Systems analysis-based assessment of post-treatment adverse events in lymphatic filariasis

Abstract

Background: Lymphatic filariasis (LF) is a neglected tropical disease, and the Global Program to Eliminate LF delivers mass drug administration (MDA) to 500 million people every year. Adverse events (AEs) are common after LF treatment.

Methodology/principal findings: To better understand the pathogenesis of AEs, we studied LF-patients from a treatment trial. Plasma levels of many filarial antigens increased post-treatment in individuals with AEs, and this is consistent with parasite death. Circulating immune complexes were not elevated in these participants, and the classical complement cascade was not activated. Multiple cytokines increased after treatment in persons with AEs. A transcriptomic analysis was performed for nine individuals with moderate systemic AEs and nine matched controls. Differential gene expression analysis identified a significant transcriptional signature associated with post-treatment AEs; 744 genes were upregulated. The transcriptional signature was enriched for TLR and NF-κB signaling. Increased expression of seven out of the top eight genes upregulated in persons with AEs were validated by qRT-PCR, including TLR2.

Conclusions/significance: This is the first global study of changes in gene expression associated with AEs after treatment of lymphatic filariasis. Changes in cytokines were consistent with prior studies and with the RNAseq data. These results suggest that Wolbachia lipoprotein is involved in AE development, because it activates TLR2-TLR6 and downstream NF-κB. Additionally, LPS Binding Protein (LBP, which shuttles lipoproteins to TLR2) increased post-treatment in individuals with AEs. Improved understanding of the pathogenesis of AEs may lead to improved management, increased MDA compliance, and accelerated LF elimination.

Fig 1. Filarial antigens increase post-treatment in individuals with moderate adverse events (AEs).

Circulating filarial antigen (CFA) levels increased significantly more post-treatment in individuals who experienced moderate adverse events (mod AEs) compared to individuals with no AEs (P < 0.05 by Kruskal-Wallis) (n = 62 with no AEs, n = 24 with mild AEs, n = 9 with moderate AEs). Boxes indicate the interquartile range (25^th and 75^th percentile of data distribution), and horizontal lines within the boxes are median values. The whiskers show 95% confidence intervals around the median values.

Fig 2. Levels of complement components and lipopolysaccharide binding protein (LBP) pre- and post-treatment (n = 9 with moderate adverse events (AEs), n = 9 with no AEs).

Fig 2A: Complement component 3 (C3) significantly decreased post-treatment in individuals with moderate AEs (*P = 0.03 by paired t-test). B: Complement component 4 (C4) did not change with treatment in either AE group. C: Complement Factor B (FB) did not change with treatment in either AE group. D: LBP levels significantly increased post-treatment in individuals with moderate AEs (***P = 0.00007 by paired t-test).

Fig 3. Post-treatment fold changes for 12 cytokines in nine participants who experienced moderate AEs and in nine participants who had no AEs.

Levels of IL-8, MCP-1, VEGF, TNF-α, MIP-1β, G-CSF and IFN-γ significantly increased post-treatment in individuals with moderate AEs, and levels of IL-1RA, IL-6, IL-10, MIP-1α and IP-10 increased significantly post-treatment in individuals with moderate AEs and in individuals with no AEs (*P < 0.05, **P < 0.001 by Wilcoxon signed-rank tests). IL-1RA, IL-6 and IL-10 increased more in individuals with moderate AEs (*P < 0.05 by Mann-Whitney U tests).

Fig 4. Overall expression patterns and enrichment pre- and post-treatment (n = 9 with moderate adverse events (AEs), n = 9 with no AEs).

Fig 4A: Sample clustering (Euclidean distance) based on gene expression profiles across all genes. Post-treatment AEs (red) samples are significantly overrepresented in the fourth group (bolded, P < 0.0001, binomial distribution). Treatment arms; A: albendazole (ALB), IA: ivermectin (IVM) and ALB, IDA: IVM, diethylcarbamazine and ALB. B: Principal component analysis of paired samples. Post-treatment AEs samples (red) are significantly different from their pre-treatment controls (yellow) (P = 0.005 by PERMANOVA). No other differences between the four groups are significant. C/D: Expression plots comparing pre- and post-treatment gene expression in individuals with moderate AEs (C) and individuals with no AEs (D). Red dots represent genes significantly upregulated post-treatment, and blue dots are genes significantly upregulated pre-treatment (P < 10⁻⁵ for the AE group (C), and P < 0.05 for the no AE group (D) from DESeq2). E: Overlap between the genes upregulated pre- and post-treatment between the two AE groups.

Fig 5. Estimated leucocyte subtypes (CIBERSORT) for the two adverse events (AEs) groups.

Fig 5A: Percent change post-treatment of leukocyte subtypes. Estimated lymphocytes decreased more and estimated neutrophils increased more post-treatment in individuals with moderate AEs (n = 9) compared to individuals with no AEs (n = 9). B: Estimated leukocyte proportions pre-treatment. Individuals that did not develop AEs (n = 9) had higher levels of estimated memory B cells pre-treatment compared to individuals that developed moderate AEs (n = 9). *P < 0.05 by Mann-Whitney U tests.

Fig 6. qRT-PCR validation of RNA-seq data.

Post-treatment fold change of the top eight genes identified by random forest analysis. Seven of the genes (DIP2B, ZCCHC6, PELI1, FNDC3B, TLR2, LTBR and NT5C2) significantly increase only in individuals with moderate adverse events (AEs) (n = 8) compared to individuals with no AEs (n = 9). *P < 0.05, **P < 0.001 by t-tests. HPRT1 is a housekeeping gene that should not change with treatment in either AE group.