Introduction

Increasingly emerging data on the effects of chemical fertilizers and pesticides to human health and environmental impact led in recent years to the sharp increase in agricultural practices and animal husbandry of organic crops and livestock. The organic farming allows the use of many natural derivates as fertilizers [Council Regulation (EC) No 834/2007 and Commission Regulation (EC) No 889/2008]. Fertilizers including animal and vegetable derivates can display estrogen-like activity and are able to bind and activate steroid receptors (Degen and Bolt 2000; Combalbert et al. 2012; Alvarez et al. 2013; Valdehita et al. 2014); some phytosanitary products, as well as products for disinfection and cleaning in the livestock and crop production, have been classified as endocrine disruptor chemicals (EDC) (Ying et al. 2002). Fertilizer of animal origin is made of bones, slaughter residues and manure that may contain residues of hormonal substances and/or their metabolites (Bartelt-Hunt et al. 2013; Valdehita et al. 2014). Indeed, vertebrates excrete conjugated or free steroid hormones that may interfere with the endocrine system, affecting the reproduction and development in wildlife (Jobling et al. 1996). It has been demonstrated that metabolites of steroid hormones persist in manure for several months (Schiffer et al. 2001). In addition, conjugated and biologically inactive forms are easily converted into free steroids by soil microrganisms (Ternes et al. 1999; Baronti et al. 2000). Escherichia coli in activated sludge is able to deconjugate human hormones, excreted in the urine as inactive glucuronide or sulphate conjugates, in the active steroids (Desbrow et al. 1998). As consequence, when manures are applied to the soil as fertilizer, these compounds may reach living organisms inducing abnormal endocrine response (Combalbert et al. 2012). Recently, bioassay of estrogenicity on water samples from streams across the United States revealed that higher estrogenic activity was frequently associated with manure application to cultivated fields (Alvarez et al. 2013).

The most common responses are activation of estrogen receptors (ERs) and biosynthesis of vitellogenin (VTG), the estrogen-dependent and sex-specific yolk protein naturally occurring only in females of oviparous vertebrates during the reproductive period (Tata and Smith 1979; Wallace 1985). In the males of oviparous species, VTG gene is silent but can be activated by experimental treatment with estradiol or by exposure to xenoestrogens. As a consequence, the presence of VTG in males is considered the most reliable biomarker of estrogen contamination (Denslow et al. 1999; Irwin and Gray 2001; Matozzo et al. 2008; Verderame et al. 2011; Del Giudice et al. 2012).

Manures of both animal and vegetal origin can contribute to heavy metal loads on agricultural soils; heavy metal contents vary considerably between manure types and origins (Menzi et al. 1993). The role of chemical fertilizers and atmospheric deposition on lead (Pb) and cadmium (Cd) accumulation in topsoils is well known (N’guessan et al. 2009), however, manure of animal origin could contain Cd concentrations exceeding those permitted in agricultural soils (Li et al. 2010; Zhou et al. 2015). Cd is a well recognized endocrine disruptor by affecting the synthesis and/or regulation of several hormones (Henson and Chedrese 2004; Darbre 2006) and exerting an estrogen-like activity in vitro and in vivo (Iavicoli et al. 2009; Hofer et al. 2010). Recent studies have shown that in mammals, Cd may also activate the estrogen receptor α (ERα) and/or mimic estrogen effects in different tissues (Johnson et al. 2003). Cadmium is also able to interfere with the estradiol (E2)-controlled gene expression of ERs and VTG genes. In liver of juvenile females of rainbow trout, Cd reduces the E2-stimulated mRNA levels of VTG and ERα in a dose-dependent manner (Vetillard and Bailhache 2005). On the other hand, in Salmo trutta exposed to high dose of E2, Cd is able to delay the ovulation but does not inhibit ERs upregulation or modify VTG gene expression (Brown et al. 1994).

Active biomonitoring of soil contamination can use bioindicator organisms such as microbial community, mosses, liverworts and, among terrestrial invertebrates, nematodes and gastropods (Bargagli et al. 1995; Kammenga et al. 2000; Weeks et al. 2004; Caruso et al. 2009; Asensio et al. 2013). But often these organisms do not provide information applicable to more complex systems such as vertebrates nor provide data on the fate of secondary consumers, responsible for controlling the population of many invertebrates as well as flora such as weeds. Lizards are soil surface animals that represent an important link between invertebrates and higher predators. The wall lizard Podarcis sicula is the most abundant lizard species present in southern Italy. As prey for many small mammals and birds, it fills an important role also as soil top predator contributing, in several areas, to the control of agricultural pests (Henle and Klaver 1986). Obviously, being part of wild fauna, they can be affected by contamination from anthropic activities. In the lizard P. sicula, VTG synthesis in females is accompanied by the expression of ERα (Verderame and Limatola 2010). In males, both VTG and ERα genes are silent; however, it has been demonstrated that both genes can be activated by estrogen stimulation and/or by EDC with estrogen-like activity (Verderame et al. 2011). In addition, EDCs have been shown to interfere with spermatogenesis (Cardone 2012; Verderame and Limatola 2015), as well as Cd ions interfere with the reproductive performance and embryonic development (Simoniello et al. 2011, 2013, 2014).

Considering the increasing consumption of organic products, the main aim of this study was to determine in the liver of P. sicula male specimens from organic farmland the presence of VTG and ERα, as good biomarkers of EDC contamination. On the same tissue, it has been also evaluated the concentration of Cd and Pb ions, and the induction of the expression of metallothionein (MT), the principal biomarker of heavy metal contamination (Kammenga et al. 2000; Dallinger et al. 2000, 2004; Trinchella et al. 2006; Simoniello et al. 2010).

Materials and methods

Animals and experimental protocol

Forty-five sexually mature male P. sicula (7–8 cm snout-vent length) were caught by noose or hand during their mating period (May–June). Thirty specimens were captured in two organic farms both located in Sorrento Peninsula (Campania, Italy); 15 animals were from a site near Gragnano (BIO-A experimental group) and 15 from a site near Agerola (BIO-B group). The two sites are certified as exclusively organic farms by the Italian Department of Agriculture. Both farms are of wide extent, perched on a hill, far distant and isolated from non-organic crops. These sites use the manure of animals (cows, sheep, horses, pigs and chicken) bred in the same farms as fertilizer. Fifteen animals, referred to as controls (Ctr group), were caught in the outskirts of Naples, in an uncultivated site. The latter represents the wild specimens on which many studies on the reproductive cycles of this species were performed; they never showed signs of estrogenic contamination (Verderame et al. 2012, 2014; Verderame 2014).

Lizards were kept individually in terraria with soil coming from each site and maintained under conditions of natural temperature and photoperiod. Water dishes were always present in the terraria and the animals were fed with uncontaminated larvae of Tenebrio molitor and fresh vegetables ad libitum. After 5 days of relaying to remove the effects of capture stress, the animals were anaesthetized by hypothermia chilling in chipped ice and killed by decapitation. Each group was analyzed separately to determine any differences between animals from the two organic farms and between them and the group of animals caught in the uncultivated field. Livers were immediately excised and processed for histological and molecular analyses. Lizards were captured under authorization No.SCN/2D/2000/9213 of the Italian Ministry of the Environment and the experiments were carried out in compliance with the ethical provisions enforced by the European Union and authorized by the National Committee of the Italian Ministry of Health on in vivo experimentation (Department for Veterinary Public Health, Nutrition and Food Safety). Efforts were made to avoid animal suffering and minimize the number of specimens used. Animals were maintained in accordance with the institutional guidelines for care and use of laboratory animals.

Histology

For each animal, small liver pieces (2–3 mm) were fixed in Bouin’s fluid, dehydrated in a graded series of alcohols and processed for paraffin embedding. A Reichert-Jung 2030 microtome was used to obtain 7 µm thick sections. Some slides were dewaxed, rehydrated in a graded series of alcohols, rinsed with distilled water and processed for immunocytochemistry (ICC) to detect VTG or MT proteins, and others were processed by in situ hybridization (ISH) to detect VTG, ERα or MT transcripts. For both ICC and ISH analyses, we processed ten slides, each one with four attached sections. All the histological results were examined by using a Nikon-MicroPhot-FXA light microscope.

Immunocytochemistry

The immunoreactive VTG localization was carried out as previously reported (Verderame and Limatola 2010). Briefly, liver sections from the animals of the three different groups were incubated with homologous primary anti-VTG antibody (1:1000) in phosphate buffer (PB) 0.1 M pH 7.4 overnight at 4 °C; washed in the same buffer and incubated with secondary polyclonal biotinylated anti-rabbit antibody (Pierce, Rockford, USA) (1:500 in PB). For ERα localization the primary antibody was an anti-hERα (1:80) (Novocastra, United Kingdom), and the secondary a biotinylated anti-mouse (1:500) (Pierce, USA). To detect antigens, we used the ultrasensitive ABC staining reagent kit (Pierce, USA) and diaminobenzidine (DAB 1 mg/mL; Sigma).

The immunoreactive MT was detect with anti-MT (Thermo Fisher) (1:400) and the antigens were identified with Novolink Max Polymer Detection System (Leica Biosystems) according to the manufacturer’s procedure. Negative controls of reactions were performed on other slides by omitting the incubation with the primary antibodies.

In situ hybridization

For in situ hybridization analysis, liver sections mounted onto Superfrost Plus slides (BDH) were dewaxed and treated with proteinase K (10 µg/mL) at 50 °C for 10 min. The Gallus gallus VTG (Verderame et al. 2011), P. sicula ERα (Verderame and Limatola 2010) and P. sicula MT (Riggio et al. 2003) cDNA fragments cloned into the pGEM-T easy plasmid (Promega) were used for generation of cRNA probes. The cRNA synthesis was performed by using the Dig-RNA (SP6/T7) labeling Kit (Roche Applied Science), according to the kit’s protocol. After plasmid linearization, the RNA probes were synthesized using RNA polymerase T7 to obtain the sense probes and RNA polymerase SP6 to obtain the anti-sense probes. Digoxigenin (DIG)-labeled anti-sense riboprobes were used at a concentration of 80 ng/100 μL, in hybridization buffer (Tris-HCl 0.02 M, pH 7.5; NaCl 0.3 M; EDTA 0.01 M; DTT 0.1 M; formamide 50 %; Denhardt’s 1×; t-RNA 100 µg/mL; ss-DNA 100 µg/mL) overnight at 48 and 50 °C, respectively. The slides were rinsed several times in 2× SSC at room temperature, washed two times in washing buffer (NaCl 0.3 M; formammide 50 %; Tris-HCl 0.02 M; EDTA 0.01 M; DTT 0.1 M) and incubated at 37 °C for 30 min in RNase mix (NaCl 0.5 M; Tris-HCl 0.01 M; EDTA 0.01 M; DTT 0.1 M; RNase 10 µg/mL) and for 30 min in the same mix without RNase. The slides were washed in 2× SSC 3 min, in 0.1× SSC 15 min at 60 °C, in 0.1× SSC 3 min at room temperature, in NTP (Tris-HCl 0,1 M pH 7.5; NaCl 0.15 M) and then incubated 1 h in 2 % blocking solution (Roche Diagnostics, Mannheim, Germany) in maleic acid buffer (0.1 M maleic acid, 0.15 M NaCl, pH 7.5). The slides were incubated overnight at 4 °C with an alkaline phosphatase-conjugated sheep anti-DIG antibody (Roche Diagnostics) (1:2 500) in blocking solution. The sections were then rinsed four times in NTP buffer for 30 min and two times in NTM buffer (Tris-HCl 100 mM pH 9.5, MgCl 50 mM, NaCl 100 mM) for 30 min. Finally, the sections were incubated in the dark at room temperature in the color detection substrate solution (BCIP/NBT Roche) in NTM until the appearance of the color. The reaction was stopped by rinsing the slides in Tris-HCl 0.1 M pH 7.5 and mounted in Acquovitrex (Carlo Erba). The negative control sections were incubated with the DIG-labeled mRNA sense probes and processed as previously described. Twin sections were treated with DNase to exclude cross-link with genomic DNA.

Determination of cadmium and lead contents in lizard livers

For the determination of total Cd and Pb content in P. sicula liver, 0.4 g fresh tissue were digested in concentrated ultrapure HNO3 (Fluka), using 1 mL of acid every 100 mg of wet tissue. The mixture was heated for 30 min at 80 °C, cooled and centrifuged for 15 min at 12,000×g. Cd and Pb content in liver supernatants (n = 15) was determined by Zeeman Graphite Furnace Atomic Absorption Spectrometry, using a Varian A220 atomic spectrophotometer, according to the manufacturer’s conditions. Working standards in 0.2 % v/v HNO3 were prepared daily by diluting known aliquots of the stock solution to the appropriate volume. Metal recovery, calculated as a percentage of known amounts of metal added to the samples, ranged from 90 to 93 % for both metals.

Real time PCR analysis of metallothionein expression

Total RNA was extracted according to the TRI-Reagent (Sigma Aldrich) protocol. The concentration and purity of RNA samples were determined by UV absorbance spectrophotometry; RNA integrity was checked on 1.2 % agarose gel electrophoresis. First-strand cDNA was synthesized from each total RNA (1 μg) using the QuantiTect reverse transcription kit (Qiagen), which allows the removal of genomic DNA contamination. Real Time PCR reactions were carried out in an Applied Biosystems 7500 Real Time System by using the Power SYBR® Green Master Mix PCR (Life Technologies). Each SYBR® Green reaction (20 μL total volume) contained 2 μL of 1:1 diluted cDNA as template, 12 μL of real-time PCR Master Mix, 1 μL of each of the forward and reverse primer (10 μM), and 4 μL of nuclease free water. Nuclease-free water for the replacement of cDNA template was used as a negative control. The thermal protocol was as follows: 1 min at 95 °C, followed by 40 cycles (15 s at 95 °C and 40 s at 60 °C). A melting curve analysis of PCR products was performed from 60 to 95 °C in order to ensure gene-specific amplification. For internal standard control, the expression of β-actin gene was quantified and used to normalize gene expression levels in each sample. Changes in the gene expression relative to the different samples were calculated according to the standard \( 2^{{ - \varDelta \varDelta C_{t} }} \) method described by Schmittgen and Livak (2008). MT primers were designed on the N-terminal (forward primer) and C-terminal (reverse primer) regions of the P. sicula MT cDNA (Riggio et al. 2003); the sequences were: forward, 5′-ATGGATCCTCAGGACTGC-3′; reverse, 5′-ATCTGTATGAAAGAGCAATTAC-3′. ß-actin primers, designed on a homologous cDNA fragment (accession n. DQ015917), were: forward, 5′-TATCCACGAGACCACCTTC-3′; reverse, 5′-TCCACCAATCCAGACAGAG-3′.

Data are presented as mean ± standard error of the mean (SEM) from three separate experiments in each sample (n = 4). Statistical analyses were carried out by StatView software (Altura Software, Inc.), as described below.

Statistical analysis

Data analysis was performed by the independent samples t test, using the program contained in the package StatView (Altura Software, Inc.) at a significance level of P < 0.05. The two organic farming sites were analysed as independent cases. Metal concentrations and amount of MT transcripts were compared between each organic farming site and the control site using t tests.

Results

Localization of VTG and ERα mRNAs and proteins in liver

In situ hybridization experiments performed on liver sections from P. sicula males captured in the two organic farms showed the presence of transcripts for VTG (Fig. 1a, c) and ERα (Fig. 2a, c) in hepatocytes cytoplasm of both BIO-A and BIO-B animals. Such cells after immunolocalization experiments also showed a positive signal at cytoplasmatic level for both proteins (Figs. 1d, f; 2d, f). Sections of livers from control males remained completely unlabeled, both after ISH (Figs. 1b, 2b) and ICC (Figs. 1e, 2e) analyses. No signals were detected on sections incubated with sense riboprobes, as well as on sections for ICC obtained by omitting the primary antibodies. No changes in ISH and ICC signals were detected in DNase-treated sections (data not shown).

Fig. 1
figure 1

Localization of VTG mRNA (a, b, c) and protein (d, e, f) in liver sections of P. sicula males. a, d BIO-A group; c, f BIO-B group; b, e Ctr control group. Both positive hybridization and immunolocalization signals, obtained as described under “Materials and methods” section, appear as brown areas (arrows). No signal was evident in the control animals. Bars 30 μm (Color figure online)

Full size image
Fig. 2
figure 2

Localization of ERα mRNA (a, b, c) and protein (d, e, f) in liver sections of P. sicula males. a, d BIO-A group; c, f BIO-B group; b, e Ctr control group. Both positive hybridization and immunolocalization signals, obtained as described under “Materials and methods” section, appear as brown areas (arrows). No signal was evident in the control animals. Bars 30 μm (Color figure online)

Full size image

Cadmium and lead concentration in liver

Cd and Pb contents were determined in liver of every animal collected at the different sampling sites. Results are summarized in Fig. 3 with data given as group mean ± SD, n = 15. Both Cd and Pb levels were significantly higher in Ctr group with respect to BIO-A and BIO-B groups, whereas no differences in metals values were found between BIO-A and BIO-B groups (Table 1). In all instances Pb levels were lower than Cd levels.

Fig. 3
figure 3

Cadmium (Cd, filled square), and lead (Pb, open square), contents in P. sicula male livers. Values are expressed as mean ± SD (n = 15). *P < 0.05 versus Ctr group; **P < 0.001 versus Ctr group (see Table 1 for details on statistical analysis)

Full size image
Table 1 Statistical values obtained from the two samples t test analysis on cadmium and lead contents in lizard livers
Full size table

Metallothionein expression in liver

To assess differences among sites in MT expression, a Real-Time PCR analysis was performed on transcripts from livers of the three different groups of lizards. Results demonstrated that in livers of animals captured in organic farms MT mRNA was less than half of the amount determined in livers of Ctr group animals, while no statistical differences were found between BIO-A and BIO-B groups (Table 2).

Table 2 Statistical values obtained from the two samples t test analysis on metallothionein mRNA content in lizard livers
Full size table

Localization of MT mRNA and protein in liver

Both in situ hybridization (Fig. 4a, b, d) and immunolocalization (Fig. 4e, f, h) experiments showed positive signals in the large Kupffer cells; some hepatocytes were also stained. No differences were observed in localization of MT mRNA and protein in the different groups of animals (Fig. 4). No signals were detected on sections incubated with sense riboprobes (Fig. 4c), as well as on sections for ICC obtained by omitting the primary antibody (Fig. 4g).

Fig. 4
figure 4

Localization of MT mRNA (a, b, d) and protein (e, f, h) in liver sections of P. sicula males. a, e BIO-A group; d, h BIO-B group; b, f Ctr control group. Both positive hybridization and immunolocalization signals, obtained as described under “Materials and methods” section, appear as brown areas (arrows). Note the positive signals also in control animals. No hybridization and immunolocalization signals are evident in the control sections incubated with MT-sense riboprobe (c) and omitting anti-MT antibody (g), respectively. Bars 30 μm (Color figure online)

Full size image

Discussion

Data herein described highlight that P. sicula males inhabiting organic farm areas show the typical hepatic biosynthetic alterations indicative of an estrogenic contamination. Hepatocytes, in fact, contain detectable amount of both VTG and ERα transcripts and proteins. On the other hand, the livers of the same animals show a very low level of the heavy metals Cd and Pb. In terrestrial vertebrates, Cd and Pb accumulation in liver is typical of organisms living in soils contaminated by anthropogenic activities such as smelting of ores, burning of fossil fuels, waste incineration; in addition, Cd reaches the soil through pesticides and herbicides (Frost and Ketchum 2000), Pb through the urban traffic (Pena-Fernandez et al. 2015). The amount of MT transcripts in these animals is significantly lower than that determined in lizards inhabiting rural sites in the outskirts of Naples. The soils of the two organic farms show a very low content of heavy metals (for BIO-A, Cd was below the detection limit and Pb was 53.6 mg/kg soil; for BIO-B, Cd = 0.41 mg/kg and Pb = 20.3 mg/kg) and phenols (BIO-A: <0.010 mg/kg; BIO-B: <0.010 mg/kg), a condition typical of soils not contaminated with chemical fertilizers. Hence, the presence of substances able to induce the expression of the known biomarkers of estrogenic pollution i.e. ERα and VTG in the liver of lizard males, is most likely due to the massive use of manure as fertilizer. It has been suggested that VTG expression in oviparous males may be interpreted as a warning of reproductive adverse consequences (Cheek et al. 2001), and that over-activation of estrogen signaling through its receptors can have detrimental effects on the fertility parameters of male rats (Dumasia et al. 2015). Xenoestrogenic substances are responsible of feminization in fish (Hamilton et al. 2014) and are able to impair spermatogenesis in terrestrial vertebrates (Verderame and Limatola 2015). In P. sicula, an induction of apoptosis in male germ cells is also observed (Verderame and Limatola 2015) hence a prolonged use of manure could lead to changes in the spermatogenic cycle of these animals, endangering their reproduction and survival. In addition, an increased amount of estrogen plasma levels could have adverse effects also in female specimens, in particular on oogenesis and oviposition. It has been demonstrated that exogenous estradiol increases shell thickness in laying hens (Wistedt et al. 2014). Moreover, yolk steroids of maternal origin may also influence sex determination and gonad differentiation (Ding et al. 2012; Navara 2013). On the whole, it can be hypothesized that exogenous estrogens could affect P. sicula embryonic development, offspring and reproductive cycle. In addition, it is not ruled out an increase of estrogens in animals bred in these organic farms (chicken, pig, sheep, cow) that could lead to reproductive failures for the same animals (Padmanabhan et al. 2010; Rosenfeld 2012) and to an undesirable higher content of estrogens in food chain, with possible risk for human consumers (Andersson and Skakkebaek 1999; Ibarreta et al. 2001; Aravindakshan et al. 2004).

In light of these data, it can be stated that organic farming avoids soil contamination by chemicals and heavy metals but, on the other hand, the use of manure as fertilizer may lead to an accumulation in soils of natural steroids and/or EDC that may affect reproductive processes. Organic farms should focus more on crop rotation and encouraging biological cycles and soil biological activity rather than using manure as fertilizer; it is also important to monitor the amount of estrogen derivatives in manure to avoid possible unwanted effects to wildlife and human health. In the recent years, it has been estimated that organic farming has grown by 8.9 % per year (Paull 2011); so, if nothing is done to mitigate the problem, soil contamination by steroids and EDC could soon represent a serious risk as currently is soil contamination by chemicals and heavy metals.