Quantitative modeling of in vitro data using an adverse outcome pathway for the risk assessment of decreased lung function in humans.

In the absence of epidemiological data, there is a need to develop computational models that convert in vitro findings to human disease risk predictions following toxicant exposure. In such efforts, in vitro data can be evaluated in the context of adverse outcome pathways (AOPs) that organize mechanistic knowledge based on empirical evidence into a sequence of molecular-, cellular-, tissue-, and organ-level key events that precede an adverse outcome (AO). Here we combined data from advanced in vitro organotypic airway models exposed to combustible cigarette (CC) smoke or Tobacco Heating System (THS) aerosol with an AOP for increased oxidative stress leads to decreased lung function. The mathematical modeling predicted reduced risk of decreased ciliary beating frequency (CBF) based on oxidative stress measurements and reduced risk of decreased mucociliary clearance (MCC) based on CBF measurements in THS aerosol- compared with CC smoke-exposed cultures. To extend the predictions to the AO of decreased lung function, we leveraged human MCC data from current smokers, nonsmokers, former smokers, and users of heated tobacco products. This approach provided a plausible prediction of diminished reduction in lung function in response to THS use compared with continued smoking. The current approach may also present a basis for an integrated approach to testing and assessment of tobacco products for future regulatory decision-making.

Tobacco Alkaloid Assessment in a DSS-Induced Colitis Mouse Model with a Fully Humanized Immune System.

Inflammatory bowel disease (IBD) refers to chronic intestinal immune-mediated diseases including two main disease manifestations: ulcerative colitis (UC) and Crohn's disease (CD). Epidemiological, clinical, and preclinical evidence has highlighted the potential anti-inflammatory properties of naturally occurring alkaloids. In the present study, we investigated the potential anti-inflammatory activities of the tobacco alkaloids nicotine and anatabine in a dextran sulfate sodium (DSS)-induced UC mouse model with a fully humanized immune system. Our results show that nicotine significantly reduced all acute colitis symptoms and improved colitis-specific endpoints, including histopathologically assessed colon inflammation, tissue damage, and mononuclear cell infiltration. The tobacco alkaloid anatabine showed similar effectiveness trends, although they were generally weaker or not significant. Gene expression analysis in the context of biological network models of IBD further pinpointed a possible mechanism by which nicotine attenuated DSS-induced colitis in humanized mice. The current study enables further investigation of possible molecular mechanisms by which tobacco alkaloids attenuate UC symptoms.

Causal biological network models for reactive astrogliosis: a systems approach to neuroinflammation.

Astrocytes play a central role in the neuroimmune response by responding to CNS pathologies with diverse molecular and morphological changes during the process of reactive astrogliosis. Here, we used a computational biological network model and mathematical algorithms that allow the interpretation of high-throughput transcriptomic datasets in the context of known biology to study reactive astrogliosis. We gathered available mechanistic information from the literature into a comprehensive causal biological network (CBN) model of astrocyte reactivity. The CBN model was built in the Biological Expression Language, which is both human-readable and computable. We characterized the CBN with a network analysis of highly connected nodes and demonstrated that the CBN captures relevant astrocyte biology. Subsequently, we used the CBN and transcriptomic data to identify key molecular pathways driving the astrocyte phenotype in four CNS pathologies: samples from mouse models of lipopolysaccharide-induced endotoxemia, Alzheimer's disease, and amyotrophic lateral sclerosis; and samples from multiple sclerosis patients. The astrocyte CBN provides a new tool to identify causal mechanisms and quantify astrogliosis based on transcriptomic data.

Causal Biological Network Model for Inflammasome Signaling Applied for Interpreting Transcriptomic Changes in Various Inflammatory States.

Virtually any stressor that alters the cellular homeostatic state may result in an inflammatory response. As a critical component of innate immunity, inflammasomes play a prominent role in the inflammatory response. The information on inflammasome biology is rapidly growing, thus creating the need for structuring it into a model that can help visualize and enhance the understanding of underlying biological processes. Causal biological network (CBN) models provide predictive power for novel disease mechanisms and treatment outcomes. We assembled the available literature information on inflammasome activation into the CBN model and scored it with publicly available transcriptomic datasets that address viral infection of the lungs, osteo- and rheumatoid arthritis, psoriasis, and aging. The scoring inferred pathway activation leading to NLRP3 inflammasome activation in these diverse conditions, demonstrating that the CBN model provides a platform for interpreting transcriptomic data in the context of inflammasome activation.

Systems Toxicology Approach for Assessing Developmental Neurotoxicity in Larval Zebrafish.

Adverse outcomes that result from chemical toxicity are rarely caused by dysregulation of individual proteins; rather, they are often caused by system-level perturbations in networks of molecular events. To fully understand the mechanisms of toxicity, it is necessary to recognize the interactions of molecules, pathways, and biological processes within these networks. The developing brain is a prime example of an extremely complex network, which makes developmental neurotoxicity one of the most challenging areas in toxicology. We have developed a systems toxicology method that uses a computable biological network to represent molecular interactions in the developing brain of zebrafish larvae. The network is curated from scientific literature and describes interactions between biological processes, signaling pathways, and adverse outcomes associated with neurotoxicity. This allows us to identify important signaling hubs, pathway interactions, and emergent adverse outcomes, providing a more complete understanding of neurotoxicity. Here, we describe the construction of a zebrafish developmental neurotoxicity network and its validation by integration with publicly available neurotoxicity-related transcriptomic datasets. Our network analysis identified consistent regulation of tumor suppressors p53 and retinoblastoma 1 (Rb1) as well as the oncogene Krüppel-like factor (Klf8) in response to chemically induced developmental neurotoxicity. The developed network can be used to interpret transcriptomic data in a neurotoxicological context.

Systems biology approach highlights mechanistic differences between Crohn's disease and ulcerative colitis.

The molecular mechanisms of IBD have been the subject of intensive exploration. We, therefore, assembled the available information into a suite of causal biological network models, which offer comprehensive visualization of the processes underlying IBD. Scientific text was curated by using Biological Expression Language (BEL) and compiled with OpenBEL 3.0.0. Network properties were analysed by Cytoscape. Network perturbation amplitudes were computed to score the network models with transcriptomic data from public data repositories. The IBD network model suite consists of three independent models that represent signalling pathways that contribute to IBD. In the "intestinal permeability" model, programmed cell death factors were downregulated in CD and upregulated in UC. In the "inflammation" model, PPARG, IL6, and IFN-associated pathways were prominent regulatory factors in both diseases. In the "wound healing" model, factors promoting wound healing were upregulated in CD and downregulated in UC. Scoring of publicly available transcriptomic datasets onto these network models demonstrated that the IBD models capture the perturbation in each dataset accurately. The IBD network model suite can provide better mechanistic insights of the transcriptional changes in IBD and constitutes a valuable tool in personalized medicine to further understand individual drug responses in IBD.

In Vivo Profiling of a Natural Alkaloid, Anatabine, in Rodents: Pharmacokinetics and Anti-Inflammatory Efficacy.

Natural alkaloids, a large class of plant-derived substances, have attracted considerable interest because of their pharmacological activities. In this study, the in vivo pharmacokinetics and anti-inflammatory profile of anatabine, a naturally occurring alkaloid, were characterized in rodents. Anatabine was found to be bioavailable and brain-penetrant following systemic administration. Following intraperitoneal (i.p.) administration (1, 2, and 5 mg/kg), anatabine caused a dose-dependent reduction in carrageenan-induced paw edema in rats; in mice, it inhibited the production of pro-inflammatory cytokines and simultaneously elevated the levels of an anti-inflammatory cytokine in a dose-dependent manner 2 h after lipopolysaccharide challenge. Furthermore, anatabine (∼10 and ∼20 mg/kg/day for 4 weeks; inhalation exposure) had effects in a murine model of multiple sclerosis, reducing neurological deficits and bodyweight loss. Comparative studies of the pharmacokinetics and anti-inflammatory activity of anatabine demonstrated its bioequivalence in rats following i.p. administration and inhalation exposure. This study not only provides the first detailed profile of anatabine pharmacokinetics in rodents but also comprehensively characterizes the anti-inflammatory activities of anatabine in acute and chronic inflammatory models. These findings provide a basis for further characterizing and optimizing the anti-inflammatory properties of anatabine.

Systems Toxicology Approach for Testing Chemical Cardiotoxicity in Larval Zebrafish.

Transcriptomic approaches can give insight into molecular mechanisms underlying chemical toxicity and are increasingly being used as part of toxicological assessments. To aid the interpretation of transcriptomic data, we have developed a systems toxicology method that relies on a computable biological network model. We created the first network model describing cardiotoxicity in zebrafish larvae-a valuable emerging model species in testing cardiotoxicity associated with drugs and chemicals. The network is based on scientific literature and represents hierarchical molecular pathways that lead from receptor activation to cardiac pathologies. To test the ability of our approach to detect cardiotoxic outcomes from transcriptomic data, we have selected three publicly available data sets that reported chemically induced heart pathologies in zebrafish larvae for five different chemicals. Network-based analysis detected cardiac perturbations for four out of five chemicals tested, for two of them using transcriptomic data collected up to 3 days before the onset of a visible phenotype. Additionally, we identified distinct molecular pathways that were activated by the different chemicals. The results demonstrate that the proposed integrational method can be used for evaluating the effects of chemicals on the zebrafish cardiac function and, together with observed cardiac apical end points, can provide a comprehensive method for connecting molecular events to organ toxicity. The computable network model is freely available and may be used to generate mechanistic hypotheses and quantifiable perturbation values from any zebrafish transcriptomic data.

Evaluation of toxicity of aerosols from flavored e-liquids in Sprague-Dawley rats in a 90-day OECD inhalation study, complemented by transcriptomics analysis.

The use of flavoring substances is an important element in the development of reduced-risk products for adult smokers to increase product acceptance and encourage switching from cigarettes. In a first step towards characterizing the sub-chronic inhalation toxicity of neat flavoring substances, a study was conducted using a mixture of the substances in a base solution of e-liquid, where the standard toxicological endpoints of the nebulized aerosols were supplemented with transcriptomics analysis. The flavor mixture was produced by grouping 178 flavors into 26 distinct chemical groups based on structural similarities and potential metabolic and biological effects. Flavoring substances predicted to show the highest toxicological effect from each group were selected as the flavor group representatives (FGR). Following Organization for Economic Cooperation and Development Testing Guideline 413, rats were exposed to three concentrations of the FGR mixture in an e-liquid composed of nicotine (23 µg/L), propylene glycol (1520 µg/L), and vegetable glycerin (1890 µg/L), while non-flavored and no-nicotine mixtures were included as references to identify potential additive or synergistic effects between nicotine and the flavoring substances. The results indicated that the inhalation of an e-liquid containing the mixture of FGRs caused very minimal local and systemic toxic effects. In particular, there were no remarkable clinical (in-life) observations in flavored e-liquid-exposed rats. The biological effects related to exposure to the mixture of neat FGRs were limited and mainly nicotine-mediated, including changes in hematological and blood chemistry parameters and organ weight. These results indicate no significant additive biological changes following inhalation exposure to the nebulized FGR mixture above the nicotine effects measured in this sub-chronic inhalation study. In a subsequent study, e-liquids with FGR mixtures will be aerosolized by thermal treatment and assessed for toxicity.

The reduction of DSS-induced colitis severity in mice exposed to cigarette smoke is linked to immune modulation and microbial shifts.

Exposure to cigarette smoke (CS) causes detrimental health effects, increasing the risk of cardiovascular, pulmonary diseases and carcinogenesis in exposed individuals. The impact of CS on Inflammatory Bowel Disease (IBD) has been established by a number of epidemiological and clinical studies. In fact, CS is associated with a higher risk of developing Crohn's disease (CD) while inversely correlates with the development, disease risks, and relapse rate of ulcerative colitis (UC). To investigate the effect of CS exposure on experimental colitis, we performed a comprehensive and integrated comparative analysis of colon transcriptome and microbiome in mice exposed to dextran sodium sulfate (DSS) and CS. Colon transcriptome analysis revealed that CS downregulated specific pathways in a concentration-dependent manner, affecting both the inflammatory state and composition of the gut microbiome. Metagenomics analysis demonstrated that CS can modulate DSS-induced dysbiosis of specific bacterial genera, contributing to resolve the inflammation or accelerate recovery. The risks of smoking far outweigh any possible benefit, thus smoking cessation must always be encouraged because of its significant health benefits. However, the inverse association between active smoking and the development of UC cannot be ignored and the present study lays the foundation for investigating potential molecular mechanisms responsible for the attenuation of colitis by certain compounds of tobacco when decoupled from combustion.

NPA: an R package for computing network perturbation amplitudes using gene expression data and two-layer networks.

BACKGROUND: High-throughput gene expression technologies provide complex datasets reflecting mechanisms perturbed in an experiment, typically in a treatment versus control design. Analysis of these information-rich data can be guided based on a priori knowledge, such as networks of related proteins or genes. Assessing the response of a specific mechanism and investigating its biological basis is extremely important in systems toxicology; as compounds or treatment need to be assessed with respect to a predefined set of key mechanisms that could lead to toxicity. Two-layer networks are suitable for this task, and a robust computational methodology specifically addressing those needs was previously published. The NPA package ( https://github.com/philipmorrisintl/NPA ) implements the algorithm, and a data package of eight two-layer networks representing key mechanisms, such as xenobiotic metabolism, apoptosis, or epithelial immune innate activation, is provided. RESULTS: Gene expression data from an animal study are analyzed using the package and its network models. The functionalities are implemented using R6 classes, making the use of the package seamless and intuitive. The various network responses are analyzed using the leading node analysis, and an overall perturbation, called the Biological Impact Factor, is computed. CONCLUSIONS: The NPA package implements the published network perturbation amplitude methodology and provides a set of two-layer networks encoded in the Biological Expression Language.

A Meta-Analysis of the Performance of a Blood-Based Exposure Response Gene Signature Across Clinical Studies on the Tobacco Heating System 2.2 (THS 2.2).

As part of emerging tobacco harm reduction strategies, modified risk tobacco products (MRTP) are being developed to offer alternatives that have the potential to reduce the individual risk and population harm compared with smoking cigarettes for adult smokers who want to continue using tobacco and nicotine products. MRTPs are defined as any tobacco products that are distributed for use to reduce harm or the risk of tobacco-related disease associated with commercially marketed tobacco products. One such candidate MRTP is the Tobacco Heating System (THS) 2.2, which does not burn tobacco but instead heats it, thus producing significantly reduced levels of harmful and potentially harmful constituents compared with cigarettes. The clinical assessment of candidate MRTPs requires the development of exposure-response markers to distinguish current smokers from either nonsmokers or former smokers with high specificity and sensitivity. Toward this end, a whole blood-derived gene signature was previously developed and reported. Four randomized, controlled, open-label, three-arm parallel group reduced exposure clinical studies have been conducted with subjects randomized to three arms: switching from cigarettes to THS 2.2, continuous use of cigarettes, or smoking abstinence. These clinical studies had an investigational period of 5 days in confinement, which was followed by an 85-day ambulatory period in two studies. Here we tested the previously developed blood-derived signature on the samples derived from those clinical studies. We showed that in all four studies, the signature scores were reduced consistently in subjects who either stopped smoking or switched to THS 2.2 compared with subjects who continued smoking cigarettes.

Construction of a Suite of Computable Biological Network Models Focused on Mucociliary Clearance in the Respiratory Tract.

Mucociliary clearance (MCC), considered as a collaboration of mucus secreted from goblet cells, the airway surface liquid layer, and the beating of cilia of ciliated cells, is the airways' defense system against airborne contaminants. Because the process is well described at the molecular level, we gathered the available information into a suite of comprehensive causal biological network (CBN) models. The suite consists of three independent models that represent (1) cilium assembly, (2) ciliary beating, and (3) goblet cell hyperplasia/metaplasia and that were built in the Biological Expression Language, which is both human-readable and computable. The network analysis of highly connected nodes and pathways demonstrated that the relevant biology was captured in the MCC models. We also show the scoring of transcriptomic data onto these network models and demonstrate that the models capture the perturbation in each dataset accurately. This work is a continuation of our approach to use computational biological network models and mathematical algorithms that allow for the interpretation of high-throughput molecular datasets in the context of known biology. The MCC network model suite can be a valuable tool in personalized medicine to further understand heterogeneity and individual drug responses in complex respiratory diseases.

How complex should an in vitro model be? Evaluation of complex 3D alveolar model with transcriptomic data and computational biological network models.

To more accurately model inhalation toxicity in vitro, we developed a tetra-culture system that combines lung alveolar epithelial cells, endothelial cells, macrophages, and mast cells in a three-dimensional orientation. We characterized the influence of the added complexity using network perturbation analysis and gene expression data. This will allow us to gain insight into the steady-state profile of the assembled, complete three-dimensional model using all four cell types and of simpler models of one, two, or three cell types. Gene expression data were analyzed using cause-and-effect biological network models, together with a quantitative network-scoring algorithm, to determine the biological impact of co-culturing the various cell types. In the assembled tetra-culture, macrophages appeared to be the largest contributors to overall network perturbations, promoting high basal levels of oxidative stress and inflammation. This finding led to further optimization of the model using rested macrophages; the addition of rested macrophages decreased the basal inflammatory and cell stress status of the co-culture. Finally, we compared transcriptional profiles from publicly available datasets of conventional in vitro models representative of the airways and of healthy human lung tissues to assess similarities between our model and other in vitro models and the human lung. On the transcriptional level, we found an increasing correlation between airway models and normal human lung tissue, particularly as cell types became more physiologically relevant and the complexity of the system increased. This indicates that the combination of multiple lung-relevant cell types in vitro does indeed increase similarity to the physiological counterpart.

Biological changes in C57BL/6 mice following 3 weeks of inhalation exposure to cigarette smoke or e-vapor aerosols.

We compared early biological changes in mice after inhalation exposures to cigarette smoke or e-vapor aerosols (MarkTen® cartridge with Carrier, Test-1, or Test-2 formulations; 4% nicotine). Female C57BL/6 mice were exposed to 3R4F cigarette smoke or e-vapor aerosols by nose-only inhalation for up to 4 hours/day, 5 days/week, for 3 weeks. The 3R4F and e-vapor exposures were set to match the target nose port aerosol nicotine concentration (∼41 µg/L). Only the 3R4F group showed postexposure clinical signs such as tremors and lethargy. At necropsy, the 3R4F group had significant increases in lung weight and changes in bronchoalveolar lavage parameters, as well as microscopic findings in the respiratory tract. The e-vapor groups had minimal microscopic changes, including squamous metaplasia in laryngeal epiglottis, and histiocytic infiltrates in the lung (Test-2 group only). The 3R4F group had a higher incidence and severity of microscopic findings compared to any e-vapor group. Transcriptomic analysis also showed that the 3R4F group had the highest number of differentially expressed genes compared to Sham Control. Among e-vapor groups, Test-2 group had more differentially expressed genes but the magnitude of gene expression-based network perturbations in all e-vapor groups was ∼94% less than the 3R4F group. On proteome analysis in the lung, differentially regulated proteins were detected in the 3R4F group only. In conclusion, 3-weeks of 3R4F exposure induced molecular and microscopic changes associated with smoking-related diseases in the respiratory tract, while e-vapor exposures showed substantially reduced biological activities.

Mechanistic Evaluation of the Impact of Smoking and Chronic Obstructive Pulmonary Disease on the Nasal Epithelium.

Chronic obstructive pulmonary disease (COPD) is one of the major causes of chronic morbidity and mortality worldwide. The development of markers of COPD onset is hampered by the lack of accessibility to the primary target tissue, and there is a need to consider other sample sources as surrogates for biomarker research. Airborne toxicants pass through the nasal epithelium before reaching the lower airways, and the similarity with bronchial histology makes it an attractive surrogate for lower airways. In this work, we describe the transcriptomics findings from the nasal epithelia of subjects enrolled in a clinical study focusing on the identification of COPD biomarkers. Transcriptomic data were analyzed using the biological network approach that enabled us to pinpoint the biological processes affected in the upper respiratory tract in response to smoking and mild-to-moderate COPD. Our results indicated that nasal and lower airway immune responses were considerably different in COPD subjects and caution should be exercised when using upper airway samples as a surrogate for the lower airway. Nevertheless, the network approach described here could present a sensitive means of identifying smokers at risk of developing COPD.

Novel approaches to develop community-built biological network models for potential drug discovery.

INTRODUCTION: Hundreds of thousands of data points are now routinely generated in clinical trials by molecular profiling and NGS technologies. A true translation of this data into knowledge is not possible without analysis and interpretation in a well-defined biology context. Currently, there are many public and commercial pathway tools and network models that can facilitate such analysis. At the same time, insights and knowledge that can be gained is highly dependent on the underlying biological content of these resources. Crowdsourcing can be employed to guarantee the accuracy and transparency of the biological content underlining the tools used to interpret rich molecular data. Areas covered: In this review, the authors describe crowdsourcing in drug discovery. The focal point is the efforts that have successfully used the crowdsourcing approach to verify and augment pathway tools and biological network models. Technologies that enable the building of biological networks with the community are also described. Expert opinion: A crowd of experts can be leveraged for the entire development process of biological network models, from ontologies to the evaluation of their mechanistic completeness. The ultimate goal is to facilitate biomarker discovery and personalized medicine by mechanistically explaining patients' differences with respect to disease prevention, diagnosis, and therapy outcome.

Aerosol from Tobacco Heating System 2.2 has reduced impact on mouse heart gene expression compared with cigarette smoke.

Experimental studies clearly demonstrate a causal effect of cigarette smoking on cardiovascular disease. To reduce the individual risk and population harm caused by smoking, alternative products to cigarettes are being developed. We recently reported on an apolipoprotein E-deficient (Apoe-/-) mouse inhalation study that compared the effects of exposure to aerosol from a candidate modified risk tobacco product, Tobacco Heating System 2.2 (THS2.2), and smoke from the reference cigarette (3R4F) on pulmonary and vascular biology. Here, we applied a transcriptomics approach to evaluate the impact of the exposure to 3R4F smoke and THS2.2 aerosol on heart tissues from the same cohort of mice. The systems response profiles demonstrated that 3R4F smoke exposure led to time-dependent transcriptomics changes (False Discovery Rate (FDR) < 0.05; 44 differentially expressed genes at 3-months; 491 at 8-months). Analysis of differentially expressed genes in the heart tissue indicated that 3R4F exposure induced the downregulation of genes involved in cytoskeleton organization and the contractile function of the heart, notably genes that encode beta actin (Actb), actinin alpha 4 (Actn4), and filamin C (Flnc). This was accompanied by the downregulation of genes related to the inflammatory response. None of these effects were observed in the group exposed to THS2.2 aerosol.

Evaluation of the tobacco heating system 2.2. Part 9: Application of systems pharmacology to identify exposure response markers in peripheral blood of smokers switching to THS2.2.

As part of current harm reduction strategies, candidate modified risk tobacco products (MRTP) are developed to offer adult smokers who want to continue using tobacco product an alternative to cigarettes while potentially reducing individual risk and population harm compared to smoking cigarettes. One of these candidate MRTPs is the Tobacco Heating System (THS) 2.2 which does not burn tobacco, but instead heats it, thus producing significantly reduced levels of harmful and potentially harmful constituents (HPHC) compared with combustible cigarettes (CC). A controlled, parallel group, open-label clinical study was conducted with subjects randomized to three monitored groups: (1) switching from CCs to THS2.2; (2) continuous use of non-menthol CC brand (CC arm); or (3) smoking abstinence (SA arm) for five days. Exposure response was assessed by measuring biomarkers of exposure to selected HPHCs. To complement the classical exposure response measurements, we have used the previously reported whole blood derived gene signature that can distinguish current smokers from either non-smokers or former smokers with high specificity and sensitivity. We tested the small signature consisting of only 11 genes on the blood transcriptome of subjects enrolled in the clinical study and showed a reduced exposure response in subjects that either stopped smoking or switched to a candidate MRTP, the THS2.2, compared with subjects who continued smoking their regular tobacco product.

Semi-Automated Curation Allows Causal Network Model Building for the Quantification of Age-Dependent Plaque Progression in ApoE-/- Mouse.

The cellular and molecular mechanisms behind the process of atherosclerotic plaque destabilization are complex, and molecular data from aortic plaques are difficult to interpret. Biological network models may overcome these difficulties and precisely quantify the molecular mechanisms impacted during disease progression. The atherosclerosis plaque destabilization biological network model was constructed with the semiautomated curation pipeline, BELIEF. Cellular and molecular mechanisms promoting plaque destabilization or rupture were captured in the network model. Public transcriptomic data sets were used to demonstrate the specificity of the network model and to capture the different mechanisms that were impacted in ApoE-/- mouse aorta at 6 and 32 weeks. We concluded that network models combined with the network perturbation amplitude algorithm provide a sensitive, quantitative method to follow disease progression at the molecular level. This approach can be used to investigate and quantify molecular mechanisms during plaque progression.