(302e) Molecular Initiating Events Linked to Carbon Nanotube Functionalization for Medical Applications

The widespread use of functionalized carbon nanotubes (CNTs) makes the understanding of potential harmful effects highly important. Numerous studies have been conducted in the last years on toxicological effects of CNTs. However, existing data are controversial and findings have been difficult to interpret in some cases. In the present study, two cell culture systems, human A549 pneumocytes and HaCaT keratinocytes, were used to assess the modulation of gene expression due to exposure to single and multi-walled CNTs. Moreover, CD-1 male mice were exposed to the CNTs tested by intra-tracheal instillation and lung samples were taken and analyzed after 1 day of exposure. Differentially functionalized CNTs (MW-COOH and MW-NH2) were tested in comparison with pristine multi-wall carbon nanotubes (MWCNTs) and single-wall carbon nanotubes (SWCNTs).

Cells were seeded in 96-well plates at the density of 1x 104/cm² in complete medium. After 24 h of cell attachment, plates were washed with 100 l/well of phosphate buffered saline (PBS). Cells were then exposed to suspensions (10 l) of the test materials (pristine MW-CNTs, MW-COOH, MWNH2) at concentration between 1 and 100 μg/ml, for 24 or 48 h. No fetal bovine serum was used in these preparations as it was proven to interact with nanotubes.

Each experiment was repeated independently three times. Per experiment all treatments were repeated six times per 96-well plate. Identical treated cultures were taken as replicate measure for statistical tests. Significant effects (p < 0.05) were determined by one-way analyses of variance (ANOVA) (software package SPSS Inc. 1999).

Three groups of four CD-1 male mice were exposed to all tested types of CNTs (both pristine and functionalized) by intra-tracheal installation for one day. SW-CNTs were used at concentrations between 0.1 and 100 μg/ml, while all multiple wall CNTs (both pristine MW-CNTs and functionalized MW-COOH, MW-NH₂) were used at concentrations between 1 and 100 μg/ml. After the exposure period lung samples were taken from the exposed mice and the respective modulation in gene expression was analysed in the whole genome as described below.

Gene expression profiling offers considerable potential for identifying chemical causation of effects induced in exposures to complex mixtures, and for understanding the mechanistic basis for their phenotypic effects. It entails a series of experimental steps, including RNA extraction and preparation, microarray hybridisation, chemiluminescent signal detection, image acquisition and image analysis of the microarrays, and microarray data analysis (in our case we used the Applied Biosystems Expression System software to extract assay signal and signal to noise values from the microarray images). Microarray data analysis identifies differentially expressed genes, applying different statistical filters to allow for only the statistically and biologically significant probes to be evaluated. Significant probe sets were evaluated for relevance to canonical pathway, molecular function and biological function using the Protein ANalysis THrough Evolutionary Relationships (PANTHER) Classification System (http://www.pantherdb.org). Statistically significant over- and under-represented annotation categories were determined by binomial statistics, using the observed number of expressed genes versus the numbers expected by chance within a certain annotation group. Categories with p-values > 10-2 are usually rejected. The results of microarray analysis were validated using quantitative real-time polymerase chain reaction (PCR); the latter determines the expression level of messenger RNA (mRNA) in differentially expressed genes. Transcriptomics analysis of the whole genome in the keratinocytes (HaCaT) selected as in vitro models in this study showed that gene expression significantly differs according to the duration of exposure. A large number of genes (between 1200 and 2200) changed their expression level when the cells were exposed for 6 hours to the different types of CNTs analysed, namely SWCNT, MWCNT, MW-NH₂ and MW-COOH. On the contrary, long-lasting exposure (48 hours) resulted in a significant reduction in the number of genes expression differentially compared to the control. Less than 500 genes were expressed differentially in the case of MWCNT, MW-NH2 and MW-COOH; only in the case of SWCNTs ca. 1800 genes showed differential expression after 48 hours of exposure. Similar results were found when we analyzed the data isolating the 25 genes involved in the molecular pathway associated to oxidative stress. The picture changes in terms of modulated gene expression, when focusing on the genes involved in the pathway of inflammation mediated by chemokine and cytokine signalling. SWCNTs and MWCNTs showed a reduction in the number of genes that had modulated their expression level after long-term exposure (48h); on the contrary, both types of functionalized CNTs showed practically no difference in the number of genes with modulated expression when comparing short and long-term exposure data. The observed differences in the level of gene expression as a function of exposure duration may indicate that different cellular mechanisms are involved, namely that short-term response indicates the onset of defensive mechanisms encompassing several molecular pathways and, thus, a large number of genes. The observed effects after 48 h of exposure would seem to indicate the molecular processes that underlie actual cellular toxic response to CNT exposure. The persistence of the induction of genes driving chemokine and cytokine-signalling-mediated inflammation showed that further attention would have to be paid to the inflammatory effects of the tested CNTs, and in particular the functionalized ones. For this reason, we analyzed the gene expression of the genes involved in the inflammation pathway for the two types of in vitro models we used, i.e. HaCaT keratinocytes and A549 lung epithelial cells. Comparative analysis showed that the gene expression of A549 cells was much higher than the gene expression of the keratinocytes after 48 h of exposure.

The in vitro observations outlined above indicate that the main pathway to be analyzed in the case of in vivo experiments is inflammation in the lungs mediated by chemokine/cytokine signalling. Identification of specific chemokines involved in inflammation in the lung would result in the discovery of specific biomarkers of effect that may be crucial for the early diagnosis of the toxic potency of CNTs, functionalized or not. The whole genome of lung tissue of CD-1 male mice has been analyzed after 1 day of exposure to CNTs via intra tracheal instillation (1 mg/kg bw).

Similar numbers of genes modulate their expression level after exposure to pristine SW and MW CNTs, whereas significant differences exist between the two functionalized CNTs: MW-NH₂ have a lesser effect than the non-functionalized MWCNTs; on the contrary, MW-COOH affect a much higher number of genes. Another interesting observation is that out of the 180-700 genes with significantly modulated expression only 15 are common across the four treatments.

The microarray results were validated with real-time PCR (polymerase chain reaction). Using bioinformatics analysis on the Integromics^® software we made a comparative assessment of the main genes related to chemokine signalling-mediated inflammation. Using a more stingent cut-off criterion for significance of modulation in gene expression (accepting as significant only the genes with more than 3-fold change in expression level) it was clear that MW-COOH had a very significant effect on the expression of all the genes of interest. Using the PANTHER database coupled to the significance analysis using Integromics^®we compared the induction of the inflammation mediated by chemokine and cytokine signalling pathway after exposure to MW-COOH vs. MWCNTs and after exposure to MW-NH2 vs. MWCNTs. It was clear that the inflammation pathway was significantly induced by MW-COOH, and it was less so by MW-NH₂ or pristine MWCNTs.

Analysis of the effects of different types of functionalized and pristine CNTs on gene expression showed that both the chemical family and the type of functionalization influence significantly the induction of key toxicity pathways such as inflammation mediated by chemokine and cytokine signalling. In particular, MW-COOH induced the over-expression of chemokine (C-C motif) ligand 2 (Ccl2) and ligand 19 (Ccl19). Both are cytokines involved in immunoregulatory and inflammatory processes. Ccl19 expression was induced significantly also after exposure to MW-NH2 and SWCNTs. Moreover, MW-COOH induced the over-expression of chemokine (C-X-C motif) ligand 10 (Cxcl10) and ligand 11 (Cxcl11), interferon gamma-induced proteins, involved in immunoregulation. More importantly, clinical observations have recently connected the abundant presence of these cytokines to hypersensitivity pneumonitis and, to a lesser extent, to lung fibrosis.

In summary, this study indicated that the biological effect of CNTs at the genomic level could be modified by chemical functionalization in association with changes in solubility and tendency of CNTs to form agglomerates. One of the goals of functionalizing CNTs is to increase their solubility in aqueous media, a feature that may make CNTs more compatible with physiological systems. Differential modulation in gene expression was observed as a function of single or multiple wall geometry and presence of specific functional groups. Comparison of gene expression between in vitro and in vivo exposure to CNTs revealed significant differences in the level of biological response induced towards endpoints such as oxidative stress or inflammation. In particular, MW-COOH induced excessive coding for specific proteins associated clinically with lung fibrosis and other inflammatory processes.