The World Health Organization has periodically classified NPC based on histology, yet all NPC appears to be variants of squamous cell carcinoma 5. Though NPC risk varies geographically, nearly all cases and types of NPC demonstrate EBV-positive cells and thus EBV is strongly implicated in the etiology of NPC 578. Over 90% of the population is EBV positive, yet EBV is rarely detected in normal human epithelial biopsies, and when it is detected, it is almost always in carcinomas and in the latent phase 6847. The EBV exhibits three latency programs 48. Latency III is the least restrictive and involves the expression of six nuclear antigens (EBNA1-6), three latent membrane proteins (LMP1, LMP2A, and LMP2B), and untranslated nuclear RNAs. Latency I is the most restrictive and expresses LMP2A and EBNA1 and is associated with memory B-cells. The Latency II program expresses EBNA1 and the three LMPs, and is the program associated with NPC 48. These latency programs and other factors compete to repress and promote the activity of the EBV lytic transcription factor Zta 47. Importantly, both EBV reactivation and EBV latency programs are thought to contribute to the etiology of NPC 6849505152. The former process may increase the risk of infection to target epithelial cells, whereas the latter may increase the risk of cell transformation. The potential influence of formaldehyde on these processes is described in the sections that follow.

In 1968, the EBV was reported to be reactivated in cells cultured in media deficient in the nonessential amino acid and NO precursor L-arginine 53. Over 25 years later, NO synthase (NOS) inhibitors were shown to reactivate EBV in human lymphocytes exhibiting Type I latency but not Type III 29. This specificity was posited to relate to the fact that LMP1 (not present in Type I latency) and NO converge on similar signaling pathways such as NF-κB 2934. NO also prevented B-cell apoptosis, which is consistent with evidence that LMP1 promotes uncontrolled B-cell proliferation 2948. This suppression of the lytic phase and apoptosis by NO was reportedly mediated through cGMP-independent signaling pathways, and likely involves nitrosothiol modification of p21 Ras 293454. NO also suppresses the lytic phase of EBV in epithelial cell lines; and moreover, the established induction of EBV reactivation by TPA (12-O-tetradecanoylphorbol-13 acetate) coincided with decreased levels of NOS 55. It is also noteworthy that NO inhibition increased the expression of the early lytic transcription factor Zta in both lymphocytes and epithelial cells 2955. The binding of this viral transcription factor to DNA is reported to be inhibited by both redox and nitrosative mechanisms that involve critical cysteine residues 28. Interestingly, Zta is reported to induce the expression of interleukin-13, which may in turn up-regulate ADH3 5657. Although further studies are needed to examine these responses together in individual cell types and in in vivo systems, the data suggest that nitrosothiol homeostasis plays a dual role in promoting the latent phase and inhibiting the lytic phase of the EBV in lymphocytes and epithelial cells expressing certain latency programs (Figure 1b).

To date, little is known about what conditions might lead to altered nitrosothiol homeostasis in vivo. It is becoming increasing clear, however, that environmental exposures to allergens can elevate ADH3 expression in airways thereby decreasing nitrosothiol levels sufficiently to cause patent changes in pulmonary physiology 415758. Moreover, evidence suggests that exposure to relatively high levels of formaldehyde vapor can lead to immune responses, elevated ADH3, and increased GSNO breakdown in rodent airways 455960. Additionally, formaldehyde can accelerate GSNO reduction several fold in cell and cell-free systems through NAD⁺/NADH cofactor recycling on ADH3, as well as decrease cellular nitrosothiol levels in vitro (Figure 1a)) 6162. Together, these data suggest that formaldehyde can accelerate GSNO reduction, which might promote EBV reactivation in cells lining the airways (Figure 1c). This reactivation, in turn, could lead to increased viral shedding in nasopharyngeal lymphocytes and subsequent infection of basal epithelial cells 8. It is notable, though circumstantial, that the relative risk of NPC among occupationally exposed workers is reported to be stronger among EBV seropositive individuals 2.

Several mechanisms are implicated in EBV-mediated cell transformation. EBNA1, EBNA2, and LMP1 have all been shown to induce oxidative stress, genomic instability, or DNA repair inhibition 63646566. In fact, LMP1 is posited to be an important oncogene and a key modulator in the pathogenesis of NPC through activation of NF-κB signaling pathways related to cell cycle control, apoptosis and transformation 51. LMP1 acts as a ligand-independent CD40 receptor that functions, in part, by preventing the ubiquitination of TNF-receptor associated factors 50. Other EBV proteins manipulate the fate of host proteins by accelerating or decelerating their degradation through the ubiquitin-proteosome system, while simultaneously decelerating their own ubiquitination and degradation 5067. These myriad effects imparted by EBV latency programs likely perturb many cellular functions in subtle ways that can increase spontaneous transformation or otherwise potentiate cells for transformation. As noted previously, the etiology of NPC involves host, viral and environmental factors. Considering that the functions of many proteins may be modulated by S-nitrosylation, it is conceivable that excess formaldehyde exposure may further alter protein functions in EBV-positive cells. Additionally, formaldehyde itself can deplete GSH levels and can react with proteins and peptides 686970. Recent genomics analyses in yeast cells clearly indicate that oxidative stress pathways are induced by formaldehyde exposure 70. These analyses also identified protein fate (processing and degradation, folding and stabilization, and ubiquitination pathways) as a major category of altered gene expression; formaldehyde induced 14-19% of genes involved in ubiquitination processes, whereas methanol (which is oxidized to formaldehyde) induced only 1% of these genes 70. While this simple eukaryotic system may not directly inform in vivo mammalian tissue responses, there is little reason to believe that it cannot provide sentinel information as to how formaldehyde interacts with cellular macromolecules. These effects on protein function may partially explain the synergistic effect formaldehyde has on genetic damage in mammalian cells 7172. Moreover, such effects on proteins may explain the association between NPC and consumption of salted fish containing nitrosamines like N-nitrosodimethylamine, as the cytochrome P450-mediated oxidation of these compounds can generate formaldehyde (Figure 1c) 5773747576.