A nairovirus isolated from African bats causes haemorrhagic gastroenteritis and severe hepatic disease in mice

Abstract

Bats can carry important zoonotic pathogens. Here we use a combination of next-generation sequencing and classical virus isolation methods to identify novel nairoviruses from bats captured from a cave in Zambia. This nairovirus infection is highly prevalent among giant leaf-nosed bats, Hipposideros gigas (detected in samples from 16 individuals out of 38). Whole-genome analysis of three viral isolates (11SB17, 11SB19 and 11SB23) reveals a typical bunyavirus tri-segmented genome. The strains form a single phylogenetic clade that is divergent from other known nairoviruses, and are hereafter designated as Leopards Hill virus (LPHV). When i.p. injected into mice, the 11SB17 strain causes only slight body weight loss, whereas 11SB23 produces acute and lethal disease closely resembling that observed with Crimean-Congo Haemorrhagic Fever virus in humans. We believe that our LPHV mouse model will be useful for research on the pathogenesis of nairoviral haemorrhagic disease.

Figure 1. Phylogenetic analysis of LPHVs and other bunyaviruses based on amino acid sequences.

The percentages of trees in which the associated taxa are clustered are shown beside the branches. The tree is drawn to scale, with branch lengths indicating the number of substitutions per site. The analysis involved 16 amino acid sequences. All positions containing gaps and missing data were eliminated. (a) Amino acid sequences of the GPC were phylogenetically analyzed, and the tree with the highest log-likelihood value (−31327.7075) is shown. There were a total of 933 positions in the final data set. (b) Amino acid sequences of the partial L protein were phylogenetically analyzed, and the tree with the highest log-likelihood value (−1582.7832) is shown. There were a total of 137 positions in the final data set. Nairoviruses were designated by the following IDs: AKAV, akabane virus; HEAV, heartland virus; HTNV, hantaan virus; OROV, oropouche virus; RVFV, rift valley fever virus; SEOV, Seoul virus; SFTSV, severe fever with thrombocytopenia syndrome; ToSWV, tomato spotted wilt virus. ToSWV was defined as the out group.

Figure 2. Comparison of deduced glycoprotein structures.

Glycoprotein precursors of the two LPHV strains and ERVV were deduced on the basis of primary structures and known information for CCHFV. Identified and deduced proteolytic sites are indicated by positions of polyproteins after tetrapeptide sequences that represent protease recognition sites. Proteases that recognize each site are indicated at the left side of the tetrapeptide sequences: c, supposed precursor convertase as described in a previous study; f, furin and s, SKI-1/S1P ref. . Deduced cleavage site after endogenous signal peptide sequence by signalase are indicated. Predicted numbers of O-glycosylated residues are indicated in the mucin-like domain. The homologies of 11SB17 and 11SB23 domains are indicated. Percentage homologies of the full-length glycoprotein precursor and each domain are indicated between the strain names and the domains.

Figure 3. Clinical indicators of LPHV infection in C57BL/6 mice.

11SB17 and 11SB23 strains at the indicated titres or PBS (NC) were i.p. injected into C57BL/6 mice. Five individuals (n=5) were challenged for each condition. (a) Body weight changes in LPHV-infected mice were monitored. Average relative body weights are indicated with s.d. (b) Survival rates of mice after LPHV challenges or PBS injections are indicated.

Figure 4. Macroscopic changes in LPHV-infected mice.

C57BL/6 mice injected with PBS (NC; a–c) and high inoculum dose of the 11SB23 strain (d–f) were dissected at 5 d.p.i. Whole body, intestine and liver photographs are shown. Bleeding region in the small intestine of 11SB23-infected mouse is indicated by black line (e), and faeces in colon and rectum were indicated by white line.

Figure 5. Histological findings of virus-infected tissues.

C57BL/6 mice injected with PBS (NC; a,d,g) and high dose of 11SB23 (b,c,e,f,h) were dissected at 5 d.p.i. Small intestine, liver and spleen tissues are shown with haematoxylin–eosin staining after fixation in 10% formalin. Small intestine and liver were observed by × 20 (panels in the left and middle column) and × 40 (panels in the right column) objective lens and that for the spleen was × 20. Scale bars are shown in the lower right side of the figures (a,b,d,e: 100 μm; c,f,g,h: 50 μm).

Figure 6. Blood tests and quantification of viremia in non-infected and LPHV-infected mice.

C57BL/6J mice were i.p. injected with PBS (NC) or 5 × 10⁶ f.f.u. of LPHV 11SB17 (17) or 11SB23 (23) strain. At the indicated time points, peripheral blood was harvested from seven individual animals (n=7). Individual values (open circle) and their averages (solid bar) were indicated. ***P<0.001 and *P<0.05 (Welch’s t-test). (a) Values of alanine aminotransferase (ALT), alkaline phosphatase (ALP), total bilirubin (TBIL) and glucose (GLU) are indicated. At 4 d.p.i. in 11SB23-infected mice, a value of TBIL affected by haemolysis was removed from the analysis (n=6). The maximum measureable value of ALT was 2,000, so values beyond 2,000 were recorded as 2,000. (b) Number of white blood cells (WBC), platelets (PLT) and red blood cells (RBC) are indicated. (c) Mean copy numbers of viral RNA in the three representative individuals of peripheral blood were measured by qRT–PCR. In each infection, this was measured at 2 and 4 d.p.i. Negative control (NC) at 0 d.p.i. indicated non-specific background levels and each value was expressed as mean±s.d. Y-axis is a logarithmic scale.