DISTRIBUTION AND NEUROCHEMICAL CHARACTERIZATION OF PROTEIN KINASE C-THETA AND -DELTA IN THE RODENT HYPOTHALAMUS

Abstract

PKC-theta (PKC-θ), a member of the novel protein kinase C family (nPKC), regulates a wide variety of functions in the periphery. However, its presence and role in the CNS has remained largely unknown. Recently, we demonstrated the presence of PKC-θ in the arcuate hypothalamic nucleus (ARC) and knockdown of PKC-θ from the ARC protected mice from developing diet-induced obesity. Another isoform of the nPKC group, PKC-delta (PKC-δ), is expressed in several non-hypothalamic brain sites including the thalamus and hippocampus. Although PKC-δ has been implicated in regulating hypothalamic glucose homeostasis, its distribution in the hypothalamus has not previously been described. In the current study, we used immunohistochemistry to examine the distribution of PKC-θ and -δ immunoreactivity in rat and mouse hypothalamus. We found PKC-θ immunoreactive neurons in several hypothalamic nuclei including the ARC, lateral hypothalamic area, perifornical area and tuberomammillary nucleus. PKC-δ immunoreactive neurons were found in the paraventricular and supraoptic nuclei. Double-label immunohistochemisty in mice expressing green fluorescent protein either with the long form of leptin receptor (LepR-b) or in orexin (ORX) neurons indicated that PKC-θ is highly colocalized in lateral hypothalamic ORX neurons but not in lateral hypothalamic LepR-b neurons. Double-label immunohistochemistry in oxytocin-enhanced yellow fluorescent protein mice or arginine vasopressin-enhanced green fluorescent protein (AVP-EGFP) transgenic rats revealed a high degree of colocalization of PKC-δ within paraventricular and supraoptic oxytocin neurons but not the vasopressinergic neurons. We conclude that PKC-θ and -δ are expressed in different hypothalamic neuronal populations.

Protein kinase C (PKC) comprises a group of serine/threonine kinases which play an important role in regulating a broad spectrum of neuronal activities including neuronal differentiation (Campenot et al., 1994), modulation of ion channels (Shearman et al., 1989), desensitization of receptors (Huganir and Greengard, 1990) and glutamate-induced neurotoxicity (Felipo et al., 1993). One class of PKCs, called the novel PKC (nPKC) includes PKC-δ, ε, η and θ. These PKCs exhibit enzymatic activity in the presence of phosphatidylserine (PS) and diacylglycerol (DAG), but are independent of intracellular calcium (Ca²⁺) levels (Ono et al., 1988; Konno et al., 1989). PKC-δ and ε isoforms have previously been demonstrated to be widely distributed in several regions of the CNS (Wetsel et al., 1992; Saito et al., 1993; Tanaka and Nishizuka, 1994). In contrast, early reports suggested PKC-θ was mainly expressed in peripheral tissues such as skeletal muscle (Osada et al., 1992), and not in the rodent brain (Tanaka and Nishizuka, 1994). Nearly a decade later, Minami and colleagues detected PKC-θ expression in the habenula of the rat brain (Minami et al., 2000). However, using in situ hybridization, they did not observe hypothalamic expression of PKC-θ. Additionally, PKC-δ immunoreactivity, while present in the brain, was also not previously observed in hypothalamic nuclei (Merchenthaler et al., 1993; Naik et al., 2000).

Recently, we and others demonstrated expression of PKC-θ (Dewing et al., 2008) and its immunoreactivity (Benoit et al., 2009) in the arcuate hypothalamic nucleus (ARC). The ARC is a key CNS site known to regulate food intake (Schwartz et al., 2000), glucose homeostasis (Anand et al., 1964), reproduction (Hill et al., 2008) and arousal (Yamanaka et al., 2003; Adamantidis and de Lecea, 2008). The ARC contains two neuronal populations with opposing actions, the anorexigenic proopiomelanocortin (POMC) neurons and the orexigenic neuropeptide Y/agouti-related protein (NPY/AgRP) neurons. Both of these express the long form of the leptin receptor (LepR-b) (Mercer et al., 1996; Cheung et al., 1997) as well as the insulin receptor (Marks et al., 1990; Niswender and Schwartz, 2003). ARC POMC neurons are activated by insulin and leptin (Cowley et al., 2001) and secrete the anorexic neuropeptide, α-melanocyte stimulating hormone (α-MSH). Orexigenic neurons expressing NPY/AgRP are inhibited by insulin and leptin (Niswender and Schwartz, 2003; van den Top et al., 2004). ARC neurons expressing α-MSH and AgRP innervate the melanin-concentrating hormone (MCH) and orexin (ORX) neurons in the perifornical region and lateral hypothalamic area (LHA) (Broberger et al., 1998; Elias et al., 1998), areas also important for the regulation of food intake and body weight (Anand and Brobeck, 1951). We recently demonstrated that ARC-specific knockdown of PKC-θ attenuated diet-induced obesity and improved insulin signaling (Benoit et al., 2009). Additionally, Ross and colleagues have implicated hypothalamic PKC-δ in mediating glucose production (Ross et al., 2008). Hence, activation of nPKCs may play a crucial role in regulating hypothalamic neuroendocrine functions.

While previous studies have described PKC-θ and -δ distribution in rodent brains, these studies were not focused on the hypothalamus. Here we systematically determined the distribution of PKC-θ and -δ immunoreactivity in the basal medial hypothalamus of rat and mouse brains and, we used double-label immunohistochemistry with green fluorescent protein (GFP) or enhanced yellow fluorescent protein (EYFP) neuronal-specific reporter mice and/or rats to investigate the neurochemical identity of PKC-θ and -δ hypothalamic neurons. Information on the distribution and co-localization with other key neuroendocrine cells will be valuable for identifying the role and function of PKC-θ and -δ in mediating metabolic functions.

EXPERIMENTAL PROCEDURES

Animals

Adult, male, Long-Evans rats (250–300 g, n=5; Harlan) and C57BL/6 mice (25–30 g, 14 wk old, n=5; Jackson Laboratories) were housed with ad libitum access to food and water in a light-(12 hours on/12 hours off) and temperature-(22–23 °C) controlled environment. The LepR-b-EGFP mice (n=3) were provided by Dr. Martin Myers Jr (University of Michigan, Ann Arbor, MI, USA); orexin-GFP mice (n=3) were provided by Dr. Masashi Yanagisawa (UT Southwestern Medical Center, Dallas, TX, USA); arginine vasopressin-enhanced green fluorescent protein (AVP-EGFP) transgenic rats (n=3) and oxytocin-EYFP (Oxt-EYFP) mice (n=3) were provided by Dr. Andrew Zinn (UT Southwestern Medical Center, Dallas, TX, USA). The animals and procedures used for these studies were in accordance with the UT Southwestern Institutional Animal Care and Use Committee.

Generation of transgenic mice

The generation of the LepRb-EGFP mice has been described previously (Leinninger et al., 2009). Briefly, the LepR-Cre mice were crossed with Gt(Rosa)-26-Sor^tm2Sho (Jackson Laboratory) and the double heterozygous progeny were intercrossed to generate LepR-Cre/Cre; Gt(Rosa)-26-Sor^{tm2Sho/tm2Sho} (i.e., LepR-b-EGFP) mice (Leshan et al., 2009). The generation of the ORX-GFP mouse has been described previously (Li et al., 2002). The Lac Z gene under control of a human hypocretin promoter was used to drive the GFP sequence (Sakurai et al., 1999). The generation of the PKC-δ KO mice has been described previously (Miyamoto et al., 2002). The target vector was constructed by replacing a 2.8-kb Hpal-BamH1 fragment containing exons I and II of the PKC-δ gene with a phosphoglycerate kinase (PGK)-neo-poly(A) cassette. Mutant embryonic stems cells were prepared and microinjected into C57BL/6 blastocysts, and the resulting male chimeras were mated with female C57BL/6 mice. Heterozygous breeding strategy was used to produce homozygous PKC-δ KO mice. The validity of the above three transgenic mouse models has been verified previously (Li et al., 2002; Miyamoto et al., 2002; Yamanaka et al., 2003; Humphries et al., 2006; Leinninger et al., 2009; Leshan et al., 2009). The generation and validation of the AVP-EGFP transgenic rats has been described previously (Ueta et al., 2005).

The Oxt-Cre transgenic mice were generated by using methods described previously (Lee et al., 2001; Balthasar et al., 2005). Briefly, Cre recombinase and a polyadenylation signal were targeted to the initiator methionine codon of the Oxt gene using recombineering techniques with bacterial artificial chromosome (RP24-388N9) (Lee et al., 2001; Balthasar et al., 2005). A fragment of genomic DNA containing the targeted Oxt gene plus 10 kb of upstream flanking sequence (Fig. 1A) and the downstream vasopressin gene lacking its initiator codon and upstream sequences was subcloned (Fig. 1B). The resulting construct was injected into pronuclei of fertilized mouse oocytes, and a transgenic line expressing Cre recombinase in the hypothalamus was obtained. Finally, this line was crossed with Rosa26-EYFP reporter mice (Jackson Laboratory, stock number 006148) to generate the Oxt-EYFP mice. We confirmed the validity of this mouse model by carrying out double label immunofluorescence for oxytocin (Fig. 1C) and EYFP (Fig. 1D). Majority of cells that expressed EYFP were also positive for oxytocin immunoreactivity (Fig. 1E). Few oxytocin immunoreactive neurons showed relatively little EYFP expression suggesting that not every oxytocin neuron expresses detectable EYFP (Fig. 1E).

Fig. 1.

Generation of the Oxt-Cre mice. (A) Map of mouse oxytocin/vasopressin locus. (B) Transgene with Cre recombinase inserted into Oxt gene and vasopressin ATG codon mutated. Double label IHC for: oxytocin neurons (C) and EYFP (D) showed a high degree of colocalization (E). Colocalized neurons are indicated with white arrows. Scale bar=200 μm for (C–E). IGR, intergenic region.

Histology

Rats and mice were deeply anesthetized with an intraperitoneal injection of chloral hydrate (7%; 350 mg/kg; Sigma, St. Louis, MO, USA) and perfused transcardially with DEPC-treated 0.9% saline followed by 4% paraformaldehyde or 10% neutral buffered formalin. Brains were removed, post-fixed in 10% formalin for 2 h and then transferred to 20% sucrose made in DEPC-treated phosphate-buffered saline (PBS), pH 7.0 at 4 °C. Brains were cut coronally at 40 μm (for rats) or 25 μm (for mice) into five equal series on a freezing microtome. Tissue was stored at −20 °C in antifreeze solution until processed for immunohistochemistry.

Preparation and specificity of primary antibodies

The PKC-θ antiserum was purchased from Santa Cruz Biotechnology (catalog no. sc-212, lot no. L2408, Santa Cruz, CA, USA). The PKC-θ antiserum is a rabbit polyclonal affinity purified antibody raised against a peptide mapping at the C-terminus of the PKC-θ of mouse origin. Previously, we confirmed the specificity of the antibody in tissue from PKC-θ KO mice (Benoit et al., 2009). We did not observe any staining in the hypothalamic sections of PKC-θ KO mice.

The PKC-δ antiserum was provided by Dr. Todd Sacktor, SUNY Downstate Medical Center, NY, USA. The preparation of the rabbit polyclonal antibody has been described previously (Naik et al., 2000). The antiserum was prepared against a synthetic peptide consisting of a NH₂-terminal cysteine followed by the sequence of the carboxyl-terminal variable region for PKC-δ: NPKYEQFLE. No staining was observed when the primary antibody was preincubated with 2 μg/mL of the immunizing peptide. We further confirmed the specificity of this antiserum in PKC-δ KO mice.

The guinea pig anti-histidine decarboxylase antibody was purchased from Alpco Diagnostics (catalog no. B-GP- 265-1, lot no. HS-4061, Salem, NH, USA) (Zhao et al., 2008a). The chicken anti-GFP antibody was purchased from Aves Lab Inc. (catalog no. GFP-1020, lot no. 1223FP03, Aves Labs Inc., Tigard, OR, USA) and its specificity has been previously confirmed (Zhao et al., 2008b). Chickens were immunized with fluorescent GFP, eggs were collected from hens after multiple injections and IgY fractions were prepared from the yolks. The manufacturer has confirmed the specificity of the antibody by western blot analysis and immunohistochemistry in transgenic mice expressing the GFP product.

Immunohistochemistry (IHC)

Rats

For each animal, a series of sections extending from the anterior olfactory nucleus to the caudal end of the ARC was subjected to a standard immunoperoxidase reaction. Sections were pretreated with 0.3% hydrogen peroxide, after which they were blocked with 3% normal donkey serum (Jackson Laboratories, West Grove, PA, USA) and 0.3% Triton X-100 (Sigma, St. Louis, MO, USA). Subsequently, sections were incubated in 1:2000 PKC-θ (catalog no. sc-212, Santa Cruz Biotechnology, CA, USA) or 1:2000 PKC-δ (Naik et al., 2000) rabbit primary anti sera overnight at room temperature. Sections were then incubated for 1 h in biotin-conjugated IgG donkey anti-rabbit (1:1000, Jackson Laboratories) and for 1 h in avidin–biotin complex (1:500, Vector Labs, Burlingame, USA). The sections were then submitted to a 4–6 min immunoperoxidase reaction with 0.03% hydrogen peroxide using diaminobenzidine tetrahydrochloride (DAB; catalog no. D5905; Sigma) as the chromogen. The reaction was terminated with rinses in PBS. Sections were mounted onto gelatin-coated slides, dehydrated, delipidated and coverslipped with DPX (Fluka; Sigma-Aldrich Chemie GmbH, Deisenhofen, Germany).

An adjacent series of sections was thionin stained to identify PKC-positive nuclear cell bodies. Briefly, the sections were mounted on gelatin-coated slides and dried overnight at room temperature. On the following day, sections were dehydrated in ethyl alcohol and cleared in xylene for 30 min. The slides were gradually rehydrated in ethanol solution and immersed in a 0.25% thionin solution for 1 min. The sections were rinsed in water, dehydrated in graded ethanols and dipped in xylene, and then slides were cover-slipped with DPX (Fluka; Sigma-Aldrich).

Mice

Both, single- and double-label immunofluorescence were performed. First, to identify brain nuclei expressing either PKC-θ or PKC-δ, we performed single-label IHC. Adjacent coronal series of sections extending from the anterior olfactory nucleus to the caudal end of the NTS were pretreated with hydrogen peroxide, incubated for 30 min in 3% normal donkey serum (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) with 0.25% Triton X-100 in PBS (PBT) with 0.02% sodium azide. Sections were incubated overnight at room temperature in 1:200 PKC-θ (catalog no. sc-212, Santa Cruz Biotechnology, CA, USA) or 1:500 PKC-δ rabbit primary antisera (Naik et al., 2000). After washing in PBS, tissue was incubated in fluorescently labeled Alexa 594 or Alexa 488 donkey secondary antisera (Invitrogen, Carlsbad, CA, USA) diluted to 1:200 in PBT for 1 h at room temperature. Tissues were then rinsed in PBS and mounted on gelatin coated slides. Dried slides were cover-slipped using Fluoromount G (catalog no. 17984-25, Electron Microscopy Sciences, Hatfield, PA, USA).

For double label IHC for EYFP and Oxt labeling, coronal brain sections were permeabilized and blocked in 3% normal goat serum/0.3% Triton X-100 for 1 h, and incubated for 48 h at 4 °C in 1:5000 rabbit anti-GFP antiserum (catalog no. A6455, Molecular Probes, Eugene, OR, USA) and 1:5000 mouse anti-Oxt antiserum (catalog no. MAB5296, Millipore Corp., Billerica, MA, USA). Sections were then incubated with fluorescein isothiocyanate goat antirabbit IgG secondary antiserum and 1:400 Cy-2 affiniPure goat antimouse IgG secondary antiserum (catalog no.115-225-166, Jackson ImmunoResearch) for 2 h at room temperature. Sections were placed in 0.2 μg/mL 4′,6-diamidino-2-phenylindole (catalog no. 236276; Roche, Indianapolis, IN, USA) for 10 min and then mounted on plus coated slides and coverslipped using Vectashield (catalog no. H-1000; Vector Laboratories, Burlingame, CA, USA). Images of sections containing paraventricular hypothalamic nucleus (PVN) were captured on an Olympus BX61 microscope using Cytovision software (Applied Imaging Corp., San Jose, CA, USA).

The sections from LepR-b-EGFP, C57BL/6 mice (n=3) and ORX-GFP (n=3) were incubated in PKC-θ antisera. Sections from oxytocin-EYFP mice (n=3) and AVP-EGFP rats (n=3) were incubated in PKC-δ antisera and then processed as described above. Sections from LepR-b-EGFP, ORX-GFP, oxytocin-EYFP mice and AVP-EGFP rats were washed in PBS, and then incubated in 1:5000 chicken, anti-GFP antibody (catalog no. GFP-1020, Aves Labs Inc., Tigard, OR, USA).

For PKC-θ/GFP and PKC-δ/YFP double-label IHC, sections were then washed in PBS, and the tissue was incubated in fluorescently labeled Alexa 488 donkey secondary antisera (Invitrogen) diluted to 1:200 in PBT for 1 h at room temperature. For the PKC-θ/histidine decarboxylase (HDC) double-label IHC brain sections were incubated in 3% normal goat serum/0.3% Triton X-100 for 1 h, and incubated for 48 h at 4 °C in 1:5000 anti-HDC antiserum (ALPCO Diagnostics) and 1:200 anti-PKC-θ antiserum. Sections were then incubated Alexa 488 donkey anti-rabbit IgG and Alexa 594 goat anti-guinea pig IgG secondary antisera (invtrogen), diluted 1:200 respectively, for 2 h at room temperature. Tissues were then rinsed in PBS and mounted on gelatin coated slides. Dried slides were cover-slipped using Fluoromount G (catalog no. 17984-25, Electron Microscopy Sciences, Hatfield, PA, USA).

Production of photomicrographs

Sections were analyzed using a Zeiss Axioplan light microscope and Zeiss Apotome microscope. Adobe Photoshop CS2 9.0 (Adobe Systems, San Jose, CA, USA) was used to adjust only sharpness, brightness and contrast as well as to combine select images into plates. Unless otherwise noted, nomenclature corresponds to the Paxinos and Franklin (2001, 2nd Edition) mouse brain atlas.

Cell counts and data analysis

Single and double-labeled cells were counted in brain regions demonstrating colocalization of PKC-θ and orexin. PKC-θ and orexin cells were counted in the LHA, perifornical area (Pfx) and ventral part of dorsomedial hypothalamic nucleus DMN (vDMN). Cells were counted from the same atlas level for all three nuclei (Paxinos and Franklin, 2001, 2nd Edition). Similarly, PKC-δ and oxytocin cells were counted in the PVN and supraoptic nucleus (SON) from the same atlas level. Cells were counted from three separate mice (n=3) and counts are presented as mean±SEM. These data were corrected for double counting by applying Abercombie’s formula, N=n(T/T+D), where N=corrected cell count, n=observed number of cells, T=section thickness (25 μm) and D=diameter of nucleus (Guillery and Herrup, 1997). For each brain area with double-labeled cells we calculated the mean nuclear diameter from 10–12 random cells. We did not use stereological techniques for cell count estimation.

RESULTS

Distribution of PKC-θ immunoreactivity in rat brain

PKC-θ immunoreactive (IR) neurons were observed in discrete regions of the CNS. In the thalamus (Fig. 2A, B), the highest concentration of PKC-θ-IR cells was in the medial habenular nucleus (MHb), and a few PKC-θ-IR cells were observed ventral to the MHb in the paraventricular thalamic nucleus (PVT) and lateral habenular nucleus (LHb). In the hypothalamus, the highest concentration of PKC-θ-IR neurons were at the tuberal level dorsal to the fornix in a region that comprises the LHA and zona incerta (Fig. 2C, D). A few PKC-θ-IR neurons were observed in the vDMN (Fig. 2E). We confirmed our previous findings (Benoit et al., 2009) demonstrating PKC-θ-IR neurons in the tuberal level of the ARC (Fig. 3A), and also found cells in the posterior level of the ARC (Fig. 3B).

Fig. 2.

Distribution of PKC-θ-IR neurons in rat brain. (A) Thionin-stained section denoting the thalamic nuclei (MHb and PVT) of the brain. (B) Adjacent section of the brain shown in (A), showing PKC-θ-IR neurons in the MHb and PVT. (C) Low-power magnification of a thionin-stained section of hypothalamic nuclei. (D) PKC-θ-IR neurons in the LHA and Pfx. (E) PKC-θ-IR neurons in the DMN. 3v, third ventricle; ARC, arcuate hypothalamic nucleus; D3v, dorsal third ventricle; DMN, dorsomedial hypothalamic nucleus; fx, fornix; LHA, lateral hypothalamic area; MHb, medial habenular nucleus; mt, mammillothalamic tract; Pfx, perifornical area; PVT, paraventricular thalamic nucleus. Scale bars=200 μm for (A); 50 μm for (B, D, E); 100 μm for (C).

Fig. 3.

Distribution of PKC-θ-IR neurons in the rat hypothalamic arcuate nucleus (ARC). (A) PKC-θ-IR neurons in the ARC. (B) PKC-θ-IR neurons in the DTM and in ventral aspects of the PMV. 3v, third ventricle; DTM, dorsal tuberomammillary nucleus; fx, fornix; PMV, ventral premammillary nucleus. Scale bars=100 μm for (A, B).

Distribution of PKC-δ immunoreactivity in rat brain

We confirmed the presence of PKC-δ-IR neurons in the thalamus, septal nuclei and hippocampus (Merchenthaler et al., 1993; Naik et al., 2000). Although, PKC-δ-IR neurons are found in several thalamic nuclei, the highest concentrations were in the ventral posterolateral (VPL) and ventral posteromedial (VPM) thalamic nuclei (Fig. 4A, B). PKC-δ-IR was also found in the hypothalamus where its distribution was limited to the PVN (Fig. 4C) and SON (Fig. 4D). Consistent with previous findings, our data suggest PKC-δ immunoreactivity is primarily neuronal (Merchenthaler et al., 1993; Naik et al., 2000).

Fig. 4.

Distribution of PKC-δ-IR neurons in rat brain. (A) Thionin-stained section denoting the thalamic nuclei (VPM and VPL) of the brain. (B) Adjacent section of the brain shown in (A) at higher magnification, showing PKC-δ-IR neurons in the VPM and VPL. (C) PKC-δ-IR neurons in the PVN. (D) PKC-δ-IR neurons in the SON. 3v, third ventricle; fx, fornix; ic, internal capsule; LV, lateral ventricle; PVN, paraventricular hypothalamic nucleus; SON, supraoptic nucleus; VPL, ventral posterolateral thalamic nucleus; VPM, ventral posteromedial thalamic nucleus. Scale bars=200 μm for (A); 200 μm for (B); 50 μm for (C, D).

Distribution of PKC-θ immunoreactivity in mouse brain

Fluorescent IHC demonstrated a similar distribution of PKC-θ-IR neurons in the mouse brain (Table 1). High concentrations of PKC-θ-IR neurons in the mouse were observed in the MHb, LHA, Pfx and tuberomammillary nucleus. Similar to the rat, PKC-θ-IR neurons were also observed in the ARC and vDMN.

Table 1.

Distribution pattern of PKC-theta versus PKC-delta in the mouse brain

Brain area	PKC-θ-IR cells	PKC-δ-IR cells	PKC-θ-IR fibers	PKC-δ-IR fibers
Cerebral cortex
Piriform cortex	−	−	+	−
Primary somatosensory cortex, forelimb region	−	−	+	−
Primary somatosensory cortex, hindlimb region	−	−	+	−
Amygdala
Anterior amygdaloid area, dorsal	−	+++	−	−
Anterior amygdaloid area, ventral	−	++	−	−
Hippocampus and septum
Dentate gyrus	−	+	−	−
CA1 field	−	+	−	−
CA3 field	−	++	−	−
Lateral septal nucleus	−	++	+	++
Thalamus
Paraventricular thalamic nucleus	+	−	++	−
Medial habenular nucleus	+++	−	−	−
Lateral habenular nucleus	+	−	+	−
Reticular thalamic nucleus	−	+++	−	−
Ventral posterolateral thalamic nucleus	−	+++	−	−
Ventral posteromedial thalamic nucleus	−	+++	−	−
Ventrolateral thalamic nucleus	−	++	−	−
Ventromedial thalamic nucleus	−	++	−	−
Hypothalamic nuclei
Medial preoptic nucleus	−	−	++	++
Medial forebrain bundle	−	−	++	++
Supraoptic nucleus	−	+++	−	−
Suprachiasmatic nucleus, dorsal	−	++	−	−
Anterior hypothalamic area	−	+	+++	−
Paraventricular hypothalamic nucleus	−	+++	+++	+
Perifornical area	+++	−	+++	−
Lateral hypothalamic area	+++	−	+++	−
Dorsomedial hypothalamic nucleus, ventral	++	−	++	−
Ventromedial hypothalamic nucleus	−	−	++	−
Median eminence	−	−	−	+++
Arcuate hypothalamic nucleus	++	−	++	−
Posterior hypothalamic area	−	−	++	−
Ventral premamillary nucleus	+	−	+	−
Ventral tuberomammillary nucleus	+	−	++	−
Dorsal tuberomammillary nucleus	++	−	−	−
Medial mamillary nucleus	−	−	++	−
Supramammillary nucleus	−	−	++	−
Midbrain, pons and medulla oblongata
Dorsal raphe	−	−	++	−
Interpenduncular nucleus	−	−	++	−
Periaqueductal gray matter	−	−	++	−
Substantia nigra, pars compacta	−	−	++	−
Ventral tegmental area	−	−	+++	−
Nucleus of the solitary tract	−	−	+	−

Qualitative estimates of number of labeled PKC-θ and -δ cells and immunoreactive fibers. The following four point density scale was used: +++, high density; ++, moderate density; +, low density; −, absent.

Fibers from PKC-θ-IR-positive neurons were found throughout the mouse brain (Table 1). A few immunoreactive fibers were observed in the piriform cortex and primary somatosensory cortex; moderate-density immunoreactive fibers were found in the medial preoptic nucleus, medial forebrain bundle, periventricular area, lateral septal nucleus, ventromedial hypothalamic nucleus (VMN), ARC, posterior hypothalamic area (PH), supramammillary nucleus (SuM), and medial mamillary nucleus (MM); and dense innervation was observed in the anterior hypothalamic area (AHA) and PVN. Moderate density of PKC-θ-IR fibers was observed in both the anterior and posterior portions of the PVT (Table 1).

In the midbrain, a high density of PKC-θ-IR fibers was observed in the ventral tegmental area (VTA), while a moderate density was observed in the interpenduncular nucleus, substantia nigra (SN), dorsal raphe (DR) and periaqueductal gray (PAG). In addition, we found a few fibers in the nucleus of the solitary tract (NTS), hypoglossal nucleus (12 N), lateral reticular nucleus (LRt), and the ventral part of medullary reticular nucleus (MdV).

Distribution of PKC-δ immunoreactivity in mouse brain

As indicated in Table 1, PKC-δ-IR neurons were observed in several thalamic nuclei, with the densest immunoreactivity in the reticular thalamic nucleus (Rt), VPM and VPL (Fig. 5A, B). No staining for PKC-δ was found in these thalamic nuclei in PKC-δ KO mice (Fig. 5C), thus validating the specificity of the antibody. Similar to the rat, there were moderate density PKC-δ-IR neurons in the LS and CA3 field, and few PKC-δ-IR neurons were observed in the dentate gyrus and CA1 field (Table 1).

Fig. 5.

Validation of PKC-δ antibody. (A) Low power magnification of PKC-δ fluorescent IHC in the thalamic nuclei (VPM and VPL) of the mouse brain. (B) Higher power view of boxed area in (A). (C) Lack of staining in section from same atlas level as (B) in PKC-δ-KO mouse. ic, internal capsule; Rt, reticular thalamic nucleus; VPL, ventral posterolateral thalamic nucleus; VPM, ventral posteromedial thalamic nucleus. Scale bars=100 μm for (A); (B, C)=50 μm.

In the hypothalamus, there was a high concentration of PKC-δ-IR neurons in the PVN, SON and dorsal portion of the suprachiasmatic nucleus (SCh). We also observed a few PKC-δ-IR neurons in the AHA region between the PVN and SON. In the amygdala, a higher concentration of PKC-δ-IR neurons was in the dorsal part of the anterior amydaloid area (AAD) than in the ventral portion.

In comparison to PKC-θ-IR fibers, the distribution of PKC-δ-IR fibers was very limited in the mouse brain. Specifically, PKC-δ-dense immunoreactive fibers were found in the ME and SCh, while moderate density fibers were observed in the periventricular area and LS (Table 1).

Neurochemical characterization of hypothalamic PKC-θ expressing cells

We also determined the neurochemical identity of the hypothalamic PKC-θ-IR neurons. We had previously demonstrated 20% colocalization of PKC-θ with ARC NPY neurons (but not POMC neurons) and ~22% colocalization of PKC-θ with LepR-b in the ARC (Benoit et al., 2009). LepR-b neurons have also been demonstrated in the Pfx and LHA of mouse brain (Leinninger et al., 2009; Scott et al., 2009), and, we found no colocalization of GFP-IR and PKC-θ-IR in the Pfx or LHA (Fig. 6A–C) of LepRb-EGFP mice.

Fig. 6.

Double-label IHC for PKC-θ-IR neurons (A, denoted by white arrow heads) and LepR-b-EGFP in mouse hypothalamus (B, denoted by white lines) showed no colocalization (C) in the Pfx. fx, fornix; Pfx, perifornical area. Scale bar=50 μm for (A–C).

Orexin neurons are exclusively found in the Pfx and LHA of the mouse brain and have no colocalization with LepR-b neurons in the LHA (Li et al., 2002; Leinninger et al., 2009). Double-label IHC for ORX-GFP and PKC-θ-IR (Fig. 7A–C) revealed a high percentage of colocalization in the Pfx (86.7±1.9% of total PKC-θ-IR neurons) and LHA (82.8±8.6% of total PKC-θ-IR neurons) (Table 2). PKC-θ immunoreactive fibers were also observed in the PVN and VTA (Table 1), which are also known to receive dense innervation from ORX fibers (Fig. 7D–I) (Date et al., 1999; Nambu et al., 1999). We also observed a high degree of PKC-θ and ORX-immunoreactive fibers in the PVN (Fig. 7D–F) and VTA (Fig. 7G–I). Apparent appositions of PKC-θ and ORX immunoreactive fibers were observed in both the PVN (Fig. 7F) and VTA (Fig. 7I); however, overlapping fiber distribution was only observed in the PVN (Fig. 7F). We also observed apparent innervations of histamine neurons by PKC-θ immunoreactive fibers in the ventral tuberomammillary nucleus (VTM) (Fig. 8A–C).

Fig. 7.

Double label IHC for: PKC-θ-IR neurons (A) and ORX-GFP (B) in mouse hypothalamus showed a high degree of colocalization (C). Colocalized neurons are indicated with white arrows. High density of PKC-θ-IR (D) and ORX-GFP-IR fibers (E) in the PVN. Overlapping PKC-θ-IR and ORX-GFP-IR fibers (F, denoted by white arrows). High density of PKC-θ-IR (G) and ORX-GFP-IR fibers (H) in the VTA. No overlapping distribution pattern of PKC-θ-IR (denoted by white arrowheads) and ORX-GFP-IR fibers were observed (I). The fibers lay adjacent to each other and made several appositions. Points where appositions are made have been indicated by *. 3v, third ventricle; IPF, interpeduncular fossa; LHA, lateral hypothalamic area; MM, medial mammillary nucleus; Pfx, perifornical area; PVN, paraventricular hypothalamic nucleus; VTA, ventral tegmental area. Scale bars=50 μm for (A–C); 100 μm for (D–F); 100 μm for (G–I).

Table 2.

Double-labeled (PKC-θ+Orexin) neurons in mouse hypothalamic nuclei

Regions	Total PKC-θ	Total orexin	Doubles	% Doubles/total PKC-θ	% Doubles/total orexin
LHA	56±5.69	52±5.86	45.7±4.17	82.8±8.57	88.4±3.53
vDMN	11.3±0.88	9.33±1.15	9±0.58	84.2±11.6	96.7±3.33
Pfx	15.7±2.33	13.7±2.19	13.7±2.33	86.7±1.92	97±3.03

Values represent estimates of mean counts of cells±SEM (n=3). All cells were counted from atlas level 46 of Paxinos and Franklin (2001, 2nd Edition) Mouse Brain Atlas. LHA, lateral hypothalamic area; vDMN, ventral part of the dorsomedial hypothalamic nucleus.

Fig. 8.

Double label IHC for: PKC-θ-IR (A, denoted by white arrow heads) and histidine decarboxylase-IR (B, denoted by white lines) in mouse hypothalamus showed several appositions (C). Points where appositions are made have been indicated by *. Scale bar=100 μm.

Neurochemical characterization of hypothalamic PKC-δ expressing cells

Double-label IHC for GFP-like immunoreactivity (Fig. 9A, B) and PKC-δ-IR in Oxt-EYFP mice revealed a high degree of colocalization of PKC-δ-IR and GFP-IR in PVN (Fig. 9C–E) and SON (Fig. 9F–H) neurons. In the PVN, 83.4±2.9% of PKC-δ-IR neurons were colocalized with Oxt neurons while in the SON 84.5±5.3% had colocalization (Table 3). Although, results from double label IHC for PKC-δ and YFP immunoreactivity in oxytocin-EYFP mice (Fig. 1C–E) might not necessarily represent the accurate number of colocalized neurons we next determined whether the ~15% non-oxytocin-PKC-δ neurons are AVP neurons. We did not observe any colocalization of PKC-δ-IR neurons with AVP neurons in either the PVN (Fig. 10A–C) or the SCh (Fig. 10D–F) suggesting that PKC-δ neurons in these two regions of the mouse hypothalamus are mainly oxytocin neurons.

Fig. 9.

Colocalization between Oxt-YFP-IR and PKC-δ-IR neurons in the mouse hypothalamus. Lower power magnification of Oxt-YFP-IR neurons in PVN (A) and SON (B). Double-label IHC for PKC-δ-IR (C) and Oxt-YFP-IR (D) in boxed area of (A) showed a high degree of colocalization in the PVN (E). Double-label IHC for PKC-δ-IR (F) and Oxt-YFP-IR (G) showed a moderate degree of colocalization in the SON (H). Examples of colocalized neurons are indicated by white arrows. 3v, third ventricle; PVN, paraventricular hypothalamic nucleus; SON, supraoptic nucleus. Scale bars=100 μm for (A–G).

Table 3.

Double-labeled (PKC-δ+Oxytocin) neurons in mouse hypothalamic nuclei

Regions	Total PKC-δ	Total oxytocin	Doubles	% Doubles/total PKC-δ	% Doubles/total oxytocin
PVN	51.7±5.24	66±5.13	43.3±5.61	83.4±2.87	65.3±4.43
SON	20.67±1.2	34.7±2.91	17.3±0.33	84.5±5.28	50.7±4.24

Values represent estimates of mean counts of cells±SEM (n=3). All cells were counted from atlas level 38 of Paxinos and Franklin (2001, 2nd Edition) Mouse Brain Atlas. PVN, paraventricular hypothalamic nucleus; SON, supraoptic nucleus.

Fig. 10.

Colocalization between PKC-δ-IR and AVP-IR in mouse hypothalamus. Double-label IHC for PKC-δ-IR neurons (A, denoted by white arrow heads) and AVP-EGFP in mouse hypothalamus (B, denoted by white lines) showed no colocalization (C) in the PVN. There was no PKC-δ-IR at this level of the SCh (D), only AVP-EGFP was observed (E) and hence no colocalization was observed (F). 3v, third ventricle; PVN, paraventricular hypothalamic nucleus; SCh, suprachiasmatic nucleus. Scale bars=100 μm for (A–C); 50 μm for (D–F).

DISCUSSION

We have characterized the distribution and neurochemical identity of PKC-θ and PKC-δ neurons in the CNS, focusing on the hypothalamus. Our findings suggest that PKC-θ and PKC-δ neurons have distinct distribution patterns in the rat and mouse brain. We replicated previous findings demonstrating the presence of PKC-θ immunoreactivity in the ARC of the mouse and rat (Benoit et al., 2009) and PKC-δ immunoreactivity in the septal nuclei, thalamus and hippocampus of the rat brain (Merchenthaler et al., 1993; Naik et al., 2000). Additionally, we have demonstrated the presence of PKC-θ immunoreactivity in the LHA and tuberomamillary nucleus of the hypothalamus, whereas PKC-δ immunoreactivity was found predominately in the PVN and SON. Finally, we found that PKC-θ is highly colocalized with orexin neurons in the Pfx and LHA, while PKC-δ is highly colocalized with oxytocin neurons in the PVN and SON.

Comparison of PKC-θ versus PKC-δ distribution

As indicated in Table 1, the distribution patterns of cell bodies containing PKC-θ and PKC-δ were distinct. In the thalamus, PKC-θ immunoreactive neurons were mainly restricted to the MHb and PVT, while PKC-δ cell bodies were extensively distributed in several thalamic nuclei with high concentrations in the VPL and VPM. In opposition, we observed a high number of PKC-θ immunoreactive neurons in the perifornical area, LHA, and tuberomamillary nucleus of the hypothalamus, while PKC-δ immunoreactive neurons were mainly limited to the PVN and SON.

Functional implications: PKC-θ in orexin neurons

For the first time, we have demonstrated a high degree of colocalization of PKC-θ in orexin neurons, suggesting that PKC-θ may be an important downstream mediator of orexin activity, which has been implicated in feeding and arousal (Sakurai et al., 1998; Yamanaka et al., 2000; Tsujino and Sakurai, 2009).

Feeding

A critical role has been identified for LHA orexin neurons in feeding and energy homeostasis (Sakurai et al., 1998; Yamada et al., 2000; Yamanaka et al., 2000). Central administration of an anti-orexin antibody or orexin receptor 1 (OX1R)-selective antagonist reduces food intake in rats (Haynes et al., 2000; Yamada et al., 2000), and prepro-orexin KO mice and orexin KO mice are hypophagic compared to WT mice (Hara et al., 2001; Willie et al., 2001). Orexin neurons receive innervation from NPY/AgRP and α-MSH immunoreactive fibers (Broberger et al., 1998; Elias et al., 1998). Central administration of AgRP results in increased c-Fos expression in orexin but not MCH neurons of the lateral hypothalamus (Zheng et al., 2002). These findings collectively suggest that orexin has an important role in the regulation of feeding behavior. Lastly, the majority of orexin neurons are glutamatergic (Rosin et al., 2003; Torrealba et al., 2003) and not GABAergic (Rosin et al., 2003) although we did not assess these transmitters in this analysis.

Activation of PKCs has been demonstrated to cause desensitization of receptors (Huganir and Greengard, 1990), glutamate-induced neurotoxicity (Felipo et al., 1993), and neuropeptide release (Liu et al., 1992). PKC activation is downstream of G-protein coupled receptors (GPCRs), and orexin neurons may express GPCRs involved in feeding behavior, including Y4 receptors (Campbell et al., 2003), Y5 receptors (Gerald et al., 1996), melanocortin-4-receptors (MC4R) (Huszar et al., 1997), and orexin autoreceptors (Shiraishi et al., 2000). Hence, we hypothesize that activation of GPCR, by their respective agonists such as pancreatic polypeptide, NPY, AgRP or orexin, leads to coupling with excitatory G protein G_q/11. This activation presumably leads to agonist-induced hydrolysis of inositol phospholipids by phospholipase C (PLC) and results in an increase of DAG; and potentially activation of PKC-θ resulting in translocation to neuronal membranes where PKC-θ causes phosphorylation of calcium (Ca²⁺) channels, leading to influx of Ca²⁺, membrane depolarization, and potentially release of orexin, dynorphin and/or glutamate.

Other mechanisms by which PKC-θ may regulate orexin and glutamate release include phosphorylation-dependent inhibition of K⁺ channels (Hermans and Challiss, 2001) and phosphorylation of a conserved threonine residue (T180) of the Kir6.2 subunit of the potassium dependent ATP (K_ATP) channel (Light et al., 2000), therefore increasing the effects of orexin. Lastly, orexin released from LHA orexin neurons may control feeding by acting on ARC NPY/AgRP (Horvath et al., 1999; Yamanaka et al., 2000), VMN and/or PVN neurons (Marcus et al., 2001). Here we observed moderate innervation of the ARC and VMN by PKC-θ immunoreactive fibers, whereas the PVN had dense innervation of PKC-θ immunoreactive fibers, some of which are colocalized with orexin fibers. These data further suggest a potential role for PKC-θ in modulating the hypothalamic NPY/AgRP-orexin feeding circuit.

Arousal

Orexin also regulates sleep/wake cycles (Chemelli et al., 1999; Lin et al., 1999). The tuberomammillary nucleus contains a high percentage of histaminergic neurons, heavily expresses orexin receptor 2 (OX2R) mRNA (Marcus et al., 2001), and receives dense orexin innervation (Chemelli et al., 1999; Lin et al., 1999; Torrealba et al., 2003). In addition, at least 90% of orexin-IR terminals in the tuberomammillary nucleus are glutamatergic (Torrealba et al., 2003). Importantly, we found PKC-θ immunoreactivity in areas surrounding the dorsal tuberomammillary nucleus and apparent innervations of VTM histaminergic neurons by PKC-θ immunoreactive fibers. One model in which PKC-θ may modulate histamine release is through a G_q/11-PKC-θ mechanism via activation of Ca²⁺ channels. Another mechanism by which PKC-θ may modulate tuberomammillary nucleus neuronal function is through glutamate activation of group I metabotropic glutamate receptors (mGluIR) which are coupled with G_q/11 proteins. PKC-θ activation of mGluIR (Hermans and Challiss, 2001) may result in inhibition of K⁺ or activation of Ca²⁺ channels leading to membrane depolarization and histamine release. Prolonged or repeated exposure of mGluIR to glutamate may result in desensitization of the receptor by phosphorylation of mGluIR by PKC-θ and hence prevent glutamate excitotoxicity in these neurons (Alaluf et al., 1995). Therefore, orexin and glutamatergic innervation to the tuberomammillary nucleus histaminergic neurons (presumably modulated by PKC-θ) may regulate feeding-related excitatory input neurons and may promote arousal in anticipation of feeding time.

Functional implications: PKC-θ in non-orexin neurons

We also observed a high concentration of PKC-θ-IR neurons in the medial habenular nucleus. The MHb has been implicated in regulating various behaviors including sleep and circadian rhythms (Haun et al., 1992; Zhao and Rusak, 2005). Importantly, the MHb contains high levels of GABA_B receptors and MHb neurons are inhibited by GABA (Wang et al., 2006; Kim and Chung, 2007). The MHb innervates the interpenduncular nucleus (IPN), and these projections are known to be glutamatergic (Qin and Luo, 2009). Our results indicate a moderate distribution of PKC-θ-IR fibers in the IPN and surrounding midbrain nuclei. Therefore, one possibility is that PKC-θ may modulate glutamate release from MHb neurons.

We also observed moderate density PKC-θ-IR fibers in the VTA. These immunoreactive fibers were in close proximity with the orexin immunoreactive fibers. However, we found no colocalization with orexin fibers, suggesting that PKC-θ-IR fibers in the VTA may have originated from non-orexin neurons, such as neuronal populations in the MHb or tuberomammillary nucleus.

Functional implications: PKC-δ in oxytocin neurons

Oxytocin is involved in lactation, feeding and sexual behavior (Gimpl and Fahrenholz, 2001). In the brain, oxytocin is mainly synthesized in the magnocellular neurons of the PVN and SON (Silverman and Zimmerman, 1983). The bulk of the peptide is transported to the posterior pituitary via the internal zone of the median eminence (Wiegand and Price, 1980; Smith and Armstrong, 1990). In contrast, oxytocin neurons in the parvocellular subdivision of the PVN project to other brain regions such as the NTS and subfornical organ (Sawchenko and Swanson, 1982). Importantly, our findings indicate a high degree of colocalization of PKC-δ immunoreactivity in the PVN and SON oxytocin neurons, suggesting that PKC-δ may be an important modulator of oxytocin release. Furthermore, we also observed a dense innervation of PKC-δ fibers in the median eminence. No immunoreactive fibers were observed in the rest of the brain, with the exception of the lateral septal nucleus and suprachiasmatic nucleus. These results suggest that PKC-δ may be involved in modulation of oxytocin release to the posterior pituitary but not to other brain nuclei.

Hypothalamic oxytocin neurons receive dense glutamatergic innervation (van den Pol et al., 1990) and glutamate can regulate hypothalamic oxytocin release through activation of mGluIR (Pampillo et al., 2001). Importantly, PVN and SON magnocellular neurons have been demonstrated to express mGluIR (Meeker et al., 1994), and activation of mGluIR has been demonstrated to reduce K⁺ currents in SON magnocellular neurons (Schrader and Tasker, 1997). Furthermore, activation of mGluIR participates in the stimulatory effects of glutamate on hypothalamic oxytocin release (Pampillo et al., 2001). As previously discussed, we hypothesize PKC-δ may modulate glutamate-dependent oxytocin release in a similar fashion, through mGluIR-G_q/11 dependent K⁺ or Ca²⁺ channel phosphorylation. Lastly, oxytocin neurons also produce several other neuropeptides such as galanin (Skofitsch and Jacobowitz, 1986), and it is possible PKC-δ may regulate the release of these peptides as well.

CONCLUSION

Here we demonstrate PKC-θ and -δ neuronal distribution and their projections in the mouse and rat hypothalamus. We confirmed the presence of PKC-δ in several non-hypothalamic sites (Merchenthaler et al., 1993; Naik et al., 2000) and PKC-θ immunoreactivity in the ARC (Dewing et al., 2008; Benoit et al., 2009). We further found PKC-θ in hypothalamic orexin neurons and histamine-containing tuberomammillary neurons, suggesting that PKC-θ may be an important modulator of feeding behavior in these neuronal populations. Additionally, the distribution of PKC-δ in PVN and SON oxytocin neurons suggests a role for PKC-δ in regulating feeding, lactation and/or other neuroendocrine functions. Importantly, even though these novel PKC isoforms are found in different hypothalamic neuronal populations, they may have a common mechanism by which they modulate neuropeptide release. Future research will be required to determine the specific contribution of these two PKC isoforms in regulating hypothalamic neuroendocrine functions.

Abbreviations

α-MSH

α-melanocyte stimulating hormone

ARC

arcuate hypothalamic nucleus

AVP

arginine vasopressin

DAG

diacylglycerol

EYFP

enhanced yellow fluorescent protein

GFP

green fluorescent protein

IPF

interpeduncular fossa

LepR-b

long form of leptin receptor

LHA

lateral hypothalamic area

MHb

medial habenular nucleus

MM

medial mammillary nucleus

ORX

orexin

Oxt

oxytocin

Pfx

perifornical area

PKC

protein kinase C

PKC-δ

PKC-delta

PKC-θ

PKC-theta

POMC

proopiomelanocortin

PVN

paraventricular hypothalamic nucleus

PVT

paraventricular thalamic nucleus

SON

supraoptic nucleus

vDMN

ventral part of the dorsomedial hypothalamic nucleus

VMN

ventromedial hypothalamic nucleus

VPL

ventral posterolateral thalamic nuclei

VPM

ventral posteromedial thalamic nuclei

VTA

ventral tegmental area

3v

third ventricle