"In Litero" Screening: Retrospective Evaluation of Clinical Evidence to Establish a Reference List of Human Chemical Respiratory Sensitizers

Abstract

Despite decades of investigation, test methods to identify respiratory sensitizers remain an unmet regulatory need. In order to support the evaluation of New Approach Methodologies in development, we sought to establish a reference set of low molecular weight respiratory sensitizers based on case reports of occupational asthma. In this context, we have developed an "in litero" approach to identify cases of low molecular weight chemical exposures leading to respiratory sensitization in clinical literature. We utilized the EPA-developed Abstract Sifter literature review tool to maximize the retrieval of publications relevant to respiratory effects in humans for each chemical in a list of chemicals suspected of inducing respiratory sensitization. The literature retrieved for each of these candidate chemicals was sifted to identify relevant case reports and studies, and then evaluated by applying defined selection criteria. Clinical diagnostic criteria were defined around exposure history, respiratory effects, and specific immune response to conclusively demonstrate occupational asthma as a result of sensitization, rather than irritation. This approach successfully identified 28 chemicals that can be considered as human respiratory sensitizers and used to evaluate the performance of NAMs as part of a weight of evidence approach to identify novel respiratory sensitizers. Further, these results have immediate implications for the development and refinement of predictive tools to distinguish between skin and respiratory sensitizers. A comparison of the protein binding mechanisms of our identified "in litero" clinical respiratory sensitizers shows that acylation is a prevalent protein binding mechanism, in contrast to Michael addition and Schiff base formation common to skin sensitizers. Overall, this approach provides an exemplary method to evaluate and apply human data as part of the weight of evidence when establishing reference chemical lists.

Keywords: adverse outcome pathway; allergic asthma; chemical allergens; clinical reference list; new approach methodologies; occupational asthma; respiratory sensitisation; respiratory sensitization.

FIGURE 1

Clinical diagnosis of Allergic Asthma. The clinical inquiry following presentation of a patient suspected of chemical allergy is shown. If the symptoms occur following exposure in the workplace, the clinician may pursue the identification of any materials in the workplace suspected to be causative. If enough evidence can be collected, higher tier testing demonstrating an immune-mediated respiratory response is consistent with respiratory sensitization (Friedman-Jiménez et al., 2000; Beckett, 2008; Malo et al., 2015).

FIGURE 2

(A) Clinical diagnostic tests aligning with Key Events in the AOP for respiratory sensitization from LMW chemicals (Beckett, 2008; Sullivan et al., 2017). (B) Evidence in case reports that point to Key Events. Chemicals for which case reports included evidence of exposure, immune involvement either through direct detection of specific IgE or IgG, or indirectly through a skin prick test, as well as demonstrated airflow obstruction were determined to be clinical respiratory sensitizers. Representative chemicals from clinical respiratory sensitizer (glutaraldehyde, piperazine) and equivocal (‘glycyl compound’) categories are shown with clinical evidence leading to category decision.

FIGURE 3

Generalized “In Litero” screening approach. In order to identify clinical literature, Abstract Sifter was used to collect and “sift” the most relevant results to the top. Reports referring to clinical cases of occupational asthma associated with each chemical were tabulated and annotated. Chemicals for which relevant reports could not be identified were categorized as “No Information” and excluded from evaluation. The reports for the remaining chemicals were categorized according to the established evaluation criteria as “Clinical Respiratory Sensitizers” or “Equivocal”.

FIGURE 4

Sankey Diagram of “in litero” screening results for all putative respiratory sensitizers and related compounds. From the original review of 118 putative sensitizers, 24 clinical respiratory sensitizers (Clinical RS) were identified. An additional automated search for similarly reported chemicals identified 240 chemicals, resulting in the identification of 4 more clinical respiratory sensitizers. For nine of the total 28 clinical respiratory sensitizers identified, we found case reports demonstrating respiratory sensitization in greater than ten total patients (High N). For the remaining 19, at least one case of clinical respiratory sensitization was identified (Low N). Inconclusive (Equivocal) evidence suggestive of clinical respiratory sensitization was found for an additional 153 chemicals.

FIGURE 5

Relative distribution of occupational sector of (A) all evaluated chemicals compared with (B) the identified clinical respiratory sensitizers (Clinical RS). A small percentage (2—10%) of chemicals in each market sector represented in our investigation were found to be clinical respiratory sensitizers. Low N: 10 or fewer cases; High N: greater than 10 cases. Pharm/Med: pharmaceutical and medical. Food/Cosm: food and cosmetic. Biocides: insecticide, fungicide, antimicrobials, and other pesticides. Industrial: industrial materials, especially polymerizing agents. Not found: a market sector of use could not be identified for this compound in available public databases.

FIGURE 6

Relative distribution of protein binding mechanisms of (A) all evaluated chemicals compared with (B) the identified clinical respiratory sensitizers. Protein binding mechanisms were predicted using the OECD QSAR Toolbox. Chemicals which had no protein binding alerts except when predicted metabolites or oxidation products were profiled are included as “Pre/Pro-hapten.”