Lymphotropism of Merkel Cell Polyomavirus Infection, Nova Scotia, Canada

Lymphoid cells may be a site for virus persistence.

M erkel cell polyomavirus (MCPyV) was fi rst described in 2008 (1) as a new human virus associated with Merkel cell carcinoma (MCC), an uncommon but aggressive form of skin cancer. Subsequent studies have reported the presence of MCPyV in 24%-100% of MCCs from patients from the United States, Germany, France, the Netherlands, and Australia (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11). Findings of the clonal integration of MCPyV in tumor cell genomes, tumor-associated mutations in the large T-antigen (T-ag) gene, and large T-ag expression in tumors suggest that MCPyV is not only associated with MCC, but that it may be the causative agent (1,12,13).
However, the natural reservoir of MCPyV in infected hosts remains to be identifi ed. MCPyV DNA has been detected at low copy number in some non-MCC skin tumors, in normal tissues of skin and the gastrointestinal tract, and in a few nasopharyngeal aspirates and blood samples, including infl ammatory monocytes (1,5,6,11,(14)(15)(16).
Lymphocytes can disseminate viruses throughout a host and may provide sites of viral persistence. Human polyomaviruses are known to establish persistent infections in healthy persons, to undergo periodic reactivation and replication, and to cause disease in susceptible hosts. Some polyomaviruses are lymphotropic; BK virus, simian virus 40, and JC virus DNA sequences have been detected in human lymphoid tissues, blood cells, and lymphomas (17)(18)(19)(20). Recently, Shuda et al. (13) reported the presence of MCPyV in a low percentage (2.2%) of hematolymphoid malignancies. In this study, we investigated the presence of MCPyV in benign lymph nodes and malignant lymphomas in patients from Canada.

Patients and Samples
This study was approved by the Capital Health Research Ethics Board, Halifax, Nova Scotia, Canada, and by the Baylor College of Medicine Institutional Review Board, Houston. A total of 353 frozen specimens from various body sites were analyzed. Tissues from 196 malignant lymphomas, including 152 non-Hodgkin lymphoma (NHL) and 44 Hodgkin lymphoma (HL) samples, were retrieved from the Department of Anatomical Pathology, Queen Elizabeth II Health Sciences Center, Halifax, Nova Scotia, Canada.
All cases were diagnosed during 1994 through 2001 and classifi ed according to World Health Organization criteria by using morphologic and immunohistochemical evaluation (21). The NHL samples were classifi ed as Bcell lymphoma (n = 133), NK/T-cell lymphoma (n = 18), or were unclassifi ed (n = 1). The HL samples were classifi ed as classical HL (n = 41) or nodular lymphocyte predominant HL (n = 3). Also included in our study were 157 nonlymphoma frozen tissue specimens, including 110 benign lymph node biopsy specimens from healthy patients or patients with infl ammatory disease, 27 lymph nodes from patients with metastatic carcinoma or melanoma, 7 biopsy specimens of infl ammatory tissues, and 13 other neoplastic (non-MCC) tissue samples.
Lymphoma specimens were obtained from 96 women and 98 men; patients' ages ranged from 15 to 88 years (mean 54.5 years). Of the 110 benign lymph nodes, 53 were from women and 57 from men, ranging in age from 17 to 85 years (mean 46 years). Lymph nodes with metastatic tumors were obtained from 16 women and 10 men, ranging in age from 22 to 86 years (mean 60.6 years). Of the 7 infl ammatory tissue samples, 3 were from women and 4 were from men 19-84 years of age (mean 52 years). Finally, the non-MCC tumors were obtained from 8 men and 5 women, ranging in age from 22 to 72 years (mean 50.5 years). Data were not available for 3 patients, 2 with lymphoma and 1 with metastatic cancer. The cancers of all lymphoma patients were staged according to the Ann Arbor system (22).
Formalin-fi xed, paraffi n-embedded biopsy specimens of skin cancers (4 MCCs and 4 melanomas), obtained from the archives of the Department of Pathology, Baylor College of Medicine, Houston, Texas, USA, served as controls for PCR. Specimens were collected during 1998-2008.

Immunohistochemical Analysis
The expression of MCPyV T-ag was detected by immunohistochemical analysis by using a BenchMark XT IHC (Ventana Medical Systems, Inc., Tucson, AZ, USA) system. Tissue sections on microslides were deparaffi nized with xylene, hydrated in serially diluted alcohol, and endogenous peroxidase activity was quenched. The sections were then treated with slightly basic Tris-based buffer for 30 min for antigen retrieval. Sections next were incubated with CM2B4, a monoclonal antibody against MCPyV large T-ag (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) at a dilution of 1:250, followed by a standard polymer detection kit (diaminobenzidine) that served as the chromogen. The slides were counterstained with hematoxylin, dehydrated, and mounted for examination. A sample from a patient with MCC served as a positive control.

DNA Extraction
DNA was extracted from frozen tissues by lysis buffer/ proteinase K treatment and phenol-chloroform extraction as previously described (20), omitting the deparaffi nization step. For the paraffi n-embedded specimens, sections were deparaffi nized, and DNA was extracted by using our previously described protocol (20). During sample processing, stringent precautions were taken to avoid cross-contamination between samples. The microtome was cleaned carefully, and the blade was replaced for sectioning of each tissue. Other safety measures included working within a biosafety hood located in a dedicated room free from plasmids and viruses and using a dedicated set of pipettors and barrier fi lter tips. In addition, a negative extraction control that lacked tissue was processed in parallel.

Real-Time Quantitative PCR
Samples were fi rst screened for the single-copy human RNase P gene by real-time quantitative PCR (qPCR), as described (20,23). Briefl y, 5 μL of each DNA sample (undiluted and 1:10 diluted) was used in 50-μL qPCRs. This strategy detected potential PCR inhibitors in the DNA preparations, determined the human cell equivalents in each DNA sample, and normalized MCPyV viral loads to human cell numbers.
MCPyV was detected by qPCR with primers and Taq-Man probe (Applied Biosystems, Foster City, CA, USA) designed to detect sequences from the unique coding region of the small tumor antigen (t-ag) gene of MCPyV. The sequences of the oligonucleotides were as follows: LT3fwd primer 5′-AGTGTTTTTGCTATCAGTGCTTTA TTCT-3′, corresponding to nt 632-659 of MCPyV350 (GenBank accession no. EU375803); LT3-rev primer 5′-CCACCAGTCAAAACTTTCCCA-3′, corresponding to nt 702-682; and fl uorogenic probe 5′-FAM-TGGTTT GGATTTCCTC-MGB-3′, corresponding to nt 661-676. The pCR.MCV350 plasmid described by Feng et al. (1) was used as a positive control. Amplifi cations were performed with the ABI Prism 7700 Sequence Detection System (Applied Biosystems, Foster City, CA, USA) by using the following cycling parameters: 2 min at 50°C, 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and of 1 min at 60°C. All PCRs were performed in duplicate. Reactions were considered positive if >10 viral genome copies/reaction were detected.

Conventional PCR
Conventional PCR was performed on a subset of DNA samples to validate the presence of MCPyV. The primer set used was the following: MC_F2 (5′-CTCATCCTCTGGATCCAGTAGC-3′) and MC_R2 (5′-CAGAAGAGATCCTCCCAGGTG-3′) specifi c for a conserved region between nt positions 1142 and 1267 of the large T-ag gene of MCPyV (GenBank accession no. EU375803) and gave a 126-bp fragment. The pCR. MCV350 positive control plasmid was added to the control PCR outside the clean room after tubes containing test DNAs and negative controls (without DNA template) were closed. PCRs were performed in a GeneAmp PCR System 2700 thermal cycler (Applied Biosystems). Thermal cycling parameters were as follows: 94°C for 2 min, followed by 45 cycles at 94°C for 15 s, 60°C for 15 s, and 72°C for 15 s, with a fi nal extension step at 72°C for 7 min. Amplifi ed fragments were visualized by electrophoresis on a 3% agarose gel and stained with ethidium bromide. All PCR products were then purifi ed by using the DNA Clean & Concentrator-5 system (Zymo Research, Orange, CA, USA) and sequenced by Lone Star Labs, Inc. (Houston, TX, USA).

Statistical Analysis
The z score of difference for proportions was used to test for a difference between selected groups and an outcome. Fisher exact test was used to test the distribution of clinical stage of disease and 5-year survival category between MCPyV-status groups; a p value <0.05 was considered signifi cant.

MCPyV Detection in MCC
We tested 4 MCC and 4 melanoma samples (fi xed and paraffi n-embedded) for the presence of MCPyV by qPCR. MCPyV sequences were detected in 2 of 4 MCC samples; none were detected in the 4 melanoma samples (Table 1). MCCs contained an average of 0.29 (range 0.02-0.56) MCPyV genome copies per cell. MCPyV in 1 sample was confi rmed by conventional PCR and sequence analysis. The sequence of the amplifi ed fragment showed 98% similarity with the MCC350 reference sequence (GenBank accession no. EU375803) and 100% homology with MCC349 (Gen-Bank accession no. FJ173813).

MCPyV Detection in Malignant Lymphomas, Lymphoid Tissues, and Other Infl ammatory or Neoplastic Tissues
A total of 196 frozen malignant lymphoma samples were tested for MCPyV sequences. The classifi cation of those samples is summarized in Table 2. Viral DNA was detected in 13 (6.6%) of the lymphomas (Tables 1, 2). As determined by qPCR, the viral copy numbers were relatively low. An average of 4.6 copies/10 4 cells (range 0.16-27 copies/10 4 cells, median 0.94 copies/10 4 cells) was detected in the MCPyV-positive lymphomas. Among the lymphomas, the overall frequency of MCPyV between NHL and HL cases was similar (6.6% and 6.8%; p = 1.0) ( Table 2).
MCPyV was identifi ed most frequently in chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/ SLL) (20.8%), a tumor type known to occur in MCC patients (Table 2). Patients with CLL/SLL had a high prevalence of MCPyV compared to all other B-cell lymphomas (20.8% vs. 2.8%; p = 0.01). The virus loads in CLL/SLL were similar to those found in patients with other types of lymphoma. Characteristics of MCPyV-positive and MCPyV-negative CLL/SLL patients are shown in Table 3. None of the 24 CLL/SLL patients had MCC.
A total of 110 frozen benign lymph node samples, 27 lymph nodes with metastatic tumors, and 20 other infl ammatory or neoplastic tissues were tested in parallel ( Table 4)  ples and 3 were from 29 (10.3%) normal lymph nodes. All other samples tested were negative for MCPyV (Table 4).
Sequence analysis confi rmed the qPCR results. Conventional PCR with MCPyV-specifi c primers was performed on 9 viral DNA-positive samples (4 lymphoma, 5 reactive hyperplasia). DNA amplifi cation products of the expected size (126 bp) from the large T-ag gene were obtained, and sequence analysis confi rmed the identity of MCPyV. Three of 4 MCPyV-positive lymphoma samples shared 100% sequence homology with strain MCC344 (GenBank accession no. FJ173807), whereas sequences of the remaining sample were identical to those of MCC349 (GenBank accession no. FJ173813). Among the 5 MCPyVpositive reactive hyperplasia samples, 4 had 100% homology to strain MCC344 and 1 to strain MCC339 (GenBank accession no. EU375804).

Expression of MCPyV T-antigen
A series of 17 lymphoid specimens (7 of which were positive for MCPyV DNA by PCR), consisting of 11 lymphomas and 6 reactive hyperplasia samples, were tested for expression of MCPyV T-ag. Immunohistochemical staining with the CM2B4 monoclonal antibody was carried out without knowledge of the PCR results. One sample, classifi ed as an angioimmunoblastic T-cell lymphoma, expressed detectable T-ag in scattered lymphocytes (Figure). This sample was positive for MCPyV DNA by PCR. The remaining samples tested were negative for T-ag expression by immunohistochemical test.

Clinical Follow-Up
Five-year clinical follow-up information was available for 114 of 196 patients who had malignant lymphoma. Of those patients not included in follow-up analysis, 14 had died from other causes, 47 had received a diagnosis of lymphoma within the past 5 years, and 21 were lost to follow-up. Among the 114 patients with a 5-year followup, 49 had died of the disease and 65 were alive (24 with lymphoma and 41 without). Analyses showed no difference in the distributions among survival categories relative to MCPyV status (Table 5). Of the MCPyV-positive patients for whom follow-up data were available, 3 (42.9%) of 7 were alive and in remission compared with 38 (35.5%) of 107 MCPyV-negative patients. There were 3 (42.9%) deaths among the MCPyV-positive group and 46 (43.0%) deaths among the MCPyV-negative group (p = 0.88).    ‡Five-year follow-up survival information. Data were available for 13 patients, excluding 10 who had received a diagnosis within the past 5 y and one who was lost to follow-up. Disease refers to original diagnosis of either lymphoma or leukemia.

Discussion
This study describes the presence of MCPyV DNA in benign lymph nodes and malignant lymphomas in specimens from patients living in Canada. MCPyV was detected in 6.6% of lymphomas and in 10% of nonneoplastic lymph node samples. These results, together with those of Shuda et al. (13) and Mertz et al. (14), support the hypothesis that lymphocytes and monocytes may serve as a tissue reservoir for MCPyV infection. Because serologic assays have indicated that MCPyV primary infections frequently occur in children (24)(25)(26), we favor the interpretation that the MCPyV genomes detected in the adult lymphoid tissues refl ect the presence of persistently infected cells. Only 1 specimen among MCPyV DNA-positive samples tested expressed T-ag, which suggests that most infected lymphoid cells are not producing detectable levels of viral protein. However, because MCPyV DNA copy numbers in the samples were low, a few antigen-expressing cells in the tissues may have escaped detection in the immunohistochemical assays.
The data from this study do not suggest that MCPyV caused the lymphoid tumors that were virus positive. However, more comprehensive studies are necessary to exclude the possibility that MCPyV may have lymphomagenic potential under certain conditions. An observation of interest was the presence of MCPyV in 5 (20.8%) of 24 CLL/SLL cases. CLL is a type of leukemia that is now regarded as being identical to SLL (27). The most recent World Health Organization classifi cation scheme for hematopoietic malignancies considers CLL and SLL to be different manifestations of the same disease and combines these entities into 1 disease category (CLL/SLL) (21). Some studies have found that CLL co-exists with MCC, making the association rare but well recognized (28)(29)(30)(31)(32)(33)(34). CLL and MCC are age-related with an increased risk in those >60 years of age. Koljonen et al. (35) recently showed that MCPyV DNA is frequently present in MCCs that occur in CLL patients. The basis for the association between CLL and MCPyV is unclear. The link may be coincidental or may refl ect some infl uence of the MCPyV-infection process.
The present study provides evidence of the presence of MCPyV in samples of reactive lymphoid hyperplasia. (Reactive lymphoid hyperplasia refers to a benign, reversible enlargement of the lymph node as a consequence of proliferation of some or all of its cellular components.) This is a normal response of the lymph nodes to an antigenic stimulus, such as infection or infl ammation. Viruses, e.g., Epstein-Barr virus, induce reactivity of lymphoid cells in lymphoid tissues from healthy persons (36). In our study, 8 (13.1%) of 61 reactive hyperplasia specimens were shown to harbor MCPyV DNA at low copy number. Whether MCPyV infection prompted those cases of reactive hyperplasia is unknown.
In conclusion, our fi ndings of the presence of MCPyV in malignant lymphomas, reactive hyperplasia, and normal lymph nodes support the hypothesis that MCPyV is lymphotropic. Our fi ndings also suggest that the lymphoid system plays a role in MCPyV infection and may be a site for MCPyV persistence.