Membrane Transformations in Degenerating Rodlet Cells in Fishes of Two Teleostean Families (Salmonidae, Cyprinidae)
Article first published online: 24 OCT 2008
Copyright © 2008 Wiley-Liss, Inc.
The Anatomical Record
Volume 291, Issue 12, pages 1693–1706, December 2008
How to Cite
Bielek, E. (2008), Membrane Transformations in Degenerating Rodlet Cells in Fishes of Two Teleostean Families (Salmonidae, Cyprinidae). Anat Rec, 291: 1693–1706. doi: 10.1002/ar.20796
- Issue published online: 25 NOV 2008
- Article first published online: 24 OCT 2008
- Manuscript Received: 17 AUG 2008
- Manuscript Accepted: 15 AUG 2008
- rodlet cell;
- cell degeneration;
- crystalloid endoplasmic reticulum;
Rodlet cells (RCs) of teleosts are identified by their fibrillar capsule and peculiar inclusions, the rodlets, consisting of a club-like sac and a central dense core. Former ultrastructural studies showing signs of hypertrophy of endoplasmic reticulum (ER) were followed up in Salmonids (Oncorhynchus mykiss, Salmo trutta L.) and compared with Cyprinids (Cyprinus carpio L., Carassius auratus L., Alburnus alburnus). Focusing on membrane transformations, unusual undulations of the membranes of rodlet sacs and often apposed ER-membranes, which were observed in mature or discharging cells, increased continuously in degenerating stages and ejected cytoplasmic packages or rodlets. Tubular elements (ø 25–30 nm or 30–50 nm) or small vesicles appeared partly derived from them. Terminal stages of this development were represented by RCs retained in the epithelium, which were completely filled by stacks of tubules and cores. Convoluted membranes were also found persisting between mostly undissolved rodlets at the epithelial surfaces. In Cyprinid species, the membrane changes were less conspicuous but essentially similar, including stages with confluent ER reported only in trout up to now. The membrane transformations resemble structures known as “crystalloid ER” indicating a disturbance in the protein production. The positive immunocytochemical reaction for calreticulin in the rodlet sacs, a luminal ER chaperone mediating recycling of misfolded proteins and upregulated during stress, supports this interpretation. The ER stress-reaction is an evolutionary conservative cytoprotective mechanism during physiological, environmental, and genetic aberrations and fits the increase of RCs reported in quite different situations, although details of its triggering need further investigation. Anat Rec, 2008. © 2008 Wiley-Liss, Inc.
Although the rodlet cell (RC) of teleosts has been discovered over a century ago, its origin and function are still controversially discussed. During the last years, the rodlet cell gained renewed interest as numerous studies showed the increase of rodlet cells in adverse conditions, ranging from diverse lesions, parasitic infections, or toxins to diverse environmental influences (see review of Manera and Dezfuli, 2004). Therefore, several authors followed the suggestion of Iger and Abraham (1997) interpreting the rodlet cell as a regular constituent of the defence system, representing possibly a convenient biomarker for environmental stress (Smith et al., 1995; Dezfuli et al., 2003b; Manera and Dezfuli, 2004). From the former three main interpretations—parasitic stage, secretory cell, or blood cell—the latter theory in differing modifications has been adopted again in many studies. The interpretations vary between an unspecified defensive (and/or) secretory cell, “sharing as only cell structural features and behavior with leucocytic and epithelial secretory cells” (see Dezfuli et al., 1998, 2000, 2003a, b; 2007a, b; Manera and Dezfuli, 2004). More specific is the assumption that the RC is related to granulocytes or mast cells (i.e., in teleosts eosinophilic granulocytes or “EGCs,” respectively, see Reite, 2005; Reite and Evensen, 2006). The RC is also discussed with reference to epithelioid cells (i.e., derived from macrophages) and mesothelial cells, both able to form desmosomes, and defensive functions of epithelial cells (Dezfuli et al., 2000; Manera and Dezfuli, 2004).
Nevertheless, inconsistencies in the distribution as the lack of rodlet cells in certain species or individuals (e.g., Sulimanovic et al., 1996; Reite, 2005; Reite and Evensen, 2006) or their presence unrelated to apparent stressors (Smith et al., 1995) remain. Even in recent studies stating the increase of RCs with growing symptoms of infection (Schmachtenberg, 2007) or ovarian maturation (Jordanova et al., 2007), the occurrence of individual variations was emphasised. Moreover, part of the morphological features is contradictory to the common pattern of blood cells. As has been discussed in several investigations (e.g., Leino, 1979; Manera and Dezfuli, 2004; Bielek, 2005; Schmachtenberg, 2007), neither a thick fibrillar capsule nor the presence of tight junctions and desmosomes are found in any other vertebrate blood cell type (Rowley et al., 1988; Zapata and Cooper, 1990; Zucker-Franklin and Grossi, 2003). The analysis of the organelle development in two Salmonid species shows the known secretory cycle (Leino, 1974, 1979; Barber et al., 1979) but indicates continuing hypertrophy of the endoplasmic reticulum (Bielek, 2005). The observation of confluent ER lacunae loosing their ribosomes proves the assumption that the conspicuous vesiculation of the cytoplasm is derived from RER during the decrease of the secretory activity. Moreover, in mature cells of Salmonids ER-cisternae and rodlet sacs, the latter especially after ejection, display often undulating membranes and/or association with tubular elements (Bielek, 2005). These formations are similar to “convoluted membranes” (CM) or “crystalloid reticulum” and typical for defects of the secretory pathway or various pathological conditions (Ghadially, 1997, Vol. I, p 433–602; Mackenzie et al., 1999).
To follow up the possibility of an aberrant cell development, this angle was further investigated ultrastructurally in trout (brown trout, rainbow trout) with the focus on discharging or degenerating rodlet cells. To probe the validity of the hypothesis of an aberrant ER development in other teleostean families, the results were compared with rodlet cell stages in different tissue samples of several Cyprinid species (carp, goldfish, and bleak).
Multiple ER stress reactions are characterized by modulations of the involved chaperones (Ma and Hendershot, 2002; Ni and Lee, 2007). As preliminary experiments showed a positive staining for calreticulin, the distribution of this leading luminal ER chaperone involved in folding and oligomerization of glycoproteins was tested immunocytochemically.
MATERIAL AND METHODS
In material investigated during former haematological studies of Salmonid and Cyprinid species, rodlet cells were recurrently found in haemopoietic organs and blood to a varying extent. For this study, part of this material (i.e., kidney, head kidney) was supplemented with samples of heart, gill, gallbladder, and intestinal epithelium from adult specimen of trout (Oncorhynchusmykiss, Salmo trutta L.) and compared with probes from three Cyprinid species (Cyprinus carpio L.; Carassius auratus L., Alburnus alburnus). The fishes were obtained from commercial sources or—in a few cases—caught by hook in their natural environment (6 specimen of trout, 2 of bleak) and killed by a blow to the head.
The samples comprised brown trout (S. trutta L. : 3 adult specimen, 26–50 cm; 5 immature ones: ca. 4 cm) and trout (O. mykiss: 8 adults, 30–35 cm, 2 immature; 8 and 11 cm), and the Cyprinid species goldfish (C. auratus L, 12 adults, ca. 12–14 cm), carp (C. carpio L., 17 adults weighing between 0.8 and 2.2 kg) and bleak (A. alburnus, 2 adults, ca. 16 cm).
The fishes showed no abnormal signs in behavior or morphology controlled under a low power dissection microscope. Examination by light and electron microscopy showed occasional tissue damage, sometimes with leucocytic infiltration, but without morphologically obvious cause. Because of the limitations of the possible scanning, minor infections or, for example, protozoan parasites in other regions or organs could not entirely be ruled out, but the presence of observed rodlet cells in the present samples was not related to any visible pathology.
For electron microscopical investigations, small blocks (1–2 mm3) were dissected out and processed after standard electron microscopical methods, that is, fixation in 2% glutaraldehyde buffered in phosphate buffer or cacodylate buffer (0.1 M, pH 7.2–7.4; at 4°C) for 1–2 hr, post fixation in 1% OsO4 in the same buffer for 1 hr, followed by dehydration in a graded ethanol series and embedding in epoxy resin (Epon 812 or Glycidether).
For the immunocytochemical reactions, the blocks were embedded without postfixation with OsO4 in LR White medium resin. Sections 80 to 150 nm-thick were mounted on gold grids and incubated for 1–2 hr with rabbit polyclonal anti-calreticulin (600-101-ab4, Abcam), followed by staining with a 10 nm gold conjugate (GaR Auro Probe RPN-421) for 1 hr. Nonspecific binding sites on the sections were blocked by a preceding incubation with 2–5% BSA in PBS. Controls were also performed by incubation of the sections with PBS instead of the primary antibody.
Semi- and ultrathin sections (0.5 μm and ca. 70 nm, respectively, with LR-White embedding up to 150 nm thickness) were cut on a Reichert ultra-microtome (OM U3, Ultracut S), stained with Toluidine blue or uranyl acetate/lead citrate, respectively, and examined on a Jeol 1200 EXII electron microscope (for LM survey: Leitz DMRB).
Samples of gill and intestine from fishes especially rich in rodlet cells (i.e., in focal accumulations) were selected for ultrastructural analysis.
Salmonids (Trout: Oncorhynchus mykiss, Brown Trout: Salmo trutta L.)
The rodlet cells of the two species correspond to the descriptions in the mentioned literature. Mature cells are oval with a fibrillar capsule that displays no distinct thick filaments or marginal densities observed in some other species. In the course of cell differentiation, the nucleus changed from round to irregular or indented, whereas the active Golgi area shifted from a supranuclear position to the basal part of the cell later on. The dilatation of ER as essential feature of the beginning RC development complies with literature and a former study (Bielek, 2005), displaying extension of RER cisternae, shedding its ribosomes, and giving the cell increasingly a vesicular appearance (with vesicles growing in size up to 1–2 μm, ending especially in gill epithelium in large, confluent lacunae). The apically dislocated mitochondria were often deformed by indenting vesicles. Discharge of the rodlets through microvillar corona and/or cytoplasmic bleb results in complete expulsion of cellular organelles and sometimes whole cells. Degeneration of RCs (partly with apoptotic nuclear morphology) can be observed at differing stages. These forms comprised either mature elongated cells, showing decreasing size, dense cytoplasm, and numerous rodlets, or immature round cells with only few (1–3) rodlets or 1–2 “condensing” vacuoles with or without visible core
Cells clusters comprising developmental stages as well as degenerating forms displaying even more conspicuous features of ER–hypertrophy were encountered especially in intestine and gill but rarely in kidney and seemed to be lacking (in the present material) in other sites (e.g., heart or gall bladder). Therefore structural analysis was concentrated on RC aggregations from intestine and gill.
In focal aggregations of rodlets cells all developmental stages were found, ranging from immature cells lacking a capsule and identified by enlarged RER cisternae to fully mature, discharging RCs, including the mentioned smaller, dense cells representing probably degenerating, or apoptotic cells (Figs. 1 and 7). Occasionally, migrating leucocytes, especially lymphocytes, were present in the epithelium as well as in the lumen near damaged epithelial cells (Fig. 1). Masses of ejected rodlets (Figs. 1 and 2) and whole or spent RCs occurred in the intestinal lumen and displayed differing signs of dissolution. The rodlet sacs in discharging cells showed often membranes undulating in varying degrees, which became even more conspicuous in ejected rodlets. Tubular elements with dense or more inconspicuous membrane (ø 25–30 nm or varying between 30 and 50 nm, respectively) were often associated with cores (Figs. 2 and 3) or rodlet sac membranes. In apparently degenerating areas containing various vesicles and cell organelles, rodlet sacs displayed undulating membranes apparently generating parallel tubular foldings (Fig. 4). At this stage, the RCs were usually released from the epithelium, but in cells retained and without apparent contact to its surface, occasional associations of cores and tubular stacks were observed (Fig. 5). In rare cases, these degenerating RCs were filled with parallel cores and intermingling long tubules in between (Figs. 7 and 7a). Concentrated stack formations (Fig. 6) were presumably derived by increasing compression of such cells, with the terminal stage probably represented by phagocytosis of the remnants (Fig. 8).
In the multilayered epithelium of the primary gill lamellae, the stages of the rodlet cells ranged from undifferentiated cells to mature ones showing the same morphology as in the intestine. In the secondary lamellae, RCs were rare and usually with their long axis arranged parallel instead of perpendicular to the surface, probably due to the low height of the epithelium and similar, for example, to goblet cells in the skin (not shown). Mature cells occasionally displayed extremely dilated ER lacunae surrounding the rodlet sacs and the remnants of the organelles (Fig. 9). Cells with segregation of their nuclear components undergoing probably apoptosis seemed to occur more often than in the intestine (Figs. 10, 12, and 13), possibly reflecting a greater rate of cell sloughing. In some cases, in these cells, a definite capsule was lacking altogether (Figs. 10 and 12) or appeared to be in the process of dissolution (Fig. 13). Again, ER dilations, undulating membranes of rodlet sacs and tubular elements in the sacs or running along the sac membrane were regularly found (Figs. 10–12). Tubular elements were especially conspicuous in rodlet sacs of degenerating, exfoliating or free RCs (Figs. 10–12) as well as in discharged rodlets, independently if contained in a shed cytoplasmic package (Fig. 15) or free in a state of lysis (Fig. 16). Transformations to stacks corresponding to the compressed cells observed in the intestine were not seen, although tubular aggregations occurred in or near dissolving rodlets (Fig. 16a).
Free RCs near an epithelial abrasion of a secondary gill lamella belonged—as judged by morphological means—definitely to an apoptotic degeneration pattern, displaying not only an apoptotic nucleus (i.e., chromatin segregation) and enormous ER lacunae surrounding the rodlets, but also a dissolving fibrillar capsule and a clear cytoplasmic marginal zone free of organelles. Additionally, very dense round bodies of unknown origin and consistence were seen in the border zone and between the rodlet sacs (Figs. 13 and 14). Morphologically, they were reminiscent of aggresomes resulting from cytoplasmic segregation of proteins during defects in their synthesis.
Calreticulin as soluble chaperone is found principally in ER-cisternae and the ER-Golgi-intermediate compartment (ERGIC). In the two Salmonid species as well as in samples of goldfish and carp with which they were compared, the rodlet sacs but not the cores showed a strong positive reaction independently of tissue and species (Figs. 17, 18, and 20) with the exception of a modification in the apical part of certain rodlets with apically condensed material in the Cyprinid species (see Fig. 37). In RCs containing a still active, large Golgi apparatus and dilated ER lacunae, the staining in the ER compartment was only weak to moderate (Figs. 18 and 20). In the presumptive ERGIC zone near the outer cis-Golgi face, the reaction was also rather diffuse (Fig. 20). In cells with extremely dilated ER lacunae or vesicles, only the RC sacs were positive. Goblet cells in gill and intestine of trout (Figs. 19 and 21), which might serve as a prototype of a secretory cell, showed the expected dot-like reaction pattern in active, narrow stacks of ER lacunae, and a rather strong one in vesicles in the ERGIC-zone
Cyprinidae (C. auratus L.; C. carpio L., Alburnus alburnus)
In RC clusters of goldfish and carp, the full cycle of developmental stages with similar characteristics as in trout was encountered. Certain differences in comparison to trout species can be enumerated: the nucleus retaining its round shape throughout the cell development, the change of the active Golgi area in a supranuclear position to an inconspicuous stack pressed laterally against the capsule, and the transforming of the mitochondria to dark and long structures.
With respect to the ER, maturing RCs appeared sometimes to develop stacks of RER in the marginal zone before the beginning of the dilatation in more centrally located ER lacunae around the increasing Golgi apparatus. During the consecutive maturing of condensing vacuoles and rodlet sacs, respectively, the ER-derived vesicles stayed comparatively small and round (ø 200–500 μm compared with the often long-stretched cisternae in trout) irrespective of the tissue (compare C. auratus L., gallbladder, Fig. 22, with C. carpio L., intestine, Fig. 23). On the other hand, the ER-dilatation was sometimes continued, possibly due to retaining of the cells in the tissue (e.g., C. carpio L., haemopoietic tissue, kidney, Figs. 24 and 25a,b). These terminal stages contained larger vesicles and often shed the rodlets not only by the differing apical openings (microvillar: Fig. 24, cytoplasmic bleb: Fig. 25a) but also lateral rupture of the capsule (Fig. 25a). Occasionally, the enormously extended ER filled the space around the dislocated organelles (Fig. 25b) corresponding to quite similar stages observed in trout gill.
RC accumulations selected for analysis were frequent—apart from the gill sample of bleak—in head kidney, kidney, heart (especially in auricles, bulbus arteriosus), and occasionally aorta, but irregular in gallbladder or intestine of goldfish and carp.
Narrow apposition of small clear (ER derived) vesicles with the undulations of the rodlet sac membrane was observed in C. auratus L. as well as in C. carpio L. (compare different stages in heart, C. auratus L., Figs. 26 and 27, intestine, C. carpio L., Fig. 28)
The transformation of the sac membranes into tubular elements could also be observed, for example, in degenerating RCs in C. auratus L.: extreme membrane undulations of nuclear envelope (not shown) and rodlet sacs (Fig. 29) resulting partly in tubular and vesicular elements (ø ca. 50 nm), occurrence of tubules (ø 20–50 nm) in the cytoplasm near sacs (Figs. 30 and 30a) or cores (not shown). In one case a sac filled with tubules of smaller diameter (ø 20 nm) was found (Fig. 31) besides 2 other sacs of differing density but homogeneous appearance. Compressed stacks of tubules and cores seen in trout were only found in an area of RC-aggregation in the gill of A. alburnus (Fig. 32), but were completely similar.
Stages with dark cytoplasm probably indicating degeneration of RCs were seen in different tissues. In some of these cells, dense aggregations of differing size (ranging from very small and dot-like to large patches) were distributed along the inner side of the capsule and rarely in the cytoplasm (Fig. 33).
The structure of the rodlets was characterized by differing degrees of condensation. Although in C. auratus L. the distended lower part of the sac showed varying density (and often artificial shrinkage from the sac membrane), the rim of the sac and the apical narrow part along the core was either less dense and of cloudy appearance or sometimes clear with small round, dense aggregations (ø 40–60 nm; C. auratus L., compare Fig. 27, intracellular rodlet, and discharged ones: Figs. 34 and 36). In C. carpio L., this more floccular outer zone often condensed to a dark zone bordering the sac and the apical core, giving the rodlet a bipartite appearance (Fig. 35). Undulations of the membranes in discharged rodlets resulting in vesicular projections were less pronounced than in trout (Fig. 36).
Irrespective of the degree of condensation, the apical zone reacted in the tested carp and goldfish negative for calreticulin (Fig. 37).
The question of function and nature of the rodlet cell is still open for debate. The outstanding features of the rodlet cell, the filamentous capsule and the peculiar inclusions, the rodlets with their dense, rodlet-like core, are difficult to reconcile with any known scheme of tissue cell. The development of the secretory organelles has been documented by electron microscopical investigations, as well as the formation of true junctional complexes indicating an epithelial origin of the cell (e.g., Leino, 1974, 1979; Gruenberg and Hager, 1978; Barber, 1979; Bielek and Viehberger, 1983). Among other possible functions, an antibiotic nature of the presumed secretory product, the rodlets, was proposed (Leino, 1974, 1979), possibly stimulated by tissue injury (Leino, 1996), thereby explaining the increase of RCs, for example, in parasitic infections, inflammations or environmental stress situations as reviewed by Manera and Dezfuli (2004). The defensive effect seems to be supported by scattered immunocytochemical studies showing a positive reaction for alkaline phosphatase and peroxidase (Iger and Abraham, 1997), the stress hormone alpha MSH (melanin stimulating hormone, Mazon et al., 2007) and the antimicrobial peptide piscidin (Silphaduang et al., 2006).
This hypothesis of a regular defensive element is complicated by the distribution of the RCs in quite different tissues including connective tissue, haemopoietic organs and peripheral blood. Although an example for ubiquitous secretory cells increasing, for example, in inflammation might be found in the (epithelial) mucous cells, the distribution pattern of the RCs corresponds with a migratory (secretory) cell instead of a fixed epithelial one. Consequently, as mentioned in the introduction, the interpretation of the RC as a blood cell (first proposed by Duthie, 1939) has been taken up again (reviewed by Bielek, 2002, 2005; Manera and Dezfuli, 2004; Reite, 2005; Reite and Evensen, 2006; Mazon et al., 2007; Schmachtenberg, 2007). With respect to distribution, an inherited pattern of a “standing force” of RCs as adaption to consistent exposure to, for example, parasites, has been suggested, the RCs populating tissues as immature forms and activated by diverse stimuli in contrast to a constant “mobile force” of other cell types (see Reite, 2005; Reite and Evensen, 2006). A number of observations at different levels imply indeed similarities to migrating leucocytes, for example, the “margination” of RCs along the endothelium reported by Smith et al. (1995), or the corresponding influence of a synthetic corticosteroid (dexamethasone) on the surface molecules both of RCs and leucocytes (Manera et al., 2001). Smith et al. (1995) and El Habback et al. (1997) considered undifferentiated cells in the blood as precursors of rodlet cells. This opinion correlates in part with the ontogenetic study of Mazon et al. (2007) who described the first RCs between kidney and intestine (5 days post fertilization, i.e., 1 day post feeding), but found only mature RCs in other sites (e.g., gill). As a consequence, they proposed a colonization of different organs by precursors from haemopoietic sites. Unluckily, this brings the discussion back to the beginning as contrary to these results differentiating immature cells indicating at least differentiation in situ have been described in different epithelia (e.g., Leino 1974, 1979; Barber et al., 1979; Bielek, 2005). Moreover, a leucocyte moves or adheres by adhesion molecules, whereas—as mentioned above—immature as well as surface contacting RCs form definitely junctional complexes with epithelial cells. In blood cells of vertebrates this feature is completely atypical (see Rowley et al., 1988; Zapata and Cooper, 1990; Zucker-Franklin and Grossi, 2003) and only discussed for reticular cells or monocytic derivatives, for example, multinucleate epithelioid cells (Dezfuli et al., 2000; Tafalla et al., 2001). Because of this conflicting observations the interpretations vary considerably from an undefined defensive cell or an unknown type of granulocyte to a cell comparable or even related to mast cell equivalents and eosinophil granulocytes, respectively, (see Manera and Dezfuli, 2004; Reite, 2005; Reite and Evensen, 2006). The proposal that the RC combines the features of an leucocytic as well as an epithelial cell and represents a new blood cell type unique for teleosts (Manera and Dezfuli, 2004) overcomes the controversial arguments and accepts the contradictions.
As recent alternatives for the blood cell hypothesis the submicroscopic infection of a leucocyte (Fishelson and Becker, 1999) or the incorporation of a foreign (parasitic ?) genome into the fish genome resulting in an endogenous defensive cell (Schmachtenberg, 2007) were proposed. In the latter study, the role of the rodlets is seen as defensive weapon against other invasive organisms, a positive feed back of the assumed genetic addition.
A further hypothesis based on the ultrastructural observation of hypertrophy of the endoplasmic reticulum (ER) is the possibility of an aberrant cell development (Bielek, 2005).
In trout, the normal sequence of secretion, extension of rough ER, transport of the proteins to the intermediate ER-Golgi compartment (ERGIC), modification in the Golgi apparatus, and transport from the trans-face to the cell surface seems differentiated in the RCs (Bielek, 2005). Immature stages showed not only dilation of RER cisternae corresponding to normal protein production, but increasing extension and loss of ribosomes resulting in the vesiculation of cytoplasm typical for RCs but also representing the classical signs of degenerating ER. Occasional terminal stages display extremely dilated, confluent ER-lacunae. The implication of direct participation of ER-derived vesicles in the differentiation of the rodlet sacs and the observation of membrane undulations and tubules are indicating protein overproduction. This is supported by the present findings showing the continuity of this process ending sometimes in degenerating cells, which are completely filled with stacks of such membranous elements and cores. This pattern can be related to features found in cells with disturbed protein production (Hammond and Helenius, 1994; Kaufman, 1999).
Multiple mechanisms can induce defects in the folding or transport of the nascent proteins. The retention of malfolded products usually involves repeated recycling of the proteins between ER and cis-Golgi face until correct folding succeeds, involving the vesicles of the ERGIC or—as recently proposed—sequestered directly from the ER without local restriction to the Golgi area (Zuber et al., 2001; Fan et al., 2003; Appenzeller-Herzog and Hauri, 2006). During this process, the proteins are bound to chaperones mediating transport and folding.
Any accumulation of un- or misfolded proteins leads to ER-stress, triggering the UPR (unfolded protein reaction). Chaperones even of one group may affect and modify the translocation of the respective proteins differently, that is, transferring them to the cytosol for degradation or retaining them in the ER-lumen (Ni and Lee, 2007). Failure of this repair mechanism leads to apoptosis. Applied to the RC, this would imply the segregating of one (or several) defective or surplus protein(s) into the developing rodlets, which are then stored, partly crystallized and growing continuously till discharge of the whole cell. The observation of rodlet cells degenerating at varying stages of their development agrees with this scheme, whereas the occasional stages with apoptotic nucleus and dense round cytoplasmic bodies found in trout gill could correspond to degraded protein accumulations (i.e., aggresomes, Kopito and Sitia, 2000) not retained in ER but retranslocated back to the cytoplasm.
Apoptosis is defined as programmed cell death, which can be controlled at several points. It is essential for the regulation of cell numbers in tissue differentiation and the elimination of wrongly programmed cells (e.g., in embryonal development, metamorphosis or the removal of incorrectly primed lymphocytes in thymus). The former clear demarcation from necrosis as accidental death after insults has been complicated by the discovery of intermediate mechanisms in the form of programmed autophagic cell death (PCD) or secondary necrosis following apoptosis in the absence of final phagocytosis (Clarke, 1999; Bursch, 2001; Doonan and Cotter, 2008; Krysko et al., 2008). Apoptosis is also triggered by exogenous factors, either artificially by toxic agents or physiologically (e.g., the killing of aberrant cells by granule proteases of cytotoxic lymphocytes). Nuclear apoptosis under conditions of mercury exposure has been described in RCs and intestinal cells recently (Giari et al., 2008), but in this case was combined with intracellular oedema and abundant myelinoid bodies, that is, signs associated with PCD or necrosis.
With respect to the cell cycle of defensive cells, the disappearance of human eosinophils in inflammation seems to involve primary cytolysis and secondary necrosis instead of the formerly assumed exclusive apoptosis (Uller et al., 2004). Moreover, apoptosis as a defensive reaction of the cell might be modulated in different ways. The example of virally infected fish cells (Joseph et al., 2004) shows that viral proteins may not only kill by necrosis of the host cells but induce or inhibit apoptosis and even do both at different stages. The cell reaction depends on the cell type as well as on the mechanism aiming either on blocking apoptosis to ensue efficient virus production or cause cell death to enhance virus dissemination (Joseph et al., 2004).
In the case of the RC, the process appears to set in at various developmental stages, and therefore regulation of homeostasis might be the promoting event. The progress of cell and organelle deterioration is fitting the description of apoptosis (Joseph et al., 2004): shrinkage of cell, rounding, reorganization of the cytoskeleton, loss of cell contacts, detachment of ribosomes from the ER, at least sometimes typical nuclear condensation, and also lack of inflammatory reaction and phagocytosis by adjacent cells. However, definitive designation as apoptosis needs integrated approach by a combination of several methods (Krysko et al., 2008).
A further test for this hypothesis is the demonstration of the involved chaperones, for example, calreticulin, an unique ER luminal resident protein (Johnson et al., 1997; Ma and Hendershot, 2002; Ni and Lee, 2007). In its role as chaperone, it is involved in the release of proteins to the Golgi apparatus. In case of malfolding, it becomes upregulated and preserves terminally misfolded proteins in soluble conformation in the ER lumen. Recently, calreticulin has been identified in rainbow trout and proved to be highly conservative (Kales et al., 2004) with at least in part similar functions as in higher vertebrates (Kales et al., 2007). In this study, goblet cells, which might serve as model for a typical secretory cell, showed the expected positive reaction over marginal ER cisternae and the ERGIC region at the cis face of the Golgi apparatus. The reaction in rodlet cells was much more diffuse over ER and ERGIC and negative in the extremely dilated lacunae in degenerating cells. The material of the rodlet sacs, not cores, showed a strong reaction, thereby supporting the assumed relation to ER and malfunction.
Calreticulin has yet other functions, for example, as modulator of calcium, but it is also found at the cell surface (regulating activated lymphocytes and/or leucocytes, tumor and embryonal cells), in the nucleus (mediating nuclear export and steroid receptors) and is also linked to neurodegenerative prion diseases. In this context, it might be interesting to speculate about a possible influence of calreticulin in the presumed migratory behavior of the rodlet cell. Moreover, Balabanova (2000) registered a “common structure” of RCs but increased numbers after exposure to Ca2+- free water or in water supplemented with Cd 2+ which is known to suppress Ca2+- absorption.
With respect to the other conspicuous morphological feature, the appearance of membrane undulations and/or of tubular elements is also associated with protein overproduction (Ghadially, 1997, Vol. I, p 433–602). Generally, the occasional formation of membrane undulations may be caused by an overproduction of membrane-bound proteins or involved enzymes, serving as temporary storage and repair mechanism. The membranes may form multiple configurations, ranging from more or less curved to tubular formations or straight foldings. These arrangements may assume loose or accumulated, sometimes compact (“crystalloid”) shapes and are known under different designations: crystalloid ER, paracrystalline ER, convoluted membranes (CM), tubular reticular system (TRS), organized SER, or cubic membranes (for synonyms see Cho et al., 1994; Ghadially, 1997, Vol. I, p 433–602; Teterina et al., 1997; Snapp et al., 2003).
If the membrane transformations of ER and rodlet sacs occurring in the RCs correspond to this reaction, this could be an explanation for the observed variation of elements.
The present ultrastructural results showed a continuous increase of undulations of the membranes of rodlet sacs and of tubular elements probably derived from them. The diameter of the tubules varied (ø 25–30 nm or 30–50 nm), possibly indicating differing components, but was in the same range as the tubular elements of the crystalloid ER-aggregations described in literature. Tubules in rodlet cells have been only reported by Leino (1974, 1979), who described them as straight or bent with larger diameters (ø 50 nm) and presumably representing species specific structures. As it is known from different strains of viral infections used as experimental models that small changes in the production of the proteins induce conspicuous changes in the curvatures, changing from loose to very narrow curves or—on the contrary—parallel tubular forms (Teterina et al., 1997), similar molecular variations could be the reason for different diameters and occurrence of the tubular elements. The membrane association is seen as result of their oligomerization capacity (Fukuda et al., 2001), a low affinity protein interaction of the arrays (Snapp et al., 2003) or due to multispanning C-terminals (Fedorovitch et al., 2005).
The apoptotic rodlet cells found in the gill of trout which contained rodlets associated with numerous tubular elements but no trace of fibrillar capsule rise the question if the formation of the capsule is an essential feature necessary for active contraction and discharge as supposed in literature (see Leino, 1974; review of Manera and Dezfuli, 2004). The possibility of a cellular reaction lacking in rodlet cells, for example, in low epithelia due to a shortened or retarded development and/or heightened rate of desquamation should be considered (compare, e.g., the differentiation of myoid bands in stressed epithelial cells: Ghadially, 1997, Vol. II, p 418).
Moreover, Schmachtenberg (2007) developed an isolation method for the RCs and reported in vitro a rapid, often subsecond discharge without capsular contraction. The persistence of ejected rodlets during the 12-hr observation period supports the present results of only slow dissolution of rodlets and their membranes, although their lysis in quite different conditions (blood, intestinal lumen, mucous environment) has to be further investigated.
Comparing these results with the Cyprinid samples (goldfish, carp, bleak), some morphological differences can be enumerated, for example, the capsule differentiated into thin and thick fibrils and often marginal densities and the round (instead of the irregular) shape of the basal nucleus. The general development corresponded to scattered observations in literature as well as the sequence described for the carp by Iger and Abraham (1997), including early apoptotic stages. The manner of rodlet discharge follows in carp and goldfish also the scheme of cytoplasmic microvillar “corona” or cytoplasmic “bleb” and/or final disruption (see Gruenberg and Hager, 1978; Kramer and Potter, 2002; Manera and Dezfuli, 2004), independent from eventual modulation by external toxic factors reported by Manera et al. (2001) or lateral expulsion by a ruptured capsule observed in carp (Mazon et al., 2007) and other species, for example, Anguilla (Dezfuli et al., 1998), Xiphophorus (Kramer and Potter, 2002), Abramis (Dezfuli et al., 2003a), and trout (Bielek, 2005).
With respect to the ER, the developmental pattern is with certain limitations (i.e., smaller ER-derived vesicles, less conspicuous membrane undulations of ER and rodlet sacs) similar in the investigated Cyprinid species. Although in most maturing rodlet cells the vesicles were round and measured only about 200–500 nm (instead of long-stretched cisternae or vesicles up to 1–2 μm in maturing trout cells), in tissues where rodlet cells were apparently retained without immediate shedding the rodlets, for example, multilayered (gill) epithelium, haemopoietic tissue of kidney, or subendothelial space of the heart, rodlet cells with larger vesicles were found easily, implying continued increase. This is supported by the observation of stages with the extreme dilatation of the ER lacunae identical with similar stages up to now published only for trout in polluted water as well as in healthy appearing fishes (Iger and Abraham et al., 1997; Pawert et al., 1998; Bielek, 2005). Some rodlet cells showing degenerating organelles and sometimes an apoptotic (i.e., segregated) nucleus containing subcapsular dense accumulations might correspond to terminal stages with retranslocated material, possibly comparable to the RCs with round aggresome-like aggregations in trout.
With respect to tubular membrane transformations, quite similar membrane undulations of rodlet sacs as well as tubular elements in sac or cytoplasm (ø 20 or ca. 50 μm) could be found in discharging or degenerating cells. Although end stages with stacks of cores and tubules seen in trout were found up to now only in the gill epithelium of bleak, they displayed similar morphology.
The calreticulin reaction corresponded to the positive one in trout, that is, strong staining of the rodlet sacs with an interesting difference. In carp as well as some other species of Cyprinidae and other families, the apical part around the tip of the core and rim of the rodlet sac shows often—possibly dependent on the degree of condensing—a dense (or “coarse granular”) appearance (e.g., C. carpio L. and Tinca tinca: Bielek and Viehberger, 1983; C. carpio L.: Iger and Abraham et al., 1997; Mazon et al., 2007; Leuciscus leuciscus: Dezfuli et al., 2003b; Anguilla anguilla: Dezfuli et al., 1998). This region reacts negative to calreticulin, indicating another component and/or dissociated material. Moreover, in goldfish this zone was often electron lucent with scattered small dense aggregations, implying focal condensation of the contents.
Summarizing the ultrastructural observations, the present findings confirm the assumption of a continuous hypertrophy of ER elements during maturation of the RC. The observed membrane transformations might represent reversible signs of some metabolic change, but are so numerous in diseased states that they are often taken as indicator of pathological conditions (Ghadially, 1997). They are also developed in situations of deprivation, for example, during a defect in the LDL-dependent cortisol production, and therefore principally related to conditions of stress. Most frequently, this ER transformation is seen in genetic defects or in virally infected cells in which the secretory pathway gets corrupted. In normal cells, these different forms of crystalloid ER are reversible and have been found during a change in the production of lipidogenic enzymes, for example, in steroid or lipid producing cells as gonads or sebaceous glands (see Snapp et al., 2003; Fedorovitch et al., 2005; Almsherqi et al., 2006). Independently of the cause, the reaction represents a phylogenetic conservative cytoprotective mechanism, segregating the misfolded or superfluous proteins at least temporarily.
In the case of the rodlet cell, a frequent correlation to quite different stress conditions is conspicuous, but the triggering of the development might set in at different levels, with the exact cause still unknown. Each of the enumerated defects might result in a disturbance of the ER-function. If during this reaction to (cellular) stress the synthesis of enzymes is increased, for example, alkaline phosphatase (Iger and Abraham, 1997), which might act not only against some surplus product but also against other organisms as parasites, this would be consistent with the reported increase and effect in diverse infections. On the other hand, an ubiquitous occurrence of any modification needs further explanation. While this question should be followed up, the ER-development is not only relevant for clearing up role and function of the RC, but also the mechanisms acting in cell stress.
The author thanks E. Vanyek-Zavadil for excellent technical support; Dr. E. Noisser and Dr. R. Konecny for helping with the wild catches; and Mag. R. Schulz for correction of the English draft.
- 2006. Cubic membranes: a legend beyond the flatland of cell membrane organization. Rev J Cell Biol 173: 839–844. , , .
- 2006. The ER-golgi intermediate compartment (ERGIC): in search of its identity and function. J Cell Sci 119: 2173–2183. , .
- 2000. The ultrastructure of rodlet cells of fishes in condition of calcium deficiency in the environment. Tsitologiia 42: 1048–1052. .
- 1979. New observations on the rodlet cell (Rhabdospora thelohani) in the white sucker Catostomus commersoni (Lacepede): LM and EM studies. J Fish Biol 14: 277–284. , , .
- 2002. Rodlet cells in teleosts: new ultrastructural observations on the distribution of the cores in trout (Oncorhynchus mykiss, Salmo trutta L.). J Submicrosc Cytol Pathol 34: 271–278. .
- 2005. Development of the endoplasmic reticulum in the rodlet cell of two teleost species. Anat Rec A 283: 239–249. .
- 1983. New aspects of the “rodlet cell” in teleosts. J Submicrosc Cytol 15: 681–694. , .
- 2001. The autophagosomal–lysosomal compartment in programmed cell death. Rev Cell Death Differ 8: 569–581. .
- 1994. Membrane rearrangement and vesicle induction by recombinant poliovirus 2C and 2BC in human cells. Virology 202: 129–145. , , , , .
- 1999. Apoptosis versus necrosis. How valid a dichotomy for Neurons? In: KoliatsosVE,RatanRR, editors. Cell death and diseases of the nervous system. Totowa, NJ: Humana Press. p 3–27. .
- 1998. A description of rodlet cells from the alimentary canal of Anguilla anguilla and their relationship with parasitic helminths. J Fish Biol 53: 1084–1095. , , .
- 2007b. Rodlet cells and the sensory systems in Zebrafish (Danio rerio). Anat Rec 290: 367–374. , , , , .
- 2003a. Immunohistochemistry, ultrastructure and pathology of gills of Abramis brama from Lake Mondsee, Austria, infected with Ergasilussieboldi (Copepoda). Dis Aquat Organ 53: 257–262. , , , , .
- 2007a. The role of rodlet cells in the inflammatory response in Phoxinus phoxinus brains infected with Diplostomum. Fish Shellfish Immunol 23: 300–304. , , .
- 2003b. Alteration of rodlet cells in chub caused by the herbicide Stam(R) M-4 (Propanil). J Fish Biol 63: 232–239. , , , , .
- 2000. Rodlet cells and other inflammatory cells of Phoxinus infected with Rhaphidascaris acus (Nematoda). Dis Aquatic Org 43: 61–69. , , , .
- 2008. Morphological assessment of apoptosis. Methods 44: 200–204. , .
- 1939. The origin, development and function of the blood cells in certain marine teleosts. I. Morphology. J Anat 73: 396–412. .
- 1997. The possible origin and function of rodlet cells in Oreochromis niloticus. Egypt J Histol 20: 135–150. , , .
- 2003. Ultrastructural analysis of transitional endoplasmic reticulum and pre-Golgi intermediates: a highway for cars and trucks. Histochem Cell Biol 120: 455–463. , , .
- 2005. The dynamic ER: experimental approaches and current questions. Curr Opin Cell Biol 17: 409–414. , , .
- 1999. Rodlet cells in the head and trunk kidney of the domestic carp (Cyprinus carpio): enigmatic gland cells or coccidean parasites? Naturwissenschaften 86: 400–403. , .
- 2001. Formation of crystalloid endoplasmic reticulum induced by expression of synaptotagmin lacking the conserved WHXL motif in the C terminus—structural importance of the WHXL motif in the C2B domain. J Biol Chem 276: 41112–41119. , , .
- 1997. Ultrastructural pathology of the cell and matrix. 4th ed. Boston: Butterworths–Heinemann. .
- 2008. Histo-cytological responses of Dicentrarchus labrax (L.) following mercury exposure. Ecotoxicol Environ Saf 70: 400–410. , , , .
- 1978. Zur Ultrastruktur der “Staebchendruesenzellen” (rodlet cells, pear-shaped cells) im Bulbus arteriosus des Karpfens, Cyprinus carpio L. (Pisces: Cyprinidae). Anat Anz 143: 277–290. , .
- 1994. Quality control in the secretory pathway: retention of a misfolded viral membrane glycoprotein involves cycling between the er, intermediate compartment, and golgi apparatus. J Cell Biol 126: 41–52. , .
- 1997. Rodlet cells in the epidermis of fish exposed to stressors. Tissue Cell 29: 431–438. , .
- 1997. Protein folding in vivo: unraveling complex pathways. Minireview. Cell 90: 201–204. , . .
- 2007. A quantitative study of the hepatic eosinophilic granule cells and rodlet cells during the breeding cycle of Ohrid trout, Salmo letnica Kar. (Teloestei, Salmonidae). Fish Shellfish Immunol 23: 473–478. , , .
- 2004. Mechanism of cell death during infectious salmon anemia virus is cell type-specific. J Gen Virol 85: 3027–3036. , , , , .
- 2007. Calreticulin in rainbow trout: a limited response to endoplasmic reticulum (ER) stress. Comp Biochem Physiol B 147: 607–615. , , .
- 2004. Molecular cloning and characterization of calreticulin from rainbow trout (Oncorhynchus mykiss). Immunogenetics 55: 717–723. , , .
- 1999. Stress signaling from the lumen of the endoplasmic reticulum: coordination of gene transcriptional and translational controls. Review. Genes Dev 13: 1211–1233. .
- 2000. Aggresomes and Russell bodies. Symptoms of cellular indigestion? EMBO Rep 1: 225–231. , .
- 2002. Ultrastructural observations on rodlet-cell development in the head kidney of Southern Platyfish, Xiphophorus maculatus (Teleostei:Poeciliidae). Can J Zool 74: 1432–1436. , .
- 2008. Apoptosis and necrosis: detection, discrimination and phagocytosis. Methods 44: 205–221. , , , .
- 1974. Ultrastructure of immature, developing, and secretory rodlet cells in fish. Cell Tissue Res 155: 367–381. .
- 1979. Aspects of the fine structure, cytochemistry and distribution of teleost rodlet cells. Ph. D. Thesis. Cincinnati, Ohio: Union Institute. .
- 1996. Reaction of rodlet cells to a myxosporean infection in kidney of the bluegill, Lepomis macrochirus. Can J Zool 74: 217–225. .
- 2002. The mammalian endoplasmic reticulum as a sensor for cellular stress. Cell Stress Chaperones 7: 222–229. , .
- 1999. Markers for trans-Golgi membranes and the intermediate compartment localise to induced membranes with distinct replication functions in Flavivirus-infected cells. J Virol 73: 9555–9567. , , .
- 2004. Rodlet cells in teleosts: a new insight into their nature and functions. Review. J Fish Biol 65: 597–619. , .
- 2001. The effect of dexamethasone in the occurrence and ultrastructure of rodlet cells in goldfish. J Fish Biol 59: 1239–1248. , , .
- 2007. The first appearance of rodlet cells in carp (Cyprinus carpio L.) ontogeny and their possible roles during stress and parasite infection. Fish Shellfish Immunol 22: 27–37. , , , , .
- 2007. ER chaperones in mammalian development and human diseases. Minireview. FEBS Lett 581: 3641–3651. , .
- 1998. Ultrastructural changes in fish gills as biomarker to assess smallstream pollution. Tissue Cell 30: 617–626. , , .
- 2005. The rodlet cells of teleostean fish: their potential role in host defence in relation to the role of mast cells/eosinophilic granule cells. Fish Shellfish Immunol 19: 253–267. .
- 2006. Inflammatory cells of teleostean fish: a review focusing on mast cells/eosinophilic granule cells and rodlet cells. Fish Shellfish Immunol 20: 192–208. , .
- 1988. Fish. In: RowleyAF,RatcliffeNA, editors. Vertebrate blood cells. Cambridge: Cambridge University Press.Chapter 2, p 19–127. , , , .
- 2007. Epithelial sentinels or protozoan parasites? Studies on isolated rodlet cells on the 100th anniversary of an enigma. (Centinelas epiteliales o parásitos protozoarios? Estudios en células rodlet aisladas en el centenario de un enigma). Rev Chil Hist Nat 80: 55–62. .
- 2006. Evidence for widespread distribution of piscidin antimicrobial peptides in teleost fish. Dis Aquat Organ 72: 241–252. , , .
- 1995. Observations on rodlet cells found in the vascular system and extravascular space of angelfish (Pterophyllum scalare). J Fish Biol 46: 241–254. , , , .
- 2003. Formation of stacked ER cisternae by low affinity protein interactions. J Cell Biol 163: 257–269. , , , , , , , .
- 1996. The possible role of rodlet cells in the immune system of carp (Cyprinus carpio L.). Vet Arhiv 66: 103–109. , , , .
- 2001. Cytotoxic activity against prelabeled RTG-2 cells in the turbot, Scophthalmus maximus (L.). J Fish Dis 24: 169–175. , , .
- 1997. Induction of intracellular membrane rearrangements by HAV proteins 2C and 2BC. Virology 237: 66–77. , , , , .
- 2004. Occurrence of apoptosis, secondary necrosis, and cytolysis in eosinophilic nasal polyps. Am J Resp Crit Care Med 170: 742–747. , , , , .
- 1990. The immune system: comparative histophysiology. Chicester: Wiley. , .
- 2001. Immunolocalization of UDP-glucose: glycoprotein glucosyltransferase indicates involvement of pre-Golgi intermediates in protein quality control. Proc Natl Acad Sci USA 98: 10710–10715. , , , , , , .
- 2003. Atlas of blood cells. Function and pathology. 3rd Ed. Milano: Edi.Ermes s.r.l. , .