Numerous experimental data have shown the existence of paracrine relationships between the steroidogenic and the chromaffin tissues. In mammals glucocorticoids stimulate the activity of the tyrosine hydroxylase (TH) enzyme, the rate-limiting enzyme in catecholamine biosynthesis [14,15] and the phenylethanolamine-N-methyltransferase (PNMT) enzyme, the last of the catecholamines biosynthetic pathway, converting NE into E [16,17,18,19,20] (Figure 2). Aldosterone elevates PNMT activity too, but with less power than glucocorticoids [21].

Glucocorticoids are believed to upregulate catecholamine synthesis, storage, and secretion [20]. Moreover, they are essential for postnatal maintenance of adrenal and extra-adrenal chromaffin cells [22]. In birds, the adrenocorticotropic hormone (ACTH) influences TH enzyme [15], increases PNMT activity and the conversion of NE into E via intraglandular glucocorticoid production [19,23]. In reptiles, as in mammals, a stimulating action of glucocorticoids on the enzyme PNMT has been found [24,25,26] whereas in fish and amphibians such stimulating action has not been observed [1,27,28]. In turn, catecholamines and the neuropeptides co-released with them from the chromaffin tissue, may influence the synthesis and release of the steroid hormones from the steroidogenic tissue in a paracrine way. This paracrine relationship has been observed in many vertebrates, including fish and amphibians [1,14]. The resulting crosstalk that settles between the two components of the gland may be important to synchronize the response of this gland to stress [29]. From this point of view, the evolutionary tendency to the progressive coalescence of steroidogenic and chromaffin tissues has optimized the ability of the organisms to respond to stressful situations and supports the widely accepted opinion that the progressive concentration of the glandular tissues has been a selectively favorable process [30].

The morphology of the chromaffin and the steroidogenic tissues can be studied with techniques utilized in light and electron microscopy. As for chromaffin tissue, it should be noted that two situations have been observed: the presence of two types of chromaffin cells, one of which contains NE and the other E, or the presence of only one type of chromaffin cell that produces both catecholamines. The first type of organization is generally present: (a) in fishes; (b) in anurans amphibians, where NE and E cells are usually intermingled; (c) in reptiles, where the distribution of NE and E cells is different in the different species; and (d) in birds, where the chromaffin cells are usually mixed and do not have any preferential location. As regards urodele amphibians, two subtypes of chromaffin cells were observed [4], whereas in Triturus carnifex, a single type of chromaffin cell was found, producing both NE and E, correlated to the environmental temperature [31,32] and the reproductive cycle [33]. In mammals, generally NE and E cells have been found, but their relative quantity and their topographical location are different in different species. For example, about 75% of the chromaffin cells in humans and bovine animals, and 80–85% in the rat, are E cells. In the rat, but not in bovine animals, the NE cells are generally located at the boundary between the cortex and the medulla, whereas in other species the E cells are peripherally located [2,11,34,35,36].

The two types of chromaffin cells can be distinguished with histochemical staining, using a fixative based on potassium dichromate, the Wood’s fixative, a mixture of 2.5% potassium dichromate and 1% sodium sulphate (buffered at pH 4.1 with 5 M acetate buffer) and 10% formaldehyde. Indeed, the oxidation with potassium dichromate produces brown insoluble pigments from both NE and E, which can be subsequently distinguished using two histochemical stains: the Wood stain and the Giemsa stain [37,38]. The Wood stain, a mixture of eosin and aniline blue, buffered at pH 4 with 5 M acetate buffer, stains the NE cells gold and the E cells orange red. A similar staining is also obtained with a trichromic histological stain, the Mallory staining, with which the NE cells appear gold yellow, and the E cells appear red. Moreover, this staining has in addition the advantage of staining well also the steroidogenic tissue, that appears light pink. Finally, the Giemsa solution, modified according to Pearse [38], a mixture of eosin blue and methylene blue, stains the NE cells dark green and the E cells light green [25,26] (Figure 3).

When only one type of chromaffin cell, synthesizing both catecholamines, is present, the histochemical stains produce a mixed staining. In that case, the best method to study the chromaffin cells is electron microscopy, with which the simultaneous presence of the two types of granules in the same cell, as well as the presence of separate NE and E chromaffin cells, can be observed. Indeed, such a distinction can be obtained by using a first fixation with glutaraldehyde followed by a post-fixation with osmium tetroxide. Glutaraldehyde induces the formation of an insoluble complex with NE, which will be subsequently stained with osmium tetroxide. On the contrary, E is largely lost during the fixation and dehydration because it does not form such a complex. Therefore, E granules show only a light electron-density [39]. Under the electron microscope, the NE granules appear polymorphic and very electron-dense, with a core closely adhering to the limiting membrane, whereas the E granules are roundish, moderately electron-dense and characterized by a clear halo between the core and the limiting membrane [40,41] (Figure 4). A study, performed in E-deficient mice, generated by knocking out PNMT gene, has recently shown that the E granules still retain their shape and general appearance, despite the lack of E [42].

However, the chromaffin cells can also be studied with many other histological, histochemical and immunocytochemical methods [43,44]; for example, the NE cells can be identified through the immunopositivity to the enzyme, tyrosine hydroxylase (TH), converting l-tyrosine to l-3,4-dihydroxyphenylalanine (l-DOPA) (Figure 2), and by staining with Harris hematoxylin after treatment with citrate buffer at pH 6 [36]. The E cells possess the enzyme PNMT, converting NE into E, and therefore can be identified through the immunopositivity to this enzyme. An exception is represented by birds, in which all chromaffin cells possess the enzyme PNMT, but it is active only in the E cells. Therefore, in birds the PNMT enzyme cannot be used to discriminate the E cells from the NE cells [19].

Under light microscopy, the steroidogenic tissue is characterized by cells with a generally weakly stained and finely spongy cytoplasm, due to the presence of lipid droplets, composed of esters of cholesterol, the precursor of steroid hormones (Figure 3). At the electron microscopic level, the steroidogenic cells are generally characterized by a poor rough endoplasmic reticulum and few free ribosomes while they are rich in smooth endoplasmic reticulum, Golgi complex and mitochondria, having inner membranes forming tubular or vesicular cristae. Moreover, the steroidogenic cells have lipid droplets, the quantity of which varies according to the species and the functional stage of the tissue [45] (Figure 4). In non-mammalian vertebrates, the steroidogenic tissue is usually arranged in cords forming interconnected networks, without preferential organization. However, in birds and in some reptiles, just below the connective capsule, the cords of steroidogenic cells assume a ring or basket arrangement at the periphery of the gland, and parallel more deeply, suggesting a functional zonation, that has been observed in birds [8,46,47] and reptiles, in which the existence of several sub-populations of adrenocortical cells is supposed [8,48]. In mammals, the adrenal cortex is organized in an outer glomerular zone, mainly producing mineralocorticoid hormones; a deep fasciculata zone, mainly producing glucocorticoid hormones, and an inner reticularis zone, mainly producing androgen hormones. However, this standard model of adrenocortical zonation may be variable; in the rat, for example, a white, or undifferentiated zone (ZU) was observed between glomerular and fasciculata zones. The ZU zone was considered a stem/progenitor cell zone [49], but its presence in other species is not clear [50].

The possibility of distinguishing NE cells from E cells by light microscopy makes it possible to calculate the numerical ratio between NE and E cells, or NE/E cell ratio. This ratio reflects the proximity relationships between the steroidogenic and the chromaffin tissues, since, in mammals, birds and reptiles, glucocorticoids stimulate the activity of the enzyme PNMT, that methylates NE converting it into E (Figure 2). Therefore, high E levels indicate an efficient methylation process due to a close contact between the chromaffin and the steroidogenic cells, and therefore high glucocorticoid levels available to stimulate the enzyme PNMT [29]. The values of this ratio are very variable in birds and reptiles, and attempts were made to attribute a phylogenetic significance to its variations. For example, it was found that birds with more primitive ancestry, as cormorant and egret, had more NE, while recently evolved species, as passerine birds, had more E; a NE/E cell ratio around 1/1 was typical of species, as pigeon and cuckoo, having an intermediate evolutionary position [51]. However, this opinion is no more supported by recent bird phylogenetic reconstructions [52]. Also in reptiles, the NE/E cell ratio is highly variable. For example, in Squamata (lizards, snakes and amphisbaenians), high values of this ratio correspond to a high degree of separation between the steroidogenic and the chromaffin tissues, whereas low values of this ratio correspond to a higher degree of admixing of the two tissues. As for birds, a correlation between the value of NE/E cell ratio and the phylogenetic history of different species was assumed [2,51]. This correlation was confirmed by numerous biochemical, karyological and paleontological data [53,54,55,56]. However, in recent years the hypotheses on Squamata phylogeny have changed radically, especially if both morphological and molecular data are considered [57,58,59,60,61,62]. Therefore, the hypothesis of a correspondence between the value of the NE/E cell ratio and the greater or lesser antiquity of the species in Squamata should be verified with further studies.

Before describing the adrenal gland of reptiles, it must be emphasized that, from a systematic point of view, there is no agreement on the taxon Reptilia, since morphological and molecular studies have questioned its composition and its nomenclature [63]. The discussion on the systematic position and the nomenclature of this taxon goes beyond the scope of this paper, in which the traditional and better-known term of “reptiles” will still be used. Moreover, it must be said that in this review are considered also papers published many years ago, relating to species whose nomenclature has changed over time. Where possible, the current nomenclature in brackets right after the specific name, utilized in the mentioned papers, will be quoted [64].

In reptiles, the steroidogenic and the chromaffin tissues are associated to form a gland in close relationship with the gonads and the genital ducts (Figure 1), except for chelonians, where the gland is related to the ventral surface of the kidney. The degree of admixing of the two tissues is highly variable and it is not possible to define a scheme valid for all reptiles [1,2,7]. Indeed, in Chelonia and Crocodilia, a close intermingling of the steroidogenic and the chromaffin cells, without any concentration of chromaffin tissue on the dorsal region of adrenal gland, can be observed [6,65,66]. The arrangement of the adrenal gland of Chelonia, recently considered diapsid reptiles (for review see [63]), is like that of amphibians, forming a stripe on the ventral surface of the kidney, whereas the arrangement of the adrenal gland of Crocodilia, characterized by strands of chromaffin cells intermingled with interrenal cords, is like that of birds, where the chromaffin tissue forms clumps or strands of cells mixed with blood vessels and interrenal steroidogenic cords, both in subcapsular zone and inner part of the gland [6]. In Rhynchocephalia and Squamata, the chromaffin tissue is mainly distributed on dorsal part of the gland. The adrenal gland of Rhynchocephalia has a parenchyma of steroidogenic tissue also containing islets of chromaffin cells; a major part of chromaffin tissue forms a dorsal mass that sends expansions into the steroidogenic parenchyma, whereas some clusters of chromaffin cells are also present on the ventral surface of the gland [2,6,67].

In Squamata (lizards, snakes and amphisbaenians), the general arrangement of the adrenal gland is like that of Rhynchocephalia; indeed, the adrenal gland shows a steroidogenic parenchyma and a dorsal mass made of chromaffin tissue. The latter is often present also in the parenchyma, where it forms islets, more or less numerous, having few chromaffin cells. Unlike Rhynchocephalia, usually the adrenal gland of Squamata does not show chromaffin tissue on the ventral surface of the gland [65,66,67,68,69]. However, there are many differences in the distribution of the chromaffin cells, which can form a continuous envelope around the steroidogenic parenchyma, with few or no inner islets, as well as they can be concentrated to form islets scattered in the parenchyma and a very reduced dorsal mass. Above all, this great variability in the distribution of the two types of tissue can be observed not only between different families, but also within the same family or within the same genus, as has been observed studying many species belonging to different families, subfamilies, orders, and infraorders (for review see [2]).