Leishmania aethiopica field isolates bearing an endosymbiontic dsRNA virus induce pro-inflammatory cytokine response

Abstract

Background: Infection with Leishmania parasites causes mainly cutaneous lesions at the site of the sand fly bite. Inflammatory metastatic forms have been reported with Leishmania species such as L. braziliensis, guyanensis and aethiopica. Little is known about the factors underlying such exacerbated clinical presentations. Leishmania RNA virus (LRV) is mainly found within South American Leishmania braziliensis and guyanensis. In a mouse model of L. guyanensis infection, its presence is responsible for an hyper-inflammatory response driven by the recognition of the viral dsRNA genome by the host Toll-like Receptor 3 leading to an exacerbation of the disease. In one instance, LRV was reported outside of South America, namely in the L. major ASKH strain from Turkmenistan, suggesting that LRV appeared before the divergence of Leishmania subgenera. LRV presence inside Leishmania parasites could be one of the factors implicated in disease severity, providing rationale for LRV screening in L. aethiopica.

Methodology/principal findings: A new LRV member was identified in four L. aethiopica strains (LRV-Lae). Three LRV-Lae genomes were sequenced and compared to L. guyanensis LRV1 and L. major LRV2. LRV-Lae more closely resembled LRV2. Despite their similar genomic organization, a notable difference was observed in the region where the capsid protein and viral polymerase open reading frames overlap, with a unique -1 situation in LRV-Lae. In vitro infection of murine macrophages showed that LRV-Lae induced a TLR3-dependent inflammatory response as previously observed for LRV1.

Conclusions/significance: In this study, we report the presence of an immunogenic dsRNA virus in L. aethiopica human isolates. This is the first observation of LRV in Africa, and together with the unique description of LRV2 in Turkmenistan, it confirmed that LRV was present before the divergence of the L. (Leishmania) and (Viannia) subgenera. The potential implication of LRV-Lae on disease severity due to L. aethiopica infections is discussed.

Figure 1. LRV is found in L. aethiopica field isolates.

A. LRV detection by dot blot using an anti-dsRNA antibody (J2). Promastigotes were directly spotted onto a nitrocellulose membrane (2 µg of total protein/spot) before dsRNA detection by J2. A ponceau red staining of the membrane before J2 addition is added to demonstrate that similar amounts of parasites were loaded. B. Viral genomic dsRNA visualization from nucleic acid extracts. Total promastigote nucleic acids (5 µg for Lg and 20 µg for Lae) were digested with ssRNase and DNase I then migrated in a 0.8% agarose gel (upper panel). An undigested control of each sample acts as a loading control (lower panel).

Figure 2. LRV-Lae dsRNA localization by immunofluorescence microscopy.

Promastigotes were fixed with formaldehyde and spread on poly-lysine coated slides before visualization of viral dsRNA with the J2 antibody (standardized exposure time in all images: 200 ms). Scale bars: 10 µm.

Figure 3. LRV detection in the L. aethiopica L494 strain.

In addition to the three L. aethiopica cryobank lines tested, Lg M4147 LRV1+ was added as a positive control. A. PCR amplification. A portion of the LRV capsid protein open reading frame (489 and 486 bp for LRV1 and LRV2 respectively) was amplified from total cDNA using LRV universal primers (Table S1). As a cDNA quality control, a 372 bp fragment of the beta-tubulin gene was also amplified. B. LRV dsRNA visualization. Total RNA was analyzed on agarose gel. Ribosomal RNA (rRNA) and the complete 5.3 kb LRV genomic dsRNA are indicated.

Figure 4. LRV2-Lae genome comparison to previously sequenced LRVs.

A. Comparison of LRV1-Lg, LRV2-Lmj and LRV2-Lae open reading frames (ORFs). The two major ORFs encoding the capsid protein (CP) and the RNA-dependent RNA polymerase (RdRp) are represented by gray and black boxes respectively. The short ORFs upstream of the CP genes are represented by open boxes with nucleotide positions indicated in italics. The two short ORFs (15–233) and (117–299) upstream of LRV1-Lg capsid are deduced from LRV1-Lg CUMC1 and LRV1-Lg M4147 viral genomes respectively. B–D. Phylogenetic analysis of L. aethiopica LRVs and previously sequenced LRV genomes. Total RNA genome sequence (B), and the deduced capsid (C) and RdRp amino acid sequences (D) of LRVs isolated from the indicated strains were analyzed by Jalview. The average distances (using BLOSUM62) are indicated on the trees.

Figure 5. A unique LRV2-Lae genomic organization in the capsid protein/RdRp open reading frames switching region.

The LRV genomic region coding for the end of the capsid protein (CP) and the beginning of the RdRp is shown for L. guyanensis M4147 (LRV1-Lg), L. major ASKH (LRV2-Lmj) and L. aethiopica 303 (LRV2-Lae). The corresponding amino acids of CP and RdRp are above and below the cDNA sequence respectively. CP stop codon is indicated by *. The overlapping region is shown in grey.

Figure 6. The presence of LRV2-Lae leads to TLR3-dependent production of pro-inflammatory cytokines by in vitro infected macrophages.

C57BL/6 (in black) and TLR3 knock-out (in grey) murine bone marrow derived macrophages were infected by Leishmania promastigotes (parasite/macrophage ratio 10∶1), and the level of IL-6 (A) and TNF-α (B) in culture supernatants was measured by ELISA 24 hours post-infection. Non inf.: non-infected macrophages. The cut-off line was calculated as 3 standard deviations (SD) above the mean absorbance of the uninfected macrophage control. Average values presented were obtained from two independent experiments performed in duplicates.