Abstract
Purpose
This study was designed to find a reliable Epstein-Barr virus (EBV) immunoglobulin (Ig) G–based diagnostic/screening test for nasopharyngeal carcinoma (NPC) able to demarcate between the NPC-related seropositivity of EBV IgG antibodies and that of other head and neck cancer (HNCA) and control groups. The NPC-associated immunosuppression affects EBV IgA much more than IgG, leading to inconsistent detection of NPC using EBV IgA antibodies.
Materials and methods
One hundred twenty-two HNCA patients, 42 NPC, 66 laryngeal carcinoma, and 14 hypopharyngeal carcinoma and 3 groups of 100 control subjects were enrolled in this study. Enzyme-linked immunosorbent assay (ELISA) was used to find a specific cutoff value for the NPC-related seropositivity of EBV IgG antibodies.
Results
NPC group showed higher serum level of EBV IgG antibodies than control and other HNCA groups ( P < .05). However, the traditional cutoff value, mean + 2 SDs of control subjects, failed to demarcate the seropositives of NPC patients from those of healthy population ( P > .05). The new cutoff value, mean + 2 SDs of the seropositives group of control subjects who had already been grouped by the traditional cutoff value, proved successful. It succeeded to demarcate between the NPC-related EBV IgG seropositivity and that issued from the persistent, latent, or reactivated EBV infection in the population ( P < .05). The sensitivity/specificity of NPC detection by the new cutoff-based ELISA kit, 76.19% and 86%, was close or higher than that of EBV IgA antibodies.
Conclusion
EBV IgG-based ELISA could be used for the diagnosis of NPC using a new cutoff threshold that excludes the population baseline of EBV IgG seropositivity.
1
Introduction
Nasopharyngeal carcinoma (NPC) is unique and distinct from other malignant tumors arising from head and neck because of its epidemiology, histopathologic spectrum, clinical characteristics, and biological behavior . Epstein-Barr virus (EBV) is a common herpesvirus that preferentially infects humans and is harbored by most adult population . Viral infection can lead to self-limiting lymphoproliferative disease of infectious mononucleosis and is associated with 2 human cancers, which are Burkitt lymphoma and NPC . The EBV establishes a lifelong persistent infection in more than 90% of the human adult population worldwide. EBV has been linked to the development of a variety of human malignancies of lymphoid and epithelial origin, including Burkitt lymphoma, Hodgkin disease, NPC, and oral hairy leukoplakia .
The consistent association of EBV with NPC was established by the following findings: the presence of EBV DNA, RNA, and proteins in the tissues of NPC; evidence that the tumor cells are composed of single EBV clone–infected cell; and increased antibodies against EBV antigens in NPC patients . Elevated antibody titers to EBV antigens, for example, early antigens, viral capsid antigens (VCA), are frequently found in NPC patients and have been considered as important markers in the diagnosis and prognosis of the EBV-associated diseases . However, it was found that EBV serum immunoglobulin (Ig) G and IgA are independent from the circulating EBV nucleic acids that conforms a clear confusion . Also, it was stated that there is a great diversity of the serum and mucosal IgG and IgA response against Epstein-Barr nuclear antigen, early antigen, and VCA in serum samples from patients with NPC, leading to inconsistent prediction and diagnosis process of NPC by serology . Despite the fact that EBV serum Igs response was considered one of the reliable indicators for the occurrence of NPC, no reliable and simple clinical immunological kit has been used to diagnose NPC and/or survey population . This might be attributed to the notion that EBV serum IgG and IgA antibodies are present in highly diversified levels in the healthy population due to the childhood exposure to EBV, presence of lifelong latent infection, or recent reactivation of latent infection. Therefore, it is not straightforward to demarcate which serum level of EBV antibody should be related to NPC. Moreover, using EBV serum IgA antibodies in NPC detection revealed a major drawback. Although EBV serum IgA proved to be more sensitive than EBV serum IgG , EBV serum IgA in NPC detection is unfortunately more liable to NPC-related immune suppression than EBV serum IgG antibodies . This might be attributed to the long life span of the already synthesized IgG antibodies before or during NPC development when immune suppression was at its early stages. This necessitated a new approach to revive the use of EBV serum IgG antibodies in the detection or screening of NPC patients, especially most of NPC patients worldwide showed considerable immune suppression where the serology of EBV IgA antibodies turns less valuable.
Unfortunately, previous research did not convince health bodies to adopt one of the EBV serum IgG-based kits for the consistent detection of NPC. Apart from the high sensitivity, the used kits of EBV serum IgG antibodies were of low specificity. The reason behind this is that cutoff values of EBV serum IgG or IgA antibodies were averaged from healthy populations, which might not differentiate the seropositivity of EBV antibodies related to NPC from other causes of EBV seropositivity such as previous EBV infection, latent infection, or recent reactivation of latent infection. In this study, we intended to formulate a novel approach of assigning a more reliable cutoff value of EBV serum IgG by which simply diagnose NPC and survey high-risk groups of NPC without interference of the population baseline seropositivity of EBV antibodies. It was intended to formulate a high-sensitivity/specificity NPC detection kit that could rival the efficacy of EBV IgA–based enzyme-linked immunosorbent assay (ELISA) kits. Therefore, a large number of head and neck cancer (HNCA) patients and control subjects were examined by ELISA to elucidate a newly designed methodology to calculate a cutoff value able to determine exclusively the seropositivity level of EBV IgG antibodies in NPC patients rather than in healthy population.
2
Methods
2.1
The population of the study
One hundred twenty-two HNCA patients were selected without any bias to any type of HNCA, in the period between January 2006 and January 2008 from the University Hospital of The Medical College of Alnahrain University and Radiotherapy Reference Center in Baghdad, Iraq, and from Alhussein Hospital in Amman, Jordan. The samples processing and the study were totally conducted in University Putra Malaysia in Kuala Lumpur.
Patients with HNCA were involved in the study after the diagnosis was established. Primary HNCA was selected rather than secondary or recurrent cases. Patients with HNCA were 42 NPC patients, 66 carcinoma of larynx (CL), and 14 hypopharyngeal carcinoma (HPC) patients. Patients with HNCA and control subjects in Iraq and Jordan were native Middle Eastern people of Arabic ethnicity. Twenty patients of other HNCA types that were of small sample size, less than 6, were neglected for statistical purposes such as tongue, retromalar, tonsillar, epiglottic, and postpharyngeal carcinomas. The age of HNCA patients ranged between 16 and 74 years, with median age of 53 years and mean age of 51.8 years. The median age of NPC patients was 54.4 years and the mean age was 54 years. For HPC patients, the median age was 62.5 years and the mean age was 62.16 years. For CL patients, the median age was 50.5 years and the mean age was 50.81 years. For HNCA patients, 89 (72.9%) were males and 33 (27%) were females, with male/female ratio of 2.6:1. For NPC patients, 9 (21.4%) were females and 33 (78.5%) were males, revealing male predominance at 3.6:1 ratio. For HPC patients, 12 (85.7%) were females and 2 (14.3%) were males, with females predominance at sex ratio 1:6. For CL patients, 6 (9%) were females and 60 (91%) were males, with clear male predominance at sex ratio 10:1. HNCA staging was based on TNM system by which HNCA patients were classified into 4 stages, depending on T (the extent of primary tumor), N (regional LN spread), and M (distant metastasis) . Accordingly, the studied HNCA patients were diagnosed at different stages of the disease: 68% at stage IV, 21% at stage III, 7% at stage II, and 4% at stage I. Cases of NPC were all found to be undifferentiated carcinoma of types II and III.
Three groups of age- and sex- matched control subjects were selected, each group was composed of 100 subjects. They showed normal blood and biochemical laboratory tests, namely, differential blood count, hemoglobin level, total serum proteins, liver function tests, erythrocytes sedimentation rate, kidney function tests, and C-reactive proteins, as well as were medically tested by a specialist and found free of HNCA. The first control group was used to calculate the first cutoff value of EBV serum IgG seropositivity, which was applied on the second control group demarcating the control subjects into seronegatives and seropositives. Out of the seropositive subjects of the second control group, the second cutoff value was calculated and applied on the third group of controls. The second cutoff value is supposed to demarcate between the normal EBV seropositivity of the population and the abnormal NPC-related EBV seropositivity. The reason behind using 3 groups of control subjects was to avoid the circular definition of calculating and applying the cutoff values on the same group. Sera were sampled from HNCA patients before any chemotherapy or surgery and from control subjects after taking full written consent to estimate the serum level of EBV IgG antibodies. This study was carried out according to the guidelines of the institutional ethics committee of biomedical research.
2.2
Enzyme-linked immunosorbent assay
Nonsandwich ELISA was modified and standardized from the routine procedure of Schuurs and Weeman . Nonsandwich ELISA was applied rather than sandwich ELISA after a series of pilot runs that showed the nonsandwich ELISA yielded the most precise readings. EBV VCA solution (Wellcome, Beckenham, England) was used as a solid phase of ELISA. The VCA antigen was diluted 1:10 in carbonate-bicarbonate coating buffer (1.59 g/L carbonate and 2.93 g/L bicarbonate; Sigma, St. Louis, MO) at concentration 10 μ g/mL. This concentration was used consistently in all runs to avoid the affection of the results by concentration variations. After a series of standardization steps, 50 μ L/well of the antigen in coating buffer was added for overnight at 4°C. Next day, the remaining fluid was aspirated and wells were washed twice with washing buffer (phosphate-buffered saline from Sigma, 1% bovine serum albumin wt/vol from Merck, Darmstadt, Germany, 0.05% vol/vol Tween 20 from Merck). Blocking buffer (phosphate-buffered saline, 1% bovine serum albumin wt/vol) was then added for 1 hour. The second layer of ELISA composed 50 μL/well of nondiluted serum of HNCA patients and control subjects. The ELISA plate was incubated for 2 hours at 37°C. After washing step, the third layer of ELISA was composed of 50 μL/well of horseradish peroxidase–labeled rabbit antihuman IgG antibodies diluted 1:40 at original concentration of 2 mg/mL (ICN Immunologicals, Lisle, IL). After washing, 50 μL OPD-H 2 O 2 (Sanofi Diagnostics, Lyon, France) was added for 15 minutes as chromogen-substrate for the development of the enzymatic reaction of horseradish peroxidase. Enzyme-linked immunosorbent assay readings were immediately measured in terms of optical density (OD) using a microtiter reader A-1764A (Beckman, Fullerton, CA) at a wavelength of 492 nm. A standard curve was generated using serial dilutions of known amounts of standard human IgG antibodies (Wellcome). The minimal detection limits for anti-EBV VCA IgG antibodies was 0.1 ng/mL.
2.3
Calculation of cutoff values for EBV IgG seropositivity
Cutoff value is considered as the upper limit, above which all readings were considered as positive. Therefore, the first and second cutoff values were calculated as follows:
Cutoff value = Mean ELISA OD readings of 100 controls + 2 SDs
2.4
Statistical analysis
Statistical analysis was conducted using SPSS software version 10 and MS Excel 2000. After proving that all studied samples obey the normal distribution pattern using Kolmogorov and Smirnov normalization tests, parametric tests of significant difference were used, namely, single-factor analysis of variance (ANOVA) and multivariate Student t tests. χ 2 Test of independence was used for evaluating the significant association of the studied groups with EBV IgG seropositivity depending on the observed frequencies of EBV IgG seropositives in each group. P values less than .05 were considered as significant. The sensitivity and specificity of ELISA assay to detect NPC out of control subjects or out of other HNCA groups were calculated by considering NPC patients as the positive control and control subjects or other HNCA patients as the negative control. Two sets of ELISA sensitivity and specificity were used. The first and second sets of ELISA sensitivity and specificity for detecting NPC were calculated using the first and second cutoff values respectively and as follows:
Specificity = number of negative control samples appeared negative by our test/number of negative control samples as a whole.
Sensitivity = number of positive control samples appeared positive in our test/number of positive control samples as a whole.
2
Methods
2.1
The population of the study
One hundred twenty-two HNCA patients were selected without any bias to any type of HNCA, in the period between January 2006 and January 2008 from the University Hospital of The Medical College of Alnahrain University and Radiotherapy Reference Center in Baghdad, Iraq, and from Alhussein Hospital in Amman, Jordan. The samples processing and the study were totally conducted in University Putra Malaysia in Kuala Lumpur.
Patients with HNCA were involved in the study after the diagnosis was established. Primary HNCA was selected rather than secondary or recurrent cases. Patients with HNCA were 42 NPC patients, 66 carcinoma of larynx (CL), and 14 hypopharyngeal carcinoma (HPC) patients. Patients with HNCA and control subjects in Iraq and Jordan were native Middle Eastern people of Arabic ethnicity. Twenty patients of other HNCA types that were of small sample size, less than 6, were neglected for statistical purposes such as tongue, retromalar, tonsillar, epiglottic, and postpharyngeal carcinomas. The age of HNCA patients ranged between 16 and 74 years, with median age of 53 years and mean age of 51.8 years. The median age of NPC patients was 54.4 years and the mean age was 54 years. For HPC patients, the median age was 62.5 years and the mean age was 62.16 years. For CL patients, the median age was 50.5 years and the mean age was 50.81 years. For HNCA patients, 89 (72.9%) were males and 33 (27%) were females, with male/female ratio of 2.6:1. For NPC patients, 9 (21.4%) were females and 33 (78.5%) were males, revealing male predominance at 3.6:1 ratio. For HPC patients, 12 (85.7%) were females and 2 (14.3%) were males, with females predominance at sex ratio 1:6. For CL patients, 6 (9%) were females and 60 (91%) were males, with clear male predominance at sex ratio 10:1. HNCA staging was based on TNM system by which HNCA patients were classified into 4 stages, depending on T (the extent of primary tumor), N (regional LN spread), and M (distant metastasis) . Accordingly, the studied HNCA patients were diagnosed at different stages of the disease: 68% at stage IV, 21% at stage III, 7% at stage II, and 4% at stage I. Cases of NPC were all found to be undifferentiated carcinoma of types II and III.
Three groups of age- and sex- matched control subjects were selected, each group was composed of 100 subjects. They showed normal blood and biochemical laboratory tests, namely, differential blood count, hemoglobin level, total serum proteins, liver function tests, erythrocytes sedimentation rate, kidney function tests, and C-reactive proteins, as well as were medically tested by a specialist and found free of HNCA. The first control group was used to calculate the first cutoff value of EBV serum IgG seropositivity, which was applied on the second control group demarcating the control subjects into seronegatives and seropositives. Out of the seropositive subjects of the second control group, the second cutoff value was calculated and applied on the third group of controls. The second cutoff value is supposed to demarcate between the normal EBV seropositivity of the population and the abnormal NPC-related EBV seropositivity. The reason behind using 3 groups of control subjects was to avoid the circular definition of calculating and applying the cutoff values on the same group. Sera were sampled from HNCA patients before any chemotherapy or surgery and from control subjects after taking full written consent to estimate the serum level of EBV IgG antibodies. This study was carried out according to the guidelines of the institutional ethics committee of biomedical research.
2.2
Enzyme-linked immunosorbent assay
Nonsandwich ELISA was modified and standardized from the routine procedure of Schuurs and Weeman . Nonsandwich ELISA was applied rather than sandwich ELISA after a series of pilot runs that showed the nonsandwich ELISA yielded the most precise readings. EBV VCA solution (Wellcome, Beckenham, England) was used as a solid phase of ELISA. The VCA antigen was diluted 1:10 in carbonate-bicarbonate coating buffer (1.59 g/L carbonate and 2.93 g/L bicarbonate; Sigma, St. Louis, MO) at concentration 10 μ g/mL. This concentration was used consistently in all runs to avoid the affection of the results by concentration variations. After a series of standardization steps, 50 μ L/well of the antigen in coating buffer was added for overnight at 4°C. Next day, the remaining fluid was aspirated and wells were washed twice with washing buffer (phosphate-buffered saline from Sigma, 1% bovine serum albumin wt/vol from Merck, Darmstadt, Germany, 0.05% vol/vol Tween 20 from Merck). Blocking buffer (phosphate-buffered saline, 1% bovine serum albumin wt/vol) was then added for 1 hour. The second layer of ELISA composed 50 μL/well of nondiluted serum of HNCA patients and control subjects. The ELISA plate was incubated for 2 hours at 37°C. After washing step, the third layer of ELISA was composed of 50 μL/well of horseradish peroxidase–labeled rabbit antihuman IgG antibodies diluted 1:40 at original concentration of 2 mg/mL (ICN Immunologicals, Lisle, IL). After washing, 50 μL OPD-H 2 O 2 (Sanofi Diagnostics, Lyon, France) was added for 15 minutes as chromogen-substrate for the development of the enzymatic reaction of horseradish peroxidase. Enzyme-linked immunosorbent assay readings were immediately measured in terms of optical density (OD) using a microtiter reader A-1764A (Beckman, Fullerton, CA) at a wavelength of 492 nm. A standard curve was generated using serial dilutions of known amounts of standard human IgG antibodies (Wellcome). The minimal detection limits for anti-EBV VCA IgG antibodies was 0.1 ng/mL.
2.3
Calculation of cutoff values for EBV IgG seropositivity
Cutoff value is considered as the upper limit, above which all readings were considered as positive. Therefore, the first and second cutoff values were calculated as follows:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
