). These genes were expected to participate in blood pressure control and the pathogenesis of stroke. Moreover, 2 enriched G-3 genes, Apoh and Hrg, were categorized as GO:0051918 (negative regulation of fibrinolysis) (Table II, G-3). Since Apoh has been implicated in various physiological pathways, including lipoprotein metabolism, coagulation and the production of antiphospholipid autoantibodies, we hypothesized that it may participate in the genesis of atherosclerosis and stroke. Hrg possesses antimicrobial activity, and the incorporation of Hrg into fibrin clots facilitates bacterial entrapment and killing and promotes inflammation. Since vascular inflammation is known to trigger atherosclerosis, Hrg influences atherosclerosis and susceptibility to strokes.

Two out of the 5 enriched G-4 genes, Apoh and Hrg were categorized as GO:0051918 (negative regulation of fibrinolysis), while the remaining 3 genes, chemokine (C-C motif) receptor 1 (Ccr1), leukocyte immunoglobulin-like receptor B-3-like (Lilrb3l) and Zbtb16 were categorized as GO:0030097 (hemopoiesis) (Table II, G-4): Ccr1 encodes a member of the β-chemokine receptor family. Knockout studies on the mouse homolog suggested roles for this gene in host protection from inflammatory responses, and susceptibility to viruses and parasites. Lilrb3l is a receptor for the major histocompatibility complex class I antigen (MHC-I), and may play a physiological role in the brain for neuronal circuitry stability by inhibiting synaptic plasticity. Zbtb16 encodes a protein which is located in the nucleus. It is involved in cell cycle progression and interacts with a histone deacetylase.

We previously examined gene expression profiles in the brain, and found that 6 SHRSP-specific genes isolated from the rats at 6 weeks of age (Agtr1b, Arc, Egr2, Fos, Hspa1b and Snca) were annotated to ‘behavior' and were suggested to participate in the genesis of ADHD (5). In the present study, we investigated gene expression profiles in the kidneys, and unexpectedly found that 6 SHRSP-specific genes isolated from the rats at 6 weeks of age (Banp, Ephx2, Rbfox1, Rxrg, Ryr1 and Zbtb16) were annotated to ‘Huntington's disease' (Table IV, G-4). Tp53 was also found to be involved in ‘Huntington's disease' (33,52). These findings suggested the participation of common genes in the genesis of symptoms related to ADHD and Huntington's disease (Table IV, G-4).

SHR-specific genes isolated from the kidneys of 3-week-old rats included possible candidate genes that trigger hypertension (Bcl6 and Sox2), and SHRSP-specific genes isolated from the kidneys of 3-week-old rats included possible candidate genes that trigger stroke, such as Agt, Agtrap and Apoh. The results obtained from SHRSP-specific genes isolated from the kidneys of 6-week-old rats included 6 genes that have been functionally annotated to Huntington's disease (Banp, Ephx2, Rbfox1, Rxrg, Ryr1 and Zbtb16). These results implicate these genes in the involuntary movement associated with Huntington's disease as well as ‘attention-deficit hyperactivity' observed in ADHD.