Cervical Carcinoma

Michael Chen,* Hong Wang,* Craig D. Woodworth,* Paolo Lusso,t Zwi Berneman,t Douglas Kingma,§ Gregory Delgado,11 and Joseph A. DiPaolo* From the Laboratory ofBiology and the Laboratory of Tumor Cell Biology,t National Cancer Institute, Bethesda, Maryland, the Laboratory ofExperimental Hematology,* University ofAntwerp, Antwerp, Belgium; the Laboratory of Pathology,§ National Cancer Institute; and the Department of Obstetrics and Gynecology,11 Loyola University Medical Center, Stritch School ofMedicine, Maywood, Illinois

A subset of human papilomaviruses (HPVs) is associated with the majority ofcervical cancers; however, cofactors appear to be requiredfor carcinogenicprogression ofHPV-induced neoplasia. As human herpesvirus-6 (HHV-6) was recently shown to infect cervical epithelial ceUs in vitro and activate transcription of HPV-transforming genes, human cervical dysplasia and cancers were analyzedfor the presence ofHHV-6 by multiple methods, including polymerase chain reaction, slot blot, Southern blot, and in situ hybridization. HHV-6 DNA sequences were detected in 6 of 72 cases ofsquamous cervical carcinoma and cervical intraepithelial neoplasia. HPV-16 was found in four of the HHV-6-positive cases (two squamous cervical carcinomas and two cervical intraepithelial neoplasias). None ofthe 30 normal cervices and biopsies ofpatients with cervicitis waspositiveforHHV-6DNA. These results are the first suggestion ofan in vivo association between HHV-6 and some cervical neoplasia (Am J Pathol 1994, 145:1509-1516) A subgroup of human papillomaviruses (HPVs) is important in the etiology of cervical cancer; however, they are not sufficient for the development of the disease. 1 The long latency between HPV infection and development of cervical cancer plus the low prevalence of cancer relative to the number of women positive for HPVs suggest that other etiologic agents and cellular factors must be involved in the progression to malignancy. Experimental evidence suggests that cellular changes including gene and chromosomal alterations may coincide with the acquisition of tumorigenicity. Gene and chromosomal alterations such as deletions of 1 1 p and 1 7p2 that eliminate tumor suppressor genes3 or loss of heterozygosity of genes on 3p4 contribute to the process of tumor development. These alterations demonstrating the recessive nature of molecular events associated with tumor progression, coupled with preferential HPV integration at 8q24 and its consequential activation of the myc gene,57 provide the basis for a model for understanding the genetic changes in cervical neoplasia.
Recently, human herpesvirus 6 (HHV-6) has been shown to productively infect HPV-immortalized or transformed cervical carcinoma cell lines, transactivate HPV gene expression, enhance expression of HPV RNAs encoding the viral oncoproteins E6 and E7, and accelerate tumor growth in nude mice.8 HHV-6 was first isolated in 1986 from the peripheral blood of patients with lymphoproliferative disorders and AIDS. 9 Seroepidemiologic studies indicate that HHV-6 infection is highly prevalent in the human population10 and is commonly acquired early in childhood.1" HHV-6 is the etiologic agent of exanthem subituM12 and is primarily a T-lymphotropic virus both in vitro13 and in vivo.14 However, other cell types also show variable susceptibility to infection. 15 In particular, epithelial cells from bronchial mucosa16 and salivary glands16'17 were found to be infected by HHV-6 in vivo. The detection of HHV-6 viral DNA sequences in some non-Hodgkin's18'19 as well as Hodgkin's lym-phomas20 suggests a possible role for HHV-6 in human neoplasia.
In symptomatic HIV-1 -infected women, a strong association between HPV infection and squamous intraepithelial lesions has been demonstrated.21 HIVinfected women have a higher risk for cervical cancer than women who are HIV negative. Both squamous cell neoplasia of the cervix and HIV-1 infection are, in part, sexually transmitted diseases. HHV-6 was suggested as a possible cofactor in HIV infection,22 and recent evidence indicates that HHV-6 infection is active and widespread in terminal AIDS patients. 23 The sum of the observations raises the possibility that HHV-6 may also interact with HPV in HIV-infected women and possibly contributes to the development of squamous cervical carcinoma (SCC). To determine the presence of HHV-6 in SCC, DNA or tissues from patients with cervical diseases were examined for the presence of HHV-6 and HPV DNA sequences.

Material and Methods Samples
Two groups of SCC and cervical intraepithelial neoplasia (CIN) from different geographic regions were analyzed. The first consisted of DNA obtained from 62 fresh surgical specimens from the Shanxi Cancer Hospital (China), 90% (56/62) from patients diagnosed as having SCC and 10% (6/62) with CIN Ill. In addition, DNA from biopsies of 14 Chinese patients with cervicitis was also analyzed. Tissues were homogenized and lysed, and high molecular weight DNA from the Chinese specimens was extracted in China according to a standard sodium dodecyl sulfate proteinase K procedure. 24 The second group consisted of paraffin-embedded tissue specimens from 10 American patients with SCC. Ten sections, 5 p thick, were obtained from each paraffin block with a new razor blade used for each tissue block. DNA was extracted from the paraffin tissue sections as previously described.25 As a control DNA was also obtained from the cervices of 16 normal American patients after hysterectomies for nonmalignant diseases (ie, fibroids or endometriosis). plification, with primer pairs HV-7 and HV-8 and probe HV-12.28 HPV DNA was detected by PCR amplification with either the HPV-16 E7-P1 and E7-P2 primer pair or HPV consensus primers MYO9 and Myl 1.3 HPV genital types were differentiated by restriction fragment analysis of the PCR products30 PCR products were electrophoresed through 2% MetaPhor agarose gels (FMC BioProducts, Rockland, ME) and visualized after ethidium bromide staining (1 pg/ml for 10 minutes). The gels were photographed with an MP4 Polaroid camera system. In each gel 100-bp DNA markers (Life Technologies, Gaithersburg, MD) were included in the first or last lane.

Southern Blot Hybridization
Genomic DNA from an SCC sample was digested with Hindill and electrophoresed in a 0.7% agarose gel, alongside a 1-kb DNA marker. The gel was stained with ethidium bromide, and DNA was transferred to nylon membranes (Genescreen Plus, New England Nuclear, Boston, MA). The membranes were hybridized with QuikHyb rapid hybridization solution (Stratagene, La Jolla, CA) containing 1 x 1 06 cpm/ml of denaturated [32P]-labeled pZVH-14,31 pZVB-70, or [32P]-labeled pBluescript probe and hybridized at 65 C for 3 hours as described.8 Subsequently, the membranes were washed under stringent conditions (0.1 x SSC, 0.1 % sodium dodecyl sulfate 65 C for 30 minutes) before autoradiography.
The amplified PCR DNA products were prehybridized for 20 minutes at 52.5 C in QuikHyb solution and hybridized in the same buffer with the addition of 2 x 1 06 dpm/ml of T4-polynucleotide kinase-labeled oligo HV-9 or S probe28'29 at 52.5 C for 2 hours, washed with 0.1 x SSC, 0.1% sodium dodecyl sulfate at room temperature for 20 minutes, and stringently at 55 C for 20 minutes. For HHV-7 detection, HV-12 oligo probe was used at a hybridization temperature of 55.5 C.28 DNA Amplification DNA was amplified by polymerase chain reaction (PCR) following standard protocols. 26 The presence of HHV-6 DNA in the samples was demonstrated by PCR, with external primer pairs H6-8 and H6-6 and internal primer pairs H6-7 and H6-6/7.27 For nested PCR 10 pl of external PCR reaction products were amplified with internal primer pairs H6-7 and H6-6/ 7,28 and Southern blot24 was performed to confirm the identity of the HHV-6 sequence. HHV-6 strain variants were differentiated by restriction analysis of PCR fragments and hybridized with S probe, as previously described.29 HHV-7 DNA was detected by PCR am-Slot Blot Hybridization SCC or CIN DNAs or recombinant HHV-6 plasmid DNAs were resuspended, serially diluted in 20 pl of TE buffer, and denatured by adding 200 pl of 0.5 N NaOH in 1.5 mol/L NaCI for 30 minutes, heated to 98 C for 3 minutes, and placed on ice. The DNA samples were transferred to nylon membrane filters with the use of a Minifold II slot blot system (Schleicher and Schuell, Keene, NH) and neutralized by washing with 200 pi of a 1 mol/L Tris, pH 7.4, 1.5 mol/L NaCI solution three times. DNA was crosslinked to the 1200 p Joules by using a UV Stratalinker (Stratagene, La Jolla, CA), and filters were hybridized with [32P]-labeled DNA probe as described.

In Situ Hybridization
Pathological sections from HHV-6-positive patients were examined by in situ techniques to determine the specific localization of HHV-6. As a negative control, sections were included from an Epstein-Barr virus (EBV)-positive carcinoma proven to be HHV-6 negative by PCR. As a positive control, an HPV-1 8-positive cervical carcinoma cell line, C4-1, was infected with the GS strain of HHV-6 and analyzed in parallel.32'33 HHV-6 clones pZVB-5634 and pZVH-1431 were used as a template for synthesis of labeled RNA by means of T7 RNA polymerase, with either [35S]-labeled GTP and CTP or digoxigenin-labeled UTR35 In situ hybridization with the isotopic probe was performed according to the method of Harper et al36 with some modification, and the hybridization with nonisotopic probe was performed as previously described.37

Detection of HHV-6 DNA Sequences in Cervical Cells of Patients with Cervical Diseases
Of the 72 patients diagnosed with SCC or CIN 111, 6 (8%) were positive for HHV-6 ( Table 1), as determined by PCR amplification and slot blot hybridization. None of the 14 patients with cervicitis or the 16 patients with normal cervices were positive by PCR for HHV-6 DNA. Results obtained with the 6 patients are given in Figure 1. The 549-bp fragments were obtained when external primer pair H6-8 and H6-6 was used ( Figure  1A, lanes 3 to 8). With internal primer pair H6-7 and H6-6/7, a 11 2-bp fragment was amplified (Figure 1 A, lanes 1 1 to 16). To confirm the specificity of PCR products, a nested PCR was performed in which the prod-uct amplified with the external primers was used as a template for PCR with the internal primers. A 1 1 2-bp fragment was produced as a result of nested PCR ( Figure 1A, lanes 17 to 22). HHV-6-infected HSB-2 cells served as a positive control ( Figure 1A, B, lanes 1 and 9) and noninfected H9 cells as a negative control (lanes 2 and 10). Southern blot hybridization of the amplified DNA with a specific oligonucleotide probe also confirmed that the PCR-amplified bands were HHV-6 specific (Figure 1 B). PCR analysis for the presence of HHV-7 sequence was negative for all samples (data not shown).
Recent evidence indicates the existence of at least two subgroups (variants) of HHV-6, designated A and B, which differ genetically, biologically, and immunologically. Group A isolates are similar to GS strain, and group B isolates are similar to Z-29 strain.33 HHV-6 strains can be differentiated into variants A or B by restriction fragment analysis of PCR products generated with the primer pairs A and C.29 Therefore, the six HHV-6-positive samples were amplified with primers A and C.
Two of the six HHV-6-positive patients (2 and 6) had variant A; and patients 1, 3, 4, and 5 had variant B (summarized in Table 1). Results of three of the six positive cases (lanes 3 to 8) and HSB-2 control (lanes 1 and 2) are shown (Figure 2A, B). The 830-bp PCR products were purified and digested with Hindlll. Restriction fragments were visualized with ethidium bromide staining (panel A) or Southern blotting and hybridization with S probe (panel B). The nondigested 830-bp PCR products from the HSB-2-positive control and three HHV-6-positive cases are shown in lanes 1, 3, 5, and 7 (Figure 2A, B). Restriction analysis with Hindlil was consistent with the lack of an internal restriction site, indicating that the virus in HSB-2 and patient 2 belonged to variant A (lanes 2 and 4). DNA samples of patients 1 and 5 were cleaved by Hind ll into 220-bp and 610-bp fragments, indicating the presence of variant B (lanes 6 and 8). The 610-bp fragments hybridized with the S probe but not with the  Subtyping ofHHV-6 variants by endonuclease restriction of PCR prodtucts. tising primers A and C, 100 ng of DNA from HSB-2 (positive control), two SCCs and one CIN uere amplified by PCR. Products were digested with restriction enzymes, electrophoresed, and visualized by ethidium bromide staining (A) or Souithern blot transfer; hybridization with [-2P]-labeled probe S (B). Lanes I and 2 are from HSB-2-positive control; lanes 3, 4, 5, and 6from SCCs ofpatients 2 and I (see Table 1), respectively; lanes 7 and 8from a CIN III ofpatient 5 (Table 1). Lane M, 100-bp marker. Lanes 1, 3, 5, and 7 unere amplified 830-bp products uwithout restriction enzyme digestion. Lanes Figure  5B) revealed nuclear staining with nucleolar sparing, a characteristic staining pattern previously described in hybridization studies for other herpesviruses, namely EBV. 35 The HHV-6 riboprobe showed no cross-reactivity with EBV, as demonstrated by the absence of hybridization to an EBV-positive nasopharyngeal carcinoma metastatic to lymph node ( Figure  5D). As a positive control, the C4-1 SCC cell line32 was infected with HHV-68 and examined at subpassage 16 ( Figure 5C). The C4-1 HHV-6 line has episomal HHV-6 as demonstrated by fluorescent visualization on interphase nuclei and confirmed by Gardella gel technique.8 These cells were analyzed with a radiolabeled probe. The presence of silver grains (arrowheads) indicates HHV-6 ( Figure 5C).

HPV Detection
The incidence of HPV genital types in 72 patients with SCC or CIN IlIl demonstrated by using HPV Li consensus primer pairs for the amplification of genital HPV sequences was 81 % (58/72), of which 81% (47/ 58) were HPV-16. Of the 14 patients with cervicitis, 3 were positive for HPV, with one having HPV-16. None of 16 normal American patients was positive for HPV. : . samples was confirmed by using HPV-1 6 E7 PCR amplification, which produced 142-bp fragments (data not shown).

Discussion
This is the first report demonstrating the presence of both HHV-6 and HPV-16 in SCC and CINs in vivo. No evidence for presence of HHV-6 was found in normal cervices or biopsies from women with cervicitis. HHV-6 is a ubiquitous virus as indicated by seroepidemiological results and by detection by PCR in peripheral blood leucocytes.38 After primary infection, HHV-6 becomes latent and persists in specific tissues for life. The current data examining nonmalignant cervical-derived material suggests that the cervix is probably not a normal reservoir for HHV-6 because all 30 samples tested were negative. The prevalence of HHV-6 viral sequences detected in Chinese and American patients was 4/62 and 2/10, respectively. It will be interesting to determine whether a larger number of specimens from American patients will yield higher positivity than noted for China. Although the number is low, a similar low level of HHV-6-positive cases was found in Hodgkin's lymphoma (12%; 3/25)14 and non-Hodgkin's lymphoma (3.8%; 2/53).19 HHV-6 can directly infect cervical epithelial cells immortalized by HPV and SCC cells in vitro and can transactivate the HPV-16 and HPV-18 promoters, resulting in increased expression of the HPV oncogenes E6 and E7.8 In addition, SCC cells infected in vitro with HHV-6 exhibited accelerated tumorigenicity in vivo after inoculation into immunodeficient mice. After PCR, 1/10 of the product was either undigested (lane 1) or digested (lanes 2 to 8) with the following restriction enzymes: BamHI, DdeI, HaeIII, Hinfp, PstI, RsaI, and Sau3A. After electrophoresis, the gels were stained and photographed. M, 100-bp DNA molecular weight marker.
However, because of the small number of HHV-6positive cases including two HPV-negative cases, the contribution of HHV-6 to cervical cancer will require further investigation.
The possible role of HHV-6 in human neoplasia is still a matter of controversy. To date, no definitive etiological association has been established. However, in vitro evidence conclusively showed that the completion of the lytic cycle in lymphocytes results invariably in cell death whereas no convincing in vitro transformation was ever achieved in human T lymphocytes, the major target cells for HHV-6.13 14 Epithelial cells, like fibroblasts (28 and Lusso, unpublished observation), do not efficiently sustain replication of HHV-6. As previously observed,8 only a small proportion of cells exposed to HHV-6 exhibit signs of viral replication, and such signs are usually undetectable after a few S-0 passages in vitro or xenotransplantation into mice. Thus, HHV-6 is likely to establish a latent or nonproductive infection in cervical epithelial cells but still be able to transactivate cellular and/or viral oncogenes present in the host cell.
In recent years several studies have demonstrated that SCC has an enhanced incidence and a more severe disease course in women infected by HIV.39 This association is unlikely to be merely secondary to the immunosuppression of AIDS because several other neoplasias do not show a similar trend in HIVinfected individuals. Thus, reactivation of latent HHV-6 and/or de novo HHV-6 infection may have a role in HIV-associated cervical malignancy. Further investigation, including studies in experimental animal models, is needed to definitively elucidate the possible role of HHV-6 in human cervical cancer.