Role of EBV infected lymphocytes in the pathogenesis of Nasopharyngeal carcinoma

Summary

We investigated the association between EBV-infected lymphocytes and the pathogenesis of Nasopharyngeal carcinoma (NPC). In situ hybridisation (ISH) and Immunohisto chemistry (IHC) were used to test the frequency and distribution of EBV-infected lymphocytes in specimens of tonsils from patients with NPC. In situ hybridisation was used for EBV encoded RNAs (EBER) and Immunohistochemical expression of LMP-1 protein was examined by Immunohistochemistry in formalin- fixed, paraffin embedded specimens of NPC.

Introduction

Epstein-Barr virus (EBV) is a lymphotropic herpes virus associated with several lymphoid malignancies including Burkitt’s lymphoma, Hodgkin’s disease and Post Transplant Disease (PTLD), which itself is a lymphoma (Cruchley et al 1997). It is also associated with Nasopharyngeal carcinoma (NPC), an epithelial neoplasm prevalent in southern china. NPC is rare in Europe and North America. The undifferentiated type of NPC is known to be associated with genomic Epstein-Barr virus (EBV) DNA. EBV was first discovered in 1964 by Epstein et al in cultured lymphocytes obtained from patients with Burkitt’s lymphoma (Vasef et al 1997). Reinhard et al study in 2002 concluded that undifferentiated NPC are strongly associated with EBV. The cytoskeleton of undifferentiated NPC reveals no specific pattern of cytokeratins expression.
A unique characteristic of EBV is its ability to infect and transform primary resting B Lymphocytes in vitro into permanently growing lymphoblastoid cell lines (Cruchley et al 1997). This effect is associated with constitutive expression of a limited set of viral genomes. Few infected cells go on to produce new viruses and die in the process. Other EBV-transformed cells secrete immunoglobulins (Cruchley et al 1997). Epithelial cells and T cells, as well as B cells, have also been shown to be infectible with EBV, since the viral genome can be found in the epithelial tumour cells of nasopharyngeal carcinoma, epithelial lesions, Hodgkin’s disease and peripheral T-cell tumours. It is widely believed that similar events take place in vivo and this may probably explain the maintained levels of EBV in an individual throughout his life (Cruchley et al 1997). This means that the virus live along with its host (B cell), through alternated stages of latency and growth, keeps the virus continuously regenerating as well as maintaining the same frequency of virus in each individual. Throughout these stages, the virus has been found to express at least eight viral proteins each one at completely different stage. Additionally EBV was found to encode for RNAs (EBERs), which is expressed in great amount, regardless of the stage of latency of the virus unlike the other EBV-encoded proteins (Cruchley et al 1997).
We were keen to know whether events similar to in vitro are taking place in vivo which may explain the maintained levels of EBV in an EBV seropositive individual throughout his life? And we wondered, If EBV is a lymphotropic virus, how can it infect epithelial cells?
The association between EBV and Nasopharyngeal carcinoma (NPC) has been confirmed by detection of EBV antigens in NPC by immunoblotting, detection of EBV DNA by in situ hybridisation, detection of EBV genome by polymerase chain reaction and demonstration of monoclonality of EBV in NPC (Plaza et al 2002).
Number of studies have been conducted in different countries to study the association between EBV and NPC. Bar Sela et al studied EBV and NPC in Israel particularly with reference to the latency effect. Immunohisto chemistry (IHC) for latent membrane protein 1 (LMP-1) and in situ hybridisation (ISH) for EBV encoded RNA (EBER) were used to evaluate the prevalence. They concluded that NPC in Israel is highly associated with EBV latency as detected by EBER ISH (Sela et al 2004). Mirzamani et al study in 2006 in Iran concluded that there is a strong association between Epstein-Barr virus and Nasopharyngeal carcinoma (Mirzamani et al 2006). Immunohisto chemistry was used to detect the association between EBV and NPC. Plaza et al in 2002 studied the association between EBV and NPC in Spanish patients. They evaluated the association of nasopharyngeal carcinoma (NPC) and Epstein-Barr virus (EBV) in Spanish patients, and studied the expression of EBV products (latent membrane protein-1 (LMP-1) and ZEBRA proteins) by NPC cells and its possible prognostic value. In situ hybridization (ISH) for EBV-encoded nonpolyadenylated RNAs (EBERs) and immunohisto chemical expression of LMP-1 and ZEBRA proteins by immunohisto chemistry were examined in formalin-fixed, paraffin-embedded NPC specimens from 30 patients, and a survival analysis was done by the Kaplan-Meier method. They detected EBERs by ISH in 96.67% of the NPC cases, and detected expression of LMP-I in 43.33% of the NPC cases and expression of ZEBRA protein in 6.67% of the NPC cases. They concluded that ISH for expression of EBERs is an adequate method for detection of EBV in NPC. In our study we used in situ hybridisation and the same principle.
Earlier studies on EBV have found out that the frequency of the virus varies significantly among the majority of EBV-seropositive individuals. But it is interesting to note that this frequency was found to be stable over period of time. In the studies of EBV and Hodgkin’s disease, frequency of EBV-infected non-malignant lymphocytes has been shown to be significantly higher as compared to healthy EBV-seropositive individuals (Khan et al 2005). Our study is based on the hypothesis that there is a similar kind of association between increased frequency of EBV-infected lymphocytes and Nasopharyngeal carcinoma. We wanted to explore that association by using In situ hybridisation and Immunohisto chemistry. Similar study was conducted last year by another student who used infected patient with no disease symptoms as a negative control. Data from previous study is not been used here.
In-situ Hybridisation initially made use of gene segment probe from the EBV genome (repeat sequence). Later it was established that during latent infection with EBV, at least 8 virally encoded proteins are expressed. Epstein Barr encoded RNAs (EBER1- EBER2) are also present and by far the most abundant. This makes them the targets of choice for In-situ hybridisation using cloned EBERs riboprobes. After the tissue sample is taken, the methods involved fixation of the tissues and microtomy, in which samples were cut and set on TEPSA-coated slides.

Materials and Methods

The tissue samples: 80 slides of tonsillar tissue from a deceased patient were already made of two main tissue blocks fixed in paraffin (H18326/03/A1, H18326/03/D1).
Microtomy: The samples were 80 slides of tonsillar tissue already microtomed because we did have neither the time nor the expertise to do the job. However the tissue samples were kept in a wax of paraffin.
Probes End-labelling by digoxigenin:
Oligos were from sigma
ABC detection kit from vector laboratories
Digoligonucleotide tailing kit was from roche
Controls:they were from a PTLD (Post transplant disease) case. Positive as it is and the negative when we skipped the primary antibody
Pre Hybridisation (sample preparation)
The slides were incubated at 60ºc in a heated cupboard, for 1hour prior to use, to assist with dewaxing the sections. The slides were then transferred whilst warm into slide baths of xylene and were left to dewax for at least 5 minutes before being transferred to another final xylene bath for another 5 minutes. Slides were inspected thoroughly to ensure that all paraffin had been completely removed before advancement. The slides were dehydrated in graded alcohols, 70%, 90% and absolute for 1 minute in each bath. The endogenous peroxidase was blocked by an acid alcohol bath (0.5% hydrogen peroxide in methanol) for 20 minutes. This step saturates and denatures the tissue bound peroxidase especially present in the connective tissue. 150 µl of Proteinase K type XXVIII (Sigma) digestion of the tissue was incubated for 15 minutes at 37ºc in heat resistant hybridization tray, with 35 ml of water, to limit the drying of the sections. Digestion was terminated by immersion of the slides in distilled water for 1 minute and repeated 3 times. The graded alcohol dehydration of the slides was repeated and the slides were allowed to air dry thoroughly. 20 µl of the digoxigenin-labelled EBER probe in hybridization buffer solution at concentration of 1 ng/ µl was added to each slide except the negative control and covered with the 32 x 22 mm cover slip.

Hybridisation

35 ml of 2x SSC was added into the hybridization tray, prior to loading the slides, taking careful precautions to avoid contact of the slides directly with the solution to prevent the dilution of the hybridization mix. The hybridization tray was inserted into the microwave and microwave was set to medium for 1 minute, followed by low for a further 7 minutes. The heat of the hybridization tray required caution during the removal of the lid. The condensation of the 2 x SSC was poured back into the tray, again avoiding contact with the slides. With the lid replaced, the hybridization tray was transferred to an incubator preset at 42ºc and left overnight.

Detection by ImmunoHisto Chemistry

The slides were moved from the hybridisation tray and the cover slips were removed by the immersion into 2x SSC. The slides were left in the SSC for 5 minutes, the SSC replaced and again immersed for another 5 minutes. The stringency wash disrupts any electrostatic interaction between probe and non- EBER sequences; this was done using 0.1x SSC incubated at 55ºc for 20 minutes, replacing the 0.1 x SSC solutions at 10 minutes. The slides were then immersed in 2 x SSC for 5 minutes before the solution was replaced and left for a further 5 minutes. Without allowing the slides to dry, 150 µl of primary mouse anti-digoxin antibody was applied to each section. This was made up from 1/5000 of anti-digoxin monoclonal mouse antibodies, 1/50 horse serum and balance 1 x phosphate buffered saline (PBS). The slides were covered and left to incubate for 30 minutes. The incubation was terminated by 3 consecutive washes with PBS each lasting 5 minutes before the solution was replaced. The 150 µl of secondary biotinylated anti-mouse antibody was applied to each section; made up from 1/500 of anti mouse antibody, 1/50 horse serum and PBS. The slides were again covered for 30 minutes incubation. The incubation was terminated by 3 consecutive washes with PBS each lasting 5 minutes before the solution was replaced.
The avidin biotinylated peroxidase complex was prepared from 1/50 avidin, 1/50 biotinylated peroxidase and PBS. It is essential that the complex be provided 30 to 40 minutes of time to form prior to application. As consequence the solution was made up during the incubation of the secondary antibody. The incubation was terminated by 3 consecutive washes with PBS each lasting 5 minutes before the solution was replaced. The chromogen was prepared from 1/50 Buffer, 1/25 DAB, 1/50 hydrogen peroxide and made up with distilled water. 150 µl of DAB was applied to each section, and left to incubate for 7 minutes. The incubation was terminated by 3 consecutive washes with tap water each lasting 5 minutes before the solution was replaced. A haematoxylin counter stain was completed to by 45-second bath in haematoxylin followed by 1 minute in tap water, 1 minute in acid alcohol and 5 minutes in tap water. The slides were dehydrated in graded alcohols, 1 minute in each solution. The slides were immersed in xylene, and left for at least 5 minutes before transfer to another xylene bath for another 5 minutes. The slides were mounted with DPX and cover slip

Statistical Analysis

Results

This was a retrospective study using archival surgical pathology tissue blocks from St. Georges Hospital, tooting, London, UK. Ethical approval was obtained in prior.
ISH for EBERs: We detected EBER expression in most of the slides. The positive hybridisation for EBERs was demonstrated as nuclear staining that involved most of the neoplastic cells in most specimens, whereas surrounding reactive lymphoid tissue showed no staining (Plaza et al 2002).
Immunohisto chemistry: Expression of LMP-1 was observed to be positive. But only isolated cases showed membrane expression, and characteristically, the intensity of the staining was weak in most cases. LMP-1 expression was restricted to scattered cells in several areas of the specimens. Plasma cell staining was considered cross-reactivity of LMP-1. EBV proteins were only expressed by tumoral cells (Plaza et al 2002)

Discussion

Genetic predisposition, environmental factors and EBV infection have all been implicated in the pathogenesis of NPC (Vasef et al 1997), but EBV appears to be the strongest and most consistent related factor. Occupational exposure to smoke, fumes and chemicals have also been linked to NPC (Vasef et al 1997). Ingestion of salted fish is also considered to be one of the risk factors in the pathogenesis of NPC. But EBV gene expression is considered to be closely associated with the pathogenesis of NPC. Among EBV genes expressed in NPC, EBV-encoded non-polyadenylated RNAs, termed EBERs are the most abundant transcripts of EBV in NPC (Yoshizaki et al 2007). Yoshizaki et al concluded that EBERs are believed to induce the initial transformation of epithelial cells, thus contributing to the oncogenesis of NPC. Expression of abundant EBERs is considered to be critical for this transforming property of EBERs. According to the current sensitive in situ hybridization methods for the detection of EBV-encoded small RNAs (EBER), almost 100% of cases of NPC, irrespective of their histologic subtypes, have demonstrable EBERs in the nuclei of the tumor cells (Vasef et al 1997).
The EBER in situ hybridisation probe and immunohisto chemical detection, as used by Khan et al 1992 was reproduced for this investigation with minor changes and shown to provide clear and consistent staining results. Of the 100 slide sample only 8 sections required exclusion, either due to loss of section material or significant damage during the staining process. All cells presenting with positive stain were identifiable as of lymphoid lineage presenting the characteristic morphology and nuclear to cytoplasmic ratio. The consistent results of the positive and negative control PTLD tissue further support the specificity of this technique, and shown to be easily reproducible.
Involvement of Identified EBV Infected Cells in Latency Mechanism (Babcock et al 1998) speculates that direct infection of EBV activates naïve B cells, presenting the EBNA-2 dependant growth program to enter the germinal centre reaction; initially proliferating as centroblasts expressing the EBNA-1. The CD40L activity of the LMP1 and the BCR mimicry by the LMP2a ITAM will allow the infected centrocyte, which has developed from centroblast to survive independent of selection by CD4+ T helper cells. Given the association of naïve B cells with the mantel zone of the tonsillar tissue it is possible to infer that the cells which are positive within the germinal centres, are either migrating naïve B cells or germinal centre centroblasts or centrocytes. Once differentiated, the memory B cells migrate from the germinal centre to peripheral blood. The isolated positive cells observed in the extrafollicular compartment may represent these latently infected memory B cells. The virus maintains a silent latency transcribing the latency program, only consistently transcribing LMP2a to maintain the inactivity of the BCR and the transient expression ENBA-1 product during the division of the latency infected memory B cell. The independence of the latency from CTL elimination within this cellular compartment would account for presence of 99% of the EBV genome in the peripheral blood.
Memory B cells have a half life of 2 -3 weeks; once the virus has gained entry to this cellular compartment, the limited life span of memory B cells would require the periodic antigen reactivation of the BCR to sustain the recirculation of these latently infected cells. The latently infected memory B cells will periodically return to the marginal zones of the tonsillar tissue preferentially and reactivate the growth program undergoing the same germinal centre reaction (Joseph et al 2000).
‘Absence of EBER Positive Germinal Centre Cells in IM’: In this study by Anagnostopoulos et al utilising a similar EBER in situ hybridisation detection of infected cells during IM, found positive cells of similar lymphoid morphology restricted from the germinal centres with very few exceptions. Although absence of epithelial infected cells was in concurrence, the strong infiltration of the tonsillar epithelium was however distinctly absent from our study. The predating study by Niedobitek ET a corroborates the exclusion of EBER positive cells from the germinal centres during IM.
The conflict with this investigation can be directly attributed to the atypical cellular infection of EBV during acute IM. Kurth et al failed to find infected naïve B cells or the somatic hypermutation of the V region immunoglobin genes associated with germinal centre reaction in cases of IM.
However, the direct infection of the memory and germinal centre B cells was found to be causative of the clonal expansion, identified to express the growth program. During IM it is probable that direct infection of memory B cells will by-pass the infection of naïve B cells without utilisation of the germinal centre to provide the long-term latency in the peripheral blood.

Frequency of EBV infected cell

The frequency was very fluctuant between the two samples A1 and D1 as well as within the sample D1. This could have been an observation error as the protocol may not be even to all the other slides. This highlights the importance of the evenness in the protocol in IHC and the human error factor. Further studies like Image Analysis and digital computer assisted techniques would be definitely beneficial to validate this routinely used visual technique. Digital computer assisted quantitative method could resolve disagreement between two observers about the quality of staining intensity because the digital method does not classify the results into groups, but rather provides a numerical value for each individual case and, thus, increases the diagnostic and, above all, prognostic sensitivity of the immunohisto chemical analysis. It is not always easy to differentiate between infected epithelial cell and an infected lymphocyte while using immunohisto chemistry and mistakes can happen more often than not, mainly because of human errors. Errors due to improper staining, observation errors on the part of the individual conducting the study are common while using Immunohisto chemistry. The experience and the expertise of the observer also play a major role in the successful use of techniques like ISH and IHC. Though IHC is the most commonly used technique, it is always worth considering other staining techniques available to improve the contrast.
I would like to highlight some of the errors in our study. I think that the EBERs-positive epithelial cells could have been confirmed better by double staining of cytokeratine and also it is important to remember that the comparing the frequencies of the virus in few slides of the NPC case’s tonsils is very different from comparing the frequency in several people. The latter is more representative and relevant in terms of sampling.
Optimum use of IHC and ISH is very important in studies like these and when used effectively these methods can yield great results. I would like to further emphasise on the effective use and intended benefits of IHC and ISH by following examples.