Human Infections by Non–Rodent-Associated Hantaviruses in Africa

Abstract

Infections by hantaviruses (family Bunyaviridae) are found worldwide causing up to 200 000 clinical cases in humans annually. Hantavirus disease is a febrile illness with renal and/or pulmonary impairment, also called hemorrhagic fever with renal syndrome or hantavirus cardiopulmonary syndrome. Case fatality rates range between 0.1% and 50% [1]. For many years, rodents (order Rodentia) were considered the only hantavirus reservoirs that transmit the virus to humans. Recent studies revealed that insectivores (shrews, moles; order Soricomorpha) and bats (order Chioptera) can also carry novel, genetically unique hantaviruses; however, there is no clear evidence that these viruses can infect humans. A single study described the serologic detection of Thottapalayam virus–specific immunoglobulin (Ig) G in a patient with fever of unknown origin from Thailand [2]. Moreover, in vitro studies showed that shrew-borne hantaviruses, Thottapalayam virus and Imjin virus, induce proinflammatory responses in human macrophages comparable to those induced by pathogenic Hantaan virus, suggesting that shrew-associated hantavirus species potentially could cause disease in humans [3].

The African continent turned out to be a focus for the discovery of novel, highly diverse hantaviruses since Sangassou virus (SANGV) was identified in a captured African wood mouse (Hylomyscus simus) in Guinea in 2006. So far, this is the only African hantavirus isolated from biological material and grown in cell culture [4]. Specific IgM (in acute serum) and neutralizing antibodies (in convalescent serum) against SANGV were also detected in Guinean patients. Together with the occurrence of clinical symptoms of hemorrhagic fever with renal syndrome, the detection of anti-SANGV IgM and IgG implied that SANGV (or a closely related virus) can cause not only hantavirus infection in humans but also disease [5]. Subsequent studies revealed the occurrence of other hantaviruses in a broad range of insectivores and even bats (for review, [6]). Guinea was also the place of the first detection of Tanganya virus and Bowé virus (BOWV), carried by Therese's shrew (Crocidura theresae) and Doucet's musk shrew (Crocidura douceti), respectively [6, 7]. In the East African country Tanzania, Uluguru virus (ULUV) was discovered in the Geata mouse shrew (Myosorex geata) [8].

The ongoing discovery of more and more hantaviruses in “nonconventional” (nonrodent) hosts raises the question whether these viruses can infect humans and induce disease as rodent-borne hantaviruses do. Acute and past human infections can be demonstrated by detection of specific IgG using different screening and confirmatory assays (eg, enzyme-linked immunosorbent assay [ELISA], immunoblot, and immunofluorescence [IF]), whereas additional finding of specific IgM in the context of relevant clinical symptoms speaks for a disease association. In almost all cases, antibodies of IgG and IgM classes can be detected at the onset of clinical symptoms; IgM titers usually disappear after few months [1].

Our current study aimed to elucidate the relevance of 2 representative African shrew-borne hantaviruses, BOWV and ULUV. In parallel, we searched for human infections by a rodent-borne African hantavirus, SANGV.

METHODS

We analyzed 346 human serum samples collected from healthy persons positive (n = 140) or negative (n = 206) for human immunodeficiency virus in Gabon [9], as well as 330 serum samples originating from Côte d'Ivoire. The latter samples were taken from patients with fever of unknown origin and symptoms of renal or pulmonary disease, putatively indicating a hantavirus infection (Supplementary Table 1). The study was approved by the National Ethics Panel, Côte d'Ivoire (permission 014/MSLS/CNER-dkn) and the Institutional Review Board, Centre de Recherche Médicale de Lambaréné, Gabon (permission CEI-CERMEL 004/2016). All participants gave informed consent before sampling.

A panel of diagnostic tests was developed to screen for specific antibodies, based on recombinantly expressed nucleocapsid proteins of BOWV, ULUV, and SANGV, according to published protocols [10–12]. IgG-specific ELISAs were conducted for an initial screening. Reactive serum samples were further investigated by means of Western blot (WB). Thereafter, IF microscopy was applied to confirm WB-reactive serum samples.

The following definitions were used to characterize a serum sample as (1) hantavirus antibody positive and (2) specifically anti-BOWV, anti-ULUV, or anti-SANGV positive. A sample was considered hantavirus antibody positive if results of all 3 tests (ELISA, WB, and IF microscopy) were positive with any of the used antigens of BOWV, ULUV, or SANGV. If a sample consistently reacted with the antigen of 1 specific virus at ELISA, WB, and IF microscopy, it was further tested using antibody and antigen competition ELISAs to prove the specificity of the reaction. Only if these assay results were also positive was the serum sample rated as specifically BOWV, ULUV, or SANGV positive. These samples were also tested with IgM-specific ELISA, with results confirmed again by IgM competition ELISA (see Supplementary Material for the working scheme [Supplementary Figure 1] and detailed assay protocols).

RESULTS

Of 330 serum samples collected in Côte d'Ivoire, 19 (5.8%) were hantavirus antibody positive (Table 1). Of these, 6 samples were typed as specifically BOWV positive and 1 as SANGV positive because they reacted with BOWV and SANGV antigens, respectively, in all assays (ELISA, WB, IF microscopy) as well as in the antibody and antigen competition ELISAs.

The sample collection from Gabon comprised 16 of 346 hanta-positive samples (4.6%). Among them, 3 samples were specifically BOWV positive, 1 was ULUV positive and 3 were typed as SANGV positive according to ELISA, WB, IF microscopy, and competition ELISA results.

In the antibody and antigen competition assays, the reactivity of specifically BOWV-, ULUV-, or SANGV-positive samples was exclusively blocked by preincubation of human serum samples with rabbit anti-BOWV, anti-ULUV, and anti-SANGV antibodies, respectively (Figure 1A). Preadsorption of serum samples to soluble BOWV, ULUV, or SANGV antigen also resulted in virus-specific decrease of ELISA reactivity (Figure 1B).

Figure 1.

Competition enzyme-linked immunosorbent assay (ELISA) analysis to prove specific binding of human immunoglobulin (Ig) G to Bowé virus (BOWV), Uluguru virus (ULUV), or Sangassou virus (SANGV) antigens (see Table 1). A, For antibody competition ELISA, human serum samples were preincubated with rabbit hyperimmune serum samples (anti-BOWV, anti-ULUV, or anti-SANGV) in dilutions of 1:100 000 to 1:1000 or with rabbit preimmune serum (1:1000) before use in the ELISA reaction. B, For antigen competition ELISA, the same serum samples were preadsorbed for 1 hour to soluble BOWV, ULUV, or SANGV nucleocapsid proteins (0.4 to 40 nmol/L) or hexahistidine (6xHis) peptides (40 nmol/L) as controls. Dashed lines represent ELISA cutoff. *Virus-specific reaction not confirmed by competition ELISA (see Table 1).

Open in new tabDownload slide

All samples designated as distinctly BOWV, ULUV, or SANGV positive were also analyzed with IgM-specific ELISA to prove possible acute infections. Remarkably, 1 patient from the Côte d'Ivoire cohort, who had fever (40°C) and kidney failure requiring hemodialysis, was positive for BOWV-specific IgM (Supplementary Table 2). The specificity of this result was confirmed by IgM competition ELISA; preincubation of the serum sample with either rabbit anti-BOWV antibody or a lysate of BOWV nucleocapsid protein-expressing eukaryotic cells clearly reduced the ELISA reactivity (Supplementary Figure 2).

DISCUSSION

Our findings provide serological evidence of human hantavirus infections in the West and Central African countries of Côte d'Ivoire and Gabon, respectively. Previous studies on the epidemiology of hantavirus infections in Africa were conducted using antigens of European and Asian origin (sometimes also SANGV). By using antigens derived from BOWV and ULUV, both endemic in Africa, our study showed for the first time that shrew-borne hantaviruses can also infect humans.

The principle of competitive ELISAs was recently used for the exclusion of shrew-borne Seewis virus infections in Finland [13]. In the current study, using comparable assays, we were able to show the specificity of antibodies directed against BOWV, ULUV, and SANGV antigens.

Because exact genotyping of the viruses cannot be assured with the applied tests, one can conclude that the BOWV- and ULUV-specific infections were caused by these viruses or by viruses serologically closely related to them. Remarkably, those related viruses all fall into the phylogenetic groups of non–rodent-associated hantaviruses (Supplementary Figure 3). SANGV- and BOWV-carrying reservoirs, Hylomyscus sp. and Crocidura sp., respectively, are widespread in West and Central Africa [7, 14]. An emergence of SANGV and BOWV infections in countries other than Guinea seems to be consequential. In contrast, M. geata or related host species, responsible for ULUV transmission, are not known to be endemic in the study sites. Therefore, it is not surprising that we identified only a single ULUV infection.

IgM ELISA screening revealed a single patient with acute BOWV infection among those with suspected hantavirus from Côte d'Ivoire. No viral RNA was detected in this patient by means of reverse-transcription polymerase chain reaction. This result is not unlikely, because viremia during hantavirus infections is short lived and frequently outlasted by measurable IgM antibodies [15]. The patient was a 39-year-old woman from Abidjan with febrile illness and renal failure, which implies that BOWV infection could have caused this clinical picture. Closer investigation of the pathogenesis of BOWV and ULUV is therefore required to further assess the human pathogenic potential of these viruses.

Notes

Financial support. This work was supported by the German Research Foundation (grants FL1813/7-1 and KR1293/12-1) and the German Federal Office of Bundeswehr Equipment, Information Technology and In-Service Support (project E/U2AD/CD056/DF551).

Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

Author notes