Immunohistochemical study of pituitary cells in wild and captive Salminus hilarii (Characiformes: Characidae) females during the annual reproductive cycle

Salminus hilarii (Valenciennes, 1850), commonly named tabarana, is one important migratory freshwater fish of the Brazilian ichthyofauna. It is a potamodromous (rheophilic) and carnivorous fish species occupying the top of the food chain and has a geographical distribution from the upper Paraná River Basin to São Francisco, Tocantins, Alto Amazonas, and Alto Orinoco Basins (Mirande 2010). In S. hilarii, like other migratory fish species (Brycon orbignyanus, S. brasiliensis, Pseudoplatystoma corruscans), the upstream migration is necessary to complete the development of their gonads and reproduce (Agostinho et al. 2003; Andrade et al. 2006), considering that the spawning behavior in S. hilarii takes place in the upper region of minor tributaries in this Basin. When migration is not possible, tabarana females do not ovulate, and consequently do not spawn (Honji et al. 2009 2011). Additionally, S. hilarii in the region of the upper reaches of the Tietê River (the river included in Paraná River Basin) has a great ecological importance given its high degree of environmental selectivity, it can be used as an accurate environmental indicator, considering the limited occurrence of this species just in a few unpolluted regions of the Paraná River Basin (Lima-Junior 2003; Villares-Junior et al. 2007). Moreover, this species is currently classified as “almost threatened” in the São Paulo State of Brazil (São Paulo 2008). The endangered status of this species deserves special attention and urgent action addressing knowledge of its reproductive physiology, which is the basic premise for the program of fish restocking in the Tietê River (Honji et al. 2011).

Changes in the natural environment due to threats, such as pollution (domestic, industrial, and agricultural pollution), and alteration or obstruction of river flows (like dam’s construction) are severe in the upper reaches of the Tietê River (Silva et al. 2006), resulting in a serious impact on the ichthyofauna, mainly in rheophilic fish in South America. In addition, overfishing is evident in some parts of the Paraná Basin, and obstructions in rivers (i.e., construction of dams) affect the flow of many waterways in South America, which are often so deeply dammed that they sometimes look like a chain of reservoirs (Silva et al. 2006).

Some studies have focused on the effects of environmental changes on the endocrine system. The environmental factors considered include fluctuations in salinity and variations in temperature and photoperiod, which induce internal adjustments within the fish (Vargas-Chacoff et al. 2009; Onuma et al. 2010). Changes in reproductive status also lead to adjustments, such as an altered hormone profile during the annual reproductive cycle (Mousa and Mousa 2000; Onuma et al. 2010). Other studies have focused on animals in differing conditions, such as wild and cultured broodstocks, with the objective of identifying possible endocrine failure, which might be related to the reproductive dysfunction observed in fish farmed broodstocks (Guzmán et al. 2009).

Additionally, as described by Lubzens et al. (2010), in recent years, due to the decline in natural resources due to overfishing and a growing demand for the diversification of marketable products, there is an urgent need for new aquaculture species. However, many commercial fish farms depend on the collection of wild broodstock from the natural environment and their transfer to captivity for artificial reproduction. Unfortunately, the maintenance of broodstock in captivity is not completely successful because many fish exhibit reproductive dysfunctions when reared in captivity (Mylonas et al. 2010). The reproductive dysfunction in captive conditions is the common problem faced in the aquaculture and/or conservation aquaculture of neotropical migratory fish, including S. hilarii. Therefore studies of neotropical migratory fish in both wild and captive environments have relevance.

These facts emphasize the importance of the study of neotropical migratory fish in both wild and captive environments. Nevertheless, there are few studies that aim to characterize the reproductive cycle (Andrade et al. 2006; Honji et al. 2009) and offspring (Araújo et al. 2012) of S. hilarii, and only preliminary data on the endocrine features of this important potamodromous species (Amaral et al. 2007). Thus, more information regarding the reproductive physiology of S. hilarii during gonadal maturation in both wild and captive broodstocks is needed. Additionally, any new insights into the neural and hormonal processes that control the reproductive activities can be applied in conservation programs.

Taking these points into consideration, our studies have focused on identifying and localizing the different pituitary cell types in the adenohypophysis of S. hilarii female, including a morphometric analysis that compares the pituitary cells of females from wild with captive females during the previtellogenic, vitellogenic and regression phases of the reproductive annual cycle. These topics can be used as the basis for future studies in this species, since S. hilarii females present an endocrine dysfunction when transferred from wild environment to captivity.

Independent of the environment in which the animal was captured, changes in S. hilarii ovary size during gonadal recrudescence were macroscopically observed. During the reproductive cycle, a significant increase in GSI was observed from the previtellogenic to the vitellogenic stage (P < 0.01 for both groups), and a significant decrease occurred in the regression stage (P < 0.05 only in wild group (Table 1)). The comparison among females from both groups showed a higher GSI index in the captive group in the regression stage when compared to the wild group (P < 0.05) (Table 1).

The correlation analysis between the environmental variables and pituitary cells morphometry showed that there are no significant correlations among them.

Histological and immunohistochemical analyses of the pituitary

The pituitary gland of S. hilarii appeared to be attached to the ventral region of the hypothalamus, to which it was connected by a thin pituitary stalk. The pituitary gland of S. hilarii consisted of two components, the NH (neurohypophisis) and the ADH (adenohypophysis) with the last one subdivided in three distinguished areas: the rostral pars distalis (RPD), the anterior part of the gland; the proximal pars distalis (PPD), the central part, and the pars intermedia (PI), the posterior part (Figure 2b, c, d). Additionally, the gland presented distinct cells with different tinctorial properties (Figure 2b, c), and the identification and distribution of the pituitary cells were similar to those described in a wide number of teleost species and will be discussed latter.

The RPD was invaded by thin branches of NH extensions and prolactin cells (PRL) were found in this region. PRL cells were not easily identified by the histochemical methods such as Mallory trichrome, and they were PAS-negative. Thus, we decided to use immunohistochemistry to identify these cells. PRL-ir cells were identified using anti-chum salmon PRL (Figure 2e). This antiserum did not cross-react with other ADH cells; no PRL cells were found outside the RPD region. The PRL cells (cell area 60.87 ± 5.73 μm and 57.66 ± 4.57 μm, in the wild and captive groups, respectively) were weakly acidophilic, and organized in a cordonal arrangement.

In the PPD region, the NH showed ramified branches close to the cells, and three different cellular types were recognized. The PPD was characterized by the presence of strongly immunoreactive for growth hormone (GH) cells (Figure 2f, g, h), and two weakly immunoreacting gonadotropins cells (β-follicle-stimulating hormone, β-FSH; and β-luteinizing hormone, β-LH, Figure 3a-h). GH cells were immunoreactive with anti-chum salmon (Figure 2f, g, h), which did not cross-react with other ADH cells. GH-ir cells were found throughout the entire PPD region (clusters of GH cells and isolated GH cells were detected in PPD, as depicted in Figure 2g and 2h, respectively), and they were not found in other locations. These cells were PAS-negative, acidophilic cells, showed an oval shape (65.35 ± 4.99 μm and 69.48 ± 2.06 μm, in the wild and captive groups, respectively), contained an eccentric large, irregular and oval nucleus, and were identified next to the NH processes and blood vessels. The gonadotropes were the most abundant and therefore most easily located cells in the PPD, with some cells also found in the PI. These cells were widely distributed in this region (PPD), and histochemical methods revealed that these cells were PAS-positive, basophilic, and contained vacuoles.

We used nine different antisera to identify the gonadotropin cells (GtHs) (three for β-FSH, three for β-LH, and three for general GtHs) by immunocytochemistry. Two types of antibodies (anti-chum salmon, Figure 3a-h) for β-FSH/β-LH (provided by Dr. Kawauchi and Dr. Shimizu, Figure 3a-d and 3e-h, respectively) were used; GtHs cells were weakly immunoreactive to the anti-chum salmon gonadotropins, and the GtHs-ir cells reacted more weakly to anti-chum salmon than to other anti-chum salmon antisera (i.e., anti-chum salmon GH, PRL and somatolactin). Furthermore, the anti-chum salmon β-LH (Figure 3d, h) showed a stronger immunoreaction than β-FSH (Figure 3c, g), in both antibodies.

We also identified cells showing immunoreactivity to anti-carp α-β-GtH and anti-carp β-GtH (Figure 4a-d), and these cells had a greater reaction when compared with the immunoreactivity to anti-salmon gonadotropin (Figure 3a-h). Additionally, we also used the anti-mummichog antisera (β-FSH, β-LH and α-β-GtH), though, GtH-ir cells could not be recognized using these antibodies (data not shown).

The PI region was characterized by its large, numerous NH branches and the presence of PAS-positive cells. Somatolactin (SL) cells that were strongly reactive with anti-chum salmon antibody were observed (Figure 4e, f), and no SL cells were found outside the PI. The cells (cell area 69.41 ± 2.53 μm and 89.90 ± 4.93 μm, in the wild and captive groups, respectively) were weakly acidophilic, and PAS-positive (Figure 4g). They showed an oval or elongated shape with an evident nucleus (Figure 4f), and were distributed throughout the entire PI region, next to the NH processes. Anti-chum salmon SL antibody surprisingly immunostained the GtH cells (Figure 4h) in the PPD region (as a result of cross-reaction of the antisera against chum salmon SL with gonadotropins cells in PPD region). On the other hand, SL belongs to the growth hormone family, no cross-reaction between the SL antibody with GH cells were observed, only with weakly cross-reaction with PRL cells (Figure 4i) in the RPD region. In the same way, SL-positive cells were not immunostained with the GH or PRL antibodies.

Analysis of pituitary cells optical density of staining, cellular area and nuclear area

The weak immunostaining of gonadotropes (compared with the other immunostained, i.e., PRL, SL and GH cells), added to the fact that it was difficult to distinguish if β-FSH and β-LH immunostained different or the same cell types, and did not permit semiquantitative analyses of GtHs.

The optical density of GH during the reproductive cycle (semiquantitative analyses, Figure 5a) was higher in both groups in the previtellogenic phase, decreased in the vitellogenic stage (P < 0.05), and at the regression stage remained constant in the wild females and increased in captive ones (P < 0.01). On the other hand, the cell and nucleus area values increased from the previtellogenic to the vitellogenic stages in the wild group (P < 0.01), and remained constant in this group at regression stage; but showed no changes in the captive animals (Figure 5b, c). Additionally, between the groups, the nucleus area and nucleus/cell ratio was higher in the wild than in the captive females at all stages (Figure 5c, d; P < 0.01).

Semiquantitative analyses of SL showed significant differences during the reproductive cycle and between environments. The optical density showed similar patterns in females from both groups during the reproductive cycle, it increased from the previtellogenic to the vitellogenic stage and then decreased in the regression stage (Figure 5e) (P < 0.01) to the values previously found in the previtellogenic phase in the wild group. Despite decreasing in the regression stage, the optical density remained lower in females from the wild than in captive females (P < 0.01). We also observed an increase in the cell and nucleus area with the onset of vitellogenesis in the captive group (Figure 5f, g). At this stage, between the groups, both the cell and the nucleus areas were larger in captive animals than in wild females (Figure 5f, g) (P < 0.01) and the nucleus area then decreased during the regression phase, but only in captive females (P < 0.01). The same patterns of variation were found in the nucleus/cell ratio (Figure 5h) in captive females. Furthermore, the nucleus/cell ratio was higher in wild than in captive females during the previtellogenic and regression stages (P < 0.01).

Semiquantitative analysis of PRL showed no significant differences neither throughout the reproductive cycle, in both groups, nor between wild and captive females.

Previous studies on the reproductive cycle of S. hilarii showed that animals captured in different points of Tietê River did not differ in terms of morphological or anatomical arrangement of ovaries (Honji et al. 2009), therefore, we assumed that they belong to the same group. Additionally, as expected, GSI values increased independently of the environment, wild or captivity (Honji et al. 2009). In this way, as the ovaries grow in S. hilarii, the largest GSI values were found in animals with fully developed ovaries (vitellogenic stage), and the lowest GSI values were registered during the previtellogenic stage, suggesting a normal ovarian development even in captivity. On the other hand, captive females showed a dysfunction in the endocrine system when compared to wild animals. It seems that the maintenance in captivity blocks spontaneous ovulation and consequently reproduction, demonstrating the importance of basic studies of the pituitary-gonad axis, as suggested by Amaral et al. (2007) and Honji et al. (2009 2011), for S. hilarii females.

Although the ADH and the distribution of adenohypophyseal cells in S. hilarii were similar to those described in a large number of teleost species (for reviews, Kawauchi and Sower 2006; Levavi-Sivan et al. 2010), the identification and detailed distribution of the various ADH cell types in reophilic fish, have not been performed yet in South American teleost species. Beyond the identification of these cells, the comparison of them between wild vs. captive females throughout the reproductive annual cycle and different environments have never been performed in potamodromous fish.

GtHs play critical roles in the regulation of reproductive processes, optimal rates of synthesis and secretion of GtHs are necessary and critical for successful gonadal maturation and spawning (Zohar et al. 2010; Levavi-Sivan et al. 2010). In captivity, neotropical females tend to exhibit reproductive dysfunctions, which include the failure to undergo final oocyte maturation and an absence of egg production or reduction in egg quality and quantity (Zohar et al. 2010; Bobe and Labbé2010), as observed in S. hilarii (Honji et al. 2009 2011). Furthermore, failure to spawn in captive females has been identified to be due to dysfunctions in the release of GtHs from the pituitary gland (Guzmán et al. 2009). Therefore, the endocrine system of females is adversely affected, and as the normal functioning of GtHs release is changed, the reproductive system fails (Mylonas et al. 2010; Levavi-Sivan et al. 2010). However, before planning the manipulation of ADH hormone secretion, which will enhance fish reproduction in captivity, it is first necessary to establish the morphometrical parameters and cell physiology of all cell types involved in the reproductive process.

To determine if β-FSH and β-LH are produced by only one type of cell or by different cell types in the S. hilarii pituitary gland, two adjacent sections of the pituitary were immunostained with β-FSH and β-LH antibodies and the staining was compared at high magnification. Nevertheless, it was difficult to distinguish if β-FSH and β-LH immunostained different cells or the same cell type, because different adjacent sections generally showed different patterns of immunostaining in many cells in the PPD region; additionally, in others cells, there were also cells immunostained with both antibodies.

In the present study, we showed that pituitary gonadotropes are more immunoreactive to β-LH antibodies than to β-FSH antibodies. This result can be associated to the fact that the molecular structure is more conserved, both in structure and primary sequence, in fish β-LH subunit than in β-FSH (Levavi-Sivan et al. 2010). Further research with the aim to investigate this question, using specific antisera for Characidae family, will be necessary to answer this question regarding gonadotropins in South America fishes, including S. hilarii. Additionally, due to lack of strong positive immunoreactions with GtHs cells in S. hilarii, our group decided to study in parallel with the present study, the β-FSH and β-LH gene expression in wild and captive females (Moreira et al. in preparation).

GH, SL, and PRL constitute a family of structurally related hormones (Kawauchi and Sower 2006) that have multiple physiological functions. A number of investigations on the GH family in fish were focused on their pituitary distribution and the characterization of these hormones. However, few studies have been conducted on seasonal variations in these molecules or the influence of environmental changes on the GH family in non-salmonid fish species (Vargas-Chacoff et al. 2009; Fiszbein et al. 2010), and no studies on the GH family in South America fish (reophilic Neotropical fish species), have assessed the impact of a migration impediment on these cells. In this study, we used anti-chum salmon GH, which showed a quite specific immunostaining within the GH cells of S. hilarii, as demonstrated similarly in other teleost fishes. The distribution of GH cells followed the patterns already described in others fish species (Vissio et al. 1997; Segura-Noguera et al. 2000; Pandolfi et al. 2001a; Sánchez-Cala et al. 2003; Borella et al. 2009). GH-ir cells have been located in the PPD region, either isolated or clustered, and no cross-reactivity of other cell types occurred in S. hilarii. However, some species showed cross-reactivity of GH antisera with PRL cells and of PRL antisera with GH cells or with PAS-positive cells of the PI (SL cells) (García-Hernández et al. 1996; García-Ayala et al. 1997).

Our results clearly demonstrate that GH is potential involved in the S. hilarii reproductive cycle. GH cells in vitellogenic females were significantly larger cells and had larger nucleus areas than did GH cells in previtellogenic females in the wild. Additionally, females from both environments showed a lower optical density from the previtellogenic to vitellogenic stages. This result could be correlated with major cellular synthesis activity, and probably with an increase in hormone release. In Sparus aurata, the pattern of GH protein expression showed differences during the seasons, with the highest values recorded during the summer (reproductive period) and the lowest values recorded during autumn (Vargas-Chacoff et al. 2009). In the same species, plasma GH levels increased progressively during late spring-early summer, whereas during autumn/winter, circulating GH remained low (Mingarro et al. 2002). Considering that S. hilarii vitellogenic females were captured mainly during spring/summer and the previtellogenic/regression animals were caught mainly in the autumn/winter, our results were consistent with the variation of GH profile reported during the reproductive cycle. Additionally, such increases in GH cell activity in the vitellogenic stage could be related to increased feeding during this period, as was observed in other teleost fish (Holloway and Leatherland 1998). It is also probable that in this species, GH secretion is regulated by melatonin, as observed in rainbow trout (Oncorhynchus mykiss), and/or by GnRH as suggested by Canosa et al. (2008), who demonstrated that the circulating levels of LH and GH increase at the time of ovulation in goldfish.

In this study, SL cells in S. hilarii were small, weakly acidophilic, and PAS-positive. They showed an ovoid or elongated shape with a conspicuous nucleus and were distributed in all regions of the PI, close to the NH processes. Whereas the SL cells were PAS-positive in S. hilarii, PAS methods in several other teleosts have revealed two different types of SL-ir cells in the PI region. Biochemical analyses have demonstrated both the existence and the absence of N-glycosylation sites in the SL. Briefly, in salmonid fish, SL-ir cells were PAS-negative as a result of the non-glycosylated SL form (Rand-Weaver et al. 1991; Kaneko 1996). PAS-negative (non-glycosylated form) SL-ir cells were also observed in some other non-salmonids such as S. aurata (Villaplana et al. 1997), Diplodus sargus (Segura-Noguera et al. 2000), Trachinus draco (Sánchez-Cala et al. 2003) and Arapaima gigas (Borella et al. 2009). However, PAS-positive cells, as a result of the glycosylated SL form, were identified in several fishes (Rand-Weaver et al. 1991; Kaneko 1996). In addition, SL hormone in S. aurata, was observed as both the glycosylated and non-glycosylated form in adults (Cavari et al. 1995). These data are similar to those reported for Solea senegalensis, which showed PAS-positive and PAS-negative cells in the same species (Pendón et al. 1998). Different forms of SL hormones were also observed during the ontogeny of Cichlasoma dimerus, with the glycosylated form of SL being expressed in adults, and a non-glycosylated form of SL appearing in larvae (Pandolfi et al. 2001b). In S. hilarii, SL it is probably glycosylated, considering that SL-ir cells were PAS-positive in this species.

The distribution and localization of SL cells in S. hilarii is similar to other teleosts using antisera against chum salmon SL (see review of Kaneko 1996; Kawauchi and Sower 2006). However, some species showed reactions to anti-GH and anti-PRL in the PI. These results could possibly be due to a cross-reaction of these antisera (GH and PRL) with SL cells because of structural similarities. Although the migration of cells (GH and PRL) to the PI region during the pituitary development cannot be excluded (Farbridge and Leatherland 1986), our results showed a cross-reaction of antisera against chum salmon SL with PRL cells in the RPD region and GtHs cells in PPD region. Our pre-adsorption tests and histology method (PAS stained) confirmed this cross-reaction among the SL antisera with PRL and GtHs cells. Similar cross-reactions were reported with anti-gonadotropin in SL cells in Xiphophorus maculatus and Poecilia latipinna (Margolis-Kazan et al. 1981; Batten et al. 1975; Batten 1986). According to these authors, the SL and GtHs hormone can contain antigenically similarity, which might be difficult to remove during biochemically purification.

The larger nuclear area and more intense optical density in SL cells in pituitary of vitellogenic S. hilarii suggest a possible role of SL in the reproductive physiology of this species. Taking into consideration the variation of SL hormone during the reproductive cycle, Mousa and Mousa (1999 2000) identified seasonal variations in the synthesis and secretion of SL during sexual maturation and spawning in Oreochromis niloticus and Mugil cephalus. In O. niloticus the synthetic and secretory activity of the SL-ir cells increased during sexual maturation and spawning (Mousa and Mousa 1999). In coho salmon, SL levels increased during smoltification and decreased when the smoltification was complete. They then remained low until prespawning, reaching a peak at spawning (Rand-Weaver et al. 1992). During the vitellogenesis process, there is an increasing demand for calcium due to the involvement of this cation in the transport of vitellogenin from the liver to support the development of the embryo (Lubzens et al. 2010). This suggests that SL is indirectly or directly linked with reproductive physiology (Kakizawa et al. 1993 1995). Our results regarding the cellular morphometrical parameters of SL in captive animals are consistent with studies in M. cephalus, in which the synthetic activity of SL was higher in captivity than wild animals (Mousa and Mousa 2000). Other studies suggest that SL may be responsible for the regulation of associated processes that indirectly affect reproductive activity (Vissio et al. 2002), i.e., metabolism (especially, with lipogenic enzymes and lipid mobilization), temperature variation (Mingarro et al. 2002; Vargas-Chacoff et al. 2009), feeding, and ionoregulation (Kakizawa et al. 1995).

The present results also show that chum salmon PRL antiserum can be recommended for the identification of PRL cells in S. hilarii. The distribution of these cells followed the patterns already described in other teleost fishes (Kawauchi and Sower 2006). PRL-ir cells have been located in the RPD region, and no cross-reactions with other cell types occurred in this species. However, some other species showed cross-reaction of PRL antisera with GH cells (Pandolfi et al. 2001a). Additionally, our results indicated no changes in PRL pituitary content during sexual maturation and also no differences between the groups (wild and captive females). Several studies have analyzed the pituitary PRL in teleost species, and it was established that PRL is an important hormone involved in osmoregulatory processes in these animals. It is essential for ion uptake as well as for reducing ion and water permeability of osmoregulatory surfaces (Sakamoto and McCormick 2006). Furthermore, when important differences in temperature and photoperiod occur, variations in PRL gene and hormone expression levels were observed in fish, suggesting that the seasonal cycle of PRL is also influenced by seasonal acclimatization (Figueroa et al. 1997; Vargas-Chacoff et al. 2009). In C. dimerus, pituitary levels of PRL were significant higher when animals were exposed to a long photoperiod rather than a short photoperiod (Fiszbein et al. 2010). Further, in Cyprinus carpio, it has been shown that photoperiod constitutes a particularly relevant modulator in the neuroendocrine cascade that activates PRL transcription (Figueroa et al. 1997), due to the modulation of PRL secretion by melatonin, a highly conserved feature in vertebrates (Falcón et al. 2007). In S. hilarii, even considering the annual variations of water temperature and photoperiod, there was no difference in PRL content, as described in the other fish species.

In conclusion, this study in S. hilarii broadly supports the idea that beyond the gonadotropes, other adenohypophyseal hormones are involved in reproduction in teleost fish and may be responsible for the regulation of associated processes that indirectly affect reproductive status. Our data, together with the findings previously described by Amaral et al. (2007), who observed dysfunction in GtHs and in progestagen synthesis, in captive females, suggest that the reasons for the reproduction failure of migratory fish in captivity are complex. Thus, analyses of the brain, especially the higher control of the hypothalamic gonadotropin-releasing hormone (GnRH), should also be conducted.