Detection of Human cytomegalovirus UL55 Gene and IE/E Protein Expression in Colorectal Cancer Patients in Egypt

Human cytomegalovirus (HCMV) is a ubiquitous virus with population seropositivity reaching up to 100% in Africa and Asia [1]. It attains a lifelong latency in haematopeitic progenitor cells. HCMV causes a wide range of infections ranging from asymptomatic to mild infections and occasionally severe infections in the immunocompromised as well as teratogenicity during pregnancy [2]. Correlation between HCMV and different human cancers has been widely explored with mixed results between studies [3].

HCMV encodes a wide range of proteins which have major oncomodulatory effect on changing cell survival behavior and alteration of immune surveillance [4]. HCMV US28 gene is implicated in the HCMV-generated angiogenic phenotype by secreting vascular endothelial growth factor (VEGF), a key molecular for angiogenesis [5, 6]. HCMV encodes Immediate early (IE) proteins including UL37 exon 1, UL36, and UL38 proteins that interfere with antiapoptotic genes e.g. retinoblastoma, p53 and cyclins as well as induction proapoptotic genes e.g. fos and myc [7]. Also, the level of Cyclooxygenase 2 (COX-2) increases in viral infected cells. COX-2 is a main source of prostaglandins which promotes inflammatory conditions [8]. Moreover, HCMV infection changes the formulation of matrix metalloproteinases, that enhance tumor growth by destroying matrix barriers and increasing angiogenesis, thus are crucial in cancer metastasis. Chromosomal aberration and instability are also described [9, 10]. Moreover, HCMV impacts the anticancer immunity. It enables tumors to evade immune surveillance by encoding viral proteins as US28, UL111A, and UL144 resulting in HCMV-induced immunological tolerance, which promotes tumor growth [11, 12]. Additionally, HCMV has been recently categorized into low risk (LR) strains causing oncomodulation and high risk (HR) strains with direct oncogenic potential. The molecular mechanisms behind HCMV-induced oncogenesis are complex, with cellular stress, PGCCs (Polypoid giant cancer cells), and genomic instability all playing a role [11].

HCMV can infect any part of the gastrointestinal tract from the mouth till the rectum with a spectrum of illnesses. Colorectal cancer (CRC), a multifactorial disease with infectious agents incriminated, is a leading cause of cancer with high morbidity and mortality worldwide [13].

Molecular based methods are the most widely used diagnostic tests for detection of HCMV infection as well as detecting reactivation [14]. Immunohistochemistry (IHC) for demonstration of HCMV proteins in tissue specimens is not routinely done but can give important topographical information about the viral location inside the cell [15].

The relation between HCMV and CRC pathogenesis is of a high research interest and being increasingly explored aiming to unravel the precise viral role. The aim of the current study was to detect HCMV by PCR and IHC in colorectal cancer samples as well as to correlate HCMV presence to different CRC demographic, clinical and pathological parameters.

Overall, Colorectal cancer is the second leading cause of cancer mortality and third in incidence. The incidence of CRC is rising in low-medium income countries (LMIC) due to changes in lifestyle [21]. HCMV is highly ubiquitous reaching almost 100% seroprevalence in adults in developing countries with a lifelong latency [1, 22]. In the current study, HCMV presence was assessed in paired tumorous and ANNT tissues of 50 seropositive CRC patients by two techniques; PCR and IHC. UL55 gene primer, encoding highly conserved HCMV gB, was chosen according to a HCMV primer comparative study which reported PCR sensitivity to be inversely proportional to the length of the PCR product i.e. sensitivity being higher when the product is less than 100 bp [14]. DNA was extracted from tissues with minimal formalinization and before fixation to preserve DNA integrity [19]. A significantly higher detection rate was observed in PCR results in tumorous tissues versus ANNT group (64% vs 26%, P < 0.001). Two metaanalyses, one investigating HCMV presence in gastrointestinal cancers and the other focusing on CRC, significantly detected HCMV in tumoral tissues [23, 24]. On the other hand, recently Chelbi et al. showed no difference of UL55 gene detection in tumor versus peritumor tissues [25]. Also, earlier studies found no relation between HCMV and CRC [26‐28]. This might be due to the small number of included cancerous samples, non-optimization of PCR technique, or due to testing FFPE as tumor specimens which was avoided in this study. HCMV infection at certain stages of tumor development followed by its clearance is also a possible cause.

IHC is a valuable technique to detect viral protein expression and thus differentiate latent from active infection as well as demonstrate the specific site inside the cell. HCMV IE/E proteins are transactivators for expression of early and late viral genes. They are essential to establish lytic infection and viral re-activation dysregulating cellular functions [5, 29].

In the present study, a significantly higher detection of HCMV IE and E proteins was observed in tumorous tissues versus ANNT group (P < 0.008). Taher et al. reported in 2 studies higher prevalence of HCMV IE DNA and protein expression in CRC and breast cancer and their metastases to the brain with poor outcome [30, 31]. Recently, two studies reported that HCMV gB DNA detection or expression of HCMV IE in breast cancer was found to be associated with shorter survival [32, 33]. HCMV IE and E proteins were always detected in the cytoplasm in this study. Chen et al. and Dimberg et al. also demonstrated that viral antigens are preferentially detected in the tumor cytoplasm [34, 35]. Two possible explanations were hypothesized for HCMV viral protein expression in tumoral tissue, either HCMV is carried to the tumor and become reactivated due to suppressed immunity or, more likely, the inflammatory process reactivates the already existing virus [18].

In the present work, a lower HCMV detection rate was obtained by IHC than PCR which could be due to expression of a mutated protein, [36] or due to masking viral proteins by increased N/C ratio in malignant cells which at many times reached a ratio of 1:1.

When comparing PCR and IHC results, a moderate agreement was detected (Kappa coefficient = 0.572, P < 0.001). One study comparing PCR and IHC to explore the relation between HCMV and breast cancer showed a fair concordance (kappa = 0.345; P = 0.003) [37]. In fact, combination of PCR and IHC to explore HCMV role in CRC in Egyptian population and reporting the significant association between HCMV presence and CRC as well as detection of the moderate agreement between the two techniques is distinctive in the current study. Only a few earlier studies combined IHC to PCR for HCMV detection and found no relation to CRC. Interestingly, the fact that immunostaining was detected in the cytoplasm and not in the nucleus, as in cases of acute infection, was their reason to consider it as nonspecific staining [26, 28]. Nevertheless, “the hit and run” theory of oncogenesis could not be ruled out. On the contrary, a pioneer study by Harkins et al. detected a significant relationship and postulated an oncogenic rule [38]. A recent study of the effect of IE viral protein on CRC derived stem cell discovered cytoplasmic localization to be the most commonly detected [39].

In the present study, 70% of the enrolled CRC patients were females and 36% were elderly (> 60y). A recent meta-analysis reported higher incidence of CRC in males, and in elderly patients [40]. Adenocarcinoma was the prevalent histopathological type (92%) with Grade 2, higher stages (T3 + T4) and nodal involvement accounting for (64%, 64% and 56%) respectively. Our findings could be attributed to scarcity of screening programs so that the disease is only discovered after seeking medical advice on complaining.

When demographic and clinicopathological parameters were correlated to HCMV DNA presence, a statistically significant relationship in patients aged > 60 years, with higher CRC stages and tumors with nodal involvement (P = 0.033, P = 0.006, P = 0.015) respectively.

In multivariate analysis, significant HCMV DNA detection was retained in CRC higher stages (T3 + T4) and nodal involvement (P = 0.010, P = 0.008). Thus, the most independent factor affecting HCMV detection was nodal involvement followed by stages 3 and 4 in colorectal cancer. This study’s results thus suspect a HCMV role in disease progression even in the absence of a clear role.

In a study of CRC progression in the elderly, Chen et al. demonstrated poorer outcome in the presence of HCMV independent of other factors and reported no correlation with demographic, clinical or pathological factors as grading, local extension and nodal involvement to HCMV presence. They suggested that decreased immunity with age may be an important risk factor to HCMV reactivation, nevertheless, a causal relation could not be confirmed [36]. In another study for the non-elderly, Chen et al. found an opposite effect suggesting a dual oncomodulatory age-dependent effect for HCMV presence [41]. This may be due to stimulation of the immune response to HCMV infection in young age [42].

Interplay between oncomodulatory/oncogenesis role of HCMV was suggested in the context of identification of high risk HCMV strains versus low risk strains showing difference in disease progression, survival rate and response to treatment [5, 43]. Surprisingly, on the other hand, a recent clinical analysis study detected a statistically significant correlation between HCMV infection and lower CRC incidence [44]. The detection of HCMV DNA as well as active transcription of IE protein in CRC tumor tissues points to a substantial oncomodulatory profile in the elderly.

CRC tumor tissues are more infected by HCMV than ANNT. A significant association of HCMV presence with a higher CRC tumor stage and nodal involvement in an age-dependent manner was detected. HCMV oncomodulatory and a disease progression role is suspected.