Diabetes alters immune response patterns to acute melioidosis in humans

Abstract Diabetes mellitus (DM) is a serious global health problem currently affecting over 450 million people worldwide. Defining its interaction with major global infections is an international public health priority. Melioidosis is caused by Burkholderia pseudomallei, an exemplar pathogen for studying intracellular bacterial infection in the context of DM due to the 12‐fold increased risk in this group. We characterized immune correlates of survival in peripheral blood of acute melioidosis patients with and without DM and highlight different immune response patterns. We demonstrate the importance of circulating NK cells and show that CX3CR1 expression on lymphocytes is a novel correlate of survival from acute melioidosis. Furthermore, excessive serum levels of IL‐15 and IL‐18BP contribute to poor outcome independent of DM comorbidity. CD8+ T cells and granzyme B expression in NK cells are important for survival of non‐DM patients, whereas high antibody titers against B. pseudomallei and double‐negative T cells are linked to survival of DM patients. Recall responses support a role of γδ T‐cell‐derived IFN‐γ in the establishment of protective immunity in the DM group. Defining the hallmarks of protection in people with DM is crucial for the design of new therapies and vaccines targeting this rapidly expanding risk group.

The aim of this study was to identify immune correlates of survival in acute human melioidosis and possible differences based on diabetes co-morbidity by analysing circulating immune cell populations and functional properties thereof.

Date
Experiments were performed between March 2016 and July 2017.

Conclusions
The study of circulating immune cells showed differences in immune pathways important for survival from acute melioidosis in patients with and without diabetes.

Quality control measures
The performance of the flow cytometer was analysed daily by running quality control beads according to the manufacturer's recommendations (for details see section 2.4.1). Appropriate fluorescence minus one (FMO) controls were used as quality control for software computed compensation and to set gate boundaries.

Sample Characteristics
Expected sample characteristics are the main circulating immune cell subsets (T cells, B cells, NK cells, NKT cells, monocytes, dendritic cells) typically found in peripheral blood. In case of ICCK the expected sample characteristic is IFN-γ secretion upon stimulation with cognate antigen.

Sample treatment Description
 For ex vivo phenotyping, cells were used immediately after thawing.
 For ICCK assays, cells were rested overnight and then cultured in the presence of bacterial antigen and co-stimulants for 6 hours. Media served as negative control and staphylococcal enterotoxin B as positive control for cytokine secretion. In some cases cells were concomitantly treated with cyclosporine A or vehicle control (DMSO). Brefeldin A was added for the last 4 hours of incubation to inhibit cytokine transport from the endoplasmatic reticulum to the Golgi apparatus.
 Multicolour flow cytometry staining: PBMC were resuspended in MACS buffer and incubated for 20min with near-infrared live/dead fixable stain and fluorochrome-conjugated primary human-specific antibodies in the presence of human FcR blocking reagent at 4⁰C. After washing with MACS buffer, cells were resuspended in IC fixation solution or subjected to intracellular staining. For the latter, cells were fixed with fixation/permeabilization solution for 20 min at 4⁰C , washed with permeabilization buffer followed by incubation with fluorochrome-conjugated humanspecific antibodies in the presence of FcR blocking reagent. After washing with permeabilization buffer the samples were resuspended in 1xPBS and acquired on a MACSQuant Analyzer 10 or stored at 4⁰C in the dark for up to 24 hours prior to acquisition. Data analysis was performed with FlowJo Version 10 (FlowJo, LLC, Oregon, USA) and specific gating strategies can be found in the Supplementary information.

Compensation
Compensation was performed using beads (see details in 2.4.1), using single stains for each antibody and one unstained sample. Samples were acquired uncompensated and compensation was later computed in FlowJo V10. FMO controls were used to check accurate compensation.

List-mode Data Files
FSC files can be obtained by contacting Prof. Susanna Dunachie (see contact details above) after this work has undergone peer-review and publication.

Compensation
The compensation matrices shown below are representative examples.

Data Transformation Description
FlowJo Version 10 has been used for analysis and visualization of the data.
FSC and SCC were acquired and displayed with linear scaling, all fluorescence parameters were acquired with log scaling and bi-exponentially transformed for data visualization.

Gate statistics
Count and frequency of gated population within parent population were exported and used to calculate frequency within the live cell gate. For absolute frequencies, these frequencies were applied to the live PBMC yield (live cell frequency determined by FACS was applied to the number of PBMC isolated per ml of whole blood).