Frequent detection of human cytomegalovirus in neuroblastoma: A ...

IJC International Journal of Cancer

Frequent detection of human cytomegalovirus in neuroblastoma: A novel therapeutic target? Nina Wolmer-Solberg1, Ninib Baryawno2, Afsar Rahbar1, Dieter Fuchs1, Jenny Odeberg1, Chato Taher1, Vanessa Wilhelmi1, €rnsson2,4, John Inge Johnsen2, Jelena Milosevic2, Abdul-Aleem Mohammad1, Tommy Martinsson3, Baldur Sveinbjo 2 1 €derberg-Naucle r Per Kogner and Cecilia So 1

Department of Medicine, Center for Molecular Medicine, Karolinska University Hospital in Solna, Karolinska Institutet, Stockholm, Sweden Department of Women’s and Children’s Health, Childhood Cancer Research Unit, Karolinska Institutet, Stockholm, Sweden 3 €teborg The Sahlgrenska Academy at the University of Gothenburg, Genomics Core Facility, Go 4 €, Tromso €, Norway Immunology Research Group, Faculty of Medicine, University of Tromso

Neuroblastoma is the most common and deadly tumor of childhood, where new therapy options for patients with high-risk disease are highly warranted. Human cytomegalovirus (HCMV) is prevalent in the human population and has recently been implicated in different cancer forms where it may provide mechanisms for oncogenic transformation, oncomodulation and tumor cell immune evasion. Here we show that the majority of primary neuroblastomas and neuroblastoma cell lines are infected with HCMV. Our analysis show that HCMV immediate-early protein was expressed in 100% of 36 primary neuroblastoma samples, and HCMV late protein was expressed in 92%. However, no infectious virus was detected in primary neuroblastoma tissue extracts. Remarkably, all six human neuroblastoma cell lines investigated contained CMV DNA and expressed HCMV proteins. HCMV proteins were expressed in neuroblastoma cells expressing the proposed stem cell markers CD133 and CD44. When engrafted into NMRI nu/nu mice, human neuroblastoma cells expressed HCMV DNA, RNA and proteins but did not produce infectious virus. The HCMV-specific antiviral drug valganciclovir significantly reduced viral protein expression and cell growth both in vitro and in vivo. These findings indicate that HCMV is important for the pathogenesis of neuroblastoma and that anti-viral therapy may be a novel adjuvant treatment option for children with neuroblastoma.

Neuroblastoma is an embryonal tumor that arises in the sympathetic nervous system, including the adrenal medulla, sympathetic ganglia, and paraganglia. These tumors are the most common and deadly tumors of childhood, and clinically and biologically heterogeneous—one favorable subset is prone to spontaneous apoptosis or differentiation over time with no or little therapy—whereas most neuroblastomas are metastasizing high-risk tumors that are difficult to cure with current treatment options.1 Except for a small minority of patients with specific germline mutations of the anaplastic lymphoma

kinase gene,2 there is no clear evidence for a familial predisposition to neuroblastoma. Thus, modifiable risk factors, including viruses as an additional triggering factor, may be important in the etiology of these tumors. Relapse of neuroblastoma has been related to active infection with human cytomegalovirus (HCMV).3 Viral particles in urine and HCMV-specific antibodies and DNA in the blood at time of diagnosis have been reported.3,4 In longterm in vitro culture, neuroblastoma cells persistently infected with HCMV exhibit enhanced Mycn expression, tumor

Key words: human cytomegalovirus, neuroblastoma, valganciclovir Additional Supporting Information may be found in the online version of this article. The first two authors contributed equally to this work and share first and second authorship. Last two authors contributed equally to this work and share senior authorship. Grant sponsor: Torsten and Ragnar S€oderbergs Stiftelse and Ragnar S€oderbergs Foundation; Grant sponsor: The Swedish Children’s Cancer Foundation; Grant numbers: PROJ12/115 and PROJ10/121; Grant sponsor: The Swedish Cancer Society; Grant number: 100637; Grant sponsor: The Swedish Research Council; Grant numbers: K2010-56X-12615-13-3 and K2009-56P-20937-03-1; Grant sponsors: Stichting af Jochnicks Foundation, BILTEMA Foundation, Sten A. Olssons Foundation, the M€arta and Gunnar V. Philipson Foundation, The Mary Beve Foundation, The Hans and M€arit Rausing Charitable Fund, The Dðmman Foundation, Swedish Society for Medical Research (SLS), Goljes Memory Foundation, Magnus Bergvalls Foundation, Swedish Society for Medical Research (SSMF) and Tore Nilsons Foundation. DOI: 10.1002/ijc.28265 History: Received 4 Dec 2012; Accepted 24 Apr 2013; Online 10 May 2013 Correspondence to: Cecilia S€oderberg-Naucler, CMM L8:03, Group of Cellular and Molecular Immunology, Karolinska Institutet, Karolinska University Hospital, SE-171 76 Stockholm, Sweden, Tel.: 146-8-517-79896, Fax: 146-8-31-3147, E-mail: [email protected]

C 2013 UICC Int. J. Cancer: 133, 2351–2361 (2013) V

Infectious Causes of Cancer

2

2352

Human cytomegalovirus in neuroblastoma


What’s new? Relapse and invasiveness of neuroblastoma, a frequently fatal cancer of early childhood, may be linked to the presence of human cytomegalovirus (HCMV), one of the most common congenital viral infections known. In this study, HCMV was observed in primary neuroblastoma tumors and in six neuroblastoma cell lines. Although no infectious virus was isolated from tumors, the HCMV-specific drug valganciclovir significantly reduced viral protein expression and tumor cell growth both in vitro and in vivo. The results suggest that HCMV may be important in the pathogenesis of neuroblastoma and that antiviral therapy may represent a possible future treatment option for affected children. We have shown that all examined primary neuroblastoma tumors and six neuroblastoma cell lines were infected with HCMV, but no infectious virus was isolated from tumors. The HCMV-specific drug Valganciclovir significantly reduced viral protein expression and tumor cell growth in vitro and in vivo. Thus, HCMV may be important in the pathogenesis of neuroblastoma and anti-viral therapy may provide a novel treatment option for children with neuroblastoma.

growth, invasiveness, and metastasis.5,6 About 20 to 30% of neuroblastomas exhibit MYCN amplification, which is associated with aggressive, poorly differentiated tumors that progress rapidly and have a poor prognosis.1 HCMV infects 70 to 90% of the world’s population, and is found in 0.5 to 2.2% of newborns, making it the most common congenital viral infection. Symptoms of a primary HCMV infection are generally mild or asymptomatic in immunocompetent individuals but may be life-threatening in immunocompromised patients such as transplant recipients and AIDS patients. The virus is spread through all bodily fluids and establishes a life-long latent/persistent infection. Reactivation from latency appears to be triggered by inflammation, which the virus can itself initiate by inducing cytokine and chemokine production and by enhancing leukotriene and prostaglandin synthesis, which induce expression of 5-lipoxygenase and cyclooxygenase-2 (COX-2). We recently demonstrated that HCMV positive tumor cells express COX-2 in medulloblastoma tumors. Anti-viral therapy in combination with a COX-2 inhibitor resulted in a 72% reduction in tumor growth.7 The prevalence of HCMV is surprisingly high in several forms of cancer, including brain tumors and cancers of the colon, breast, and prostate,8–12 and rhabdomyosarcoma.13 The active viral infection is restricted to tumor cells and present in over 90% of these tumor types; healthy tissues from the same patients are consistently HCMV negative.8,9,11,12 Low levels of HCMV infection in glioblastomas is associated with improved survival,14,15 which suggest a role of HCMV in tumor progression. HCMV is not considered to be oncogenic, but it may exert oncomodulatory effects that contribute to cancer development.16 The 230-kb genome of HCMV has 252 open reading frames was believed to encode approximately 180 proteins. Of these, only about 50 are known to be essential for viral replication. Thus, most HCMV proteins fulfill other functions in the viral life cycle. Through sophisticated mechanisms, viral proteins control cellular and immunological functions relevant to tumor biology and immune surveillance.17 HCMV proteins control cellular differentiation, proliferation, and epigenetic functions, induce

migration and angiogenesis, inhibit apoptosis, and control immunological functions that help virus-infected tumor cells to avoid detection and elimination by the immune system and to support efficient viral replication. In this study, we report a high prevalence of HCMV in human neuroblastoma samples, cell lines, and xenografts. Inhibition of viral replication with ganciclovir significantly inhibited tumor growth both in vitro and in vivo, suggesting anti-HCMV treatment as a novel adjuvant therapeutic option for neuroblastoma patients.

Material and Methods Human neuroblastoma tumor tissue

Tumor tissue was collected from 40 patients, 36 with neuroblastoma including three non-tumor adrenal gland samples, and four with ganglioneuroma undergoing surgery at Astrid Lindgren Children’s Hospital, Stockholm, Sweden. Ethical approval was obtained from the Karolinska University Hospital Research Ethics Committee. Parents to eligible pediatric neuroblastoma patients gave informed consent to donate tumor tissue to be used for research purpose (approval no. 2009/1369-31/1, 03-736). Ethical permission to examine these tumor samples specifically for CMV was also obtained (approval no. 2008/628-31/2). The tissue was snap frozen in liquid nitrogen and stored at –80 C. For immunohistochemistry, tissues were thawed, fixed in paraformaldehyde, embedded in paraffin, sectioned at 5 lm, and placed on Superfrost Plus glass slides. Frozen tumor tissue from 24 patients was kept at –80 C until extraction of DNA, or RNA extraction (16 samples). Cell cultures

Human neuroblastoma cell lines (SK-N-BE(2), SK-N-AS, SKN-DZ, Kelly and IMR-32) and human chronic myelogenous leukemia cells (K562) were cultured in RPMI 1640 medium (Gibco BRL, Sundbyberg, Sweden). SH-SY5Y neuroblastoma cells and human fibroblasts were grown in MEM medium (Life Technologies, Paisley, Scotland). All media were supplemented with 10% fetal calf serum (Sigma-Aldrich, Solna, Sweden), 2 mM L-glutamine (Gibco BRL), and 100 U/ml C 2013 UICC Int. J. Cancer: 133, 2351–2361 (2013) V

2353

Wolmer-Solberg et al.

Virus isolation

Frozen neuroblastoma tissue (n 5 8) and neuroblastoma cell xenografts (of six cell lines) was homogenized by ultra sonication. Homogenate (0.5–1 ml) suspended in serum free medium and added to human fibroblasts, grown in six-well plates to a confluency of 50 to 70%. After 24 hr low serum medium was added. Half of the medium was exchanged twice weekly and after 2 weeks the cells were washed in PBS, fixed in cold methanol:acetone 1:1 for 10 min in room temperature and stained for IEA. Chemicals

Ganciclovir (Cymvene, Roche, Basel, Switzerland) and valganciclovir (Valcyte; Roche, Stockholm, Sweden) were dissolved in water according to manufacturer’s guidelines. All drugs were further diluted in Opti-MEM (Gibco BRL) to the desired in vitro concentration. Detection of HCMV in tumor specimens

For detailed methods regarding detection of HCMV proteins, DNA and RNA in tumor tissues, please see Supporting Information. Clonogenic assay

Tumor cell growth was assessed with in vitro clonogenic assays as described.18 After surface attachment, cells were treated with 150 or 300 lM ganciclovir in Opti-MEM (Life Technologies, Paisley, Scotland) for 48 hrs. The medium was then replaced with drug-free complete medium. When control plates reached sufficient number of colonies, cells were fixed in formaldehyde and stained with Giemsa (GIBCO BRL). The number of colonies consisting of >75 cells with 50% plating efficiency were counted manually. For each treatment combination, the surviving fraction was calculated as the ratio of the mean plating efficiency of treated cells to that of untreated control cells. Cell viability assay

Cell viability was measured using a colorimetric assay for 96well plates with 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4disulfophenyl)-2H-tetrazolium monosodium salt (WST-1) reagent (Roche). Neuroblastoma cells were seeded at a concentration of 1,250 cells per well, allowed to attach for 24 h and treated with ganciclovir daily for 7 days at the concentrations 300, 600, 900 or 1,800 lM. After 7 days, 10 ll of WST-1 were added and cells were incubated for an additional 3 hr. Cytotoxicity of ganciclovir at different concentrations was expressed as percentage of viable cells compared to controls. C 2013 UICC Int. J. Cancer: 133, 2351–2361 (2013) V

Xenografts and in vivo administration of valganciclovir

Female NMRI nu/nu mice (Taconic Laboratories, Ejby, Denmark) 4 to 6 weeks of age were maintained at five mice per cage and were given sterile water and food ad libitum. Each mouse was subcutaneously injected with SH-SY5Y neuroblastoma cells (20 3 106) in both hind legs. Treatment was started when palpable tumors reached a volume of 0.15 ml. Mice were randomized to receive either no treatment (n 5 10) or valganciclovir at 20 mg/kg (n 5 7) or 30 mg/kg (n 5 7). Valganciclovir was administered twice daily through a gastric feeding tube. Tumors were measured daily, and tumor volume was calculated as (width)2 3 length 3 0.44. Tumor volume index was calculated as the measured volume divided by the volume measured at start of treatment. Tumor weight was recorded at autopsy, after which a part of the tumor were fixed in formaldehyde and a part was frozen. All animal experiments were approved by the regional ethics committee for animal research (nos. N231/05 and N304/08) in accordance with animal protection regulations (SFS 1988:534) and regulations of the Swedish National Board for Laboratory Animals (SFS 1988:541). A total of 55 mice were used in this experiment. Statistical analysis

All statistical analyses were performed with GraphPad Prism software (GraphPad Software, San Diego, CA). For in vitro studies, the t-test (two-tailed) was used to determine whether the mean of a single sample differed significantly from control. Tumor growth and tumor weight was analyzed by Mann Whitney test. p < 0.05 was considered significant.

Results High prevalence of HCMV in primary neuroblastoma tumors

The link between HCMV and neuroblastoma was first proposed by Nigro et al., who observed a correlation between an active HCMV infection and neuroblastoma patients who excreted HCMV in urine at the time of clinical diagnosis.3,19 Here, we examined primary tumor biopsies from 36 children with neuroblastoma for HCMV. All 36 expressed HCMV immediate-early (IE) protein, and 33 (92%) expressed late protein; IE and late proteins were widely expressed in tumor cells (Fig. 1 and Table 1). The number of positive cells was estimated manually, and samples were graded according to percentage of cells infected with IE and late proteins (Fig. 1i and Table 1). Patients of all ages and genetic subsets and clinical risk groups were included20 and all contained HCMV. We also investigated four samples from differentiated ganglioneuromas and three samples from healthy adrenal tissues, two of which from children with neuroblastoma. All these samples were also positive for HCMV DNA and protein expression (Table 1). In the lack of an HCMV negative neuroblastoma tissue sample, we included an HCMV negative glioblastoma tumor sample as a control for the


penicillin/streptomycin (Gibco BRL) at 37 C in 5% CO2/95% air. The authenticity of the cell lines was verified by using the STR DNA profiling test (Genetica DNA Laboratories, Cincinnati). The latest verification was performed in September 2010.

2354


Table 1. HCMV protein expression in neuroblastic tumors Protein expression IE proteins/L proteins 1

Diagnosis

n

Age (mo) median; range

Sex M/F

0

11

21

31

41

Neuroblastoma2

36

18; 0–136

20/16

0/3

9/20

11/5

5/4

11/4

Low/intermediate Risk

26

14; 0–123

14/12

0/1

6/14

6/4

5/4

9/3

High risk3

10

23; 12–136

6/4

0/2

3/6

5/1

0/0

2/1

Ganglioneuroma2

4

20; 10–137

2/2

0/1

2/

0/1

1/0

0/0

Healthy adrenal6

3

24; 12–25

1/2

0/0

1/1

1/2

1/0

0/0

3

1

Age in completed months at diagnosis (median; range). According to INSS diagnostic criteria (International Neuroblastoma Staging System, Brodeur et al., 1993).20 3 According to INRG criteria (International Neuroblastoma Risk grouping, Cohn et al., 2009).21 4 Metastatic stage M and age >18 months, or MYCN-amplification. 5 Benign peripheral neuroblastic tumor. 6 Non-tumoral adrenal medulla.


2

immunohistochemical staining (Figs. 1d and 1h). Expression of HCMV IE (Fig. 1k) and late proteins was also demonstrated in two fresh neuroblastoma samples by flow cytometry, which revealed HCMV proteins in 14% of cells in one sample and 54% of cells in the other. We next performed western blot analyses of seven frozen primary neuroblastoma samples using an antibody to the IE proteins IE72 and IE86 (clone 810R) and found that IE72 but not IE86 were expressed in all primary neuroblastoma examined (Fig. 1l). However, we detected a band that is predicted to be IE55, which is a splice variant of IE86, in seven of seven samples (Fig. 1l). DNA was extracted from 24 frozen neuroblastoma specimens; 22/24 were positive for HCMV IE and/or pp150 genes by PCR (data not shown). Sixteen of 16 frozen samples were positive for HCMV RNA (data not shown). In five samples, HCMV IE DNA PCR products were sequenced and found to contain the HCMV genome, which all were distinctly different from four laboratory isolates (Merlin, AD169, TB40E and VR1814) (data not shown) and excluded the possibility of contamination. However, viral cultures from eight frozen samples were consistently negative (data not shown), implying defective HCMV replication in neuroblastomas, or very low virus presence, but the possibility of failure to grow the tumor associated virus by conventional techniques for virus isolation can not be excluded. These observations suggest that HCMV proteins rather than production of infectious virus may be involved in the pathogenesis of neuroblastoma. Detection of HCMV in neuroblastoma cell lines and neuroblastoma xenografts

We next analyzed six neuroblastoma cell lines (SH-SY5Y, SK-N-BE(2), SK-N-AS, SK-N-DZ, Kelly, and IMR-32) and found that all six were positive for HCMV DNA by PCR (Supporting Information Table 1). Fluorescence in situ hybridization (FISH) with an HCMV probe (Fig. 2a and Supporting Information Table 1) also showed HCMV DNA in all six cell lines. Human chronic myelogenous leukemia cells

(K562) and human fibroblasts (MRC-5) infected with HCMV AD169 served as negative and positive controls, respectively (Fig. 2a). The location of viral DNA differed in neuroblastoma cells compared to HCMV-infected fibroblasts; viral DNA was located in the periphery of the nucleus in tumor cells but was centrally located (sometimes in owl’s eye formation) in infected fibroblasts. Flow cytometric analysis of all cell lines further demonstrated HCMV IE antigen and pp65 expression in 0 to 84% of the cells (data not shown and Figs. 2b and 2c). The number of protein-positive cells varied considerably between cell lines and at different sampling times; within each cell line, the number of positive cells identified with antibodies against IE or pp65 was relatively constant at individual sampling occasions (Figs. 2b and 2c). At least once over 9 months, IMR-32, Kelly, and SY-SH5Y were negative for both HCMV IE antigen and pp65 by flow cytometry (n 5 2–7 sampling occasions for individual cell lines and was generally low after thawing of frozen cell lines), which implies a high variation of HCMV protein expression in neuroblastoma cell lines over time. We hypothesized that the consistent presence of HCMV DNA and proteins in long-term cultures of neuroblastoma cell lines may be linked to viral persistence in a cancer stem cell or a tumor initiating cell. No clear marker has been defined for tumor initiating cells in neuroblastoma tumors. We double stained neuroblastoma cell lines for HCMV IEA and CD133 or CD44; two potential stem cell markers that could be used for double staining with HCMV proteins. Only SK-N-DZ demonstrated high CD133 expression (84%), whereof 32% were double positive for CD133 and HCMV IEA (Fig. 2c). CD44 was present on 23% of SKN-BE(2), 32% of IMR-32, 2.2% of SH-SY5Y, 69% of SK-NAS and 4.1% of SK-N-SH cells. Double immunofluorescence revealed that 14% of SK-N-BE(2), 29% of IMR-32, 2% of SH-SY5Y, 34% of SK-N-AS and 4% of SK-N-SH coexpressed HCMV IE and CD44 (representative examples in Fig. 2c). Double staining for CD133 and HCMV IE


2355

Figure 1. HCMV proteins are detected in 100% of human neuroblastoma samples. (d, h) HCMV-IE staining in tissue section obtained from a GBM patient. Neuroblastoma samples were analyzed by immunohistochemistry for IE (a, e) and L (b, f) HCMV proteins and a-actin as isotype control (c, g). Original magnifications: 203 (a–d) and 403 (e–h). (i) IE proteins were found in 100% of analyzed neuroblastoma tissues (n 5 36/36), and late proteins were found in 92% (n 5 33). (j) Stained tissues were graded 0–4 depending on the percentage of positive cells. (k) Flow cytometric analyses of a primary neuroblastoma tumor reveal HCMV IEA protein expression in a majority of the cells (n 5 2, one representative example is shown using the 810R clone detecting both the HCMV IE55, IE72 and IE86 proteins). (I) Western blot analyses of seven frozen neuroblastoma tumor samples (lane 1–7) and control (C) using the using an HCMV IE specific antibody (810R clone) recognizing IE72, IE86 and IE55. Marker sizes are indicated on right. Bars a–d; 800 lm. e–h; 100 lm.

demonstrated a few IE-positive cells expressing CD133 in frozen neuroblastoma tissue samples (Fig. 2d). CD44 staining was not successful in tissue samples. To further investigate whether neuroblastoma cell lines could be super-infected with HCMV and establish a productive infection, we infected three neuroblastoma cell lines; Kelly; SK-N-AS and SH-SY5Y with VR1814, MOI 5, and collected samples for different analyses of CMV transcripts, proteins, and infectious virus at 1 and 7 days post-infection (dpi). All three cell lines were susceptible to infection; and demonstrated very low HCMV protein expression and transcript levels in cultures not superinfected by the virus (Supporting Information Figs. 1a and 1b). IE transcripts were detected at 1 dpi and approximately 20 to 25% of cells were CMV IE positive at 1 dpi. We collected both cells and supernatants of HCMV infected SK-N-AS and SH-SY5Y at 7 dpi, which were positive for HCMV IE and gB DNA (Supporting Information Table 2) and transferred either supernatants or C 2013 UICC Int. J. Cancer: 133, 2351–2361 (2013) V

cells/supernatants to the same corresponding uninfected cells. The following day, the cells were collected for IE transcript analyses or stained for IE protein expression. We observed a dose depended increase of IE transcripts in both cell lines (Supporting Information Table 2), and higher frequency of HCMV IE positive cells in neuroblastoma cultures infected with cells and supernatants (data not shown). Thus, HCMV can establish a productive infection in neuroblastoma cell lines at low levels, and the infection appeared to be mainly cell associated. HCMV DNA, RNA and IE and late proteins were also detected by PCR and immunohistochemistry in xenografts of SH-SY5Y, SK-N-BE(2), SK-N-AS, SK-N-DZ, Kelly, and IMR32 xenografts (Supporting Information Table 1, Kelly demonstrated in Fig. 3a). When Kelly cells were xenografted subcutaneously in NMRI nu/nu mice, HCMV IE and pp65 protein expression slightly increased as determined by flow cytometry in xenografts compared to the cells in culture (Fig. 3b).





2356

Figure 2. HCMV DNA and proteins are found in neuroblastoma cell lines. (a) FISH analysis demonstrates HCMV DNA in SH-SY5Y cells and SK-N-AS cells. The viral DNA was located in the periphery of the cell nucleus in neuroblastoma cells compared to a central location in HCMV AD169 infected MRC-5 fibroblasts. K562 cells provide a negative control demonstrating no detection of HCMV DNA. (b) Flow cytometric analyses of the neuroblastoma cells SK-N-DZ, SH-SY5Y and Kelly revealed different levels of expression of HCMV IE and pp65 proteins (n 5 2–7 sampling occasions for individual cell lines). (c) Neuroblastoma cells as exemplified by SK-N-DZ cells co-expressed HCMV IEA and the stem cell markers CD133 or CD44. The experiment was repeated three times. Left panels indicate isotype IgG control. Middle panels demonstrate CMV IE single staining and right panels double staining of HCMV IEA and CD133 (SK-N-DZ) or CD44 (IMR-32). (d) Double staining of a primary neuroblastoma tissue sample section with HCMV IE (green) and CD133 specific antibodies (red) revealed some double positive cells indicated with arrows, DAPI positive cells in blue.

Ganciclovir inhibits neuroblastoma growth in vitro and in vivo

Since HCMV DNA and proteins were detected in neuroblastoma cell lines, we tested the effects of ganciclovir on cell growth and the clonogenic capacity of neuroblastoma cell

lines SH-SY5Y, SK-N-BE(2), and SK-N-AS. Both neuroblastoma cell growth and colony formation was significantly inhibited (Fig. 4), whereas no growth inhibitory effects were observed when human MRC-5 fibroblasts were incubated with the same doses of gancilovir (Fig. 4). However, the


Figure 3. HCMV DNA and proteins are found in neuroblastoma xenografts. (a) Immunohistochemistry of neuroblastoma xenograft established of Kelly cells demonstrates HCMV IEA and protein expression. (b) Flow cytometric analyses of neuroblastoma Kelly cells and Kelly xenografts established in NMRI nu/nu mice demonstrate HCMV IEA and pp65 expression by flow cytometry. The experiment was repeated three times.

extent of inhibition varied between the cell lines. At the highest ganciclovir concentration (300 lM), the colony formation was inhibited by a mean of 49% in SH-SY5Y, 65% in SK-NBE(2), and 28% in SK-N-AS (Fig. 4). To determine whether oral valganciclovir affects tumor growth in vivo, we established human xenografts of SH-SY5Y in NMRI nu/nu mice. In all xenograft biopsy specimens examined, FISH, PCR and immunohistochemistry revealed the presence of HCMV DNA, RNA as well as IE and late proteins (Supporting Information Table 1). In mice bearing xenografts of SH-SY5Y, treatment with valganciclovir (20 or 30 mg/kg twice daily) for 12 days significantly reduced tumor growth (p < 0.05, Mann-Whitney test, Figs. 5a and 5b). The HCMV protein UL97 activates ganciclovir by phosphorylation, conferring DNA polymerase inhibitor activity that prevents DNA replication and expression of late HCMV proteins in infected cells. Consistent with this activity, valganciclovir lowered the expression of HCMV late proteins in the xenografts by a mean of 88% (p 5 0.0003, Figs. 5c and 5d); IE expression was not affected.

Discussion HCMV DNA and proteins have recently been detected in several different tumors of different origin. HCMV proteins C 2013 UICC Int. J. Cancer: 133, 2351–2361 (2013) V

2357

Figure 4. Ganciclovir inhibits neuroblastoma cell growth and colony formation of neuroblastoma cell lines. (a) Gancilovir induces a dose-dependent inhibition on the growth of SKN-AS, SK-N-BE(2) and SH-SY5Y neuroblastoma cells. No growth inhibition was observed in human MRC-5 fibroblast incubated with the same gancilovir doses. (b–d) In clonogenic assays, ganciclovir inhibited colony formation. The effect was strongest at the highest concentration used. Values are mean 6 SD. *p < 0.05, **p < 0.001, ***p < 0.0001 (one-way ANOVA). The experiment was repeated three times.

may have profound effects on tumor development, as suggested by numerous reports of viral antigens in tumor tissues and the large number of studies demonstrating an array of potential oncomodulatory mechanisms by specific viral proteins. Here, we found that IE, pp65, and late HCMV proteins are expressed in primary neuroblastomas, cell lines and in neuroblastoma xenografts. The HCMV protein expression level is evidently lower than during an acute viral infection, as optimized immunohistochemistry protocols including antigen retrieval are necessary to detect HCMV proteins in fixed, paraffin-embedded tumor specimens. However, in fresh tumor tissue samples, cell lines, and xenografts, HCMV protein expression was readily detected by flow cytometry using several different antibodies. The viral proteins IE72 and IE86 induce mutations in cellular genes and can induce oncogenic transformation in cooperation with adenovirus E1A through a “hit and run” mechanism.22 IE72 binds to p53 and can control cell cycle progression,23,24 and affects ataxia-telangiectasia mutated (ATM) kinase-mediated phosphorylation and nuclear accumulation of p53/p21.25 In glioblastoma cells, IE72 downregulates p53




2358

mRNA and protein expression.26 The HCMV protein IE86 can directly interact with p53 and inhibit p53-mediated apoptosis. UL44 inhibits the transcriptional activity of p53. In glioma cells, HCMV infection increases proliferation, which is associated with IE72-mediated inhibition of p53 and Rb proteins and activation of the PI3K/Akt signaling pathway.24 HCMV UL97 phosphorylates the Rb protein and thereby prevents apoptosis,27 and HCMV pp71 induces its degradation by the proteasome (as reviewed in Ref. 28). Other HCMV proteins inhibit differentiation, control epigenetic functions, and modulate the immune system, enabling the virus to avoid recognition and elimination (as reviewed in Ref. 28). Thus, HCMV regulates a plethora of factors that are important for the initiation of aberrant cell growth, cell transformation and immune surveillance. HCMV infection also results in chromosomal aberrations and chromosome breaks at 1q21 and 1q42,29–31 and virus-infected cells exhibit impaired DNA repair mechanisms.32 Interestingly, a poor prognostic subset of high-risk neuroblastoma displays a specific chromosome instability phenotype. Furthermore, chromosome 1 deletions are common in neuroblastoma and a marker of poor prognosis,33,34 and copy number variations (CNV) in chromosome 1q21 have recently been linked to these tumors.35 We therefore scrutinized the extensive Swedish material of neuroblastoma tumors examined in detail with SNP based aCGH as previously described36 including the sets of tumors in this study. However, we were unable to detect the 1q21 CNV as described by Diskin et al.35 (data not shown). Certain HCMV proteins may also be involved in oncogenesis. The HCMV-encoded chemokine receptor US28, which is constitutively expressed in virus-infected cells, induces tumor formation in a murine model,37 and transgenic VS28 mice with targeted expression of US28 to the intestinal epithelium develop intestinal hyperplasia and tumors.37 US28 expression in intestinal epithelial cells inhibited glycogen synthase-3b (GSK-3b) function, promoted accumulation of b-catenin and increased expression of Wnt target genes involved in the control of cell proliferation,37 which provide a direct molecular link between HCMV US28 and oncogenesis. US28 also stimulates angiogenesis by inducing expression of vascular endothelial growth factor (VEGF) via the cyclooxygenase-2 signaling pathway in vivo and in vitro.38,39 US28 induces phosphorylation of STAT3, resulting in interleukin-6 production, a phenomenon observed in glioblastomas and linked to patient survival.40 Furthermore, murine CMV infection of Trp53 heterzygote neonatal mice resulted in tumors at high frequency at 9 months of age. Most of the tumors were rhabdomyosarcoma, and were positive for MCMV DNA, RNA and protein.13 Trp53 heterzygote mice develop various spontaneous tumors, but generally after 9 months of age. As only 3% of infected mice developed tumors if they were infected at 4 weeks of age, it was speculated that MCMV accelerates tumor formation in Trp53 heterzygote mice and that preferentially rhabdomyosarcoma developed, although some mice also developed lymphomas. This observation support the hypothesis that CMV may


promote tumorogenesis in genetically predisposed individuals and may provide explanations to why only few HCMV carriers develop rare tumors. We have further showed that the HCMV protein IE72 induces high telomerase activity through a direct interaction with SP1 binding sites in the hTERT promoter;41 interestingly, in primary glioblastomas, only cells positive for HCMV IE expressed hTERT. Induced telomerase activity is a common theme for oncogenic viruses, as increased telomerase activity is a prerequisite for tumor cells to divide indefinitely.42 Our observations suggest that only HCMV protein– positive glioblastoma cells have an unrestricted ability to divide. In support of this statement, we discovered that the level of HCMV infection in malignant glioblastomas has a high prognostic value.14,15 Patients with a low-grade infection in the tumor lived significantly longer than those with a high-grade infection.14,15 This finding implies that the virus has a pathogenetic role and is not merely an epiphenomenon also in malignant glioblastomas. Owing to the limited number of patients and their heterogeneity, we were unable to fully evaluate the potential clinical impact of HCMV infection in neuroblastomas in this study. However, no clear difference was found between HCMV grade and different clinical and biological subsets of neuroblastomas included. Nevertheless, we found a high prevalence of HCMV in childhood neuroblastomas; all examined primary tumors exhibited HCMV RNA and protein expression. Viral cultures from tumor tissues were negative, and healthy tissues around the tumors appeared negative for HCMV proteins. However, we also detected HCMV proteins in differentiated ganglioneuromas and healthy adrenal tissue samples obtained from neuroblastoma patients, which imply that the virus is not only detected in malignant tumors, nor is it restricted to the tumor in neuroblastoma patients. The finding that HCMV proteins are detected in all subsets of neuroblastoma indicates that additional chromosomal aberrations and/or epigenetic alterations are important for the biological behavior and clinical outcome. On the other hand, our results in vitro and in vivo clearly show that HCMV may be a novel adjuvant therapy for neuroblastomas. Our observations further suggest that HCMV is present in long-term cultures; although the HCMV protein expression levels varied between cell lines sampled at different times. We suspect that the virus is present in a population of stem cells within the tumor. In support of this possibility, we found that HCMV proteins were expressed by neuroblastoma cells carrying CD133 or CD44, two potential markers defining stem cells within the tumor. In SH-SY5Y xenografts, HCMV protein expression was low as shown by immunohistochemistry (Fig. 4c). SH-SY5Y cells had very few cells double positive for CD44 and CMV IE proteins and the IE expression varied in culture; 34, 8, 0.1, 0, 0.1, 54, 61% at seven different sampling occasions during 1 year. We observed a few weakly HCMV IE positive cells that also expressed CD133 in frozen neuroblastoma samples (Fig. 2d). C 2013 UICC Int. J. Cancer: 133, 2351–2361 (2013) V

2359

Figure 5. Valganciclovir inhibits neuroblastoma growth in mice. (a) Female NMRI nu/nu mice carrying subcutaneous neuroblastoma xenografts were randomized to receive valganciclovir (20, n 5 7, or 30 mg/kg, n 5 7) twice daily or no treatment (n 5 10). Tumors were measured every day, and tumor volume was calculated as (width)2 3 length 3 0.44. The tumor volume index was calculated as the measured volume divided by the volume at the start of treatment. Tumor volume index at day 12 show significant growth reduction in animals treated with valganciclovir. Tumor volume index values are presented as mean. *p < 0.05 (Mann-Whitney test). (b) Tumor weight recorded at autopsy correlated with results on tumor volume index. Tumor weight is shown as median. *p < 0.05 (Mann-Whitney test). (c) The number of cells positive for HCMV IE and late proteins were rather few in the xenografts, yet consistent results for IE and late expression were obtained. While valganciclovir did not affect the expression of IE proteins, it significantly reduced the expression of late protein in the xenografts. Values are mean 6 SD (t test). The animal experiment was repeated two times. The bar represents 50 lm.

A link between clinical HCMV infection and neuroblastoma has been proposed earlier3,19 but was not demonstrated in tumor tissue specimens. HCMV is the most common pathogen to cause birth defects through congenital infections, which occur in 0.5 to 2% of all live births in the developed world.43,44 At 1 year of age, 25 to 40% of children are seropositive for CMV.43 Neuroblastomas consist of poorly or undifferentiated neural cells that are believed to originate from a developmental defect and are therefore considered to be embryonic tumors. We and others have reported that HCMV infection of neuronal progenitor cells prevents their differentiation.45–47 An early block in differentiation may be an important step in tumor transformation.48 As HCMV can infect progenitor cells in the developing fetus, it may be an early step toward tumor transformation in genetically predisposed individuals. C 2013 UICC Int. J. Cancer: 133, 2351–2361 (2013) V

Metastatic high-risk neuroblastomas, in particular the chromosome unstable subset, are often resistant to treatment even with potent multimodal therapy. Our results suggest that this virus may play a role in neuroblastoma pathogenicity and therefore represent a therapeutical target in this as well as other HCMV-positive tumors. We here demonstrate that anti-viral therapy prevents the growth of neuroblastomas in vitro and in vivo. We also recently found that HCMV is highly prevalent in medulloblastomas and that anti-viral treatment inhibits growth of HCMV-positive medulloblastoma xenografts in immunodeficient mice.7 In the previous study, we established xenografts from two HCMV-negative cell lines PC-3 established from a bone metastasis of a grade IV prostatic adenocarcinoma and BxPC-3 established from a pancreatic adenocarcinoma. In clonogenic assays, no significant effects in the clonogenic capacity were observed when PC-3 or BxPC-3 cells were treated with



2360


37.5-, 75- or 150-lM ganciclovir.7 Treatment of NMRI nu/nu mice carrying established PC-3 or BxPC-3 xenografts with valganciclovir (14 mg/kg twice daily po) had no significant effect on tumor growth.7 Although we were unable to identify an HCMV-negative neuroblastoma cell line, our previous observations suggest that the effects of valganciclovir in neuroblastoma are HCMV specific. Furthermore, we have performed the world’s first clinical study evaluating anti-viral therapy in 42 glioblastoma patients randomized to valganciclovir or placebo. Although the study failed its primary endpoint demonstrating no significant difference in tumor volumes at 3 and 6 months, patients who received at least 6 months of valganciclovir treatment demonstrate an unexpectedly high overall survival; 50% at 2 years at 27% at 4 years.49

HCMV infection of neuroblastoma cells in vitro confers resistance to chemotherapeutic agents such as cisplatinum and etoposide;50 this effect was blocked by ganciclovir, which targets HCMV replication. Thus, targeting the virus in tumor cells may decrease potential viral effects on cellular functions and also increase sensitivity to chemotherapy. Since ganciclovir significantly reduced tumor growth both in vitro and in vivo, this study indicate that antiviral treatment for HCMV may be a novel, generally well-tolerated option for treating children with neuroblastoma.

Acknowledgements We thank Stephen Ordway for editorial help.

References 1.

2.


3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

Brodeur GM. Neuroblastoma: biological insights into a clinical enigma. Nat Rev Cancer 2003;3: 203–16. Mosse YP, Laudenslager M, Longo L, et al. Identification of ALK as a major familial neuroblastoma predisposition gene. Nature 2008; 455:930–5. Nigro G, Schiavetti A, Booth JC, et al. Cytomegalovirus-associated stage 4S neuroblastoma relapsed stage 4. Med Pediatr Oncol 1995;24: 200–3. Wertheim P, Voute PA. Neuroblastoma, Wilms’ tumor, and cytomegalovirus. J Natl Cancer Inst 1976;57:701–3. Scholz M, Blaheta RA, Wittig B, et al. Cytomegalovirus-infected neuroblastoma cells exhibit augmented invasiveness mediated by beta1alpha5 integrin (VLA-5). Tissue Antigens 2000;55:412–21. Cinatl J, Jr, Vogel JU, Cinatl J, et al. Long-term productive human cytomegalovirus infection of a human neuroblastoma cell line. Int J Cancer 1996;65:90–6. Baryawno N, Rahbar A, Wolmer-Solberg N, et al. Detection of human cytomegalovirus in medulloblastomas reveals a potential therapeutic target. J Clin Invest 2011;121: 4043–55. Cobbs CS, Harkins L, Samanta M, et al. Human cytomegalovirus infection and expression in human malignant glioma. Cancer Res 2002;62: 3347–50. Harkins L, Volk AL, Samanta M, et al. Specific localisation of human cytomegalovirus nucleic acids and proteins in human colorectal cancer. Lancet 2002;360:1557–63. Harkins LE, Matlaf LA, Soroceanu L, et al. Detection of human cytomegalovirus in normal and neoplastic breast epithelium. Herpesviridae 2010;1:8. Mitchell DA, Xie W, Schmittling R, et al. Sensitive detection of human cytomegalovirus in tumors and peripheral blood of patients diagnosed with glioblastoma. Neuro-oncology 2008;10:10–18. Samanta M, Harkins L, Klemm K, et al. High prevalence of human cytomegalovirus in prostatic intraepithelial neoplasia and prostatic carcinoma. J Urol 2003;170:998–1002. Price RL, Bingmer K, Harkins L, et al. Cytomegalovirus infection leads to pleomorphic

14.

15.

16. 17.

18.

19.

20.

21.

22.

23.

24.

25.

rhabdomyosarcomas in Trp531/- mice. Cancer Res 2012;72:5669–74. Rahbar A, Stragliotto G, Orrego A, et al. Low levels of human cytomegalovirus infection in glioblastoma multiforme associates with patient survival—a case-control study. Herpesviridae 2012;3:3. Rahbar A, Orrego A, Peredo I, et al. Human cytomegalovirus infection levels in glioblastoma multiforme are of prognostic value for survival. J Clin Virol 2013;57:36–42. Soroceanu L, Cobbs CS. Is HCMV a tumor promoter? Virus Res 2011;157:193–203. Soderberg-Naucler C. Does cytomegalovirus play a causative role in the development of various inflammatory diseases and cancer? J Intern Med 2006;259:219–46. Baryawno N, Sveinbjornsson B, Eksborg S, et al. Tumor-growth-promoting cyclooxygenase-2 prostaglandin E2 pathway provides medulloblastoma therapeutic targets. Neuro Oncol 2008;10: 661–74. Nigro G, Adler SP. Cytomegalovirus infections during pregnancy. Curr Opin Obstet Gynecol 2011;23:123–8. Brodeur GM, Pritchard J, Berthold F et al. Revisions of the international criteria for neuroblastoma diagnosis, staging and response to treatment. J Clin Oncol 1993;11:1466–77. Cohn SL, Pearson AD, London WB, et al. The International Neuroblastoma Risk Group (INRG) classification system: an INRG Task Force report. J Clin Oncol 2009;27:289–97. Shen Y, Zhu H, Shenk T. Human cytomagalovirus IE1 and IE2 proteins are mutagenic and mediate “hit-and-run” oncogenic transformation in cooperation with the adenovirus E1A proteins. Proc Natl Acad Sci USA 1997; 94:3341–5. Speir E, Modali R, Huang ES, et al. Potential role of human cytomegalovirus and p53 interaction in coronary restenosis. Science 1994;265:391–4. Cobbs CS, Soroceanu L, Denham S, et al. Modulation of oncogenic phenotype in human glioma cells by cytomegalovirus IE1-mediated mitogenicity. Cancer Res 2008;68:724–30. Castillo JP, Frame FM, Rogoff HA, et al. Human cytomegalovirus IE1-72 activates ataxia telangiectasia mutated kinase and a p53/p21-mediated growth arrest response. J Virol 2005;79: 11467–75.

26. Lee K, Jeon K, Kim JM, et al. Downregulation of GFAP, TSP-1, and p53 in human glioblastoma cell line, U373MG, by IE1 protein from human cytomegalovirus. Glia 2005;51:1–12. 27. Hume AJ, Finkel JS, Kamil JP, et al. Phosphorylation of retinoblastoma protein by viral protein with cyclin-dependent kinase function. Science 2008;320:797–9. 28. Soderberg-Naucler C. HCMV microinfections in inflammatory diseases and cancer. J Clin Virol 2008;41:218–23. 29. Fortunato EA, Dell’Aquila ML, Spector DH. Specific chromosome 1 breaks induced by human cytomegalovirus. Proc Natl Acad Sci USA 2000; 97:853–8. 30. Siew VK, Duh CY, Wang SK. Human cytomegalovirus UL76 induces chromosome aberrations. J Biomed Sci 2009;16:107. 31. Albrecht T, Deng CZ, Abdel-Rahman SZ, et al. Differential mutagen sensitivity of peripheral blood lymphocytes from smokers and nonsmokers: effect of human cytomegalovirus infection. Environ Mol Mutagen 2004;43:169–78. 32. Luo MH, Rosenke K, Czornak K, et al. Human cytomegalovirus disrupts both ataxia telangiectasia mutated protein (ATM)- and ATM-Rad3-related kinase-mediated DNA damage responses during lytic infection. J Virol 2007;81: 1934–50. 33. Martinsson T, Sjoberg RM, Hedborg F, et al. Deletion of chromosome 1p loci and microsatellite instability in neuroblastomas analyzed with short-tandem repeat polymorphisms. Cancer Res 1995;55:5681–6. 34. Caron H, van Sluis P, de Kraker J, et al. Allelic loss of chromosome 1p as a predictor of unfavorable outcome in patients with neuroblastoma. New Engl J Med 1996;334: 225–30. 35. Diskin SJ, Hou C, Glessner JT, et al. Copy number variation at 1q21.1 associated with neuroblastoma. Nature 2009;459:987–91. 36. Caren H, Kryh H, Nethander M, et al. High-risk neuroblastoma tumors with 11q-deletion display a poor prognostic, chromosome instability phenotype with later onset. Proc Natl Acad Sci USA 2010;107:4323–8. 37. Bongers G, Maussang D, Muniz LR, et al. The cytomegalovirus-encoded chemokine receptor US28 promotes intestinal neoplasia in transgenic mice. J Clin Invest 2010;120:3969–78.


2361


42. Opitz OG. Telomeres, telomerase and malignant transformation. Curr Mol Med 2005;5:219–26. 43. Ahlfors K. IgG antibodies to cytomegalovirus in a normal urban Swedish population. Scand J Infect Dis 1984;16:335–7. 44. Forbes BA. Acquisition of cytomegalovirus infection: an update. Clin Microbiol reviews 1989; 2:204–16. 45. Odeberg J, Wolmer N, Falci S, et al. Human cytomegalovirus inhibits neuronal differentiation and induces apoptosis in human neural precursor cells. J Virol 2006;80:8929–39. 46. Odeberg J, Wolmer N, Falci S, et al. Late human cytomegalovirus (HCMV) proteins inhibit differentiation of human neural precursor cells into astrocytes. J Neurosci Res 2007;85:583–93.

47. Han GP, Li L, Kosugi I, et al. Enhancement of susceptibility of adult mouse brain to cytomegalovirus infection by infusion of epidermal growth factor. J Neurosci Res 2007;85:2981–90. 48. Sanai N, Alvarez-Buylla A, Berger MS. Neural stem cells and the origin of gliomas. N Engl J Med 2005;353:811–22. 49. Stragliotto G, Rahbar A, Solberg NW, et al. Effects of valganciclovir as an add-on therapy in patients with cytomegalovirus-positive glioblastoma: a randomized, double-blind, hypothesis-generating study. Int J Cancer 2013, doi: 10.1002/ijc.28111. 50. Cinatl J, Jr, Cinatl J, Vogel JU, et al. Persistent human cytomegalovirus infection induces drug resistance and alteration of programmed cell death in human neuroblastoma cells. Cancer Res 1998;58:367–72.


38. Maussang D, Langemeijer E, Fitzsimons CP, et al. The human cytomegalovirus-encoded chemokine receptor US28 promotes angiogenesis and tumor formation via cyclooxygenase-2. Cancer Res 2009; 69:2861–9. 39. Maussang D, Verzijl D, van Walsum M, et al. Human cytomegalovirus-encoded chemokine receptor US28 promotes tumorigenesis. Proc Natl Acad Sci USA 2006;103:13068–73. 40. Slinger E, Maussang D, Schreiber A, et al. HCMV-encoded chemokine receptor US28 mediates proliferative signaling through the IL-6STAT3 axis. Sci Signal 2010;3:ra58. 41. Straat K, Liu C, Rahbar A, et al. Activation of telomerase by human cytomegalovirus. J Natl Cancer Inst 2009;101:488–97.
