Asbestos is a natural mineral fibre with physical and chemical properties that have led to its widespread use for various insulation and construction purposes. Asbestos-exposure is associated with malignancies of the lung and pleura, and tobacco smoking and asbestos-exposure are known to have a synergistic effect on lung cancer risk [1]. Although the use of asbestos is nowadays forbidden or under strict control in most developed countries, asbestos-associated malignancies continue to be a major health problem worldwide due to the long latency period and the extensive usage in the past.

Asbestos fibre characteristics such as length and chemical properties contribute to their toxicity. As macrophages attempt to engulf the fibres, reactive oxygen and nitrogen species (ROS/RNS) are produced. The iron content of asbestos further contributes to the production of these potentially damaging species. ROS/RNS account for several types of DNA and chromosomal damage including formation of mutagenic 8-OHdG adducts and DNA double strand breaks, as well as alterations in signal transduction pathways and apoptosis [2–4].

A major part of the research on asbestos-carcinogenesis is based on animal and in vitro models. These studies, including ours on human cell lines exposed to asbestos [5], have indicated numerous changes in several key pathways. The most studied signalling cascades induced by asbestos include the MAPK/ERK and NF-κB pathways [6, 7]. Activation of the MAPK cascade affects processes such as cell proliferation, apoptosis, differentiation, and inflammation [7, 8]. The apoptotic pathways can be further distorted by alterations in the normal mitochondrial functions, such as induction of cytochrome C release and caspase 9 activation, following asbestos-exposure [9]. Asbestos fibres may also directly interact with the cell cycle machinery that could affect normal cell division [10]. In spite of these observations a comprehensive view on the carcinogenic effects of asbestos fibres still remains largely unknown.

We have recently shown that specific gene copy number and gene expression changes can be detected in lung tumours of asbestos-exposed patients [11, 12]. Although several differentially expressed genes were revealed, the single gene approaches that were applied were not suitable for identification of deregulated pathways. To gain further insight into the pathways that are differentially regulated in lung tumours of asbestos-exposed and non-exposed patients we have now performed in silico pathway analysis. Differences were sought both in the normal and tumour tissue. While a single gene might not show a significant difference according to asbestos-exposure, moderate differences in a number of genes operating in the same pathway could indicate differential regulation of the whole pathway. Therefore, this approach has the potential of producing a wider mechanistic view on the asbestos-related effects.

We observed several down-regulated pathways that were related to protein ubiquitination in both normal and tumour tissues, whereas the GO terms related to ion transport dominated among the up-regulated pathways. Protein ubiquitination was chosen for further investigation owing to its pivotal role in controlling the replicative potential of a cell. Ubiquitin-activating enzyme E1 (UBA1) and ubiquitin-activating enzyme E1-like (UBA7) were analyzed further due to their roles at the early stages of protein ubiquitination and ubiquitin-like modification processes. To our knowledge, this is the first time when primary tissue samples from asbestos-exposed and non-exposed lung cancer patients have been utilized to perform a comprehensive pathway analysis.

Here we have presented, for the first time, comprehensive pathway analysis on asbestos-related lung cancer. Fresh frozen tissue samples were analyzed from well-defined patient material consisting of asbestos-exposed and non-exposed lung cancer patients. Gene expression measurements together with gene annotation data from the Gene Ontology database were utilized to identify potential asbestos-associated pathways.

The pathway analysis provides an attractive means for detecting gene expression changes that have occurred in a number of genes involved in the same biological processes or molecular functions, or located in the same cellular compartment. At the gene level, the changes in expression do not have to be significant, opposed to the single gene approaches that are commonly used to analyze high-throughput data. Moreover, as the samples in both the exposed and non-exposed groups represented different tumour types, the analysis was insensitive to the differences between the histological tumour types.

We observed a multitude of differences between the asbestos-associated and non-associated lung cancer in several key cellular events. In order to pinpoint the truly meaningful molecular changes and to reduce the chance of false findings resulting from the large number of statistical tests that are being made the significance thresholds should be stringent enough. To pinpoint the most relevant findings in our data, we performed a two-step analysis, where the significance of each category was first assessed by enrichment of differentially expressed genes in the category. The significance of the categories was evaluated by means of hypothesis testing using 50,000 permutations of the group labels. Both p-values from the t-test and fold-change between the group medians were used as measures of differential expression. Thereafter, we required that at least three components of a GO branch had to be significant according to the two measures. In addition to the challenges related to performing multiple statistical tests, it is noteworthy that the pathway analysis is based on the current knowledge of gene functions and processes occurring in the cellular environment, which has been derived using versatile methods. Therefore, future findings may unravel further differences between the asbestos-exposed and non-exposed or provide deeper insight into the differences between these two groups.

Some of the characteristics observed in the asbestos-exposed group have already been presented in the literature, indicating the utility of this approach. Interestingly, ion channel related GO terms dominated the list of up-regulated pathways. Asbestos could be linked to up-regulation of ion channels through ROS, which has been shown to trigger the opening of mitochondrial channels and mitochondrial depolarization [18]. Ion channel defects have been observed to underlie various diseases [19–21], but also suggested to have a significant role in tumourigenesis and tumour progression [[22], reviewed in [23]]. Ion channels could be an important target for further studies, also in relation to asbestos, because they contribute to virtually all basic cellular processes such as proliferation, differentiation, and apoptosis [[24], reviewed in [25]].

Among the down-regulated pathways, biological processes and molecular functions related to protein ubiquitination, a versatile process regulating various key cellular events such as DNA repair, cell cycle, and apoptosis [26], were the most intriguing findings. Furthermore, many of the other significantly altered pathways identified here are connected to ubiquitin-mediated cellular trafficking and degradation by the proteasome. These include, e.g., clathrin-coated vesicles and modification-dependent protein catabolism as well as pathways implicated in transport between Golgi and endoplasmic reticulum. Also, asbestos could have a direct interaction with cellular constituents [7] such as the nucleosome and spliceosome complexes. Of note, deregulated ubiquitination has been linked to carcinogenesis [27], especially to mesothelioma [[28, 29] (unpublished data)], a disease tightly linked to asbestos-exposure.

We chose protein ubiquitination as a target for further analysis due to its versatile role in many aspects of cellular homeostasis and, e.g., the replicative potential of a cell and as it therefore may in fact underlie several of the previously presented molecular alterations in asbestos-associated carcinogenesis. Due to limited availability of ample fresh frozen samples we chose two targets for further research. Namely, the protein levels of UBA1 and UBA7 in the tumour and respective normal tissue of asbestos-exposed and non-exposed patients were analyzed due to their role at the early stages of protein ubiquitination and ubiquitin-like modification processes. UBA1, an E1 enzyme, initiates the addition of ubiquitin signalling tags to target proteins. E1 enzyme transfers ubiquitin to E2 enzyme, which together with E3 enzyme subsequently tags the substrate [26]. Proteins involved in ubiquitin-like modification processes, such as UBA7, seem also to play a role in the ubiquitin-proteasome system [30, 31]. UBA7 strikes as an interesting candidate also because of the low expression levels detected in lung cancer cell lines and the location of the gene on 3p21, a region frequently deleted in lung carcinomas associated with tobacco smoking and asbestos-exposure [11, 32, 33].

According to our analysis of a rather small number of normal and tumour tissue samples the asbestos-related deregulation of ubiquitination seems neither to be caused by UBA1 nor UBA7. However, several other enzymes are also involved in protein ubiquitination besides UBA1 and UBA7. To identify the asbestos-associated alterations of the pathway, more holistic approaches analyzing a multitude of associated proteins in a large number of samples could be useful. Mass spectrometry analyses or immunohistochemistry performed on tissue arrays, applied on normal and tumour tissues of asbestos-exposed and non-exposed lung cancer patients, might reinforce the data presented here. Furthermore, imaging techniques on cell lines could provide additional information on the possible deregulations in ubiquitination.

We observed under-expression of UBA1 and UBA7 in SCCs, but not in ACs, both at the mRNA and protein level. This finding appears to contradict some studies, which have suggested a higher risk of AC than SCC in association with asbestos-exposure [34]. Several epidemiological studies on asbestos-related cancer have shown, however, similar risk ratios for all major histological types of lung cancer [35–37]. The results thus indicate both asbestos and cancer type specific alterations in protein ubiquitination with different components of ubiquitination pathway potentially contributing to different types of lung cancer. Future work is needed to understand the alterations connected with asbestos and histological cancer types

The pathway analysis depicts a thorough picture of the molecular changes induced by asbestos-exposure and indicates that there are distinct differences between asbestos-related and non-related lung tumourigenesis. Specifically, up-regulation of the pathways associated with ion transport and down-regulation of the functions related to protein ubiquitination were found in the normal and tumour tissue of exposed patients. The data will provide the research community important insight into the relevant targets underlying asbestos-associated lung cancer.