TY - JOUR
T1 - Unsupervised network mapping of commercially available immunoassay yields three distinct chronic rhinosinusitis endotypes
AU - Divekar, Rohit
AU - Rank, Matthew
AU - Squillace, Diane
AU - Kita, Hirohito
AU - Lal, Devyani
N1 - Funding Information:
This study was approved by the institutional research board at Mayo Clinic, Phoenix, AZ. Thirty-two patients undergoing sinonasal surgery with the senior author (D.L.) at Mayo Clinic, Phoenix, AZ, were prospectively enrolled into the study. Written informed consent was obtained from all subjects. Data was collected prospectively, including demographic information, history, nasal endoscopy findings (nasal polyp status; Lund-Kennedy score), clinical diagnoses, smoking status, asthma status, 22-item Sino-Nasal Outcome Test (SNOT-22) scores and sinus computed tomography (CT) scan score (Lund-Mackay [LM] stage). Subjects were defined as CRS (26 subjects) and non-CRS controls (n = 6). Control subjects included: 3 patients with nasal obstruction from deviated nasal septum and inferior turbinate hypertrophy; 1 patient with nasal septal deviation, sarcoidosis (no rhinologic manifestations), and recurrent acute sinusitis; 1 patient with isolated unilateral sphenoid sinus fungus ball; and 1 undergoing endoscopic dacryocystorhinostomy. Patients were classified into controls and those with CRS using 2007 American Academy of Otolaryngology?Head & Neck Surgery (AAO-HNS) guidelines. CRS patients were subclassified into CRSwNP (n = 13) and CRSsNP (n = 13) based on office nasal endoscopy. Surgical specimens were handled according to standards of clinical care. Additionally, sinonasal tissue defined as ?surgical waste? was collected at the time of surgery for the purposes of this research. In CRS patients, ethmoid tissue for was preferentially collected and ethmoid mucosal tissue was used to be consistent across groups. Mucosal tissue sample was taken either from the anterior ethmoidal complex or the posterior ethmoidal complex. The location of control tissues varied based on the surgery performed. Unilateral or bilateral tissue collection was done based on the type of surgery. The specimens were delivered directly from the operating room to the pathology department. After immediate macroscopic review, the tissue was sectioned into 2 halves. In 1 half, tissue histopathology was performed by a pathologist and reported as the standard-of-care pathology synoptic report. The other part was biobanked. The time from the start of harvest to freezing was approximately 10 to 15 minutes. The biobank samples were quickly frozen in 7-mL polycarbonate tubes (Sarstedt 71.9923.610) in a ?90?C bath of Novec engineered fluid (3M HFE-7000) cooled in a HistoChill freezing bath (SP Scientific HC80A0). No embedding medium or preservative was added to frozen tissues. Unique identification numbers were assigned to each individual container with bar code labels. Biospecimens were stored in ThermoFisher Ultra Low Temp upright freezers at ?80?C until retrieval for the cytokine analysis. Standardized histopathology reporting was performed as published by Snidvong, et?al. Tissue eosinophilia was graded based on numbers of eosinophils per high power field (hpf). The gradation was as follows: grade 1, <10 eosinophils per hpf; grade 2, 10 to 100 eosinophils per hpf; and grade 3, >100 eosinophils per hpf. The levels of cytokines and chemokines (41-plex) in the nasal tissue lysates were determined by a Millipore multiplex kit (Billerica, MA) and a Bio-Rad MAGPIX multiplex reader (Hercules, CA) following the procedure recommended by the manufacturers. In addition, the levels of eosinophil major basic protein (MBP) in the tissue lysates were analyzed by an in-house sandwich radioimmunoassay using the methodology published previously by Ohnuki et?al. Multiplex data was normalized and feature scaled between 0 and 1 to preserve the proportionality between the analytes and allow for comparisons. This transformed data was used to generate network modeling (using Gephi 0.8.2, graphic visualization platform) and to visualize data in an unsupervised manner and reveal association cytokine-chemokine (?analyte?) clusters and patient groups. Statistical comparisons for categorical data were performed with ?2-test or Fischer's exact test or test of comparison of proportions where applicable. Continuous data was analyzed using Wilcoxon-rank sum test. Two-sided p values <0.05 were considered significant.
Publisher Copyright:
© 2017 ARS-AAOA, LLC
PY - 2017/4/1
Y1 - 2017/4/1
N2 - Background: Endotyping chronic rhinosinusitis (CRS) through simplified cytokine assays may help direct individualized therapy such as corticosteroids, antibiotics, or biologics. We performed an unsupervised network analysis to endotype CRS and control subjects using a commercially available cytokine-chemokine immunoassay. Methods: A 41-plex cytokine-chemokine array along with major basic protein (MBP) assay was performed on sinonasal surgical tissue of 32 adults. Subjects were defined as non-CRS controls (n = 6), CRS with nasal polyps (CRSwNP; n = 13), and CRS without nasal polyps (CRSsNP; n = 13). Unsupervised network modeling was performed to reveal association cytokine-chemokine (“analyte”) clusters and “subject” groups. Results: Network mapping and unsupervised clustering revealed 3 analyte clusters and 3 subject groups. Analyte cluster-1 was composed of T helper 1 (Th1)/Th17 type markers, analyte cluster-2 Th2 markers, and analyte cluster-3 chemokines (CC) and growth factors (GF). Subject group-1 was devoid of CRSwNP, had fewer asthmatics, and was associated most strongly with analyte cluster-3 (CC/GF) (p < 0.001). Subject group-2 was characterized with the most asthmatics (86%) and CRSwNP (100%) patients, and was associated with analyte cluster-2 (Th2; p < 0.001). Subject group-3 was associated with both analyte cluster-1 (Th1/Th17) and analyte cluster-3 (CC/GF) (p < 0.001), and had the highest proportion of CRSsNP patients (62.5%). Tissue levels of MBP, eosinophilia, and computed tomography (CT) scores were significantly higher in subject group-2 vs other groups (p ≤ 0.05). Conclusion: An unbiased network-mapping approach using a commercially available immunoassay kit reveals 3 distinct tissue cytokine-chemokine signatures that endotype CRS patients and controls. These signatures are prominent even in a limited number of patients, and may help formulate individualized therapy and optimize outcomes.
AB - Background: Endotyping chronic rhinosinusitis (CRS) through simplified cytokine assays may help direct individualized therapy such as corticosteroids, antibiotics, or biologics. We performed an unsupervised network analysis to endotype CRS and control subjects using a commercially available cytokine-chemokine immunoassay. Methods: A 41-plex cytokine-chemokine array along with major basic protein (MBP) assay was performed on sinonasal surgical tissue of 32 adults. Subjects were defined as non-CRS controls (n = 6), CRS with nasal polyps (CRSwNP; n = 13), and CRS without nasal polyps (CRSsNP; n = 13). Unsupervised network modeling was performed to reveal association cytokine-chemokine (“analyte”) clusters and “subject” groups. Results: Network mapping and unsupervised clustering revealed 3 analyte clusters and 3 subject groups. Analyte cluster-1 was composed of T helper 1 (Th1)/Th17 type markers, analyte cluster-2 Th2 markers, and analyte cluster-3 chemokines (CC) and growth factors (GF). Subject group-1 was devoid of CRSwNP, had fewer asthmatics, and was associated most strongly with analyte cluster-3 (CC/GF) (p < 0.001). Subject group-2 was characterized with the most asthmatics (86%) and CRSwNP (100%) patients, and was associated with analyte cluster-2 (Th2; p < 0.001). Subject group-3 was associated with both analyte cluster-1 (Th1/Th17) and analyte cluster-3 (CC/GF) (p < 0.001), and had the highest proportion of CRSsNP patients (62.5%). Tissue levels of MBP, eosinophilia, and computed tomography (CT) scores were significantly higher in subject group-2 vs other groups (p ≤ 0.05). Conclusion: An unbiased network-mapping approach using a commercially available immunoassay kit reveals 3 distinct tissue cytokine-chemokine signatures that endotype CRS patients and controls. These signatures are prominent even in a limited number of patients, and may help formulate individualized therapy and optimize outcomes.
KW - CRS
KW - chronic rhinosinusitis endotypes
KW - immunoassay
KW - network analysis
UR - http://www.scopus.com/inward/record.url?scp=85007607590&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85007607590&partnerID=8YFLogxK
U2 - 10.1002/alr.21904
DO - 10.1002/alr.21904
M3 - Article
C2 - 28042687
AN - SCOPUS:85007607590
SN - 2042-6976
VL - 7
SP - 373
EP - 379
JO - International Forum of Allergy and Rhinology
JF - International Forum of Allergy and Rhinology
IS - 4
ER -