We identified the CC-5 mutation in 3 patients and in 0 controls, a frequency comparable to that of the PPP3CC/schizophrenia study (i.e. 3 of 210 tested patients and 0 of 75 Caucasian controls; [2
]). For CC33 and CCS3, Hardy-Weinberg-equilibrium was tested separately in cases and controls. We also estimated linkage disequilibrium values D' and r2
]. We tested for disease-haplotype association by likelihood-based approaches using the EH-plus software [16
]. No deviation from Hardy-Weinberg-equilibrium was observed for CC33 and CCS3, in both groups. Association between BPAD and CC33 genotype was borderline significant (p = 0.07), with an higher risk to develop bipolar disorder for 'CT' or 'TT' carriers versus 'CC' carriers (OR = 1.8 [1.01–3.0]; p = 0.05) and a significant allelic association was observed with the 'T'-allele of CC33 (p = 0.05, see Table for details). Our results show also a significant association between BPAD and CCS3 (p = 0.003), with an higher risk to develop bipolar disorder for 'AG' or 'GG' carriers versus 'AA' carriers (OR = 2.8 [1.5–5.2]) and a significant allelic association between the 'G'-allele of CCS3 and BPAD was observed (p = 0.002, see table for details). The CC33 and CCS3 polymorphisms were observed in significant linkage disequilibrium (D' = 0.91, r2
= 0.72). Haplotype frequencies were significantly different in BPAD patients than in controls (Likelihood ratio test: Chi2
(3 ddf) = 9.08, p = 0.03), and a significant over-transmission of the PPP3CC haplotype 'TG' (p = 0.001) was observed for bipolar patients. HapMap LD data for the PPP3CC gene support the association being due to a variant in PPP3CC and not a nearby gene.
Genotypic, Allelic and Haplotype association study of PPP3CC -CC33 and -CCS3 polymorphisms
In order to test an association between psychotic symptoms and PPP3CC polymorphisms, we have compared genotypic repartition in psychotic (N = 59) versus non-psychotic (N = 56) patients for both CC33 and CCS3 polymorphism. No difference was observed. Moreover, psychotic and non-psychotic patient samples shown similar association when compared to controls, suggesting that these variants are implicated in bipolar disorder per se and are not specific of psychotic symptoms.
As male/female ratio was significantly different between patients and controls, we have performed genotypic and haplotypic analyses stratified on gender. Similar results were observed: a higher risk to develop bipolar disorder for 'CT' or 'TT' carriers for CC33: OR = 1.45 [0.67–3.15] for males and females OR = 2.54 [1.10–5.91]; association between bipolar disorder and CC33 being significant only for females (p = 0.03). For CCS3 polymorphism, 'AG' or 'GG' carriers have a higher risk than 'AA' carriers to develop disease, (males OR = 1.29 [0.92–1.83] and females OR = 2.57 [1.57–4.21]). However, association between bipolar disorder and CC33 is significant only for females (p < 0.001). Also association of CC33 haplotypes was significant in females (Chi2 (3 df) = 50.6; p < 0.001) and not significant in males (Chi2 (3 df) = 4.4; p = 0.22). An over transmission of 'TG' haplotype was observed for bipolar males and female patients.
Our results suggest that the PPP3CC gene may contribute to the etiology of BPAD. This is in accordance with preclinical work that relates calcineurin related signaling pathways to the pathophysiology of affective and cognitive disorders. Specifically it was shown that phosphatase activity/amount could predict whether a signaling system will operate in a proportional-response-mode, or as an all-or-none-switch [17
] susceptible to generate disproportional network responses. This might be relevant to the bidirectional behavioral dysfunction (depressed or agitated mood) characterizing BPAD. Furthermore, as stated above, preclinical studies have implicated calcineurin in cognitive function as well as in affective responsiveness [1
]. Our present findings suggest that PPP3CC might be a shared susceptibility gene for schizophrenia and BPAD. Calcineurin dysregulation, therefore, might be associated with transnosographic phenotypical aspects, such as affective and cognitive pathology, seen in both disorders.