Rotondo Dottore et al. Page 1

Transcription

1 ANTIOXIDANT ACTIONS OF SELENIUM IN ORBITAL FIBROBLASTS: A BASIS FOR THE EFFECTS OF SELENIUM IN GRAVES ORBITOPATHY Giovanna Rotondo Dottore 1 *, Marenza Leo 1 *, Giamberto Casini 2, Francesco Latrofa 1, Luca Cestari 2, Stefano SellariFranceschini 3, Marco Nardi 2, Paolo Vitti 1, Claudio Marcocci 1 and Michele Marinò 1 *Equal contributors 1 Department of Clinical and Experimental Medicine, Endocrinology Units, University of Pisa and University Hospital of Pisa, Via Paradisa 2, 56124, Pisa, Italy 2 Department of Surgical, Medical and Molecular Pathology, Ophthalmopathy Unit I, University of Pisa and University Hospital of Pisa, Via Paradisa 2, 56124, Pisa, Italy 3 Department of Surgical, Medical and Molecular Pathology, ENT Unit I, University of Pisa and University Hospital of Pisa, Italy, Via Paradisa 2, 56124, Pisa, Italy Corresponding Author: Michele Marinò. Department of Clinical and Experimental Medicine, Endocrinology Unit I, University of Pisa and University Hospital of Pisa, Via Paradisa 2, 56124, Pisa, Italy. Phone: ; Fax: ; michele.marino@med.unipi.it Contact Information Giovanna Rotondo Dottore: Endocrinology, Via Paradisa 2, 56124, Pisa; ge.rd@hotmail.it Marenza Leo: Endocrinology, Via Paradisa 2, 56124, Pisa; marenzaleo@libero.it Giamberto Casini: Ophthalmology, Via Paradisa 2, 56124, Pisa; giamberto.c@alice.it Francesco Latrofa: Endocrinology, Via Paradisa 2, 56124, Pisa; latrofaf@libero.it Luca Cestari: Ophthalmology, Via Paradisa 2, 56124, Pisa; cestariluca1981@hotmail.it Stefano SellariFranceschini: ENT Unit, Via Paradisa 2, 56124, Pisa; stefano.sellari@unipi.it Marco Nardi: Ophthalmology, Via Paradisa 2, 56124, Pisa; marco.nardi@med.unipi.it Paolo Vitti: Endocrinology Unit, Via Paradisa 2, 56124, Pisa; paolo.vitti@med.unipi.it Claudio Marcocci: Endocrinology, Via Paradisa 2, 56124, Pisa; claudio.marcocci@med.unipi.it Running Title: selenium in orbital fibroblasts Key Words: selenium, orbitopathy, ophthalmopathy, Graves, fibroblast Rotondo Dottore et al. Page 1

2 Abstract Background A recent clinical trial has shown a beneficial effect of the antioxidant agent selenium in Graves orbitopathy (GO). In order to shed light on the cellular mechanisms on which selenium may act, here we investigated its effects in cultured orbital fibroblasts. Methods Primary cultures of orbital fibroblasts from 6 GO patients and 6 control subjects were established. Cells were treated with H 2 O 2 to induce oxidative stress, after preincubation with selenium (Methyl)selenocysteine (SeMCys). The following assays were performed: glutathione disulfide (GSSG), as a measure of oxidative stress, glutathione peroxidase (GPX) activity, cell proliferation, hyaluronic acid (HA) and proinflammatory cytokines. Results H 2 O 2 induced an increase in cell GSSG and fibroblast proliferation, which were reduced by SeMCys. Incubation of H 2 O 2 treated cells with SeMCys was followed by an increase in glutathione peroxidase activity, one of the antioxidant enzymes into which selenium is incorporated. At the concentrations used (5 µm), H 2 O 2 did not affect significantly HA release, which was however reduced by SeMCys. H 2 O 2 determined an increase in endogenous cytokines involved in the response to oxidative stress and GO pathogenesis, namely TNFα, IL1β and IFNγ. The increases in TNFα and IFNγ were rescued by SeMCys. Whereas the effects of SeMCys were similar in GO and control fibroblasts concerning oxidative stress and cytokines, they were exclusive to GO fibroblasts concerning proliferation and HA. Conclusions Selenium, in the form of SeMCys, rescues from some of the effects of oxidative stress in orbital fibroblasts, namely increased proliferation and proinflammatory cytokines. SeMCys reduces HA release in GO fibroblasts in a manner that seems at least in part independent from H 2 O 2 induced oxidative stress. Some effects of SeMCys are specific for GO fibroblasts. Our findings reveal some cellular mechanisms by which selenium may act in patients with GO. Rotondo Dottore et al. Page 2

3 Introduction Graves orbitopathy (GO) is an autoimmune syndrome observed in ~2530% of patients with Graves disease (GD) and more rarely in patients with hypothyroid autoimmune thyroiditis or in the absence of overt thyroid diseases (14). Moderatetosevere and active forms of GO can be effectively treated with glucocorticoids, orbital irradiation, or both, with positive responses ranging from 60 to 80% of patients depending on the studies (26). However, patients with milder forms of GO, which to some extent can improve spontaneously (7), are generally not given any treatments, except for local measures (i.e., artificial tears, ointments, and prisms). Nevertheless, similar to those with moderatetosevere GO, also patients with mild GO have a substantial impairment in their quality of life, as shown either by general health related qualityoflife questionnaires or by a GO specific qualityoflife questionnaire (GOQOL) (8, 9). In addition, although a discrete proportion of patients with mild GO experience a spontaneous improvement of eye symptoms, a small proportion of them (~15%) progress to a moderatetosevere disease (7, 10). Because of these reasons, some kind of treatment should be offered also to patients with mild GO, in whom the benefits of high dose glucocorticoids are not sufficient to justify the risks that the treatment carries (1114). An ideal treatment for mild GO should be effective, affordable, well tolerated, and widely available. A recent study has provided evidence that selenium fulfils these requirements (15). Selenium is a trace mineral and an essential nutrient for selenocysteine synthesis (16). The latter is incorporated into several selenoproteins, most of which are enzymes. Within selenoproteins, selenium acts as a reduction oxidation center and exerts antioxidant actions. In this regard, several in vitro studies have shown that increased generation of oxygen free radicals plays a role in the pathogenesis of GO (1719). To investigate whether selenium has a beneficial effect in GO, Marcocci et al. carried out a multicenter, randomized, doubleblind, placebocontrolled trial selenium vs placebo, and found that treatment with selenium is associated with an improvement of eye signs and symptoms and of the quality of life of patients with mild GO (15). The beneficial effects of selenium in GO may reflect either a direct effect in orbital tissues, where selenium may counteract oxidative stress, or an effect on the immune system, which has been shown to take place both in GD and autoimmune thyroiditis (16, 2025). Here we investigated some of the possible mechanisms responsible for the action of selenium in GO using primary cultures of orbital fibroblasts subjected to oxidative stress. Materials And Methods Primary cultures of fibroblasts Orbital adipose tissue samples were collected from 6 GO patients who underwent orbital decompression. Normal orbital tissue samples were collected from 6 patients who underwent eye surgery for conditions other than GO not involving fibroadipose tissue. Signed, fully informed Rotondo Dottore et al. Page 3

4 consent was obtained from all patients. Shortly prior to surgery, GO patients underwent, among others, the following assessments and measurements: i) ophthalmological evaluation, including evaluation of the Clinical Activity Score (CAS) according to Mourits (26), and ii) assay for serum autoantibodies against the TSHreceptor (TRAb) (RSR Ltd Cardiff, UK). Primary cultures of orbital fibroblasts were prepared as described (27, 28). Briefly, tissue samples were minced, dispersed in Medium 199 (Sigma, St. Louis, MO) containing 20% fetal bovine serum (Invitrogen Corporation, Carlsbad, CA), penicillin (Sigma) and gentamycin (Sigma). Tissue samples were then kept in 10 cm Petri dishes in cell incubators at 37 C for 2 weeks. Cells attached to dishes were detached by trypsinization and seeded onto 10 cm Petri dishes (2x10 6 cells per dish) in Medium 199 containing 10% fetal bovine serum and antibiotics. Medium was replaced by fresh medium on alternate days. Every passage was performed at confluence, which occurred every 510 days depending on the cells. For all the experiments cells were seeded onto 96 well plates (5,000 cells per well) in complete medium. Experiments were started after 24h, when ~50% confluence had been reached. Induction of oxidative stress and treatment with selenium To induce oxidative stress, cells were incubated for 24 hours at 37 C with fresh, complete medium containing H 2 O 2 at various concentrations, after which the medium was replaced with fresh complete medium and cells were kept in cell incubators at 37 C for 48h. In experiments aimed at assessing the effects of selenium, cells were preincubated for 2 days in cell incubators at 37 C with complete medium without compounds, or with medium containing selenium in the form of Se(Methyl)selenocysteine hydrochloride, (SeMCys, Sigma) or, as a control, methilcysteine (MCys) [S(Methyl)Lcysteine, Sigma] (Sigma) at various concentrations. Then, cells were treated with H 2 O 2 5 µm, as described above. Preparation of cell extracts Cells were washed with PBS and incubated on ice for one hour in lysis buffer (1% Triton X100, 1% deoxycholate in H 2 O). Samples were spun for 10 minutes at 10,000 x g, pellets were discarded and supernatants were collected. Protein concentrations were measured with the Bradford method. Cell assays After incubation with SeMCys or MCys, and/or H 2 O 2, performed as detailed above, media were collected and cell vitality was assessed using a colorimetric reagent (Invitrogen Corporation, Carlsbad, CA), according to the manufacturer s instructions. Glutathione disulfide (GSSG) was measured in cell extracts using a commercial assay (Trevigen, Gaithersburg, MD) according to the manufacturer s instructions. Glutathione peroxidase (GPX) activity was measured using a commercial assay (Enzo Life Scinces, New York, NY) according to the manufacturer s instructions. Rotondo Dottore et al. Page 4

5 Cell proliferation was measured using a commercial assay (Roche Applied Science, Mannheim, DEU) according to the manufacturer s instructions. Hyaluronic acid HA was measured in cell media using a commercial ELISA (Echelon Sciences, Salt Lake City, UT), according to the manufacturer s instructions. TNFα, IL1β and IFNγ were measured in cell media using commercial ELISAs (Invitrogen Corporation, Carlsbad, CA), according to the manufacturer s instructions. Data presentation and statistics Results were normalized for the amounts of proteins in cell extracts, unless otherwise specified. Data are presented as median and IQR. The following tests were performed as appropriate: i) Chisquare; ii) ttest; iii) Wilcoxon; iv) Friedman; v) ANOVA for repeated measures. Results Preliminary tests Primary cultures of orbital fibroblasts were established from 6 patients with GO and 6 patients without GO who underwent eye surgery. The demographical and clinical features of patients are reported in Table 1. GO and control patients were similar in terms of age and gender distribution. All GO patients had a long standing, inactive eye disease, as shown by clinical activity score values 2/7 and by undetectable, or in two cases low, serum TRAb concentrations. All GO patients underwent orbital decompression for disfiguring proptosis. Control patients underwent eye surgery for eyelid ectropion or entropion. There were no patientrelated difference in the results reported below (not shown). In order to induce oxidative stress, we incubated fibroblasts with H 2 O 2. To determine the optimal concentrations of H 2 O 2 to be used, we performed cell vitality assays after treatment of fibroblasts with increasing concentrations of H 2 O 2 for 24 hours. As shown in Fig. 1a, the majority of cells survived up to a H 2 O 2 concentration of 5 µm. At this concentration, which was the lowest at which significant oxidative stress was observed (see below), the median vitality was 86.2% in GO and 79.4% in control fibroblasts. Overall, H 2 O 2 had a significant, dosedependent effect on cell vitality regardless of the cell origin, reflecting cytotoxicity at concentrations 10 µm. The effect of H 2 O 2 on cell vitality was similar in GO and control fibroblasts. We used GSSG as a measure of oxidative stress following incubation of fibroblasts with H 2 O 2. As shown in Fig. 1b, after incubation with increasing concentrations of H 2 O 2, GSSG was detectable in the cell extract. Although there was some GSSG increase already at 2.5 µm, the effect became quite evident at a 5 µm concentration of H 2 O 2, which was slightly lower than the 6.25 µm concentration used in a previous study (29). The effect somehow reached a plateau between 5 and 10 µm. Overall, the effect of H 2 O 2 on GSSG was dosedependent and statistically significant. The apparent difference Rotondo Dottore et al. Page 5

6 between GO and control fibroblasts was not statistically significant. In view of the findings obtained in cell vitality and GSSG release experiments, we decided to use a 5 µm concentration of H 2 O 2 in the experiments reported below. In order to determine the optimal concentration of SeMCys to be used, we assessed the effects of SeMCys on cell vitality by incubating fibroblasts with increasing concentrations of the compound. As shown in Fig. 1c, SeMCys was overall very poorly cytotoxic. Thus, a very slightly reduced cell vitality could be observed only at concentrations 500 µm. Overall, no statistically significant effect was observed and there was no difference between GO and control fibroblasts. The highest vitality was observed at a 10 µm concentration of SeMCys for both GO and control fibroblasts, corresponding to a concentration of atomic selenium of 4.3 µm, because of which we used this concentration in the experiments reported below. Selenium prevents oxidative stress To determine the effects of SeMCys on the oxidative stress caused by H 2 O 2, GSSG was measured in cell extracts following incubation of fibroblasts with H 2 O 2, in cells preincubated or not with SeMCys. As shown in Fig. 2a, the increase of GSSG induced by H 2 O 2 was abolished by preincubation with SeMCys, in a statistically significant manner both in GO and control fibroblasts. MCys, used as a negative control, had no effect. The apparent difference between GO and control fibroblasts was not statistically significant. Selenium is known to exerts its functions after incorporation into several selenoproteins, especially GPX (16, 2425). As shown in Fig. 2b, following incubation of fibroblasts with H 2 O 2, treatment with SeMCys, but not with MCys, resulted in a significant, dosedependent increase in GPX activity, with a peak at a SeMCys concentration of 10 µm (the concentration used in all the experiments reported below), followed by a plateau. The apparent difference between GO and control fibroblasts was not statistically significant. Selenium prevents cell proliferation Fibroblast proliferation is probably the major pathogenetic mechanism of GO (1). Therefore, we investigated the effects of H 2 O 2 and SeMCys on cell proliferation. As reported previously (29), in spite of its cytotoxic effects at high concentrations (see Fig. 1a), the relatively low concentration of H 2 O 2 used (5 µm), which, as reported above was not cytotoxic, determined a significant increase in cell proliferation in GO, but not in control fibroblasts (Fig. 3). SeMCys, but not MCys, abolished the increased proliferation induced by H 2 O 2 in GO fibroblasts. Overall, proliferation was significantly greater in GO than in control fibroblasts, regardless of incubation conditions. Selenium prevents HA release Associated with their increased proliferation, orbital fibroblasts in GO release great amounts of glycosaminoglycans, especially HA, which contributes the increased orbital content (1). We Rotondo Dottore et al. Page 6

7 investigated the effect of H 2 O 2 and SeMCys on HA release by orbital fibroblasts. As shown in Fig. 4, HA in cell media was significantly greater in GO than in control fibroblasts, regardless of the incubation conditions. H 2 O 2 did not affect significantly HA both in GO and control fibroblasts. However, SeMCys reduced significantly HA compared with cells treated with H 2 O 2 in GO fibroblasts, whereas there was no effect in control fibroblasts. MCys had no effect in both cell types. Effect of selenium on proinflammatory cytokines Oxidative stress is known to be associated with increased release of proinflammatory cytokines by fibroblasts, which is likely one of the mechanisms through which oxidative stress contributes GO pathogenesis (1, 30, 31). We investigated the effects of SeMCys on cytokine release following treatment with H 2 O 2. For this purpose, we assessed TNFα, IL1β and IFNγ in the cell medium. As shown in Fig 5a, TNFα increased significantly upon H 2 O 2 treatment, both in GO and control fibroblasts. The effect was virtually abolished by SeMCys, but not by MCys, in both cell groups, to a significant extent, with no difference between GO and control fibroblasts. In contrast, SeMCys had no effect on IL1β. Thus, although treatment with H 2 O 2 was followed by a significant increase of IL1β in both GO and control fibroblasts, SeMCys, and also MCys, did not affect this cytokine (Fig. 5b). There was no difference between GO and control fibroblasts. Finally, we performed similar experiments to investigate the effects of H 2 O 2 and SeMCys on IFNγ. As shown in Fig 5c, IFNγ was significantly increased by H 2 O 2 both in GO and control fibroblasts. SeMCys, but not MCys, abolished the effect of H 2 O 2 in both cell groups to a significant extent, with no difference between GO and control fibroblasts. Discussion In the present study we show an antioxidant action of selenium, in the form of SeMCys, in orbital fibroblasts, which provides a cellular basis for the effects of selenium in vivo in patients with GO (15). Selenium is known to be a quite potent inhibitor of oxidative stress (16), one of the mechanisms involved in the pathogenesis of GO (1719, 30). Here we found that administration of SeMCys to primary cultures of orbital fibroblasts from GO patients or control subjects reduces oxidative stress, which is associated with protection from increased cell proliferation, one of the major pathogenetic mechanisms of GO (1, 30). In addition, SeMCys reduces HA release, another major mechanism involved in GO pathogenesis (1, 30). Finally, SeMCys prevents some of the microenvironmental changes induced by oxidative stress, namely the increased release of some of the cytokines involved in GO pathogenesis (1, 2931), which is likely one of the mechanisms through which oxidative stress, and therefore selenium, act on GO. Findings supporting our conclusions are summarized as follows. H 2 O 2 had an oxidative effect in orbital fibroblasts, reflected by a dosedependent increase in GSSG, a known measure of cell response to free radicals (32). Preincubation of cells with SeMCys Rotondo Dottore et al. Page 7

8 abolished the effects of H 2 O 2, therefore providing evidence for an antioxidant action of SeMCys in orbital fibroblasts. The effects of SeMCys were observed both in fibroblasts from GO patients or from subjects without GO, suggesting that fibroblasts are susceptible to oxidative stress and to protection by SeMCys regardless of the underlying conditions and/or of their genetic background. Treatment of cells with SeMCys was followed by an increase in GPX activity in both GO and control fibroblasts. As mentioned above, selenium acts after incorporation into various selenoproteins, including GPX, where selenium is located on the catalytic site of the enzyme, being an essential component of its antioxidant action (16, 2425). Our findings indicate that GPX is very likely involved in protection from oxidative stress by selenium in orbital fibroblasts. It remains to be investigated whether also other selenoproteins are involved, which, however, was beyond the purpose of the present study. As mentioned above, fibroblast proliferation is one of the two major mechanisms responsible for the increased orbital volume in GO (1, 30). In confirmation of a previous study (29), a relatively low concentration of H 2 O 2 (5 µm) determined a significant increase in fibroblast proliferation, indicating that oxidative stress induces cell proliferation. The findings indicate that, concerning this particular feature, and as reported below also concerning the release of HA, cells from GO patients are somehow different. This may reflect a different response to an initial pathogenetic event, presumably an autoimmune insult, or a different genetic background that renders them more proliferative. Obviously, these explanations are entirely speculative and further studies are needed to clarify the mechanisms responsible for this different cell behaviour. In GO fibroblasts, the effect of H 2 O 2 in terms of proliferation was counteracted by SeMCys, which brought proliferation back to the levels observed in untreated cells. It is worth noting that H 2 O 2 seems to exert a dual effect on fibroblasts. If used at doses exceeding 5 µm, it displays cytotoxic actions, as observed in cell vitality experiments, reflecting, its well known, direct cytotoxic effect on cells. On the contrary, if used at concentrations 5 µm, H 2 O 2 is poorly cytotoxic and it rather stimulates cell proliferation, presumably indirectly, as a consequence of the induction of oxidative stress. Similar findings were obtained by Tsai et al, who used a 6.25 µm concentration of H 2 O 2 (29). H 2 O 2 did not affect significantly HA release in cultured fibroblasts. Nevertheless, in GO fibroblasts SeMCys reduced significantly HA compared with cells treated with H 2 O 2, bringing HA release to levels that were even lower than those observed in untreated cells. The finding seems to indicate that SeMCys acts on HA release at least in part regardless of the oxidative stress induced by H 2 O 2, through mechanisms that remain to be investigated. SeMCys had no effect on orbital fibroblasts from subjects without GO, a possible explanation of which is the same reported above concerning proliferation. To gain some insights into the mechanisms induced by oxidative stress and involved in GO pathogenesis on which selenium may act, we investigated the release of proinflammatory cytokines Rotondo Dottore et al. Page 8

9 upon oxidative challenge (treatment with H 2 O 2 ) in orbital fibroblasts. In a previous study it was shown that, in addition to immunocompetent cells, also fibroblasts are capable of releasing proinflammatory cytokines, in particular TNFα, IL1β and IFNγ, in response to oxidative stress (29). In turn, these cytokines stimulate HA synthase, leading to increase HA release (30). In addition to these actions, IFNγ is known to regulate fibroblast growth and collagen biosynthesis, and TNFα is able to stimulate fibroblast proliferation (30). As reported previously (29), H 2 O 2 determined an increased release of TNFα, IL1β and IFNγ in both GO and control fibroblasts. The effect of H 2 O 2 on TNFα and IFNγ, but not on IL1β, was abolished by SeMCys in both cell types, suggesting that SeMCys acts on the oxidative stressinduced release of some proinflammatory cytokines, which may be responsible for some of the beneficial effects of selenium in GO fibroblasts. The different behaviour of proinflammatory cytokines in response to selenium is quite intriguing and requires further investigations to shed light on the responsible mechanisms. In this regard, a highly speculative hypothesis involves the IL1receptor antagonist (IL1RA), a natural competitor for binding to the IL1receptor (IL1R) that is synthesized and released by the same cells that produce IL1 (31). Selenium may determine independently an increase in IL1RA, which, by competing for IL1R, may result in the persistence of increased IL1β in the cell medium following exposure of the cells to H 2 O 2. A similar phenomenon cannot occur for TNFα and IFNγ, as they do not have natural antagonists. The fact that IL1RA itself has antioxidant actions is also quite intriguing (32). Clearly, further studies are needed to investigate this hypothesis. Because the respective role and relative importance of the various proinflammatory cytokines in the cellular and biochemical milieu of orbital tissue in GO is not established, we do not know whether the different effects of selenium on proinflammatory cytokines have any implications on its actions in vivo in patients with GO. IL1β is more effective in eliciting HA synthase in skin fibroblasts compared with TNFα and especially with IFNγ (29), but, to our knowledge, no data are available in orbital fibroblasts, nor it is known if and to what extent orbital tissue cytokines of GO patients are affected by selenium. Our study carries the obvious limitations of in vitro investigations. In our cell system only orbital fibroblasts were present, unlike in vivo where immune and inflammatory cells determine a much more complex environment, which, among other not less important actors, probably includes also autoantibodies, especially against the TSHreceptor, which is largely considered the most important autoantigen in GO (1). A further development of our investigations may foresee the use of more complex culture systems, for example with the addition of the well known monoclonal antitshreceptor antibody named M22 (33). As mentioned above, here we used selenium in the form of SeMCys. However, this is only one of the various seleniumcontaining compounds, which are known to be metabolized by different biochemical cell pathways and have eventually different extents of action in cells (16, 23, 24, 34). Rotondo Dottore et al. Page 9

10 Thus, it is possible that different forms of selenium may be less or more effective than SeMCys in fibroblasts, which will be the subject of further studies. On the same line, selenium is known to be incorporated into selenoproteins, which in turn act as reduction oxidation centers (16, 24, 34). Here we demonstrated an increased GPX activity induced by selenium in orbital fibroblasts, but additional studies are needed to investigate if other selenoproteins contribute the actions of selenium. As mentioned above, our investigations stemmed on the observation of a beneficial effect of selenium compared with placebo in patients with GO, as shown by a multicenter, doubleblind, randomized clinical trial conducted by the European Group on Graves Orbitopathy (EUGOGO) (15). Treatment with selenium was followed by an improvement of the overall GO outcome, reflecting amelioration of soft tissue involvement and eyelid aperture. The quality of life of patients increased accordingly. The EUGOGO study included only patients with mild GO (15), but in view of our findings indicating a direct action in orbital fibroblasts, we postulate that selenium may exert some beneficial effect also in moderatesevere GO. Further studies are needed to investigate this possibility. In conclusion, our study provides some cellular bases for the beneficial effects of selenium observed in patients with mild GO (15). Thus, selenium has a direct action in orbital fibroblasts, presumably the major target of the immune system in GO. It remains to be investigated whether the effects of selenium in patients with GO reflect also its known actions on the immune system (16, 24, 25). Rotondo Dottore et al. Page 10

11 Acknowledgments Unconditionally supported by IBSA Farmaceutici Italia, Lodi, Italy Disclosure Statement Giovanna Rotondo Dottore, Marenza Leo, Giamberto Casini, Francesco Latrofa, Luca Cestari, Stefano SellariFranceschini, Marco Nardi, Paolo Vitti, Claudio Marcocci and Michele Marinò declare that they do not have any commercial association that might create a conflict of interest in connection with this manuscripts. Rotondo Dottore et al. Page 11

12 References 1 Bahn RS 2010 Graves' ophthalmopathy. N Engl J Med 362: Bartalena L 2013 Diagnosis and management of Graves disease: a global overview. Nat Rev Endocrinol 9: Piantanida E, Tanda ML, Lai A, Sassi L, Bartalena L 2013 Prevalence and natural history of Graves' orbitopathy in the XXI century. J Endocrinol Invest 36: Leo M, Menconi F, Rocchi R, Latrofa F, Sisti E, Profilo MA, Mazzi B, Albano E, Nardi M, Vitti P, Marcocci C, Marinò M 2015 Role of the underlying thyroid disease on the phenotype of Graves orbitopathy in a tertiary referral center. Thyroid 25: Sisti E, Menconi F, Leo M, Profilo MA, Mautone T, Mazzi B, Rocchi R, Latrofa F, Nardi M, Vitti P, Marcocci C, Marinò M 2015 Longterm outcome of Graves' orbitopathy following highdose intravenous glucocorticoids and orbital radiotherapy. J Endocrinol Invest 38: Menconi F, Leo M, Sabini E, Mautone T, Nardi M, Sainato A, SellariFranceschini S, Vitti P, Marcocci C, Marinò M 2016 Natural history of Graves orbitopathy after treatment. Endocrine. [in press]. 7 Menconi F, Profilo MA, Leo M, Sisti E, Altea MA, Rocchi R, Latrofa F, Nardi M, Vitti P, Marcocci C, Marinò M 2014 Spontaneous Improvement of Untreated Mild Graves' Ophthalmopathy: Rundle's Curve Revisited. Thyroid 24: Gerding MN, Terwee CB, Dekker FW, Koornneef L, Prummel MF, Wiersinga WM 1997 Quality of life in patients with Graves' ophthalmopathy is markedly decreased: measurement by the Medical Outcomes Study instrument. Thyroid 7: Terwee CB, Gerding MN, Dekker FW, Prummel MF, Wiersinga WM 1998 Development of a disease specific quality of life questionnaire for patients with Graves' ophthalmopathy: the GOQOL. Br J Ophthalmol 82: Perros P, Crombie AL, KendallTaylor P 1995 Natural history of thyroid associated ophthalmopathy Clin Endocrinol (Oxf) 42: Zang S, Ponto KA, Kahaly GJ 2011 Clinical review: Intravenous glucocorticoids for Graves' orbitopathy: efficacy and morbidity. J Clin Endocrinol Metab 96: Sisti E, Coco B, Menconi F, Leo M, Rocchi R, Latrofa F, Profilo MA, Mazzi B, Albano E, Vitti P, Marcocci C, Brunetto M, Marinò M 2015 Intravenous glucocorticoid therapy for Graves ophthalmopathy and acute liver damage: an epidemiological study. Eur J Endocrinol 172: Sisti E, Coco B, Menconi F, Leo M, Rocchi R, Latrofa F, Profilo MA, Mazzi B, Vitti P, Marcocci C, Brunetto M, Marinò M 2015 Age and Dose Are Major Risk Factors for Liver Damage Associated with Intravenous Glucocorticoid Pulse Therapy for Graves' Orbitopathy. Thyroid 25: Moleti M, Giuffrida G, Sturniolo G, Squadrito G, Campennì A, Morelli S, Puxeddu E, Sisti E, Trimarchi F, Vermiglio F, Marinò M 2016 Acute liver damage following intravenous glucocorticoid treatment for Graves' ophthalmopathy. Endocrine [in press] Rotondo Dottore et al. Page 12

13 15 Marcocci C, Kahaly GJ, Krassas GE, Bartalena L, Prummel M, Stahl M, Altea MA, Nardi M, Pitz S, Boboridis K, Sivelli P, von Arx G, Mourits MP, Baldeschi L, Bencivelli W, Wiersinga W 2011 Selenium and the Course of Mild Graves' Orbitopathy. N Engl J Med 364: Rayman MP 2000 The importance of selenium to human health. Lancet 356: Burch HB, Lahiri S, Bahn RS, Barnes S 1997 Superoxide radical production stimulates retroocular fibroblast proliferation in Graves' ophthalmopathy. Exp Eye Res 65: Heufelder AE, Wenzel BE, Bahn RS 1992 Methimazole and propylthiouracil inhibit the oxygen free radicalinduced expression of a 72 kilodalton heat shock protein in Graves' retroocular fibroblasts. J Clin Endocrinol Metab 74: Lu R, Wang P, Wartofsky L, Sutton BD, Zweier JL, Bahn RS, Garrity J, Burman KD 1999 Oxygen free radicals in interleukin1betainduced glycosaminoglycan production by retroocular fibroblasts from normal subjects and Graves' ophthalmopathy patients. Thyroid 9: Hoffmann PR, Berry MJ 2008 The influence of selenium on immune responses. Mol Nutr Food Res 52: Negro R. Selenium and thyroid autoimmunity 2008 Biologics 2: Nacamulli D, Mian C, Petricca D, Lazzarotto F, Barollo S, Pozza D, Masiero S, Faggian D, Plebani M, Girelli ME, Mantero F, Betterle C 2010 Influence of physiological dietary selenium supplementation on the natural course of autoimmune thyroiditis. Clin Endocrinol (Oxf) 73: Vrca VB, Skreb F, Cepelak I, Romic Z, Mayer L 2004 Supplementation with antioxidants in the treatment of Graves' disease: the effect on glutathione peroxidase activity and concentration of selenium. Clin Chim Acta 341: Köhrle J 2015 Selenium and the thyroid. Curr Opin Endocrinol Diabetes Obes 22: Duntas LH, Benvenga S 2015 Selenium: an element for life. Endocrine 48: Mourits MP, Prummel MF, Wiersinga WM, Koornneef L 1997 Clinical activity score as a guide in the management of patients with Graves' ophthalmopathy. Clin. Endocrinol. (Oxf) 47: Botta R, Lisi S, Marcocci C, SellariFranceschini S, Rocchi R, Latrofa F, Menconi F, Altea MA, Leo M, Sisti E, Casini G, Nardi M, Pinchera A, Vitti P, Marinò M 2013 Enalapril reduces proliferation and hyaluronic acid release in orbital fibroblasts. Thyroid 23: Lisi S, Botta R, Lemmi M, SellariFranceschini S, Altea MA, Sisti E, Casini G, Nardi M, Marcocci C, Pinchera A, Marinò M 2011 Quercetin decreases proliferation of orbital fibroblasts and their release of hyaluronic acid. J Endocrinol Invest 34: Tsai CC, Wu SB, Kao SC, Kau HC, Lee FL, Wei YH 2013 The protective effect of antioxidants on orbital fibroblasts from patients with Graves' ophthalmopathy in response to oxidative stress. Mol Vis 19: Campo GM, Avenoso A, Campo S, Angela D, Ferlazzo AM, Calatroni A 2006 TNFalpha, IFNgamma, and IL1beta modulate hyaluronan synthase expression in human skin fibroblasts: synergistic effect by concomital treatment with FeSO4 plus ascorbate. Mol Cell Biochem 292: Rotondo Dottore et al. Page 13

14 31 Garlanda C, Dinarello CA, Mantovani A 2013 The interleukin1 family: back to the future. Immunity 39: Wahab F, SantosJunior NN, de Almeida Rodrigues RP, Costa LH, Catalão CH, Rocha MJ 2016 Interleukin1 Receptor Antagonist Decreases Hypothalamic Oxidative Stress During Experimental Sepsis. Mol Neurobiol 53: Bahn RS 2015 Current Insights into the Pathogenesis of Graves' Ophthalmopathy. Horm Metab Res 47: Furmaniak J, Sanders J, Núñez Miguel R, Rees Smith B 2015 Mechanisms of Action of TSHR Autoantibodies. Horm Metab Res 47: Misra S, Boylan M, Selvam A, Spallholz JE, Björnstedt M 2015 Redoxactive selenium compoundsfrom toxicity and cell death to cancer treatment. Nutrients 7: Rotondo Dottore et al. Page 14

15 Figure Legends Figure 1. a) Effects of various concentrations of H 2 O 2 on cell vitality in fibroblasts from patients with Graves orbitopathy (GO fibroblasts) or from control subjects. *P< by Friedman test; **P< by Friedman test; P=NS between GO and control fibroblasts by ANOVA for repeated measures; b) Effects of various concentrations of H 2 O 2 on Glutathione disulfide (GSSG). *P= by Friedman test; **P=0.001 by Friedman test; P=NS between GO and control fibroblasts by ANOVA for repeated measures; c) Effects of various concentrations of selenium(methyl)selenocysteine (SeMCys) on cell vitality. *P=NS by Friedman test; **P=NS by Friedman test; P=NS between GO and control fibroblasts by ANOVA for repeated measures. Figure 2. a) Combined effects of H 2 O 2 (5 µm), selenium(methyl)selenocysteine (SeMCys) (10 µm) or its control methilcysteine (MCys) (10 µm) on Glutathione disulfide (GSSG) in fibroblasts from patients with Graves orbitopathy (GO fibroblasts) or from control subjects. * and **P=0.02 vs H 2 O 2 by Wilcoxon; and P=NS vs H 2 O 2 by Wilcoxon; P=NS between GO and control fibroblasts by ANOVA for repeated measures. b) Effects of increasing concentrations of SeMCys or MCys (10 µm) on glutathione peroxidase (GPX) activity in fibroblasts from GO patients or from control subjects. *P=0.03 by Wilcoxon between SeMCys 0 and 10 µm in GO and control fibroblasts. P=NS between GO and control fibroblasts by ANOVA for repeated measures. Figure 3. Combined effects of H 2 O 2 (5 µm), selenium(methyl)selenocysteine (SeMCys) (10 µm) or its control methilcysteine (MCys) (10 µm) on cell proliferation in fibroblasts from patients with Graves orbitopathy (GO fibroblasts) or from control subjects. *P=0.02 vs untreated cells by Wilcoxon; **P=0.02 vs H 2 O 2 by Wilcoxon; P=0.003 between GO and control fibroblasts by ANOVA for repeated measures. Figure 4. Combined effects of H 2 O 2 (5 µm), selenium(methyl)selenocysteine (SeMCys) (10 µm) or its control methilcysteine (MCys) (10 µm) on hyaluronic acid (HA) release in fibroblasts from patients with Graves orbitopathy (GO fibroblasts) or from control subjects. *P=0.02 vs H 2 O 2 by Wilcoxon; P=0.02 between GO and control fibroblasts by ANOVA for repeated measures Figure 5. Combined effects of H 2 O 2 (5 µm), selenium(methyl)selenocysteine (SeMCys) (10 µm) or its control methilcysteine (MCys) (10 µm) on proinflammatory cytokines in fibroblasts from patients with Graves orbitopathy (GO fibroblasts) or from control subjects. a) effects on tumor necrosis factor α (TNFα). * and **P=0.02 vs untreated cells by Wilcoxon; and P=0.02 vs H 2 O 2 by Wilcoxon; b) effects on interleukin1β (IL1β). *P=0.03 and **P=0.03 vs untreated cells by Wilcoxon; c) effects on interferonγ IFNγ. * P=0.02 and **P=0.02 vs H 2 O 2 by Wilcoxon; P=0.023 and P=0.05 vs H 2 O 2 by Wilcoxon. P=NS by ANOVA for repeated measures between GO and control fibroblasts for all three panels. Rotondo Dottore et al. Page 15

16 a Cell vitality (% absorbance/µg of cell protein) GO fibroblasts Control fibroblasts 0 2, H 2 O 2 (µm) b GSSG [(pm/well)/cell number] GO fibroblasts Control fibroblasts 0 0 2, H 2 O 2 (µm) c Cell vitality (% absorbance/µg of cell protein) GO fibroblasts Control fibroblasts SeMCys (µm) Rotondo Dottore et al. Fig. 1 Rotondo Dottore et al. Page 16

17 Rotondo Dottore et al. Fig. 2 Rotondo Dottore et al. Page 17

18 Rotondo Dottore et al. Fig. 3 Rotondo Dottore et al. Page 18

19 300 GO fibroblasts Hyaluronic Acid (ng/mg of cell protein) Control fibroblasts * 150 H 2 O 2 SeMCys MCys Rotondo Dottore et al. Fig. 4 Rotondo Dottore et al. Page 19

20 a 15 GO fibroblasts TNFα [(pg/ml) µg of cell protein] * ** Control fibroblasts 0 H 2 O 2 SeMeCys MCys No treatment H2O2 SeMeSeCys MeCys b 60 IL1β [(pg/ml) µg of cell protein] * ** GO fibroblasts Control fibroblasts 0 H 2 O 2 SeMCys MCys No treatment H2O2 SeMeSeCys MeCys c 40 IFNγ [(pg/ml) µg of cell protein] * ** GO fibroblasts Control fibroblasts 0 H 2 O 2 SeMCys MCys No treatment H2O2 SeMeSeCys MeCys Rotondo Dottore et al. Fig. 5 Rotondo Dottore et al. Page 20

21 Table 1. Features of patients at baseline. Age is reported as mean±sd. P values were obtained by Chisquare (gender) and ttest (age). GO: Graves orbitopathy; TRAb: antitshreceptor autoantibodies. Feature GO patients Control patients P Gender Men: 2; Women: 4 Men: 3; Women: 3 NS Age (yr) 60.1± ±8.4 NS Patient 1: 2/7 Patient 2: 2/7 Clinical Activity Score Patient 3: 1/7 Patient 4: 2/7 Patient 5: 0/7 Patient 6: 1/7 N/A N/A Serum TRAb (U/L) Patient 1: 6.6 Patient 2: <0.3 Patient 3: <0.3 Patient 4: 5.3 Patient 5: <0.3 Patient 6: <0.3 N/A N/A GO duration (years) Diagnosis/Indication to surgery Patient 1: 8 Patient 2: 13 Patient 3: 7 Patient 4: 5 Patient 5: 2 Patient 6: 11 Disfiguring proptosis N/A Eyelid ectropion (4 patients) Eyelid entropion (2 patients) N/A N/A Rotondo Dottore et al. Page 21