Cough Capture as a Portal Into the Lung

Hypothesis: Cough can serve as a valid non-invasive surrogate for the lung for large molecules such as DNA and proteins, potentially for a variety of clinical and public health purposes.

Proposal in Brief: For development of an epidemiological and clinically-applicable platform for large molecules captured uniquely from deep lung, the investigator team will pursue a new approach that includes cough collection, and testing the efficacy and practicality of partitioning cough specimen for deep-lung specific extra-cellular vesicles (EVs). This cough specimen will be compared to that from invasively collected deep lung samples BAL/bronchial brushings, and to the potential contaminating mouthrinse, all from the same individuals. The study team initially proposes to examine these crude cough specimens, and then EV partitioned airway samples, for somatic mutations by SMM bulk sequencing for somatic mutations, state-of-art techniques developed in the Vijg/Maslov labs and for proteins known to be specific for lung, as opposed to oral cavity/saliva, in the Sidoli/proteomics core.

Specific Aims:

1. Cough will be captured by aerochamber or equivalent/optimized device in 5 healthy volunteers, and in 5 current smokers, in parallel with mouthrinse and with BAL. Lung-specific EVs from these same samples via deep lung capture (Loudig lab) will be performed in parallel.
2. Cough will undergo DNA mutation analyses using single molecule mutation sequencing (SMM-seq).
3. Cough, mouthwash and BAL will undergo proteomic analysis using LC-MS in our Core facility.

Aims 1 and 2 will be studied in a cross-sectional manner, enrolling a convenience sample of PLWHA smokers attending the Montefiore Medical Center (MMC) Infectious Diseases (ID) Clinic. During Aim 1, group EBC DNA methylation and microRNA and metabolomics patterns will be determined by high resolution tagged bisulfite genomic sequencing of three tumor suppressor genes (DAPK, RASSF1A, PAX5B), according to smoking status (current, former, never), and lifelong dose (pack-years) in a sample of HIV-infected smokers. Aim 2 will pilot the comparison of EBC DNA methylation and microRNA and metabolomics patterns from HIV infected individuals to that from uninfected individuals, collected in other projects funded by other mechanisms. Aim 3 will employ a prospective design, collecting two specimens over time from a subset of the overall study sample to examine the stability of EBC DNA and microRNA and metabolomics patterns methylation patterns over time in HIV-infected smokers.

Approach:

Human Lung Studies protocol practices, in general: Ongoing subject enrollment is part of a longstanding IRB-approved Einstein-Montefiore clinical case-control protocol (2007-407). Individuals from pulmonary practices eligible to enroll are: Age > 21 years old, any smoking status, myriad comorbid status except contraindications to BAL, planned for clinically-indicated lung collection procedure (bronchoscopy) as part of routine care. This amounts to >100-150 bronchoscopy and related former smoker donors recruits per year. Over 150-200 such pre-existing bronchoscopic sample sets, and attendant non-invasive EBC, along with MW, buccal brush, others] are banked and potentially available for the current study. All data/specimens are collected pre-diagnosis and pre-procedure; each subject provides EBC (and other non-invasives such as Cough) samples pre- bronchoscopy. As such, this active protocol minimizes (i) recall bias; and (ii) contamination of non-invasive specimens from the disruptions and spillage of the lung tissue inherent to these lung procedures. Each subject will be followed for 3 months in the medical record to gather updated diagnostics/testing, and minimize case-control misclassification. Extensive demographic, exposure and clinical data are available: (1) demographic information; (2) smoking details: (3) markers of underlying lung disease: (4) existing imaging (5) clinical information/pathological findings. For this ICTR pilot:

Cough will be captured in aerochamber (common, disposable handheld spacer device for asthma meter dose inhalers/pumps). The investigator team has found that it is an adequate cough (forced exhalation) capture device, pending further optimization. Under a planned amendment, the patient is instructed to cough every 30 seconds into the handheld plastic chamber, during an overall period of 10 minutes. All inside surfaces of the entire chamber is then rinsed with 2-4 ml of PBS, and the rinse/wash is snap frozen. The rinse is subsequently processed according to the analyte to be studied. For DNA, there is an Isopropanol/EtOH precipitation step, followed by resuspension in smaller volume, and extraction by Qiagen/Zymo DNA isolation kit. For protein, the crude extract in PBS is handled directly in the core by precipitation, drying, resuspension in buffer, and injection in the LC-MS platform for shotgun analyses.

The EV partitioning procedure is published. EV-CATCHER™ can be customized to target unique surface membrane proteins particular to the desired cell type and thus to direct capture of cell-specific EVs. The Loudig lab has compiled extensive data demonstrating, for example, that using antibodies targeting deep lung cell unique and specific proteins (e.g., Clara Cell Specific protein (CCSP), Alveolar type 2 cell Surfactant Protein-C (SFTPC)) can specifically purify EVs from Bronchoalveolar lavages (BAL) and exhaled breath condensates (EBC) and obtain similar miRNA expression profiles. Finally, by adding an enzymatically degradable uracylated double-stranded biotin-labeled DNA linker between immobilizing platform (polystyrene 96-well plate) and the activated antibody, the intact release of immunopurified EVs by incubation with uracil glycosylase (UNG) can be facilitated.

Approach: Aliquots from the 10 initial donors providing cough, mouthrinse, and BAL samples (5 current smokers, 5 never smokers) totaling 30 initial samples among all sample types for EV partitioning. Then, both whole-unpartitioned and EV-partitioned samples will be generated (totaling 60 samples) to be triaged to the two main assays, SMM for somatic mutation, and shotgun proteomics, below, both for feasibility demonstration, and pilot evaluation of surrogacy of cough for lung.

Aim 2: Cough will undergo DNA mutation analyses using SMM. Direct analysis of genome sequence integrity in normal cells. A key problem in studying genome sequence integrity in normal human tissues is the random nature of DNA mutations in the genome. While some mutations are clonally amplified, the vast majority are unique to each cell, and therefore rare and variant, which necessitates either single-cell whole genome amplification or in vitro clonal amplification assays. Both approaches have been associated with artifacts. Recently, the Maslov/Vijg lab developed reliable methods to quantitatively assess somatic mutations in single cells or in clonally amplified single cells. Using these methods it was demonstrated that base substitution mutations increase with smoking (and age) in proximal bronchial epithelial basal cells. A newer method for the quantitative analysis of somatic mutations, Single Molecule Mutation-seq (SMM-seq), is much more cost effective and can be applied to small amounts of bulk DNA. This latest SMM-seq method is significant because it allows, among other considerations, a practical means to accurately detect mutations in cytologically normal cells.

Approach: The 20 cough samples (10 unfractionated, 10 fractionated) will undergo DNA extraction, and library preparation and sequencing as described in the protocol. Somatic mutations among all the detected variants occurs by subtraction of germline polymorphisms found by analysis of regular sequencing library from DNA of the same individual. The number of somatic variants is normalized to the number of callable bases, i.e., bases with coverage more than 7X in SMM-seq libraries and with coverage more than 20X in regular libraries. Somatic mutation frequency will be expressed as a number of mutations per genome. The remaining 40 samples (20 from mouthwash and 20 from BAL) will be processed under other ongoing funding mechanisms, given budgetary constraints.

Aim 3: Cough, mouthwash, and BAL will undergo proteomic analysis using LC-MS in Einstein's Core facility. Prior proteomics attempts in cough do not exist to the investigator team's knowledge, perhaps due to technical limitations of low abundance material. The proteomics core (Sidoli, et al) will optimize a strategy to enhance sensitivity and speed for quantifying low abundance proteins/peptides. The sample preparation protocol will be modified to minimize sample loss and miniaturize nano-liquid chromatography coupled with state-of-the-art MS to quantify proteins/peptides in BAL, MW, and EBC specimens.

Approach: Sample preparation for proteomics: Cough specimen is dried to completion and then resuspended in 5 µL of 50 mM ammonium bicarbonate (pH = 8) for canonical proteomics sample preparation (20 ng of trypsin). The entire sample is then analyzed using a Dionex RSLC Ultimate 300 (Thermo Scientific) nano-liquid chromatographer coupled online with an Orbitrap Fusion Lumos (Thermo Scientific) mass spectrometer. For separation, the lab will utilize an in-house packed analytical column with minimal internal diameter (50 µm ID, 25 cm length) so that molecules with a \~100 nL/min flow-rate can be separated. The low flow-rate provides enhanced pre-concentration of the sample eluting from the column, leading to sub-attomole sensitivity.

Data analysis, including data transformation, is then performed using a pipeline in the Sidoli lab. Briefly, proteomics MS spectra are identified and quantified using Proteome Discoverer (v2.5, Thermo Scientific) with thresholds more stringent than currently recommended by the community for false discovery rate (\<1% FDR for spectrum matches, peptides and proteins). Quantitative values are then transformed and normalized as described by the Sidoli lab. Statistical regulation is assessed using parametric statistics (p\< 0.05). Chromatograms are manually inspected, given the relatively short list of candidates.

All cough samples (10 unfractionated, 10 fractionated x 3 specimen types=60 samples) will undergo proteomic analysis as described.

Cross-compartment comparisons: Determining cough surrogacy for lung can be performed in analogous fashion to that performed for EBC surrogacy for the lung. Within a given platform, the levels (and signature) of SMM mutation will be compared. In the recent past in the exhaled microRNA environment (not proposed here), this has been performed by Spearman correlations of microRNA-seq data clustering in principal components, PCA (not shown) and Spearman correlations (not shown). For the proteomics environment, this surrogacy of cough for lung can be shown for cough similarly by clustering heatmaps, PCA, and methods such as Spearman correlation.

Outcome: This initial pilot testing of cough as a new platform biospecimen is high risk, high reward pragmatic translational work. If this pilot project shows cough to be a valid surrogate for lung, on the two platforms examined, then that will provide strong preliminary evidence for supporting further federal funding. This potential for an airway based portal to the lung could be envisioned for a wide variety of acute and chronic clinical and public health purposes.