Zika Virus Outbreak in Haiti in 2014: Molecular and Clinical Data

Abstract

Background: Zika virus (ZIKV), first isolated in Uganda in 1947, is currently spreading rapidly through South America and the Caribbean. In Brazil, infection has been linked with microcephaly and other serious complications, leading to declaration of a public health emergency of international concern; however, there currently are only limited data on the virus (and its possible sources and manifestations) in the Caribbean.

Methods: From May, 2014-February, 2015, in conjunction with studies of chikungunya (CHIKV) and dengue (DENV) virus infections, blood samples were collected from children in the Gressier/Leogane region of Haiti who presented to a school clinic with undifferentiated febrile illness. Samples were initially screened by RT-PCR for CHIKV and DENV, with samples negative in these assays further screened by viral culture.

Findings: Of 177 samples screened, three were positive for ZIKV, confirmed by viral sequencing; DENV-1 was also identified in culture from one of the three positive case patients. Patients were from two different schools and 3 different towns, with all three cases occurring within a single week, consistent with the occurrence of an outbreak in the region. Phylogenetic analysis of known full genome viral sequences demonstrated a close relationship with ZIKV from Brazil; additional analysis of the NS5 gene, for which more sequences are currently available, showed the Haitian strains clustering within a monophyletic clade distinct from Brazilian, Puerto Rican and Guatemalan sequences, with all part of a larger clade including isolates from Easter Island. Phylogeography also clarified that at least three major African sub-lineages exist, and confirmed that the South American epidemic is most likely to have originated from an initial ZIKV introduction from French Polynesia into Easter Island, and then to the remainder of the Americas.

Conclusions: ZIKV epidemics in South America, as well as in Africa, show complex dissemination patterns. The virus appears to have been circulating in Haiti prior to the first reported cases in Brazil. Factors contributing to transmission and the possible linkage of this early Haitian outbreak with microcephaly remain to be determined.

Fig 1. Virus-specific CPE in simian kidney cell line LLC-MK2.

Non-inoculated cells (A) and cells inoculated with plasma specimen 1225/2014, 8 days post-inoculation (B). Perinuclear vacuoles are evident. Original images taken at 400x magnification; insets at approx. 800X.

Fig 2

A) Transmission electron micrograph detail of a ZIKV-infected LLC-MK2 cell. The large arrow points out an area containing typical flavivirus-induced paracrystalline arrays/convoluted membranes in a ZIKV-infected LLC-MK2 cell. (B) Transmission electron micrograph detail of ZIKV-infected LLC-MK2 cell. Crystalline arrays of virus cores (large arrow) are shown in association with membrane vesicles.

Fig 3. Maximum-Likelihood tree of ZIKV complete genome sequences.

The tree was obtained using the best fitting nucleotide substitution model (TN93+G) selected by a hierarchical likelihood ratio test. Branches are drawn to scale in nucleotide substitutions per site according to the bar at the bottom of the tree. Percentage bootstrap (out of 1000 replicates) support values are given along branches. The Haiti sequence is in bold.

Fig 4. Maximum clade credibility (MCC) tree with Bayesian phylogeography reconstruction of ZIKV NS5 gene region.

Branches are scaled in time and colored according to the legend to the left where each color represents the geographic location of the sampled sequence (tip branches), as well as of the ancestral lineage (internal branches) inferred by Bayesian phylogeography. The molecular clock was calibrated by using ZIKV strains known sampling times and enforcing a relaxed molecular clock with a Bayesian skyline plot demographic prior (see Supplementary Methods). For further clarity, the country of origin of the main strains in the MCC tree is also indicated to the right of each major clade (the MCC tree with full names of each isolate is provided in S2 Fig). Significant posterior probability support (p≥ 0.9) is indicated by the number along the branch. The Haitian sequences are in bold.