Letters to the Editor 118 Zucker Hillside Hospital/Feinstein Institute/AECOM, response to elevated glucose levels along with State University of New York at Stony Brook, approximately 100 other proteins, including residual Stony Brook, NY, USA proinsulin, the conversion intermediates des31,32- E-mail: [email protected] proinsulin and des64,65-proinsulin, C-peptide and chromogranin-like molecules.5 Here we have inves- References tigated whether the circulating levels of these major secreted proteins are altered in schizophrenia 1 Egan MF, Kojima M, Callicott JH, Goldberg TE, Kolachana BS, patients. Bertolino A et al. Cell 2003; 112: 257–269. 2 Pezawas L, Verchinski BA, Mattay VS, Callicott JH, Kolachana BS, Molecular studies of chronic schizophrenia pa- Straub RE et al. J Neurosci 2004; 24: 10099–10102. tients can be confounded since routinely used 3 Sohrabji F, Lewis DK. Front Neuroendocrinol 2006; 27: 404–414. antipsychotic medications have several side effects, 4 Murty VP, Sambataro F, Das S, Tan HY, Callicott JH, Goldberg TE such as dysregulated glucose homeostasis. We cir- et al. J Cogn Neurosci 2008; e-pub ahead of print 29 September cumvented this problem by carrying out analyses of 2008. 5 Webster MJ, Herman MM, Kleinman JE, Shannon Weickert C. Gene serum and plasma samples from first-onset, antipsy- Expr Patterns 2006; 6: 941–951. chotic naı¨ve patients. Recruitment of such patients is 6 Hariri AR, Goldberg TE, Mattay VS, Kolachana BS, Callicott JH, challenging, as large clinics diagnose only 20–30 such Egan MF et al. J Neurosci 2003; 23: 6690–6694. patients each year and few centers follow standard operating procedures for sample collection. To Supplementary Information accompanies the paper on the Molecular Psychiatry website (http://www.nature.com/mp) achieve adequate numbers of well-characterized first-onset patients, we recruited subjects from four independent clinical centers over 2006–2008. Patients were diagnosed using the Diagnostic and Statistical Manual of Mental Disorders-IV criteria for Increased levels of schizophrenia and bipolar disorder. Schizophrenia patients were acutely psychotic (Positive and circulating insulin-related Negative Symptoms Scale total = 87±16; n = 66) and peptides in first-onset, bipolar disorder subjects were euthymic (Young Mania Ratings Scale total = 3.9±4.7, Hamilton antipsychotic naı¨ve Depression scale total = 4.6±5.4; n = 10) at the time of sample collection. Euthymic bipolar disorder schizophrenia patients patients were chosen, as such subjects often experi- ence cognitive deficits similar to those observed in schizophrenia and this can be a potential means of 6 Molecular Psychiatry (2010) 15, 118–119; misdiagnosis. Control subjects (n = 78) were matched doi:10.1038/mp.2009.81 for age, gender, BMI and smoking. Insulin, proinsulin and des31,32-proinsulin were determined through two-site time-resolved fluores- Here, we show that circulating insulin-related pep- cence assays using combinations of monoclonal tides are elevated in first-onset schizophrenia sub- antibodies that discriminate between the specific jects, with no relative difference in glucose levels. forms of the molecule.7 C-peptide and chromogranin This suggests that hyperinsulinemia may have a A were measured using commercially available role in the development of schizophrenia, and that immunoassays. Insulin, proinsulin, des31,32-pro- measurement of these peptides may have utility insulin and C-peptide were found to be present in diagnosis or stratification of patients before anti- at significantly elevated levels in the serum and psychotic treatment. Moreover, drugs that improve plasma from schizophrenia patients (Table 1). In insulin action may represent a novel treatment addition, chromogranin A was also found at signi- strategy. ficantly elevated levels. We, along with others, have identified alterations of Glucose levels were unchanged in schizophrenia metabolic biomarkers in schizophrenia patients, patients, except for cohort 3, which showed a slight which are indicative of perturbations in glucoregula- elevation (Table 1). Therefore, the observed changes tory pathways.1 In addition, schizophrenia patients in insulin-related molecules occurred against a back- show increased prevalence of impaired glucose ground of relatively normal glucose levels, suggesting tolerance2 and metabolic syndrome, irrespective of that at least some patients show signs of insulin whether they received antipsychotic treatment.3 Dys- resistance. Another factor that should be considered regulation of glucose metabolism is normally accom- is the lack of gender balance in some cohorts. How- panied by hyperinsulinemia, because of increased ever, covariate analyses showed that this factor had secretory demands on pancreatic b-cells to maintain no influence on analyte levels (data not shown). This homeostasis.4 Insulin is the major glucoregulatory shows the reproducibility of the findings irrespective hormone produced in b-cell secretory granules by of confounding factors such as glycemic status and complete proteolytic cleavage of proinsulin. The gender. In contrast, no significant differences were mature hormone is released into the circulation in found regarding insulin-related molecules in bipolar Molecular Psychiatry Letters to the Editor 119 Table 1 Increased circulating levels of insulin-related that drugs that improve insulin signaling may repre- peptides in schizophrenia subjects sent a novel treatment strategy. In this regard, the insulin-related molecules identified here, and poten- FC P-value tially other co-secreted insulin-secretory granule proteins, may have utility as biomarkers for patient Schizophrenia 1 stratification and for monitoring the responses to Glucose 1.07 0.520 existing and novel therapeutic treatment strategies. Insulin 3.77 0.034 1 1 1 2 1 Schizophrenia 2 PC Guest , L Wang , LW Harris , K Burling , Y Levin , Glucose 0.92 0.169 A Ernst1, MT Wayland1, Y Umrania1, M Herberth1, Insulin 2.36 0.020 D Koethe3, JM van Beveren4, M Rothermundt5, Proinsulin 2.01 0.015 G McAllister6, FM Leweke3,7, J Steiner8 and S Bahn1 des31,32-proinsulin 1.78 0.023 1Institute of Biotechnology, University of Cambridge, Cambridge, UK; 2NIHR Biomedical Research Centre, Schizophrenia 3 Department of Clinical Biochemistry, Glucose 1.30 0.017 University of Cambridge, Cambridge, UK; Insulin 4.00 0.030 3 Proinsulin 2.67 0.033 Department of Psychiatry and Psychotherapy, des31,32-proinsulin 5.48 0.010 University of Cologne, Cologne, Germany; 4 C-peptide 2.26 0.019 Department of Psychiatry, Erasmus University Medical Centre, Rotterdam, Netherlands; Schizophrenia 4 5Department of Psychiatry, University Medical Glucose ND ND Faculty, Mu¨nster, Germany; 6Psynova Neurotech Ltd, Proinsulin 2.59 0.003 St John’s Innovation Centre, Cambridge, UK; 7Central C-peptide 1.54 0.006 Institute of Mental Health, University of Heidelberg, Chromogranin A 1.52 0.043 Mannheim, Germany and 8Department of Psychiatry, University of Magdeburg, Magdeburg, Germany Bipolar disorder Glucose 0.89 0.453 E-mail: [email protected] Proinsulin 0.80 0.484 des31,32-proinsulin 0.78 0.582 References 1 Holmes E, Tsang TM, Huang JT et al. PLoS Med 2006; 3: e327. ND, not determined. 2 Spelman LM, Walsh PI, Sharifi N, Collins P, Thakore JH. Diabet Med Subjects were matched for age (schizophrenia (S)1 (serum) = 2007; 24: 481–485. 25±7 years; S2 (plasma) = 23±4 years; S3 (plasma) = 34±12 3 Newcomer JW. Am J Manag Care 2007; 13(Suppl): S170–S177. years; S4 (serum) = 30±7 years; bipolar disorder (BD) 4 Shanik MH, Xu Y, Skrha J, Dankner R, Zick Y, Roth J. Diabetes Care (serum) = 37±11 years) and body mass index (24±5kgmÀ2 2008; 31: S262–S268. across all cohorts). Cohorts were comprised (male/female) 5 Guest PC, Bailyes EM, Rutherford NG, Hutton JC. J Biol Chem 1992; of the following: S1—controls (cont) = 12/5, schiz = 10/0; 274: 73–78. S2—cont = 21/0, schiz = 26/0; S3—cont = 6/4, schiz = 7/3; 6 Ferrier IN, Stanton BR, Kelly TP, Scott J. Br J Psychiatry 1999; 175: S4—cont = 9/11, schiz = 10/10; BD—cont = 5/5; BD = 5/5. 246–251. 7 Sobey WJ, Beer SF, Carrington CA et al. Biochem J 1989; 260: Glucose was determined using the Dimension RXL system 535–541. (Dade Behring, Dallas, TX, USA) (control levels = 4.6±0.8 mM). 8 Taguchi A, Wartschow LM, White MF. Science 2007; 317: 369–372. Insulin was determined using a two-step time-resolved fluoro- 9 Andersen L, Dinesen B, Jørgensen P, Poulson F, Røder M. Clin Chem metric (TRF) assay9 (control levels = 75.1±58.6 pM). Proinsulin 1993; 39: 578–582. and des31,32-proinsulin were determined using two-site TRF assays7 (control levels = 4.9±4.4 and 5.9±5.2 pM, respectively). C-peptide was measured using the Immulite system (DPC Diagnostic Products Corp, Los Angeles, CA, USA) (control levels = 1.9±1.3 nM). Chromogranin A was measured using a CACNA1C (rs1006737) is two-site immunoassay kit (ALPCO Diagnostics, Salem, NH, USA) (control levels = 29.6±21.3 ng mlÀ1). Fold change (FC) associated with was calculated as the disease:control ratio of analyte levels. Statistical significance (P-value) was determined by two-tailed schizophrenia t-tests. Significant FC values are indicated in bold font. disorder patients (Table 1), suggesting that these Molecular Psychiatry (2010) 15, 119–121; molecules are not altered in all neuropsychiatric doi:10.1038/mp.2009.69 disorders. Taken together, these findings show that hyper- In a large collaborative study combining three insulinemia may have a role in the onset of schizo- separate whole-genome association studies, the CAC- phrenia. This has important implications, as elevated NA1C gene (rs1006737) was recently found to display insulin levels can have deleterious effects on brain a genome-wide significant association with bipolar function.8 In addition, this suggests the possibility disorder (BPD). Here, we report for the first time the Molecular Psychiatry.