Radiation Effects in the Lung

The lung has long been recognized clinically as being highly radiosensitive, resulting in its consideration as a dose-limiting organ in many radiation therapy protocols. Because of this sensitivity, radiation-induced lung injury is of significant medical concern, not only with respect to localized radiotherapy-related exposures, but also for its potential role following radiation accident scenarios. These might include exposure from a nuclear detonation or from a radiological dispersion device, which could result in inhalation of radioactive particles. As supportive care protocols have improved, individuals surviving ARS have ultimately succumbed to a multiple organ dysfunction syndrome, with pneumonitis and/or pulmonary fibrosis playing a prominent role (4). For example, following the Chernobyl incident, seven patients were diagnosed with pneumonitis and, of those, two died. Of a subset of six patients that died later, all had some degree of respiratory disorder (5).

Principal radiation-induced normal tissue events in the lung are an acute phase alveolitis/pneumonitis and a late chronic pulmonary fibrosis (6), with multiple lung-related symptoms being expressed as part of injury progression. Following the initiating radiation injury, progression to late lung disease involves a complex network of cellular processes and interacting signals (7). During the acute phase, inflammation is the predominant histological and physiological feature (6), culminating in the pneumonitic and/or fibrotic phases (7). Classically, prevention of normal tissue late effects has focused on pre-treatment, including free radical scavengers such as amifostine (8). However, this approach has failed to gain significant headway in the clinic, and pre-treatment protocols have limited relevance in a mass-casualty exposure scenario. Another widely accepted target for mitigating strategies for radiation-induced lung injury has been inflammation. Clinical evidence suggests that limiting the inflammatory reaction may be a reasonable approach, given beneficial effects seen following glucocorticosteroid drug administration (9, 10). However, in general, these approaches have a poor track record for decreasing long-term injury (11). General classes of drugs currently under study within NIAID's program for mitigation of radiation-induced lung damage include anti-oxidant drugs (12, 13), statins with anti-inflammatory properties² (14); oxidized glutathione variants (15), ACE inhibitors^3,4 (16, 17), and nutraceuticals⁵ (18, 19). Other novel approaches include drugs to enhance mucociliary clearance of inhaled radionuclides (20), and a substance P analog (21).

Animal Models and Development of Mitigators for Radiation-Induced Lung Damage

It is generally accepted that the FDA Animal Rule licensure pathway will require proof of efficacy in large-animals models as well as in rodents (22). For this reason, animal model development is being pursued in several species, including mice, rats, canines and non-human primates (NHPs). With regard to selection of strain, some researchers have noted that certain mouse strains may be more predictive of the human lung radiation response than other strains. For example, the C57BL/6J strain may not be the best model of lung injury due to an observed development of pleural effusions. This pathology, while not widely observed in humans, was noted in patients who received accidental radiotherapy overexposures (23). At issue is that these effusions are normally not the cause of death in humans (whose symptoms can be resolved with steroid treatment). However, they can be the cause of death in small animal models. An excellent overview on mouse strain differences in lung effects and potential animal model limitations to mimic human lung syndromes is available (24). Rats are also under study for radiation-induced lung injury, since their larger size allows for greater ease of manipulation, including achieving clearer computed tomography (CT) scanning. Because medical countermeasures will also be needed for use in children, development of pediatric mouse models of lung damage is being undertaken, and animal age appears to play an important role in the outcome of the exposure (25). In addition to rodent models of radiation exposure, larger animals, such as canines and NHPs, are being developed for potential use as models of the human pulmonary response⁶. Pig models may also be representative of radiation-induced damage to the human lung due to the structural similarity of pig lungs to those of humans (2). In any model selected, careful consideration of an end point (e.g., pneumonitis or fibrosis) should be made in the context of the chosen animal.

As part of a large-scale nuclear or radiological event, it is likely that the majority of victims will experience a total body (TBI), albeit heterogeneous, exposure. However, in the absence of supportive care, external TBI, delivered at levels necessary to cause late lung damage, would be acutely lethal. To more effectively model expected scenarios, several researchers are using high-dose TBI exposures, relying on bone marrow transplants to minimize mortality resulting from hematopoietic complications. Other researchers have adopted high-dose exposure protocols, in which only a part of the animal is irradiated (partial-body irradiation, PBI). A drawback with some of these models is the potential absence of hematopoietic and/or gastrointestinal damage, which might affect the progression of radiation-induced pulmonary damage. To address this issue, several models are under development in which animals either have 5% shielding of their bone marrow⁷, or are exposed to a sub-lethal TBI radiation exposure with an additional, localized “top-up,” high-dose radiation exposure to the thorax (26). Still other investigators are developing “two-hit” models, in which radiation exposure is delivered concomitantly with other stressors (such as skin burns or infection) as a combined injury (27).

Because government guidance dictates that medical countermeasure delivery from the stockpile will likely not occur within hours or even days after an incident, dose regimens beginning at times >24 h to 72 h post-exposure represent appropriate testing windows for drugs to treat radiation effects, and investigators are strongly encouraged to consider testing using even greater administration delays. Only drugs licensed by the FDA for any indication can be stockpiled. Drugs can be released as therapies for radiation injuries only if they are FDA-approved for a radiation indication or through issue of an emergency use authorization by the FDA (28). In either of these pathways, it is important that data be obtained in appropriate, Good Laboratory Practice (GLP)-compliant animal models, as per FDA's Animal Rule (3). Because the ultimate goal of NIAID's program is licensure of a drug to treat late radiation-induced lung complications following exposure during a radiation incident, efforts have focused on developing animal models that would be acceptable for licensure via FDA's Animal Rule pathway. This pathway is utilized for “approval of new drugs when human efficacy studies are not ethical or feasible” (22). In contrast to traditional drug development licensure, this pathway requires: (1) a reasonably well-understood mechanism of radiation damage; (2) that effects of radiation exposure and medical countermeasure mitigation are demonstrated in at least one animal species expected to react with a response predictive of humans; (3) animal study outcomes that are clearly related to the desired benefit in humans; and (4) data on pharmacokinetics (PK) and pharmacodynamics (PD) to allow selection of an effective dose in humans. It is critical to link together data obtained in each animal model and show that the selected models are appropriate representations of the damage observed in humans. It must also be shown that amelioration of damage by the drug in the model is predictive of the response observed in humans. For example, if the mechanism of action of a particular treatment in the animal model is through inhibition of a particular pathway in a cell or tissue, then that pathway must also exist in humans, must be affected similarly by radiation damage, and must be modified by the drug in the same fashion. Precisely how this requirement will be implemented by the FDA in practice is still unclear. Although the Animal Rule has been implemented in the U.S., it does not currently have an equivalent in most countries. However, the European Medicines Agency has issued guidelines on a regulatory procedure for the “Granting of a Marketing Authorization under Exceptional Circumstances” [Article 14 (8) of regulation (EC) NO 726/2004] that might be applicable in some countries (29, 30). This approval can be used when a sponsor is “unable to provide comprehensive data on the efficacy and safety under normal conditions of use”. Additional information on considerations for investigators pursuing licensure for a radiation indication (including who to contact at the FDA) is discussed elsewhere (31).

Whatever animals are selected for study, it is critical to standardize the models and validate them for their ability to detect medical countermeasure efficacy. Validation should include comparing proposed new treatments to current standards of supportive care, and perhaps testing the new treatment concurrent with administration of what would be expected care in a mass-casualty scenario. For example, dexamethasone is used clinically to treat lung injury (32), and in the case of radiation-induced damage, could be used off-label by a physician. Therefore, researchers might consider including this standard therapy concurrently, especially in larger animal models of radiation lung damage. Whether supportive care is needed, and to what extent, is especially important for certain models of radiation-induced lung damage (e.g., the 5% shielded model), since animals must survive multiple syndromes before manifesting lung injury. Only discussions with the FDA can delineate the animal care protocol that should be employed for development of a particular medical countermeasure. It is also important to precisely define radiation exposure parameters, and to ensure that the radiation dose delivered to the animals is accurate. Other factors, such as time of day of irradiation, anesthesia, use of acidified water, and food composition can also affect radiation responses. Because lung damage can result from internal contamination (from fallout) and/or external exposure (from prompt radiation), depending on the proposed mechanism of action of the medical countermeasure, it may be important to consider if a mitigators/treatment will work for lung injury caused by both routes of radiation exposure. Finally, it is critical to identify appropriate end points (primary and secondary) to assess efficacy of lung medical countermeasures.

FDA requires evaluation of end points that indicate clinical benefit, which for many indications is survival. Generally, an improvement in survival of 25–30% is desirable in a drug for which licensure for this indication is being sought. It is, however, important to note that use of a survival end point is sometimes problematic for researchers, especially in large animals, due to Institutional Animal Care and Use Committee concerns. Reduction of major morbidity by the medical countermeasure may represent another possible end point. To date, FDA's stance has been that “Primary study end points, which should be specifically discussed with the review division, generally are the enhancement of survival or prevention of major morbidity” (22). The dose response for these end points should be explored fully and established. Although secondary end points can provide useful information about the animal model, and the activity of the product as studied in the animal model, ordinarily, only primary end points can serve as the basis of approval. Determination of major morbidity in the lung is a difficult clinical end point, and improvements in “quality of life” of affected patients are challenging to quantify in humans, and even more so in animal models. Therefore, it is important to show clinically-relevant treatment outcomes if a survival end point is not possible. For example, human lung damage is often assessed by CT, so similar scans of irradiated animals might be warranted to relate those findings to a clinically-relevant outcome. Lung compliance and/or respiratory rate could also represent surrogate end points. A more thorough consideration of potential end points can be found elsewhere (26). Further studies are still needed to better understand the impact of age, gender and immune competency on lung sensitivity to radiation exposure, as well as responses to therapeutics. There is also still a need to further validate existing models for radiation damage to the lungs, and to better understand the mechanisms responsible for the radiation damage as well as cellular and organ responses to the injury. For this reason, NIAID plans to continue to fund studies in these areas.