Gut morphology and metallothionein immunoreactivity in Liza aurata from different heavy metal polluted environments

Abstract

A critical analysis was made as to whether changes in morphology and histomorphology of fish intestinal mucosa could be appropriate for rapid application in field monitoring programs for heavy metal pollution. Equivalent gut samples of the gold grey mullet Liza aurata from different polluted environments were simultaneously treated using morphological, histomorphological and immunohistochemical methods. The morphological aspects of the mucosal folds and the characteristics of the mucous goblet cells seemed to vary according to the environmental pollution, as well as to the presence and distribution of metallothionein immunoreactivity. On the basis of these findings, the use of the gut fold morphology test is suggested as an expertise‐independent, cost‐effective and rapid prognostic biomarker for field heavy metal monitoring programs.

Keywords:

• Wild mullet
• heavy metal pollution
• gut morphology
• metallothioneins

Introduction

It is well known that contaminated environments can produce numerous morphological and functional alterations in marine organisms (Wendelaar Bonga & Lock 1992; Dang et al. 1999; de la Torre et al. 2000; Fossi et al. 2002; de Oliveira Ribeiro et al. 2002; Kilemade et al. 2002; Stentiford et al. 2003). The absorbed chemical compounds can interact at the cellular level with endogenous substances, causing dangerous biological effects which can impair the life quality not only of the exposed organisms, but of the whole ecosystem. The analysis of these effects on different organisms can furnish relevant data about future chronic pathological changes. Different biomarkers have been used to evidence the biological effects of various pollutants on aquatic organisms, both in natural and experimental conditions. The utility and validity of these biomarkers are determined by the possibility they afford to establish a definite correlation between easily verifiable biological alterations and contaminant exposure, and to give information on the biological effects of pollutants rather than on pollution environmental levels (Au 2004). Besides chemical and physical parameters, biochemical, physiological and histocytopathological biomarkers have been increasingly recognized as important tools to evaluate the impact of environmental pollutants. In particular, fishes are promising models for environmental risk assessment, and the use of some fish cytological biomarkers has been recommended for pollution monitoring in coastal environments (Au 2004). Fishes can be considered as the end point of the trophic chain, since they can accumulate toxicants not only from water but also from food (van der Oost et al. 2003). Previous studies evidenced morphological and histomorphological alterations in fish gills, especially following heavy metal exposure (Mauceri et al. 2002; Fracacio et al. 2003; Stentiford et al. 2003), and in kidney and liver (see Au 2004 for a review), but few of them addressed the gut (Radecki et al. 1992; Banerjee & Bhattacharya 1994; Pederzoli et al. 1996; Kamunde et al. 2001). External morphological biomarkers concerning fin erosion or skeletal malformations were also recommended in various monitoring projects (International Council for the Exploration of the Sea, ICES, 1996).

In the present research we detect morphological and histomorphological alterations in the gut of the gold grey mullet Liza aurata in natural heavy metal polluted environments, and compare them with the presence and distribution of mucous cells, as well as of immunohistochemically detected metallothioneins (MTs). MTs are a family of metal binding proteins which play a pivotal role in the process of metal detoxification. Because there is a close relation between exposure to heavy metals, their accumulation and MTs induction, these proteins have been proposed as a useful biomarker for eco‐toxicological studies of aquatic animals (Chan 1995; Pedersen et al. 1997; Schlenk et al. 1997; Irato et al. 2001). Liza aurata feeds principally on small benthic organisms and detritus (Ben‐Tuvia 1986), and, due to these feeding habits, seems to be particularly adapted to accumulate sediment‐associated contaminants. Furthermore, this species has been previously used as a biological indicator for the detection of heavy metals (Mzimela et al. 2003) and metallothioneins (Filipović & Raspor 2003). The widespread distribution of L. aurata and its resistance to different environmental conditions make it a good candidate for the role of biomarker model, especially for chronic toxicant exposure. For the present study, different specimens of L. aurata were sampled from the brackish lakes of Faro and Ganzirri (Capo Peloro, Messina, Italy). The chemical characteristics and the heavy metal presence in the sediments of these lakes were previously studied (Giacobbe et al. 1996; Munaò et al. 2000; Mauceri et al. 2002). As a reference site, the protected area of Lake Verde was chosen (Regione Sicilia 2002). The knowledge of the gut morphological alterations, both at macroscopic and microscopic level, can be used as a rapid, easy to detect, and cost‐effective biomarker for a preliminary assessment of field contamination.

Materials and methods

A total of 44 immature specimens (length 18–20 cm) of the gold grey mullet L. aurata (Perciformes Mugilidae)—9 from Lake Verde, 15 from Lake Faro and 20 from Lake Ganzirri (Messina, Italy)—were collected during spring and autumn in 2001, 2002, and 2003.

Lakes Faro and Ganzirri (Capo Peloro, Messina, Italy) were previously studied from a chemical point of view, and are considered to be highly polluted environments, in particular for the presence of elevated concentrations of heavy metals (Giacobbe et al. 1996; Mauceri et al. 2002). The Faro brackish lake is characterized by the presence of hydrogen sulphide and many photosynthetic sulphur micro‐organisms. During summer, the highest water temperature at the interface is 29°C while at a depth of 20 m the temperature is 19°C. The following concentration of metals were detected: Al 36.56 µg l⁻¹; Fe 20.7 µg l⁻¹; Cd 48.4 µg l⁻¹; Pb 5.2 µg l⁻¹, Hg 55.5 µg l⁻¹.

Ganzirri is a brackish lake 7 m deep; during summer the highest water temperature is 29.8°C. The metals present are: Al 21.6 µg l⁻¹; Fe 74.7 µg l⁻¹; Cd 68.4 µg l⁻¹; Pb 1.2 µg l⁻¹; Hg 32.5 µg l⁻¹; no organic compounds or herbicides were detected (Munaò et al. 2000). Lake Verde is routinely monitored as a protected area of the Regione Sicilia (Regione Sicilia 2002) and for this reason is considered as a reference site. Fishes were killed with a lethal dose of 3‐aminobenzoic acid ethyl ester‐methanesulfonate, 1‰ in sea water (SERVA, Germany), and their gut was fixed in 4% p‐formaldehyde in phosphate buffer saline pH 7.4 (PBS) and thoroughly rinsed in PBS. The next treatments varied according to the technique used. For the present research, the intestinal portion near the openings of the pyloric caeca was used.

Gross morphology

Stereo microscope

The PBS‐rinsed gut samples were positioned flat on a cork support and observed through Stemi 2000‐C stereo microscope (Zeiss, Germany) equipped with a Camedia 3030 digital camera (Olympus, Japan). The gross morphology of the intestinal surface, the shape of the intestinal folds and their arrangements were taken into account as morphological markers of environmental pollution.

Scanning electron microscope (SEM)

The PBS‐rinsed gut samples were dehydrated through graded ethanol to hexamethyldisilazane (HMDS, Carlo Erba, Italy), gold‐coated and examined with a Leo Stereoscan 440 Scanning Electron Microscope (LEO Electron Microscopy Ltd, UK). The morphology of the folds and of their surface was recorded and photographed.

Histomorphology

PBS‐rinsed gut samples, from the same specimens as those utilized for the gross morphology, were dehydrated and Paraplast‐embedded. Five‐micron‐thick transverse sections were stained with Haematoxylin–Eosin (H–E, Bio‐Optica, Italy) and Alcian Blue pH 2.5–PAS (Bio‐Optica, Italy). The sections were examined at a BX60 Olympus microscope, visualized through the Color‐View Camera (Olympus, Japan) and acquired by the software AnalySIS (Soft Imaging System, USA). Public domain software Image J (http://rsb.info.nih.gov/ij/) was used to measure the epithelial thickness and to count the different stained goblet cells. The epithelium of the mid‐height of the intestinal fold was used to measure the epithelium thickness on H–E sections. Measurements were made for three intestinal villi or ridges per section, and repeated in five different sections for each specimen. The mean and the standard deviation were calculated, to determine significant differences among the three groups from Lake Faro, Lake Ganzirri, and Lake Verde.

An Alcian Blue pH 2.5–PAS reaction was carried out to reveal neutral and acidic mucous secretions in the goblet cells. The number, percentage, and density of the different stained goblet cells (Alcian Blue‐, PAS‐, and Alcian–PAS‐positive cells, respectively blue, red, and violet) were calculated with the Image J program, considering 4 mm of the epithelial linear profile in the Alcian–PAS‐stained intestinal wall sections. The count was repeated for three sections for each specimen.

Immunohistochemistry

Dewaxed sections were rinsed in PBS, treated with 2% H₂O₂ in PBS to stop endogenous peroxidase activity, and incubated overnight in a moist chamber at room temperature with the polyclonal anti‐MT antiserum (1:100 in PBS, biomol, Germany). After a rinse in PBS (30 min), the sections were incubated for 1 h at room temperature with a peroxidase or FITC conjugated sheep anti‐rabbit IgG (Sigma, USA), diluted 1:100 in PBS. The peroxidase was stained by a fresh solution of 3.3‐diaminobenzidine‐4HCl (DAB, 30 mg 100 ml⁻¹) and H₂O₂ (0.01%) in PBS. The specificity of the immunoreaction was controlled by substitution of the primary antiserum with PBS or replacement of the primary antiserum with non‐immune rabbit serum diluted 1:200.

Results

Gross morphology

Stereo microscope

The aspects of the intestinal surface varied according to the site of sampling and, to a lesser extent, to the specimen used. The different morphology and the arrangements of the intestinal folds (villi and/or ridges) were subdivided into four groups.

	Group 1: normal leaf‐shaped villi with a rounded apex (figure 1a);
	Group 2: short lamellar ridges with a flattened apex (figure 1b);
	Group 3: long lamellar ridges with a flattened apex and random distribution (figure 1c);
	Group 4: long extended lamellar ridges, with a flattened apex, arranged in a herring‐bone pattern (figure 1d).

Figure 1 Intestinal fold morphology from the three different lakes.a, Lake Verde. Leaf‐shaped villi with a rounded apex, characteristic of Group 1; b, Lake Faro. Short lamellar ridges with a flattened apex, characteristic of Group 2; c, Lake Ganzirri. Long lamellar ridges with a flattened apex, characteristic of Group 3; d, Lake Ganzirri. Long lamellar ridges with a herring‐bone pattern, characteristic of Group 4. Scale bar, 200 µm.

In Table I, the numbers of fish from the different lakes falling within the four groups are reported, and in figure 2 the percentages of the four groups detected in fishes from the different lakes are shown.

Table I. Number of fishes from the different lakes showing fusion of the finger‐shaped folds to form ridges (Groups 1 to 4).

	L. Verde	L. Faro	L. Ganzirri
Group 1	3	1	3
Group 2	3	6	5
Group 3	3	4	4
Group 4	0	4	8

Figure 2 Percentages of the four intestinal fold morphology groups from the three different lakes: a, Lake Verde, b, Lake Faro and c, Lake Ganzirri.

Scanning electron microscope

The observations of equivalent parts of intestine through the SEM confirmed the gross morphology descriptions concerning the aspect of the normal leaf‐shaped villi and of the lamellar ridges with a flattened apex (figure 3a–c). The surface of the folds did not seem to present any difference.

Figure 3 Intestinal fold aspects at the SEM.a, Lake Verde, leaf‐shaped villi, Group 1; b, Lake Faro, short ridges with flattened apex, Group 2; c, Lake Ganzirri, V‐shaped long ridges, Group 3. Scale bars,100 µm.

Histomorphology

The proximal intestinal mucosa showed folds lined by a columnar absorptive epithelium characterized by intercalating mucous goblet cells. The number of the mucous cells and their Alcian Blue–PAS staining, as well as the epithelial thickness, differed according to the site of sampling (figure 4a, b). In the specimens from Lake Verde, the mucous goblet cells were variously stained with Alcian Blue–PAS, and their secretion appeared red, light blue, or violet. Most of the mucous goblet cells detected in the specimens from Lake Faro and Lake Ganzirri seemed to be smaller and empty. Furthermore, they appeared alcianophilic only, and for this reason light‐blue‐stained (figure 4c, d). The counting of goblet cells showed that there is a marked increase in the percentage of the Alcian Blue‐stained cells in samples from Ganzirri and Faro, reaching percentages of 84% and 90%, respectively, and a decrease in the Alcian Blue–PAS‐stained cells of from 34%, found in Lake Verde to 13% in Lake Ganzirri, and 9% in Lake Faro. The percentage of Alcian Blue‐ or PAS‐positive goblet cells found in the intestinal epithelium (4 mm of its linear extension) is reported in figure 5a–c.

Figure 4 Histomorphological aspects of the intestinal wall. Alcian–PAS. a, Lake Verde, normal intestinal folds (scale bar, 100 µm); b, Lake Ganzirri, ruffled intestinal folds (scale bar, 100 µm); c, Lake Verde, 24‐µm‐thick intestinal epithelium with numerous roundish mucous goblet cells. Arrowheads: Alcian Blue–PAS‐positive mucous cells, arrow: Alcian Blue‐positive mucous cells (scale bar, 20 µm); d, Lake Ganzirri, 34‐µm‐thick intestinal epithelium with small depleted Alcian Blue‐positive mucous cells (scale bar, 20 µm).

Figure 5 Percentage of the Alcian Blue‐, PAS‐ and Alcian Blue–PAS‐positive mucous goblet cells from the three different lakes: a, Lake Verde, b, Lake Faro and c, Lake Ganzirri.

The surface epithelium seemed to become thicker in fishes from Lake Ganzirri and, to a lesser extent, in those from Lake Faro (Table II).

Table II. Intestinal epithelial thickness (µm) of Liza aurata from the different environments (mean±SD).

L. Verde	L. Faro	L. Ganzirri
26.5±2.4	30.6±3.2	42.3±2.1

Immunohistochemistry

MTs immunoreactivity was only detected in the gut epithelium of fishes collected in Lake Faro and Lake Ganzirri. The immunoreactivity was located in the epithelial cells, mainly in those at the base of the intestinal folds. Some of the immunoreactive cells appear to be, from a morphological point of view, similar to mucous goblet cells, but others to the intestinal absorptive cells. More numerous cells were observed in samples from Lake Ganzirri (figure 6a) than in fishes from Lake Faro (figure 6b). No immunoreactivity was observed in samples from the control lake. The specificity controls gave negative results.

Figure 6 MTs immunoreactive cells in the intestinal epithelium.a, Lake Ganzirri, numerous immunoreactive cells along the intestinal fold epithelium. Immunofluorescence; b, Lake Faro, a few immunoreactive cells at the base of the intestinal folds. Immunoperoxidase. Scale bars, 50 µm.

Discussion

The results reported in this study evidence the fact that environments considered at a low level of pollution (Lake Faro is generally used for the depuration of commercially used mussels) have a fish population presenting morphological alterations, both at macroscopic and microscopic levels. The morphological aspects of the intestinal folds were graded in four groups. The morphological alterations seemed to parallel the concentrations of heavy metals previously studied in the brackish environments of Lake Faro and Lake Ganzirri (Giacobbe et al. 1996; Mauceri et al. 2002). In particular, no specimen from the protected area of Lake Verde displayed the Group 4 affected folds, even if minor alterations could be detected in a few specimens. The Group 4 impairment was found in 27% of the Lake Faro specimens (low level of pollution) but in 40% of those of Lake Ganzirri. At the latter site, the highest levels of heavy metal pollution were found, in particular for Cd and Fe. The flattened appearance of the tips of the intestinal villi (ascribed to Group 2) and the presence of intestinal ridges (Groups 3 and 4) were probably caused to the fusion of leaf shaped villi. Abnormal proliferative and apoptotic processes were observed in the intestinal epithelium of L. aurata from the same polluted area (Ferrando et al. 2005); the intestinal epithelium renewal was impaired, gut surface morphology showed the herring‐bone pattern typical of Group 4, and the intestinal surface and nutrient uptake were reduced. Furthermore, a thicker epithelium was observed in the polluted lakes. Intestinal villus fusion was also observed in the proximal intestinal portion of the fish Channa punctatus, experimentally exposed for long periods to different kind of pollutants (Banerjee & Bhattacharya 1995), but the present report is the first one on feral fish from a natural polluted environment. Gill abnormality for secondary lamellae fusion was, instead, reported in flounder captured from British estuaries differently impacted by contaminants (Stentiford et al. 2003). The number of mucous cells was lowest in Lake Ganzirri specimens, as was the quantity of the neutral mucous substances revealed by the Alcian Blue–PAS method. Alcianophilic mucous cells were more numerous in the Lakes Faro and Ganzirri specimens, and this feature agrees well with previous observations that acidification of mucous film enhances the protective function of mucous against pathogens (Sanchez et al. 1997). These differences might be related to the heavy metal differences observed in the two polluted lakes.

All the reported data, intestinal folds impairment, epithelium thickness, number and acidity of mucous cells, parallel the frequency of MTs immunoreactive cells in the gut of Liza aurata from the polluted area of Lake Faro and Lake Ganzirri (Giacobbe et al. 1996; Munaò et al. 2000; Mauceri et al. 2002).

A great variability was observed among the fishes belonging to the same environment; this can be ascribed either to individual variability, or to the different season considered (Ferrando et al. 2003, Abstract in 18th Wilhelm Bernard Workshop on Cell Nucleus: 39). In fact, a seasonal bioaccumulation pattern according to metal concentration in water and sediments was previously demonstrated in fishes (Luoma & Carter 1991; Mzimela et al. 2003). Whether different association of heavy metals, usually found in the different periods of the year, might induce different morphofunctional aspects remains an open question. Thus, considerable further work is required in this field. However, on the basis of our results, we can propose the use of the villus morphology test as an expertise‐independent, cost‐effective and rapid prognostic biomarker for field heavy metal monitoring programs.

Acknowledgments

This research was supported by MIUR (Grant N° 2001058987) Cofin2001.