Portal de Periódicos da CAPES

Elevated regulatory T cells at diagnosis of Coccidioides infection associates with chronicity in pediatric patients

2018; Elsevier BV; Volume: 142; Issue: 6 Linguagem: Inglês

10.1016/j.jaci.2018.10.022

1097-6825

Dan Davini, Fouzia Naeem, Aron Phong, Mufadhal Al-Kuhlani, Kristen Valentine, James McCarty, David M. Ojcius, David M. Gravano, Katrina K. Hoyer,

Immune Response and Inflammation

Coccidioidomycosis, also known as Valley fever, is caused by a fungal pathogen endemic to the San Joaquin Valley in California, Arizona, northern Mexico, and arid areas of South America.1Johnson L. Gaab E.M. Sanchez J. Bui P.Q. Nobile C.J. Hoyer K.K. et al.Valley fever: danger lurking in a dust cloud.Microbes Infect. 2014; 16: 591-600Crossref PubMed Scopus (25) Google Scholar Early mouse studies indicate a need for TH1 responses in protective immunity to Coccidioides.2Hung C.Y. Castro-Lopez N. Cole G.T. Card9- and MyD88-mediated gamma interferon and nitric oxide production is essential for resistance to subcutaneous Coccidioides posadasii infection.Infect Immun. 2016; 84: 1166-1175Crossref PubMed Scopus (26) Google Scholar More recent studies highlight a role for TH17 cells and IL-17 cytokine responses.3Wuthrich M. Gern B. Hung C.Y. Ersland K. Rocco N. Pick-Jacobs J. et al.Vaccine-induced protection against 3 systemic mycoses endemic to North America requires Th17 cells in mice.J Clin Invest. 2011; 121: 554-568Crossref PubMed Scopus (181) Google Scholar Effective TH1 responses have been linked to resolution of human Coccidioides infection4Duplessis C.A. Tilley D. Bavaro M. Hale B. Holland S.M. Two cases illustrating successful adjunctive interferon-gamma immunotherapy in refractory disseminated coccidioidomycosis.J Infect. 2011; 63: 223-228Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar; however, TH17 effectors are critical for clearance of many other fungal pathogens. An appropriate balance in T-effector and regulatory T (Treg) cells allows effective immune responses while limiting tissue damage. TH17 and Treg cell differentiation are inversely regulated, with IL-6 promoting TGF-β–induction of TH17 cells and inhibiting TGF-β–induction of Treg cells.5Bettelli E. Carrier Y. Gao W. Korn T. Strom T.B. Oukka M. et al.Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells.Nature. 2006; 441: 235-238Crossref PubMed Scopus (5557) Google Scholar In mice infected with Coccidioides posadasii, Treg cell expansion correlated with reduced survival.6Gonzalez A. Hung C.Y. Cole G.T. Absence of phagocyte NADPH oxidase 2 leads to severe inflammatory response in lungs of mice infected with Coccidioides.Microb Pathog. 2011; 51: 432-441Crossref PubMed Scopus (15) Google Scholar Thirty pediatric patients with a diagnosis of coccidioidomycosis (15 inpatients and 15 outpatients) and 20 healthy controls were enrolled. Blood was collected from patients at the time of Coccidioides diagnosis, during acute disease stage when disease outcome was unknown. Of the 30 patients with coccidioidomycosis, 25 had pulmonary involvement and 5 developed disseminated disease including 1 patient who died because of disease severity and was excluded from analysis. Resolved disease was observed in 13 (44.8%) patients and persistent disease in 16 (55.2%), including 12 with pulmonary and 4 with disseminated disease. Age and ethnicity, as well as the incidence of chest pain, night sweats, weight loss, fever, and fatigue symptoms, and serum antibody titers and isotypes were similar between the groups, and did not correlate with disease outcome (see Table E1, Table E2, Table E3 in this article's Online Repository at www.jacionline.org). As observed previously, erythema nodosum mildly correlated with better disease prognosis (P = .0722). On evaluating complete blood cell count measurements, we found that inpatients had higher neutrophil, eosinophil, white blood cell, and platelet numbers, and lower lymphocyte numbers compared with outpatients (see Fig E1, A, in this article's Online Repository at www.jacionline.org). Eosinophil, platelet, and creatinine levels were elevated and lymphocyte counts reduced in patients with persistent disease compared with patients who resolved infection (Fig E1, B). Analysis of immune parameters based on hospital status revealed several differences between inpatients and outpatients in both innate and adaptive responses. Plasmacytoid dendritic-cell frequency was reduced in inpatients relative to healthy controls and outpatients (Fig E1, C). CD4+, CD8+, and B-cell frequency was reduced in inpatients, and mild elevation of TH1 was observed (Fig E1, D and E). These differences likely reflect the more severe illness, elevated inflammatory responses, and ineffective adaptive and TH-cell immunity of the inpatients. A larger proportion of inpatients versus outpatients were unable to resolve Coccidioides infection, including patients with dissemination, suggesting that the more severe disease at diagnosis, the more likely that disease will persist. No differences were observed in the frequency or total number of innate immune populations based on disease outcome (not shown). Patients who developed persistent disease tended to have lower adaptive immune cell frequency at diagnosis, in particular a significant reduction in B-cell frequency relative to healthy controls and patients who resolved infection (Fig 1, A). Current paradigm suggests that a strong TH1 response is required for controlling Coccidioides infection. TH1 frequencies were similar between patients with resolved and persistent disease (Fig 1, B), whereas patients with persistent disease had lower TH17 and higher Treg cell frequencies than did patients with resolved disease. Immune cells secrete inflammatory and effector cytokines to induce cell migration, differentiation, and function. We evaluated the concentration of 26 inflammatory and TH cytokines in serum collected at diagnosis (Fig 1, C); only cytokines above the limit of detection are shown. Significant differences in IL-6, IL-18, and IL-12, and mild difference in IL-10, were observed between patients with resolved and persistent disease, and relative to healthy controls (Fig 1, D). IL-6, IL-18, and IL-12 are produced by antigen-presenting cells and promote T-effector differentiation. IL-10, an immunosuppressive cytokine, is expressed by several immune populations, including Treg cells. IL-18 inhibits TH17 and enhances Treg cell differentiation and function,7Harrison O.J. Srinivasan N. Pott J. Schiering C. Krausgruber T. Ilott N.E. et al.Epithelial-derived IL-18 regulates Th17 cell differentiation and Foxp3(+) Treg cell function in the intestine.Mucosal Immunol. 2015; 8: 1226-1236Crossref PubMed Scopus (131) Google Scholar which could explain the Treg cell expansion and persistent infection in these patients. Next, chi-square automatic interaction detection, a decision tree technique, was used to define clinical and immune parameters that associate with disease outcome. Based on this analysis, Treg cell frequency had the greatest impact on disease outcome of all clinical and immune parameters (Fig 2, A). Treg cell frequency at diagnosis distinguished disease outcome in 11 of 13 patients with resolved disease and 12 of 16 patients with persistent disease. Eosinophil, neutrophil, and TH17 numbers further defined patients within these disease outcomes (not shown). Principle-component analysis using the parameters identified by chi-square automatic interaction detection in the first 2 levels of separation (%CD4+Treg, eosinophils, and neutrophils) associated with disease outcome in 26 of 29 (89.7%) patients (Fig 2, B). We applied high-dimensional flow cytometry analysis tools to identify cellular populations associated with clinical outcome. Comparing patients with resolved versus persistent disease, CITRUS identified several cell populations that stratified the data, including a population representing Treg cells (Fig 2, C). CITRUS clusters revealed CCR5 as expressed more highly in patients with persistent disease; specifically, elevated CCR5 stratified Treg cells. CCR5+ Treg cells, known to have enhanced suppressive capacity,8Soler D.C. Sugiyama H. Young A.B. Massari J.V. McCormick T.S. Cooper K.D. Psoriasis patients exhibit impairment of the high potency CCR5(+) T regulatory cell subset.Clin Immunol. 2013; 149: 111-118Crossref PubMed Scopus (42) Google Scholar had higher frequency in patients with persistent disease than in those who resolved infection (Fig 2, D). On performing viSNE, a dimensionality reduction and data visualization tool, we confirmed a region of Treg cells that could be divided on CCR5 expression (Fig 2, E). CCR5+ Treg cells were significantly elevated in patients with persistent disease as compared with patients who resolved infection, whereas CCR5neg Treg cell frequency was unchanged (Fig 2, E). High-dimensional analysis confirms a Treg cell disparity based on disease outcome, and reveals CCR5 as a functional marker associated with chronicity. CCR5-deficient mice have elevated TH17 response to Histoplasma infection, resulting in a Treg cell/TH17-cell imbalance.9Kroetz D.N. Deepe Jr., G.S. CCR5 dictates the equilibrium of proinflammatory IL-17+ and regulatory Foxp3+ T cells in fungal infection.J Immunol. 2010; 184: 5224-5231Crossref PubMed Scopus (59) Google Scholar We speculate that Treg cell expansion and elevated functionality prevent appropriate TH17 effector responses allowing persistent Coccidioides infection. Overall, the data suggest that pediatric patients with persistent disease develop an inappropriately prolonged innate immune response and ineffective adaptive immune activation to Coccidioides during acute infection. Expanded neutrophils and eosinophils indicate that the innate immune system in these patients mounts an inflammatory response, in contrast to reduced lymphocyte frequencies. Because lymphocytes and appropriate CD4+ T-cell responses are required for immune control of Coccidioides, patients with persistent infection likely do not have sufficient activated lymphocyte numbers to control infection. Resolution of infection appears to require a TH17, but not a TH1, response, which is reduced in patients with persistent infection. Few studies on host response to Coccidioides infection include pediatric patients. We identified Treg cell frequency, eosinophil numbers, neutrophil numbers, and TH17 frequency as correlates of clinical outcome in Coccidioides-infected children. Elevated Treg cell frequency provided the strongest association with clinical disease outcomes at the time of diagnosis, suggesting that Treg cells may suppress effective immune responses during acute Coccidioides infection. Alternatively, because Treg cells and TH17 cells differentiate through reciprocal pathways, Treg cells may be generated at the expense of TH17 cells required for fungal clearance, resulting in persistent infection.5Bettelli E. Carrier Y. Gao W. Korn T. Strom T.B. Oukka M. et al.Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells.Nature. 2006; 441: 235-238Crossref PubMed Scopus (5557) Google Scholar Defining host immune responses and clinical parameters that indicate persistent infection may allow clinicians to identify patients earlier in their disease course who would benefit from prolonged treatment course, and close monitoring for disease relapse once off therapy. These parameters warrant further testing and evaluation in predictive studies of children and adults infected with Coccidioides. Treg cell frequency is described as increasing with age; thus, this association with clinical outcome may be more pronounced in older patients. We thank Padma Desai and Christine Banda for clinical enrollment support, our patients for participating in the study, Dr Larry Fong for helpful discussions, the UC Merced Stem Cell Instrumentation Foundry for assistance in flow cytometry design, and the UC Merced Biostatistics and Data Support Core for consultation on statistical methods. Patients were enrolled and samples collected at Valley Children's Healthcare, a 355-bed Children's Hospital serving as tertiary referral center for 10 counties in central California. Thirty children aged 2 to 18 years with coccidioidomycosis diagnosed by positive coccidioidal serology or cultures demonstrating Coccidioides species were included. Of these, 15 were enrolled as inpatients and 15 as outpatients at the time of diagnosis. Inpatients were hospitalized at the discretion of the attending physician on the basis of the severity of illness and were then diagnosed with Valley fever and enrolled in the study. Twenty healthy siblings of hospital patients, with negative coccidioidal serology, were enrolled as controls. Children known to be pregnant, immunocompromised, and/or on immunosuppressive medications, with severe underlying illness, cystic fibrosis, or inflammatory diseases, were excluded. The study was approved by the Valley Children's Healthcare institutional review board. Written informed consent was obtained from legal guardians and participants older than 7 years. Baseline demographic, clinical, laboratory, radiographic, antifungal treatment, and outcome data were collected. Outcomes were defined as resolved or persistent infection at the time of study conclusion. Enrollment occurred over the course of 2 years, with study follow-up ranging from 4 to 24 months postenrollment. Resolution of symptoms and normal radiographic and clinical laboratory findings defined disease resolution. Persistent disease was defined as still on therapy without complete resolution of symptoms and/or persistent abnormal radiographic or laboratory findings. Patients were characterized by 42 clinical parameters (Table E4), 17 immune cell populations (Table E5), and 26 serum proteins. Frequency and total number per milliliter of innate and adaptive immune populations were evaluated in the peripheral blood at the time of enrollment. We compared percentages, total number and activation state of immune cells in patients based on disease outcome (healthy control, resolved, persistent), and hospital status at time of blood draw (inpatient, outpatient). One milliliter of whole blood collected in sodium heparin tubes was stored at room temperature and processed within 24 hours for flow cytometry. One milliliter collected in red-cap plug tubes was centrifuged at 3500 rpm for 15 minutes and then transferred into new tubes for storage at −20°C. Antibody staining was performed in PBS/2% FBS. Hundred microliter of whole blood was incubated with 5 μL of Human TruStain FcX (Biolegend, San Diego, Calif) for 5 minutes before staining. Three flow cytometry panels were designed to profile PBMCs (see flow cytometry panels below). A single blood sample was processed 3 times such that 50 μL of antibody for each panel was added and incubated for 20 minutes. Red blood cells were lysed in 1-step Fix/Lyse Solution (eBioscience, San Diego, Calif) for 25 minutes and resuspended with 300 μL PBS/2% FBS and 100 μL of 123count-eBeads (eBioscience) for flow cytometry analysis. Stained PBMCs were acquired on an LSRII (BD Biosciences, San Jose, Calif), and analyzed using FCS Express (DeNovo Software) FACS Diva (BD Biosciences). Three flow cytometry panels were designed to profile PBMCs. The "general leukocyte panel" contained the antibodies CD4 FITC (RPA-T4, Biolegend), CD3 APC (UCHT1, Biolegend), CD8a PE-Cy7 (RPA-T8, Biolegend), CD20 eflour450 (2H7, eBioscience), CD25 PE (M-A251, Biolegend), CD34 PE-eFluor 610 (4H11, eBioscience), CD45 PerCP-Cy5.5 (HI30, Biolegend), CD56 APC-Cy7 (HCD56, Biolegend), HLA-DR BUV395 (G46-6, BD Biosciences), and LIVE eflour506 (eBioscience). The "T-effector panel" consisted of the following panel of antibodies: CD4 FITC (RPA-T4, Biolegend), CD25 APC (M-A251, Biolegend), CD45 PerCP-Cy5.5 (HI30, Biolegend), CD127 eflour450 (eBioRDR5, eBioscience), CD185 PE-Cy7 (J252D4, Biolegend), CD193 PE (eBio5E8-G9-B4, eBioscience), CD195 APC-Cy7 (J418F1, Biolegend), CD196 PE-eFluor 610 (R6H1, eBioscience), HLA-DR BUV395 (G46-6, BD Biosciences), and Fixable Viability Dye eflour506 (eBioscience). The "innate immune panel" included CD57 FITC (TB01, eBioscience), CD3 APC (UCHT1, Biolegend), CD11C PE-eFluor 610 (3.9, eBioscience), CD14 PerCP-Cy5.5 (M5E2, Biolegend), CD16 PE-Cy7 (3G8, Biolegend), CD20 eflour450 (2H7, eBioscience), CD56 APC-Cy7 (HCD56, Biolegend), CD123 PE (6H6, eBioscience), HLA-DR BUV395 (G46-6, BD Biosciences), and Fixable Viability Dye eflour506 (eBioscience). All antibodies were purchased from eBioscience, Biolegend, or BD Biosciences as noted. To ensure consistent cytometer calibration and data collection, voltage and compensation parameters were standardized using SPHERO Ultra Rainbow Calibration Particle Kits (Spherotech, Inc, Lake Forest, Ill). Consistent MFI (mean fluorescence intensity) for each fluorophore was maintained by small adjustments to voltage parameters if baseline MFI changes were more than 10% between experiments. Fluorophore compensation was maintained with antibody single stains mixed from 1 μL of each antibody into UltraComp eBeads (eBioscience), washed with PBS/2% FBS, incubated for 15 minutes, and resuspended in 500 μL PBS/2% FBS. FCS files were loaded into Cytobank (www.cytobank.org), and pregated to remove counting beads, debris, doublets, and dead cells. Remaining cells were gated on CD4+CD45+, grouped on outcome, and analyzed using CITRUS and viSNE on markers CD127 (IL-7Rα), CXCR5 (CD185), CCR3 (CD193), CCR5 (CD195), CCR6 (CD196), CD25 (IL-2Rα), and HLA-DR to cluster the data.E1Bruggner R.V. Bodenmiller B. Dill D.L. Tibshirani R.J. Nolan G.P. Automated identification of stratifying signatures in cellular subpopulations.Proc Natl Acad Sci U S A. 2014; 111: E2770-E2777Crossref PubMed Scopus (291) Google Scholar, E2Amir e.l.-A.D. Davis K.L. Tadmor M.D. Simonds E.F. Levine J.H. Bendall S.C. et al.viSNE enables visualization of high dimensional single-cell data and reveals phenotypic heterogeneity of leukemia.Nat Biotechnol. 2013; 31: 545Crossref PubMed Scopus (1098) Google Scholar, E3Liu W. Putnam A.L. Xu-Yu Z. Szot G.L. Lee M.R. Zhu S. et al.CD127 expression inversely correlates with FoxP3 and suppressive function of human CD4+ Treg cells.J Exp Med. 2006; 203: 1701-1711Crossref PubMed Scopus (2081) Google Scholar File internal compensation was used, clusters characterized on abundance, event sampling equalized per file, minimum cluster size set to 0.5%, and significance analysis of microarrays correlative model used. CITRUS analysis was repeated 3 independent times to confirm stratifying signatures and statistically significant populations (discovery rate, <1%). viSNE map was generated from 373,781 total events (12,889/patient) using 2,500 iterations, a perplexity of 90, and a Theta of 0.3. Thawed serum was centrifuged at 1000g for 5 minutes to remove particulates. Serum cytokine concentrations were determined using LEGENDplex Human 13-plex kits, "Th Cytokine Panel," and "Cytokine Panel 2" bead-based immunoassay kits (Biolegend) in duplicate according to the manufacturer's instructions. Samples were diluted 2-fold with kit Assay Buffer before assay initiation. Samples were analyzed by flow cytometery and processed using LEGENDplex Data Analysis Software. Standard curves were generated using a 5-parameter curve-fitting model, and cytokine levels were calculated as the average of the duplicate measurements. Leukocyte subset total numbers were calculated using counting beads in combination with each leukocyte percentage determined by flow cytometry. Statistical analyses were performed using Prism software v6.0 (GraphPad Software, La Jolla, Calif). One-way ANOVA with Bonferroni correction was used to compare multiple groups with a 95% CI. The Fisher exact test was used to compare the distribution of groups and percentages. Principle-component analysis and chi-square automatic interaction detection analysis were performed using XLSTAT .Table E1Demographic characteristics of patients based on outcome statusCharacteristicHealthy controls (n = 20)Resolved (n = 13)Persistent (n = 16)Age (y), median (range)11 (3-16)12 (6-18)13 (2-17)Sex, n (%) Male11 (54)3 (23.1)5 (31.3) Female9 (45)10 (76.9)11 (68.8)Ethnicity, n (%) Hispanics14 (70)11 (84.6)14 (87.5) Non-Hispanics6 (30)2 (15.4)2 (12.5)Hospitalization status, n (%) Inpatients02 (15.4)12 (75.0) Outpatients20 (100)11 (84.6)4 (25.0) Open table in a new tab Table E2Clinical features of pediatric Coccidioides-infected patientsClinical featureResolved (n = 13)Persistent (n = 16)Fever, n (%)11 (94.6)13 (18.8)Cough, n (%)7 (53.8)10 (62.5)Chest pains, n (%)5 (38.5)8 (50.0)Night sweats, n (%)3 (23.1)5 (31.3)Fatigue, n (%)4 (30.8)5 (31.3)Weight loss, n (%)1 (7.7)1 (6.3)Erythema nodosum, n (%)7 (53.8)3 (18.8)No. of clinical parameters per patient, mean ± SD2.9 ± 1.22.8 ± 1.4No. of medications per patient, n (%) 01 (7.7)0 112 (92.3)9 (56.3) 204 (25.0) 303 (18.8)First medication, n (%) Fluconazole12 (92.3)11 (68.6) Ambisome05 (31.3)X-ray, n (%) Infiltrates6 (46.2)9 (56.3) Consolidation4 (30.8)10 (62.5) Adenopathy1 (7.7)4 (25.0) Not tested02 (12.5) Open table in a new tab Table E3Coccidioides antibody assaysAntibody assayHealthy controls (n = 20)Resolved (n = 13)Persistent (n = 16)Enzyme immunoassay, n (%) Negative20 (100)00 IgM only001 (6.3) IgG only002 (12.5) IgM & IgG07 (53.8)13 (81.3) Not tested060Serum immunodiffusion, n (%) NegativeNA00 IgM onlyNA4 (30.8)5 (31.3) IgG onlyNA1 (7.7)2 (12.5) IgM & IgGNA8 (61.5)8 (50.0) Not tested2001Serum complement fixation, n (%) NegativeNA5 (38.5)7 (43.8) PositiveNA8 (61.5)8 (50.0) Not tested2001 Highest antibody titer during treatment, median (range)NA6 (0-128)12 (0-128)CSF, Cerebrospinal fluid; NA, not applicable/available. Open table in a new tab Table E4Clinical parameters evaluatedMedical testMeasurementsCoccidioidal antibody assaysEnzyme immunoassay, serum and cerebrospinal fluid immunodiffusion, serum and cerebrospinal fluid complement fixation, highest complement fixation titer during studyExam questionsFever, cough, chest pain, night sweats, fatigue, weight loss, erythema nodosumTreatment and evaluationDate of diagnosis, date of resolution or stabilization, complications, antifungal prescribed, date of medication changeComplete blood cell countHemoglobin, hematocrit, white blood cell, platelets, neutrophils, lymphocytes, eosinophilsBlood chemistry panels for kidney and liver functionCreatinine, blood urea nitrogen, alanine amino transferase, aspartame amino transferase, total and conjugated bilirubinChest x-ray or CT scan as neededInfiltrates, consolidation, pulmonary nodules, adenopathyCT, Computed tomography. Open table in a new tab Table E5Immune populations and antibodies used for identificationImmune populationIdentifying flow markersClassical monocytesCD3−, CD20−, CD14+, CD16low/intermediateNonclassical monocytesCD3−, CD20−, CD14+, CD16hiConventional/myeloid dendritic cellsCD3−, CD20−, CD14+, CD56−, CD11c+, HLA-DR+Plasmacytoid dendritic cellsCD20+, CD11c+/intermediateNK cellsCD20−, CD56+NK T cellsCD3+, CD20−, CD56+NK CD4 T cellsCD3+, CD20−, CD56+ CD4+, CD8a−NK CD8 T cellsCD3+, CD20−, CD56+ CD4−, CD8a+CD3 T cellsCD3+, CD45+CD4 T cellsCD3+, CD45+, CD4+, CD8a−CD8 T cellsCD3+, CD45+, CD4−, CD8a+B cellsCD20+CD4 TH1CD45+, CD4+, CD195+, HLA-DR−CD4 TH2CD45+, CD4+, CD193+, HLA-DR+CD4 TH17 (and TH22)CD45+, CD4+, CD196+, HLA-DR−CD4 TfhCD45+, CD4+, CD185+, HLA-DR+CD4 TregCD45+, CD4+, CD25+, CD127lowNK, Natural killer. Open table in a new tab CSF, Cerebrospinal fluid; NA, not applicable/available. CT, Computed tomography. NK, Natural killer.