Spatial Dispersal of Epstein-Barr Virus in South America Reveals an African American Variant in Brazilian Lymphomas

Abstract

Epstein−Barr virus (EBV) is a saliva-borne ɣ-herpesvirus associated with benign and malignant lymphoproliferation. EBV-mediated tumorigenic mechanisms are not fully understood and may be related to viral genetic variations. In this work, we characterize the genetic diversity of EBV from Brazil, assessing 82 samples derived from saliva from asymptomatic carriers (n = 45), biopsies of benign reactive hyperplasia (n = 4), and lymphomas (n = 33). Phylogenetic and phylogeographic analysis of the entire coding region of the LMP-1 was performed. Additionally, type 1/type 2 distinction by the EBNA3C gene and Zp variants were evaluated. Our results revealed a high diversity of EBV in Brazil, with the co-circulation of four main clades, described here as: Mediterranean (40.2%, n = 33), Raji/Argentine (39%, n = 32), B95-8 (6.1%, n = 5), and Asian II (1.2%, n = 1). The Raji/Argentine and Mediterranean clades were the most prevalent in South America (45% and 28%, respectively). The Raji/Argentine clade was associated with polymorphisms I124V/I152L, del30 bp, and ins15 bp (p < 0.0001, to all clades) and with a high haplotype diversity related to EBV type and Zp variants. We found that a Raji/Argentine subclade spread primarily from Brazil and later to other South American countries. Although no LMP1 variant has been directly associated with disease, the Raji/Argentine clade was predominantly clustered with lymphomas (61%) and the Mediterranean clade with non-malignant cases (59%) (p = 0.1). These data highlight the high genetic diversity of EBV circulating in Brazil, calling attention to a Raji-related variant with great recombination potential in Brazilian lymphomas.

Keywords: EBV diversity; Epstein–Barr virus; LMP-1; South America; phylogeography.

Figure 1

Maximum likelihood phylogenetic tree of the complete coding region of LMP1 oncogene sequences from asymptomatic carriers and pathological cases. Reference sequences included in the analysis belong to variants: B95-8 (green); Med+/−, Mutu, and Daudi (red); AG876 (dark red); Akata, GD1, China1, Cao, GD2, M81, and HKNPC (pink); China2, Alaskan, and NC (purple); Raji (blue); and Argentine (light blue). Five highly supported main clades harboring reference sequences (grey shadows) were identified. The tip labels of our samples represent: case type (AC: asymptomatic carrier; cHL: classic Hodgkin lymphoma; BL: Burkitt lymphoma); EBV type (T1 for type 1 or T2 for type 2); and Zp variant (V1 or V3). The samples carrying haplotype T1 + V3 are named in bold. aLRT support values are indicated at the main branches, and the branch lengths are to scale with the lower bar indicating nucleotide substitutions per site. The tree was rooted at the midpoint.

Figure 2

Genetic map of the LMP-1 polymorphisms found in Brazil related to the prototype B95-8 strain (GenBank number NC_007605). (A) Structure of the LMP1 gene highlighting the position of XhoI site loss, SNPs, insertions, and deletions. The dotted region denotes the intron positions. The color of the triangles represents the clade associated with the polymorphism, following the legend. (B) Structure of the LMP1 protein showing the domains. The triangles indicate the approximated position of each analyzed polymorphism, colored according to the associated clade (legend on the right). The color of the polymorphism is associated with the most prevalent clade (black indicates no association).

Figure 3

Maximum likelihood phylogenetic tree of the LMP1 gene from the Brazilian samples in this work (n = 82, red circles) and those available in the GenBank database (n = 472, gray circles). Six main clades were observed, marked by different shadings, according to the legend at center (Mediterranean—grouping the references Med+, Med− Daudi, and Mutu; Asian II—grouping the references HKNPC1, M81, Cao, Akata, NPC, and China1; Asian II—grouping the NC, Alaskan, and China2 references; America/Europa—grouping sequences from different countries without a reference strain. The aLRT branch support values are displayed in key branches. Branch lengths are scaled with the bar indicating nucleotide substitutions per site.

Figure 4

Dispersal of the Raji/Argentine subclade circulating in South America. Bayesian MCC time-scaled discrete phylogeographic tree for the Raji/Argentine clade. The colors of the branches represent the most likely location of its descending nodes, with the key colors shown on the legend. The posterior probabilities (PP) and the posterior state probabilities (PSP) are annotated in the basal nodes. All horizontal branch lengths are drawn on a scale of years. The inset represents the viral migrations routes of the Raji/Argentine clade based on the location probability of the MCC tree. The arrows represent the viral migrations routes of Raji/Argentine clade based on the location probability of the MCC tree, and are colored according to time as indicated by the legend.