Fibrotic gene expression coexists with alveolar proteinosis in early indium lung.

Abstract Occupational inhalation of indium compounds can cause the so-called “indium lung disease”. Most affected individuals show pulmonary alveolar proteinosis (PAP) and fibrotic interstitial lung disease. In animal experiments, inhalation of indium tin oxide or indium oxide has been shown to cause lung damage. However, the mechanisms by which indium compounds lead to indium lung disease remain unknown. In this study, we constructed a mouse model of indium lung disease and analyzed gene expression in response to indium exposure. Indium oxide (In2O3, 10 mg/kg, primary particle size <100 nm) was administered intratracheally to C57BL/6 mice (male, 8 weeks of age) twice a week for 8 weeks. Four weeks after the final instillation, histopathological analysis exhibited periodic acid-Schiff positive material in the alveoli, characteristic of PAP. Comprehensive gene expression analysis by RNA-Seq, however, revealed expression of fibrosis-related genes, such as surfactant associated protein D, surfactant associated protein A1, mucin 1, and collagen type I and III, was significantly increased, indicating that fibrotic gene expression progresses in early phase of indium lung. These data supported the latest hypothesis that PAP occurs as an acute phase response and is replaced by fibrosis after long-term latency.


Introduction
Indium compounds are materials used in the manufacture of transparent conductive films for flat panel displays (Yoshimura et al., 2013). Their demand has been rapidly increasing since the 2000s. Indium tin oxide (ITO) is the most commonly used compound; it consists of indium oxide and about 10% (wt) tin oxide.
Inhalation of indium compounds is known to cause respiratory damage. Occupational inhalation of ITO is linked to a new occupational disease named ''indium lung disease'' (Cummings et al., 2012). The first case of indium lung disease was confirmed in 2003, which manifested with interstitial lung disease accompanied by emphysema (Homma et al., 2003). This indium-induced emphysema resulted in the bilateral pneumothorax that eventually lead to the death of the patient, which raised the concern of the Industrial Safety and Health Department in Ministry of Health, Labor, and Welfare of Japan in 2004 (Ministry of Health, Labour and Welfare, 2004). Epidemiological studies of indium-exposed workers also revealed a dose-dependent relationship between indium exposure and lung disease characterized by interstitial fibrosis, emphysematous changes and alveolar inflammation (Chonan et al., 2007;Nakano et al., 2009Nakano et al., , 2014. According to a report in 2012, at least 10 cases of indium lung disease had been reported worldwide (Cummings et al., 2012). Among these nine cases were exposed to ITO, whereas one case was exposed to indium oxide, which is a toxicological equivalent of ITO. Most of these cases were radiologically and pathologically diagnosed with interstitial fibrosis, however, some cases showed characteristic of fibrosis with pulmonary alveolar proteinosis (PAP). In two cases diagnosed as PAP in initial diagnosis, fibrosis developed during followup, indicating indium inhalation cause fibrosis after PAP. As we have one unpublished indium lung case caused by indium oxide exposure, we have focused the toxicity of indium oxide.
In animal experiments, intratracheal administration of ITO or indium oxide caused hamsters to develop respiratory disorders including fibrotic proliferation (Tanaka et al., 2002(Tanaka et al., , 2010. Inhalation of either of ITO and indium oxide resulted in inflammation or PAP at the early stage, and long-term observation additionally showed fibrosis (Nagano et al., 2011a,b,c). Despite the histological data gathered in these animal experiments, these findings did not fully explain how indium lung disease begin with PAP and progress to fibrosis at the molecular level. In this study, we established an indium lung mouse model manifesting PAP by intratracheal administration of indium oxide to elucidate gene expression changes in the early phase of indium lung, or what we call early indium lung. We performed an exhaustive gene expression analysis by high-throughput sequencing, and RNA-sequencing (RNA-Seq) data were used to identify important genes in pulmonary proteinosis and fibrosis. Additionally, we explored other differentially expressed genes.

Animals and indium oxide administration
To establish a mouse model of indium lung disease, 10 mg/kg of indium oxide (In 2 O 3 , primary particle size5100 nm, secondary particle size 462 ± 236 nm (Supplementary Figure  S1) (Sigma, St. Louis, MO) was administered intratracheally to C57BL/6 mice (male, 8 weeks of age; Japan SLC, Inc., Shizuoka, Japan) twice a week for 8 weeks. For the control group, saline was intratracheally administered instead. Four weeks after the final administration, mice were euthanized and autopsied. All experiments were approved by the Ethical Committee for Animal Experimentation, Kochi Medical School, Kochi University, Japan.

Histopathology
The left lung was excised and immersed in 10% formalin, dehydrated, and embedded in paraffin. Four-micrometre tissue sections were stained with hematoxylin and eosin (H&E), periodic acid-Schiff (PAS), or Masson trichrome.

RNA extraction and RNA-Seq
The right lung tissue was homogenized with TRI Reagent (Molecular Research Center Inc., Cincinnati, OH), and total RNA was extracted according to the manufacturer's instructions. Library constructions and sequencing were provided as a custom service of Eurofins Genomics K.K. (Tokyo, Japan). Sequencing was performed using the Illumina HiSeq 2000 platform with a 2 Â 100 bp module of version 3 chemistry. Sequencing results were mapped using the TopHat v2.0.13 software and the mouse genome reference sequence mm10. The mean mRNA expression value (FPKM) of each gene was calculated using the Cuffdiff program in the Cufflinks v.2.2.1 software. A q-value (false discovery rate-adjusted p value) 50.05 compared with the saline control sample was considered statistically significant, as calculated using Cuffdiff (Trapnell et al., 2012).

Histopathology of the lungs of indium oxide-administered mice
The excised lung tissue from indium oxide administered mice was stained with H&E or PAS ( Figure 1) and observed microscopically. Presumed indium oxide particles were evident in some areas. There were also areas of cellular infiltration were evident. The accumulation of PAS-positive material was evident in the alveoli, characteristic of PAP. However, no cholesterol clefts or interstitial fibrosis were observed.

RNA-Seq analysis in indium oxide-administered mice
RNA-Seq analysis was performed on the homogenized lung tissue of indium oxide-administered mice to examine changes in gene expression compared with saline control mice (n ¼ 4 per group) ( Figure 2) (DDBJ Sequence Read Archive accession number DRA DRA004749). The average number of total reads was 40,408,945, and on average 88.7% of each fragment was mapped. In total, 3730 genes showed significant differences in expression between the indium oxide and control samples. These genes included the PAP-or fibrosisrelated genes (Supplementary Table S1).

Fibrosis-related gene expression
The expression of the Muc1 (mucin 1), Sftpd (surfactant protein D), and Sftpa1 (surfactant protein A1) genes, known fibrosis marker protein coding genes, was significantly increased in indium oxide-administered mice (Figure 4a and c). The expression of Tgfb1 (transforming growth factor beta 1), a key factor in fibrosis, was also increased. Expression of the Smad genes, Smad6 and Smad7, was decreased ( Figure 4b). Expression of the collagen genes, Col1a1 (collagen type I A1), Col1a2 (collagen type I A2), and Col3a1 (collagen type III A1), that accumulate in fibrosis was significantly increased in indium oxide-administered mice ( Figure 4a). Timp1 (tissue inhibitor of metalloproteinase 1), another fibrosis-related gene was also increased in expression in exposed mice ( Figure 4d). Tjp1 (tight junction protein 1), an epithelial marker known to be decreased in fibrosis, was significantly decreased in expression ( Figure 4e), by contrast, the expression of mesenchymal genes such as S100a4 (S100 calcium binding protein A4), and Vim (vimentin) was increased in indium oxide-administered mice (Figure 4f and g). Some genes showed no significant difference in expression between control and indium oxideadministered mice including the mesenchymal and epithelial marker genes Acta2 (a-SMA), and Cdh1 (E-cadherin), respectively ( Figure 4 h and i). Expression of the mesenchymal marker Cdh2 (N-cadherin), however, was significantly decreased (Figure 4j).

Discussion
This is the first study to report comprehensive gene expression analysis using RNA-Seq in a mouse model of indium lung disease, which we established using intratracheal instillation. Animal models have previously been employed to investigate indium lung disease using either intratracheal or inhalational administrations. Although the toxicity of indium Figure 1. Histopathology of the lung tissue from indium oxide-administered mice compared with control mice. H&E (a-d), PAS (e, f) or Masson trichrome (g, h) staining of the lung sections of mice administered with saline (a, e, g) or indium oxide (In 2 O 3 ) (b-d, f, h). Black arrows indicate In 2 O 3 particles (b, f). Arrowheads indicate PAS-positive material in the alveoli (f). Arrow indicates macrophage including In2O3 particles (d). Bars indicate 200 mm (a,b,g,h), 50 mm (c, e, f) or 10 mm (d), respectively.
oxide was first reported in 1961 (Leach et al., 1961), that of ITO was reported in 2002 after first patient of indium lung disease died. In this report, 8-week-old hamsters were administered 6.0 mg/kg of ITO particles by intratracheal instillation along with ether once a week, for a total of 16 times. Exposure to ITO resulted in an inflammatory response, with the infiltration of alveolar macrophages, necrotic cell debris, and inflammatory cells into the lung and the accumulation of alveolar macrophages in the alveolar spaces (Tanaka et al., 2002). Another report using the same animal model investigated indium lung disease using indium oxide instead of ITO. They administered 2.7 mg/kg or 5.4 mg/kg of indium oxide to hamsters intratracheally twice a week for 8 weeks. Using the higher dose, they observed interstitial fibrotic proliferation 40 weeks after the final instillation (Tanaka et al., 2010). This type of cellular proliferation was not however evident 16 weeks after the final instillation, suggesting that some effects of indium oxide exposure may have long latency. In a separate study using a mouse model, mice were exposed to ITO or indium oxide aerosol for 6 hours/day, 5 days/week for 2 weeks at a concentration of 10 or 100 mg/m 3 using inhalation exposure chambers, and PAP was observed in the lungs after the exposure duration (Nagano et al., 2011b). In the current study, we focused indium oxide known as one of the causes of indium lung disease, and administered indium oxide into mice      intratracheally in the same schedule with previous study which used hamsters (Tanaka et al., 2010). The concentration of indium oxide was 10 mg/kg, about two times higher than this previous study. Our preliminary study revealed the significant increasing of relative lung weight in 10 mg/kg comparing other lower concentrations (0, 1.25, 2.5, and 5 mg/kg indium oxide), which suggests that inflammation is clearly caused in this concentration. This is the first report of mouse model of indium lung disease using intratracheal administration method, although there is the limitation that the character of particle, dose and mode of exposure would be different with the situation in the workplace handling indium. In our present study, histopathology of the lung tissue revealed PAP and the accumulation of PAS-positive material in the alveoli, confirming the findings of previous studies on indium lung disease. These findings confirmed that intratracheal administration and inhalation via exposure chambers both provided similar results suggesting that intratracheal administration is a cost-effective method to study indium lung disease in the mouse model. Furthermore, mice have been better defined genetically than other rodent species, i.e. hamsters and rats. Therefore, we proposed that this mouse model was the most suitable for analyzing the genetic mechanisms involved in indium lung disease. PAP was first described in 1958 (Rosen et al., 1958), and by 2002, over 400 cases were reported (Seymour & Presneill, 2002). PAP occurs in three major distinct forms. The first form is congenital PAP caused by mutation of surfactant protein genes (Seymour & Presneill, 2002). The second one is autoimmune PAP caused by anti-GM-CSF auto-antibodies (Borie et al., 2011;Carey & Trapnell, 2010;Huffman et al., 1996;Kitamura et al., 1999). The last one is PAP that does not fit in the other two categories, for example, lysinuric protein intolerance, immunodeficiency disorders (Seymour & Presneill, 2002), and caused by inhalation of materials, such as silica (Xipell et al., 1977), aluminium (Miller et al., 1984), titanium (Keller et al., 1995). Mechanisms of congenital and autoimmune PAP (Trapnell et al., 2003) are well known, however, there are little information about molecular mechanism of PAP caused by inhalation of inorganic materials (Badding et al., 2014(Badding et al., , 2015Lison et al., 2009). In order to confirm indium-induced PAP in our mouse model is not autoimmune PAP, we analyzed the expression of genes related to GM-CSF pathway. GM-CSF is a glycoprotein cytokine that stimulates proliferation and maturation of macrophages (Gasson, 1991) and is necessary for surfactant clearance (Yoshida et al., 2001). Anti-GM-CSF antibodies neutralize GM-CSF and inhibit the downstream of the GM-CSF pathway, resulting in the impairment of the surfactant clearance by alveolar macrophage (Trapnell et al., 2003). Csf2 (encoding GM-CSF) gene expression was found to be increased in the lungs of mice exposed to indium oxide. Additionally, the expression of genes downstream of the GM-CSF pathway, the receptors of Csf2, and downstream genes of GM-CSF pathway, namely Spi1 (encoding PU.1), Cd14, Cd180, and Irak3 (Carey & Trapnell, 2010;Trapnell & Whitsett, 2002), were also increased. These results suggested that the GM-CSF pathway is intact and correspond with the findings in many human indium lung disease cases. Acute silicosis, fibrosis caused by silica exposure, is recognized as silicoproteinosis because of histologic resemblance to PAP (Castranova & Vallyathan, 2000;Xipell et al., 1977). Silicoproteinosis is seen as a form of not congenital or autoimmune PAP (Xiao et al., 2015). Early indium lung in our model also shows histopathological features of PAP, which would not be congenital or autoimmune PAP. This study will be a starting point to understand the molecular mechanism of exposure-related PAP.
In pulmonary fibrosis, TGF-b1 is an important key regulator in the progression of fibrosis. In our mouse model of indium lung disease, Tgfb1 (encoding TGF-b1) gene expression was significantly increased. By contrast, Smad6 and Smad7, downstream inhibitors of the TGF-b1 signaling pathway (Schmierer & Hill, 2007), were significantly decreased in expression. These Smad genes have previously been reported to be decreased in expression in a bleomycininduced fibrosis mouse model (Peng et al., 2013). Additionally, the expression of type I and type III collagen genes, known to accumulate during fibrosis, was significantly increased in the indium lung disease model. However, no collagen deposits were detected in indium oxide-administered mice by histopathology. Genes encoding proteins used as a biomarkers of fibrosis, Timp1 (Peng et al., 2013), S100a4 (Tanjore et al., 2009), and Vim (Kalluri & Weinberg, 2009) were also significantly increased in expression, as were Muc1, Sftpd, and Sftpa1, the blood markers of fibrosis Krebs von den Lungen (KL-6), sialylated carbohydrate antigen related to Muc1, surfactant protein (SP)-A, and SP-D cording genes, respectively (Greene et al., 2002;Ishikawa et al., 2012;Kuroki et al., 1998;Samukawa et al., 2012). High level of serum KL-6 was observed in the patient of indium lung disease (Homma et al., 2005) and indium-processing workers (Chonan et al., 2007). These results suggested that the process of fibrosis had been initiated in terms of the gene expression profile but not on the histopathological features in early indium lung. The expression of several epithelial or mesenchymal marker genes, such as Cdh1 or Acta2, was not changed suggesting that epithelial-to-mesenchymal transition, a key phenomenon in fibrogenesis (Kalluri & Weinberg, 2009;Noguchi et al., 2015), had not occurred or was undetectable, possibly due to RNA extraction from the whole lung. In histopathology, there were no collagen accumulation characteristics of pulmonary fibrosis, and there were no cholesterol clefts, observed in many human indium lung cases (Cummings et al., 2012), and observed in long-term observation of ITO-or indium oxide-treated hamsters (Tanaka et al., 2010). That is, fibrosis did not develop completely but partially at the gene expression level. Although our data on PAP and fibrosis-related gene expression indicated partial PAP and fibrosis in the subacute phase, they also supported the latest hypothesis that fibrosis occurred after PAP in indium-exposed human cases (Cummings et al., 2012). This is the first report to reveal that exposure of indium compounds causes fibrotic gene expression changes as well as the histopathology of PAP in early indium lung. Long-term observations in our indium lung model may uncover the histopathological features of fibrosis.
Gene expression changes were observed not only in genes directly related to fibrosis or proteinosis. In our model, inflammatory cytokines were also significantly increased in indium oxide-administered mice. The expression of chemokine genes, such as Ccl4 and Cxcl10 (Solomon et al., 2013), was also increased. These findings indicated that indium oxide exposure can cause inflammation as shown in previous in vitro studies (Badding et al., 2015;Jeong et al., 2015). Chil3, a gene known to be abundant in the mouse lung and increased during allergy (Hung et al., 2002;Webb et al., 2001), was significantly decreased in expression in indium oxide-administered mice. Although the fibrosis-or PAPrelated function of this gene is unknown, it may be related to mechanism of the indium lung. Expression of Ear1 and Ear2 was also significantly decreased in indium oxide-administered mice. While these two genes have previously been shown to be increased in expression during pulmonary inflammation (Cormier et al., 2002), they may not be required in the inflammatory response caused by indium exposure. The expression of oxidative stress-related genes was significantly increased in our model suggesting that indium oxide exposure induces oxidative stress, a known cause of pulmonary fibrosis (Kinnula et al., 2005;Lison et al., 2009;Liu et al., 2010Liu et al., , 2012. It is clear that inflammation and the oxidative stress response contributed to the development of fibrosis, suggesting that indium oxide administration promoted fibrosis. Moreover, there were many non-coding RNAs that displayed significantly changed expression. For example, non-coding RNA Fendrr showed a significant decrease in expression; this non-coding RNA is related to heart development (Grote et al., 2013) and pulmonary fibrosis (Sakamoto et al., 2014). These results indicated the possibility that many more non-coding RNAs including imprinted genes, such as Peg13, Meg3, and Kcnq1ot1, are involved in PAP or fibrosis. RNA-Seq analysis revealed many genes that were changed in expression that were not known as PAP-or fibrosis-related genes. Therefore, this method may facilitate the identification of new genes related to indium lung disease.

Conclusions
In summary, we have established a mouse model of indium lung disease using intratracheal administration of indium oxide. Histopathology of mouse lung tissue excised in the subacute phase revealed that PAS-positive material accumulated in the alveoli space, confirming PAP. From gene expression analysis, the data for PAP-related genes indicated that the GM-CSF pathway is intact in indium oxide-exposed mice. Fibrosisrelated genes were differentially expressed in indium oxideexposed mice compared with control mice. These data suggested that fibrosis had been initiated in early indium lung under the subacute PAP response, as predicted by human indium lung disease cases. Early detection of PAP signal will enable early treatment of disease by inhalation of indium.