Mechanisms of Carbon Nanotube-Induced Pulmonary Fibrosis: A Physicochemical Characteristic Perspective

Abstract

Carbon nanotubes (CNTs) are engineered nanomaterials (ENMs) with numerous beneficial applications. However, they could pose a risk to human health from occupational or consumer exposures. Rodent models demonstrate that exposure to CNTs via inhalation, instillation, or aspiration results in pulmonary fibrosis. The severity of the fibrogenic response is determined by various physicochemical properties of the nanomaterial such as residual metal catalyst content, rigidity, length, aggregation status, or surface charge. CNTs are also increasingly functionalized post-synthesis with organic or inorganic agents to modify or enhance surface properties. The mechanisms of CNT-induced fibrosis involve oxidative stress, innate immune responses of macrophages, cytokine and growth factor production, epithelial cell injury and death, expansion of the pulmonary myofibroblast population, and consequent extracellular matrix accumulation. A comprehensive understanding of how physicochemical properties affect the fibrogenic potential of various types of CNTs should be considered in combination with genetic variability and gain or loss of function of specific genes encoding secreted cytokines, enzymes, or intracellular cell signaling molecules. Here we cover the current state of the literature on mechanisms of CNT-exposed pulmonary fibrosis in rodent models with a focus on physicochemical characteristics as principal drivers of the mechanisms leading to pulmonary fibrosis.

Graphical/Visual Abstract and Caption

Pulmonary exposure to carbon nanotubes can result in direct interaction with cells in the lung and result in an increased myofibroblast population thereby enhancing fibrogenesis, the severity of which is determined by physicochemical characteristics of the nanotubes and genetic susceptibility.

Introduction

Carbon nanotubes (CNTs) are a class of engineered nanomaterial (ENM) that comprise a major portion of the nanotechnology market. CNT production rates increase each year for incorporation into a variety of consumer products. Of interest, CNTs have unique optical, physical, and conductive properties that enhance the functionality of polymers, batteries, and electronics¹. Single-walled CNTs (SWCNTs) are a rolled graphene sheet with a diameter similar to that of a DNA double helix (1 to 4 nm), whereas multi-walled CNTs (MWCNTs) are composed of multiple concentric layers of graphene and may have a diameter typically between 10 and 100 nm. Both SWCNTs and MWCNTs may have lengths in the micrometer range. CNTs are synthesized by processes, such as chemical vapor deposition, that require high temperatures and metal catalysts to initiate the reaction process to ‘grow’ a forest of CNTs. Increased manufacturing of CNTs implicates an increasing risk of occupational exposure and development of pulmonary diseases. Exposure to CNTs occurs primarily at the manufacturing level where they are first synthesized². However, the potential for exposure also occurs during or after incorporation into consumer products, during recycling, or after disposal. Therefore, potential human exposure to CNTs throughout their life cycle may be of concern. Experimental evidence in rodents shows that inhalation or aspiration exposure to CNTs causes pulmonary fibrosis, a disease characterized by excessive deposition of collagen and progressive lung tissue scarring^3–5. Pulmonary fibrosis is defined by the American Thoracic Society as the production and deposition of collagen in the lung resulting in the buildup of scar tissue, thereby reducing the exchange of oxygen and carbon dioxide between the alveolar airspace and pulmonary capillaries⁶.

From a historical perspective, it is well-established that inhalation of specific types of particles and fibers such as silica, metals, coal dust, or asbestos leads to the development of pulmonary fibrosis^7–10. Thus, there is already a fundamental understanding of some of the cellular and molecular mechanisms of pulmonary fibrosis caused by particle or fiber exposure. However, due to the unique physicochemical characteristics of CNTs, including but not limited to their nanoscale dimensions, these ENMs may interact with the intracellular microenvironment or extracellular matrix to mediate fibrotic reactions in the lungs or other tissues through novel mechanisms that remain to be elucidated. As CNTs are relatively new in terms of their emergence into society, there is no comprehensive epidemiologic data to convincingly support the conclusion that all types of CNTs will cause pulmonary fibrosis, as human exposures thus far have been limited, and the development of pulmonary fibrosis in humans may take decades to manifest respiratory symptoms after the initial exposure due to the long latency period of the disease. Understanding specific CNT physicochemical characteristics that are important for initiating and perpetuating lung injury will be important towards determining the relative risk for pulmonary fibrosis. Deposition of an inhaled ENMs in the lung depends on many factors, one of which is size. Particles on the nanomaterial scale are able to be inhaled deep into the lung, reach the alveolar region, and interact directly with alveolar macrophages and epithelial cells^11,12. Because of their small size, CNTs also have the potential to be transported from the lungs into the systemic circulation or lymphatic system and reach organs such as the liver, kidney, heart, brain, and the thymus^11,13. With increasing production and use of CNTs, it is imperative that we understand mechanisms of CNT-induced pulmonary fibrosis in order to evaluate a wide spectrum of different CNT types for relative risk in order to prevent a future respiratory disease. The aim of this review is to overview the evidence in experimental animal models and cultured cells demonstrating that CNTs cause pulmonary fibrosis, and to discuss mechanisms of susceptibility to CNT-induced fibrosis with an assessment of the growing diversity of CNT physicochemical characteristics.

Exposure Methodology as a Determinant of CNT-Induced Pulmonary Fibrosis

Fibrosis has been documented in the lungs of rodents after exposure to CNTs delivered by several methods, including inhalation, instillation, or oropharyngeal aspiration. Inhalation exposure is ideal as it represents a more realistic exposure in terms of deposition patterns in the lung that would occur in occupational settings. The deposition of inhaled CNTs is determined by several factors including size, shape, electrostatic charge, and aggregation state. Inhalation exposure to well-dispersed MWCNTs results in deposition into the distal regions; i.e., alveolar duct bifurcations and alveolar epithelial surfaces of the lungs of mice or rats^3,14. Less aggregated or more dispersed nanotubes are more biologically available for macrophage uptake and clearance from the lung. Inhalation of dry or aerosolized CNTs in surfactant-containing media cause diffuse interstitial fibrotic lesions within the alveolar and subpleural regions of the lung. Exposure by instillation or oropharyngeal aspiration, which involves a bolus delivery of CNTs suspended in aqueous media, can also result in deposition at distal sites in the lungs if the nanoparticles are well-dispersed but typically stimulate focal granuloma formation. Furthermore, experiments conducted in multiple laboratories at different institutions using harmonized methods and identical sources of CNTs have demonstrated interlaboratory reproducibility in deposition patterns and pro-inflammatory responses in the lungs of mice and rats¹⁵. Many studies have reported granulomas in the lungs of rodents resulting from agglomerated SWCNTs or MWCNTs lodged within small airways, but this is not observed in most inhalation studies probably due to better dispersion of the respirable fraction of inhaled CNTs, which primarily includes singlet CNTs or small agglomerates of CNTs. Methods for dispersing CNTs in an aqueous suspension using surfactant-containing media prior to instillation or aspiration in rats or mice have greatly improved, making this route of exposure generally acceptable¹⁶. Moreover, long term in vivo studies demonstrate that while CNTs clear from the lung to some extent via the mucociliary escalator or pulmonary lymphatics, longer nanotubes remain in the lung tissue of rodents over time due to their biopersistent nature^17–19. Rodent exposure methods (e.g., inhalation vs aspiration) should be taken into consideration when comparing the relative fibrotic effects of CNTs.

Physicochemical Properties of CNTs as Determinants of Fibrosis

Physicochemical properties of CNTs are key in determining their reactive potential; the size, charge, length, rigidity, residual metal catalyst content, and surface functionalization can each greatly impact the toxicity and fibrogenic potential of these materials.

Residual Metal Catalyst Content

A variety of metals are used as catalysts in the manufacturing of CNTs by chemical vapor deposition (CVD). For instance, cobalt is used as a catalyst in the synthesis of MW- and SW-CNTs, while nickel or iron have been used as catalysts in the synthesis of MWCNTs. These same metals are known to mediate pulmonary fibrosis in humans in occupational settings²⁰. For example, nickel is known to cause occupational asthma and contact dermatitis, whereas iron and cobalt cause interstitial pulmonary fibrosis in occupations related to mining and metallurgy. Metal catalysts used in the CVD manufacturing process become integrated into the carbon structure of nanotubes and mediate at least some of the proinflammatory effects seen after exposure to MWCNTs in rodents. For example, activation of macrophage inflammasomes and subsequent interleukin- (IL-) 1β production induced by MWCNT-exposure is due at least in part to residual nickel contamination²¹. Acid washing of MWCNTs removes some, but not all, of residual nickel and interlaboratory comparisons of the pro-inflammatory effects of acid-washed MWCNTs versus pristine MWCNTs show that neutrophilic inflammation in the lungs of mice is reduced by partial removal of nickel^22,23. The biological availability of trace metals on or in CNTs limits their interactions in the lung and resulting pathology in vivo. Residual metal content has the potential to drive CNT pulmonary toxicity and should be a consideration in the design of CNTs.

Length

CNTs can have a length that greatly exceeds their nanoscale diameter (10–100 nM) and this high length to width aspect ratio is an important determinant of their toxic potential. Only a few studies have addressed variations in length as a determinant of fibrosis. One study comparing two MWCNTs, a long (5–15μm) and a short (350–700nm) MWCNT, found pulmonary fibrosis to be length-dependent. In this study, mice treated with long MWCNTs via instillation resulted in a significant increase in expression of profibrotic mediator transforming growth factor (TGF) -β1 and collagen deposition in the lungs compared to those treated with the shorter MWCNTs²⁴. Another study examined the effects of a long (~12μm) and a short (~1μm) SWCNT on fibroblasts in vitro and found the longer SWCNTs to induce greater reactive oxygen species, collagen, and TGF-β¹⁸. These data were also validated in vivo by oropharyngeal aspiration exposure of mice to the long or short SWCNTs; while both treatments resulted in collagen deposition in the lungs of these mice, greater fibrosis was measured in the lungs from the mice treated with the longer SWCNTs¹⁸. Furthermore, instillation of long (20–50μm) MWCNTs in rat lungs resulted in macrophage activation and profibrotic mediator (TGF-β1) production as well as greater fibroblast proliferation, collagen production, and granuloma formations compared to rats treated with short (0.5–2μm) MWCNTs²⁵. Interestingly, direct instillation of two types of long or two types of short CNTs into the parietal pleura resulted in greater inflammation and fibrogenesis from both the long CNTs and greater clearance of the shorter CNTs²⁶. Length is clearly a property of CNTs that plays a key role in determining the exposure risk of CNT pulmonary fibrogenesis and toxicity.

Rigidity

Toxicity and fibrogenic potential of CNTs can also be derived from how long, rigid or tortuous they are versus how short, flexible and pliable they are. MWCNTs from different manufacturing sources possess different degrees of rigidity even though they may have similar width and length; some are ‘curly’ whereas others are straight. This comparison is reminiscent of the comparison between asbestos fiber types; chrysotile asbestos is a curly fiber whereas crocidolite asbestos is a more toxic straight rigid fiber. A useful metric for assessing rigidity that has been adopted by the International Standards Organization is the static bending persistence length (SBPL) and bending ratio, which are derived by measuring convolutions within the nanotube structure from transmission electron microscope (TEM) images²⁷. Thick straight MWCNTs delivered to the lungs of female mice via intratracheal instillation cause similar inflammatory responses yet result in more severe pulmonary fibrosis and interstitial pneumonia compared to thinner, more curled MWCNTs²⁸. Long, rod-like MWCNTs disrupt macrophage function because their length makes them more resistant to compaction within the cell resulting in CNT protrusion from the macrophage causing frustrated phagocytosis, the disruption of cell membranes, and cell death^29,30. Cell membrane disruption also causes leakage of cellular constituents (e.g., reactive oxygen species, enzymes, cytokines) that can cause injury to surrounding cells and tissues. Rod-like MWCNTs also result in mucous cell metaplasia alone while tangled MWCNTs do not^31–34. While CNT length is an important determinant of clearance rate from the lungs, length alone does not necessarily determine CNT persistence^25,35. For example, long SWCNTs or tangled MWCNTs can be compacted and contained within phagolysosomes after uptake by macrophages without causing frustrated phagocytosis (Fig. 1). Rigid, rod-like, MWCNTs disrupt macrophage function if the nanotube length exceeds the width of the engulfing phagocyte, however short, rigid MWCNTs are capable of being taken up and cleared by macrophages without causing frustrated phagocytosis. Therefore, both rigidity and length are important determinants when considering the fibrogenic potential of these ENMs.

Figure 1.

Macrophage host responses to carbon nanotubes (CNTs) differs based on selected physicochemical properties. Longer, tangled or short, rod-like CNTs are able to be engulfed and contained within a phagolysosome while long, rod-like CNTs result in frustrated phagocytosis, ROS production, and activation of signaling pathways resulting in the production of profibrogenic cytokines.

Surface Functionalization

Functionalization refers to surface modification (e.g. coatings or side-chain additions) of CNTs and is important in enhancing specific properties of CNTs. These modified nanomaterials can be useful as polymer composites, sensors, and for biomedical applications. Functionalization of CNTs can be accomplished via two predominant methods; atomic layer deposition (ALD) or molecular layer deposition (MLD), however others are being developed or optimized³⁶. CNTs have been functionalized with carboxyl (R-COOH), amine (R-NH3), polysaccharide hyaluronic acid groups, metal oxides (aluminum oxide, zinc oxide, titanium dioxide) and silver (summarized in Table 1). The nature of the coating can influence the biocorona, which refers to proteins, lipids, and other biomolecules that form around the nanomaterial³⁷. CNT coatings alter profibrotic mediators released from THP-1 monocytes (IL-1β) and epithelial BEAS-2B cells (TGF-β1, platelet derived growth factor (PDGF)-AA) in vitro, suggesting that these coatings will similarly modify cytokine release and pulmonary fibrosis in vivo³⁸. Nearly all the published work on CNTs have used ‘pristine’ or unmodified CNTs. These studies are relevant to human occupational exposures where individuals will be exposed to CNTs directly after synthesis. However, most CNTs will likely undergo some sort of post-synthesis functionalization to modify or enhance their unique properties. This means that most consumer and environmental exposures will be related to some type of functionalized CNTs. Surface functionalization can also influence agglomeration status of CNTs. For example, agglomeration of CNTs due to electrostatic forces presents a problem for the purposes of many engineering designs and a variety of dispersal agents (e.g., surfactants) have been employed to improve dispersion by reducing electrostatic attraction^16,39. More dispersed CNTs delivered to mice cause a more severe chronic interstitial fibrosis in the lungs along with elevated levels of growth factors (PDGF and TGF-β1) that play important roles in the promotion of fibrogenesis⁴⁰. However, some functionalization’s can reduce the fibrogenic response of CNTs, even though they remain dispersed. For example, carboxylated (COOH)-CNTs cause similar rat lung inflammation and fibrosis^41,42 but reduced mouse neutrophilic lung inflammation as compared to unmodified CNTs¹⁵, and cause less pulmonary fibrosis^38,43. Likewise, aluminum oxide (Al₂O₃) ALD coating of MWCNTs results in decreased fibrosis compared to uncoated MWCNTs in vivo⁴⁴. MWCNT coating with polyethylene glycol (PEG) also results in decreased fibrotic mediators IL-1β, TGF-β1, and PDGF-AA compared to pristine exposure³⁸ furthermore, hyaluronic acid coating results in decreased airway fibrosis and inflammation⁴⁵. In contrast, polyetherimide (PEI) functionalized CNTs results in increased collagen deposition and fibrosis in the lungs of mice³⁸. Zinc oxide (ZnO) ALD coating of MWCNTs also results in acute systemic inflammation in mouse lungs in vivo and proinflammatory cytokine (e.g., IL-6) production in THP-1 monocytes in vitro. However, ZnO-MWCNT exposure in the lung in vivo does not result in an altered fibrotic response compared to uncoated MWCNTs⁴⁶. Therefore, some functionalization processes reduce CNT-induced toxicity and lung injury while others increase toxicity and the potential for fibrogenesis. A comparison of different types of functionalized MWCNTs and their relative fibrotic potential is shown in Table 1. Thus, to best assess human health effects, consideration should be given to both acute and chronic effects of functionalized CNTs. Consumer exposure to specific types of functionalized CNTs in products is largely unknown based on this information being classified ‘proprietary’ by CNT manufacturers and their clients. An increasing variety of functionalization’s are employed to modify specific characteristics of CNTs, which makes assessing the toxicity each type of functionalized CNTs in rodents impractical. In response to this growing challenge, high through-put in vitro screening techniques will be necessary to test the vast majority of engineered nanomaterials for toxic signatures⁴⁷.

Table 1.

Summary of fibrogenic effects of CNT surface functionalizations in rodent models.

MWCNT Functionalization	Pathological Outcome Compared to Pristine MWCNT	Rodent/Strain	Dose	Exposure Time	Exposure Method	Reference
-COOH	Reduced neutrophilia	Mouse:C57BL/6 Rat:F344, Sprague-Dawley	Mouse: 10, 20, 40 μg Rat: 10, 50, 200 μg	21 days	Mouse: oropharyngeal aspiration Rat: intratracheal instillation	Bonner et al 2013 [15]
	Decreased TGF-β1 and PDGF-AA	Mouse: C57BL/6	50 μg	21 days	Oropharyngeal aspiration	Wang et al 2014 [43]
	No change in TGF-β1 or IL-6	Rat: Sprague-Dawley	1 mg/kg	16 days	Intratracheal instillation	Roda et al 2011 [41]
	No change in IL-6, TGF-β1, or collagen	Rat: Sprague-Dawley	1 mg/kg	16 days	Intratracheal instillation	Coccini et al 2013 [42]
	Decreased PDGF-AA, TGF-β1, and collagen	Mouse:C57BL/6	2 mg/kg	21 days	Oropharyngeal instillation	Li et al 2013 [38]
-NH3	Slightly increased TGF-β1 and collagen	Rat: Sprague-Dawley	1 mg/kg	16 days	Intratracheal instillation	Roda et al 2011 [41]
	Increased IL-6 and TGF-β1	Rat: Sprague-Dawley	1 mg/kg	16 days	Intratracheal instillation	Coccini et al 2013 [42]
	Increased TGF-β1, but similar IL-1β and PDGF-AA	Mouse:C57BL/6	2 mg/kg	21 days	Oropharyngeal instillation	Li et al 2013 [38]
-PEG	No change in PDGF-AA, TGF-β1, or collagen	Mouse:C57BL/6	2 mg/kg	21 days	Oropharyngeal instillation	Li et al 2013 [38]
-PEI	Increased TGF-β1 and collagen	Mouse:C57BL/6	2 mg/kg	21 days	Oropharyngeal instillation	Li et al 2013 [38]
-Hyaluronic acid	Decreased OPN, TNF-α, collagen, and airway fibrosis	Mouse:C57BL/6J	1.5 mg/kg	21 days	Oropharyngeal aspriation	Hussain et al 2016 [45]
-Al2O3	Reduced fibrosis	Mouse:C57BL/6	4 mg/kg	28 days	Oropharyngeal aspiration	Taylor et al 2015 [44]
-ZnO	No change in fibrosis	Mouse:C57BL/6	10 mg/kg	28 days	Oropharyngeal aspiration	Dandley et al 2016 [46]

Mechanisms of CNT-Induced Fibrosis

It is well established that rodents exposed to CNTs via inhalation, instillation, or aspiration develop pulmonary fibrosis^{12,39,48–50}. CNTs initiate an early inflammatory response in the lung characterized by increased cytokine and chemokine production by resident lung cells (e.g., macrophages, epithelial cells) followed by neutrophil recruitment and infiltration. The persistence of CNTs in the lung leads to chronic activation of pulmonary cells to produce pro-fibrogenic growth factors (e.g., TGF-β1, PDGF) and increased production of extracellular matrix (ECM) proteins (e.g., collagens, fibronectin) that lead to progressive fibrogenesis. The mechanism of CNT-induced fibrogenesis is driven largely by the production of reactive oxygen species (ROS) which serve to activate intracellular signaling pathways that favor the increased production of pro-fibrogenic growth factors and their signaling pathways (Fig. 1). The resultant increase in pro-fibrogenic cytokines also expands the resident fibroblast population by recruiting circulating fibrocytes which differentiate into myofibroblasts in the lung, and by stimulating the differentiation of resident lung epithelial or fibroblasts into myofibroblasts.

Oxidative Stress as an Initiator of CNT-Induced Fibrosis

Oxidative stress is a major driver of inflammation and fibrosis. CNTs have the potential to generate reactive oxygen species (ROS) directly in the absence of cells, possibly due the presence of residual metal catalysts (e.g., Fe, Co, Ni). Alternatively, ROS can be generated by lung cells stimulated with CNTs via activation of mitochondrial nicotinamide adenine dinucleotide phosphate (NADPH) oxidases, as shown in vitro with macrophages, fibroblasts, and alveolar epithelial cells^51–54. Treatment of RAW264.7 macrophages with SWCNTs results in the induction of inflammatory cytokine production (TNFα, IL-1β, IL-6), NADPH oxidase activation, and NF-κB activation⁵². Interestingly, when comparing SW- versus MW-CNTs, ROS production in vitro activates the same pathways however, SWCNTs are more acutely toxic compared to MWCNTs due to enhanced ROS generation⁵². Alveolar macrophages in particular are notorious for releasing a ‘respiratory burst’ of ROS after activation with particles, fibers, or bacterial products. Generated ROS can then activate the inflammasome and induce the release of the pro-inflammatory IL-1β, a leukocyte pyrogen that is secreted after pro-IL-1β is cleaved by caspase⁵⁵. Elevated ROS also causes the release of profibrotic cytokines like PDGF and TGF-β1⁵⁶. Systemic oxidative markers of stress are increased after MWCNT treatment of mouse lungs showing an increase in DNA 8-hydroxy-2′-deoxyguanosine adducts in the urine as well as increased lactate dehydrogenase, tumor necrosis factor-α, IL-1β, mucin, and surfactant protein -D in the lavage fluid after one-day of exposure; all levels decrease after a week except mucin and surfactant protein-D⁵⁷. The initial burst of CNT-induced oxidative stress causes cellular damage and is pivotal in perturbing inflammation and the fibrogenic potential of these materials. A decrease in ROS, achieved via antioxidant N-acetyl cysteine (NAC) treatment, results in suppressed fibrosis in the lungs of MWCNT-treated mice⁵⁵. NADPH oxidase, a membrane associated enzyme which produces ROS when internal metabolism of CNTs occurs, adds to intracellular oxidative stress. The removal of a critical subunit of NADPH oxidase via induced knockout reduced ROS generation and resulted in decreased bleomycin-induced fibrosis as measured by a hydroxyproline assay for collagen deposition in mouse lungs compared to wild type counterparts⁵⁸. Suppression of ROS by nuclear factor erythroid 2-related factor (Nrf2) suppressed ROS, inflammation, and fibrosis in the lungs of mice treated with MWCNTs as Nrf2 knockout mice exhibit increased basal collagen deposition and immune cell infiltration in the lung⁵⁹. Oropharyngeal aspiration of MWCNTs into mouse lungs resulted in increased Nrf2 expression⁵⁹. ROS can also be generated by macrophages as a result of frustrated phagocytosis, where excess ROS from the phagolysosome is released into the lumen and surrounding tissue causing inflammation and promoting fibrosis (Fig. 1)⁶⁰. Elevated levels of ROS can trigger redox sensitive switches within a cell to activate redox responsive cellular signaling (i.e. NFκB, ERK). Initial ROS generation from CNT-cell interactions initiates the inflammatory response and drives the beginning steps of fibrosis.

Inflammasome Activation

Increasing evidence demonstrates that CNTs and other fiber-like materials (e.g., asbestos, silica), cause an inflammatory response via activation of the macrophage inflammasome. Inflammasomes are intracellular protein scaffolds that incorporate activated caspase-1 to cleave pro-IL-1β to a mature, secreted form of IL-1β. IL-1β has a variety of key function in inflammation, including the recruitment of neutrophils to sites of lung injury. CNTs have been reported to stimulate inflammasome activation in macrophages^55,61,62. Inflammasomes likely play an important role in host defense to pathogens and inhaled nanomaterials (including CNTs), but have also been implicated in a variety of disease states⁶³. Inflammasome activation is a two-step process where step 1 involves induction of pro-IL-1β by stimulation of toll-like receptors (e.g., TLR4 activated by LPS). It has also been suggested that high-mobility group box 1 (HMGB1) can serve to initiate step 1 of the inflammasome mechanism⁶⁴. Step 2 involves organization of the inflammasome scaffold and cleavage of pro-IL-1β to mature IL-1β and is initiated by fiber-like agents such as asbestos, silica, and CNTs. MWCNTs stimulate inflammasome activation (i.e., step 2) through lysosomal disruption and ROS production⁵⁵. Activation of the inflammasome has been proposed as a mechanism of CNT-induced pulmonary fibrosis. Inflammasome activation and IL-1β release is clearly important for recruiting neutrophils to the lung to participate in microbial killing⁶⁵. Moreover, inflammasome activation occurs primarily in classically activated macrophages (CAMs) that function in microbial killing. In contrast, “alternatively activated macrophages” (AAMs) are the predominate phenotype in fibrosis and are polarized by Th2 cytokines, such as IL-4 and IL-13⁶⁶. Interestingly, inflammasome activation and IL-1β production is suppressed by IL-4 and IL-13 in human THP-1 macrophages in vitro and in the lungs of mice sensitized with house dust mite allergen prior to MWCNT exposure by oropharyngeal aspiration⁶⁷. The mechanism of inflammasome suppression by these Th2 cytokines is through a STAT6-dependent decrease in pro-caspase-1, the precursor to caspase-1 which serves as the key inflammasome component to cleave pro-IL-1β to mature IL-1β⁶⁷. This study also showed that MWCNT exposure exacerbated allergen-induced airway fibrosis and yet reduced IL-1β and neutrophils in the lung, suggesting that inflammasome activation was not a mechanism of airway fibrosis in this model system. Neutrophilia is a common response to MWCNT exposure in the lung. An important study showed that IL-1 receptor knockout (IL-1R KO) mice do not display neutrophilia and yet develop pulmonary fibrosis to a greater degree than wild type mice⁶⁸. Recently, inflammasome activation by MWCNTs in human airway epithelial cells in vitro was reported as a possible mechanism of driving pro-fibrogenic responses in fibroblasts⁶⁹. The role of inflammasomes and IL-1β in CNT-induced fibrosis remains controversial and whether IL-1β is pro-fibrogenic or anti-fibrogenic may depend on the temporal expression in the lung, which in turn could determine the duration of neutrophilic inflammation in the lung.

Box 1. Epidemiology of CNT Exposure.

Products of the emerging nanotechnology industry are relatively new and therefore epidemiologic information related to human health outcomes from CNT exposure is relatively scarce. CNT manufacturing began early in the 2000’s and only a handful of epidemiology studies have been published on CNT exposure and health risks. Engineered nanomaterials (ENMs) include a diversity of types that may produce different pathological outcomes based on their physicochemical properties. Often employees manufacture products that contain more than one type of ENM (e.g., CNT plus nanometal catalyst) resulting in exposure to a mixture of ENMs. Currently, published epidemiology studies have included relatively low numbers of participants, lack of data on duration of worker employment/exposure, and only collected data from companies who gave permission^70–72. So far, epidemiology studies have not adequately addressed which physicochemical characteristics of CNTs cause human health effects. The similarities to asbestos make CNTs a concern for human health. Epidemiology studies test individuals currently working at a CNT factory but have not followed individuals who have changed occupations. Most importantly, asbestos related lung diseases have a latency of 30–40 years, suggesting that CNT-induced pulmonary diseases could take decades before we begin to see symptom presentation⁷³. A few studies have been successful in identification of potential biomarkers of human lung inflammation and fibrosis; these offer important opportunities to follow up on in the future^74,75.

Canonical TGF-β1 Pathway

A major mechanism of collagen deposition in the lung involves production, activation and cell signaling via TGF-β1. CNTs stimulate pulmonary cells (e.g., epithelial, macrophage, or fibroblast) to produce latent TGF-β1 which can be sequestered in the ECM by thrombospondin 1 or can be cleaved to an active form via proteolysis^76–79. Active TGF-β1 then binds to a transmembrane tetramer receptor consisting of two-type I and two-type II receptors on mesenchymal cells such as fibroblast or myofibroblasts. These TGF-β1 ligand receptors activate the Smad anchor for receptor activation (SARA) to recruit transcription factors Smad2 and/or Smad3. The type I receptor is a serine/threonine kinase which phosphorylates and activates Smad2/3. A trimer of two phosphorylated Smad2/3 molecules and co-activator Smad4 then translocate into the nucleus where they recognize the Smad response element (SRE) and activate transcription of ECM mRNAs (col1a1, col1a2) which eventually become translated collagen proteins that can be secreted by the cell. Negative feedback occurs as the SRE also activates transcription and expression of Smad7, a repressor of Smad2/3 signaling. The activity of the phosphorylated Smad complex is primarily abolished by phosphatase activity or to a lesser extent ubiquitination of the complex⁸⁰. TGF-β1 signaling inhibits proliferation of most cell types, while also functioning to initiate the differentiation of fibroblasts into myofibroblasts⁸¹. While TGF-β1 is pro-fibrogenic, it also has beneficial immunoregulatory properties by suppressing excessive inflammation; TGF-β1 knockout mice have a short survival time and ultimately succumb to systemic inflammation⁸². TGF-β1 is increased in the lung lavage fluid of mice and rats in vivo several weeks after exposure to SWCNTs or MWCNTs by aspiration^48,83–85. Both SWCNTs and MWCNTs also induce TGF-β1 production from RAW264.7 macrophages and BEAS-2B lung epithelial cells in vitro^{25,51,83,84,86}. Physicochemical characteristics of CNTs, like length, can determine the degree of TGF-β1 induction. For instance, long but not short MWCNTs enhance TGF-β1 and phospho-Smad2 as measured by immunohistochemistry of the lungs of mice exposed via intratracheal instillation²⁴. It has also been demonstrated that the TGF-β1/Smad pathway is necessary for collagen production in mice in vivo and in fibroblasts in vitro after exposure to CNTs^25,48,87. The mechanism for initiating increased TGF-β1 is not yet elucidated but is most likely a response to cellular oxidative damage and stress. While TGF-β1 appears to play a central role in fibrosis, it is unclear whether it will be a useful biomarker of exposure in humans. For example, workers exposed occupationally to MWCNTs express similar levels of TGF-β1 in sputum or serum as compared to unexposed control individuals, while sputum levels of other pro-inflammatory or pro-fibrotic cytokines (e.g., IL-1β, IL-4, IL-5, IL-6, TNF-α) were significantly increased compared to unexposed workers⁷⁵. Other epidemiology studies observe increased C-C motif ligand 20, basic fibroblast growth factor and IL-1 receptor II⁷⁴.Taken together, there is strong evidence to support a role for TGF-β1 driven collagen synthesis as a key mechanism of CNT-induced fibrogenesis, yet it is unclear whether TGF-β1 will be a useful early biomarker of fibrosis to monitor occupational exposure to CNTs.

Epithelial to Mesenchymal Cell Transition

Pulmonary fibrosis is the result of a disruption in the homeostatic balance of epithelial and mesenchymal cell survival in the lung⁴⁹. The process of epithelial to mesenchymal cell transition (EMT) occurs when epithelial cells are stimulated to undergo differentiation to a myofibroblast phenotype (Fig. 2)⁸⁸. During this process epithelial cells lose their adhesion strength and polarity while gaining invasive and migratory properties⁸⁹. EMT induced by MWCNTs is can be mediated via TGF-β1 stimulation and the resultant immediate- but transient- activation of Smad2^24,90. Activation of Smad2 by long, but not short, MWCNTs results in an increase of collagen I and III 30 days after intratracheal instillation exposure in rats²⁵. Wang et al. also demonstrates that long MWCNTs interact directly with epithelial cells in vivo and in vitro to activate the TGF-β/Smad2 signaling pathway, resulting in alveolar type II epithelial cell (RLE-6TN) loss of E-cadherin, an epithelial cell specific adhesion molecule, and a gain of fibronectin expression thereby inducing EMT⁹⁰. Likewise, these MWCNTs are shown to directly interact with fibroblasts in vivo and in vitro and enhance expression of fibroblast-to-myofibroblast specific marker expression of fibroblast specific protein (FSP-1), α-smooth muscle actin (α-SMA), and collagen III in 3T3-L1 fibroblasts⁹⁰. Alternatively, human bronchial epithelial cells under TGF-β1-induced EMT are initiated by rod-like MWCNTs through Smad-independent activation of the AKT/GSK-3β/SNAIL signaling pathway⁹¹. The GSK/SNAIL pathway is an established pathway that regulates the process of EMT⁹². In other studies these same rod-like MWCNTs have been observed to directly promote EMT²⁵. In C57BL6 female mice, SWCNT exposure results in epithelial-derived fibroblasts composing almost half of the fibroblast population in the lung, demonstrating that epithelial-derived fibroblasts contribute significantly to CNT-induced pulmonary fibrosis⁵⁰. In vitro studies of human epithelial cells treated with low doses of MWCNTs demonstrate an altered morphology of epithelial cells towards a mesenchymal cell phenotype⁹³. Analysis of epithelial or mesenchymal specific markers E-cadherin, vimentin, α-SMA, and fibronectin protein expression can clarify the extent of differentiation of the epithelial-derived fibroblasts in EMT in human bronchoalveolar cells⁹³. CNT-induced pulmonary fibrosis is driven by an increase in the myofibroblast population, some of which can be derived from epithelial cells through the process of EMT.

Figure 2.

Cell signaling in the lung after carbon nanotube (CNT) exposure resulting in expansion of the myofibroblast population through three possible mechanisms: fibroblast-to-myofibroblast differentiation, epithelial-to-mesenchymal transition (EMT), and recruitment and differentiation of circulating fibrocytes.

Expansion of the Resident Lung Myofibroblast Population

In addition to EMT, the resident lung fibroblast population can be amplified via growth factor-induced proliferation (i.e., hyperplasia) by stimulation of fibroblast differentiation into myofibroblasts (Fig. 2). This can occur in fibroblasts in vitro directly in the presence of TGF-β1²⁵. PDGFs are an important family of growth factors that also drive fibroblast proliferation⁹⁴. This family of glycoproteins are composed of two chains of PDGF-AA, -BB, -AB, -CC, and -DD. PDGFs act as both mitogens and chemoattractants for fibroblasts. Transgenic overexpression of the PDGF-B gene in rat lungs causes increased fibroblast proliferation and collagen deposition⁹⁵. Exposure of rats or mice to MWCNTs by intratracheal instillation or oropharyngeal aspiration, respectively, increases PDGF-AA in the bronchoalveolar lavage fluid as well as in bronchiolar epithelial cells and macrophages as determined by immunohistochemical staining^5,48,96. Priming of fibroblasts in vitro with low doses of growth factors TGF-β1, PDGF, or epithelial growth factor (EGF) and subsequent treatment of the fibroblasts with MWCNTs promotes proliferation through prolonged ERK1/2 signaling⁵⁶. Additionally, this study also showed that the ability of several different types of MWCNTs to stimulate proliferation was correlated with prolonged ERK1/2 signaling specifically in fibroblasts⁵⁶. Furthermore, rat pleural mesothelial cells treated in vitro with MWCNTs or nickel nanoparticles, a residual catalyst present in some MWCNTs, caused prolonged PDGF-induced ERK1/2 signaling and synergistically enhanced PDGF-induced chemokine production⁹⁷. Mesothelial cell signaling of fibroblasts via chemokines or growth factors could be a mechanism of subpleural fibrosis observed in mice exposed to MWCNTs by inhalation⁵. Treatment of mouse RAW264.7 macrophages with MWCNTs results in the production of ROS, inflammatory cytokines (IL-1β, IL-10, IL-6), and profibrogenic growth factors (PDGF and TGF-β1) that collectively promote the proliferation and transformation of lung fibroblasts-to-myofibroblasts through paracrine signaling⁵¹. Taken together, these data indicate that MWCNTs increase growth factor production (TGF-β1 and PDGF) and enhance growth factor-induced cell signaling via ROS generation, resulting in expansion of the resident lung myofibroblast population.

Circulating Fibrocyte Recruitment

Fibrocytes are mesenchymal progenitor cells derived from the bone marrow that migrate towards sites of fibrosis and can play an active role in the development of fibrosis (Fig. 2). Fibrocytes express cell-surface markers related to leukocyte progenitor cells, and fibrocytes can differentiate into fibroblasts, myofibroblasts, or adipocytes. They have been demonstrated in both mouse and rat models to migrate from the bone marrow to sites of lung injury following bleomycin exposure through a mechanism involving the CXCR4 receptor and the release of its chemokine ligand CXCL12, as well as release of PDGF from the lung^98,99. Fibrocytes are identified by unique cell surface markers like hematopoietic stem cell marker CD34, leukocyte marker CD45, mesenchymal marker collagen I (COLI), and chemokine receptor (CCR7)¹⁰⁰. Regulation of the CCR7 signaling pathway decreases differentiation and migration of human circulating fibrocytes¹⁰¹. The number of circulating fibrocytes is positively correlated with pulmonary inflammation, collagen content, and severity of fibrosis¹⁰². Furthermore, patients with interstitial lung diseases have higher levels of circulating fibrocytes, decreased lung function, and interstitial pneumonitis associated with collagen vascular disease compared to patients who do not experience these diseases¹⁰³. It is clear that circulating fibrocytes contribute to pulmonary fibrosis, but currently there is no research on their migration and differentiation after MWCNT exposure. The current state of research suggests circulating fibrocytes contribute to pulmonary fibrosis due to induction of lung PDGF production, and it is known that MWCNTs induce PDGF in the lung. However, further studies need to be conducted to better understand the role of fibrocytes in the development of pulmonary fibrosis from CNT exposures.

Alternative Macrophage Activation

Pulmonary exposure to CNTs can result in a Th1 or a Th2 driven immune microenvironment. A Th1 driven response conventionally is stimulated by a viral or bacterial infection polarizing macrophages towards a classically activated M1 function, compared to an alternatively activated M2 macrophage which is polarized under the influence of a Th2 microenvironment. A Th1 driven M1 macrophage drives the inflammatory response by producing INF-γ, IL-12, IL-6, and other pro-inflammatory cytokines. MWCNTs that are more flexible or tangled promote a Th1 response in the lungs of mice exposed via oropharyngeal aspiration marked by neutrophilia³³. MWCNTs that have been functionalized with a carboxyl group have the ability to activate phospholipase C (PLC) and recruit macrophages through the PLC/IP3/Calcium release-activate calcium channel signaling pathway¹⁰⁴. M2 macrophages are classically driven by helminth infection and TGF-β1 stimulation. M2 macrophages also produce more TGF-β1 in response to stimuli. Interestingly, a rod-like MWCNT treatment via oropharyngeal aspiration results in a more Th2 response as indicated via mucous cell metaplasia and/or eosinophilia^31,34. M2 macrophages produce Th2 cytokines, like IL-4, IL-5, and IL-13 which drive the adaptive immune response and are upregulated in bleomycin-induced fibrosis¹⁰⁵. IL-17 is also upregulated in bleomycin treated lungs which can down-regulate the Th1 response and drive production of growth factors, like TGF-β1, and result in excess collagen deposition into the ECM and decreased collagen metabolism thereby disrupting homeostasis¹⁰⁵. Therefore, the microenvironment cultivated by CNT exposure can drive differential immune responses based on the physicochemical properties of the CNTs.

Genetic Susceptibility to Pulmonary Fibrosis

There are multiple genes that determine susceptibility to pulmonary fibrosis. Variability in rodent genetic strain and sex differences are two very important susceptibility factors to consider when first designing a study. In addition, studies with transgenic “knockout” mice have provided evidence for specific genes in lung fibrosis caused by CNT exposure. A loss or deficiency of genes coding for key factors involved in the process of the immune response, cell to cell adhesion, and wound healing pathways can determine the susceptibility and severity of CNT-induced fibrogenesis. Though the literature provides a plethora of these deficiencies with regard to pulmonary fibrosis in general, only a handful have been investigated specifically with CNT exposures in experimental animals and these are discussed here (Fig. 3).

Figure 3.

Protein modulators of carbon nanotube (CNT)-induced pulmonary fibrosis identified from transgenic mouse studies and how they regulate inflammation and fibrosis.

Rodent Strain Variation

Inherent variability between genomes of mouse or rat species is an important determinant in fibrogenesis after pulmonary injury by a variety of agents. Walkin et al reviewed strain differences of mice in the development of fibrosis and concludes that strain variation is a determinant for organ specific fibrosis severity¹⁰⁶. A study examining pulmonary inflammation and fibrosis following intratracheal instillation of an alkylating agent, like melphalan or mustard gas, in six strains of rats (DA, PVG, PVG.1AV1, WF, F344, LEW) exemplifies the genetic component of the severity and response timing of inflammation and fibrosis following lung injury¹⁰⁷. Interestingly, after 90 days post exposure, F344 rats have little to no increase in collagen content compared to controls, while all five other strains do¹⁰⁷. Likewise, studies of bleomycin-induced pulmonary fibrosis in multiple strains of mice finds the severity of fibrosis and collagen deposition in the lungs to be strain dependent as well¹⁰⁸. Specifically, the C57BL/6 strain responds greater than DBA/2, Swiss, and BALB/c mice, and the least response was observed from BALB/c mice¹⁰⁸. Another study examining the fibrotic outcome of pulmonary exposure to vanadium pentoxide in two mouse strains found increased inflammatory and collagen content in DBA/2J mice compared to C57BL/6J mice¹⁰⁹. Therefore, strain susceptibility to pulmonary fibrosis may depend on the agent used to induce fibrosis. Of interest, the sex of mice has also been linked to the inflammatory response experienced by mice following MWCNT aspiration exposure where female mice, especially those with reduced glutathione levels, have reduced neutrophilia and MWCNT clearance from their lungs¹¹⁰. Currently, there is a lack of information on sex differences in response to pulmonary exposure to CNTs. The strain of the rodent model used for CNT exposures is an important consideration when analyzing the effects of CNTs in the lung, and when evaluating for fibrotic endpoints the study design should utilize susceptible mouse strains to truly assess the fibrogenic potential of these nanomaterials.

Signal Transducer and Activator of Transcription-1 (STAT1)

The signal transduced and activator of transcription-1 (STAT1) plays a key regulatory role in suppressing the development and progression of CNT-induced fibrogenesis. STAT1 is a transcription factor that is responsive to interferon (IFN) -α, -β, and -γ stimulation. IFN-γ activates the transmembrane interferon receptor (IFNAR) -1 and -2 which have an internal Janus kinase (JAK) activity to recruit and activate STAT1. Activation and dimerization of STAT1 leads to translocation of the dimerized transcription factors and expression of genes that are involved in growth inhibition. Individuals with idiopathic fibrosis and systemic sclerosis have decreased STAT1 transcription and expression as well as fibroblast hyper-proliferation and apoptotic resistance¹¹¹. STAT1 knockout mice exhibit an increased susceptibility to bleomycin induced fibrosis¹¹². Furthermore, these STAT1 knockout mice experience MWCNT-induced exacerbation of fibrosis and asthma demonstrating STAT1 as protective against fibrosis^33,34. Specifically, STAT1 was found to negatively regulate TGF-β1 and its downstream signaling molecule Smad2/3 phosphorylation and activation³⁴. STAT1 knockout fibroblasts isolated from mouse lungs are hyper-responsive to TGF-β1 treatment, transcribing significantly more collagen mRNA and producing more soluble collagen than wild type primary mouse lung fibroblasts³³. STAT1 antagonizes the activity of other STAT family members (STAT3 and STAT6) that have been shown to play roles in promoting fibrogenesis. Therefore, a reduction or loss of STAT1 may also result in less restriction of STAT3 or STAT6. STAT3 is implicated in cell growth and carcinogenesis, while STAT6 is responsible for promoting a more allergic phenotype. TGF-β1 treatment of lung myofibroblasts has been shown to activate STAT3 and promote proliferation adjacent of epithelial damage¹¹³. Either STAT dysregulation would result in an abnormal phenotype and could result in increased proliferation or exacerbation of fibrosis or asthma³³.

Peroxisome Proliferator-Activated Receptor (PPAR) -α, -β/δ, and -γ

Peroxisome proliferator-activated receptors (PPAR) are nuclear receptor proteins that heterodimerize to retinoid X receptor (RXR) and function as transcription factors to regulate processes of fibrogenic inflammation, cellular differentiation, and wound healing¹¹⁴. There are three isoforms of PPARs -α, -γ, and -β/δ. Each of these have been found to play a role in fibrogenesis. PPARα is found to be highly expressed in the heart, muscle, kidney, and liver while PPARβ/δ and γ are ubiquitously expressed^115,116. In cases of idiopathic pulmonary fibrosis there is an incidence of down-regulated peroxisomal biogenesis and metabolism¹¹⁷. A reduction in peroxisomes from decreased biogenesis results in more severe inflammation and dysregulation of the wound healing process and exacerbation of fibrogenesis. PPARα knockout mice develop greater inflammation and fibrosis from bleomycin treatment than wild type mice, while PPARα agonist treatment paired with bleomycin resulted in decreased scaring of mouse lungs¹¹⁸. Treatment of fibroblasts with a PPARα agonist results in a reduction of TGF-β1 induction of myofibroblast differentiation and collagen production¹¹⁷. PPARβ/δ mainly plays a role in regulating the timing of inflammation to wound healing, however little is known about its role in pulmonary fibrosis¹¹⁹. The most influential PPAR in fibrogenesis appears to be PPARγ. PPARγ is a negative regulator of inflammatory cytokines and knockout mouse models exhibit increased inflammation and inflammatory cytokine expression compared to controls¹²⁰. Activation of PPARγ following pulmonary injury from bleomycin has been observed to promote the resolution of both inflammation and fibrosis by suppressing tumor necrosis factor α, procollagen I and connective tissue growth factor expression and was mainly found to be localized to alveolar macrophages and some parenchymal cells^121,122. Treatment of hypertrophic scar fibroblasts with a PPARγ agonist demonstrated that this treatment resulted in decreased expression of Smad3 and Collagen 1 mRNA through induction of miR-145¹²³. Furthermore, PPARγ ligands inhibit TGF-β-induced myofibroblast differentiation in a dose dependent manner through the P13K/AKT pathway^124,125. PPARγ knockout mice have higher expression of Twist1, a transcription factor responsive to classical M1 and not M2 macrophage stimulation¹²⁶. Alveolar macrophages from mice 60 days post-MWCNT exposure exhibit a decrease in PPARγ expression and knockout mice exposed this way experience increased inflammation and larger granuloma formation¹²⁰. Furthermore, PPARγ knockout mouse skin fibroblasts experience an enhanced rate of dermal wound closure, and increased activation of Smad3, AKT, and ERK suggesting that PPAR functions to control fibroblast activation and function following injury¹²⁷. As prominent regulators of inflammation, differentiation, fibrogenesis and wound healing, any PPAR deficiency combined with a fiber insult like CNT exposure could result in a more severe fibrosis.

Myeloperoxidase (MPO)

Myeloperoxidase (MPO), a lysosomal peroxidase enzyme produced and stored in neutrophil granulocytes and released with the purpose of quickly producing hypochlorous acid during the neutrophil-mediated respiratory burst to kill pathogens. Peroxidases have also been implicated in the biodegradation of CNTs¹²⁸. Bronchoalveolar lavage fluid from B6 mice following a 24 hour exposure to uncoated or aluminum oxide coated MWCNTs experienced a significant increase in MPO, indicating that it is involved in the CNT host immune response¹²⁹. Changes to CNTs, by adding a metal oxide coating, may alter the biodegradation process and should be studied further. Degradation of CNTs by neutrophilic MPO can decrease inflammation, as modeled from in vitro testing; a loss or reduction of MPO could then result in increased inflammation and resulting fibrosis from a CNT insult¹³⁰. MPO knockout mice exhibit a 2.5 fold increase in lung collagen content compared to wild type counterparts treated with SWCNTs¹³¹. MPO knockout mouse lungs also exhibit decreased clearance of SWCNTs compared to wild type exposed mice¹³¹. Epidemiologic studies have identified a MPO polymorphism in women with hepatitis C to have increased severity of hepatic fibrosis¹³². MPO is key in the mechanism of CNT degradation which decreases tube length, thereby rendering CNTs more manageable for macrophage phagocytosis and clearance. A decrease or absence of MPO has the potential to extend the biopersistence of CNTs and prolong the inflammatory and fibrogenic effects of these materials in the lung. As a potentially important mechanism involved in the biodegradation of CNTs, the presence and function of MPO is likely important for the extent of inflammation and the severity of fibrosis.

Tissue Inhibitor of Metalloproteinase (TIMP) -1, -2, -3, and -4

Tissue inhibitor of metalloproteinases (TIMP) are a group of four proteinase inhibitors (TIMP-1, -2, -3, and -4) functioning to inhibit metalloproteinase degradation of the extracellular matrix (ECM) proteins. The regulatory balance of ECM production and deconstruction following injury is extremely important in the wound healing and resolution process. Lungs from patients with idiopathic pulmonary fibrosis have overexpression of all four TIMP proteins and result in the accumulation of ECM and enhanced fibrogenesis¹³³. However, in a mouse model with bleomycin-induced pulmonary fibrosis, only TIMP1 was found to be significantly increased at both the mRNA and protein levels¹³⁴. Similarly, TIMP1 mRNA and protein levels are rapidly induced in a time- and dose- responsive manner from MWCNT exposure in a mouse model; TIMP1 knockout mice display significantly less fibrotic foci, collagen, fibroblast recruitment and differentiation indicating it has a key role in pulmonary fibrosis to MWCNT exposure¹³⁵. Fibroblast expression of TIMP1 can be activated by epithelial cell inflammasome activation and production of IL-1β and IL-18 thereby promoting a profibrogenic environment in vitro⁶⁹. TIMP1 complexes with CD63 and integrin β1 on the surface of lung fibroblasts consequently activating ERK1/2 and promoting fibroblast activation and proliferation¹³⁵. An imbalance in the ratio of TIMP to MMP can be utilized as a measure of fibrosis. For example, bleomycin-treatment of rats results in a reduction in MMP-9/TIMP-1 ratio; or a reduction in MMP-9 and increase in TIMP-1 during the promotion of fibrosis ¹³⁶. Unfortunately, while the studies referenced here did look at specific matrix metalloproteinases, they did not examine the ratio of TIMPs to MMPs to examine how these vary between controls and CNT-exposed lungs. The increased expression of TIMP proteins, specifically TIMP1, is associated with CNT injury and without it, tissue metalloproteinases continue to degrade the ECM subsequently dysregulating the balance of ECM production and deconstruction in favor of continued inflammation.

Conclusion

Physicochemical characteristics of CNTs determine the severity of pulmonary fibrosis in rodent models. However, there is still a gap in our knowledge as to which specific characteristics are important towards predicting the human health risks of CNTs. Better characterization of newly synthesized or functionalized nanomaterials may aid in determination of safety or risk of inhalation exposure to such materials. Currently, the state of science in the field of nanotoxicology emphasizes the need to consider physicochemical characteristics in the safe design of CNTs. However, this approach is only effective if consumer products use the ‘safer’ CNTs. Metal catalyst content, length, rigidity, agglomeration status and surface functionalization coatings modify fibrogenic responses in the lungs of rodents and therefore should be considered in the design of CNTs. While many studies have utilized a single bolus CNT- exposure by oropharyngeal aspiration, more studies comparing different types of CNTs should be performed using inhalation exposure and relevant doses that represent occupational settings. Variations in exposure methodology make comparisons between studies difficult. However, as the body of literature keeps growing on this topic we may be able to more effectively elucidate the appropriate physicochemical properties that are of concern.

CNTs promote fibrogenesis through proximal oxidative stress mechanisms which then triggers the activation of intracellular signaling pathways including activation of kinases, transcription factors, enzymes and the inflammasome. These signaling events within cells stimulate the production and release of growth factors like TGF-β1 and PDGF which then drive the recruitment, differentiation, and proliferation of myofibroblasts. Interestingly, some CNTs promote an alternative macrophage phenotype and alter the immune response towards a Th2 response while others activate a more classical macrophage population and resulting Th1 immune response. Studies with transgenic mice have elucidated specific genetic determinants of susceptibility. Future directions of CNT research focused on issues of susceptibility and comparison of CNTs with differing physicochemical properties could improve our mechanistic understanding of CNT-induced pulmonary fibrosis. Overall, the fibrogenic responses to different types of CNTs is due to both the physicochemical characteristics of the nanomaterial and genetic susceptibility of the host.