We've updated our Privacy Policy to make it clearer how we use your personal data. We use cookies to provide you with a better experience. You can read our Cookie Policy here.


The Future of Toxicology: An Industry Case Study

Listen with
Register for free to listen to this article
Thank you. Listen to this article using the player above.

Want to listen to this article for FREE?

Complete the form below to unlock access to ALL audio articles.

Read time: 8 minutes

The field of toxicology is going through a radical transformation. Technological advances are changing the ways in which we generate data and the emergence of systems toxicology has allowed traditional study endpoints to be enhanced with far deeper levels of analysis. In combination, new experimental and computational methods offer the potential for more effective, efficient and reliable toxicological testing strategies. However, these advances also create new challenges. The rapid growth of data, tools and resources in toxicology means it can be hard to reach consensus on which methodologies perform best. Similarly, assessing the reproducibility of scientific findings requires full access to protocols and data, and effective systems to manage that data.

At Philip Morris International (PMI), we are developing a range of new technologies, techniques and initiatives in an effort to advance the field of toxicology and address some of the challenges faced by the community. This article describes some of our new in vitro approaches to toxicology, as well as initiatives we are running to facilitate the verification of methods and conclusions, and to enhance data sharing and scientific discovery. 


One of the cornerstones of 21st century toxicology, as set out in 2007 in the US National Research Council’s landmark vision and strategy report,1 is the development of advanced in vitro models which offer new opportunities for physiologically relevant testing. Such models have the potential to reduce the necessity for animal testing and offer more cost-efficient and timely results, as well as a deeper understanding of the biological processes underlying toxicity. They have important implications for industry as well as for regulatory bodies responsible for the scientific oversight of chemicals, pharmaceuticals and many consumer products.

At PMI, we are developing a range of novel, advanced in vitro techniques for the scientific assessment of Reduced-Risk Products (RRPs)*, which are non-combustible nicotine and tobacco products that have the potential to significantly reduce individual risk and population harm compared to cigarettes. These include three-dimensional organotypic models and innovative organ-on-a-chip models, each of which allow us to apply systems toxicology techniques to reveal deeper layers of biological understanding.

Three-Dimensional, Organotypic Models

To assess the biological impact of aerosol generated from one of PMI’s candidate RRPs, the Tobacco Heating System (THS) 2.2, studies have been conducted using human oral, nasal, and gingival epithelial cells grown in three-dimensional culture systems.2-4 The conditioning of the cultures allowed them to develop ‘organotypic’ tissue complexity which closely resembles that found in the human oral cavity, airways, and gums, respectively. Cultures were exposed to either cigarette smoke or aerosol from THS at various concentrations. The biological effects of these exposures were assessed at various time-points using a combination of well-established in vitro testing procedures, -omics measurements and novel computational techniques. Multiple experimental repetitions were conducted (five in the nasal study, four in the oral study, three in the gingival study) to obtain robust, reliable and reproducible measurements.

The experimental design of these studies included the analysis of multiple endpoints, including measurements of cytotoxicity, alterations in tissue morphology, secretion of inflammatory mediators, impact on processes involved in the metabolism of toxicants, and perturbation of mRNA and miRNA expression profiles. The adopted computational techniques for the interpretation of mRNA expression profiles changes involved the processing of the collected data through a set of literature-supported biological network models5 that are known to be related to respiratory disease (29 networks in the nasal study, 28 in the oral and gingival studies). The biological impact of cigarette smoke and THS aerosol was then quantified through the computational evaluation of Network Perturbation Amplitude (NPA)6 scores and complemented by standard gene-set analysis.

PMI has applied similar three-dimensional, organotypic in vitro methodologies to the assessment of the biological impact of the aerosol generated by the Carbon Heated Tobacco Product, another RRP.7-11 These studies have looked at the impact on human oral, nasal, gingival, small airway, and bronchial epithelial cells.

Organ-On-A-Chip Models

Advances in microengineering have enabled miniaturised devices to be incorporated into three-dimensional in vitro models to reproduce the complex microenvironments of specific organs. Known as organ-on-a-chip models, many have now been developed and applied to various biomedical applications. However, organ-on-a-chip models are only able to represent a single organ, and therefore cannot be used to study the many organ-to-organ interactions observed in the human body.12 To address this, efforts are underway to incorporate multiple organs in a single model.

One of the main challenges in developing multiple-organ-on-a-chip models is to successfully connect the tissues from two separate organs and hold them in a stable condition for a sufficient amount of time to allow for toxicological testing. PMI has recently developed a combined lung-liver-on-a-chip that has been demonstrated to hold both lung and liver tissues in a stable state for at least four weeks.13 This is an important development for the toxicological assessment of airborne compounds. While such compounds are absorbed through the lung, the toxicity of some compounds is the consequence of their metabolic activation by the liver.

Multiple-organ-on-a-chip models have the potential to reduce the need for animal testing and provide more accurate, detailed and timely data. By using human tissues, they avoid the challenges involved in translating results observed in one species to another. Multiple-organ-on-a-chip models may also unlock new approaches to drug development by providing an improved understanding of dose responses, enabling the detection of drug resistance, and highlighting potential side effects. 


Toxicology, and in particular systems toxicology, is characterised by increasingly large datasets and a proliferation of novel tools and techniques for data analysis. The question of how scientific methods and conclusions can be validated and trusted is therefore highly pertinent.

The traditional peer-review system, while remaining one of the most important mechanisms for quality control of scientific papers, is limited in its ability to accurately and comprehensively validate methods and conclusions in systems toxicology. It is open to error and bias from both authors and reviewers. Authors may be selective in what data and analyses they report in their papers, they may only choose metrics that give the best outcomes. These errors and biases can occur either intentionally or unintentionally. Reviewers meanwhile are only able to assess whether a paper’s conclusions are supported by the data and results that are shown in the paper.

In an effort to address this, PMI inaugurated the sbv IMPROVER project, with the aim of developing a robust methodology for verifying scientific methods and results in the context of industrial and academic research.14,15 sbv IMPROVER stands for Systems Biology Verification: Industrial Methodology for PROcess VErification in Research. Based on the principles of crowd-sourcing and collaborative competition, the project is designed as a series of open scientific challenges where methods and conclusions related to specific scientific problems are rigorously and impartially scrutinised.

sbv IMPROVER is designed to ensure credibility and the reproducibility of scientific conclusions. In sbv IMPROVER challenges:

Many contributors are involved, each bringing independent methods and knowledge 

Different solutions can be applied to various aspects of a complex problem

Solutions can be combined to identify the optimal approach

Benchmarking is unbiased

Conclusions are fully reproducible and thus can be applied with a high degree of confidence

To date, sbv IMPROVER challenges have looked at the identification of biomarkers of disease, the extent to which biological processes are conserved across species, the comprehensiveness of biological network models, and the predictive value of blood gene expression data.

The current sbv IMPROVER challenge – the Microbiomics Challenge – explores fundamental questions relating to the function and composition of the microbiome. Studying the microbiome requires advanced sequencing technologies and computational analysis techniques. However, choosing the most appropriate computational approach can be difficult, since many different methodologies are available and it is unclear how they can be objectively benchmarked. The first phase of the sbv IMPROVER Microbiomics Challenge – the Microbiota Composition Prediction Challenge – is designed to address this by evaluating computational methodologies for their ability to predict microbial composition.


In toxicology, as in many biomedical sciences, technological advances over the past few decades have led to a rapid growth in data generation. In the interest of scientific discovery and to facilitate reproducibility of conclusions, PMI believes that the future of toxicology lies in the full disclosure and reusability of this data. Given the fact that we now typically deal with large, multi-faceted and highly complex datasets, it is important that data is shared in a way that is practical for other scientists to search and scrutinise.

At PMI, the concepts of scientific reproducibility and data reusability are particularly important. Recognising the need for our RRP-related science to be reviewed, scrutinised and verified by the external scientific community, as well as by regulatory bodies such as the U.S. Food and Drug Administration, it is crucial that our methods and data are transparently shared in a way that allows easy review and understanding.

It is for these reasons that PMI has launched INTERVALS, an inhalation toxicology repository for RRP data and research.16 The INTERVALS website hosts comprehensive, annotated datasets that have been generated by PMI as part of their development and assessment of RRPs. It has been designed to facilitate the practical reuse of methods and data, and thus to evaluate the reproducibility of scientific findings. We are also inviting other industries and institutions producing data relevant to RRPs to submit that data to the INTERVALS platform for the benefit of the wider community.

Alongside raw data, INTERVALS also provides rich metadata to describe experiments, data production, data processing, and additional relevant information.

In order to interpret these multi-faceted datasets, INTERVALS utilises Garuda technology which connects different databases, applications and services using a language-independent interface.17 It allows users to build customisable research workflows and bespoke tools by which to visualise and analyse data. An open, community-driven platform, Garuda is similarly aligned with the principles of transparency that underlie the INTERVALS platform. Ultimately, it enables researchers to integrate and extract meaningful information from large scale, multi-faceted datasets.

While INTERVALS is designed specifically for inhalation toxicology, the concept has the potential to be useful in many other areas, for example, assessment of pharmaceutical products, air-pollution risk management, and nanotechnology innovation. Essentially, any field which deals with large and complex data could benefit from this approach as the issues concerning transparency, data reusability and reproducibility are general across the sciences.

Julia Hoeng is Director of Systems Toxicology, Biological Systems Research, Philip Morris International
Manuel C. Peitsch is Chief Scientific Officer, Philip Morris International

* Reduced-Risk Products (“RRPs”) is the term we use to refer to products that present, are likely to present, or have the potential to present less risk of harm to smokers who switch to these products versus continued smoking.  We have a range of RRPs in various stages of development, scientific assessment and commercialisation.  Because our RRPs do not burn tobacco, they produce far lower quantities of harmful and potentially harmful compounds than found in cigarette smoke.


1. National Research Council. Toxicity Testing in the 21st century: a vision and a strategy. 2007, National Academy Press, Washington, DC.

2. Zanetti, F, et al. 2016. Systems toxicology assessment of the biological impact of a candidate modified risk tobacco product on human organotypic oral epithelial cultures. Chem. Res. Toxicol, 2016. Available online at: http://dx.doi.org/10.1021/acs.chemrestox.6b00174.

3. Iskandar, A, et al. 3-D nasal cultures: systems toxicological assessment of a candidate modified-risk tobacco product. ALTEX, 2016. Available online at: http://dx.doi.org/10.14573/altex.1605041

4. Zanetti, F, et al. Comparative systems toxicology analysis of cigarette smoke and aerosol from a candidate modified risk tobacco product in organotypic human gingival epithelial cultures: A 3-day repeated exposure study. Food and Chemical Toxicology, 2017. Available online at: http://www.sciencedirect.com/science/article/pii/S0278691516304884?via%3Dihub#fig1 

5. Boué, S, et al. Causal biological network database: a comprehensive platform of causal biological network models focused on the pulmonary and vascular systems. Database, 2015. Available online at: http://dx.doi.org/10.1093/database/bav030 

6. Hoeng, J, et al. Case study: the role of mechanistic network models in systems toxicology. Drug Discovery Today, 2014. Available online at: http://dx.doi.org/10.1016/j.drudis.2013.07.023 

7. Iskandar, A, et al. Reduced impact of a heat-not-burn product aerosol in comparison with cigarette smoke: a systems toxicology assessment on organotypic buccal epithelial cultures. SOT 56th, March 12–16, 2017, Baltimore, MD, USA.

8. Iskandar, A, et al. An application of a systems toxicology approach to assess the biological impact of carbon-heated tobacco product aerosol using in vitro organotypic human nasal cultures. SOT 56th, March 12–16, 2017, Baltimore, MD, USA.

9. Iskandar, A, et al. In vitro systems toxicology assessment of exposure to aerosol from a Carbon Heated Tobacco Product as compared with exposure to cigarette smoke: the impact on nasal and small airway epithelial cultures. 10th World Congress of Alternatives and Animal Use in the Life Sciences, Seattle, Washington, USA, August 20-24, 2017.

10. Zanetti, F, et al. Systems toxicology assessment of repeated exposure to cigarette smoke and a candidate modified-risk tobacco product aerosol on gingival organotypic cultures. 53rd Congress of the European Societies of Toxicology, Bratislava, Slovakia, September 10-13, 2017.

11. Gonzalez-Suarez, I, et al. In vitro systems toxicology assessment of a carbon-heated tobacco product in normal human bronchial epithelial cells. SOT 56th, March 12–16, 2017, Baltimore, MD, USA.

12. Bovard, D, et al. Organs-on-a-chip: A new paradigm for toxicological assessment and preclinical drug development. Toxicology Research and Application, 2017. Available online at: http://journals.sagepub.com/doi/full/10.1177/2397847317726351 

13. Bovard, D. Establishment and characterization of a lung/liver organ-on-a-chip model. 10th World Congress on Alternatives and Animal Use in the Life Sciences, August 20-24, 2017, Seattle, Washington, USA.

14. Meyer, P, et al. Verification of systems biology research in the age of collaborative competition. Nature Biotechnology, 2011;29(9):811-5. Available online at: https://www.nature.com/nbt/journal/v29/n9/full/nbt.1968.html 

15. Meyer, P, et al. Industrial methodology for process verification in research (IMPROVER): toward systems biology verification. Bioinformatics, 2012. Available online at: https://academic.oup.com/bioinformatics/article-lookup/doi/10.1093/bioinformatics/bts116 

16. Boué, S, et al. Supporting evidence-based analysis for modified risk tobacco products through a toxicology data-sharing infrastructure. F1000Research, 2017. Available online at: http://dx.doi.org/10.12688/f1000research.10493.1 

17. Ghosh, S, et al. Software for systems biology: from tools to integrated platforms. Nature Reviews Genetics, 2011. Available online at: http://www.nature.com/nrg/journal/v12/n12/abs/nrg3096.html