Unique cytokine and chemokine patterns in bronchoalveolar lavage are associated with specific causative pathogen among HIV infected patients with pneumonia, in Medellin, Colombia

作者：Keynan Yoav^*; Rueda Zulma V; Aguilar Yudy; Trajtman Adriana; Velez Lazaro A

来源：Cytokine, 2015, 73(2): 295-301.

DOI：10.1016/j.cyto.2015.03.003

• Pneumocystis jirovecii
• Mycobacterium tuberculosis
• HIV
• Pneumonia
• Pro-inflammatory cytokines

We wanted to investigate the pro-inflammatory cytokine/chemokine profile associated with the etiological agents identified in HIV patients. Immunosuppressed patients admitted to two hospitals in Medellin, Colombia, with clinical and radiographic diagnosis of pneumonia were enrolled in the study. After consent, bronchoalveolar lavage (SAL) was collected for bacterial, mycobacterial and fungal diagnosis. All patients were followed for a year. A stored BAL sample was used for cytokine/chemokine detection and measurement using commercial, magnetic human cytokine bead-based 19-plex assays. Statistical analysis was performed by assigning cytokine/chemokine concentrations levels into <25 percentile (lower), 25-75 percentile (normal) and >75 percentile (higher). Principal component analysis (PCA) and Kruskal-Wallis analysis were conducted to identify the clustering of cytokines with the various infectious etiologies (fungi, Mycobacterium tuberculosis - MTB, and bacteria). Average age of patients was 35, of whom 77% were male, and the median CD4 count of 33 cells/mu l. Of the 57 HIV infected patients, in-hospital mortality was 12.3% and 33% died within a year of follow up. The PCA revealed increased IL-10, IL-12, IL-13, IL-17, Eotaxin, GCSF, MIP-1 ot, and MIP-13 concentrations to be associated with MTB infection. In patients with proven fungal infection, low concentrations of IL-1RA, 1L-8, TNF-alpha and VEGF were identified. Bacterial infections displayed a distinct cytokine pattern and were not misclassified using the MTB or fungi cytokine patterns (p-value <0.0001). Our results indicate a unique pattern of pro-inflammatory cytokine/chemokine, allowing differentiation between bacterial and non-bacterial pathogens. Moreover, we found distinct, if imperfectly discriminatory, cytokine/chemokine patterns associated with MTB and fungal infections.

• 出版日期2015-6