|Home | About | Journals | Submit | Contact Us | Français|
A set of astrocytic process associated with altered myelinated axons is described in the forebrain of normal adult rodents with confocal, electron microscopy, and 3D reconstructions. Each process consists of a protuberance that contains secretory organelles including numerous lysosomes which polarize and open next to disrupted myelinated axons. Because of the distinctive asymmetric organelle distribution and ubiquity throughout the forebrain neuropil, this enlargement is named paraxial process (PAP). The myelin envelope contiguous to the PAP displays focal disruption or disintegration. In routine electron microscopy clusters of large, confluent, lysosomes proved to be an effective landmark for PAP identification. In 3D assemblies lysosomes organize a series of interconnected saccules that open up to the plasmalemma next to the disrupted myelin envelope(s). Activity for acid hydrolases was visualized in lysosomes, and extracellularly at the PAP-myelin interface and/or between the glial and neuronal outer aspects. Organelles in astrocytic processes involved in digesting pyknotic cells and debris resemble those encountered in PAPs supporting a likewise lytic function of the later. Conversely, processes entangling tripartite synapses and glomeruli were devoid of lysosomes. Both oligodendrocytic and microglial processes were not associated with altered myelin envelopes. The possible roles of the PAP in myelin remodeling in the context of the oligodendrocyte-astrocyte interactions and in the astrocyte's secretory pathways are discussed.
Understanding cellular domains has been an essential preliminary step to define functional properties inherent to each cell type. An increasing amount of evidences for the existence of cytoplasmic domains in astrocytes (ACs) has recently been accumulated. Earlier descriptions of the AC established two broad subtypes as identified in silver-impregnated specimens. According to the cell location, size, and structure of its radial processes, two AC subtypes are identified: protoplasmic and fibrous ACs. Protoplasmic ACs are commonly associated with nuclei and neuron layers, whereas fibrous ACs lie between axonal fascicles in the white matter. While the overall branching and distribution of AC processes is radial, thereby occupying rounded to oval fields (Del Río Hortega, 1914; Butt and Ransom, 1993), regional variations of astrocytic processes (APs) that reside in elliptical and/or overlapping fields are frequent (Butt et al., 1994; López-Hidalgo et al., 2016). Although primary APs generally follow a straight course, they each originate numerous shorter formations whose submissive accommodation among the neighboring neural and stromal elements imprint to them assorted, frequently, unpredictable shapes (Göbel, 1966; Valverde and López-Mascaraque, 1991). Perhaps the best-known exception regarding the signed pleomorphism of the APs is where they interact with the outer aspect of the blood vessels. In fact, one or a set of APs terminate in distinct cytoplasmic expansions, or end-feet, that surround capillary blood vessels, thereby creating a continuous sheet of APs lining the blood vessels throughout most areas of the central nervous system. Sites of the brain where capillaries are devoid of this astrocytic cuff are commonly permeable to dyes and other high molecular substances that are otherwise kept in the vascular compartment. This association of the end-feet with the capillary wall provides the structural basis for the blood-brain barrier (Brightman and Reese, 1969; Peters et al., 1976; Broadwell and Sofroniew, 1993). Additional support to the functional role of AC's end-feet is their reciprocal inductive influences with the under laying vascular endothelium (Sperri et al., 1997). Furthermore, localization of aquaporin in the AC end-feet has been implicated in the pathogenesis of neuromyelitis optica autoimmune (Jarius and Wildeman, 2010), a histopathological variant of multiple sclerosis. Another instance recognized earlier by Ramón y Cajal (1904) relates to a group of APs termed pericellular pedicles (“pedículos pericelulares”) that ramify to surround neighboring neuronal perikarya (Del Río Hortega, 1914). More recent studies have shown that this pericellular framework influences the plasticity of synaptic inputs that arrive to the neuron's soma (Garcia-Segura et al., 1994). It has been further demonstrated that the sites of interaction between the AC processes and synaptic terminals noted earlier (Göbel, 1966), define substrates for chemical and functional reciprocal neuron-glia interactions, the so-called tripartite synapses (Araque et al., 1999; Ventura and Harris, 1999; Witcher et al., 2010). Furthermore, glycogen that is stored and provided by the AC exhibits unique regional and cytoplasmic patterns of distribution (Cali et al., 2016) that concentrate next to axon boutons and postsynaptic spines, in order of frequency, supporting the notion that ACs modulate chemical synaptic transmission. Lastly, there is a set of cytoplasmic digitations arising from the AC perikaryon (see Del Río Hortega, 1914) or the end-foot proper, termed perinodal processes that directly interact with the Ranvier node to share gap junctions with the bounding paranodal loops of the oligodendrocyte (Ol; Black et al., 1984; Waxman and Black, 1984; Black and Waxman, 1988). The possibility that this AC-Ol interaction plays a role in myelin homeostasis has been put forth (Hildebrand et al., 1993).
Astrocytes play a key role in a variety of cellular responses including: cell death (He et al., 2006), developmental and adult synaptic pruning (see Chung et al., 2014), nerve degeneration (Fried and Hildebrand, 1982; Hildebrand, 1989), experimental lesioning of neurons and axon terminals (Price and Powell, 1970a; Larriva-Sahd, 2008; Marcellino et al., 2012; Anderson et al., 2016), and phagocytosis of debris (see Tasdemir-Yilmaz and Freeman, 2014; Iram et al., 2016). The AC is also a cardinal cell type for the ontogenetical process of myelin synthesis by the Ol. Thus, the AC interacting directly with the Ol precursor cell, promotes viability (Corley et al., 2001) of this cell type to further migrate (Bauman and Pham-Dinh, 2001) and differentiate to the myelin-synthesizing Ol (see Clemente et al., 2013). It is also accepted that the AC proper is directly involved in myelin retrieval in the plasticity of the developing nervous system, leading to the possibility that ACs are responsible for the process of myelin remodeling during the adulthood (Hildebrand and Aldskogius, 1976). Indeed, available studies in normal adult mammals showed that microglia and perivascular, and astroglial cells, in decreasing frequency, interact and internalize fragmented myelin envelopes (Hildebrand and Aldskogius, 1976). Furthermore, interaction of the normal AC bearing lysosomes positive to acid hydrolases with altered myelinated fibers, was noted earlier (Hildebrand, 1971; Hildebrand and Skoglund, 1971). To our knowledge these authors demonstrated by first time that lysosome-bearing AC coexists with altered myelinated fibers in the normal brain. Thus far, it is uncertain how lysosomes interact with either phagocytized or extracellular myelin. An important recent study of the amphibian optic nerve defined that ACs remodel the redundant myelin envelope of axons that results from metamorphosis (Mills et al., 2015). While in the normal mammalian central nervous system the Ol is readily involved in myelin sorting at the molecular domain (Maier et al., 2008) the role of Ols, microglia, and macrophages in myelin clearance appears to be reserved to pathological, pharmacological or experimental (Nathaniel and Pease, 1963; Hildebrand et al., 1993, 1994) conditions leading to myelin damage. Present study combines cyto-, immuno-histochemical, and electron microscopic techniques to define how astrocytic processes interact with myelin. Then, series of ultrathin sections were used to depict a 3D structure at the subcellular level. Regular identification of a set of APs based on its vesicular contents and its association with myelinated fibers piercing the neuropil confirmed their ubiquitous distribution. Present observations define the existence of a ubiquitous group of APs containing numerous lysosomes and other secretory organelles (see Zorec et al., 2017) that interact with the altered myelin envelope of nearby axons, suggesting that they are involved in normal myelin remodeling in the adult forebrain. Experimental work describing the APs response to axon degeneration will be separately presented.
Normal adult albino rats and mice where utilized. Mice belonged the colony FVB/N-Tg (GFAP-EGFP)GFEA-FKi that display EGFP (eGFP) fluorescence in astrocytes from multiple areas of the CNS (Nolte et al., 2001). Founders of the colony [GFAP-EGFP] were kindly provided by Dr. Helmut Kettenmann (Max Delbruck Center for Molecular Medicine, Berlin, Germany). All animals were males of 10 weeks of age that were raised in a pathogen-free colonies in accordance with animal care policies in our vivarium. Both the study and endeavor of animal handling and experimental manipulations were reviewed and approved (Code number: 2016-60) by Comité de Ética en Investigación, Instituto de Neurobiología, Universidad Nacional Autónoma de México. This committee was appointed by our director and consisted of five professional scientific researchers trained in ethics and familiar with the current international guide-lines of Biomedical research (see Ethics Committee SfN home-page). Animals were killed by vascular aldehyde perfusion (vide infra) under deep anesthesia (i.e., 30 mg/kg pentobarbital) or decapitation as separately described below.
To label acidic organelles (i.e., lysosomes) LysoTracker red (Invitrogen-Molecular Probes, Carlsbad, CA, USA) was used as described elsewhere (Weis et al., 2014). Briefly, seven GFAP-EGFP (eGFP) mice were decapitated and brains were immediately removed from the skull. Then, under the dissecting microscope, each brain was transversally divided into three thirds with a razor blade. Blocks of tissue were incubated for 5 min in a solution of phosphate buffered saline (0.1M, pH 7.4) and LysoTracker (1:400). Following a brief rinse in PBS, blocks of tissue were left in a 4% paraformaldehyde fixative diluted in the same vehicle for 12 h at 4°C. Tissues were transferred to 30% sucrose solution for 2 days and subsequently embedded in Tissue-Tek (OCT, Sakura) to obtain 50 μm sagittal sections with a cryostat (Leica Biosystems). Sections were transferred to PBS, blocked in PBS containing 5% normal goat serum (NGS) and 0.1% Triton X-100 for 45 min and incubated overnight at 4°C in myelin basic protein (MBP) antibody (Molecular Probes, Invitrogen) diluted 1:1,000 in PBS. Then, sections were immersed in a secondary rabbit Cy5 diluted 1:1,000 in PBS for 2 h. Following a 5-min wash in PBS, sections were incubated in DAPI diluted 1:4,000 for 5 min and mounted in Mowiol medium. Observations and image acquisition were performed using a Zeiss 780 LSM confocal microscope. Confocal micrographs were acquired at 0.37 μm interval at 1,024 × 1,024 pixel resolution and further processed and edited with the “Image J” and “Adobe Photoshop” software, respectively. Illustrative 3D videos from Figures 2B,D, were obtained with the “Amira” software (Supplemental Material).
As virtually all lysosomal enzymes are acidic hydrolases (DeDuve, 1983), specimens from five brains were incubated with beta-glycerolphosphate, a well-known substrate to these enzymes (Welsh, 1966; Nichols et al., 1971; Hildebrand, 1982). An alternative technique was utilized, as the enzyme products of the incubation are both chromogenic and electron opaque (Hildebrand and Skoglund, 1971; Bencosme et al., 1979). Thus, sites that are positive to the reaction are visualized under both a conventional light- and an electron microscopes. To accomplish this, five adult rat brains were used. Animals were deeply anesthetized and perfused through the left ventricle with 250 ml of Karnovsky's fixative cooled at 4°C. The fixative consisted of a 4% paraformaldehyde and 2% glutaraldehyde dissolved in 0.1M sodium cacodylate buffer at pH 7.4. Then, each brain was removed from the skull and transferred to fresh Kanrnovsky fixative, where it remained overnight at 4°C. The next day each brain was cut with a vibratome in the sagittal plane at 100 μm of nominal thickness. Sections were incubated in a medium at low-pH containing a beta-glycerolphosphate substratum as detailed elsewhere (Bencosme et al., 1979). Sections incubated in absence of the substratum were left as controls. Demonstrative sections were chosen for light microscopy on the basis of their completeness; then, they were mounted in glass slides and coverslipped with glycerin or further processed for electron microscopy.
Brains from five adult rats were used for this part of the study. Rats were deeply anesthetized and perfused through the left ventricle with Karnovsky's fixative and processed as previously described (Larriva-Sahd, 2006, 2008). Small tissue blocks (~1 mm per side) from the olfactory bulb, ventral hippocampus, and frontal and parietal isocortices were dissected-out from these brains and from sections that had previously been incubated for cytochemistry to acid phosphatases (vide supra). Samples of tissue were postfixed for 1 h in 1% osmium tetraoxide dissolved in 0.1 M cacodylateHCl buffer; then they were dehydrated in graded acetone, and embedded in epoxy resin. Additional tissue blocks from the ventro-medial hypothalamic and medial preoptic nuclei, and the piriform cortex were taken from prior work and processed as described underneath. One micrometer-thin sections were obtained from the tissue blocks in a Leica ultramicrotome utilizing with glass knives. The sections were stained with toluidine blue and coverslipped. Areas to be studied from these blocks included the part of the gray matter with higher occurrence of myelinated fibers and lysosome-bearing processes (see Hildebrand and Skoglund, 1971) as it follows: Layers VI and V of the cerebral cortex (Tomassy et al., 2014), Layer I of the piriform cortex (Vargas-Barroso and Larriva-Sahd, 2013), outer aspects of the medial preoptic (Larriva-Sahd and Gorski, 1987) and ventromedial hypothalamic (Larriva-Sahd et al., 1995) nuclei, intersection of the olfactory bulb cortex and medulla (Larriva-Sahd, 2014), and the ilium of the dentate gyrus. Silver sections at 70 nm of nominal thickness were obtained with a Leica ultramicrotome and mounted on copper grids. Thinner sections approximating 50–55 nm of thickness were utilized to depict organelle interactions at magnifications above of 50,000 x. In either case, sections were sequentially stained with uranium and lead salts and they were observed under a JOEL 1010 electron microscope operated at 80 kV equipped with an automated goniometer.
Thirty-one series of thin sections encompassing 2,141 sections (Table (Table1)1) were utilized to perform 3D reconstructions as described in prior work (Larriva-Sahd, 2014). To test the effect of lysosomal structure in astrocytic processes associated with myelinated axons, 10 additional series, including four pyknotic cells in the rostral migratory stream (Schmechel, 1999; He et al., 2006; Iram et al., 2016), four tripartite synapses (Araque et al., 1999), and three glomeruli (Price and Powell, 1970b) were included. Areas of interest were sequentially photographed from the electron microscope with a Gatan digital camera. Images from each series were aligned and assembled with a free Reconstruct software (http://synapses.clm.utexas.edu/tools/reconstruct/reconstruct.stm.). Astroglial processes containing rough endoplasmic reticulum, mitochondria, Golgi apparatus, electron-dense granules, and intermediate filaments, termed here “paraxial processes” (PAPs), were identified and then successively photographed throughout the series (Table (Table1).1). Axonal and lysosomal dimensions were obtained from the digital micrographs that were utilized for the 3D reconstructions aided by a PC running the Reconstruct software. The transverse diameter of all myelinated fibers interacting with the astrocytic process was defined by direct measurements taken from the Reconstruct software and expressed as the mean axonal diameter per brain area to be studied. A minimum of 30 axons and 800 electron-dense granules was considered for each case.
The incidence of glial processes and myelinated axons in series of 40 serial sections (see in Section Electron Microscopy) was quantified into three forebrain areas. Specimens from olfactory bulb, temporal isocortex, and ventromedial hypothalamic nucleus were submitted to a modified protocol as described elsewhere (Granados-Rojas et al., 2002). Briefly, numbers of PAPs (see Section A Set of Astrocytic Processes Contain Secretory Organelles) and oligodendrocytic processes (see Section Oligodendrocytes and Microglia) were counted through the series at the electron microscope operated at a variable magnification ranging from 5,000 to 40,000x. Double counts were prevented by omission of structures overlapping the X-Y site in the successive sections. Then, the volume of studied tissue was calculated by multiplying the total area of each section by its nominal thickness (i.e., 70 nm) and, then, by the number of sections. For comparison purposes a total volume of 70,000 μm3 was considered. As three series per brain area were included a total volume of 21 × 104 μ3m was studied. Last, the number of myelinated fibers per volume was calculated as it follows. Six electron micrographs at 3,000x magnification were obtained from random areas of the first and last sections. From each digital micrograph the number of axons was obtained with the “Axonseg” open software (Zaimi et al., 2016). The photographed area was normalized to the total area per section and expressed as the mean number of fibers of the first and the last section in the volume of the series.
Terms used throughout the text to designate frequency of occurrence of the structures in more than two-thousand sections (Table (Table1)1) under the electron microscope have the following connotations: “Frequent,” “usual,” or “not infrequently” = present in all or in most sections; “relatively common” = present in at least three or more sections per brain area; “Occasional(ly)” or “uncommon(ly)” = less than three per brain area; “rare” = one to three times throughout the study in a given brain area.
The well-known limitation to define a granule containing electron-dense, often pleomorphic, material as a lysosome (Fawcet, 1981; DeDuve, 1983) is prevented here with a straightforward nomenclature. Thus, whenever a suspected lysosome is observed it is termed “electron-dense granule” or “electron-opaque granule.” Meanwhile, the term “lysosome” is applied to a granule that has an identical structure to the one visualized first in specimens incubated with the substratum (i.e., beta-glycerolphosphate) to detect acid phosphatases. In line with an earlier notion (Holtzman and Dominitz, 1968; Fawcet, 1981) the presence of membranous and/or lipid droplet inclusions in presumptive lysosomes is used as a structural landmark to designate “secondary lysosomes” throughout the text. The term “lysosome” is also used to designate granules positive to the Lyso-Tracker reagent under the confocal microscope (Weis et al., 2014). Finally, the terms “apoptotic cell” and “apoptosis” apply exclusively to shrunken cells in the olfactory bulb medulla which have an overall electron-opaque appearance and varying degrees of nuclear and/or cytoplasmic disruption. These characteristics match with previous observations of apoptotic cells during adult neurogenesis (He et al., 2006), and they are thoroughly analyzed elsewhere (Schmechel, 1999).
Our observations are directed toward the neuropil of various forebrain areas, with an emphasis on the astrocytic processes contained therein. Since cytological characteristics of the AC and neuropil proper have been extensively presented elsewhere (Göbel, 1966; Butt and Ransom, 1993; Butt et al., 1994; Ventura and Harris, 1999; Witcher et al., 2010), only a brief account will be provided for structural context.
Specimens submitted to enzyme histochemistry (Figure (Figure1),1), or Lyso-Tracker fluorescence (Figure (Figure2),2), depict presumptive cell processes containing numerous rounded lysosomes that have assorted sizes from 0.3 to 0.5 μ in diameter. Lysosomes are distributed throughout the neuropil and white matter with a seemingly higher incidence in the former and at sites of transition between them (Figure (Figure1C),1C), as noted earlier in specimens from other mammals (Hildebrand and Skoglund, 1971; Hildebrand, 1982). Processes in the gray matter contain rows of lysosomes paralleling myelinated fibers that pierce through it, although clusters of three or more granules are not infrequent (Figures 1C,E). In both cases lysosome-laden processes bind polygonal or square-shaped areas of the neuropil (Figure (Figure1E).1E). Lysosomes are sparse within the white matter and lie between unstained bundles of myelinated fibers (Figure (Figure1F).1F). Occasional cells exhibit both nuclear and cytoplasmic labeling to acid phosphatase. Specifically, the nuclei of presumptive ACs (Figure (Figure1I)1I) and Bergman's radial glia (Figures 1G,H) are pale and homogeneous, contrasting the dark granular appearance of the cytoplasm due to numerous clusters of lysosomes. In appropriate sections, rows of lysosomes may be seen in the AP. Microglia exhibits a central dark nucleus with numerous lysosomes filling the oval cytoplasm (Figures 1B,J). Neurons usually contain an either weak or absent nuclear reactivity; however, a moderate to large number of small (i.e., <0.2 μm) lysosomes are identified throughout the perikaryon and proximal dendrites (Figure (Figure1C1C).
To define the distribution of lysosomes in the AC cytoplasm, we utilized specimens from a mouse strain displaying eGFP positivity in this cell type, and we further processed them to visualize both lysosomes and myelinated fibers (Figure (Figure2).2). This is accomplished through Lyso-Tracker fluorescence followed by incubation with the myelin basic protein (MBP) antibody (Figures 2A–C). It is noted that the eGFP-positive, distal astrocytic processes in such specimens regularly co-fluoresce with scattered or clustered lysosomes. These double labeled elements co-label focally with MBP-positive fibers, thereby suggesting an actual apposition of distal APs with the latter, although a similar pattern of co-labeling is occasionally observed next to the perikaryon. In sections throughout the olfactory bulb medulla, lysosomes contained by presumptive APs accumulate about pyknotic (Figure (Figure2D)2D) or fragmented nuclei. Since the photic light and confocal microscopes provide a limited resolution of cell membranes and organelles, or interactions between them are frequently overlapped, appropriate specimens were used for the electron microscope as described next.
Survey views of the neuropil (Figure (Figure3)3) reveal the presence of a group of medium-sized APs (i.e., 3–10 μm by 2–4 μm in the longest and transverse axes, respectively) that contain numerous electron-opaque granules, surrounded by free ribosomes, narrow cisterns of the rough endoplasmic reticulum, and two to four Golgi apparatus. Among these organelles, one or two distinct bundles of intermediate filaments course in assorted directions. These outgrowths are eccentric and, although they contain most organelles normally observed in the perikaryon (see, Larriva-Sahd, 2014), some differences are noticeable. The differences at first glance are the higher density, larger size, uneven profiles, and coalescence of the electron-dense granules in processes (Figures (Figures3,3, 4A,B). Another characteristic common to these enlargements is their apposition to one or two myelinated axons coursing through the neighboring neuropil (Figures (Figures33–6). Given that the organelles in these processes have a distinct asymmetrical distribution, protruding sidewise, and are common in the neuropil of most specimens utilized for conventional electron microscopy, they will be referred to as “para-axial processes” (PAPs) (Table (Table1).1). They are more frequent in samples from the olfactory bulb medulla (Figures 3A,B, 4B,C) and, in descending order of frequency, in the hippocampus (Figures (Figures3G,3G, 7), piriform- (not shown), deep parietal- (Figure (Figure3E),3E), and frontal- (Figure (Figure4A)4A) isocortices; and the ventro-medial hypothalamic (Figure (Figure3D),3D), and medial preoptic nuclei (Figure (Figure3C).3C). Attempts to find PAPs in the deep cortical white matter or in optic nerves were unsuccessful. The foregoing description which derives from observations in single or assorted sections, precludes to ascribe the lysosomal nature of granules containing electron-opaque matrices as well as a possible specialization(s) of the PAP along its intersection with myelinated fibers. These methodological drawbacks are approached by combining histochemical and cytochemical techniques with 3D reconstructions from successive sections for electron microscopy as presented next.
Presumptive primary lysosomes (PLs) are small (170–210 nm in diameter), rounded granules that contain a homogeneous matrix of moderated electron-density. In the PAP PLs usually lie next to the trans region of the Golgi apparatus (Figures (Figures3B,3B, 10B; Essner and Novikoff, 1962; Novikoff, 1963; Nichols et al., 1971; Fawcet, 1981). Secondary lysosomes (SLs; Figures Figures33–10) are the most distinctive organelles of the PAP. SLs are larger (<0.5 μm) and barely oval due to large, convex excrescences that give them an overall overinflated appearance. SL content is generally electron-dense, but it varies from one SL to the next (Figures (Figures2,2, ,4A).4A). The granular matrix fills up the entire granule; however, lipid droplets and/or membranous debris, are not infrequent, approximating 3 out of 10 SLs (Figures (Figures3B,3B, 7A). While the presence of lipid-like and membrane inclusions is a feature that SLs (see Fawcet, 1981) and lipofuscin granules (see (Peters et al., 1976; Saprunova et al., 2010)) have in common, the larger size, pleomorphic shape, and contents of the latter (Figures 10B–D) allow identification. A membranous appendage is observed at the surface of about a third of SLs, which consists of a loop made up of a thin duct (i.e., 30 nm in width) outlining a circular area of 0.3 μm in diameter (Figure 6A, sections +1, −3; Figure 6C). Serial sections through these “loops” reveal that they may correspond to a conical, funnel-like compartment that occasionally anchors with the PAP cell membrane (Figure 6D).
SLs coalesce with one another via latero-lateral anastomoses (Figures (Figures3B,3B, 4A,C) and/or via tiny tubules of varying diameters (Figures 4A,C). The latter consists of a finger-like duct whose lumen may be tubular (30 nm) or uneven (30–100 nm) in width and have also a variable length (i.e., 0.04–0.3 μm). Observation of series makes evident that SLs located next to the part of the PAP that interact with a myelinated fiber (see below) open to the limiting plasma membrane through one or two of the narrow ducts just described (Figures (Figures4C,4C, 5A,B, 6B,C, 9B) or by direct fusion of the SL with the plasma membrane (Figures (Figures5C,5C, 8E). Summarizing, SL clusters structure a membranous, communicating system that opens up to the PAP vicinity, usually next to one or two myelinated fibers.
When beta-glycerophosphate is used as the substrate, both PLs and SLs exhibit an electrondense precipitate that may underscore the granule membrane or be diffuse, thus confirming their lysosomal nature (DeDuve, 1983; Figures 8, 9A,B, 10E). Additionally, the reaction product is associated with AC's nuclear envelope and with the Golgi apparatus (see Essner and Novikoff, 1962; Novikoff, 1963; Goldfisher et al., 1964; not shown). Focal electron-dense deposits are common in the myelin envelope of certain axons (Figures 8B–D, 10E). Serial sections through areas of lysosome-myelin interface reveal continuity of reactive deposits bridging them (Figure 8E). Lysosomes in APs surrounding pyknotic cells exhibit both intense reaction products (Figure 8D). Finally, like for the PAP-myelin intersections, the reaction product for acid phosphatase can be seen within the extracellular aspect of the plasma membrane bounding astrocytes and neurons (Figures 8B,Bi).
Axons in the vicinity of the PAP vary in dimensions as a function of brain area (Table (Table1),1), and they have a mean transverse diameter of 0.98 μm. Most axoplasms (i.e., 85%) contained the usual, undisrupted organelles, without structural evidence of damage or improper fixation (Peters, 1970). However, varying degrees of internodal myelin disruption are found in axons interacting with the PAP. Areas of myelin collapse are well circumscribed and measure between 0.08 and 1.5 μm in length. Furthermore, they present two basic structural patterns. The most common pattern consists of focal substitution of both intraperiod and major dense lines by amorphous and electron-dense material (Figure (Figure7A,7A, section 10). Embedded in these areas of rarefaction, pleomorphic, and donut-like figures of electron-opaque material may be seen (Figure (Figure6A,6A, sections +4 and +4i). Observations of areas of the PAP-altered myelin interphase reveal that they coexist with the part of the former where SLs coalesce with the cell membrane; series in Figure Figure66 illustrates the site where an SL fuses with the PAP unitary membrane and the aforementioned myelin alterations arise (Figure (Figure6A,6A, sections +4 and +4i); however, the myelin envelope that is located separately or in distal sections from the myelin rarefaction, acquires the distinct, alternating pattern, observed in normal myelin (Figure (Figure6A,6A, sections +1 and −3; Figure Figure7A,7A, sections 1, 5, and 8). A second alteration of the myelin envelope contiguous to a PAP, which is observed in one third of the series, consists of a transverse interruption of the internodal myelin sheath (Figure (Figure7,7, panels 9–11; Figure Figure8A,8A, arrow). As shown in Figure Figure7,7, transmural transverse dissolution of the myelin leaving denuded areas of the axolemma defines this sort of alteration. The myelin stumps at either side of the gap display rarefaction in a similar fashion to the one just described for the previous alteration. Beyond this focal area of myelin disruption, again, the structure of the normal myelin is observed (Figure (Figure7,7, insets). Series throughout areas of myelin disappearance reveals that they may be conical or bowl-shaped and coexists with that part of the PAP where SLs appose or lead to its plasma membrane (Figure (Figure7A,7A, section 10, and Figures 7B–D).
Myelinated fibers displaying severe internodal disruption are occasionally observed [i.e., 10 times in the olfactory bulb (Figure 10B), three in the cerebral cortex (Figure 10A), and one in the ventromedial hypothalamic nucleus (no shown)]. Internodal myelin of these relative infrequent axons show widespread tumefaction, dissolution, and “ballooning” (Peters, 2009a,b; Figure Figure8A,8A, circle). Phagocytosis of presumptive fragmented myelin next to these severely damaged axons (Figures 10A,B) mimics that observed in senescence (Peters, 2009a) or in advanced stages of experimental axonal transection (Nathaniel and Pease, 1963; Franson and Ronnevi, 1984). Reconstructions throughout PAPs and AC perikarya display both progressive increase in lysosomal number and occurrence contiguous to the disrupted myelin envelope (Figures (Figures9C,9C, 10C,D).
If observations regarding localization, fusion, and acid phosphatase positivity of the PAP counterpart with a lytic process, they may be mimicked by other known lytic phenomena (Tasdemir-Yilmaz and Freeman, 2014) or be absent in APs normally associated with synapses. This is tested first in the olfactory bulb by observation of phagocytosis of apoptotic neuroblasts by the AC (He et al., 2006). In sections incubated with beta-glycerolphosphate lysosomes gather and fuse with the plasma membrane and both the lysosomes and the shrunken, presumptively apoptotic cell, display positivity to acid phosphatases (Figures (Figures2D2D and and8D).8D). To define if this takes place within a phagosome or, extracellularly between the AC and its processes, observation of series throughout the AC-shrunken cell(s) intersection was performed. Figure Figure1111 reproduces an example from two pyknotic cells trapped between two ACs next to the capillary wall. Reconstruction of all electron-dense granules in the AC's cytoplasm throughout the series (Figures 11C,D), reveals that lysosomes lie, and some of them coalesce, with the plasma membrane surrounding pyknotic cells, thereby interacting directly with that of the disrupted cell in the extracellular domain.
Additional series through tripartite synapses (Figures 11E,F; Araque et al., 1999; Ventura and Harris, 1999; Witcher et al., 2010) and glomeruli (Figures 11G,H; Price and Powell, 1970b; Peters et al., 1976) in the olfactory bulb medulla confirm their distinct 3D and inner structure as defined elsewhere; although, in our sample tripartite synapses may occasionally be seen entangled by the AC's end-foot (Figures 11E,F). APs surrounding axodendritic terminals or glomerular complexes are devoid of electron-dense granules as defined by 3D reconstructions (Figures 11F–H).
Observations of oligodendrocytes (Ols) in the neuropil of assorted specimens, depict the well-known interaction of this cell type with the intact myelin envelope (Peters et al., 1976). As shown in panel Figure Figure12,12, the Ol issues two or more processes that course radially in intimate apposition to intact myelinated axons. It is noted that some Ol processes resolve into distinct rounded structures, termed here round terminal knobs (RTK). A RTK contains a likewise rounded, solitary, granule that contains a pleomorphic matrix. The latter measures 0.3–0.7 μm in diameter and it contains lipid-droplets, membranous, and/or cotton-like inclusions. Granular contents are embedded in a homogeneous matrix of a moderate electron-density. When incubated with the appropriate medium to acid phosphatases (see above) the RTK granule exhibits focal positivity due to electron-opaque deposits (Figure 12B). Bias observation of 250 RTKs, depict that most of them (87%) appose to the myelin envelope of axons in the RTK's vicinity (Figures 12C–E). These axons are invariably invested by an intact myelin envelope. Although, the solitary lysosome of the RTK frequently interacts with the limiting plasma membrane and this with the myelin envelope proper, no fusion or continuity or enzyme reactivity were noted between the former and the latter.
Lastly, microglial cells (MGs) are relatively infrequent throughout the neuropil (see Peters et al., 1976). From the thirty six MGs perikarya identified here, most of them lie next to the capillary wall (Figures 13A,E); whereas a low proportion of them is encased by the neuropil (Figures 13B,C,E,F) or, rarely, in direct apposition to neurons (Figure 13D). Generally, MGs do not interact directly with myelinated axons and, if so, envelopes are well-preserved. To emphasize is that, like the MG's cell body, organelles in distal processes are surrounded by an organelle-free area cytoplasm underscoring the cell membrane (Figures 13E,F). Interaction of the MG's presumptive lysosomes with the cell membrane is virtually absent in the normal neuropil.
Early light microscopists assumed that small clusters of myelin-like material (i.e., “myelinoid bodies”) in the white matter might represent sites of metabolic myelin turnover (Elzholtz, 1898; Jacob, 1912). This notion was strengthen with the advent of the electron microscope. To the best of our knowledge, two pioneer studies (Hildebrand and Skoglund, 1971; Hildebrand, 1977) in normal animals provide cytological evidence for the interaction of glial processes and myelinoid bodies. From these studies it was clear that glial cells internalize disrupted myelin figures. Hildebrand (1977) concluded that: “if these bodies represent a morphological expression of a myelin turnover, the oligodendroglial cell should be the main element responsible for their breakdown,” adding: “in the rabbit and Guinea pig myelinoid bodies are also present in the astrocytic cytoplasm.” Indeed, the interaction of lysosomes in the astrocyte's perikaryon with myelinoid bodies was illustrated for the first time (33a and 33b, Hildebrand and Skoglund, 1971). Recent work in adult frogs (Mills et al., 2015), has demonstrated the direct involvement of ACs in myelin remodeling during metamorphosis. Present observations support the belief that AC in the normal rat may normally be involved in myelin modeling. In fact, a group of astrocytic processes (APs) was identified in a variety of random specimens of the forebrain of healthy adult rodents. Due to the distinct structure, high incidence, and ubiquity, of this process, it was named “paraxial process” (PAP). PAP is defined as a differentiated collection of secretory organelles protruding from the AC process, which interacts with the disrupted myelin envelope of nearby axons (Figure (Figure14).14). Distinct clusters of lysosomes linked both among themselves and with the limiting cell membrane, surrounded by secretory organelles allow PAP identification. With confocal microscopy performed on specimens from the mouse model expressing fluorescence to glial fibrilar acidic protein (eGFP), presumptive PAPs exhibited co-labeling to myelin basic protein, and lysosomes (Lyso-Tracker). While a PAP was readily identified in specimens intended for routine electron microscopy, the chances of visualizing the full PAP structure were diminished due to the drawback of observing assorted sections. In series, the presence of clusters of confluent SLs proved to be an effective landmark for PAP identification, as revealed by further inspection of the consecutive sections. This approach enabled PAP identification throughout most brain areas (Table (Table1)1) suggesting a ubiquitous distribution. The PAP appears to correspond to the normal counterpart of further differentiated glial processes induced by experimental crushing (Hildebrand et al., 1994) or immune (Gotch and Löhler, 1990; Ponath et al., 2017) damage to myelinated axons. While lysosomal activity in the PAP appears to be recruited in homeostatic myelin modeling, at the present, descriptive stage, confirmatory functional studies (see Tasdemir-Yilmaz and Freeman, 2014; Ponath et al., 2017; Zorec et al., 2017) are required to define that the lysosomal contents released in the myelin envelope cannot be deleterious and cause myelin disruption itself. So far, our contention in favor of the direct involvement of the AC in myelin plasticity along the lifespan is relevant in that myelin clearance throughout adulthood has customarily been ascribed to microglia (see Safaiyan et al., 2015) and macrophages in both the normal and the diseased central nervous system (see Nave and Werner, 2014), respectively.
The vesicular contents and the common association of the PAP with the disrupted myelin envelope of contiguous axons (Figure (Figure14)14) suggests that the PAP represents an active site of myelin turnover. A PAP contains the necessary organelles for synthesis and exportation of secretory products as may be suggested from the previous observation that ACs express genes associated with cell secretion (Cahoy et al., 2008, vide infra). In the supramolecular domain, SLs in the PAP sculpt confluent units, as each one of them is interconnected with its respective homolog which in turn opens to the processes surface next to the altered myelin envelope. The reaction product to acid phosphatase within the SLs and membranes of the altered myelin, coupled with the focal reactivity observed in the intercellular space, provide complementary evidence for a lysosome-mediated myelin-lysis (see Tasdemir-Yilmaz and Freeman, 2014). While in neurons lysosomes are directly involved in extracellular dendritic spine plasticity (Chung et al., 2014; Padamsey et al., 2017), the release of lysosomal contents by ACs into the extracellular compartment, degrading materials therein, is uncommon for the normal nervous system. Previous observations of the normal, specialized, connective tissue recapitulate this process. In effect, the osteoclast, a large multinucleated cell of the normal bone tissue, lies in direct apposition to the bone surfaces exposed to the periosteum or endosteum (Ham, 1969). Upon systemic endocrine requirement, growth, or trauma, osteoclasts concur to lysate the bone matrix to be removed. The success of this process relies on the massive exocytosis of the lysosomal hydrolases next to the exposed bone matrix (Göthlin and Ericson, 1971; Matsuda, 1992; Zaidi et al., 1993; Szewczyk et al., 2013). Furthermore, lysosome occurrence, location, and fusion in our control series throughout the AC involved in engulfing presumptive apoptotic cells (Figures 11A–D; Nixon and Cataldo, 1993; Schmechel, 1999; He et al., 2006), replicate those observed for the PAP. On the other hand, astrocytic processes surrounding tripartite synapses (Figures 11E,F) and glomeruli (Figures 11G,H; Price and Powell, 1970b), both of which are functional constituents of the neuropil, no electron-dense granules are observed throughout the APs. Thus, the use of appropriate positive and negative inner controls of lysis coupled with the well-known mechanisms of intercellular matrix turnover, suggests that the PAP may normally be involved in the AC myelin remodeling.
Regarding the structural alterations of the myelin envelope of axons adjacent to presumptive sites of lysosome secretion, present and previous observations support that they correspond to sites of myelin collapse. The fact that myelin splitting may be an artifact of fixation and plastic embedment (see Peters, 1970; Möbius et al., 2016), led us to adopt the conservative criteria of considering fusion and dissolution of the myelin membranes as indicative of myelin damage (Nathaniel and Pease, 1963; Hildebrand and Aldskogius, 1976; Franson and Ronnevi, 1984; Pineas et al., 1984; Hildebrand et al., 1994; Dupree and Popko, 1999; Han et al., 2008; Peters, 2009a,b). Axon measurements and quantification (Tables (Tables1,1, ,2)2) disclosed that those associated with PAP are the most numerous throughout the neuropil (not shown). Since physiological properties of axons correlate with their diameter (Jack, 1975) we speculate that myelin-disruption compromises sets of axons associated with distinct functional task(s).
While evidences provided here and those gathered from prior work speak in favor of a lytic involvement of the PAP (see above), the possibility that other secretory product(s) ascribed to the AC are released though a lysosomal route remains to be defined. First of all, acid hydrolases represent only one-half of the total protein content of the lysosome (Softing and Klumperman, 2009), suggesting that at least some non-enzymatic products may also be secreted (Cotrina et al., 2000; Zhang et al., 2003) by this route. Second, although of numerous transmitter, trophic, homeostatic, and immunological secretory products are released by the AC (Sperri et al., 1997; Sobue et al., 1999; see Cahoy et al., 2008; Tavaggia et al., 2010; Lundgaard et al., 2014; Nave and Werner, 2014; Anderson et al., 2016), the involvement of lysosomes in this process has also been recognized. For instance, direct evidence by Zhang et al. (2003, 2007) supports that the ATP, a nucleotide with modulatory effect that is released by the AC (Queiróz et al., 1997), utilizes the lysosome as its secretory pathway (Cotrina et al., 2000). Furthermore, recent work (Padamsey et al., 2017) in pyramidal neurons puts forth the secretory pathway of lysosomes in dendritic spine plasticity. Thus, the possible involvement of PAPs as a secretory substratum should be left as an open, though testable, issue.
As expected, the number of myelinated axons varied as a function of site, being more numerous in the olfactory bulb medulla and comparable in the VMHN and deep temporal isocortex; however, the number of PAPs and RTKs was directly proportional to the axon density. Interestingly, the PAP/RTK ratio was comparable between the three brain areas. At a time that we ignore the actual functional meaning of either type of process, their coexistence and association with the disrupted and intact myelin envelope, respectively, lead to infer that they assume opposite, yet, interrelated roles in myelin turnover.
A fundamental feature of glial cells is their synergic involvement in myelin genesis. Thus, ACs influence the onset of ontogenetical myelination by induction of the precursor cell to the myelin- synthesizing Ol (see Clemente et al., 2013). Furthermore, during adulthood, both cell types are united by gap junctions for which the existence of a functional glial framework has been supported (Black and Waxman, 1988; Orthmann-Murphy et al., 2008). In this context, previous and present observations demonstrating that both cell types interact directly with the myelin envelope opens the possibility that myelin remodeling results from a likewise, dynamically balanced interaction (Clemente et al., 2013; Zorec et al., 2017). In view of the absence of myelin phagocytosis or lysosome secretion by the normal Ol, it has been accepted that myelin renewal involves the molecular sorting and trafficking at sites of the Ol-myelin continuity (see Maier et al., 2008). In consonance with this, we found no altered myelin sheaths at sites of Ol-myelin juxtaposition and PAP-RTK frequencies, although different in magnitude, paralleled to each other. Given the signed association of PAPs with the disrupted myelin envelope and that of the Ol-RTK with the intact myelin sheath, we may infer that they represent putative substrata for a balanced myelin renewal and removal, respectively. Results of our ongoing experimental trials testing possible involvement of glial processes in response to axonal damage should enhance the likelihood of the latter postulation.
All experiemntal animals were males of 10 weeks of age that were raised in a pathogen-free colonies in accordance with animal care policies in our vivarium. An ad hoc ethics committee on animal experimentation approved the manipulation and sacrifice of animals, which was carried out according to the Animal Research Committee guidelines of the Instituto de Neurobiología.
AV designed the light microscopic study, performed observations at the confocal microscope and wrote the manuscript's resulting section. VB retrieve the electron microscopic images and performed morphometry and 3D reconstructions. CL realized the inmuno- and histo-chemical techniques and processed images and videos from confocal microscopy. JL, designed the study, performed observations, illustrations, and wrote the final version of the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The transgenic hGFAP-GFP mouse line was a generous gift from Dr. Helmut Kettenmann (Max Delbrück Center for Molecular Medicine, Berlin). Authors thank Gema Martínez Cabrera, Lourdes Palma, Elsa Nydia Hernández-Ríos, Martín García, Rafael Olivares for their technical assistance, and Jessica González Norris for the thorough revision of our manuscript.
The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fnana.2017.00075/full#supplementary-material