Differential effects of experimental and cold-induced hyperthyroidism on factors inducing rat liver oxidative damage

It is well documented that a hypermetabolic state induced by thyroid hormone administration to experimental animals is accompanied by oxidative stress in several target tissues (Videla,2000). Studies on liver, heart and muscle suggest that such oxidative stress is due to increased mitochondrial release of reactive oxygen species (ROS) (Fernández and Videla,1993; Venditti et al.,2003a; Venditti et al.,2003b; Venditti et al.,2003c) and reduced global efficacy of the antioxidant defence system (Venditti et al., 1997; Videla, 2000).

Hyperthyroid state can also be induced in homeothermic animals by physiological modifications of thyroid activity. In rat, exposure to a cold environment is associated with increased serum 3,5,3′-triiodothyronine(T₃) levels (Nejad, 1972), which is thought to be the main factor responsible for the heat produced by non-shivering thermogenesis(Jansky, 1963). Although in rodents the changes in metabolic activity involved in such process occur mostly in brown adipose tissue (BAT)(Himms-Hagen, 1983), they are also found in other tissues, including liver(Goglia et al., 1983), and cardiac (Venditti et al.,2003b) and skeletal (Guernsey and Stevens, 1977) muscles.

The effects of cold-induced hyperthyroidism on tissue oxidative damage have been scarcely investigated. However, indications have been obtained for an increase in lipid peroxidation products in several rat tissues, including BAT(Barja de Quiroga et al.,1991), lung (Tnimov et al.,1984), muscle (Venditti et al., 2004a), liver and heart(Venditti et al., 2004a; Kolosova et al., 1995). Furthermore, it has been recently reported that the liver from 10-day cold-exposed rats exhibits increases in ROS released by mitochondria(Venditti et al., 2004b) and susceptibility to oxidative challenge(Venditti et al., 2004a).

Because such modifications are similar to those induced in the liver by T₃ administration, we proposed that this hormone brings about the biochemical changes underlying tissue oxidative damage found in the two conditions (Venditti et al.,2004a; Venditti et al.,2004b). However, unlike cold exposure, T₃administration strongly decreases serum levels of thyroxine (T₄),which has been reported to have intrinsic biological activity in the cold(Cageao et al., 1992). Thus, it is possible that T₄ contributes to some changes underlying tissue thermogenesis and oxidative stress associated with cold exposure. If so,differences should have to be found in the tissue responses to treatments which differentially affects circulating T₄ levels.

Therefore, we compared the effects of 10 days of thyroid hormone(T₃ or T₄) treatment and cold exposure on oxidative metabolism and the extent of oxidative damage in rat liver. To obtain information on oxidative damage to lipids and proteins we measured levels of hydroperoxides and protein-bound carbonyls, respectively. Furthermore, we measured levels of markers of non-enzymatic protein modifications that form under oxidative conditions.

Mitochondrial and microsomal ROS production and parameters determining the susceptibility to oxidative challenge were also assessed. For this purpose,glutathione peroxidase (GPX) and glutathione reductase (GR) activities,vitamin E (Vit E), coenzyme Q (CoQ), and reduced glutathione (GSH) content,lipid composition, and response to oxidants in vitro of the tissues were determined.

All chemicals used (Sigma Chimica, Milano, Italy) were of the highest grades available. Response to oxidative stress was determined by using reagents and instrumentation of the commercially available Amerlite System(Ortho-Clinical Diagnostics, Milano, Italy). Serum levels of free triiodothyronine (FT₃) and thyroxine (FT₄) were determined by using commercial RIA kits (DiaSorin, Salluggia, Italy).

The experiments were carried out on 60-day-old male Wistar rats (Rattus norvegicus albius Berkenhault 1769), supplied by Nossan (Correzzana,Italy) at day 45 of age. From day 50, animals were randomly assigned to one of four groups: euthyroid control rats (C), and rats made hyperthyroid by T₃ (HT₃) or T₄ (HT₄) treatment (10 days of daily intraperitoneal injections of 10 μg 100 g^–1body mass of T₃ or T₄, respectively) or by cold exposure(10 days at 4±1°C) (CE). C, HT₃ and HT₄ rats were kept at room temperature of 24±1°C. All rats were subjected to the same conditions (one per cage, constant artificial circadian cycle of 12 h:12 h L:D, and 50±10% relative humidity), and fed the same diet of a commercial rat chow purchased from Nossan, and water on an ad libitumbasis.

The treatment of animals in these experiments was in accordance with the guidelines set forth by the University's Animal Care Review Committee.

The animals were sacrificed by decapitation while under ether anaesthesia. Arterial blood samples were collected and later analysed to determine plasma levels of FT₃ and FT₄ by radioimmunoassay. Liver was rapidly excised and placed into ice-cold homogenisation medium (HM) (220 mmol l^–1 mannitol, 70 mmol l^–1 sucrose, 1 mmol l^–1 EDTA, 0.1% fatty acid-free albumin, 10 mmol l^–1 Tris, pH 7.4). Then, the tissue was weighed, finely minced and washed with HM. Finally, liver fragments were gently homogenised(20% w:v) in HM using a glass Potter-Elvehjem homogeniser set at a standard velocity (500 r.p.m.) for 1 min. Aliquots of liver homogenates were used for analytical procedures and preparation of mitochondrial and microsomal fractions.

The homogenates, diluted 1:1 with HM, were freed of debris and nuclei by centrifugation at 500 g for 10 min at 4°C. The resulting supernatants were centrifuged at 10 000 g for 10 min. The mitochondrial pellets were washed twice with isolation medium (IM; 220 mmol l^–1 mannitol, 70 mmol l^–1 sucrose, 1 mmol l^–1 EGTA, 20 mmol l^–1 Tris, pH 7.4),resuspended in the same solution and used for determination of cytochrome oxidase (COX) activity and H₂O₂ release. The 10 000 g supernatants were centrifuged at 105 000 gfor 60 min and the resulting microsomal pellets were suspended in IM and used for determination of glucose-6-phosphatase activity and H₂O₂ production.

The protein content of liver preparations was determined, after solubilization in 0.5% deoxycholate, by the biuret method(Gornall et al., 1949) with bovine serum albumin as standard.

Cytochrome oxidase activity of homogenates and mitochondrial suspensions was determined by the procedure of Barré et al.(Barré et al., 1987). Glucose-6-phosphatase activity was determined in liver homogenates and microsomal preparations as described by Katewa and Katyare(Katewa and Katyare,2003).

Liver oxygen consumption was monitored at 30°C by a Gilson respirometer in 1.6 ml of incubation medium (145 mmol l^–1 KCl, 30 mmol l^–1 Hepes, 5 mmol l^–1KH₂PO₄, 3 mmol l^–1 MgCl₂,0.1 mmol l^–1 EGTA, pH 7.4) with 50 μl of homogenate and succinate (10 mmol l^–1), plus 5 μmol l^–1rotenone (Rot), or pyruvate/malate (10/2.5 mmol l^–1) as substrates, in the absence (state 4) and in the presence (state 3) of 500μmol l^–1 ADP.

The extent of the peroxidative processes in tissue homogenates was determined by measuring the level of lipid hydroperoxides according to the method of Heath and Tappel (Heath and Tappel, 1976). Tissue protein oxidation was assayed by the reaction of 2,4-dinitrophenylhydrazine with protein carbonyls as described by Reznick and Packer (Reznick and Packer,1994).

Concentrations of markers of non-enzymatic protein modifications, such as glutamic (GSA) and aminoapidic semialdehyde (AASA) (resulting from direct protein oxidation), N^ϵ-(carboxymethyl)lysine (CML) (resulting from both lipid peroxidation and glycoxidation),N^ϵ-(carboxyethyl)lysine (CEL) (resulting from glycoxidative damage), and N^ϵ-(malondialdehyde)lysine (resulting from malondialdehyde attachment to protein lysine residue) were detected and measured by gas chromatography/mass spectrometry (GC/MS) as previously described (Pamplona et al.,2005).

The rate of mitochondrial and microsomal H₂O₂ release was measured at 30°C following the increase in fluorescence (excitation at 320 nm, emission at 400 nm) resulting from oxidation of p-hydroxyphenylacetate (PHPA) by H₂O₂ in the presence of horseradish peroxidase (HRP)(Hyslop and Sklar, 1984) in a computer-controlled Jasko fluorometer equipped with a thermostatically controlled cell holder. For measurement of H₂O₂ produced by the respiratory chain, the reaction mixture consisted of 0.1 mg ml^–1 mitochondrial proteins, 6 U ml^–1 HRP,200 μg ml^–1 PHPA and 10 mmol l^–1succinate, plus 5 μmol l^–1 rotenone, or 10 mmol l^–1 pyruvate/2.5 mmol l^–1 malate added last to start the reaction in a medium containing 145 mmol l^–1KCl, 30 mmol l^–1 Hepes, 5 mmol l^–1KH₂PO₄, 3 mmol l^–1 MgCl₂,0.1 mmol l^–1 EGTA, pH 7.4. Measurements with the different substrates in the presence of 500 μmol l^–1 ADP were also performed. For measurement of H₂O₂ produced by monoamine oxidase (MAO) the reaction mixture consisted of 0.1 mg ml^–1mitochondrial proteins, 6 U ml^–1 HRP, 200 μg ml^–1 PHPA and 0.2 mmol l^–1 tyramine added last to start the reaction in the same medium used for respiration-linked H₂O₂ release. For microsomal preparations the reaction mixture consisted of 0.422 mg ml^–1 microsomal proteins, 6 U ml^–1 HRP, 200 μg ml^–1 PHPA in 0.1 mol l^–l phosphate buffer, pH 7.4. The H₂O₂produced was determined by the fluorescence change 10 min after addition of 0.35 mol l^–1 NADPH.

Known concentrations of H₂O₂ were used to establish the standard concentration curve. Preliminary experiments studied the effect of catalase addition on the measured rates of H₂O₂production. They showed a dose-dependent drop of the fluorescence in the presence of the enzyme.

Liver GPX activity was assayed at 37°C according to the method of Flohé and Günzler (Flohéand Günzler, 1984), with H₂O₂ as substrate. GR activity was measured at 30°C according to Carlberg and Mannervik (Carlberg and Mannervik,1985).

Ubiquinols (CoQH₂) from 0.5 ml of 10% homogenate were oxidized to ubiquinones (CoQs) with 0.5 ml of 2% FeCl₃ and 2.0 ml of ethanol. The total content of CoQs (CoQH₂ + CoQ) was then determined as described by Lang et al.(1986). Vit E content was determined using the HPLC procedure of Lang et al.(Lang et al., 1986). GSH concentration was measured as described by Griffith(Griffith, 1980).

Fatty acyl groups were analysed by GC/MS as previously described(Pamplona et al., 1998). Fatty acyl composition of lipids was expressed as mol%.

The following fatty acyl indices were also calculated: saturated fatty acids (SFA); unsaturated fatty acids (UFA); monounsaturated fatty acids(MUFA); polyunsaturated fatty acids from n-3 and n-6 series (PUFAn-3 and PUFAn-6); average chain length (ACL)=[(∑%Total₁₄×14)+(∑%Total₁₆×16)+(∑%Total₁₈×18)+(∑%Total₂₀×20)+(∑%T otal₂₂×22)]/100; double bond index (DBI)=[(1×∑mol%monoenoic)+(2×∑mol% dienoic)+(3×∑mol% trienoic)+(4×∑mol% tetraenoic)+(5×∑mol%pentaenoic)+(6×∑mol% hexaenoic)], and peroxidizability index(PI)=[(0.025×∑mol% monoenoic)+(1×∑mol%dienoic)+(2×∑mol% trienoic)+ (4×∑mol%tetraenoic)+(6×∑mol% pentaenoic)+(8×∑mol% hexaenoic)].

Response to oxidative challenge was determined as previously described(Venditti et al., 1999). Briefly, samples of 10% (w:v) homogenates were obtained by diluting the 20%homogenates with equal volumes of 0.2% Lubrol in 15 mmol l^–1Tris, pH 8.5. Several dilutions of samples up to a tissue concentration of 0.002% were prepared in 15 mmol l^–1 Tris (pH 8.5). The assays were performed in microtiter plates. Enhanced chemiluminescence reactions were initiated by addition of 250 μl of the reaction mixture to 25 μl of the samples. The plates were incubated at 37°C for 30 s with continuous shaking and then transferred to a luminescence analyser. The emission values were fitted to dose-response curves using the statistical facilities of the Fig.P graphic program (Biosoft, Cambridge, UK).

The data, expressed as means ± standard error, were analyzed with a one-way analysis of variance method (ANOVA). When a significant Fratio was found, the Student-Newman–Keuls multiple range test was used to determine the statistical significance between means. Probability values(P)<0.05 were considered significant. In Fig. 2 the results of the experiments are presented as sample curves.

Changes in thyroid state were documented by modifications in heart mass/body mass ratio, and plasma levels of FT₃ and FT₄. The body masses of C, CE, HT₃ and HT₄ rats, which were 267±4, 253±6, 251±6 and 252±8 g, respectively,were not significantly affected by any of the treatments. Conversely, the heart masses increased in all hyperthyroid rats so that heart/body mass ratios of these animals (3.25±0.05, 3.40±0.11 and 3.17±0.07 mg g^–1 for CE, HT₃, and HT₄ rats,respectively) were higher than control values, without any significant difference among them. Plasma levels of FT₃ increased in all hyperthyroid rats, but they were lower in HT₄ than in CE and HT₃ rats, whereas the FT₄ levels increased in CE and HT₄ rats and decreased in HT₃ rats(Fig. 1).

Cold exposure and hormonal treatments were associated with increases in COX activities in both homogenates and mitochondria, which were lower in HT₄ rats (Table 1). Homogenate COX activities were not significantly different after cold exposure or T₃ treatment, whereas mitochondrial COX activities reached the highest value after T₃ treatment. The in vitro COX activity has been positively correlated to the maximal oxygen consumption(Simon and Robin, 1971) so that its changes provided information on effects of the treatments on aerobic metabolic capacity of the biological preparations.

Glucose-6-phosphatase activities of homogenates and microsomes increased in all hyperthyroid rats, but the highest activity in the homogenates was found in that from HT₃ rats (Table 1).

The ratio between the cytochrome oxidase activities of homogenates and mitochondria and that between glucose-6-phosphatase activities of homogenates and microsomes provided rough estimates of tissue content of mitochondrial and microsomal proteins, respectively.

Mitochondrial protein content was higher after cold exposure and T₄ treatment, but not after T₃ treatment, whereas microsomal protein content was not affected by hormonal treatments and cold exposure (Table 1).

The rates of both succinate- and pyruvate/malate-supported oxygen consumption are reported in Fig. 2. Those supported by succinate were increased by all treatments and were higher in HT₃ than in CE group in state 4, and higher in HT₃ than in other two groups in state 4. Those supported by pyruvate/malate were increased in all hyperthyroid groups and, in state 3,reached the highest value in HT₃ group.

The levels of hydroperoxides and protein-bound carbonyls were higher in hyperthyroid than control rats (Fig. 3). The highest and lowest hydroperoxide levels were found in HT₃ and CE preparations, respectively, whereas the highest carbonyl levels were found in CE preparations.

Of the non-enzymatic protein modification markers, GSA levels were increased by cold exposure and T₃ treatment, whereas that increase did not reach significant levels after T₄ treatment. The steady-state levels of AASA increased significantly by all treatments,reaching the highest and lowest magnitudes in HT₃ and HT₄ groups, respectively. CEL and CML levels were increased only by T₃ treatment. MDAL levels increased in all treatment groups and reached the lowest value in HT₄ group(Table 2).

The effect of cold exposure and T₃/T₄ treatment on H₂O₂ release by succinate (complex II-linked substrate)-and pyruvate/malate (complex I-linked substrates)-supplemented mitochondria are showed in Fig. 4. With complex II-linked substrate the rates of H₂O₂ release during state 4 respiration increased in all treatment groups and reached the greatest value in HT₃ group. During state 3 the rates increased only after hormonal treatments and reached the highest value in the HT₃ group. With complex I-linked substrates, during state 4, the rates significantly increased in all hyperthyroid groups and were greater in the HT₃ than in the CE group, whereas during state 3 they increased only in hormone-treated groups, reaching the highest value in the HT₄ group. The rates of H₂O₂ production by monoamine oxidase in the presence of tyramine were not significantly modified by cold exposure and hormonal treatment. Their values were 3.39±0.08,3.35±0.17, 3.24±0.04, 3.49±0.12 nmol min^–1 mg^–1 protein for C, CE,HT₃, and HT₄ preparations, respectively. The rate of NADPH-dependent H₂O₂ production by liver microsomes,which was 119.1±3.7 pmol min^–1 mg^–1protein in the C group, increased in all treatment groups. However, the value for the HT₃ group (219.4±8.7 pmol min^–1mg^–1 protein) was significantly lower than those for the CE and HT₄ groups (259.7±6.4 and 247.5±10.6 pmol min^–1 mg^–1 protein, respectively).

The content of both enzymatic and low molecular mass antioxidants in the liver are showed in Table 3. GPX activity was increased by T₃ and decreased by T₄treatment, but was not affected by cold exposure. GR activity was increased by cold exposure and to a greater extent by T₃ treatment, whereas it was not modified by T₄ treatment. Vit E content was increased by all treatments, reaching the highest value in the HT₃ group. CoQ9 levels were slightly increased by cold exposure, more strongly increased by T₃ treatment, and not modified by T₄ treatment, whereas CoQ10 levels were increased only by T₃ treatment. GSH levels were significantly reduced by all treatments.

Effects of cold exposure and hormonal treatments on fatty acid profiles and indexes are shown in Table 4. Globally, the amount (%) of saturated fatty acids and unsaturated fatty acids is maintained constant among the different experimental groups. However, all experimental groups showed a change in the distribution of unsaturated fatty acids. Thus, cold exposure, T₃ and T₄ diminished the content of monounsaturated fatty acids and increased the content of polyunsaturated fatty acids (specially PUFA from the n-6 series). Those changes led to a significant increase in the double bond index and peroxidizability index in all hyperthyroid groups. The increase in PUFA n-6 was mainly due to the increase in the arachidonic acid content in all groups,along with 20:5 and 22:5 fatty acids specifically for the HT₄group. The increase in the PUFA content also led to a slight, but significant,increase in the average chain length.

The luminescence response to changes of concentration of the homogenates(Fig. 5) has previously been described by the equation: E=a C/exp(b C)(Di Meo et al., 1996; Venditti et al., 1999). The a value depends on the concentration of substances, such as cytochromes, able to react with H₂O₂ to produce ^•OH radicals that induce the luminescent reaction. Conversely,the b value depends on the concentration of substances, particularly water-soluble ones, able to prevent the formation or interacting with ^•OH radicals, thus reducing the levels of light emission. Such levels, and particularly the emission maximum(E_max=a/e b), can be considered an index of the susceptibility of the preparations to oxidative challenge(Venditti et al., 1999). Thus,the curves in Fig. 5 indicate that tissue susceptibility to oxidants increases in hyperthyroid preparations reaching the highest and lowest values in T₃- and T₄-treated rats, respectively. These qualitative evaluations are confirmed by the E_max values(Table 5), the increases being the result of higher a values and lower b values. The differences in the a and b parameters between preparations from control- and hyperthyroid rats are consistent with the differences in COX activities and GSH levels, respectively.

We investigated the role of iodothyronine in rat liver response to cold,comparing it with those elicited by T₃ or T₄ treatments,which made serum FT₄ levels lower and higher, respectively, than those found in cold-exposed rats.

The changes in COX activities reveal enhanced liver oxidative capacity in CE rats, which can be attributed to high serum levels of T₃,because a similar increase was obtained by T₃ treatment. However,the changes found in both mitochondrial oxidative capacity and mitochondrial proteins in the liver suggest that the mechanisms underlying the tissue oxidative capacity increase is different in CE and HT₃ rats. In the CE group, this increase seems to be due to proliferation of mitochondria,which show little increase in their oxidative capacity, in agreement with the observation that the light mitochondrial fraction, characterized by low oxidative capacity, increases in chilled liver(Venditti et al., 2004c). Conversely, in the HT₃ group it is caused only by increase in mitochondrial oxidative capacity. An intriguing possibility is that T₃ induces COX activation (and probably its synthesis) as well as mitochondrial proliferation, which, however, also requires T₄levels higher than those present in HT₃ rats. This is supported by the finding that in HT₄ rats, in the presence of high FT₄ levels, moderate increases in FT₃ levels are associated with small increases in liver COX activity and mitochondrial proteins.

Unlike COX activity, liver state 3 respiration was differently affected by cold exposure and T₃-treatment. The moderate enhancement in tissue O₂ consumption produced by cold treatment could be explained by the small increase in mitochondrial COX activity causing a small increase in O₂ consumption by the respiratory chain, which is not compensated for by the increase in mitochondrial proteins. This is supported by the observation that tissue O₂ consumption, like mitochondrial COX activity, is not significantly different in HT₄ and CE rats.

The results concerning indicators of oxidative damage to lipid (HPs and MDAL) and proteins (carbonyls groups, GSA, and AASA) confirm the strict association between hypermetabolic state and oxidative stress occurring in hyperthyroid animals (Videla,2000). However, they do not clarify what treatment causes the greatest oxidative damage in liver since the highest levels of hydroperoxides and protein-bound carbonyls were found in the HT₃ and CE groups,respectively, and their surrogate GC/MS markers in protein do not show similar behaviour in response to these treatments. In fact, carbonyl compound levels depend on the content of amino acids able to generate carbonyl groups in tissue proteins. Moreover, amino acids, such as lysine, can be subjected either to oxidative damage by ROS or to attachment by reactive carbonyl compounds formed by carbohydrate and fatty acid oxidation. The carbonyl-amine reactions can interfere with oxidative reactions in a measure dependent on extent of tissue peroxidative and glycooxidative processes. Therefore, the lower levels of protein-bound carbonyls in HT₃ than in CE rats should be consistent with the higher levels of hydroperoxides, CML and CEL found in T₃-treated animals. Nonetheless, concentration of protein adducts depends on both formation and breakdown, and in a severe hyperthyroid state protein degradation is so fast that there is a decrease in steady state CML and MDAL levels in rat liver (Guerrero et al., 1999). The lack of information on relative changes in protein degradation in the present experiments meant that it was not possible to quantify the interference between direct and indirect protein modifications. Thus, it is not clear what role, if any, T₄ plays in determining the differences in oxidative damage induced by cold or T₃ treatment. Moreover, the oxidative effects of the different treatments are mainly due to biochemical changes affecting free radical production and the antioxidant defence system.

In agreement with a previous report(Venditti et al., 2004b) we found that in CE rat mitochondria the rate of H₂O₂generation increases only during state 4 respiration. Moreover, probably due to lower levels of autoxidizable electron carriers, this rate was significantly lower than in HT₃ rats. Despite this, mitochondrial ROS release strongly contributes to oxidative stress in cold liver because of the increase in mitochondrial proteins. The differences in free radical activity between HT₃ and CE rats do not seem to be due to T₄, in the light of the increased rate of mitochondrial H₂O₂ release during state 3 respiration see in T₄-treated rats.

Within the cell, in addition to the mitochondrial respiratory chain, there are other relevant sources of ROS, such as monoamine oxidase and microsomal monooxygenases. Oxidative deamination of biogenic amines catalysed by MAO, is a large source of H₂O₂(Cadenas and Davies, 2000), the production of which, according to our results, should increase in CE and HT₄ rat livers because of their increased mitochondrial protein content.

The overall microsomal H₂O₂ production in liver from control rats (6.30 nmol min^–1 g^–1 liver) was 87.5% and 36.8% of the mitochondrial productions during state 4 respiration sustained by succinate and pyruvate/malate (7.2 and 17.1 nmol min^–1 g^–1 liver, respectively). Conversely,in all treatment groups, overall microsomal H₂O₂production during pyruvate/malate-sustained state 4 respiration was about 55%of mitochondrial H₂O₂ production, suggesting an increased contribution by endoplasmic reticulum to tissue oxidative damage in the hyperthyroid state.

Treatments produced unbalanced and sometimes opposite changes in antioxidant enzyme activities and scavenger concentrations. Although the extent of some changes is directly related to serum T₃ levels and inversely to T₄ levels, this probably reflects a more important role for T₃ than for T₄ in oxidative protection. However, the GPX activity was lower in the HT₄ than in the control group, suggesting that it might be negatively regulated by T₄.

There was no clear relationship between lipid pattern and iodothyronine serum levels. Moreover, the degree of lipid unsaturation was not differentially increased by treatments, but the greatest susceptibility to peroxidative reactions was displayed by T₃- and T₄-treated rats.

Conversely, T₄ treatment induced the smallest decrease in liver capacity to oppose oxidative damage. The tissue susceptibilities to oxidants were in part related to the values of the parameter a, which depends on the tissue concentrations of substances, such as the hemoproteins, which are able to produce ^•OH radicals(Halliwell and Gutteridge,1990). Thus, the low a value found in HT₄preparations was consistent with our observation that T₄ treatment slightly increased COX activity.

Overall, our results lead to the conclusion that the hyperthyroid state, by whatever treatment it is elicited, gives rise to increased liver oxidative capacity and oxidative damage, attributable to an action of T₃,which is the only iodothyronine for which circulating levels increase in all treatment groups. This idea is further supported by the observation that hepatic tissue, in which type I iodothyronine deiodinase is preferentially expressed, maintains thyroid hormone concentrations similar to those in plasma(Escobar-Morreale et al.,1997). However, it is apparent that there are differences in the size of the effects and underlying mechanisms, found in cold-exposed and T₃-treated rats. It is not clear whether this depends on differences in T₄ serum levels. In fact, such a conclusion could be drawn, even in animals treated with deiodinase inhibitors, only if T₄ gave rise to changes opposite or greater than those produced by T₃. Unfortunately, this is difficult to find when measuring parameters such as oxidative damage extent, which depend on numerous factors,for which the relative contribution is not well defined. Thus, the only results suggesting differential effects of T₃ and T₄concern parameters, such as GPX activity and mitochondrial protein content,more directly dependent on gene activity. Although the T₄-induced changes can supply an explanation of the different changes in the above parameters found in cold-exposed and T₃-treated rats, there is not enough evidence to indicate a role of T₄ in tissue response to cold exposure. However, it is a significant starting point for further experimental work, possibly performed on rats treated with T₄, which can be prevented from converting to T₃ by deiodinase inhibitors.

This work was supported by grants from the Spanish Ministry of Science and Technology (BFI2003-01287) and the Generalitat of Catalunya (2001SGR00311) and Italian Ministry of University and Scientific and Technological Research.