Skip to main content
NCBI home page
Journal of Anatomy logoLink to Journal of Anatomy
. 2021 Jun 20;239(5):1196–1206. doi: 10.1111/joa.13491

Localization of PIP5Kγ selectively in proprioceptive peripheral fields and also in sensory ganglionic satellite cells as well as neuronal cell membranes and their central terminals

Surang Chomphoo 1,✉, Hiroyuki Sakagami 2, Hisatake Kondo 1,3, Wiphawi Hipkaeo 1
PMCID: PMC8546504  PMID: 34151437

Abstract

Based on a previous study by others reporting that PIP5Kγ (phosphatidylinositol 4‐phosphate 5‐kinase γ) and its product, phosphatidylinositol 4,5 bisphosphate (PIP2), are involved in the regulation of nociception, the present immunohistochemical study examined the localization of PIP5Kγ‐immunoreactivity in dorsal root ganglia (DRG) and their peripheral and central terminal fields. PIP5Kγ‐immunoreactivity was localized for the first time in the muscle spindles, in which it was found in I‐bands of polar regions of intrafusal muscle fibers and also in sensory nerve terminals abutting on equatorial regions of the muscle fibers. This finding indicates the involvement of PIP5Kγ in the proprioception and suggests somehow complicated mechanisms of its involvement because of its heterogeneous localization in intra‐I‐band structures. In DRG, on the other hand, PIP5Kγ‐immunoreactivity was shown to be localized heterogeneously, but not evenly, over apposed plasma membranes of both neurons and ganglionic satellite cells in immune electron microscopy. In addition, no peripheral nerve terminals of DRG showing its distinct immunoreactivity were found in most peripheral fields of nociception and any other sensory perception except for the proprioception through muscle spindles. In contrast, numerous central terminals of DRG in the spinal posterior horn were immunoreactive for it. This finding leads us to consider the possibility that the regulation by PIP5Kγ of nociception is dominantly exerted in DRG and sensory neural tracts central, rather than peripheral, to DRG.

Keywords: muscle spindle innervation, neuronal plasma membrane, PIP5Kγ, satellite cell

PIP5Kγ‐immunoreactivity was localized in I‐bands of polar regions of intrafusal muscle fibers and sensory nerve terminals abutting on equatorial regions of the muscle fibers, indicating the involvement of PIP5Kγ in the proprioception.

graphic file with name JOA-239-1196-g003.jpg

1. INTRODUCTION

Phosphatidylinositol 4‐phosphate 5‐kinase (PIP5K) catalyzes the production of phosphatidylinositol bisphosphate (PIP2) by phosphorylating the 5‐position on the inositol ring of phosphatidylinositol 4‐phosphate. This phospholipid PIP2 is generally known to function as a lipid anchor, a component involved in membrane trafficking and a regulator of intramembranous channels/transporters, and also to function in apoptosis and autophagy, as well as a precursor of such second messengers as diacylglycerol, inositol triphosphate and phosphatidylinositol 3,4,5‐triphosphate in the phosphoinositide (PI) signal (Balla et al., 2009; Delage et al., 2013; Doughman et al., 2003; Oude Weernink et al., 2004).

Among three isoforms, α, β, γ, of PIP5K so far identified and termed (Kanaho et al., 2007), PIP5Kγ is known to be expressed in the brain at the highest level (Akiba et al., 2002; Wenk et al., 2001), and it has recently been shown to be involved in the regulation of nociception in electrophysiology using mice with the global heterozygous knockout of its gene (Wright et al., 2014). Their study was based on immuno‐light microscopic findings that its immunoreactivity was localized in almost all sensory neuronal somata of spinal dorsal root ganglia (DRG), especially distinct in the cellular rims delineating the somata in wild mice. Subsequently, using two conditional gene knockout mice, the one selectively deleting Pip5Kγ in spinal DRG and the other deleting it in the brain as well as DRG, the same research group has shown that PIP5Kγ exerts its regulatory role on nociception widely in the central nervous system rather than the peripheral sensory neurons of DRG alone (Loo & Zylka, 2017).

Although their molecular biological findings themselves are clear, several morphological questions remain to be clarified for the validity of their interpretation: one of the questions comes from a lack of detailed localization of PIP5Kγ in DRG at ultrastructural levels. It is because the immunoreactive neuronal rims represent not only the plasma membrane itself and subjacent cytoplasm of neuronal somata but also those of satellite glial cells (SGCs) enclosing the neurons intimately and thinly. Although it is difficult to identify their thin enclosure in light microscopy, the SGCs have recently been shown to be an essential element in the sensory signaling pathway including pain (Hanani, 2005; Gu et al., 2010). It is thus critical to clarify at ultrastructural levels whether or not the distinct immunoreactivity appearing in forms of neuronal somatic rims is ascribed solely to the neuronal plasma membranes. Another question is whether or not this isozyme molecule is localized in peripheral sensory nerve terminals themselves and is involved in the start of nociception and/or any other sensory modalities. The other questions are whether or not such distinct localization of PIP5Kγ is the case also in cranial sensory ganglia.

In order to clarify these questions above, the present immunohistochemical study was undertaken to examine at light and electron microscopic levels, detailed localization of distinct immunoreactivity for PIP5Kγ in DRG at different levels of the body axis, in several cranial sensory ganglia such as the trigeminal and facial genu ganglia, and in various peripheral tissues rich in sensory innervation such as submucous, subepithelial, and dermal areas as well as skeletal muscles.

2. MATERIALS AND METHODS

2.1. Animals and sample collection

A total of ten male ICR 10 weeks postnatal mice were obtained from Nomura Siam International with licenses to produce “JCL‐brand” animals (RRID: MIG_5652875), and they were given ad libitum access to food and water and were kept under the condition of a 12‐hour dark/light cycle. All experiments were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals at Northeast Laboratory Animal Center, Khon Kaen University, Thailand. The study was reviewed and approved by the Animal Ethics Committee of Khon Kaen University based on the Ethical Principles and Guidelines for the Use of Animals by the National Research Council of Thailand with the ethics number IACUC‐KKU‐89/63.

Under anesthesia by peritoneal injection of pentobarbital sodium (60 mg/kg body weight), mice were transcardially perfused with 25 ml of 4% paraformaldehyde/0.1 M phosphate buffer (PB). DRG at cervical, thoracic, and lumbar levels and corresponding portions of the spinal cord as well as the extirpated trigeminal and facialis genus ganglia as representative cranial sensory ganglia were then extirpated. In addition, the following tissues were extirpated: the oral lip and various portions of the alimentary tract, larynx and trachea, renal pelvis and ureter, genital and digital skins, and several skeletal muscles such as the erector spine muscle, triceps brachii and gastrocnemius muscles, and extraocular muscles. All the specimens were then postfixed with the same fixative overnight for immunohistochemical analysis.

2.2. Immuno‐light microscopy

The extirpated specimens were immediately immersed in 4% paraformaldehyde/0.1 M PB. They were subsequently immersed into 30% sucrose/PB for cryoprotection. Sections of 20‐μm thickness were made on cryostat and mounted on glass slides and treated with 0.1% TritonX‐100/phosphate‐buffered saline (PBS) for 30 min at room temperature, then with 0.3% H2O2/methanol for 10 min, followed by 5% normal goat serum/PBS for 30 min. The sections were incubated overnight at room temperature with a rabbit antibody against mouse PIP5Kγ (1 μg/ml) (RRID: AB2819240) or, when necessary, with goat polyclonal to CGRP (calcitonin gene‐related peptide) antibody (Cat. #ab36001, RRID: AB725807, Abcam) at a concentration of 1:500. PIP5Kγ‐antibody was generated and its specificity was reported by one (HS) of the present authors (Hara et al., 2013), and it was the same as that employed in the previous study by Wright et al (2014) who were given it by one (HS) of the present authors. The sections were subsequently incubated for 1 h at room temperature with biotinylated anti‐rabbit IgG (Cat. #ab64256, RRID: AB_2661852, Abcam). They were then treated for the DAB (diaminobenzidine) reaction using an Elite ABC kit (Cat. #PK‐6100, RRID: AB_2336819, Vector Laboratories). In the antigen absorption test acting as a control for the immunohistochemistry, the antibody was preabsorbed with PIP5Kγ at 20 µg/ml, and cryostat tissue sections were incubated with the absorption solution for 1 h at room temperature and subsequently treated by the same procedure as the regular immunoreaction described above.

For double immunofluorescence light microscopy, some of the cryostat sections were incubated first with the PIP5Kγ‐antibody at the same concentration as used for single immunostaining described above, and then with goat polyclonal to CGRP (calcitonin gene‐related peptide) antibody (Cat. #ab36001, RRID:AB725807, Abcam) at a concentration of 1:500. The antigen‐antibody reaction sites were visualized with Alexa Fluor 488‐labeled anti‐rabbit IgG (Cat. #35552, Thermo Scientific) and Rhodamine labeled anti‐goat IgG (Cat. #31620, Thermo Scientific).

2.3. Immuno‐electron microscopy

Some of the sections immunostained with PIP5Kγ‐antibody on poly‐ l ‐lysine‐coated plastic slides were postfixed with 0.5% OsO4 for 30 min and embedded in Epon (Cat. #14120, Electron Microscopy Sciences) after en bloc staining with 1% uranyl acetate. Ultrathin sections were observed under a transmission electron microscope (JEM1010, JEOL).

3. RESULTS

3.1. Peripheral tissue fields rich in sensory innervation including muscle spindles

In immuno‐light microscopy, no nerve terminals distinctly immunoreactive for PIP5Kγ were found in the dermis of skins from the dorsal covering and sole of the extremities, the lip, the nose, and the perineum, in the submucosa of the larynx, trachea, esophagus, and all portions of the alimentary tract, or in the renal pelvis and ureter (data not shown).

In skeletal muscles, while a majority of thick regular muscle fibers were immunonegative, a few immunoreactive thin muscle fibers, about 0.9–1.2 μm in width at their equatorial regions, were randomly found among thick muscle fibers (Figure 1a). The occurrence of such immunoreactive thin muscle fibers was variable with the erector spinae muscle the highest, followed by the triceps brachii (Figure 1b) and gastrocnemius (Figure 1c) muscles as well as the extraocular muscles, but none of such muscle fibers were actually found in the tongue muscle, laryngeal muscle, and pharyngeal muscle.

FIGURE 1.

FIGURE 1

Open in a new tab

(a‐g) Light micrographs of strongly immunoreactive intrafusal (IF) and immunonegative extrafusal (EF) muscle fibers of the erector spinae muscle (a, d‐g), triceps brachii (b), and gastrocnemius (c) of mice. Areas corresponding to those indicated by a rectangle of broken lines in (a), a rectangle of solid lines in (d), and multiple thick arrows in (d) are shown at higher magnification in (f), respectively. Thick arrows in (d) and (f) indicate immunoreactive sensory (annulospiral) nerve terminals abutting on equatorial regions of IF muscles. The outlines of two adjacent IF muscles are indicated by white brackets (e). Note intensely immunoreactive striations in polar regions of IF muscles, running perpendicular to the longitudinal axis of IF muscles, taking a rather constant interval, and having various thicknesses with round shapes often at sarcolemmal sites and tapering ends along the course (*). Striations close to the sarcolemma may occasionally be fused with each other as indicated by (⋆). Some portions of sarcolemma are immunoreactive and appear as dense lines bridging adjacent striation ends (broken lines), while some other portions are immunonegative, resulting in appearance of immunonegative spaces intervening between any two adjacent striation ends and looking open into interstitial spaces (arrowheads). Figure g represents negative control by antigen absorption test. Bars represent 200 µm (a‐c), 50 µm (d, f‐g), 30 µm (e)

In polar regions of the thin muscle fibers, the intense immunoreaction appeared in forms of dark striations running transversely to the long axis of the fibers and arranged at a regular periodicity of intervals with thin immunonegative spaces intervening. No such immunoreactive striations were evident in equatorial regions of thin muscle fibers (Figure 1d). Some immunoreactive striations fully spanned the entire width of thin muscle fibers, while some others had tapering ends along their spanning course. The sarcolemmal ends of striations often took a round or dot shape. The immunoreactivity was also sporadically found along the sarcolemma in the forms of immunoreactive lines extending over several striations, where the ends of immunoreactive striations looked to be fused with the immunoreactive sarcolemma at right angles. At immunonegative portions of the sarcolemma, the immunoreactive striations ended freely, resulting in an appearance of immunonegative interval spaces between two adjacent striations looking open to the interstitial space. The immunoreactive striations might be occasionally crossed to each other close to the sarcolemma (Figure 1e).

On the other hand, in equatorial regions of thin muscle fibers without such intensely immunoreactive striations, nerve terminals immunoreactive for PIP5Kγ were found in the forms of thick threads entwining the thin muscle fibers (Figure 1d,f). In immuno‐electron microscopy, PIP5Kγ‐immunoreactive thin muscle fibers were loosely surrounded by thin fibroblastic capsular cell processes. Aggregations of nuclei were often seen in equatorial regions of the thin muscle fibers directly apposed to nerve terminals immunoreactive for PIP5Kγ (Figure 2a). The immunoreactive nerve terminals were rich in small mitochondria and they were directly apposed to the thin muscle fibers without membrane infoldings of muscle fibers or basal lamina intervening (Figure 2b). Immunoreactive materials were deposited on the plasma membranes and intracellular membranes of small vesicles. No specific intracellular structures immunoreactive for PIP5Kγ were found within the equatorial regions of the thin muscle fibers.

FIGURE 2.

FIGURE 2

Open in a new tab

(a‐b) Immuno‐electron micrographs of IF muscle equatorial regions indicated by arrows in (d) and (f) (a) and an area corresponding to that enclosed by a rectangle in (a) at higher magnification (b). Note sensory (annulospiral) nerve terminals (*) rich in mitochondria (m) abutting on IF muscles containing scanty myofibrils (mf) and voluminous nuclear bag (NB) with narrow, empty, and smooth (no infoldings) intercellular space intervening (arrows). Thin cytoplasmic process of capsular cell (Ca) loosely surrounds IF muscle fiber. N: nucleus. Bars represent 2 µm (a), 800 nm (b)

In contrast to their equatorial regions, in polar regions of the thin muscle fibers without apposition to immunoreactive nerve terminals, electron‐dense materials due to the immunoreaction were localized on most Z‐lines at the midlines of I‐bands, while it was not detected on some Z‐lines (Figure 3a‐d). Both immunoreactive and immunonegative Z‐lines might occur in contiguity and might be arranged next to each other on the same myofibrils (Figure 3b). No significant differences were found in their width between immunonegative and immunoreactive Z‐bands. The immunoreactive materials were also associated with some, but not all, of thin myofilament bundles, the sarcoplasmic reticulum, and triads in the same I‐band domains (Figure 3a‐d). Heterogeneous localization of the immunoreactive materials in such various intra‐I‐band structures occurred randomly without any specific rules noted at present. In contrast, no significant immunoreactivity was seen in A‐band domains. The immunoreactive materials were also associated with the plasma membranes in an intermittent pattern regardless of I‐ or A‐band domains (Figure 3a,b).

FIGURE 3.

FIGURE 3

Open in a new tab

(a‐e) Immuno‐electron micrographs of polar regions of IF muscle fibers. Note, in contrast to immunonegative A‐band domains (white brackets), deposits of electron‐dense immunoreactive materials in most I‐band domains (asterisks) with regional differences in immunoreaction density at lower magnification (a). Areas corresponding to those marked by asterisks in (a) are shown at higher magnification in (b‐d). Also note differences in immunoreaction density among intra‐I‐band organelles even in the same sarcomere: Immunopositive Z‐lines and sarcoplasmic reticulum labeled by Z and S (capital letters), and immunonegative Z‐lines and sarcoplasmic reticulum labeled by z and s (small letters), especially in (b) whose upper and lower domains represent immunonegative and positive ones, respectively. Arrows in (d) indicate individual thin myofilaments exhibiting higher electron density due to the immunoreaction deposits, resulting in appearance of their thickness similar to thick myofilaments of A‐bands. Solid arrowheads indicate intermittent deposits of immunoreactive materials along the plasma membranes of IF muscles, in contrast to immunonegative domains of the membranes (white arrowheads) between them. (e) shows IF muscles in negative control by antigen absorption test. Note the electron density of I‐band domains including Z‐lines almost similar to that of immunonegative domains labeled by z and s (small letters) and significantly lower than that of immunopositive domains labeled by Z and S (capital letters) in (b‐d). IS: interstitial space, m: mitochondria, Tr: triad. Bars represent 1 µm (a), 500 nm (b‐e)

3.2. DRGs at cervical, thoracic, and lumber levels and cranial sensory ganglia

In immuno‐light microscopy, immunoreactivity for PIP5Kγ was moderate to intense in almost all ganglionic cells, and the periphery of individual neuronal somata was especially distinct, resulting in an appearance of dense immunoreactive rings along the contours of neuronal somata (Figure 4a). These features were basically in accord with those already reported by others (Wright et al., 2014). All the features of immunoreactivity described above were essentially the same in any of the DRG regardless of the spinal level and also in the trigeminal and facialis genu ganglia of the cranial nerves (data not shown).

FIGURE 4.

FIGURE 4

Open in a new tab

(a‐g) Immuno‐light micrographs of first lumbar spinal cord and DRG for PIP5Kγ (a, b), both PIP5Kγ and CGRP (c‐f), and negative control by antigen absorption for PIP5Kγ (g). A higher magnification view of an area enclosed by a thick‐lined rectangle in (a) is shown in (b). Note weak immunoreactivity for PIP5Kγ in both peripheral (P) and central (C) processes of neurons. In double immunofluorescence micrographs for PIP5Kγ (green, c) and CGRP (red, d) in spinal posterior horn (PH) enclosed by a yellow rectangle in 4a, a substantial number of nerve terminals are simultaneously immunoreactive for the two molecules in the marginal zone (yellow dots, e‐f) is a higher magnification view of an area marked by rectangle in €. AH: spinal anterior horn, DRG: dorsal root ganglion. Bars represent 200 µm (a, g), 100 µm (b, c‐f)

On the other hand, weak immunoreactivity for PIP5Kγ was seen at similar levels in both central and peripheral processes of neuronal somata in DRG (Figure 4a‐b). In the spinal posterior horn, immunoreactivity for PIP5Kγ was evident in the lamina marginalis (I) and the substantia gelatinosa (II) of the posterior horn (Figure 4a). In double immunostaining for both PIP5Kγ and CGRP, a substantial number of tiny dots simultaneously immunopositive for both molecules were found in the lamina marginalis (Figure 4c‐f).

In immuno‐electron microscopy, immunoreactive materials for PIP5Kγ were deposited discontinuously along the plasma membranes of both neuronal somata and SGCs and the subjacent thin cytoplasm, resulting in an appearance of dominant immunoreactive domains of the plasma membranes with immunonegative domains of them interposed (Figure 5a,c‐g). The deposits were also distinct in neuronal short cytoplasmic projections to the intercellular spaces with SGCs (Figure 5d). At the immunopositive plasma membrane domains, omega‐shaped invaginations and small vesicles with immunoreactivity were often seen (Figure 5c‐g). In SGCs, immunoreactive materials were deposited in the Golgi apparatus (Figure 5a,b) but were not seen on their plasma membranes facing the interstitial spaces (Figure 5b‐g).

FIGURE 5.

FIGURE 5

Open in a new tab

(a‐g) Immuno‐electron micrographs of neuron somata (N) and enclosing ganglionic satellite cells (S), and their two apposed plasma membranes in DRG. Extensions of thin envelopings of satellite cells (arrows) are indicated in (a). In a higher magnification image of an area enclosed by a rectangle in (a) is shown in (b), note deposits of immunoreactive materials in Golgi apparatus (G) of ganglionic satellite cell (S). At higher magnification of the two apposed plasma membranes and adjacent thin cytoplasmic domains (c‐g), immunoreactive materials are deposited (⋆) with immunonegative plasma membrane portions (underlined by yellow broken lines) interposed. Short projections (P) of neurons are also immunoreactive. Small vesicles and membranous invaginations (white arrowheads in neurons and dark arrowheads in satellite cells) are embedded within or associated with deposits of immunoreactive materials. IS: interstitial space. Bars represent 2 µm (a), 1 µm (b, g), 800 nm (c‐f)

In the antigen absorption test as a negative control for PIP5Kγ‐immunoreactivity, no significant immunoreactivity was found in any portions of tissues examined at both light and electron microscopic levels (Figures 1g, 3e, 4g).

4. DISCUSSION

In accord with the previous studies by others (Loo & Zylka, 2017; Wright et al., 2014), intense immunoreactivity in almost all neuronal somata of spinal DRG was faithfully reproduced in the present immuno‐light microscopy. Since the antibody for PIP5Kγ employed in their studies was a gift from one (HS) of the present authors to them, this reproduction is reasonable and the authenticity of the immunoreaction in the present study can thus be regarded as having already been confirmed by their Western blot findings reported in their studies. In addition, the present study clarified such a light microscopic localization of the immunoreactivity in neural somata to be the same in the cranial sensory ganglia as well as DRG at various levels of the spinal cord (data not shown).

A major finding in this study is the disclosure of thin muscle fibers and nerve terminals abutting on and entwining the muscles, both of which were distinctly immunoreactive for PIP5Kγ in contrast to immunonegative thick skeletal muscle fibers of the regular type. The two immunopositive structures were identified as intrafusal muscle fibers and innervating sensory (annulospiral) nerve terminals of the muscle spindles based on their established criteria in light microscopic histology and ultrastructure, especially the absence of the basal lamina intervening between muscles and nerve terminals and of the foldings of apposed muscular plasma membranes (Ovalle, 1972; Winarakwong et al., 2004). The latter two criteria rule out the possibility that the immunoreactive terminal is motor in nature. Since PIP5Kγ‐immunoreactivity was at negligible levels in para‐ and prevertebral sympathetic ganglia (data not shown), the possibility that the immunoreactive entwining nerve terminals were sympathetic is ruled out.

Since proprioceptive neurons are known to be included in DRG (Inoue et al., 2002), and because distinct PIP5Kγ‐immunoreactivity in almost all of DRG neuronal somata were reported by the previous authors (Wright et al., 2014) and confirmed in the present study, it is reasonable to conclude that the immunoreactive nerve terminals abutting on the immunoreactive intrafusal muscle fibers originate from DRG. In this regard, a finding of the previous molecular biological study (Loo & Zylka, 2017) should be noted: the detection of early onset of proprioceptive deficit in sensory neuron‐selective conditional Pip5kγ knockout mice. The present finding of PIP5Kγ‐immunoreactive sensory nerve terminals is the morphological basis for the phenotype in proprioception of the conditional gene‐knockout mice. Furthermore, there has been a study reporting that the third phenotype of lethal congenital (autosomal recessive) contractural syndrome (LCCS3) results from a mutation of PIP5Kγ (Narkis et al., 2007). Judging from the occurrence of severe muscle wasting and atrophy in association with multiple joint contractures among its clinical phenotypes, it is possible to speculate the disturbance of proprioception in this disease. In this regard, the present finding on the localization of PIP5Kγ‐immunoreactivity in the muscle spindle is highly significant to understand the pathogenic phenotype of this congenital disease.

Concerning the light microscopic striations showing intense immunoreactivity in intrafusal muscle fibers, varieties in the length and thickness should be noted along their trajectory although their constant periodicity was largely maintained. In immuno‐electron microscopy, such varieties were revealed to be ascribed to different degrees of deposits of the immunoreactive materials on individual intra‐I‐band structures, including Z‐lines, sarcoplasmic reticulum, and thin myofilament bundles. From the good correspondence of these immunoreactive substructures between light and electron microscopy and their consistent occurrence, this apparently irregular deposits in the ultrastructure of the immunoreactivity on these various intra‐I‐band structures are unlikely to be owed to any technical artifacts. It is thus tempting to consider that this phenomenon represents different levels of activities of the individual substructures, all of which may be regulated by either one of a wide variety of roles played by PIP5K and its product PIP2 as described in the Introduction.

As for PIP5Kγ‐immunoreactivity in such membrane‐less structures as Z‐bands and thin myofilaments, although PIP5K and other PI‐signal‐related enzyme molecules were originally functional on plasma membranes and intracellular membranes, the localization itself is not peculiar at present because there has been ample evidence for the PI‐signaling molecules to be localized in membrane‐less intranuclear domains (Choi et al., 2019; Divecha & Irvine, 1995; Keune et al., 2011). The presence of Z‐bands with or without PIP5Kγ‐immunoreactivity suggests the molecular heterogeneity in Z‐bands. While there have been data showing heterogeneity of Z‐bands connecting anti‐parallel actin filaments formed by α‐actinin, reflecting the difference in thickness of Z‐bands (Luther et al., 2003), no significant differences in their thickness were found between PIP5Kγ‐immunopositive and PIP5Kγ‐immunonegative Z‐bands in the present study. In this regard, a recent finding reported by the present authors (Chomphoo et al., 2020) should be noted. In that study, immunoreactivities for PIP5Kγ as well as Arf6, an upstream activator of this isozyme, together with EFA6 (exchange factor for Arf6) A, an activator of Arf6, were found to be confined to I‐bands in regular skeletal muscle fibers of pre‐ and perinatal mice, and all these molecules were thereafter progressively attenuated as postnatal development proceeded. In addition, EFA6‐ and Arf6‐immunoreactivities were localized on thin myofilaments in immune‐electron microscopy, although ultrastructural localization of PIP5Kγ‐immunoreactivity was not clarified because of its weak reaction in that study. Therefore, it is tempting to speculate that the requirement of PIP5Kγ in some yet‐unidentified functions of prenatal developing skeletal muscle cells is maintained in adult intrafusal muscles different from adjacent extrafusal muscles.

Another main finding of this study is related to the localization of PIP5Kγ in DRG. The first to be noted in this regard is no detection of nerve terminals showing significant levels of the immunoreactivity, except for the proprioceptive terminals abutting on intrafusal muscle fibers, in several representative sites of the body involved in nociception. All of the sites examined are well known to be rich in nerve fibers and terminals immunoreactive for such sensory terminal markers as CGRP (Franco‐Cereceda et al., 1987), which was confirmed in the present study (data not shown).

This failure of detection of PIP5Kγ‐immunoreactive nerve terminals at these sites of peripheral sensory perception suggests several possibilities as follows: One possibility is that the antigenicity of PIP5Kγ in nerve terminals of pain perception is changed from that in neuronal somata by some causes including posttranslational modification during the axoplasmic flow, with its bioactive level in these terminals maintained sufficiently for its involvement in the regulation of nociception at the periphery. The present study, however, showed distinct immunoreactivity for PIP5Kγ in numerous nerve terminals of the central process of DRG neurons in the marginal layer of spinal cord posterior horn. The possibility of the antigenicity change of PIP5Kγ in the periphery is thus less likely. Another possibility is that the axoplasmic transport of PIP5Kγ in DRG neurons is not so active to reach its level sufficient to be detected by immunohistochemistry in peripheral terminals, different from that in the central process of DRG neurons. The almost similar levels of immunoreactivity for PIP5Kγ in both central and peripheral nerve processes at their proximal sites close to DRG neuron somata is not in favor of this possibility. On the other hand, there have been accumulating evidence for the importance of PIP2 in the regulation of sensory perception‐implicated receptors such as TRPV (Transient Receptor Potential Vanilloid) and bradykinin receptors in peripheral terminals (Braun, 2008; Cao et al., 2013; Prescott & Julius, 2003; Ufret‐Vincenty et al., 2015). The functional significance of PIP5Kγ‐immunoreactivity in nerve terminals abutting on intrafusal muscle fibers can be understood in this way, but that in nociceptive nerve terminals is difficult to be understood in this way because of the failure of detection of its immunoreactivity there.

On the other hand, the present study showed numerous PIP5Kγ‐immunoreactive nerve terminals in the marginal zone of spinal posterior horn, a substantial number of which were simultaneously CGRP‐immunoreactive, inferring colocalization of the two molecules in the central nerve terminals of DRG. CGRP‐immunoreactive terminals in the spinal marginal zone have already been reported to come from DRG by others (Gibson et al., 1984), and the occurrence of distinct PIP5Kγ‐immunoreactivity in almost all DRG neurons was reported by others and confirmed in the present study. Therefore, it is thus likely that PIP5Kγ is present in the terminals of central processes of PIP5Kγ‐immunoreactive DRG neuron somata.

Under admission of the difficulty at present to extrapolate function from a lack of immunohistochemical signals for PIP5Kγ in peripheral pain perception fields, and based on the present findings together with the high content of PIP5Kγ in the brain already reported by two (HS & HK) of the present authors (Akiba et al., 2002), it is tempting to consider that the regulation by PIP5Kγ of nociception is exerted dominantly in DRG and sensory neural tracts central, rather than peripheral, to the ganglia. This possibility is compatible with the second report (Loo & Zylka, 2017) by the same group as those who initially published the involvement of PIP5Kγ of DRG in nociception (Wright et al., 2014) as cited in Introduction.

The second to be noted about the DRG is the distinct localization of immunoreactivity for PIP5Kγ along the contours of DRG neuronal somata in light microscopy as already reported in the previous study (Wright et al., 2014). The present ultrastructural study newly revealed such distinct localization of immunoreactivity along the neuronal contours to be ascribed to the plasma membranes of not only neurons but also SGCs thinly enclosing neurons. The localization of immunoreactivity for PIP5Kγ on SGC membranes was supported by its localization also in Golgi apparatus of SGC somata. This colocalization finding is compatible with the recent idea that that neuronal somata of DRG activates SGC, which in turn exert excitatory or inhibitory modulation of neuronal activity, resulting in active involvement of SGCs in the process of afferent signaling (Gu et al., 2010; Hunani, 2005; Huang & Nehar, 1996; Zhang et al., 2005, 2007). The present ultrastructural study also disclosed the frequent association of immunoreactive small vesicles and membranous invaginations at the apposed plasma membrane sites of both cells, and the immunoreaction deposits in short and small somatic projections of neurons whose presence is already known without information on functional roles (Pannese, 1981). It is thus suggested that PIP5Kγ and PIP2 play roles through active membrane trafficking in both DRG neuron and SGC, and that the projections of neuronal somata are implicated in roles played by these molecules.

Regarding the distinct immunoreactivity for PIP5Kγ along the contours of DRG neuronal somata, this feature is discrete from rather simple and diffuse localization throughout the somata of immunoreactivities in single sections for a variety of bioactive molecules including even receptors and ion channels which have so far been reported in DRG neurons by others (Aikawa & Martin, 2003; Funakoshi et al., 2010; Giudici et al., 2006; Grubb & Evans, 1999; Hanani, 2005; Huang et al., 2013; Suadicani et al., 2010). Ultrastructural feature of immunoreactive domains of the plasma membranes with their immunonegative domains interposed, indicates that PIP5Kγ is associated selectively with such heterogeneously immunoreactive membrane domains on demand to catalyze PIP to PIP2. Since PIP2 has been shown to be localized in cholesterol‐rich lipid rafts in cells in vitro (Morenilla‐Palao et al., 2009), and because there has been evidence for the role of lipid rafts in Ca2+‐gating of some channel molecules in sensory neurons (Horváth et al., 2021; Sághy et al., 2015), it is tempted to speculate the possibility that the heterogeneously immunoreactive membrane domains include such lipid rafts in the apposed plasma membranes of DRG.

Taken all together, it is suggested that pseudo‐unipolar sensory neuronal somata are not only sites for an action potential initiated in their peripheral/distal process terminals to bypass for continuous propagation along the central process but also implicated in the sensory signaling process in intimate collaboration with SGCs in a more complicated way than previously thought. Further study on localization of PIP5Kγ in DRG of female mice remains to be elucidated for more authenticity of the present findings and interpretation regardless of the sex.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

AUTHORS’ CONTRIBUTIONS

SC contributed to the experimental design, data analysis and interpretation, and drafted the manuscript. HS contributed to the synthesis of antibody, data analysis and interpretation, and made critical revisions to the drafted manuscript. HK contributed to the experimental design, interpretation, and made critical revisions to the drafted manuscript. WH contributed to the experimental design and made critical revisions to the drafted manuscript.

ACKNOWLEDGMENTS

Sincerely thanks are extended to Mr D Hipkaeo for his technical supports. We would like to acknowledge Emeritus Prof. James A. Will, University of Wisconsin, Madison for editing the manuscript via Publication Clinic, Khon Kaen University, Thailand.

Chomphoo, S. , Sakagami, H. , Kondo, H. & Hipkaeo, W. (2021) Localization of PIP5Kγ selectively in proprioceptive peripheral fields and also in sensory ganglionic satellite cells as well as neuronal cell membranes and their central terminals. Journal of Anatomy, 239, 1196–1206. 10.1111/joa.13491

Funding information

This study was granted by Faculty of Medicine, Khon Kaen University, Thailand (Grant Number IN64131 to SC).

REFERENCES

  1. Aikawa, Y. & Martin, T.F. (2003) ARF6 regulates a plasma membrane pool of phosphatidylinositol(4,5)bisphosphate required for regulated exocytosis. The Journal of cell biology, 162(4), 647–659. 10.1083/jcb.200212142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Akiba, Y. , Suzuki, R. , Saito‐Saino, S. , Owada, Y. , Sakagami, H. , Watanabe, M. et al. (2002) Localization of mRNAs for phosphatidylinositol phosphate kinases in the mouse brain during development. Brain Research. Gene expression patterns, 1(2), 123–133. 10.1016/s1567-133x(01)00023-0. [DOI] [PubMed] [Google Scholar]
  3. Balla, T. , Szentpetery, Z. & Kim, Y.J. (2009) Phosphoinositide signaling: New tools and insights. Physiology (Bethesda, MD.), 24, 231–244. 10.1152/physiol.00014.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Braun, A.P. (2008) BK channels and the expanding role for PIP2‐mediated regulation. Channels (Austin, TX.), 2(4), 229. 10.4161/chan.2.4.6873. [DOI] [PubMed] [Google Scholar]
  5. Cao, E. , Cordero‐Morales, J.F. , Liu, B. , Qin, F. & Julius, D. (2013) TRPV1 channels are intrinsically heat sensitive and negatively regulated by phosphoinositide lipids. Neuron, 77(4), 667–679. 10.1016/j.neuron.2012.12.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Choi, S. , Chen, M. , Cryns, V.L. & Anderson, R.A. (2019) A nuclear phosphoinositide kinase complex regulates p53. Nature Cell Biology, 21(4), 462–475. 10.1038/s41556-019-0297-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chomphoo, S. , Sakagami, H. , Kondo, H. & Hipkaeo, W. (2020) Discrete localization patterns of Arf6, and its activators EFA6A and BRAG2, and its effector PIP5kinaseγ on myofibrils of myotubes and plasma membranes of myoblasts in developing skeletal muscles of mice. Acta Histochemica, 122(3), 151513. 10.1016/j.acthis.2020.151513. [DOI] [PubMed] [Google Scholar]
  8. Delage, E. , Puyaubert, J. , Zachowski, A. & Ruelland, E. (2013) Signal transduction pathways involving phosphatidylinositol 4‐phosphate and phosphatidylinositol 4,5‐bisphosphate: convergences and divergences among eukaryotic kingdoms. Progress in Lipid Research, 52(1), 1–14. 10.1016/j.plipres.2012.08.003. [DOI] [PubMed] [Google Scholar]
  9. Divecha, N. & Irvine, R.F. (1995) Phospholipid signaling. Cell, 80(2), 269–278. 10.1016/0092-8674(95)90409-3. [DOI] [PubMed] [Google Scholar]
  10. Doughman, R.L. , Firestone, A.J. & Anderson, R.A. (2003) Phosphatidylinositol phosphate kinases put PI4,5P(2) in its place. The Journal of Membrane Biology, 194(2), 77–89. 10.1007/s00232-003-2027-7. [DOI] [PubMed] [Google Scholar]
  11. Franco‐Cereceda, A. , Henke, H. , Lundberg, J.M. , Petermann, J.B. , Hökfelt, T. & Fischer, J.A. (1987) Calcitonin gene‐related peptide (CGRP) in capsaicin‐sensitive substance P‐immunoreactive sensory neurons in animals and man: distribution and release by capsaicin. Peptides, 8(2), 399–410. 10.1016/0196-9781(87)90117-3. [DOI] [PubMed] [Google Scholar]
  12. Funakoshi, Y. , Hasegawa, H. & Kanaho, Y. (2010) Activation mechanisms of PIP5K isozymes by the small GTPase ARF6. Advances in Enzyme Regulation, 50(1), 72–80. 10.1016/j.advenzreg.2009.11.001. [DOI] [PubMed] [Google Scholar]
  13. Gibson, S.J. , Polak, J.M. , Bloom, S.R. , Sabate, I.M. , Mulderry, P.M. , Ghatei, M.A. et al. (1984) Calcitonin gene‐related peptide immunoreactivity in the spinal cord of man and of eight other species. Journal of Neuroscience, 4(12), 3101–3111. 10.1523/JNEUROSCI.04-12-03101.1984. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Giudici, M.L. , Lee, K. , Lim, R. & Irvine, R.F. (2006) The intracellular localisation and mobility of Type Igamma phosphatidylinositol 4P 5‐kinase splice variants. FEBS letters, 580(30), 6933–6937. 10.1016/j.febslet.2006.11.052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Grubb, B.D. & Evans, R.J. (1999) Characterization of cultured dorsal root ganglion neuron P2X receptors. The European Journal of Neuroscience, 11(1), 149–154. 10.1046/j.1460-9568.1999.00426.x. [DOI] [PubMed] [Google Scholar]
  16. Gu, Y. , Chen, Y. , Zhang, X. , Li, G.W. , Wang, C. & Huang, L.Y. (2010) Neuronal soma‐satellite glial cell interactions in sensory ganglia and the participation of purinergic receptors. Neuron Glia Biology, 6(1), 53–62. 10.1017/S1740925X10000116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hanani, M. (2005) Satellite glial cells in sensory ganglia: from form to function. Brain Research Reviews, 48(3), 457–476. 10.1016/j.brainresrev.2004.09.001. [DOI] [PubMed] [Google Scholar]
  18. Hara, Y. , Fukaya, M. , Tamaki, H. & Sakagami, H. (2013) Type I phosphatidylinositol 4‐phosphate 5‐kinase γ is required for neuronal migration in the mouse developing cerebral cortex. The European Journal of Neuroscience, 38(5), 2659–2671. 10.1111/ejn.12286. [DOI] [PubMed] [Google Scholar]
  19. Horváth, A. , Payrits, M. , Steib, A. , Kántás, B. , Biró‐Süt, T. , Erostyák, J. et al. (2021) Analgesic effects of lipid raft disruption by sphingomyelinase and myriocin via transient receptor potential vanilloid 1 and transient receptor potential ankyrin 1 ion channel modulation. Frontier Pharmacology, 11, 593319, 10.3389/fphar.2020.593319. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Huang, L.Y. , Gu, Y. & Chen, Y. (2013) Communication between neuronal somata and satellite glial cells in sensory ganglia. Glia, 61(10), 1571–1581. 10.1002/glia.22541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Huang, L.Y. & Neher, E. (1996) Ca(2+)‐dependent exocytosis in the somata of dorsal root ganglion neurons. Neuron, 17(1), 135–145. 10.1016/s0896-6273(00)80287-1. [DOI] [PubMed] [Google Scholar]
  22. Inoue, K. , Ozaki, S. , Shiga, T. , Ito, K. , Masuda, T. , Okado, N. et al. (2002) Runx3 controls the axonal projection of proprioceptive dorsal root ganglion neurons. Nature neuroscience, 5(10), 946–954. 10.1038/nn925. [DOI] [PubMed] [Google Scholar]
  23. Kanaho, Y. , Kobayashi‐Nakano, A. & Yokozeki, T. (2007) The phosphoinositide kinase PIP5K that produces the versatile signaling phospholipid PI4,5P(2). Biological and Pharmaceutical Bulletin, 30(9), 1605–1609. 10.1248/bpb.30.1605. [DOI] [PubMed] [Google Scholar]
  24. Keune, W.J. , Bultsma, Y. , Sommer, L. , Jones, D. & Divecha, N. (2011) Phosphoinositide signalling in the nucleus. Advances in Enzyme Regulation, 51(1), 91–99. 10.1016/j.advenzreg.2010.09.009. [DOI] [PubMed] [Google Scholar]
  25. Loo, L. & Zylka, M. (2017) Conditional deletion of Pip5k1c in sensory ganglia and effects on nociception and inflammatory sensitization. Molecular Pain, 13, 1744806917737907. 10.1177/1744806917737907. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Luther, P.K. , Padrón, R. , Ritter, S. , Craig, R. & Squire, J.M. (2003) Heterogeneity of Z‐band structure within a single muscle sarcomere: implications for sarcomere assembly. Journal of Molecular Biology, 332(1), 161–169. 10.1016/s0022-2836(03)00883-0. [DOI] [PubMed] [Google Scholar]
  27. Morenilla‐Palao, C. , Pertusa, M. , Meseguer, V. , Cabedo, H. & Viana, F. (2009) Lipid raft segregation modulates TRPM8 channel activity. Journal of Biological Chemistry, 284(14), 9215–9224. 10.1074/jbc.M807228200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Narkis, G. , Ofir, R. , Landau, D. , Manor, E. , Volokita, M. , Hershkowitz, R. et al. (2007) Lethal contractural syndrome type 3 (LCCS3) is caused by a mutation in PIP5K1C, which encodes PIPKI gamma of the phophatidylinsitol pathway. American Journal of Human Genetics, 81(3), 530–539. 10.1086/520771. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Oude Weernink, P.A. , Schmidt, M. & Jakobs, K.H. (2004) Regulation and cellular roles of phosphoinositide 5‐kinases. European Journal of Pharmacology, 500(1–3), 87–99. 10.1016/j.ejphar.2004.07.014. [DOI] [PubMed] [Google Scholar]
  30. Ovalle, W.K. Jr (1972) Fine structure of rat intrafusal muscle fibers. The equatorial region. The Journal of Cell Biology, 52(2), 382–396. 10.1083/jcb.52.2.382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Pannese, E. (1981) The satellite cells of the sensory ganglia. Advances in Anatomy, Embryology, and Cell Biology, 65, 1–111. 10.1007/978-3-642-67750-2. [DOI] [PubMed] [Google Scholar]
  32. Prescott, E.D. & Julius, D. (2003) A modular PIP2 binding site as a determinant of capsaicin receptor sensitivity. Science (New York, N.Y.), 300(5623), 1284–1288. 10.1126/science.1083646. [DOI] [PubMed] [Google Scholar]
  33. Sághy, É. , Szőke, É. , Payrits, M. , Helyes, Z. , Börzsei, R. , Erostyák, J. et al. (2015) Evidence for the role of lipid rafts and sphingomyelin in Ca2+‐gating of Transient Receptor Potential channels in trigeminal sensory neurons and peripheral nerve terminals. Pharmacological Research, 100, 101–116. 10.1016/j.phrs.2015.07.028. [DOI] [PubMed] [Google Scholar]
  34. Suadicani, S.O. , Cherkas, P.S. , Zuckerman, J. , Smith, D.N. , Spray, D.C. & Hanani, M. (2010) Bidirectional calcium signaling between satellite glial cells and neurons in cultured mouse trigeminal ganglia. Neuron Glia Biology, 6(1), 43–51. 10.1017/S1740925X09990408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Ufret‐Vincenty, C.A. , Klein, R.M. , Collins, M.D. , Rosasco, M.G. , Martinez, G.Q. & Gordon, S.E. (2015) Mechanism for phosphoinositide selectivity and activation of TRPV1 ion channels. The Journal of General Physiology, 145(5), 431–442. 10.1085/jgp.201511354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Wenk, M.R. , Pellegrini, L. , Klenchin, V.A. , Di Paolo, G. , Chang, S. , Daniell, L. et al. (2001) PIP kinase Igamma is the major PI(4,5)P(2) synthesizing enzyme at the synapse. Neuron, 32(1), 79–88. 10.1016/s0896-6273(01)00456-1. [DOI] [PubMed] [Google Scholar]
  37. Winarakwong, L. , Muramoto, T. , Soma, K. & Takano, Y. (2004) Age‐related changes and the possible adaptability of rat jaw muscle spindles: immunohistochemical and fine structural studies. Archives of Histology and Cytology, 67(3), 227–240. 10.1679/aohc.67.227. [DOI] [PubMed] [Google Scholar]
  38. Wright, B.D. , Loo, L. , Street, S.E. , Ma, A. , Taylor‐Blake, B. , Stashko, M.A. et al. (2014) The lipid kinase PIP5K1C regulates pain signaling and sensitization. Neuron, 82(4), 836–847. 10.1016/j.neuron.2014.04.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Zhang, X. , Chen, Y. , Wang, C. & Huang, L.Y. (2007) Neuronal somatic ATP release triggers neuron‐satellite glial cell communication in dorsal root ganglia. Proceedings of the National Academy of Sciences of the United States of America, 104(23), 9864–9869. 10.1073/pnas.0611048104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Zhang, X.F. , Han, P. , Faltynek, C.R. , Jarvis, M.F. & Shieh, C.C. (2005) Functional expression of P2X7 receptors in non‐neuronal cells of rat dorsal root ganglia. Brain Research, 1052(1), 63–70. 10.1016/j.brainres.2005.06.022. [DOI] [PubMed] [Google Scholar]

Articles from Journal of Anatomy are provided here courtesy of Anatomical Society of Great Britain and Ireland

RESOURCES