Phosphodiesterase 11A in brain is enriched in ventral hippocampus and deletion causes psychiatric disease-related phenotypes

Abstract

Phosphodiesterase 11A (PDE11A) is the most recently identified family of phosphodiesterases (PDEs), the only known enzymes to break down cyclic nucleotides. The tissue expression profile of this dual specificity PDE is controversial, and little is understood of its biological function, particularly in the brain. We seek here to determine if PDE11A is expressed in the brain and to understand its function, using PDE11A(-/-) knockout (KO) mice. We show that PDE11A mRNA and protein are largely restricted to hippocampus CA1, subiculum, and the amygdalohippocampal area, with a two- to threefold enrichment in the ventral vs. dorsal hippocampus, equal distribution between cytosolic and membrane fractions, and increasing levels of protein expression from postnatal day 7 through adulthood. Interestingly, PDE11A KO mice show subtle psychiatric-disease-related deficits, including hyperactivity in an open field, increased sensitivity to the glutamate N-methyl-D-aspartate receptor antagonist MK-801, as well as deficits in social behaviors (social odor recognition memory and social avoidance). In addition, PDE11A KO mice show enlarged lateral ventricles and increased activity in CA1 (as per increased Arc mRNA), phenotypes associated with psychiatric disease. The increased sensitivity to MK-801 exhibited by PDE11A KO mice may be explained by the biochemical dysregulation observed around the glutamate alpha-amino-3-hydroxy-5-methyl-4-isozazolepropionic (AMPA) receptor, including decreased levels of phosphorylated-GluR1 at Ser845 and the prototypical transmembrane AMPA-receptor-associated proteins stargazin (gamma2) and gamma8. Together, our data provide convincing evidence that PDE11A expression is restricted in the brain but plays a significant role in regulating brain function.

Fig. 1.

PDE11A is enriched in ventral hippocampus. (A) PDE11A knockout (KO) mice from Deltagen were generated using a Lac-Z neo cassette to delete catalytic domain amino acids 693–717 (corresponding to bases 2123–2199 and indicated by “X”). (B, 1) Nissl stain of brain section that corresponds to (B, 2) autoradiographic in situ hybridizations for PDE11A mRNA. Comparison of the autoradiographs with the Nissl stain suggests that PDE11A mRNA is enriched in ventral hippocampus CA1, subiculum, and the amygdalohippocampal area in wild-type mice (WT) and is absent in knockout mice (KO). Labeling in rat shows an identical pattern (Fig. S2). (C) Quantitative PCR analyses of PDE11A mRNA [primers: 2707–2728/2775–2755, normalized to GAPDH mRNA and expressed as percentage of wild-type (%WT) dorsal hippocampus (DHIPP)] and (D) Western blotting of PDE11A protein (100 μg total protein; 2 ng purified human PDE11A4-His as positive control, FabGennix 112AP antibody to amino acids 454–468, normalized to actin) confirms that PDE11A mRNA and protein are enriched in WT mouse ventral hippocampus (VHIPP) and absent in KO hippocampus. (E) PDE11A protein (normalized to actin and expressed as %WT DHIPP) appears threefold higher in VHIPP vs. DHIPP of WTs. (F) PDE11A protein also appears to be expressed in human hippocampus (60 μg total protein; 1 ng purified PDE11A4). (G) Fractionation of mouse hippocampus shows that PDE11A expression is equally distributed between cytosolic (CYTO) and membrane (MEMB) fractions in WT mice. (H) Finally, Western blotting experiments show that PDE11A protein is developmentally regulated with increasing levels of expression from postnatal day 7 (P7) to adulthood. In all Western blotting experiments, the PDE11A antibody was able to detect purified PDE11A4-His at the appropriate molecular weight (∼98 kDa), and the signal disappeared when the primary antibody was preincubated with a blocking peptide or when tested against KO samples, suggesting specificity of the signals observed. In addition, similar results were observed with two additional PDE11A antibodies (Fig. S1). PFC, prefrontal cortex; STR, striatum; CBLM, cerebellum.

Fig. 2.

PDE11A knockout mice exhibit subtle behavioral abnormalities consistent with ventral hippocampal dysfunction. (A) PDE11A knockout (KO) and heterozygous (HT) mice are normal relative to wild-type (WT) littermates on measures of startle and prepulse inhibition (PPI) of acoustic startle, suggesting intact sensorimotor gating. (B) PDE11A KO and HTs also exhibit normal improvement across days of rotarod training (effect of days: F_(2,222) = 130.47, P < 0.0001), suggesting intact motor coordination and procedural learning. (C) In contrast, PDE11A KO mice are hyperactive in a novel open field (effect of genotype: F_(2,320) = 4.59, P = 0.014; post hoc KO vs. WT and HT: P = 0.024–0.0005); however, they do habituate equally over time. (D) Following habituation to the open field, PDE11A KO mice do not show hyperactivity after saline injection. PDE11A KO mice do, however, show significantly increased locomotion following injection of 0.3 mg/kg MK-801 (F_(2,1653) = 3.32, P = 0.043; post hoc WT vs. KO: P = 0.01; WT vs. HT: P = 0.064), suggesting alterations in glutamatergic signaling. PDE11A KO mice also show alterations in socially based behaviors. (E) During social odor recognition (SOR) training Trial 1, PDE11A WT, HT, and KO mice spend more time sniffing beads scented with a stranger (stranger1) vs. themselves (home cage; effect of bead: F_(1,180) = 28.12, P < 0.0001) and spend less time sniffing the novel scented beads during Trial 2 (effect of trial: F_(1,180) = 28.03, P < 0.0001). This suggests that the ability to detect, identify, and learn about novel odors remains intact in PDE11A knockout mice. Twenty-four hours later, PDE11A WT mice exhibit a robust memory for the stranger1 bead that they learned about during training, as indicated by a significant difference in the percentage of time spent sniffing stranger1 vs. a bead from novel stranger2 (genotype × bead: F_(4,155) = 2.46, P = 0.048; post hoc within WT, donor 2 vs. donor 1: P < 0.0001). PDE11A HT mice also exhibit memory for the stranger1 bead (HT, stranger2 vs. stranger1, P = 0.015); however, this memory appears less robust (stranger2 WT vs. stranger2 HT: P = 0.022). In contrast, PDE11A KOs fail to show significant recognition memory for stranger1. (F) Despite showing a normal approach to a stranger-scented bead during SOR training, PDE11A KO males avoid a chamber containing an actual stranger mouse (genotype × sex × chamber: F_(4,102) = 2.80, P = 0.03; post hoc vs. WT-F and WT-M: P = 0.049–0.013; vs. stranger: P = 0.001). WT, n = 20–41; HT, n = 16–37; KO, n = 18–39. F, females; M, males. Post hoc: WT vs. KO across time—*P = 0.024–0.01; vs. stranger (1)—^#P = 0.015 to <0.001; vs. WT within bead/compartment—^@P = 0.022–0.013. Data graphed are mean ± SEM.

Fig. 3.

PDE11A knockout mice exhibit significantly enlarged ventricles. (A) Thionin-stained sagital sections from brain suggest that PDE11A knockout (KO) mice show limited neuroanatomical alterations relative to wild-type (WT) littermates. PDE11A KO mice exhibit a small, but significant, enlargement of the lateral ventricles in sections taken (B) 2.76 mm lateral from Bregma (genotype × region: F_(6,126) = 2.32, P = 0.037) and (C) 1.32 mm lateral from Bregma (genotype × region: F_(10,205) = 2.36, P = 0.012). WT, n = 19; heterozygous (HT), n = 15; KO, n = 20. Post hoc vs. WT: *P < 0.001. Data graphed are mean ± SEM.

Fig. 4.

PDE11A KO mice show biochemical alterations consistent with glutamatergic hypofunction. (A) Across sexes, PDE11A knockout (KO) mice exhibit a small but significant reduction in the ratio of phosphoGluR1(Ser845):totalGluR1 (an AMPA receptor subunit), relative to wild-type (WT) littermates (effect of genotype: F_(1,30) = 4.32, P = 0.046). In contrast, PDE11A KOs show normal levels of phosphoGluR2/3 (S880/S891):total GluR2/3 (AMPA subunit), phosphoNR2B(Y1336):total NR2B (NMDA receptor subunit), and total NR2A (NMDA subunit). (B) Across sexes, PDE11A KO mice also show a reduction in levels of the transmembrane AMPA receptor associated regulatory proteins stargazin (γ2) and γ8 (effect of genotype: F_(1,34) = 5.89, P = 0.021), but not the NMDA receptor scaffolding protein PSD-95. (C) Consistent with the disinhibition that would be predicted to occur as a consequence of glutamatergic hypofunction, PDE11A KO mice show increased neural activity in ventral CA1, as mapped by Arc mRNA (genotype × region × sex: F_(6,132) = 2.45, P = 0.028). WT, n = 19–20; heterozygous (HT), n = 16; KO, n = 20. F, female; M, male. Post hoc vs. WT: *P = 0.046–0.01. Data graphed are mean ± SEM.