- Email: [email protected]
Human cytomegalovirus infection levels in glioblastoma multiforme are of prognostic value for survival
Journal of Clinical Virology 57 (2013) 36–42
Contents lists available at SciVerse ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
Human cytomegalovirus infection levels in glioblastoma multiforme are of prognostic value for survival Afsar Rahbar a , Abiel Orrego b , Inti Peredo c , Mensur Dzabic a , Nina Wolmer-Solberg a , Klas Strååt d , Giuseppe Stragliotto c , Cecilia Söderberg-Nauclér a,∗ a
Department of Medicine Solna, Experimental Cardiovascular Research Unit, Center for Molecular Medicine, Karolinska Institute, Stockholm, Sweden Cancer Center, Karolinska University Hospital, Stockholm, Sweden c Department of Neurosurgery, Karolinska University Hospital, Stockholm, Sweden d Department of Cell and Molecular Biology (CMB), Stockholm, Sweden b
a r t i c l e
i n f o
Article history: Received 27 September 2012 Received in revised form 20 December 2012 Accepted 21 December 2012 Keywords: HCMV GBM Survival TTP
a b s t r a c t Background: Patients with glioblastoma multiforme (GBM) generally live 12–15 months after diagnosis, despite maximal surgical resection, adjuvant radiotherapy, and chemotherapy. HCMV has been detected in 90–100% of GBMs. We recently found that low grade HCMV infection in GBM tumours was highly associated with survival over 18 months (case–control study). Here, we sought to determine whether low-grade HCMV infection in GBMs is associated with prolonged survival in a consecutive patient cohort, analysed retrospectively. Study design: Tumour samples from 75 consecutive GBM patients treated surgically at Karolinska University Hospital in 2004–2005 were examined by immunohistochemistry (IHC) and in situ hybridization for HCMV proteins and DNA, respectively. Tumours were graded 1–4, depending on the percentage of positive cells by IHC. Low-grade HCMV was defined as grade 1 (<25% of HCMV infected tumour cells). Time to tumour progression (TTP) and survival data were analysed with Cox regression and Kaplan–Meier models. Results: HCMV infection was detected in 74 of 75 tumours (99%). In patients with low-grade HCMV infection, median survival was 20 months longer than in patients with high-grade infections (P = 0.036, HR: 2.2), and TTP was 8 months longer (P = 0.1, HR: 1.8). Two-year survival was much higher in patients with low-grade HCMV infection (63.6% vs. 17.2%, P = 0.003). Conclusion: The longer survival in patients whose tumours had low-grade HCMV infection suggests that the level of HCMV infection in GBMs has a prognostic value and that HCMV may contribute to the pathogenesis of GBM. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Glioblastoma multiforme (GBM) is the most malignant primary tumour of the central nervous system.1 The median survival after diagnosis is 12–15 months.2 Although 90–100% of GBMs are infected with human cytomegalovirus (HCMV),3–5 it is unclear whether this virus contributes to tumour development or progression. HCMV is a herpesvirus carried by 70–100% of the world’s population. In healthy persons, HCMV infection is generally asymptomatic; however, the virus remains latent in the bone marrow
∗ Corresponding author at: Department of Medicine, Centre for Molecular Medicine, Karolinska Institute, S-171 76 Stockholm, Sweden. Tel.: +46 8 51779896; fax: +46 8 313147. E-mail address: [email protected] (C. Söderberg-Nauclér). 1386-6532/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jcv.2012.12.018
and peripheral blood and can be reactivated by inflammation.6–8 Emerging evidence demonstrate that HCMV can be detected in certain malignant tumours, such as GBM,3–5 colon cancer,9 breast cancer10 and prostatic carcinoma,11 mucoepidermoid carcinoma of salivary glands12 and rhabdomyosarcomas.13 Using immunohistochemistry, flow cytometry, PCR and in situ hybridization techniques, we and others have detected HCMV in tumour tissue specimens.14 However, some investigators have not confirmed the finding of HCMV in tumours.15–17 In our own experience, we can readily detect the virus in frozen tumour specimens by indirect immunofluorescence or in paraffin embedded tumour tissue specimens using the high sensitive immunostaining protocols, which include optimal tissue fixation, antigen retrieval, use of appropriate antibodies for paraffin embedded tissues and blocking of nonspecific binding.3–5,9–12,14 However, we do not detect the virus in tumours using regular immunohistochemistry protocols. Therefore, different detection methods used by different investigators, in
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42 Table 1 Characteristics of 75 consecutive patients with GBM. Characteristics Age <50 ≥50 Sex Male Female RPA class III–IV V–VI Extent of surgery Partial Radical Gamma knife No Yes RT + temozolomide No Yes RT + adjuvant chemotherapy No Yes
n (%) 17 (23) 58 (77) 42 (56%) 33 (44) 34 (45) 41 (55) 20 (27) 55 (73) 69 (92) 6 (8) 60 (80) 15 (20) 26 (35) 49 (65)
Three patients were alive at study closure after a mean follow-up of 82.3 months. RT: radio therapy; RPA: recursive partition analysis.
particular the staining procedures, may be the main reasons why some studies have failed to detect HCMV in tumour tissues.15–17 Although one HCMV protein, US28, has an oncogenic potential,18 this virus is not considered to be oncogenic. Rather, it may be an oncomodulator that contributes to cancer development by modifying tumour cell biology.19 In such scenario, the level of HCMV infection in the tumour may affect disease progression and the prognosis of patients with HCMV positive tumours. In support of this hypothesis, we recently observed that a low-grade HCMV infection was strongly associated with long-term survival in GBM patients (P = 0.0006)5 in a case–control study. Here, we aimed to determine if low-grade HCMV infection in a consecutive cohort of GBM patients was associated with prolonged overall survival. We also defined the prevalence of Epstein–Barr virus (EBV) and herpesvirus 6 (HHV-6) as controls for viral persistence.
late protein of 47–55 kDa) (both antibodies IgG2a, Chemicon International), EBV (BZLF1 protein, IgG1; DakoCytomation, Glostrop, Denmark) andHHV-6 (IgG1, Abcam, Cambridge, UK). Antibodies against smooth muscle cell alpha actin (IgG2a, Biogenex, San Ramon, CA) and von Willebrand factor (IgG1, DakoCytomation) served as isotype controls. The specimens were examined by AR and AO; neither had access to the clinical records of the patients. These samples were scored for HCMV proteins in the specimens as previously described5 (Table 2). Tumour cell proliferation (Ki-67), p53 mutation, mitosis, expression of glial fibrillary acidic protein, epidermal growth factor receptor (EGF-R), and VEGF-R were assessed with automated immunohistochemical staining protocols at our hospital, and the findings were analysed for associations with high- and low-grade HCMV infection (Table 2). 2.3. Polymerase chain reaction (PCR) and DNA sequencing Primary cultures were established from the tumours of five newly diagnosed GBM patients. DNA samples were prepared with QIAamp DNA mini-kits (Qiagen, Valencia, CA), and analysed by PCR for the HCMV major immediate-early (MIE) gene.3,21 DNA from PCR products were sequenced with an ABI 3730 DNA analyser. Sequence data were analysed by BLAST searches. As controls, DNA samples were analysed for herpes simplex virus (HSV)-1 (gpD, 188 bp), HSV2 (gpG, 149 bp), and HHV-6 (U67) by real-time PCR at our hospital. 2.4. Statistical analysis Overall survival and time to tumour progression (TTP) were analysed by a medical statistician (Fredrik Hansson, Norma, Lund, Sweden) using Cox regression models; the covariates were lowgrade infection (no infection or grade 1) or high-grade infection (grades 2–4), age, extent of surgery, and RPA class. All patients who were alive or did not have progression at the time of investigation were censored in the analysis (April 16, 2012). The results are presented as hazard ratios with 95% confidence intervals (CIs). Survival graphs were created with the Kaplan–Meier life-table method. The median times to end points were estimated from the Kaplan–Meier curves, with 95% CIs.
2. Methods
3. Results
2.1. Patient samples
3.1. HCMV infection level has a high prognostic value for GBM patients
Paraffin-embedded tumour specimens were obtained from 75 consecutive GBM patients who underwent their first surgery after diagnosis at Karolinska University Hospital in 2004–2005 (two patients were excluded due to lack of biopsy material, 51 patients were analysed in our previously published case–control study5 ). All patients had World Health Organization (WHO) grade IV GBM and received standard treatment (Table 1). The diagnosis was independently confirmed by two experienced neuropathologists (mainly AO). Patients were classified according to the adapted European Organization for Research and Treatment of Cancer recursive partition analysis classification (Table 1).20 2.2. Immunohistochemistry and in situ hybridization All samples were analysed by immunohistochemistry for HCMV-IEA and late antigen (LA) as described.3 Seventeen randomly selected samples were also stained for HHV-6 and EBV as viral specificity controls and 19 randomly selected tissue sections were analysed by in situ hybridization as described.3 Primary antibodies used were against HCMV-IEA (reacts with an immediate early non-structural antigen of 68–72 kDa), HCMV-LA (reacts with a
37
HCMV IE protein was detected in 74 of 75 biopsies (99%) and HCMV-LA in 70 of 75 (93%). Seventeen randomly selected samples were subjected for analyses of HHV-6 and EBV as specificity samples (Fig. 1A). Blood vessels within tumours were also positive for HCMV IE and HCMV-LA in 45 of 75 patients (60%) (Fig. 1A). One patient was negative for HCMV-IEA (Fig. 1A), HCMV-LA, EBV, and HHV-6 (data not shown). Noncancerous cells near tumour cells were consistently negative for HCMV. HCMV infection was confirmed by in situ hybridization in all randomly selected HCMVpositive samples but not in the single HCMV-negative sample (Fig. 1B). To further confirm the presence of HCMV in GBMs, we also established primary cultures from surgical tumour specimens of five patients with newly diagnosed GBM. DNA was extracted at passage 1 and amplified by PCR assays for the gene encoding HCMV MIE. PCR assays for the genes encoding HHV-6 (U67), HSV-1 (gpD), and HSV-2 (gpG) served as controls. All samples were positive for HCMV-MIE and negative for the other viruses (data not shown). DNA sequence analysis of HCMV MIE PCR products revealed HCMV MIE DNA sequences in all five GBM cultures (data not shown),
38
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42
Table 2 Hazard ratio according to HCMV infection level in the tumour, age, extent of resection, and RPA subclass using Cox regression model for survival data in 75 consecutive patients. Characteristic
HCMV infection (n: 11 low-grade vs. 64 high-grade) Age, years (n: 17 < 50 vs. 58 ≥ 50, n = 58) Extent of surgery (n: 55 radical vs. 20 partial) RPA class (n: 12 III and 22 IV vs. 34 V and 7 VI)
HCMV-IEA
HCMV-LA
P
HR
0.036 0.23 0.002 0.0019
2.2 1.44 2.6 2.4
95% CI 1.0–4.38 0.79–2.6 1.44–4.82 1.38–4.2
P
HR
0.6 0.23 0.009 0.0003
1.14 1.46 2.2 2.8
95% CI 0.70–1.86 0.79–2.67 1.23–3.9 1.59–4.78
HCMV: human cytomegalovirus; IEA: immediate-early antigen; LA: late antigen; n: number of patients; RPA: recursive partition analysis; HR: hazard ratio; CI: confidence interval.
but the DNA sequences were distinctly different from wild type (VR1814, TB40) or laboratory virus strains (AD169 and Towne), which excluded contamination. HCMV Infection was graded according to the estimated percentage of cells expressing HCMV IEA or HCMV Late proteins in the specimens as previously described5 ; negative or grade 1, <25%; grade 2, ≥25–50%; grade 3, ≥50–75%; and grade 4, >75% HCMV positive cells in the tumour. Patients with an HCMV negative or an HCMV grade 1 tumour were defined to have a low grade HCMV infection in their tumour, while patients with a high grade infection had a tumour graded as 2, 3 or 4 for HCMV infection. One patient was HCMV IEA negative and survived 34 months. 14.7% of patients had grade 1 HCMV-IEA infection and their median OS was 33 months (n = 11), 8% of the patients had grade 2 (median OS was 13 months, n = 6), 22.7% of patients had grade 3 (median OS was 12 months, n = 17) and 54.7% of patients had grade 4 (median OS was 13 months, n = 41). Thus, as we previously observed there was no difference in OS (Median OS was 12.6 months) between HCMV IE grade 2–4, which was categorized as high-grade HCMV infection. Negative or grade 1 was regarded as low-grade HCMV infection. Based on this categorization we next analysed whether there was an association between low-grade HCMV infection and survival. Eleven of the 75 patients (14.7%) had low-grade HCMVIEA expression, and 64 (85.3%) had high-grade infection (grades 2–4). Median overall survival was significantly higher in patients with low-grade HCMV-IEA than in those with high-grade HCMVIEA (33 vs. 13 months, P = 0.036; HR: 2.2, 95% CI: 1.0–4.38) (3 patients were censored; one with low-grade and two with highgrade HCMV-IEA infections) (Fig. 2A). The 2-years survival rate was significantly higher in patients with low-grade than in those with high-grade HCMV-IEA infection (63.6% vs. 17.2%, P = 0.003, Fisher Exact Test). HCMV-LA infection had no effect on median overall survival (14.5 vs. 13 months, P = 0.6; HR: 1.4, 95% CI: 0.70–1.86). One patient with low-grade and two with high-grade HCMV-LA infections were censored (Fig. 2B). The median TTP did not differ significantly in patients with lowgrade HCMV-IEA or LA expression than in those with high-grade IEA or LA expression (for HCMV-IEA: 14 vs. 6 months, P = 0.1; HR: 1.8, 95% CI: 0.9–3.55; (Fig. 2C) and for HCMV-LA: 8.25 vs. 5 months, P = 0.13; HR: 1.49, 95% CI: 0.89–2.47) (Fig. 2D). Multivariable analysis of overall survival was performed to evaluate HCMV infection expression in relation to other prognostic factors known to be important for survival (e.g., age, extent of surgery, and RPA class) (Table 2). The multivariable analysis in relation to HCMV infection showed that overall survival was not significantly related to age (HCMV-IEA HR: 1.44, P = 0.23; HCMV-LA HR: 1.46, P = 0.23). Patients with radical surgery had significantly improved survival (HCMV-IEA HR: 2.63, P = 0.002; HCMV-LA HR: 2.2, P = 0.009). Risk of death in patients with RPAs of V or VI was 2.4-fold (HCMV-IEA HR: 2.4, P = 0.002) and 2.8-fold (HCMVLA HR: 2.8, P = 0.0003) higher than in those with RPAs of III or IV (Table 2). Furthermore, multivariable analysis in relation to HCMV infection showed that TTP was not related to age (HCMV-IEA
HR: 0.8, P = 0.50; HCMV-LA HR: 0.72, P = 0.33). However, patients with radical surgery (HCMV-IEA HR: 2.23, P = 0.007; HCMV-LA HR: 2.21, P = 0.008) or with RPAs III-IV (HCMV-IEA HR: 2.46, P = 0.0023; HCMV-LA HR: 2.7, P = 0.0007) had significantly longer TTP (Table 2). Overall survival and TTP did not differ in patients with low-grade HCMV-IEA immunoreactivity who underwent radiotherapy (RT) or RT in combination with adjuvant chemotherapy. The cause of death was related to the tumour in all patients. 3.2. Phenotypic characterization of GBMs Further evaluation of tumour markers examined for clinical purpose revealed p53 mutations in 27 of 55 (49%) samples with high-grade HCMV-IEA infection and 2 of 9 (22%) with low-grade infection. p53 mutations were detected in 26 of 47 (55%) tumours with high-grade HCMV-LA infection and in 3 of 17 (18%) with low-grade infection. While only a trend between p53 mutations for HCMV-IEA was observed, there was a significant association between HCMV late protein expression and p53 mutations (P = 0.0075; Chi-square test, Table 3). The extent of mitosis did not differ significantly between the groups (Table 3). Expression of EGFR and VEGF-R did not differ in the low- and high-grade groups (Table 3). 4. Discussion Despite tremendous progress in understanding the molecular aspects of GBM, the aetiology of these tumours is unclear. In this study, we confirmed the previously reported high frequency of HCMV infections in GBMs. Importantly; the level of HCMV infection at the time of diagnosis had a prognostic value for patient survival. Patients whose tumours had low-grade infections lived 2.5 times longer than those with high-grade infections (33 vs. 13 months for HCMV-IEA, P = 0.036; HR: 2.2, 95% CI: 1.0–4.38) and had a higher 2-year survival (63.6% vs. 17.2%, P = 0.003). However, median TTP was not significantly longer in patients with low-grade infection (14 months vs. 6 months for HCMV-IEA, P = 0.1; HR: 1.8, 95% CI: 0.9–3.55). Several factors have shown prognostic potential in patients with GBMs. Historically, young age (<50 years) and good neurologic function (Karnovsky score >90 and WHO performance score 0) are considered as positive prognostic factors. Here, the risk of death was 2.2-fold higher (P = 0.036; HR: 2.2, 95% CI: 1.0–4.38) in patients with high-grade HCMV-IE protein expression than in those with low-grade infection and 2.4-fold higher in patients with RPA class ≥V than in those with RPA class
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42
39
Fig. 1. Detection of HCMV and HHV-6 expression by immunohistochemistry (A) and HCMV DNA by in situ hybridization (B) in GBMs, bar, 50 m.
The association between low-grade HCMV infection and improved survival imply that HCMV may contribute to GBM progression. An active HCMV infection in tumour tissues might result from a new primary infection or reactivation of a latent
infection.23,24 Consistent with the latter possibility, GBMs are highly associated with inflammation and exert a variety of local and systemic immunosuppressive effects. We earlier demonstrated that latent HCMV can be reactivated by inflammatory stimuli.8
40
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42
Fig. 2. HCMV infection levels in GBMs are of high prognostic value for patient outcome (A and B) and for TTP (C and D).
HCMV may be transmitted to tumour cells and through its unique ability to influence cellular and immunological functions, affect the malignancy of the tumour by and its sensitivity to therapy. The HCMV protein US28 activates NF-kB and phosphorylates STAT3, causing COX-2 expression, VEGF production, and IL-6, resulting in inflammation, angiogenesis and tumour formation in vivo.25–27 Transgenic mice that express US28 in intestinal epithelial cells developed adenomas and adenocarcinomas, which was associated with inhibited glycogen synthase 3 function, promoted accumulation of -catenin, and increased expression of Wnt target genes involved in the control of the cell proliferation.28 HCMV-US28 also
induces cellular migration29 and angiogenesis26 and may thereby enhance tumour invasiveness – another hallmark of GBMs. HCMV IE72 and IE86 induce expression of Rb, a key cell-cycle regulator that allows the transition from G1 to S phase and interacts with p53; under certain circumstances the IE proteins may result in uncontrolled DNA replication, transformation, and p53 mutations.30,31 We found p53 mutations in 26 of 47 samples with high-grade HCMV-LA infection and in 8 of 17 with low-grade infection. Expression of HCMV-LA was significantly associated with p53 mutations (P = 0.0075, Chi-square test), suggesting a potential viral effect on p53.30 Recently, we also showed that HCMV-IE72
Table 3 Expression of HCMV and available clinical data on the expression of glial fibrillary acidic protein, epidermal growth factor receptor, vascular endothelial growth factor receptor, the presence of p53 mutations, and tumour cell proliferation in GBMs. HCMV-IEA
Consecutive patients (n = 75) p53 mutation (n = 64) GFAP (n = 75) Ki-67 (n = 71) Ki-67 MIB index interval (%) Mitosis (n = 50) EGF-R (n = 60) VEGF-R (n = 44)
HCMV-LA
Grades 0–1 n (%)
Grades 2–4 n (%)
Grades 0–1 n (%)
Grades 2–4 n (%)
11 (15) 2/9 (22) 11/11 (100) 10/10 (100) 5–40 10/11 (91) 7/7 (100) 7/8 (88)
64 (85) 27/55 (49) 64/64 (100) 61/61 (100) 10–90 38/39 (97) 44/53 (83) 35/36 (97)
32 (43) 3/17 (18)a 22/22 (100) 21/21 (100) – 10/11 (91) 15/15 (100) 7/8 (88)
43 (57) 26/47 (55)a 53/53 (100) 50/50 (100) – 38/39 (97) 36/45 (80) 35/36 (97)
IEA: immediate-early antigen; LA: late antigen; GFAP: glial fibrillary acidic protein; EGF-R: epidermal growth factor receptor; VEGF-R: vascular growth factor receptor. a P = 0.0075 (Chi-square test).
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42
may be involved in oncogenesis. HCMV IE72 stimulated telomerase activity, which is important for cellular immortalization and transformation.32 hTERT expression co-localized with HCMV proteins in GBMs, and hTERT was highly expressed in tumour areas with high-grade HCMV infection.32 Furthermore, the protooncogenes c-myc, c-fos, and c-jun, cyclin E, cyclin A, and Cdk2 are all rapidly activated in HCMV infected cells.27 HCMV infection in S-phase cells results in specific chromosome 1 breaks in 1q21 and 1q42, and deletions in 1q42 have been associated with GBM.33–36 Hence, reactivation of HCMV may occur in GBM patients and promote the oncogenic potential of HCMV, possibly through a mechanism involving Wnt, VEGF, STAT3 phosphorylation/IL-6 production, and induced telomerase activity, oncogene expression and DNA damage. Thus, HCMV infection may result in a more aggressive tumour phenotype with a poor prognosis. Importantly, the presence of HCMV in glioblastoma may provide a new target for therapy. We recently detected HCMV/COX-2 expression in primary medulloblastoma tumours and in medulloblastoma cell lines; the growth of medulloblastoma xenografts in vivo (nude mice) was inhibited by 72% using a COX-2 inhibitor in combination with an anti-HCMV drug.14 No effect of anti-viral therapy was observed in HCMV negative tumours. As HCMV infected tumour cells have been shown to exhibit increased resistance to chemotherapy in vitro,37 anti-viral therapy may enhance the clinical response to chemotherapy. In summary, our findings confirm our recent observation of a strong association between low-grade HCMV infection and longterm survival in GBM patients.5 We here provide further evidence that an active HCMV infection in GBM patients has a prognostic value for patient outcome in a consecutive cohort of GBM patients operated at our hospital during 2004–2005. In this cohort, we could estimate TTP and overall survival by Kaplan–Meier analyses, which was not possible in our previous case control cohort study. We found that patients with low grade infection survived 20 months longer than patients with high grade HCMV infection in their tumours at diagnosis (P = 0.036). Since HCMV exhibits oncomodulatory effects and may prevent the action of chemotherapy, treatment to control HCMV infection in tumour tissues may suppress tumour progression, and increase the tumour’s sensitivity to conventional therapy. Funding This work was supported by the Torsten and Ragnar Söderbergs’ Research Foundation, the Swedish Research Council (10350 and K2007-56X-12615-10-3), the Swedish Children Cancer Research Foundation (Project 05/100, 07/065), the Swedish Cancer Foundation (5044-B05-01XAB), Swedish Society for Medical Research (SLS), The Royal Swedish Academy of Sciences (KVA), Goljes Memory Foundation, Magnus Bergvalls Foundation, Swedish Society for Medical Research (SSMF), and Tore Nilsons Foundation. Competing interests The authors declare no conflicts of interest. However, CS-N has served as a lecturer and once in Scientific Advisory Board for Roche and holds an investigational grant from the same company for a clinical study to evaluate the effect of antiviral treatment in GBM patients. Ethical approval The study was approved by the Ethics Committee at the Karolinska Institute (2008/628-31/2).
41
Acknowledgment We thank Stephen Ordway for editorial assistance. References 1. Graham DILP, editor. Greenfield’s neuropathology. 6th ed. London: Arnold; 1996. 2. Affronti ML, Heery CR, Herndon II JE, Rich JN, Reardon DA, Desjardins A, et al. Overall survival of newly diagnosed glioblastoma patients receiving carmustine wafers followed by radiation and concurrent temozolomide plus rotational multiagent chemotherapy. Cancer 2009;115(15):3501–11. 3. Cobbs CS, Harkins L, Samanta M, Gillespie GY, Bharara S, King PH, et al. Human cytomegalovirus infection and expression in human malignant glioma. Cancer Res 2002;62(12):3347–50. 4. Mitchell DA, Xie W, Schmittling R, Learn C, Friedman A, McLendon RE, et al. Sensitive detection of human cytomegalovirus in tumors and peripheral blood of patients diagnosed with glioblastoma. Neuro-oncol 2008;10(1):10–8. 5. Rahbar A, Stragliotto G, Orrego A, Peredo I, Taher C, Willems J, et al. Low levels of human cytomegalovirus infection in glioblastoma multiforme associates with patient survival—a case–control study. Herpesviridae 2012;3:3. 6. ES M, editor. Fields virology. 4th ed. Philadelphia: Lippincott Williams and Wilkins; 2001. 7. Taylor-Wiedeman J, Sissons JG, Borysiewicz LK, Sinclair JH. Monocytes are a major site of persistence of human cytomegalovirus in peripheral blood mononuclear cells. J Gen Virol 1991;72(Pt 9):2059–64. 8. Soderberg-Naucler C, Fish KN, Nelson JA. Reactivation of latent human cytomegalovirus by allogeneic stimulation of blood cells from healthy donors. Cell 1997;91(1):119–26. 9. Harkins L, Volk AL, Samanta M, Mikolaenko I, Britt WJ, Bland KI, et al. Specific localisation of human cytomegalovirus nucleic acids and proteins in human colorectal cancer. Lancet 2002;360(9345):1557–63. 10. Harkins LE, Matlaf LA, Soroceanu L, Klemm K, Britt WJ, Wang W, et al. Detection of human cytomegalovirus in normal and neoplastic breast epithelium. Herpesviridae 2010;1(1):8. 11. Samanta M, Harkins L, Klemm K, Britt WJ, Cobbs CS. High prevalence of human cytomegalovirus in prostatic intraepithelial neoplasia and prostatic carcinoma. J Urol 2003;170(3):998–1002. 12. Melnick M, Sedghizadeh PP, Allen CM, Jaskoll T. Human cytomegalovirus and mucoepidermoid carcinoma of salivary glands: cell-specific localization of active viral and oncogenic signaling proteins is confirmatory of a causal relationship. Exp Mol Pathol 2012;92(1):118–25. 13. Price RL, Bingmer K, Harkins L, Iwenofu OH, Kwon CH, Cook C, et al. Cytomegalovirus infection leads to pleomorphic rhabdomyosarcomas in trp53+/− mice. Cancer Res 2012;72(22):5669–74. 14. Baryawno N, Rahbar A, Wolmer-Solberg N, Taher C, Odeberg J, Darabi A, et al. Detection of human cytomegalovirus in medulloblastomas reveals a potential therapeutic target. J Clin Invest 2011;121(10):4043–55. 15. Sabatier J, Uro-Coste E, Pommepuy I, Labrousse F, Allart S, Tremoulet M, et al. Detection of human cytomegalovirus genome and gene products in central nervous system tumours. Br J Cancer 2005;92(4):747–50. 16. Poltermann S, Schlehofer B, Steindorf K, Schnitzler P, Geletneky K, Schlehofer JR. Lack of association of herpesviruses with brain tumors. J Neurovirol 2006;12(2):90–9. 17. Lau SK, Chen YY, Chen WG, Diamond DJ, Mamelak AN, Zaia JA, et al. Lack of association of cytomegalovirus with human brain tumors. Mod Pathol 2005;18(6):838–43. 18. Maussang D, Verzijl D, van Walsum M, Leurs R, Holl J, Pleskoff O, et al. Human cytomegalovirus-encoded chemokine receptor US28 promotes tumorigenesis. Proc Natl Acad Sci U S A 2006;103(35):13068–73. 19. Cinatl Jr J, Vogel JU, Kotchetkov R, Wilhelm Doerr H. Oncomodulatory signals by regulatory proteins encoded by human cytomegalovirus: a novel role for viral infection in tumor progression. FEMS Microbiol Rev 2004;28(1):59–77. 20. Mirimanoff RO, Gorlia T, Mason W, Van den Bent MJ, Kortmann RD, Fisher B, et al. Radiotherapy and temozolomide for newly diagnosed glioblastoma: recursive partitioning analysis of the EORTC 26981/22981-NCIC CE3 phase III randomized trial. J Clin Oncol 2006;24(16):2563–9. 21. Larsson S, Soderberg-Naucler C, Wang FZ, Moller E. Cytomegalovirus DNA can be detected in peripheral blood mononuclear cells from all seropositive and most seronegative healthy blood donors over time. Transfusion (Paris) 1998;38(3):271–8. 22. Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M, et al. MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 2005;352(10):997–1003. 23. Dziurzynski K, Wei J, Qiao W, Hatiboglu MA, Kong LY, Wu A, et al. Gliomaassociated cytomegalovirus mediates subversion of the monocyte lineage to a tumor propagating phenotype. Clin Cancer Res 2011;17(14):4642–9. 24. Cannon MJ, Schmid DS, Hyde TB. Review of cytomegalovirus seroprevalence and demographic characteristics associated with infection. Rev Med Virol 2010;20(4):202–13. 25. Slinger E, Maussang D, Schreiber A, Siderius M, Rahbar A, Fraile-Ramos A, et al. HCMV-encoded chemokine receptor US28 mediates proliferative signaling through the IL-6-STAT3 axis. Sci Signal 2010;3(133):ra58. 26. Maussang D, Langemeijer E, Fitzsimons CP, Stigter-van Walsum M, Dijkman R, Borg MK, et al. The human cytomegalovirus-encoded chemokine receptor US28
42
27. 28.
29.
30.
31.
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42 promotes angiogenesis and tumor formation via cyclooxygenase-2. Cancer Res 2009;69(7):2861–9. Soroceanu L, Cobbs CS. Is HCMV a tumor promoter? Virus Res 2011;157(2): 193–203. Bongers G, Maussang D, Muniz LR, Noriega VM, Fraile-Ramos A, Barker N, et al. The cytomegalovirus-encoded chemokine receptor US28 promotes intestinal neoplasia in transgenic mice. J Clin Invest 2010;120(11):3969–78. Streblow DN, Soderberg-Naucler C, Vieira J, Smith P, Wakabayashi E, Ruchti F, et al. The human cytomegalovirus chemokine receptor US28 mediates vascular smooth muscle cell migration. Cell 1999;99(5):511–20. Muganda P, Carrasco R, Qian Q. The human cytomegalovirus IE2 86 kDa protein elevates p53 levels and transactivates the p53 promoter in human fibroblasts. Cell Mol Biol (Noisy-le-grand) 1998;44(2):321–31. Boldogh I, AbuBakar S, Albrecht T. Activation of proto-oncogenes: an immediate early event in human cytomegalovirus infection. Science 1990; 247(4942):561–4.
32. Straat K, Liu C, Rahbar A, Zhu Q, Liu L, Wolmer-Solberg N, et al. Activation of telomerase by human cytomegalovirus. J Natl Cancer Inst 2009;101(7): 488–97. 33. Kleihues P, editor. Pathology and genetic of tumors of the nervous system. 2nd ed. Lyon: International Agency for Research on Cancer (IARC); 2000. 34. Siew VK, Duh CY, Wang SK. Human cytomegalovirus UL76 induces chromosome aberrations. J Biomed Sci 2009;16:107. 35. Li YS, Ramsay DA, Fan YS, Armstrong RF, Del Maestro RF. Cytogenetic evidence that a tumor suppressor gene in the long arm of chromosome 1 contributes to glioma growth. Cancer Genet Cytogenet 1995;84(1):46–50. [36]. Fortunato EA, Sanchez V, Yen JY, Spector DH. Infection of cells with human cytomegalovirus during S phase results in a blockade to immediate-early gene expression that can be overcome by inhibition of the proteasome. Journal of Virololy 2002;76:5369–79. 37. Cinatl J, Scholz M, Kotchetkov R, Vogel JU, Doerr HW. Molecular mechanisms of the modulatory effects of HCMV infection in tumor cell biology. Trends Mol Med 2004;10(1):19–23.
Contents lists available at SciVerse ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
Human cytomegalovirus infection levels in glioblastoma multiforme are of prognostic value for survival Afsar Rahbar a , Abiel Orrego b , Inti Peredo c , Mensur Dzabic a , Nina Wolmer-Solberg a , Klas Strååt d , Giuseppe Stragliotto c , Cecilia Söderberg-Nauclér a,∗ a
Department of Medicine Solna, Experimental Cardiovascular Research Unit, Center for Molecular Medicine, Karolinska Institute, Stockholm, Sweden Cancer Center, Karolinska University Hospital, Stockholm, Sweden c Department of Neurosurgery, Karolinska University Hospital, Stockholm, Sweden d Department of Cell and Molecular Biology (CMB), Stockholm, Sweden b
a r t i c l e
i n f o
Article history: Received 27 September 2012 Received in revised form 20 December 2012 Accepted 21 December 2012 Keywords: HCMV GBM Survival TTP
a b s t r a c t Background: Patients with glioblastoma multiforme (GBM) generally live 12–15 months after diagnosis, despite maximal surgical resection, adjuvant radiotherapy, and chemotherapy. HCMV has been detected in 90–100% of GBMs. We recently found that low grade HCMV infection in GBM tumours was highly associated with survival over 18 months (case–control study). Here, we sought to determine whether low-grade HCMV infection in GBMs is associated with prolonged survival in a consecutive patient cohort, analysed retrospectively. Study design: Tumour samples from 75 consecutive GBM patients treated surgically at Karolinska University Hospital in 2004–2005 were examined by immunohistochemistry (IHC) and in situ hybridization for HCMV proteins and DNA, respectively. Tumours were graded 1–4, depending on the percentage of positive cells by IHC. Low-grade HCMV was defined as grade 1 (<25% of HCMV infected tumour cells). Time to tumour progression (TTP) and survival data were analysed with Cox regression and Kaplan–Meier models. Results: HCMV infection was detected in 74 of 75 tumours (99%). In patients with low-grade HCMV infection, median survival was 20 months longer than in patients with high-grade infections (P = 0.036, HR: 2.2), and TTP was 8 months longer (P = 0.1, HR: 1.8). Two-year survival was much higher in patients with low-grade HCMV infection (63.6% vs. 17.2%, P = 0.003). Conclusion: The longer survival in patients whose tumours had low-grade HCMV infection suggests that the level of HCMV infection in GBMs has a prognostic value and that HCMV may contribute to the pathogenesis of GBM. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Glioblastoma multiforme (GBM) is the most malignant primary tumour of the central nervous system.1 The median survival after diagnosis is 12–15 months.2 Although 90–100% of GBMs are infected with human cytomegalovirus (HCMV),3–5 it is unclear whether this virus contributes to tumour development or progression. HCMV is a herpesvirus carried by 70–100% of the world’s population. In healthy persons, HCMV infection is generally asymptomatic; however, the virus remains latent in the bone marrow
∗ Corresponding author at: Department of Medicine, Centre for Molecular Medicine, Karolinska Institute, S-171 76 Stockholm, Sweden. Tel.: +46 8 51779896; fax: +46 8 313147. E-mail address: [email protected] (C. Söderberg-Nauclér). 1386-6532/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jcv.2012.12.018
and peripheral blood and can be reactivated by inflammation.6–8 Emerging evidence demonstrate that HCMV can be detected in certain malignant tumours, such as GBM,3–5 colon cancer,9 breast cancer10 and prostatic carcinoma,11 mucoepidermoid carcinoma of salivary glands12 and rhabdomyosarcomas.13 Using immunohistochemistry, flow cytometry, PCR and in situ hybridization techniques, we and others have detected HCMV in tumour tissue specimens.14 However, some investigators have not confirmed the finding of HCMV in tumours.15–17 In our own experience, we can readily detect the virus in frozen tumour specimens by indirect immunofluorescence or in paraffin embedded tumour tissue specimens using the high sensitive immunostaining protocols, which include optimal tissue fixation, antigen retrieval, use of appropriate antibodies for paraffin embedded tissues and blocking of nonspecific binding.3–5,9–12,14 However, we do not detect the virus in tumours using regular immunohistochemistry protocols. Therefore, different detection methods used by different investigators, in
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42 Table 1 Characteristics of 75 consecutive patients with GBM. Characteristics Age <50 ≥50 Sex Male Female RPA class III–IV V–VI Extent of surgery Partial Radical Gamma knife No Yes RT + temozolomide No Yes RT + adjuvant chemotherapy No Yes
n (%) 17 (23) 58 (77) 42 (56%) 33 (44) 34 (45) 41 (55) 20 (27) 55 (73) 69 (92) 6 (8) 60 (80) 15 (20) 26 (35) 49 (65)
Three patients were alive at study closure after a mean follow-up of 82.3 months. RT: radio therapy; RPA: recursive partition analysis.
particular the staining procedures, may be the main reasons why some studies have failed to detect HCMV in tumour tissues.15–17 Although one HCMV protein, US28, has an oncogenic potential,18 this virus is not considered to be oncogenic. Rather, it may be an oncomodulator that contributes to cancer development by modifying tumour cell biology.19 In such scenario, the level of HCMV infection in the tumour may affect disease progression and the prognosis of patients with HCMV positive tumours. In support of this hypothesis, we recently observed that a low-grade HCMV infection was strongly associated with long-term survival in GBM patients (P = 0.0006)5 in a case–control study. Here, we aimed to determine if low-grade HCMV infection in a consecutive cohort of GBM patients was associated with prolonged overall survival. We also defined the prevalence of Epstein–Barr virus (EBV) and herpesvirus 6 (HHV-6) as controls for viral persistence.
late protein of 47–55 kDa) (both antibodies IgG2a, Chemicon International), EBV (BZLF1 protein, IgG1; DakoCytomation, Glostrop, Denmark) andHHV-6 (IgG1, Abcam, Cambridge, UK). Antibodies against smooth muscle cell alpha actin (IgG2a, Biogenex, San Ramon, CA) and von Willebrand factor (IgG1, DakoCytomation) served as isotype controls. The specimens were examined by AR and AO; neither had access to the clinical records of the patients. These samples were scored for HCMV proteins in the specimens as previously described5 (Table 2). Tumour cell proliferation (Ki-67), p53 mutation, mitosis, expression of glial fibrillary acidic protein, epidermal growth factor receptor (EGF-R), and VEGF-R were assessed with automated immunohistochemical staining protocols at our hospital, and the findings were analysed for associations with high- and low-grade HCMV infection (Table 2). 2.3. Polymerase chain reaction (PCR) and DNA sequencing Primary cultures were established from the tumours of five newly diagnosed GBM patients. DNA samples were prepared with QIAamp DNA mini-kits (Qiagen, Valencia, CA), and analysed by PCR for the HCMV major immediate-early (MIE) gene.3,21 DNA from PCR products were sequenced with an ABI 3730 DNA analyser. Sequence data were analysed by BLAST searches. As controls, DNA samples were analysed for herpes simplex virus (HSV)-1 (gpD, 188 bp), HSV2 (gpG, 149 bp), and HHV-6 (U67) by real-time PCR at our hospital. 2.4. Statistical analysis Overall survival and time to tumour progression (TTP) were analysed by a medical statistician (Fredrik Hansson, Norma, Lund, Sweden) using Cox regression models; the covariates were lowgrade infection (no infection or grade 1) or high-grade infection (grades 2–4), age, extent of surgery, and RPA class. All patients who were alive or did not have progression at the time of investigation were censored in the analysis (April 16, 2012). The results are presented as hazard ratios with 95% confidence intervals (CIs). Survival graphs were created with the Kaplan–Meier life-table method. The median times to end points were estimated from the Kaplan–Meier curves, with 95% CIs.
2. Methods
3. Results
2.1. Patient samples
3.1. HCMV infection level has a high prognostic value for GBM patients
Paraffin-embedded tumour specimens were obtained from 75 consecutive GBM patients who underwent their first surgery after diagnosis at Karolinska University Hospital in 2004–2005 (two patients were excluded due to lack of biopsy material, 51 patients were analysed in our previously published case–control study5 ). All patients had World Health Organization (WHO) grade IV GBM and received standard treatment (Table 1). The diagnosis was independently confirmed by two experienced neuropathologists (mainly AO). Patients were classified according to the adapted European Organization for Research and Treatment of Cancer recursive partition analysis classification (Table 1).20 2.2. Immunohistochemistry and in situ hybridization All samples were analysed by immunohistochemistry for HCMV-IEA and late antigen (LA) as described.3 Seventeen randomly selected samples were also stained for HHV-6 and EBV as viral specificity controls and 19 randomly selected tissue sections were analysed by in situ hybridization as described.3 Primary antibodies used were against HCMV-IEA (reacts with an immediate early non-structural antigen of 68–72 kDa), HCMV-LA (reacts with a
37
HCMV IE protein was detected in 74 of 75 biopsies (99%) and HCMV-LA in 70 of 75 (93%). Seventeen randomly selected samples were subjected for analyses of HHV-6 and EBV as specificity samples (Fig. 1A). Blood vessels within tumours were also positive for HCMV IE and HCMV-LA in 45 of 75 patients (60%) (Fig. 1A). One patient was negative for HCMV-IEA (Fig. 1A), HCMV-LA, EBV, and HHV-6 (data not shown). Noncancerous cells near tumour cells were consistently negative for HCMV. HCMV infection was confirmed by in situ hybridization in all randomly selected HCMVpositive samples but not in the single HCMV-negative sample (Fig. 1B). To further confirm the presence of HCMV in GBMs, we also established primary cultures from surgical tumour specimens of five patients with newly diagnosed GBM. DNA was extracted at passage 1 and amplified by PCR assays for the gene encoding HCMV MIE. PCR assays for the genes encoding HHV-6 (U67), HSV-1 (gpD), and HSV-2 (gpG) served as controls. All samples were positive for HCMV-MIE and negative for the other viruses (data not shown). DNA sequence analysis of HCMV MIE PCR products revealed HCMV MIE DNA sequences in all five GBM cultures (data not shown),
38
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42
Table 2 Hazard ratio according to HCMV infection level in the tumour, age, extent of resection, and RPA subclass using Cox regression model for survival data in 75 consecutive patients. Characteristic
HCMV infection (n: 11 low-grade vs. 64 high-grade) Age, years (n: 17 < 50 vs. 58 ≥ 50, n = 58) Extent of surgery (n: 55 radical vs. 20 partial) RPA class (n: 12 III and 22 IV vs. 34 V and 7 VI)
HCMV-IEA
HCMV-LA
P
HR
0.036 0.23 0.002 0.0019
2.2 1.44 2.6 2.4
95% CI 1.0–4.38 0.79–2.6 1.44–4.82 1.38–4.2
P
HR
0.6 0.23 0.009 0.0003
1.14 1.46 2.2 2.8
95% CI 0.70–1.86 0.79–2.67 1.23–3.9 1.59–4.78
HCMV: human cytomegalovirus; IEA: immediate-early antigen; LA: late antigen; n: number of patients; RPA: recursive partition analysis; HR: hazard ratio; CI: confidence interval.
but the DNA sequences were distinctly different from wild type (VR1814, TB40) or laboratory virus strains (AD169 and Towne), which excluded contamination. HCMV Infection was graded according to the estimated percentage of cells expressing HCMV IEA or HCMV Late proteins in the specimens as previously described5 ; negative or grade 1, <25%; grade 2, ≥25–50%; grade 3, ≥50–75%; and grade 4, >75% HCMV positive cells in the tumour. Patients with an HCMV negative or an HCMV grade 1 tumour were defined to have a low grade HCMV infection in their tumour, while patients with a high grade infection had a tumour graded as 2, 3 or 4 for HCMV infection. One patient was HCMV IEA negative and survived 34 months. 14.7% of patients had grade 1 HCMV-IEA infection and their median OS was 33 months (n = 11), 8% of the patients had grade 2 (median OS was 13 months, n = 6), 22.7% of patients had grade 3 (median OS was 12 months, n = 17) and 54.7% of patients had grade 4 (median OS was 13 months, n = 41). Thus, as we previously observed there was no difference in OS (Median OS was 12.6 months) between HCMV IE grade 2–4, which was categorized as high-grade HCMV infection. Negative or grade 1 was regarded as low-grade HCMV infection. Based on this categorization we next analysed whether there was an association between low-grade HCMV infection and survival. Eleven of the 75 patients (14.7%) had low-grade HCMVIEA expression, and 64 (85.3%) had high-grade infection (grades 2–4). Median overall survival was significantly higher in patients with low-grade HCMV-IEA than in those with high-grade HCMVIEA (33 vs. 13 months, P = 0.036; HR: 2.2, 95% CI: 1.0–4.38) (3 patients were censored; one with low-grade and two with highgrade HCMV-IEA infections) (Fig. 2A). The 2-years survival rate was significantly higher in patients with low-grade than in those with high-grade HCMV-IEA infection (63.6% vs. 17.2%, P = 0.003, Fisher Exact Test). HCMV-LA infection had no effect on median overall survival (14.5 vs. 13 months, P = 0.6; HR: 1.4, 95% CI: 0.70–1.86). One patient with low-grade and two with high-grade HCMV-LA infections were censored (Fig. 2B). The median TTP did not differ significantly in patients with lowgrade HCMV-IEA or LA expression than in those with high-grade IEA or LA expression (for HCMV-IEA: 14 vs. 6 months, P = 0.1; HR: 1.8, 95% CI: 0.9–3.55; (Fig. 2C) and for HCMV-LA: 8.25 vs. 5 months, P = 0.13; HR: 1.49, 95% CI: 0.89–2.47) (Fig. 2D). Multivariable analysis of overall survival was performed to evaluate HCMV infection expression in relation to other prognostic factors known to be important for survival (e.g., age, extent of surgery, and RPA class) (Table 2). The multivariable analysis in relation to HCMV infection showed that overall survival was not significantly related to age (HCMV-IEA HR: 1.44, P = 0.23; HCMV-LA HR: 1.46, P = 0.23). Patients with radical surgery had significantly improved survival (HCMV-IEA HR: 2.63, P = 0.002; HCMV-LA HR: 2.2, P = 0.009). Risk of death in patients with RPAs of V or VI was 2.4-fold (HCMV-IEA HR: 2.4, P = 0.002) and 2.8-fold (HCMVLA HR: 2.8, P = 0.0003) higher than in those with RPAs of III or IV (Table 2). Furthermore, multivariable analysis in relation to HCMV infection showed that TTP was not related to age (HCMV-IEA
HR: 0.8, P = 0.50; HCMV-LA HR: 0.72, P = 0.33). However, patients with radical surgery (HCMV-IEA HR: 2.23, P = 0.007; HCMV-LA HR: 2.21, P = 0.008) or with RPAs III-IV (HCMV-IEA HR: 2.46, P = 0.0023; HCMV-LA HR: 2.7, P = 0.0007) had significantly longer TTP (Table 2). Overall survival and TTP did not differ in patients with low-grade HCMV-IEA immunoreactivity who underwent radiotherapy (RT) or RT in combination with adjuvant chemotherapy. The cause of death was related to the tumour in all patients. 3.2. Phenotypic characterization of GBMs Further evaluation of tumour markers examined for clinical purpose revealed p53 mutations in 27 of 55 (49%) samples with high-grade HCMV-IEA infection and 2 of 9 (22%) with low-grade infection. p53 mutations were detected in 26 of 47 (55%) tumours with high-grade HCMV-LA infection and in 3 of 17 (18%) with low-grade infection. While only a trend between p53 mutations for HCMV-IEA was observed, there was a significant association between HCMV late protein expression and p53 mutations (P = 0.0075; Chi-square test, Table 3). The extent of mitosis did not differ significantly between the groups (Table 3). Expression of EGFR and VEGF-R did not differ in the low- and high-grade groups (Table 3). 4. Discussion Despite tremendous progress in understanding the molecular aspects of GBM, the aetiology of these tumours is unclear. In this study, we confirmed the previously reported high frequency of HCMV infections in GBMs. Importantly; the level of HCMV infection at the time of diagnosis had a prognostic value for patient survival. Patients whose tumours had low-grade infections lived 2.5 times longer than those with high-grade infections (33 vs. 13 months for HCMV-IEA, P = 0.036; HR: 2.2, 95% CI: 1.0–4.38) and had a higher 2-year survival (63.6% vs. 17.2%, P = 0.003). However, median TTP was not significantly longer in patients with low-grade infection (14 months vs. 6 months for HCMV-IEA, P = 0.1; HR: 1.8, 95% CI: 0.9–3.55). Several factors have shown prognostic potential in patients with GBMs. Historically, young age (<50 years) and good neurologic function (Karnovsky score >90 and WHO performance score 0) are considered as positive prognostic factors. Here, the risk of death was 2.2-fold higher (P = 0.036; HR: 2.2, 95% CI: 1.0–4.38) in patients with high-grade HCMV-IE protein expression than in those with low-grade infection and 2.4-fold higher in patients with RPA class ≥V than in those with RPA class
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42
39
Fig. 1. Detection of HCMV and HHV-6 expression by immunohistochemistry (A) and HCMV DNA by in situ hybridization (B) in GBMs, bar, 50 m.
The association between low-grade HCMV infection and improved survival imply that HCMV may contribute to GBM progression. An active HCMV infection in tumour tissues might result from a new primary infection or reactivation of a latent
infection.23,24 Consistent with the latter possibility, GBMs are highly associated with inflammation and exert a variety of local and systemic immunosuppressive effects. We earlier demonstrated that latent HCMV can be reactivated by inflammatory stimuli.8
40
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42
Fig. 2. HCMV infection levels in GBMs are of high prognostic value for patient outcome (A and B) and for TTP (C and D).
HCMV may be transmitted to tumour cells and through its unique ability to influence cellular and immunological functions, affect the malignancy of the tumour by and its sensitivity to therapy. The HCMV protein US28 activates NF-kB and phosphorylates STAT3, causing COX-2 expression, VEGF production, and IL-6, resulting in inflammation, angiogenesis and tumour formation in vivo.25–27 Transgenic mice that express US28 in intestinal epithelial cells developed adenomas and adenocarcinomas, which was associated with inhibited glycogen synthase 3 function, promoted accumulation of -catenin, and increased expression of Wnt target genes involved in the control of the cell proliferation.28 HCMV-US28 also
induces cellular migration29 and angiogenesis26 and may thereby enhance tumour invasiveness – another hallmark of GBMs. HCMV IE72 and IE86 induce expression of Rb, a key cell-cycle regulator that allows the transition from G1 to S phase and interacts with p53; under certain circumstances the IE proteins may result in uncontrolled DNA replication, transformation, and p53 mutations.30,31 We found p53 mutations in 26 of 47 samples with high-grade HCMV-LA infection and in 8 of 17 with low-grade infection. Expression of HCMV-LA was significantly associated with p53 mutations (P = 0.0075, Chi-square test), suggesting a potential viral effect on p53.30 Recently, we also showed that HCMV-IE72
Table 3 Expression of HCMV and available clinical data on the expression of glial fibrillary acidic protein, epidermal growth factor receptor, vascular endothelial growth factor receptor, the presence of p53 mutations, and tumour cell proliferation in GBMs. HCMV-IEA
Consecutive patients (n = 75) p53 mutation (n = 64) GFAP (n = 75) Ki-67 (n = 71) Ki-67 MIB index interval (%) Mitosis (n = 50) EGF-R (n = 60) VEGF-R (n = 44)
HCMV-LA
Grades 0–1 n (%)
Grades 2–4 n (%)
Grades 0–1 n (%)
Grades 2–4 n (%)
11 (15) 2/9 (22) 11/11 (100) 10/10 (100) 5–40 10/11 (91) 7/7 (100) 7/8 (88)
64 (85) 27/55 (49) 64/64 (100) 61/61 (100) 10–90 38/39 (97) 44/53 (83) 35/36 (97)
32 (43) 3/17 (18)a 22/22 (100) 21/21 (100) – 10/11 (91) 15/15 (100) 7/8 (88)
43 (57) 26/47 (55)a 53/53 (100) 50/50 (100) – 38/39 (97) 36/45 (80) 35/36 (97)
IEA: immediate-early antigen; LA: late antigen; GFAP: glial fibrillary acidic protein; EGF-R: epidermal growth factor receptor; VEGF-R: vascular growth factor receptor. a P = 0.0075 (Chi-square test).
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42
may be involved in oncogenesis. HCMV IE72 stimulated telomerase activity, which is important for cellular immortalization and transformation.32 hTERT expression co-localized with HCMV proteins in GBMs, and hTERT was highly expressed in tumour areas with high-grade HCMV infection.32 Furthermore, the protooncogenes c-myc, c-fos, and c-jun, cyclin E, cyclin A, and Cdk2 are all rapidly activated in HCMV infected cells.27 HCMV infection in S-phase cells results in specific chromosome 1 breaks in 1q21 and 1q42, and deletions in 1q42 have been associated with GBM.33–36 Hence, reactivation of HCMV may occur in GBM patients and promote the oncogenic potential of HCMV, possibly through a mechanism involving Wnt, VEGF, STAT3 phosphorylation/IL-6 production, and induced telomerase activity, oncogene expression and DNA damage. Thus, HCMV infection may result in a more aggressive tumour phenotype with a poor prognosis. Importantly, the presence of HCMV in glioblastoma may provide a new target for therapy. We recently detected HCMV/COX-2 expression in primary medulloblastoma tumours and in medulloblastoma cell lines; the growth of medulloblastoma xenografts in vivo (nude mice) was inhibited by 72% using a COX-2 inhibitor in combination with an anti-HCMV drug.14 No effect of anti-viral therapy was observed in HCMV negative tumours. As HCMV infected tumour cells have been shown to exhibit increased resistance to chemotherapy in vitro,37 anti-viral therapy may enhance the clinical response to chemotherapy. In summary, our findings confirm our recent observation of a strong association between low-grade HCMV infection and longterm survival in GBM patients.5 We here provide further evidence that an active HCMV infection in GBM patients has a prognostic value for patient outcome in a consecutive cohort of GBM patients operated at our hospital during 2004–2005. In this cohort, we could estimate TTP and overall survival by Kaplan–Meier analyses, which was not possible in our previous case control cohort study. We found that patients with low grade infection survived 20 months longer than patients with high grade HCMV infection in their tumours at diagnosis (P = 0.036). Since HCMV exhibits oncomodulatory effects and may prevent the action of chemotherapy, treatment to control HCMV infection in tumour tissues may suppress tumour progression, and increase the tumour’s sensitivity to conventional therapy. Funding This work was supported by the Torsten and Ragnar Söderbergs’ Research Foundation, the Swedish Research Council (10350 and K2007-56X-12615-10-3), the Swedish Children Cancer Research Foundation (Project 05/100, 07/065), the Swedish Cancer Foundation (5044-B05-01XAB), Swedish Society for Medical Research (SLS), The Royal Swedish Academy of Sciences (KVA), Goljes Memory Foundation, Magnus Bergvalls Foundation, Swedish Society for Medical Research (SSMF), and Tore Nilsons Foundation. Competing interests The authors declare no conflicts of interest. However, CS-N has served as a lecturer and once in Scientific Advisory Board for Roche and holds an investigational grant from the same company for a clinical study to evaluate the effect of antiviral treatment in GBM patients. Ethical approval The study was approved by the Ethics Committee at the Karolinska Institute (2008/628-31/2).
41
Acknowledgment We thank Stephen Ordway for editorial assistance. References 1. Graham DILP, editor. Greenfield’s neuropathology. 6th ed. London: Arnold; 1996. 2. Affronti ML, Heery CR, Herndon II JE, Rich JN, Reardon DA, Desjardins A, et al. Overall survival of newly diagnosed glioblastoma patients receiving carmustine wafers followed by radiation and concurrent temozolomide plus rotational multiagent chemotherapy. Cancer 2009;115(15):3501–11. 3. Cobbs CS, Harkins L, Samanta M, Gillespie GY, Bharara S, King PH, et al. Human cytomegalovirus infection and expression in human malignant glioma. Cancer Res 2002;62(12):3347–50. 4. Mitchell DA, Xie W, Schmittling R, Learn C, Friedman A, McLendon RE, et al. Sensitive detection of human cytomegalovirus in tumors and peripheral blood of patients diagnosed with glioblastoma. Neuro-oncol 2008;10(1):10–8. 5. Rahbar A, Stragliotto G, Orrego A, Peredo I, Taher C, Willems J, et al. Low levels of human cytomegalovirus infection in glioblastoma multiforme associates with patient survival—a case–control study. Herpesviridae 2012;3:3. 6. ES M, editor. Fields virology. 4th ed. Philadelphia: Lippincott Williams and Wilkins; 2001. 7. Taylor-Wiedeman J, Sissons JG, Borysiewicz LK, Sinclair JH. Monocytes are a major site of persistence of human cytomegalovirus in peripheral blood mononuclear cells. J Gen Virol 1991;72(Pt 9):2059–64. 8. Soderberg-Naucler C, Fish KN, Nelson JA. Reactivation of latent human cytomegalovirus by allogeneic stimulation of blood cells from healthy donors. Cell 1997;91(1):119–26. 9. Harkins L, Volk AL, Samanta M, Mikolaenko I, Britt WJ, Bland KI, et al. Specific localisation of human cytomegalovirus nucleic acids and proteins in human colorectal cancer. Lancet 2002;360(9345):1557–63. 10. Harkins LE, Matlaf LA, Soroceanu L, Klemm K, Britt WJ, Wang W, et al. Detection of human cytomegalovirus in normal and neoplastic breast epithelium. Herpesviridae 2010;1(1):8. 11. Samanta M, Harkins L, Klemm K, Britt WJ, Cobbs CS. High prevalence of human cytomegalovirus in prostatic intraepithelial neoplasia and prostatic carcinoma. J Urol 2003;170(3):998–1002. 12. Melnick M, Sedghizadeh PP, Allen CM, Jaskoll T. Human cytomegalovirus and mucoepidermoid carcinoma of salivary glands: cell-specific localization of active viral and oncogenic signaling proteins is confirmatory of a causal relationship. Exp Mol Pathol 2012;92(1):118–25. 13. Price RL, Bingmer K, Harkins L, Iwenofu OH, Kwon CH, Cook C, et al. Cytomegalovirus infection leads to pleomorphic rhabdomyosarcomas in trp53+/− mice. Cancer Res 2012;72(22):5669–74. 14. Baryawno N, Rahbar A, Wolmer-Solberg N, Taher C, Odeberg J, Darabi A, et al. Detection of human cytomegalovirus in medulloblastomas reveals a potential therapeutic target. J Clin Invest 2011;121(10):4043–55. 15. Sabatier J, Uro-Coste E, Pommepuy I, Labrousse F, Allart S, Tremoulet M, et al. Detection of human cytomegalovirus genome and gene products in central nervous system tumours. Br J Cancer 2005;92(4):747–50. 16. Poltermann S, Schlehofer B, Steindorf K, Schnitzler P, Geletneky K, Schlehofer JR. Lack of association of herpesviruses with brain tumors. J Neurovirol 2006;12(2):90–9. 17. Lau SK, Chen YY, Chen WG, Diamond DJ, Mamelak AN, Zaia JA, et al. Lack of association of cytomegalovirus with human brain tumors. Mod Pathol 2005;18(6):838–43. 18. Maussang D, Verzijl D, van Walsum M, Leurs R, Holl J, Pleskoff O, et al. Human cytomegalovirus-encoded chemokine receptor US28 promotes tumorigenesis. Proc Natl Acad Sci U S A 2006;103(35):13068–73. 19. Cinatl Jr J, Vogel JU, Kotchetkov R, Wilhelm Doerr H. Oncomodulatory signals by regulatory proteins encoded by human cytomegalovirus: a novel role for viral infection in tumor progression. FEMS Microbiol Rev 2004;28(1):59–77. 20. Mirimanoff RO, Gorlia T, Mason W, Van den Bent MJ, Kortmann RD, Fisher B, et al. Radiotherapy and temozolomide for newly diagnosed glioblastoma: recursive partitioning analysis of the EORTC 26981/22981-NCIC CE3 phase III randomized trial. J Clin Oncol 2006;24(16):2563–9. 21. Larsson S, Soderberg-Naucler C, Wang FZ, Moller E. Cytomegalovirus DNA can be detected in peripheral blood mononuclear cells from all seropositive and most seronegative healthy blood donors over time. Transfusion (Paris) 1998;38(3):271–8. 22. Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M, et al. MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 2005;352(10):997–1003. 23. Dziurzynski K, Wei J, Qiao W, Hatiboglu MA, Kong LY, Wu A, et al. Gliomaassociated cytomegalovirus mediates subversion of the monocyte lineage to a tumor propagating phenotype. Clin Cancer Res 2011;17(14):4642–9. 24. Cannon MJ, Schmid DS, Hyde TB. Review of cytomegalovirus seroprevalence and demographic characteristics associated with infection. Rev Med Virol 2010;20(4):202–13. 25. Slinger E, Maussang D, Schreiber A, Siderius M, Rahbar A, Fraile-Ramos A, et al. HCMV-encoded chemokine receptor US28 mediates proliferative signaling through the IL-6-STAT3 axis. Sci Signal 2010;3(133):ra58. 26. Maussang D, Langemeijer E, Fitzsimons CP, Stigter-van Walsum M, Dijkman R, Borg MK, et al. The human cytomegalovirus-encoded chemokine receptor US28
42
27. 28.
29.
30.
31.
A. Rahbar et al. / Journal of Clinical Virology 57 (2013) 36–42 promotes angiogenesis and tumor formation via cyclooxygenase-2. Cancer Res 2009;69(7):2861–9. Soroceanu L, Cobbs CS. Is HCMV a tumor promoter? Virus Res 2011;157(2): 193–203. Bongers G, Maussang D, Muniz LR, Noriega VM, Fraile-Ramos A, Barker N, et al. The cytomegalovirus-encoded chemokine receptor US28 promotes intestinal neoplasia in transgenic mice. J Clin Invest 2010;120(11):3969–78. Streblow DN, Soderberg-Naucler C, Vieira J, Smith P, Wakabayashi E, Ruchti F, et al. The human cytomegalovirus chemokine receptor US28 mediates vascular smooth muscle cell migration. Cell 1999;99(5):511–20. Muganda P, Carrasco R, Qian Q. The human cytomegalovirus IE2 86 kDa protein elevates p53 levels and transactivates the p53 promoter in human fibroblasts. Cell Mol Biol (Noisy-le-grand) 1998;44(2):321–31. Boldogh I, AbuBakar S, Albrecht T. Activation of proto-oncogenes: an immediate early event in human cytomegalovirus infection. Science 1990; 247(4942):561–4.
32. Straat K, Liu C, Rahbar A, Zhu Q, Liu L, Wolmer-Solberg N, et al. Activation of telomerase by human cytomegalovirus. J Natl Cancer Inst 2009;101(7): 488–97. 33. Kleihues P, editor. Pathology and genetic of tumors of the nervous system. 2nd ed. Lyon: International Agency for Research on Cancer (IARC); 2000. 34. Siew VK, Duh CY, Wang SK. Human cytomegalovirus UL76 induces chromosome aberrations. J Biomed Sci 2009;16:107. 35. Li YS, Ramsay DA, Fan YS, Armstrong RF, Del Maestro RF. Cytogenetic evidence that a tumor suppressor gene in the long arm of chromosome 1 contributes to glioma growth. Cancer Genet Cytogenet 1995;84(1):46–50. [36]. Fortunato EA, Sanchez V, Yen JY, Spector DH. Infection of cells with human cytomegalovirus during S phase results in a blockade to immediate-early gene expression that can be overcome by inhibition of the proteasome. Journal of Virololy 2002;76:5369–79. 37. Cinatl J, Scholz M, Kotchetkov R, Vogel JU, Doerr HW. Molecular mechanisms of the modulatory effects of HCMV infection in tumor cell biology. Trends Mol Med 2004;10(1):19–23.