Chromaffin cells are neuroendocrine cells that, like sympathetic neurons, derive from sympathoadrenal precursors in the neural crest, a transient structure formed early in embryonic development along the dorsal surface of the neural tube (Coupland and Weakley, 1968; Langley and Grant, 1999). The anatomical arrangement of chromaffin tissue is quite different throughout vertebrate classes. In mammals, it populates the adrenal gland and sympathetic paraganglia, such as the carotid body and the “organ of Zuckerkandl” (Unsicker et al., 2005). In adrenal glands, chromaffin cells accumulate in the inner part of the organ, the so-called medulla, being surrounded by a coat of steroidgenic tissue of mesodermal origin that forms the adrenal cortex. In birds, chromaffin cells localize to adrenal glands but there is no division into medulla or cortex and the chromaffin and steroidogenic cells are intermingled (Coupland, 1965). In reptiles, the relationship between steroidogenic chromaffin cells and kidney varies according to different orders; in some of them the chromaffin cells have a similar distribution to those observed in birds (Lofts, 1978; Accordi et al., 2006). In the classes of amniotes lacking a true cortex and in anamniotes, the steroidogenic cells are generally called “interrenal cells” (Gallo and Civinini, 2003). Various mingling of chromaffin and interrenal cells is observable in amphibians. Here, small dispersed islets of either type of cells are more or less combined together on the ventral surface of the mesonephric kidneys (Grassi Milano and Accordi, 1986; Accordi, 1991). The most primitive arrangement occurs in fish, where chromaffin and steroidogenic cells are mostly separated and scattered inside the renal tissue, forming the so-called adrenal homolog (Chester Jones and Mosley, 1980; Mastrolia et al., 1981, 1984; Gallo et al., 1993; Abelli et al., 1996; Gallo and Civinini, 2005). Thus, an evolutionary trend toward aggregation of chromaffin and steroidogenic cells into compact encapsulated organs, the adrenal glands, is evident throughout vertebrate classes (Gallo and Civinini, 2003). Clusters of extra adrenal chromaffin cells, however, can be recognized in different organs. In agnathans, the most primitive fishes such as hagfishes and lampreys, they have been found in the heart walls and within the great veins (Perry et al., 1993). In the more advanced species of fishes, they are recognizable in the ovaries and corpuscles of Stannius (Thakur and Shrivastava, 1975; Unsicker et al., 1977). In amphibians, reptiles, and birds, extra adrenal chromaffin cells may be observed inside nerve bundles and ganglia (Gallo and Civinini, 2003).
Although the anatomical arrangement of the chromaffin tissue is very dissimilar among the various classes of vertebrates, chromaffin cells as such appear highly conserved during evolution. They exhibit indeed the typical ultrastructural pattern of neuroendocrine cells, being typified by plentiful cytoplasmic granules showing different electron density according to the main type of catecholamine stored in them, i.e., adrenaline or noradrenaline. In most vertebrates, two subsets of chromaffin cells may be distinguished under the electron microscope after fixation in glutaraldehyde and osmium tetroxide: the adrenaline-containing cells, which exhibit rounded, moderately electron-dense secretory granules, and the noradrenaline-containing cells, which present more pleomorphic, strongly electron-dense granules (Gallo and Civinini, 2003). In the primitive amphibians, the urodeles, in some species a single type of chromaffin cell can be distinguished (Accordi, 1991); moreover, in Triturus, it undergoes an annual cycle with seasonal variations in the adrenaline/noradrenaline granule ratio (Laforgia and Capaldo, 1991).
Biochemical analysis of chromaffin granules has revealed that catecholamines colocalize with a number of other substances, such as chromogranins, ATP, and neuroactive peptides, in a complex storage package (Winkler, 1993). These molecules are secreted from chromaffin granules in response to stressful stimulations and, once released, they can act either locally in a paracrine manner on other chromaffin cells and steroidogenic cells next to them or enter the bloodstream to act on distant targets in an endocrine fashion (Coupland, 1965).
In all species examined, secretion from chromaffin cells has been postulated to occur by exocytosis (Burgoyne, 1991). This model of cell secretion implicates a process of fusion of the granule-limiting membrane with the plasma membrane, followed by discharge of granule constituents outside the cell. Fusion may be either complete and irreversible, leading to full collapse of the granule membrane into the cell membrane resulting in entire release of granule contents (“full fusion” exocytosis) or transient and reversible, allowing partial release of cargo constituents and immediate granule recapture back from the cell membrane (“kiss-and-run” exocytosis) (Burgoyne and Morgan, 2003). Recently, ultrastructural evidence of a particulate pattern of cell degranulation, called piecemeal degranulation (PMD), has been recognized in the chromaffin cells from mouse, rat, and human adrenal medulla (Crivellato et al., 2003a, 2004, 2005a, 2006). This ultrastructurally defined secretory model implicates discrete release of granule particles from storage granules without granule fusion to the plasma membrane (Dvorak, 1991). As documented in studies on basophils and mast cells, secretion occurs by translocation of shuttling loaded vesicles from the granule compartment to the cell surface. Small vesicles filled by secretory cargoes bud from granules, move to the cell periphery, and fuse with the plasma membrane, thus releasing small quanta of secretory material (Dvorak, 2005).
As PMD appears to be a general secretory process in cells involved in endocrine/paracrine functions (Crivellato et al., 2003b), we wondered whether this pattern of cell degranulation was conserved during evolution. In the present study, we undertook an ultrastructural morphometric investigation on chromaffin cells in the adrenal homolog of Aphanius fasciatus, an euryhaline teleost living in saltpans. Teleosts are the most evolutionarily advanced fishes; in these animals, the adrenal homolog represents a system involved in adaptive responses to different kinds of stressors. In this study, we have quantified a series of ultrastructural parameters, comprehensively recognized as highly distinctive for PMD, both at rest and after experimentally induced stressful condition, i.e., after animal transfer in dechlorinated tap water. The direct transfer of fish from salt water to freshwater causes an immediate strong shock due to handling and osmotic stress. Former study performed on Aphanius (Mastrolia and Gallo, 1989) has already evidenced that this stress affects the adrenal homolog as evidenced by cytological modifications of some organelles as nucleus, mitochondria, and endoplasmic reticulum of the interrenal cells. Our ultrastructural morphometric study relied on a series of well-established notions regarding the fine mechanism of PMD secretion. As widely accepted, such process implies a trend toward granule emptying associated with both granule enlargement and mobilization of vesicles from the granule compartment to the plasma membrane and vice versa. Granule changes indeed are closely related to the movement of shuttling vesicles, which actually occurs during the secretory progression: on the one hand, electron-dense vesicles filled by cargo material bud and detach from the granule-limiting membrane; on the other hand, clear vesicles are captured from the cell membrane. The former type of vesicles is responsible for mobilization of small packets of cargo material from chromaffin granules and for their transport to the cell surface. Here, vesicles fuse with the plasma membrane, allowing for release of little amounts of granule material. The latter type of vesicles is retrieved from the plasma membrane, swarms to the granule compartment, and fuses with the granule-limiting membranes. Being insertion of clear vesicles to the granule membrane generally more conspicuous than detachment of budding vesicles, the resulting outcome is that the granule gradually reduces its electron-dense content and increases its size. This ultrastructural schema was indeed what we found in our study. Results indicate that chromaffin cells in the adrenal homolog of Aphanius fasciatus express PMD in response to severe osmotic stress. In evolutionary terms, these data allow us to conclude that PMD appears to be a highly conserved mechanism of granule discharge.
MATERIALS AND METHODS
Adult specimens of Aphanius fasciatus, 5.5–6 cm in length, collected from the saltpans of Tarquinia (in the outskirts of Rome, Italy), were kept in aquariums as previously reported (Mastrolia and Gallo, 1989) for a week's acclimatation in the water of origin, whose salinity was 3.7%. Specimens of both sexes were transferred directly from salt water to dechlorinated tap water maintained at the same temperature. Each tank of 10 l of capacity contained 20 specimens that were killed at 2 and 48 hr from the beginning of the experiment. The controls were kept in the water of origin. For each stage, 10 specimens were used.
Electron Microscopy Procedure
The fishes were quickly decapitated. The head kidneys were excised and fixed for 2 hr at 4°C in 2% glutaraldehyde solution in 0.1 M cacodylate buffer, pH 7.4 (410 mOsm), suitable for marine teleosts (Saito and Tanaka, 1980). Samples were postfixed for 1 hr in 1% osmium tetroxide solution in Millonig buffer, dehydrated, and embedded in Araldite. Plastic 1 μm sections were stained with toluidine blue and examined by light microscopy. Ultrathin sections were contrasted with uranyl acetate and lead citrate solution and examined in a Philips (Eindhoven, The Netherlands) CM 12 electron microscope at 80 Kv.
Quantitative Evaluation of Electron Microscopic Data
To assess whether PMD was part of a secretory response to severe osmotic and handling stress in the chromaffin cells of Aphanius fasciatus adrenal homolog, we performed morphometric analyses both in control animals and animals maintained in dechlorinated tap water for scheduled intervals (2 and 48 hr). Ultrastructural quantitative investigations included determination of the frequency of electron-dense granules with closely adhering limiting membrane (type 1 granules), granules with reduced electron-dense content surrounded by a clear halo (type 2 granules), and empty or almost completely empty containers (type 3 granules). As PMD generally causes progressive enlargement of the granule chamber paralleled by reduction of the granule content, we calculated the mean diameter and average area of the whole granule as well as the mean area of the inner granule electron-dense content. From these data, we calculated the percentage of granule occupation, both in control and in treated samples. In addition, we also evaluated the number of small, membrane-bound, cytoplasmic vesicles having a diameter of 30–100 nm and presenting either electron-dense or -lucent contents. Some of these vesicles appeared to bud from the granule-limiting membrane. Therefore, we also calculated the proportion of secretory granules with such budding features.
Two chromaffin cells from each control fish and two chromaffin cells from each animal exposed to osmotic stress for 2 and 48 hr were selected at random and photographed at 8,000× and 30,000× magnification. Enlarged prints (18 × 24 cm) were analyzed for morphometric evaluation. The number of type 1, type 2, and type 3 granules was counted in the group of 60 low-magnification micrographs and expressed as percent of the total number of granules/containers. The mean diameter and the average area of all granules as well as the mean area of the inner granule electron-dense content were calculated in the group of 60 high-magnification micrographs using a BDS-Image software package system interfaced with a Macintosh computer. A number of 30–100 nm cytoplasmic vesicles were evaluated for each printed cell in the same group of 60 high-magnification micrographs and related to the net cytoplasmic area (expressed in square micrometers) of the corresponding cell. The net cytoplasmic area was calculated using the same software system, which allowed subtraction of the areas occupied by chromaffin granules, nucleus, Golgi region, and mitochondria from the total cytoplasmic area of a given cell. Vesicle frequency in chromaffin cells of control and stressed fishes was eventually expressed as average number of vesicles per net cytoplasmic area. The number of granules with budding features was also calculated in the group of 60 high-magnification micrographs and expressed as percent of the total number of granules/containers.
Data were expressed as mean ± SD and analyzed statistically using Student's t-test. Values of P < 0.05 were considered significant. The software package SPSS (Statistical Package Social Sciences; SPSS, Chicago, IL) was used for data analyses.
The chromaffin tissue of the adrenal homolog of Aphanius fasciatus was found near the postcardinal veins and in the surrounding hematopoietic tissue of the head kidney. In control samples, two types of chromaffin cells were recognizable in relation to the aspect of chromaffin granules. The first type (class 1 cell) contained rounded to oval, apparently empty granules with an average diameter of 162.4 ± 18.3 nm (range, 106–234 nm). Very few granules presented small aggregates of moderately electron-dense material (Fig. 1A). The second type (class 2 cell) has a predominance of pleomorphic, highly electron-dense granules often displaying an eccentric core separated from the enclosing membrane by a clear halo of variable width and containing sometimes finely granular material (Fig. 1B). The diameters of these granules ranged from 104 to 314 nm, with an average of 184.3 ± 14.8 nm and an average area of 26,663.7 ± 171.9 nm2 (Fig. 2). As the morphology of the granules was similar to that observed in the chromaffin cells of other teleost fishes (Gallo and Civinini, 2003), class 1 cells were interpreted as adrenaline-containing cells and class 2 cells as noradrenaline-storing cells. The electron-lucent aspect of the granules in class 1 cells might depend on the fact that they contained a more diffusible ground substance than adrenaline granules in mammals and other vertebrates. Due to the particularly frail structure of granule matrix in class 1 cells, which caused disappearance of granule electron density, we relied solely on class 2 cells for quantitative morphometric analysis. In resting condition, the proportion of type 1 granules in these cells was 25.6% ± 1.4% of the total granule repertoire. Type 2 and 3 granules accounted for 64.1% ± 3.6% and 10.3% ± 1.1% of total granules, respectively (Fig. 3). The electron-dense granule content occupied a mean area of 18,226.4 ± 164.3 nm2 (Fig. 2) with a percentage of occupation of 68.3%. Some vesicles, 30 to 100 nm in diameter, were seen in the cytoplasm of these cells, either free or adhering to granules. The number of these vesicles per net cytoplasmic area was 0.38 ± 0.06. Elongated projections or outpatches emerging from the perigranule membranes were occasionally recognized. The proportion of secretory granules with such budding features was found to be 3.2% ± 1.8% of total granules (Fig. 4).
After 2 Hr
The ultrastructure of class 2 chromaffin cells changed remarkably 2 hr after animal transfer in dechlorinated tap water (Fig. 1C). The overall granule size appeared in creased, being the mean granule diameter 255.4 ± 21.3 nm (range, 139–453 nm) and the mean granule area 51,204.9 ± 356.1 nm2 (P < 0.01 vs. control; Fig. 2). In addition, the granule population was more pleomorphic. The proportion of type 1 granules decreased to 8.0% ± 0.8% of total granules (P < 0.01 vs. control), while types 2 remained substantially unchanged (62.7% ± 5.2% NS) and type 3 granules augmented to 29.3% ± 3.4% (P < 0.01 vs. control; Fig. 3). The mean area of the electron-dense content inside granules was 14,376.3 ± 224.5 nm2 (P < 0.01 vs. control; Fig. 2), with a percentage of occupation of 28.1%. Besides being quantitatively reduced, the granule constituents often appeared less electron-dense, with peripheral or internal areas of content losses (Fig. 5A–C). Although granules presented often irregular distorted shape, still they maintained their individual close structure and exocytoses were exceptionally observed (Fig. 5D). Remarkably, plentiful membranous vesicotubular structures were seen in the cytoplasm of these cells (Fig. 5C). Vesicles presented smoothed surface and a diameter ranging from 30 to 100 nm. Analysis at higher magnification revealed that the majority of them appeared electron-lucent but a relevant proportion contained particles similar in structure and electron density to those that made up the core of chromaffin granules (Figs. 5A–C and 6). Many vesicles could be recognized free in the cytoplasm but a lot were found adhering to granules. Remarkably, some vesicles filled with cargo material were seen fused with the plasma membrane and apparently open to the cell exterior (Fig. 6B). Morphometric examination revealed that the number of vesicles per net cytoplasmic area was 1.46 ± 0.36, a feature significantly different from control (P < 0.01; Fig. 4). Interestingly, vesicles appeared to bud from the limiting membranes of chromaffin granules. Budding vesicles had chiefly the shape of pear-like evaginations attached to the granules by narrow elongated necks. Most of them appeared filled by the same strongly electron-dense secretory material constitutive of the substance stored inside the granules (Fig. 6A and B). Remarkably, granules with budding vesicles were almost completely empty and the residual secretory material characteristically accumulated close to the base of the budding structures and inside the evaginations themselves. In other instances, budding vesicles appeared as moderately electron-dense surface outpouches (Fig. 6C). Taken together, budding features were recognized in 16.2% ± 2.8% of total granules (P < 0.01 vs. control; Fig. 4).
After 48 Hr
At this stage, the ultrastructural morphology of class 2 cells underwent further changes (Fig. 1D). Type 1 granules were virtually absent, being 1.3% ± 0.4% of total granule compartment, while the percentage of type 2 granules decreased to 38.5% ± 6.3% and that of type 3 granules increased to 60.2% ± 12.6% (P < 0.01 vs. controls in all cases; Fig. 3). The overall granule size was further increased, the mean granule diameter being 321.4 ± 27.4 nm (range 149 to 528 nm) and the mean granule area 81,088.9 ± 589.3 nm2 (P < 0.01 both vs. control and 2 hr values; Fig. 2). The mean area of the electron-dense content inside granules was 6,852.4 ± 103.6 nm2 (P < 0.01 both vs. control and 2 hr values; Fig. 2), with a percentage of occupation of 8.4%. Despite a clear trend to size enlargement, granules still maintained their close structure and exocytoses were never seen at this stage. Therefore, the cytoplasm of chromaffin cells appeared as replenished by empty or almost completely empty granule containers (Fig. 1D). Analysis at higher magnification revealed a dramatic reduction of the number of electron-dense or clear vesicles per net cytoplasmic area, with values returning to control limit (0.41 ± 0.05 NS vs. control; P < 0.01 vs. 2 hr value; Fig. 4). A similar trend was observed as to the occurrence of granule surface projections and outpatches. Indeed, the percentage of granules showing budding features decreased to 3.9% ± 2.3% of total granules (NS vs. control; P < 0.01 vs. 2 hr value; Fig. 4).
It is well established that fish exposure to a number of stressors determines rapid incretion of adrenaline and noradrenaline into the blood stream (Reid et al., 1998). Stressing conditions must be severe, almost brutal, to effect blood catecholamine increase in these animals. In deed, the outburst of an acute humoral adrenergic response in fishes occurs under conditions of extreme physiological impairment (Perry and Bernier, 1999). In general terms, fish stressors are defined as stimuli causing an integrated adaptive response and include natural changes in the environment such as physical alterations of salinity, temperature, turbidity, water pollution by chemicals, low pH and heavy metals, as well as a series of experimental conditions including air exposure, handling, exhausting exercise, capture, confinement, artificially induced hypoxia or hypercapnia, etc. (Wendelaar Bonga, 1997). The adrenergic activation results in secondary effects mainly on the circulation, osmoregulation, and energetics.
In the present investigation, we have analyzed the ultrastructural changes of chromaffin cells in the adrenal homolog of the teleost fish Aphanius fasciatus during adaptation response to severe osmotic stress. Teleosts are bony fishes comprising both freshwater and marine species. In these animals, the adrenal homolog is composed by aminergic chromaffin and interrenal stroidogenic cells located mostly inside the head kidney that, in this taxon, generally has a hematopoietic function (Gallo and Civinini, 2003). Although the arrangement of interrenal cells is most variable, distribution of chromaffin cells is remarkably constant. These indeed are always situated within the walls of the posterior cardinal veins and/or their tributaries. In general terms, the results of our study indicate that PMD in chromaffin cells may be part of a secretory response to severe osmotic stress in these animals. There is good evidence that in teleosts, plasma catecholamines increase the permeability of gill to water (Pic et al., 1974) and also act on ionic transfer across the gills (Girard and Payan, 1980). Thus, most of the effects of stress on osmotic balance are mediated by adrenergic secretion that, in a short time, results in water or ionic overload, depending on whether fishes are held in fresh water or sea water.
PMD is regarded as a discrete, particulate kind of cell secretion that effects little by little discharge of secretory material from storage granules without granule opening to the cell exterior (Dvorak, 1991). Indeed, PMD differs from exocytosis insofar as granules never fuse with the plasma membrane but maintain their close structure during the entire course of the releasing process. This pattern of cell degranulation was initially described in basophils, mast cells, and eosinophils, and functional studies in these cells, coupled with electron microscopic analysis, soon demonstrated that granule emptying was effected by a mechanism mediated by an outward flow of shuttling vesicles loaded with cargo material. According to this model, vesicles containing bits of granule contents bud from the perigranule membrane, move through the cytoplasm, and fuse with the plasma membrane, leading to content discharge (Dvorak, 2005). Endocytic vesicles are retrieved from the plasma membrane, traverse the cytoplasm, and fuse with granules in a closely coupled inward flow. If the rate and amount of vesicular traffic are balanced, granule containers empty in a piecemeal fashion but maintain a constant size. If, on the other hand, the inward flow of the endocytic vesicles exceeds the outward flow of the exocytic vesicles, the granule chambers become enlarged. The latter event is what generally occurs. Thus, granules release their secretory constituents without direct fusion with the plasma membrane and eventually transform into large empty containers. Secretory material enclosed inside granules undergoes a unique step-by-step process of patchy losses leading to characteristic piecemeal pictures when viewed by transmission electron microscopy (Dvorak, 1991). Hence, the ultrastructural hallmark of a cell undergoing PMD is threefold. First, a clear trend to loss of secretory material from granule stores without granule opening to the cell exterior; second, a tendency to enlargement of granule size; third, an early increase of cytoplasmic vesicles associated with budding structures emerging from the granule surface.
Ultrastructural features representative of the occurrence of a PMD model of cell secretion has recently been documented in dense-core granules of neuroendocrine cells and neurons (Crivellato et al., 2002, 2005b). In particular, a series of granule and cytoplasmic changes highly indicative of PMD has been reported in chromaffin cells of the mouse, rat, and human adrenal medulla (Crivellato et al., 2003a, 2004, 2005a, 2006). The present study adds further to this concept by describing compelling evidence of PMD morphologies in chromaffin cells of Aphanius fasciatus in response to extreme osmotic stress.
Our morphometric data support indeed the occurrence of a degranulation process that can be reconciled with a PMD conceptual scheme. After osmotic stress exposure, we observed a secretory pathway typified by the following distinctive traits. First of all, a progressive loss of content material from chromaffin granules with preservation of a close granule structure. This point is exemplified by reduction of both the mean electron-dense area inside granules and the mean percentage of granule occupation during the experimental course. The former indeed decreases from 18,226.4 ± 164.3 nm2 in controls to 14,376.3 ± 224.5 nm2 and 6,852.4 ± 103.6 nm2 after 2- and 48-hr osmotic stress, respectively. The latter diminishes from 68.3% to 28.1% and 8.4%, in the same order. The second key feature is a parallel trend to granule enlargement. This point is illustrated by the progressive increase of the mean granule diameter and granule sectional area during the experimental steps. The former indeed enhances from 184.3 ± 14.8 nm in controls to 255.4 ± 21.3 nm after 2 hr and 321.4 ± 27.4 nm after 48 hr. The latter rises from 26,663.7 ± 171.9 to 51,204.9 ± 356.1 and 81,088.9 ± 589.3 nm2, in the same order. The third issue is the formation of numerous budding projections filled by electron-dense secretory material from the surface of chromaffin granules during the first 2 hr, followed by return to control value after 48 hr. In fact, the proportion of secretory granules showing budding features changes from 3.2% ± 1.8% of total granules in control samples to 16.2% ± 2.8% and 3.9% ± 2.3% in stressed animals after 2 and 48 hr, respectively. The fourth key point is the dramatic increase of 30 to 100 nm in diameter, membrane-bound vesicles during the first 2 hr after exposure to osmotic stress. These vesicles appear either electron-dense or clear and are recognizable either free in the cytoplasm or attached to granules or fused with the plasma membrane. Their number per net cytoplasmic area changes from 0.38 ± 0.06 in controls to 1.46 ± 0.36 after 2 hr. Remarkably, these features drop again to control level after 48-hr stress treatment (0.41 ± 0.05). At last, the endpoint of this degranulation pathway is the formation of large nonfused cytoplasmic containers almost completely devoid of secretory constituents. In fact, the proportion of this kind of granules (type 3 granules) changes from 10.3% ± 1.1% in controls to 29.3% ± 3 6.4% and 60.2% ± 12.6% after 2 and 48 hr, respectively. Notably, these features are again indicative of a trend to cargo discharge from granules. This assumption is also inferred by the progressive disappearance of compact electron-dense granules with adhering limiting membrane (type 1 granules), whose proportion decreases from 25.6% ± 1.4% in controls to 8% ± 0.8% and 1.3% ± 0.4% after 2 and 48 hr, respectively. Taken on the whole, our ultrastructural findings coupled with morphometric analysis are strongly indicative of a degranulation process effected by vesicle-mediated mechanism, which is what PMD actually is.
An important issue that emerges from this study is the significance of PMD in an evolutionary perspective. PMD has already been recognized in a series of cells belonging to mammals. Indeed, early studies by Dvorak (1991, 2005) concerning PMD in basophils, mast cells, and eosinophils were carried out in human as well as mouse and guinea pig samples. Studies on eosinophil PMD in animal models of asthma were performed in rats and cats (Erjefalt et al., 1998, 2001). Subsequent investigations on adrenal chromaffin cells, enteroendocrine cells, and neurons expressing PMD secretory patterns were conducted in human, mouse, and rat specimens. The first evidence of PMD in nonmammalian tissues was provided in a study concerning the neuroendocrine cells in the developing thymus of the chick embryo (Crivellato et al., 2005c). Different kinds of intrathymic granulated cells exhibited indeed the characteristic repertoire of PMD ultrastructures. These findings were further corroborated by the discovery that mast cells in the developing chicken chorionallantoic membrane as well as in the thymic anlagen also express electron microscopic features indicative of PMD (Crivellato et al., 2005c, 2005d). Here we demonstrate that chromaffin cells in the adrenal homolog of a teleost fish are also able to organize a PMD reaction in response to violent stressful stimulation. These findings emphasize the significance of PMD in the physiological response of chromaffin cell secretion and, in addition, highlight the basic relevance of PMD as a general mechanism of cell secretion. Remarkably, this mechanism appears to be accurately conserved during evolution.
An interesting issue concerns the molecular composition of electron-dense or clear vesicles and their constitution in comparison with granule content. It is conceivable that electron-dense vesicles, which appear loaded with material often identical to that stored in chromaffin granules, may contain the same kind of matrix and mediator constituents. Conversely, clear vesicles retrieved from the plasma membrane by endocytosis would lack typical components of granule-derived vesicles. A fascinating question pertains to the possibility that PMD might allow for a selective release of distinct molecular components stored inside granules. As these organelles contain a myriad of molecular constituents, among which there are different hormones and peptides, this conjectural mechanism would permit a fine control of discharged mediators by chromaffin cells and would fit well into the functional schema of a paracrine kind of cell secretion. We argue that PMD may alternatively work as a secretory mechanism that amplifies exocytosis in the course of endocrine responses or conversely function as a discrete and possibly selective pathway for mediator release in a paracrine context. Further electron microscopy immunolabeling investigations, coupled with molecular analysis of granule and vesicle contents, will hopefully shed light on this intriguing issue.
This study was supported by MIUR local funds to the Department of Medical and Morphological Research, Anatomy Section, University of Udine, and Department of Animal and Human Biology, University of Rome “La Sapienza.”