Jo Wallace, PhD
It’s All About the Dose – How to Link In Vitro and In Vivo
Moving closer to a fully in vitro system for respiratory toxicology testing
The billion-dollar question for in vitro toxicology— one we get asked on nearly a daily basis—is “how do I extrapolate from my expected human exposure to a concentration that I can use to test in vitro ?” Or how do I achieve In Vitro-In Vivo Extrapolation (IVIVE). This is the issue/question that underpins all in vitro toxicology. What is the physiological relevance of our in vitro dose?
Questions around IVIVE are being asked by the US Environmental Protection Agency (EPA) and other regulatory bodies around the world, in their drive to reduce the number of animals used in safety testing. Yet these questions are increasingly challenging as we delve deeper into complex issues. Consider also the directive issued by the EPA in 2019 to eliminate all mammalian study requests and funding by 2035 , and the pressure is higher than ever to drive use of in vitro data in regulatory decisions.
One area which has received a lot of interest is the lung, because it is a particularly complex case. Its vast surface area is a very convoluted, folded-up structure in humans measuring 50 to 75 square meters—the equivalent of 4-5 parking spaces or a fifth of an IMAX screen!! Its surface area and role of gas exchange to and from the blood makes it an ideal target for inhaled pharmaceuticals, and for toxicity. Not only do you have to deal with complex aerosols if you want a truly “physiological” delivery, but you also have to navigate a complex branched structure of multiple tissue and cell types, fluid dynamics of particles in the air flow, and anatomical differences between species. And you must contend with all of this is before you can think about mechanical stretch, and the lung’s complex protective mechanisms: the mucociliary escalator, lung surfactant, and the immune system!
In short, we have come a long way, but still have a marathon to run. Certainly, there is no single in vitro model that can capture all aspects of this complex system. Here is a tour of what we currently do have at our disposal.
Which Tissue Model?
The word “organotypic” gets thrown around a lot for in vitro toxicology.
Organotypic culture is defined as “the culture of an organ collected from an organism. It is one method allowing the culture of complex tissues or organs. It allows the preservation of the architecture of the cultured organ and most of its cellular interactions.”
Thinking about the airways, there are some advanced organotypic or co-culture models of different regions of the lung produced by the two main suppliers Epithelix (MucilAir™, SmallAir™) and MatTek (EpiAirway™ and EpiAlveolar™). An electron micrograph of MucilAir™ is shown above.
These models are grown at the air liquid interface, and are reconstructed from primary human cells, mimicking much of the structure and biological responses of the parent organ. Critical here is the “human” part. They are the perfect species for protecting human health. Models are available to reproduce the entire respiratory tract from the nose to the alveoli. Both suppliers also produce several human disease models (like cystic fibrosis, chronic obstructive pulmonary disorder (COPD), smoker-derived models and rat upper airway models to aid translational toxicology.
Identifying Relevant Dose Levels for the Chemical / Agrochemical Sector
Let’s use a pesticide as an example. I’ll call it Pestatron (Disclaimer: I made this product up to prove my point). The producer ofPestatron wants the farmers to spray a concentration of 1% (v/v) Pestatron in water over their crops. What application rate of Pestatron should they use for their in vitro test to obtain human relevant toxicity data?
We identify that they want to target the upper airway, because they have some data from read-across suggesting that this is the target for local toxicity for Pestatron.
(Read-across is when your data-poor target substance is considered similar enough to a data-rich source substance to allow you to base some of your safety assessment decisions on the source substance without testing on the target substance)
Based on what we know so far, we pick an upper airway in vitro model. But how much Pestatron will reach the upper airway of our farmer after he has spent a day in his tractor spraying Pestatron over his crops? How much should we apply to the in vitro model to give relevant results?
The MPPD model (Multiple Path Particle Dosimetry Model) is a computational model that is used for estimating human and laboratory animal inhalation particle dosimetry. The model is applicable to risk assessment, research, and education, and is freely available from ARA, the developer of the product.
We then feed the MPPD model with data about Pestatron spray. This can include concentration in water in the tank, generated in the air around the tractor cab, particle size and characteristics, the farmer’s working day, and other parameters about physiology and breathing rates. Among the wealth of data provided is the ratio of deposition to different regions of the lung vs exhalation, and a deposition rate (in µg/min/cm2) onto the different regions. We can then use this value for the upper airway region to estimate the total deposited mass onto the surface of the in vitro model. Now we have a way to bridge between an aerosol concentration, a specific region of the lung, and an in vitro model, to test at a concentration that will deliver a relevant amount of Pestatron to the cells. The more information that the computational model can be provided with, the more accurate the output. This computational modelling approach to in vitro dosimetry has been presented to the US EPA and reviewed as part of a New Approach Methodology (NAM) for inhalation risk assessment for the pesticide Chlorothalonil (Charles River participated in the NAM meetings and co-authored the paper.
Identifying Relevant Dose Levels for the Pharmaceutical Sector
MPPD modelling software was developed to look at exposure over time, so when you are wanting to extrapolate an in vitro dose level for an inhaled pharmaceutical, you need a modified approach. The key difference is that typically the drug is inhaled in a single deep breath from an inhaler (not over 6‑8 hours of normal breathing).
This time let us consider a new inhaled pharmaceutical Lungbutanide (once again, not an actual drug). The developer knows that they need to deliver 10 mg of Lungbutenide to patients as a single inhalation, twice a day for a week-long treatment course. They are wondering if this will cause any local toxicity. Of course, there is no sense in putting 10 mg of Lungbutenide onto the in vitro tissues – that 10 mg will be distributed across the whole respiratory tract, and some will be exhaled again. Remember this is 4-5 parking spaces! We take advantage of the fact that MPPD software gives an output rate per minute and tells it that our patient is only going to inhale once per minute (this of course isn’t true, but for these purposes that’s OK).
We are also going to adjust the tidal volume parameter (in a normal breath while at rest an adult human might inhale 500 mL air at a time – we’ll change this to 1 L). Plenty of data is available in peer-reviewed publications to support this choice depending on the type of inhaler and clinician’s instructions. We know that we want the 1 L of air to contain the full 10 mg of Lungbutenide, so our concentration in air will be 10 mg/L. All the other parameters are decided in the same way as the earlier example, using as much information as possible, and we get our output from the computational model in µg/min/cm2. One minute equals one breath in this hypothetical scenario. Once again, we now have an application rate that allows us to extrapolate from a clinically relevant in-human dose to a relevant in vitro concentration.
To Nebulize or Not to Nebulize?
Once the decision is made about how much of a substance to apply, the next question is how to apply it? The simplest approach is to prepare a solution or suspension and directly apply this liquid to the epithelial (upper) surface of the tissues with a pipette. This method allows the greatest degree of control over what is being applied. However, the toxicity of some materials might be affected by how they are delivered. For instance, an inhaled pharmaceutical may have propellant or solvent that is intended to evaporate before it reaches the cell surface, or a dry powder may act differently to a solution. In these cases, it is beneficial to be able to expose the cell models to aerosols. We have a Vitrocell® Cloud Chamber which allows aerosol application of solutions or suspensions (in water or saline) to the tissues. Applied mass can be quantified using a Quartz Crystal Microbalance (QCM) or by collecting samples on filters for chemical analysis.
What is next?
We are continuously working to extend our in vitro aerosol capabilities. Soon we will be able to test using a continuous flow of aerosol in addition to the single cloud. This system will also allow us to work with a greater range of aerosol types and dry powders. In line sampling for particle size is also on the cards for the near future. The science and capabilities underpinning in vitro respiratory toxicology is continuously evolving. Models are in development that include aspects of the immune system, or mechanical stretch simulating breathing. In vitro lung models can’t yet fully recreate the complex human lung, but we are getting closer and closer to a truly physiologically relevant in vitro system.
Jo Wallace, PhD, is a Study Director at Charles River Laboratories' Discovery division. Her areas of expertise are in vitro toxicology, respiratory toxicology and discovery pharmacology.