Neuronal NF1/RAS regulation of cyclic AMP requires atypical PKC activation

Abstract

Neurofibromatosis type 1 (NF1) is a common neurodevelopmental disorder in which affected individuals are prone to learning, attention and behavioral problems. Previous studies in mice and flies have yielded conflicting results regarding the specific effector pathways responsible for NF1 protein (neurofibromin) regulation of neuronal function, with both cyclic AMP (cAMP)- and RAS-dependent mechanisms described. Herein, we leverage a combination of induced pluripotent stem cell-derived NF1 patient neural progenitor cells and Nf1 genetically engineered mice to establish, for the first time, that neurofibromin regulation of cAMP requires RAS activation in human and mouse neurons. However, instead of involving RAS-mediated MEK/AKT signaling, RAS regulation of cAMP homeostasis operates through the activation of atypical protein kinase C zeta, leading to GRK2-driven Gαs inactivation. These findings reveal a novel mechanism by which RAS can regulate cAMP levels in the mammalian brain.

Figure 1.

Neurofibromin regulates cAMP in a RAS/Gα_s-dependent manner. (A) Quantification of hippocampal neuron axons lengths by Smi-312 immunostaining. Nf1+/− mouse hippocampal neuron axons are significantly shorter than WT neurons (P < 0.001; n = 200). (B and C) Measurement of cAMP generation in mouse hippocampal neurons and human NF1 patient-derived NPCs (hNF1-NPCs). (B) Nf1+/− neurons have lower cAMP levels relative to their WT counterparts (P = 0.0002; n = 5). (C) hNF1-NPCs have reduced cAMP levels compared with age- and sex-matched controls (P < 0.0001; n = 3). (D and E) Quantification of Gα_i and Gα_s activation of mouse embryonic hippocampal preparations. (D) Nf1+/− mouse hippocampal preparations show no difference in Gα_i activation relative to WT neurons (P = 0.7638; n = 5). (E) Nf1+/− mouse embryonic hippocampal preparations exhibit significantly lower Gα_s activity (Gα_s-GTP) than their WT counterparts (P = 0.0001; n = 8). (F and G) Measurement of RAS activation in mouse neurons and human NPCs. (F) Nf1+/− mouse hippocampal neurons exhibit higher levels of RAS activation (P = 0.0027; n = 5). (G) hNF1-NPCs (NF1) exhibit higher levels of RAS activation than control (CTRL) NPCs (P = 0.0002; n = 3). Data are presented as means ± SEM (n ≥ 3). ***P < 0.001; Student's t-test.

Figure 2.

Pharmacologic and genetic reduction of RAS activity corrects Nf1+/− neuronal defects. (A) Quantification of Gα_s activation and cAMP levels in mouse hippocampal neuron preparations. Nf1+/− mouse neuronal Gα_s activity (P < 0.0001; n = 6) and cAMP generation (P < 0.0001; n = 5) are restored to WT levels after lovastatin (Lov) treatment. (B) Measurement of axonal lengths by Smi-312 immunostaining. Lov administration restores Nf1+/− mouse hippocampal neuron axonal lengths to WT levels (P < 0.0001; n = 150). (C) Genetic Nras reduction restores Gα_s activation (P < 0.0001; n = 6) and cAMP levels (P < 0.005; n = 4) in Nf1+/− mouse hippocampal preparations to WT levels. (D) Smi-312 immunostaining of mouse hippocampal neurons. Nf1+/−; Nras+/− neurons exhibit axonal lengths indistinguishable from WT neurons (P < 0.0001; n = 73). Data are presented as means ± SEM (n ≥ 5). **P < 0.01; ***P < 0.001; One-way ANOVA with Bonferroni post-test correction. Scale bars 50 µm.

Figure 3.

RAS regulates Gα_s/cAMP activity in a PKCζ-dependent manner. (A) Immunoblot analysis of PKCζ activity in mouse hippocampal neuron preparations. PKCζ activity (phosphorylation) is increased in Nf1+/− neurons in vitro (n = 5). (B) Immunohistochemical detection of active PKCζ in adult mouse hippocampi. PKCζ activity is increased in Nf1+/− mouse hippocampi in vivo (n = 3). Insets depict representative immunopositive neurons. Scale bar left: 50 µm; right: 12.5 µm. (C) Smi-312 immunostaining of human NPC colony cultures. Representative images of control (CTRL) and NF1 patient (NF1) NPC colonies (top panels) and their derivative differentiated neurons (bottom panels). Scale bars: 50 µm. (D) Immunoblot analysis of PKCζ activity in human NPCs. PKCζ is activated in human NF1 patient-derived NPCs (NF1) relative to those from age- and sex-matched control (CTRL) individuals (n = 2). (E and F) Immunoblot analysis of PKCζ phosphorylation in Nf1+/− mouse neurons after pharmacologic or genetic reduction of RAS activity. (E) Lov treatment or (F) genetic Nras reduction restores PKCζ activity to WT levels (n = 3).

Figure 4.

Pharmacologic and genetic inhibition of PKCζ restores neuronal defects. (A) Immunoblot analysis of Gα_s activity of mouse hippocampal neurons following administration of the PKCζ pseudosubstrate (PKCζ-ps). Treatment of Nf1+/− mouse hippocampal neurons with PKCζ-ps increases Gα_s activity (P < 0.0001; n = 3). (B and C) Quantification of cAMP levels in mouse neurons and human-derived NPCs after PKCζ-ps treatment. PKCζ-ps administration restores cAMP in (B) Nf1+/− mouse hippocampal neurons (P < 0.005; n = 4) and (C) human NF1 patient-derived NPCs (P < 0.0001; n = 3) to control levels. (D and E) Smi-312 immunostaining of mouse hippocampal neurons and measurements of axonal length. (D) PKCζ-ps treatment corrects the axonal length defects of Nf1+/− mouse neurons to WT levels (P < 0.0001; n = 200). (E) Genetic inhibition of PKCζ with siRNA in Nf1+/− mouse neurons restores axonal length to WT levels (P < 0.001; n = 150). Data are presented as means ± SEM. **P < 0.01; ***P < 0.001; One-way ANOVA with Bonferroni post-test correction. Scale bars: 50 µm.

Figure 5.

PKCζ inhibits Gα_s/cAMP in a GRK2-dependent manner. (A) Immunoblot analysis of GRK2 in mouse neurons and human-derived NPCs. GRK2 activation (phosphorylation) and total protein expression is increased in both Nf1+/− mouse neurons and human NF1 patient-derived NPCs. (B) Immunoblot analysis of GRK2 activity in mouse neurons and human NPCs after PKCζ-ps treatment. PKCζ-ps administration normalizes GRK2 activity in mouse Nf1+/− neurons and human NF1 patient-derived NPCs. (C) Immunoblot analysis of Gα_s activity following GRK2 inhibition (GRK2-inh) in mouse hippocampal preparations. GRK2-inh treatment restores Gα_s activity to WT levels (P < 0.0001; n = 3). (D) Quantification of cAMP levels in mouse neurons and human NPCs after GRK2-inh treatment. GRK2-inh corrects cAMP in Nf1+/− mouse neurons (P = 0.0011; n = 3) and human NF1 patient-derived NPCs (P < 0.0001; n = 3) to WT levels. (E) Measurement of mouse neuron axonal length by Smi-312 immunoblotting following GRK2-inh treatment. GRK2-inh treatment corrects Nf1+/− mouse neuron axonal lengths to WT levels (P < 0.0001; n = 132). (F) Schematic representation of the proposed mechanism underlying neurofibromin cAMP regulation in mammalian CNS neurons. Data are presented as means ± SEM. ***P < 0.001; One-way ANOVA with Bonferroni post-test correction. Scale bar: 50 µm.