Re: Discrete alterations in the BZLF1 promoter in tumor and non-tumor-associated Epstein-Barr virus.

In their recent article, Gutiérrez et al. (1) identified three Epstein–Barr virus (EBV) variants in the BZLF1 promoter Zp (Zp-P [considered the prototypical sequence], Zp-V3, and Zp-V4) and postulated that these might result in different lytic potentials and be differentially expressed among individuals with nonmalignant and malignant diseases. Hong Kong and Southern China are considered to be endemic regions, with the EBV variants that are different from other localities. We analyzed EBV type and sequence variations in EBV Zp domains by direct DNA sequencing in a total of 95 EBVassociated lesions (96 isolates) from individuals from Hong Kong and Southern China (Table 1). Eighty-seven of 95 (91.6%) lesions contained type A EBV. Eight (8.4%) lesions contained type B EBV, all of which carried Zp-V3. Of the 88 type A EBV isolates, 31 (35.2%) contained Zp-P and 53 (60.2%) contained Zp-V3. From 36 nonmalignant samples with type A EBV, 19 (52.8%) contained Zp-P, and 16 (44%) contained Zp-V3. From 51 malignant samples, we obtained 52 EBV isolates with type A EBV, of which 12 (23.1%) contained Zp-P and 37 (71.2%) contained Zp-V3. In addition, from one nonmalignant and three malignant samples with type A EBV, we identified four novel Zp-P variants. Three variants differed from Zp-P at one position only: Zp-V1_104 (C-A), isolated from a healthy carrier; Zp-V1_105 (C-T), isolated from a natural killer (NK)/T-cell lymphoma; and Zp-V1_119 (G-A), isolated from a nasopharyngeal carcinoma (NPC). All these substitutions occurred within the ZIIIB domain of Zp, which can strongly bind ZEBRA (Z, EB replication activator) and contribute to the transactivation of Zp (2). A fourth variant (Zp-V3+118), isolated from an NPC, contained four nucleotide substitutions relative to Zp-P, three of which were identical to those found in Zp-V3 (1), and the fourth a C to T substitution at position –118. This position is also within the functional domain of ZIIIB. It is noteworthy that this variant with four nucleotide substitutions was different from the Zp-V4 variant described by Gutiérrez et al. (1). Taken together, these results suggest that the ZIIIB domain may be a mutational hotspot. It was suggested that only Zp-P and Zp-V4, but not Zp-V3, were found in the nonmalignant samples (1). However, in the 40 nonmalignant samples from our locality, Zp-P and Zp-V3 occurred with similar frequency (19 of 40 and 20 of 40, respectively). None of our 95 samples contained Zp-V4, suggesting that the Zp-V4 identified by Gutierrez ́ et al. (1) may be restricted to people in a particular geographic region. In our malignant NPC samples, 80.5% (33/41) contained Zp-V3 and 14.6% (6/41) contained Zp-P, whereas in the nonmalignant samples, 50% (20/40) contained Zp-V3 and 47.5% (19/40) contained Zp-P. Thus, Zp-V3 was more prevalent in NPC samples than in nonmalignant samples (P .001, chi-square test). If we considered only type A EBV isolates, then the association between NPC and Zp-V3 was still statistically significant (P .001). In our locality, the distribution of EBV variants with respect to Zp in various EBV-associated conditions is distinct from other geographic regions. Although Zp-V3 is detectable in nonmalignant samples, it is statistically significantly associated with NPC. Whether these variants have a functional consequence remains to be determined. However, the possible link(s) between different variants and lytic potential or latency-to-lytic cycle switch requires further investigation.