In our study, EBV(+) cases were observed in 90% of patients with NPC, including EBV/HPV co-infection (10% of NPC patients) using ISH. This was in concordance with a previous study in which co-infection with HPV and EBV was found in 10% of patients with NPC [16]. It was reported that the HPV infection situation in NPC was different in endemic areas and non-endemic areas. In endemic areas, where type II (differentiated non-keratinizing carcinoma) and type III (undifferentiated non-keratinizing carcinoma) mainly existed, the prevalence of HPV(+) was relatively low (7.7%), and HPV(+)/EBV(-) patients showed better prognosis after radiotherapy in Southern China [17]. In non-endemic areas, where type I (keratinizing squamous cell carcinoma) exists to a relatively high extent, HPV infection is frequently found (30%), and HPV(+) and patients with EBV(-)/HPV(-) NPC had worse outcomes than those with EBV(+) NPC [18]. Controversially, some studies have found that HPV did not influence NPC carcinogenesis [19, 20]. To date regarding the interaction between EBV and HPV, it is known that EBV latent gene product latent membrane protein 1 can block p16 expression [21]. For the tissue microarray used in this study, TNM grade for patients with NPC was grade III, and therefore, we could not assess the effect of co-infection on NPC progression. More studies are needed to investigate the role of HPV in NPC pathogenesis.
Our results showed that MIF expression in tumor nests was higher than those in the tumor stroma of NPC and inflammation tissues. Higher expression of MIF has been reported in different cancers, including endometrial cancer, lung adenocarcinoma, hepatocellular carcinoma, colon cancer, and NPC [22–26]. Increased MIF has been demonstrated to be associated with poor survival in colorectal cancer, oral squamous cell carcinoma, and gastric cancer [27–29]. Furthermore, plasma MIF can improve the diagnostic specificity for NPC patients combined with VCA-IgA [30], suggesting that high tumor-associated MIF expression may drive higher circulating levels of soluble MIF. Intracellular MIF can be stored in the cytosol or secreted into the extracellular space. MIF affects both tumor progression and tumor-associated immune responses. It is becoming increasingly evident that MIF plays an important regulatory role in governing TAM-dependent tumor initiation, progression, and metastatic disease phenotypes, although a unifying mechanism that explains how MIF contributes to this seemingly divergent M1 and M2 macrophage phenotypes is still lacking [31].
We found more CD68(+) macrophages in the tumor stroma than in the tumor nest, as well as in the non-epithelial areas of inflammation cases. Interestingly, more M1 (CD11c(+)) macrophages infiltrated the tumor nest than in the tumor stroma, and more M2 (CD163(+)) macrophages infiltrated the tumor stroma than the tumor nest. Huang et al. also reported that M2 macrophages presented high density in NPC stroma, which may be caused by the mesenchymal cells near the tumor nest attracting macrophages by generating chemotactic activity [32]. It is well known that macrophages tend to accumulate in hypoxic tumor areas, and tumor cells can release cytokines to switch microphages to M2 macrophages, which can promote tumor progression [33]. Additionally, different levels of TAMs have been reported to have a different association with prognostic parameters [34]. A recent meta-analysis demonstrated that higher CD68(+) TAMs in tumor nests predicted favorable disease-free survival, and higher M2 (CD163(+)) macrophages were associated with poor survival outcomes in NPC [35]. Considering the promotion effect of M2 macrophages on tumors by producing vascular endothelial growth factor and extracellular matrix remodeling proteins, it is well understood that high M2 macrophages indicate a poor prognosis [36]. Ohri et al. found that more M1 macrophages in tumor nests were associated with a better prognosis in non-small cell lung cancer [37]. Although there was no information on prognosis in this study, HPV-related M1 macrophage infiltration in the tumor nest may have some effects on the prognosis of NPC patients with co-infection.
MIF has been reported to influence macrophage polarization. In glioma malignancy, MIF can inhibit M1 macrophages [38]. MIF also contributes to TAM polarization to the M2 subtype in tumor-bearing mice [39]. In our study, we evaluated the correlation between MIF expression levels and TAM-associated marker levels (CD68, CD11c, and CD163); however, there was no significant correlation between them. In contrast, we found a significant positive correlation between EBER1 levels and MIF expression levels in tumor nests. This suggests that EBV may promote MIF production in the tumor nest, affecting the progression of NPC. Few studies have examined the relationship between MIF and EBV, and more research is needed to understand the relationship between them. We also found a weak positive correlation between HPV16/18 level and the score for CD11c(+) cells in the tumor nest, indicating that there may be an interaction between HPV and M1 macrophages in NPC.