Letters to the Editor 234 in women, in two cohorts of older normal individ- uals. These risk alleles have previously been Renalase, a novel soluble associated with schizophrenia, bipolar disorder, and FAD-dependent protein, social and physical anhedonia, mainly in females.1,7 Reports suggest that heritability of neuroticism, is synthesized in the brain anxiety and depression is higher in females than in males, and that the genes involved differ between and peripheral nerves men and women.8–10 We tested SNPs and models specifically chosen on the basis of earlier evidence, and hence do not believe that stringent multiple Molecular Psychiatry (2010) 15, 234–236; testing corrections are appropriate, but these results doi:10.1038/mp.2009.74 need to be replicated in other suitable cohorts before variation in DISC1 is fully accepted as contributing to normal variation in neuroticism and mood. In the central nervous system (CNS), the oxidative deamination of monoamine neurotransmitters is accomplished by two membrane-bound enzymes: Conflict of interest monoamine oxidase (of which there are two isoforms, The authors declare no conflict of interest. MAO-A and MAO-B) and semicarbazide-sensitive amine oxidase (SSAO). The combined activities of these proteins are crucial for the regulation of SE Harris1,2, W Hennah1,3,4, PA Thomson1, neurotransmitter disposition and, consequently, nor- M Luciano2, JM Starr5, DJ Porteous1 and IJ Deary2 mal brain function. It is therefore not surprising that 1Medical Genetics Section, Centre for Cognitive MAO-A and B gene polymorphisms and altered Ageing and Cognitive Epidemiology, University of expression are implicated in a variety of neurological 1–5 Edinburgh, Edinburgh, UK; 2Department of disorders. Moreover, the demonstration that MAO Psychology, Centre for Cognitive Ageing and Cognitive inhibitors, such as iproniazid, were effective anti- 6 Epidemiology, University of Edinburgh, depressant agents was pivotal in Schildkraut’s Edinburgh, UK; 3Institute for Molecular Medicine formulation of the catecholamine hypothesis of Finland FIMM, Nordic EMBL Partnership for affective disorders. Here, we report for the first time Molecular Medicine, Helsinki, Finland; 4Unit of the identification of a novel flavin adenine dinucleo- Public Health Genomics, National Institute tide (FAD)-dependent protein, renalase, in various for Health and Welfare, Helsinki, Finland and regions of the CNS. We show that the renalase gene is 5Geriatric Medicine Unit, Centre for Cognitive Ageing expressed in the hypothalamus and peripheral and Cognitive Epidemiology, University of Edinburgh, nerves. Furthermore, we reveal the existence of Royal Victoria, Edinburgh, UK several splice variants of the renalase gene, which E-mail: [email protected] potentially serve to further regulate levels of mono- amine neurotransmitters in the brain. Together, our findings provide further insight into the pathways regulating monoamine neurotransmitter disposition References in the brain. 1 Hennah W, Thomson P, McQuillin A, Bass N, Loukola A, Until recently, it was thought that MAO and SSAO Anjorin A et al. Mol Psychiatry 2008; doi:10.1038/mp.2008.22. were the only monoamine oxidases expressed in 2 Deary IJ, Whiteman MC, Starr JM, Whalley LJ, Fox HC. J Pers Soc humans. The discovery of a novel FAD-dependent Psychol 2004; 86: 130–147. protein, renalase, was reported in 2005.7 Renalase was 3 Deary IJ, Gow AJ, Taylor MD, Corley J, Brett C, Wilson V et al. BMC Geriatr 2007; 7: 28. identified using an in silico approach that aimed to 4 Goldberg L In: Mervielde I, Deary IJ, de Fruyt F, Ostendorf F (eds). discover novel proteins secreted by the kidney. The Personality Psychology in Europe, vol. 2. Tilburg University Press: renalase protein sequence contains a highly con- Tilburg, 1999, pp 7–28. served N-terminal FAD-binding domain and an amine 5 Costa Jr PT, McCrae RR. Psychological Assessment Resources: oxidoreductase domain. Renalase shares low se- Odessa, 1992. 6 Zigmond AS, Snaith RP. Acta Psychiatr Scand 1983; 67: quence identity with MAO-A and MAO-B (17 and 361–370. 20 %, respectively) but, nonetheless, its predicted 7 TomppoL,HennahW,MiettunenJ,JarvelinMR,VeijolaJ,RipattiS secondary and tertiary structures closely resemble et al. Arch Gen Psychiatry 2009; 66: 134–141. those of MAO-B.8 Recombinant renalase was shown 8 Bierut LJ, Heath AC, Bucholz KK, Dinwiddie SH, Madden PA, Statham DJ et al. Arch Gen Psychiatry 1999; 56: to generate hydrogen peroxide in the presence of 557–563. monoamines (including catecholamines), suggesting 9 Boomsma DI, Beem AL, van den BM, Dolan CV, Koopmans JR, that it may share the catecholamine-degrading activ- Vink JM et al. Twin Res 2000; 3: 323–334. ity of MAO-A and B.7 This activity was greatest in the 10 Kendler KS, Gardner CO, Neale MC, Prescott CA. Psychol Med presence of dopamine (followed by adrenaline and 2001; 31: 605–616. noradrenaline),7 but it was not inhibited by MAO Supplementary Information accompanies the paper on the Molecular inhibitors, indicating differences in the possible Psychiatry website (http://www.nature.com/mp) catecholamine-degrading actions of these proteins.

Molecular Psychiatry Letters to the Editor 235 The renalase sequence does not contain a membrane- catecholamine degradation in plasma,9 renalase tethering domain, indicating that, unlike MAO-A, B clearly has an important role in mammalian physio- and SSAO, it may be responsible for the oxidation logy, as shown by its ability to significantly improve of monoamines in the cytosol or extracellular space. hemodynamic parameters in animal models.7 Renalase was identified primarily in the human Given the putative functional similarities between kidney and heart, and was shown to be present in renalase and MAO-A and B, we investigated whether the plasma of healthy individuals.7 Although it has the tissue expression pattern of these proteins been suggested that renalase could not contribute to overlapped by searching for evidence of renalase

FV FA RV RA R CortexR MedullaUreter LV Hyp. MN LV Hyp. Pons MedullaS. cordCereb. Cortex Pit. 98 98 98 64 64 64 50 kDa kDa kDa 50 50 36 36 36

22

1 2 345 LV Adr. Hyp. -ve

1500 Wild-type 1200 Variant 1 1500

bp 1000 Variant 2 1000 bp Wild-type Variant 3 600 800 Variant 2 500

1.8% agarose 1.8% agarose

Wild-type 1 2 3 4 5 67 12 3 4 5 6 7 Variant 1

Variant 2 1 2 3 4 5 6 7

342 aa polypeptide Variant 3 1 2 34 56 7

Secretion peptide FAD Monoamine oxidase

Secretion peptide FAD Monoamine oxidase

Figure 1 (A): Western Blot analysis of human tissue samples. (i) Analysis of proteins extracted from the forearm vein (FV), forearm artery (FA), renal vein (RV), renal artery (RA), renal cortex (R. cortex), renal medulla (R. medulla), ureter and left ventricle (LV) from a single donor; (ii) analysis of proteins from the hypothalamus (Hyp.), median nerve (MN) and left ventricle (LV); (iii) analysis of proteins from various regions of the CNS, from a single individual: hypothalamus (Hyp.), pons, medulla, spinal cord (S. cord), cerebellum (cereb.), cortex and pituitary gland (pit.). In total, 20 mg of protein was loaded for each sample. A signal for renalase is observed at 37 kDa in all samples. Western blot analysis has been performed on the proteins of several donors; the figures shown are representative. (B). Reverse transcription PCR (RT-PCR) of renalase transcripts. (i) Agarose gel analysis of RT-PCR using renalase-specific primers to amplify transcripts in the left ventricle (LV), adrenal gland (Adr.) and hypothalamus (Hyp.). A no-template PCR was performed as a negative control. Four PCR products are observed: wild type (most prominent band) and variants 1–3. (ii) RT-PCR of renalase transcripts from the hypothalami of five individuals. (C). The intron–exon structure of the renalase gene. (i) Seven exons of the renalase gene encode a 342 amino acid polypeptide that includes N-terminal secretion peptides (red), flavin adenine dinucleotide (FAD)-binding domain (green) and monoamine oxidase domain (blue). (ii) Top: splice variants of the renalase gene. The splice variation between exons 5 and 7 results in a frameshift mutation, altering the coding sequence in exon 7 and is shown in green. Bottom: all splice variants still encode a polypeptide with N-terminal secretion peptide and FAD-binding site. The length of the polypeptide and its C-terminal sequence is altered in the variants.

Molecular Psychiatry Letters to the Editor 236 expression in tissues that have earlier been shown to identified as splice variants of the renalase gene. express MAO-A and B. Human tissue samples were Variant 1 (851 bp) did not contain exon 6. Instead, an obtained from the Victorian Institute of Forensic alternative splicing between exon 5 and exon 7 Medicine Tissue Donor Bank at autopsy. The in- resulted in a frameshift in the coding sequence of formed consent of the donor’s next of kin was exon 7 and a premature stop codon. Variant 2 (777 bp) obtained before the autopsy. All protocols were did not contain exons 2 and 3. In variant 3 (549 bp), approved by the Victorian Institute of Forensic exons 2, 3 and 6 were absent. We identified variants 2 Medicine Ethics Review Committee. The length of and 3 in all tissue types examined (Figure 1Bi). Only time between death and autopsy did not exceed 72 h. the hypothalamus appeared to express little to no Tissues were obtained from donors, male and female, variant 1, suggesting that there is tissue-specific who died from a variety of causes, including suicide, regulation of renalase function. Furthermore, an motor vehicle accident and drug overdose. Tissues examination of hypothalamic transcripts revealed were frozen in liquid nitrogen and ground to a fine that there are differences in the relative amounts of powder using a mortar and pestle. The tissue powder wild type and variant 2 transcripts between donors was resuspended in radioimmunoprecipitation assay (Figure 1Bii); however, these differences could not be tissue lysis buffer (25 mM Tris-HCl, pH 7.6, 150 mM ascribed to gender or mode of death. Whether this NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% relates to a difference in renalase function in these sodium dodecyl sulfate) containing protease inhi- individuals and whether the activity of the splice bitors (protease inhibitor cocktail, Roche, Basel, variants differs from that of wild-type renalase Switzerland), homogenized using a 25-G needle and warrants further investigation. then incubated on ice for 1 h. Soluble protein was Our identification of renalase, a soluble monoamine obtained by centrifugation at 10 000 Â g for 10 min at oxidase in the brain, highlights the possible existence 4 1C. Western blot analysis of soluble proteins was of other pathways regulating monoamine neurotrans- performed using a monoclonal antibody (from Ren mitter levels in the CNS. Inhibition of renalase Pharmaceuticals, CA, USA) raised against a 21 amino activity may represent a novel target in the future acid segment of the renalase protein. Results from design of therapeutics in the treatment of psychiatric various human tissues showed that, in addition to the disorders. heart and kidney, renalase is present in the forearm vein and artery, renal vein and artery and the ureter (Figure 1Ai). We also observed a signal for renalase in Conflict of interest the hypothalamus and median nerve (see Figure 1Aii). A more thorough analysis of the distribution The authors declare no conflict of interest. of renalase in the CNS revealed that, in addition to the hypothalamus, renalase is found in the pons, medulla SC Hennebry1, N Eikelis1, F Socratous1, G Desir2, oblongata, cerebellum, pituitary gland, cortex and G Lambert1 and M Schlaich1 spinal cord (Figure 1Aiii). 1Baker IDI Heart & Diabetes Institute, Melbourne, VIC, Having identified evidence of renalase protein Australia and 2Department of Medicine, Yale synthesis in various tissues, reverse transcriptase University School of Medicine, New Haven, CT, USA PCR was used to confirm renalase gene expression E-mail: [email protected] in the hypothalamus and other tissues. Total RNA was extracted from tissue using the RNEasy system from Qiagen (Germantown, MD, USA). cDNA was gener- References ated using the Superscript III cDNA synthesis kit according to the manufacturer’s instructions using 1 Brunner HG, Nelen MR, van Zandvoort P, Abeling NG, van Gennip m AH, Wolters EC et al. Am J Hum Genet 1993; 52: 1032–1039. 1 g of total RNA (Invitrogen, Carlsbad, CA, USA). 2 Brunner HG, Nelen M, Breakefield XO, Ropers HH, van Oost BA. PCR was performed using renalase-specific primers Science 1993; 262: 578–580. (forward: 50-ATGGCGCAGGTGCTGATCGTGGGC-30; 3 Jorm AF, Henderson AS, Jacomb PA, Christensen H, Korten AE, reverse: 50-CTAAATATAATTCTTTAAAGCTTCCAG- Rodgers B et al. Psychiatr Genet 2000; 10: 87–90. 30) under the following conditions: 95 1C for 15 min, 4 Lenders JW, Eisenhofer G, Abeling NG, Berger W, Murphy DL, 1 1 1 Konings CH et al. J Clin Invest 1996; 97: 1010–1019. 35 cycles of 95 C for 30 sec, 58 C for 30 sec, 72 C for 5 Shih JC, Thompson RF. Am J Hum Genet 1999; 65: 593–598. 1 min, followed by a final extension at 72 1C for 6 Schildkraut JJ. Am J Psychiatry 1965; 122: 509–522. 10 min. Left ventricle RNA was used as a positive 7 Xu J, Li G, Wang P, Velasquez H, Yao X, Li Y et al. J Clin Invest 2005; control. In the tissues examined (hypothalamus, 115: 1275–1280. 8 Hennebry SC, Eikelis N, Socratous F, Desir G, Lambert GW, adrenal gland and left ventricle), the major PCR Schlaich MP. Proceedings of the HBPRCA Annual Scientific product was of the expected size of 1026 bp. Three Meeting 2008 (abstract). Hypertension 2009; 53: 1098–1123. smaller PCR products were observed, sequenced and 9 Boomsma F, Tipton KF. J Neural Transm 2007; 114: 775–776.

Molecular Psychiatry