Measuring lung function in rodent models

Lung function measurement is how a respiratory program tests whether a candidate changes how the lung actually works, rather than only how the tissue looks under a microscope. This article covers the models in common use, the endpoints they generate, the handling variables that determine whether those endpoints are reproducible, and one measurement that is frequently misused. It is written for people commissioning nonclinical respiratory studies in the United States.

Why functional endpoints are used

Respiratory drug development has broadened from relieving symptoms such as cough and breathlessness toward agents intended to act on disease progression, including repair of lung tissue and suppression of fibrosis. An agent of that kind cannot be evaluated by symptom scores in an animal, so the programme needs a quantitative measure of how the lung is performing. Functional measurement provides that: it gives a number that changes with the severity of the model and can be compared between treated and control groups.

Functional data support a proof-of-concept decision: whether the mechanism produces a measurable physiological change in a model of the disease, and at what exposure. That is a decision about whether to continue, and it is made on animal data. It is not a measurement of clinical benefit, and no correlation between a rodent functional endpoint and a clinical outcome is asserted here.

Invasive and non-invasive measurement

Methods divide into two groups. Non-invasive measurement, principally whole-body plethysmography, records the animal breathing freely in a chamber; it can be repeated in the same animal over time and imposes little burden. Invasive measurement requires tracheostomy or intubation and mechanical ventilation, and it yields direct respiratory mechanics such as resistance and compliance. The trade-off is straightforward: the invasive method measures what you want and can normally be done only once, at the end of the study, while the non-invasive method can be repeated but measures something less direct.

The models and how they are made

Asthma

Allergen sensitization models are used to produce airway hyperresponsiveness, and the usual analysis pairs the functional measurement with airway inflammation, typically eosinophil-predominant, so that a change in responsiveness can be related to a change in inflammation. A review of antigen-driven rodent asthma models notes that differences in lung anatomy and in the cell composition of the airways persist as limitations even in humanized animals.^[S4]

The animals are sensitized to an allergen, commonly ovalbumin or house dust mite extract, by repeated intraperitoneal administration, and then challenged with the same allergen by nebulized inhalation. Depending on the schedule, this produces acute or chronic airway inflammation, with increased airway responsiveness and inflammatory cell infiltration.

Chronic obstructive pulmonary disease

These models are designed to produce alveolar destruction of the kind seen in emphysema, and airway obstruction associated with increased mucus production. Because the changes they are intended to mimic develop over a long period and do not reverse, the models generally require a long induction phase. Assessment combines compliance and airflow limitation with the extent of inflammatory infiltration and structural remodelling in tissue.

A single administration of elastase into the airway destroys alveolar structure over a short period and is widely used for that reason. The alternative is repeated exposure to cigarette smoke over weeks to months, which produces sustained inflammation together with structural change in the airways and alveoli. The choice is a trade-off between study duration and how the lesion arises.

Pulmonary fibrosis

Fibrosis models produce stiffening of lung tissue, which is measured as reduced compliance and reduced lung volumes, the pattern described clinically as restrictive. Analysis pairs those functional endpoints with histological assessment of the fibrotic change, since the two can dissociate.

Bleomycin given into the airway, by intranasal or intratracheal administration, is the standard induction. An inflammatory phase precedes fibrosis, and a review of the model reports a peak around day 14.^[S3] The same review notes that standard outcome parameters vary widely after day 21 and may return toward baseline, which is why the timing of measurement is part of the study design rather than a detail.^[S3] Functional measurement at the chosen timepoint is normally combined with collagen quantification and histological staining, so that the degree of fibrosis and the loss of function can be related to each other.

Endpoints and the handling that determines their quality

Two things determine whether respiratory functional data are usable: choosing endpoints that answer the question, and controlling the procedural variables that move those endpoints independently of any treatment. Both are covered below.

The main endpoints

Three endpoints carry most of the interpretation. Airway resistance reflects narrowing or obstruction of the conducting airways and is the endpoint of choice where the mechanism is bronchoconstriction. Dynamic compliance describes how readily the lung distends during breathing, and it moves in opposite directions in the two main structural diseases: upward where alveolar walls are destroyed, downward where tissue becomes fibrotic. Forced expiratory volume over a fixed short interval measures how much air can be expelled and reflects airflow limitation. Reading them together separates a change in the airways from a change in the lung parenchyma, which a single endpoint cannot do.

Procedural variables

Three variables move respiratory measurements enough to obscure a treatment effect. Anaesthetic depth affects breathing directly: too light and spontaneous effort contaminates the measurement, too deep and respiration is suppressed, so depth has to be held consistent across animals and groups. Body temperature affects respiratory rate and metabolism, so it must be maintained actively during the procedure. Intubation technique determines measured resistance and ventilation efficiency, so tube position and the absence of leaks matter as much as the instrument. These are worth writing into the protocol and confirming with the laboratory, because they are the most common explanation for a study that produces uninterpretable variance.

Limitations and how studies are designed around them

Species differences

The mouse lung differs from the human lung in airway branching and in the structure of the peripheral airways and alveoli, and a review of rodent asthma models identifies anatomical differences of this kind, together with differences in airway cell composition, as persistent limitations.^[S4] The same endpoint can therefore represent a different underlying change in the two species. The usual response is not to rely on one parameter: measuring the pressure-volume relationship of the lung, rather than a single point on it, describes the mechanical behaviour more completely.

Analysing the whole pressure-volume curve, rather than compliance at a single point, shows hysteresis and the shape of the relationship at both extremes of volume, which distinguishes changes that a single compliance value would report identically. Choosing endpoints with the species difference in mind is a matter of measuring the right thing in the animal, not of making the animal result stand for a clinical one.

Following the same animal over time

Invasive measurement generally ends the experiment, so it produces one observation per animal, at the end. Non-invasive approaches, including micro-computed tomography imaging and double-chamber plethysmography, can be repeated in the same animal, which allows the progression of the model and the response to treatment to be followed rather than sampled once.

Repeated measurement in the same animal lets each animal serve as its own baseline, which reduces the effect of between-animal variation and usually means fewer animals are needed for the same statistical power. That is a design and animal welfare argument as much as a data quality one.

Matching the model to the phenotype

Respiratory diseases are heterogeneous, and a model driven by a single stimulus represents one route to a lesion rather than the range of presentations seen clinically. Where the target population is defined by a particular mechanism, the model should be chosen to contain that mechanism. Aged animals are used where age-related decline and chronic inflammation are part of the question, and combined challenges, such as viral infection together with allergen exposure, are used where the target is exacerbation rather than baseline disease.

The same applies to the readouts. Function, inflammatory cell counts and tissue changes are analysed together, because a compound can change one without changing the others, and knowing which of them moved is what identifies the mechanism.

Summary

Three decisions determine what a respiratory efficacy study is worth. Which model contains the mechanism being targeted. Which endpoints measure the change that mechanism should produce, with airway resistance, compliance and forced expiratory volume answering different questions and Penh answering none of them directly. And whether anaesthesia, temperature and intubation are controlled tightly enough that the measurement reflects the treatment rather than the handling. Getting these right produces data that support a development decision. They remain animal data, and this article makes no claim about what they imply for outcomes in patients.