| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Oncology, Markers, Clinical Correlates |
Department of Pediatrics and School of Public Health, Columbia University, New York, New York [M. O.]; Instituto de Fisiología Celular, Universidad Nacional Autonoma de México, México D.F., Mexico [V. P. C.]; Departments of Pathology [C. R.], Social Work [E. L.], and Pediatric Oncology [C. L.], Instituto Nacional de Pediatría, México D.F., Mexico; Ophthalmic Oncology Center, New York Presbyterian Hospital-Cornell University Medical Center, New York, New York 10021 [D. H. A.]; and Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York 10021 [I. O., W. G., C. C-C.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
There is considerable overlap in the epidemiological risk factors for the development of retinoblastoma and HPV3 infection. The use of barrier methods of contraception is associated with a reduced incidence of both retinoblastoma (3) and HPV16/HPV18 infection (4 , 5) . As with the likelihood of having a child with retinoblastoma, the likelihood of infection with HPV is inversely correlated with a woman’s educational level and socioeconomic status (6) . Diaphragms and condoms are significantly protective for the development of cervical carcinoma (7, 8, 9, 10) , although some studies have not found them to prevent HPV infection (11) . Most cervical carcinoma is caused by infection with HPV (12) . In addition, there is considerable overlap between those countries in which the relative incidence of retinoblastoma is greatest and those in which the incidence of cervical carcinoma is highest (1 , 13 , 14) .
Retinoblastoma cells lack functional pRB (15) . pRB binds and inactivates transcription factors, such as E2F members, thereby regulating cell cycle progression (16) . Mutations in the retinoblastoma gene (RB1) result in the absence of functional pRB and a concomitant increase of unbound E2F proteins. Although most cases of retinoblastoma have been found to have RB1 mutations, it has been reported that between 17% and 80% of nonfamilial cases have an intact RB1 gene (17, 18, 19) , suggesting the existence of an alternative mechanism for pRB inactivation. The HPV E7 protein binds to and inactivates pRB (20) . Transgenic HPV16 mice overexpressing E6 and E7 oncoproteins possess an intact retinoblastoma gene, but the ectopic expression of E6 and E7 in the retina leads to pRB inactivation and the development of retinoblastoma-like lesions (21) . These epidemiological and molecular data suggest that HPV may play a role in the development of sporadic retinoblastoma. Therefore, we examined tumor DNA from children with nonfamilial retinoblastoma for the presence of HPV sequences and pRB expression status, correlating laboratory results with clinicopathological features.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Clinical stage, as defined by the St. Jude’s staging system (22) , was determined from the clinical record of the patient and from the histopathological report given for the enucleated eye. H&E stains of the enucleated eye were reviewed independently for histopathological staging by the INP study pathologist (C. R.). One child was found to have a benign retinal defect, and this child’s tissue was then used as a negative control. The specimens from the remaining 39 children were diagnosed as retinoblastoma.
Extreme precautions were taken to avoid contamination of tissue samples. The enucleated eye was handled under sterile conditions. The sterile portion of the tumor was embedded in OCT and frozen immediately in liquid nitrogen. Samples were coded before processing to protect patient confidentiality and to assure blinding of laboratory personnel with respect to the clinical characteristics of the tumors. Tissue was then processed under sterile conditions for IHC and for DNA extraction as follows: samples were sectioned using a new disposable sterile blade for each tumor specimen; the cryostat chamber was cleaned completely with alcohol and sterile gauze before processing each specimen, and the field covering was discarded after each specimen processing; and the forceps used for handling the rolls of OCT/tissue were cleansed with HCl and autoclaved after each sample was processed. In addition, no more than two tumor samples were cut on any given day. During that period, no other known HPV-containing tumors, such as cervical cancer and laryngeal papillomas, were processed in our research laboratory, thus decreasing the potential for contamination of these samples. Tumor specimens were cut in 5-µm sections such that the first and last sections were processed for H&E staining, and the remaining 10 sections were saved on coated slides for IHC. In addition, 20-µm sections were placed in sterile Eppendorf tubes for DNA extraction. The H&E-stained slides were examined independently for confirmation of tumor and histopathological evaluations including extent of tumor necrosis.
HPV DNA Analysis by PCR.
DNA was extracted from the 39 primary retinoblastoma tissues using a
nonorganic method (Oncor, Gaithersburg, MD) and filter-tipped pipettes.
Amplifiable DNA was verified in all samples using ß-globin 2 gene
primers PC04 and GH20 (Perkin-Elmer, Foster City, CA). PCR conditions
consisted of: (a) preincubation at 97°C for 4 min;
(b) 50 cycles of 94°C for 30 s, 55°C for 30 s,
and 72°C for 30 s; and (c) elongation at 72°C for 3
min. Tumor DNA (100 ng) was examined for the presence of HPV-like
sequences using PCR with HPV L1 (MY09/MY11; Ref. 23
)
degenerate consensus primers (Perkin-Elmer). One nonneoplastic
enucleated eye and DNA blank samples were used as negative controls.
DNA extracted from SiHa and HeLa cell lines was used as a positive
control. To avoid contamination, the following precautions were taken:
(a) SiHa and HeLa DNA was handled in a separate room;
(b) tumor DNA was added as the final reagent before adding
Taq polymerase (Amplitaq; Perkin-Elmer); (c) SiHa and HeLa
DNA was added after the Taq polymerase; and (d) filtered
pipette tips were used. Conditions for the HPV-specific PCR reactions
consisted of preincubation at 94°C for 5 min; 40 cycles of 94°C for
1 min, 56°C for 1 min, and 72°C for 1 min; and elongation for 5 min
at 72°C. PCR products were analyzed on 2% agarose gels visualized
with ethidium bromide. PCR-amplified fragments containing the L1
sequence were expected to appear between 448 and 454 bp in length.
Southern blotting was carried out with
[
-32P]dCTP-labeled probes that were specific
for either HPV18 or HPV16. These probes were generated by using DNA
from the SiHa (HPV16) and HeLa (HPV18) cell line in PCR reactions with
species-specific oligonucleotide primers that were internal to the
degenerate primers in L1. The primers used were checked using GenBank
and control tissues to ensure that they would not amplify any human or
other non-HPV viral sequences. No other HPV subtypes, besides HPV16 and
HPV18, were recognized with either of these primer sets. For
HPV16, the internal primer sequences were 5'-TACCTACGACATGGGGAGGA-3'
and 5'-TGACAAGCAATTGCCTGGGT-3'. The sequences for the HPV18 internal
primers were 5'-CCTGGGCAATATGATGCTAD-3' and 5'-CCTTATTTTCAGCCGGTGCA-3'.
Conditions for the PCR reactions consisted of preincubation at 95°C
for 5 min; 40 cycles of 94°C for 1 min, 56°C for 1 min, and
72°C for 1 min; and elongation for 5 min at 72°C. Membranes were
hybridized overnight with the HPV16 internal probes and then washed to
a final stringency of 0.1% SSC/0.1% SDS at 52°C for 30 min and
exposed to X-ray films (Kodak, Rochester, NY) at -80°C with
intensifying screens for approximately 40 h. The membranes were
then stripped and rehybridized with the HPV18 internal probe, using the
same procedure.
Immunohistochemical Analysis.
IHC with avidin-biotin peroxidase complexes (Vector Laboratories,
Burlingame, CA) was performed using Mabs to pRB [clone Ab-5 (detected
underphosphorylated and hyperphosphorylated pRB proteins; 1 µg/ml
dilution; Oncogene/Calbiochem, Cambridge, MA), clone G99-549 (detected
only underphosphorylated pRB products; 5 µg/ml dilution; PharMingen,
San Diego, CA), Ki67 (MIB-1; 1 µg/ml dilution; Immunotech Co.,
Marseille, France), and E2F-1 (1 µg/ml dilution; Santa Cruz
Biotechnology, Santa Cruz, CA)]. Tumors were reported to have
undetectable pRB if no tumor cell nuclei were stained with Ab-5
antibody, whereas endothelial cells and other normal cells were
stained. Tumors were scored as having detectable pRB if >1% of
tumor cell nuclei stained with Ab-5. Detection of underphosphorylated
pRB was defined as tumors having >10% detectable nuclear staining
with G99-549 antibody. Slides were examined independently by two
pathologists who were blinded to the HPV status of the tumor tissues
and to the clinical stage of the patient. For Ki67 and E2F1, tumors
were scored according to the percentage of cells stained over five
high-power fields. Cases were excluded from analysis if the slides were
not interpretable because of necrotic tissue.
Statistical Analysis.
Statistical analysis was done using SPSS version 8.0 software
(SPSS, Inc. Chicago, IL). HPV status was classified as present or
absent. Ki67 and E2F1 were analyzed as continuous variables for the
percentage of tumor cells stained with antibody. Clinical stage was
analyzed as an ordinal variable (four stages) or was grouped into two
categories, tumor confined to the globe or retina (stage < 3) and
extraocular or metastatic disease (stage 
3). The relationships
between HPV status and the presence of detectable pRB and between HPV
status and clinical stage were examined with 
2
analysis and Fisher’s exact test. The relationship between the
percentage of Ki67 expression and HPV status was examined using
Student’s t test. The relationship between Ki67 expression
and clinical stage was examined using linear regression. The
significance level of = 0.05 was chosen. Data on prenatal
exposures were obtained from parental interviews. These were compared
with data on the presence of HPV DNA in tumor DNA using a case-series
design (24)
. Demographic characteristics were compared
between cases with and without HPV DNA using 2
analysis for categorical variables and Student’s t test for
continuous variables. Those variables such as monthly income, which
were not normally distributed, were taken to the square root, and
statistical analysis was then done on the square root values. Those
variables significantly associated with the presence of HPV DNA in
univariate analysis were analyzed in logistic regression for
multivariate analysis using an level of 0.05 for significance. All
tests of statistical significance were two-sided.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), and HPV18 sequences were detected in
11 of 39 (28.2%) cases. In one case, we identified sequences of both
HPV16 and HPV18. Negative controls, DNA blank specimens, and benign
retinal disease did not demonstrate amplified DNA in the PCR
experiments using L1 primers. PCR reactions were carried out in
duplicate to verify reproducibility. PCR-amplified fragments created
with L1 primers and DNA extracted from HeLa and SiHa cells were
hybridized with the HPV probes, confirming their specificity (Fig. 1)
.
|
, although this association did
not achieve significance (OR, 3.0; P = 0.2). Three of
five (60%) tumors with detectable pRB were HPV positive, whereas only
11 of those 34 (32%) tumors without detectable pRB had HPV DNA. Three
of 14 (21%) HPV-positive tumors had detectable pRB, whereas only 2 of
24 (8%) HPV-negative tumors had detectable pRB.
|
Clinical and Demographic Associations with the Presence of HPV.
Of the 39 children, 25 (64%) had unilateral disease, and 14 (36%) had
bilateral disease. Three children (7.7%) had metastatic disease (stage
4), 14 (36%) children had extraocular disease (stage 3), 21 (56%) had
ocular disease (stage 2), and 1 (2.6%) had retinal disease (stage 1).
The Ki67 proliferative index of tumors increased significantly with
increasing clinical stage (ß = 0.39; P = 0.025).
Children with HPV-positive tumors were significantly less likely to
have advanced-stage disease (Table 1)
.
The relative risk of having extraocular disease (with or without
metastatic disease) if the tumor did not have HPV DNA was 9.0 (95% CI,
1.7–49.1; P = 0.008). Among the 14 cases positive for
HPV, no cases were stage 4 (metastatic), and only 2 cases were stage 3
(orbital), whereas the majority were localized disease [ocular (11
cases) or retinal (1 case) disease]. Among the 25 cases without HPV
DNA, 3 cases had metastatic disease (stage 4), 12 cases had orbital
disease (stage 3), and 10 cases had ocular disease (stage II). The
proportion of cases with bilateral disease was equal for tumors with
HPV DNA and without HPV DNA (OR, 1.2; 95% CI, 0.30–4.69;
P = 0.81). Mortality was not significantly increased
among the HPV-negative cases (OR, 1.7; 95% CI, 0.16–18.0;
P = 0.7). However, there were only four known deaths
among the 39 children under study.
|
Results of case-series analysis of demographic variables associated
with the presence of HPV DNA in the tumors are shown in Table 1
.
Demographic data from the questionnaire interviews demonstrated that
children with HPV DNA in their tumors had lower birth weight and were
breastfed for fewer months than children whose tumors did not contain
HPV. Mothers of children with HPV DNA in their tumors were more likely
to have been diagnosed with a vaginal infection, although this
difference was not statistically significant. There was also no
significant difference in mean per capita monthly income (360
versus 220 N. pesos), mean number of years of
maternal schooling (6.0 years versus 5.5 years), mean
maternal age at delivery, or birth order of the index child between the
two groups. Fathers of children with HPV DNA in their tumors were older
than fathers of children whose tumors did not contain HPV DNA. In
multivariate analysis using logistic regression, only birth weight
remained significantly associated with the presence of HPV DNA after
adjustment for paternal age and maternal history of vaginal infection.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A role for HPV in retinoblastoma could explain the overlap in risk factors for retinoblastoma and cervical carcinoma. Cervical cancer is the major cause of mortality among Mexican women of child-bearing age (27) , and the prevalence of asymptomatic HPV infection in Mexican women is higher than that reported in other countries in the Americas (28) . A low level of education among women and a lack of use of barrier methods of contraception are both associated with an increased risk of cervical HPV infection (6 , 29) , as well as with a significantly increased risk of having a child with retinoblastoma (3) . Poverty among urban Latin American women has been associated with increased incidence of persistent (asymptomatic) cervical HPV infection (30) and invasive cervical carcinoma (31) . Poverty has also been associated with incidence of retinoblastoma in New York City,4 in Latin America (32) , and in the geographic distribution of worldwide incidence of retinoblastoma (1) . Additionally, the association of an increasing incidence of retinoblastoma in countries severely affected by the HIV epidemic is also consistent with the possible role of HPV because women with HIV have higher rates of persistent HPV infection (33) . Therefore, the presence of HPV DNA in retinoblastoma tissues is consistent with the overlap in epidemiologically determined risk factors for both the development of this disease and the detection of HPV infection.
In our series, children whose tumors contained HPV had significantly lower birth weights than children whose tumors did not contain HPV. This finding may suggest that these children may have been nutritionally deprived during pregnancy when compared with children with retinoblastoma without HPV, or it may be consistent with perinatal infection. The shorter duration of breast feeding in the children with HPV DNA in their tumors suggests that children with HPV DNA may have been living in less rural living conditions because breast feeding is known to be prolonged in more rural communities in Mexico. It is possible that those parents living in less rural communities might have had an increased number of sexual contacts. The older age of the fathers of children with HPV might be a proxy for an increased number of lifetime sexual contacts and an increased probability of infection (34) . The fact that there was no difference in maternal age at delivery and no difference in the birth order of the index child suggests that there was no difference in the age at which the mothers began sexual activity. This is consistent with the study by Hernandez-Avila et al. (28) , who found that the age at which Mexican women began sexual activity was not related to their risk of developing cervical cancer. The nonsignificant increase in self-reported vaginal infections in the mothers does suggest the possibility that this population may be more at risk for sexually transmitted disease. The incidence of urethral infections in fathers and information on paternal sexual histories were not available. Because of the transient nature of most HPV cervical infections, we would not be able to document maternal prenatal infection with HPV retrospectively, although serological testing of these women for the presence of anti-HPV VLP antibodies might be helpful. Unfortunately, we do not have sera available for either the mothers or the children.
The proportion of HPV16 to HPV18 in these tumors is the inverse of what would be expected given the distribution of high-risk HPV species among cervical cancer specimens from Mexican women. Among Mexican women with invasive cervical carcinoma, HPV16 has been found in 30–48% of cases, whereas HPV18 has been found in only 7–10% of cases (28 , 35) . However, the presence of HPV18 DNA in cervical cancer is associated with a higher incidence of adenocarcinoma and with significantly poorer survival than the presence of HPV16 DNA (36) . The apparent excess of HPV18 in our study samples may be due to an increased oncogenic potential of HPV18 in retinal tissue. In addition, initial perinatal transmission rates for mothers infected with HPV18 are higher than those for mothers infected with HPV16 (37) .
The presence of DNA from the high-risk HPV subtypes suggests that oncogenesis in retinoblastoma could occur in part as a result of E7 inactivation of pRB. There is evidence that osteogenic sarcoma, another pediatric tumor associated with RB alterations, may develop as a result of viral inactivation of pRB. Recent studies have demonstrated the presence of the pRB pocket-binding domain of the SV40 virus in 32% of osteogenic sarcomas examined (38) . Detection of HPV DNA in retinoblastoma tissue appears to be correlated with pRB expression; however, these pRB products appear to be hyperphosphorylated and are probably deficient in their activity, as reflected by the phenotypes observed using phosphorylation and E2F1-specific antibodies. Similarly, the finding of an equivalent number of cells staining with the anti-E2F1 antibody, regardless of HPV status, is consistent with the finding that HPV E7 activates E1A-responsive promoters on the E2F1 gene, thereby causing its release from the macromolecular complex that binds it (39 , 40) .
The Ki67 proliferative index of the retinoblastoma lesions studied increased significantly with increasing stage and was significantly higher in HPV-negative cases than in HPV-positive cases. It may be possible that in some tumors, E7 binds and inactivates most but not all pRB products, such that partial pRB activity remains. This functional pRB would therefore be able to keep some of the tumor cells from entering into the cycle.
The fact that HPV-positive tumors were less advanced in stage is particularly interesting in light of recent findings by Andl et al. (41) . In their study examining the presence of HPV DNA in squamous cell carcinoma tumors of the head and neck region, tumors containing HPV DNA had a more favorable prognosis than expected, given the clinical staging (41) . Additionally, some studies have demonstrated that HPV-negative cervical carcinomas appear to progress more rapidly, have a higher metastatic potential (12) , and have a higher mortality rate (42) than HPV-positive tumors, although others have not found this association (43) .
Four cases of nonfamilial bilateral disease included in the present study displayed HPV sequences. The ratio of bilateral to unilateral disease among the tumors containing HPV DNA was the same as the ratio found in the general population of patients with retinoblastoma. Although bilateral disease is usually caused by germ-line mutations in RB1, between 17% and 77% of cases of bilateral disease that have been sequenced have not been found to contain germ-line mutations in RB1 or its promoter (17 , 19 , 44) . Again, the proportion of bilateral tumors found to have germ-line RB1 mutations appears to vary by geographic origin, suggesting that the proportion of bilateral disease caused by RB1 mutations may also vary geographically. Our findings suggest that some bilateral disease in tumors with intact RB1 may be caused by viral inactivation of pRB.
Sequences of HPV DNA are detectable by PCR-based assays in some retinoblastomas. The presence of viral DNA from the high-risk HPV types suggests that oncogenesis in certain retinoblastoma cases may occur as a result of E7 inactivation of pRB. This is consistent with the overlap in epidemiological risk factors between HPV infection (29 , 30) and retinoblastoma (1) . Detection of HPV DNA in retinoblastoma tissue appears to be associated with the presence of pRB expression, lower proliferative indices, and lower clinical stage. Additional studies are necessary to investigate the relationship of HPV and the presence of its E6 and E7 oncogenes to the development and progression of retinoblastoma.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by NIH Grants NCI-CA-47538 and
NCI-CA-47179 (to C. C-C.) and the Bowen Brooks Foundation of the New
York Academy of Medicine (M. O.). 
2 To whom requests for reprints should be
addressed, at Memorial Sloan-Kettering Cancer Center, Department of
Pathology, 1275 York Avenue, New York, NY 10021. Phone:
(212) 639-7746; Fax: (212) 794-3186; E-mail: cordon-c{at}mskcc.org
3 The abbreviations used are: HPV, human papilloma
virus; IHC, immunohistochemistry; OR, odds ratio; CI, confidence
interval; INP, Instituto Nacional de Pediatría; Mab, monoclonal
antibody.
4 M. Orjuela, unpublished observations.
Received 3/27/00; revised 8/ 8/00; accepted 8/ 8/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Chintagumpala, P. Chevez-Barrios, E. A. Paysse, S. E. Plon, and R. Hurwitz Retinoblastoma: Review of Current Management Oncologist, October 1, 2007; 12(10): 1237 - 1246. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. Gillison and K. V. Shah Chapter 9: Role of Mucosal Human Papillomavirus in Nongenital Cancers J Natl Cancer Inst Monographs, June 1, 2003; 2003(31): 57 - 65. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
