Histol Histopathol (2001) 16: 423-437 Histology and http://www.ehu.es/histol-histopathol Histopathology Cellular and Molecular Biology The hepatocytes of the brown trout (Salmo trutta f. fario): A quantitative study using design-based stereology E. Rocha112, R.A.F. Monteiro112, M.H. Oliveiral and M.W. Silva' Laboratory of Histology and Ernbryology, lnstitute of Biornedical Sciences Abel Salazar (ICBAS), and 2lnterdisciplinary Center of Marine and Environmental Research (CIIMAR), University of Oporto, Porto, Portugal Summary. A stereological study was performed on volume ratio. Interspecific similarities and differences brown trout hepatocytes aiming to disclose whether were disclosed. For example, the number of there are basic gender differences when minimal levels hepatocytes/cm3 of parenchyma of brown trout was of sex hormones exist, and also to establish a platform much lower than those reported in rainbow trout, but in for both interspecific comparisons and physiological both trouts femaies seem to have an higher cell number. correlations. We used the so-called "design-based In addition, when comparing the size of hepatocytes of stereology" (with no shape, size or orientation brown trout with that from other fish and mammals it assumptions) and also some new related statistics. Two- was suggested that major interspecific differences exist. year-old brown trout were collected in April, and the livers were fixed by perfusion. From liver slicing to Key words: Hepatocyte, Liver, Teleost, Stereology, microscopical field selection, systematic sampling was Disector used. Stereology was applied at light and electron microscopy. Target parameters were the relative and total hepatocyte number, the mean individual hepatocyte lntroduction volume and surface, and also both relative and total volumes, and surfaces, either of organelles or of cell For severa1 years, we have been characterizing the compartments. Observed variability was usually high, normal histology of the brown trout liver, aiming at ' but the precision of estimates was proved to be globaily sound bases for histopathological and physiological adequate facing the true biological variation amongst approaches. The gathered insight also has a great interest specimens. Females had more hepatocytes per liver for comparative morphology. We have made both (1.79~10~vs. 1.12~10~).Considering the individual qualitative and quantitative accounts, at light and hepatocytes, whereas no gender differences were electron microscopy leve1 (Rocha et al., 1994a,b, 1995, detected in the cell volume, males had higher values of 1996, 1997, 1999). Although highly valuable, qualitative nuclear volume (199 vs. 151 pm3) and surface (170 vs. data are nonetheless subjective and not adequate for 131 pm2), endoplasmic reticulum volume 1300 vs. 824 statistics; so, stereological approaches are also desirable. pm3), and microvilli volume (82 vs. 54 pmI ) and surface From al1 liver cells, hepatocytes have been the major (1445 vs. 975 pm2). However, when dealing with target for quantification, either in mammals (e.g., Loud, quantities per liver, gender differences were found only 1968; Weibel et al., 1969; Blouin et al., 1977; Bioulac- in the volumes of dense bodies (56 vs. 97 mm3) and of Sage et al., 1984; Jack et al., 1990a,b) or in fish. In these, residual cytoplasm (169 vs. 341 mm3) - both volumes most data carne from rainbow trout, Oncorhynchus mykiss were higher in females. Functional implications of data (e.g., Arnold et al., 1995, 1996a,b; Hampton et al., 1989; are discussed, namely that females seem to have basic Moutou et al., 1997). Despite proved relevant in structural traits for coping with the later demands of toxicology (Braunbeck, 1998), stereological data on other breeding. Data also support that structural remodelling fish hepatocytes are rather scarce, but some are available, of hepatocytes occurs after breeding, urging to pursue e.g., for medaka, Oryzias latipes (Hinton et a1.,1984; seasonal studies (namely on lysosomes). We advanced Braunbeck et d., 1992), for golden ide, Leuciscus idus the hypothesis that genders differ in microvilli surface (Segner and Braunbeck, 1990), for immature brown trout just to maintain an optimal physiological surface-to- (Schramm et al., 1998), and for an immature demersal fish, Solea ovata (Au et al., 1999). Offprint requests to: Eduardo Rocha, Laboratory of Histology and Since the breakthrough of the disector (Sterio, Ernbryology, lnstitute of Biornedical Sciences Abel Salazar (ICBAS), 1984), stereology moved from what is now called "the Largo Prof. Abel Salazar no. 2, 4099-003 Porto, Portugal. Fax: 351 22 old assumption-based methods" to the "design-based 206 22 32 or 1 801 992 2122. e-rnail: [email protected] approaches". Essentially, the old techniques for Stereological study of brown trout hepatocytes estimating particle size and number relied upon strict had a total length of 28 cm [26-311, whereas the females assumptions of shape, size and orientation of the weighted 280 g [200-4001 and had a total length of 29 particle, whereas the design-based methods do not, cm [26-331. The animals were being hand fed twice a assuring a priori unbiased estimates not needing day with highest quality commercial trout pellets validation studies (Gundersen et al., 1988; Howard and (composition on the label: protein 42%, lipid 18%, Reed, 1998). When correctly used by the experimenter, cellulose 1.5%, ash 11%, phosphorus 1.8% - from design-based stereology grants accuracy, i.e., estimates Alpis - A. Coelho & Castro Lda., Póvoa de Varzim, will always vary around the true value. On the contrary, Portugal). Collection took place during late April when measuring with model-based estimators - the (middle of previtellogenesis period of female breeding assumptions being very rarely tested - we might, by cycle). Ovarian histology was typical of that period, chance, stumble across the true answer without knowing showing mainly growing oocytes mostly in a primordial it, for bias is invisible and undetectable within any one cortical alveoli stage, and thus with a scarce perinuclear experiment (Howard and Reed, 1998). Even when yolk deposition (Selman et al., 1987; Washburn et al., comparing control versus experimental groups, bias can 1990). Testes were atrophic, displaying only be significantly different between groups, despite using spermatogonia. exactly the same assumption-based methods (e.g., Mendis-Handagama, 1992). A word of caution is in Fixation and tissue processing order to remember that either using the old or the more recent methods, systematic biases can be induced by To avoid eventual diurna1 variations in the non-stereological procedures (such as shrinkage), hepatocyte content, al1 collections were made during the stemming from tissue processing. morning. Before fixation, the trout were anaesthetized In mammals, only two studies seem to exist that by immersim in 0.3 mVL aqueous solution of ethylene have used more recent stereological principles on rat glycol monophenyl ether, and then weighed. Liver was hepatocytes (Jack et al., 1990a,b). As to fish fixed by perfusion via hepatic portal vein or one of its hepatocytes, and except in one of our studies in which major tributaries. Details of the procedure are given modern concepts were introduced (Rocha et al., 1999), elsewhere (Rocha et al., 1994a). Briefly, 2-4 m1 of a al1 the other studies were solely based on classical chilled (15 "C) and heparinized isosmotic Ringer-like stereology and sampling theories (Weibel, 1979). The solution were perfused at physiological flow rate of situation in hepatology contrasts with that in about 5.2 ml/min/kg body weight (Schrnidt and Weber, neuroscience, where design-based approaches have been 1973; Hampton et al., 1985). Subsequently, 20-40 m1 of largely used; to a point that editors urge both the authors chilled fixative (15 "C) were introduced as described; and the referees to adhere to the design-based unbiased fixative was 2.5% glutaraldehyde in 0.1M phosphate estimates (Coggeshall and Lekan, 1996; Saper, 1996). buffer @H 7.3). This study was focused on hepatocytes, evaluating After removal, the liver was quickly weighed and its their number, volume, surface, and dissecting their volume was determined by Scherle's method (1970). The organeollar compartments in relative and absolute liver was then chopped into 3 mm thick slices. A series quantities. We used a design-based approach which, in of systematic random sampling steps (subsequently association with recent statistical practices, enabled us to applied to groups of smaller tissue fragments) was evaluate ow procedure, extracting guidelines for further carried out to obtain a final sample of pieces (<OS mm quantification of brown trout hepatocytes in future in diameter) to be handled for transmission electron studies. As in a previous approach (Rocha et al., 1997), rnicroscopy (TEM). This systematic scheme is known young adults were sacrificed at the previtellogenesis for its high efficiency, assuring a priori equal sampling period of the female breeding cycle, found to be the best probabilities for al1 organ zones (Gundersen, 1986; moment for seeking intrinsic sexual differences not too Gundersen and Jensen, 1987). closely related either with vitellogenesis or with The tiny fragments were further immersed in the spawning. same fixative as that used in perfusion for 3h (at 4 "C) and then rinsed for 2h in 0.1M phosphate buffer (at Materials and methods 4 "C). Postfixation was done with 1% Oso4 in 0.1M phosphate buffer, for 2h (at 4 "C). After dehydration in Animals ethanol (50%, 75%, 90%, 95%, at 4 "C, and 100% p.a., at room temperature) and two passages in propylene -o-year-old male (n=7) and female (n=7) brown oxide, the pieces were embedded in epoxy resin at no trout (Salmo truffa fario Linnaeus, 1758) were randomly specific orientation. For achieving this goal, the pieces fished with a net within the aquaculture pools of the were rolled inside the resin with the help of dissection Posto Aquícola do Torno (Amarante, Portugal), and fish needles. Thus, our procedure aiiowed the generation of were sacrificed in four consecutive days.