Human Polyomavirus-6 Infecting Lymph Nodes of a Patient With an Angiolymphoid Hyperplasia With Eosinophilia or Kimura Disease

Abstract

Polyomaviruses have been known to exist since the 1970s and are found in a wide variety of hosts including mammals and birds. However, only 2 human polyomaviruses (HPyVs) were known to exist before 2007 when, with new sequencing technologies, new HPyVs started to be discovered [1]. To date, the following 13 HPyVs have been described: BK polyomavirus (BKPyV), JC polyomavirus (JCPyV), Merkel cell polyomavirus (MCPyV), trichodysplasia spinulosa-associated polyomavirus (TSPyV), WU polyomavirus (WUPyV), KI polyomavirus (KIPyV), MW polyomavirus (MWPyV), STL polyomavirus (STLPyV), human polyomavirus 6 (HPyV6), human polyomavirus 7 (HPyV7), human polyomavirus 9 (HPyV9), human polyomavirus 12 (HPyV12), and New Jersey polyomavirus (NJPyV). Of these viruses, a reliable link to disease has been observed in only 4, causing Merkel cell carcinoma (MCPyV), progressive multifocal leukoencephalopathy (JCPyV), trichodysplasia spinulosa (TSPyV), and nephropathy (BKPyV) [2–5].

HPyVs have been isolated from the nasopharynx and lung, skin, and the gastrointestinal tract [6]. Some studies have also shown the presence of HPyV's DNA in tonsillar tissues [7, 8]. HPyVs are frequently found in immunocompromised patients [1, 6, 9, 10]. Such is the case of HPyV6, which was first described as a natural inhabitant of healthy human skin but was later detected by polymerase chain reaction (PCR) at other body sites such as the cerebrospinal fluid of a human immunodeficiency virus (HIV)–infected patient [10], at low prevalence in tonsillar tissue (5% of cases) [8], and in nasopharyngeal aspirates (1.7% of cases) [11]. HPyV6 was also found at high prevalence in epithelial proliferations induced by BRAF inhibitor therapy [12].

Here, we report the first case of HPyV6 lymph node infection in a patient who presented with angiolymphoid hyperplasia with eosinophilia in a form that could also be classified as Kimura disease due to an overlap between 2 clinical syndromes. Written informed consent was obtained from the patient for publication.

A 51-year-old woman was first referred to the Dermatology Department of the Timone Hospital (Marseille, France) in January 2014 for a history of multiple angiomatous lesions on her scalp (Figure 1A) together with palpable cervical nodes, which had developed in the last 5 years. Several diagnoses were considered including lymphoma and sarcoidosis, which were not confirmed by 3 successive biopsies, all pointing to “inflammatory angiomas.”

Figure 1.

Detection of human polyomavirus 6 (HPyV6) in lymph node. A, Photographs showing the multiple angiomatous lesions on the patient's scalp. B, Histopathological analysis of cervical lymph node. The lymph node architecture is preserved with florid germinal centers and hyperplasia of T-dependent areas (left). At high magnification of a T-dependent area (right), a marked eosinophilic infiltrate with hyperplasia of postcapillary venules (hematoxylin-eosin-saffron, original magnification ×25 and ×100, respectively) is noted. C, Classification of metagenomic reads (Illumina shotgun sequencing) from viral-like particles purified by cesium chloride gradient. Sequences were analyzed using Basic Local Alignment Search Tool for nucleotides (expect value < 0.00001) against the National Center for Biotechnology Information (NCBI) database, and then lower common ancestors were estimated using MEtaGenome ANalyzer. The abundances of all taxa found are expressed as percentage of total sequences after cleaning human contamination (remaining human sequences after cleaning are shown). D, Neighbor-joining phylogenies were constructed with representative genomes of all known HPyVs (left) and also with all HPyV6 available genomes from other studies (right) using clustal-aligned sequences. NCBI accession numbers (GenInfo Identifiers [GIs]) are indicated on the left of each name at the tips of the trees and can be consulted for further information on each sequence source. Bootstrapping was conducted with 1000 repetitions and node support of relevant branches is indicated. E, Polymerase chain reaction (PCR) confirmation of HPyV6 in lymph node. The PCR for HPyV6 (product of 99 bp) from a blood sample processed parallel to lymph node, with the same procedures and using the same reagents, was negative (1) and positive for the lymph node extracted DNA (2). PCR blank control was also negative (3). DNA ladder (M) corresponded to the 100 bp DNA ladder from Promega. F, Fluorescence in situ hybridization (FISH) on thin sections of the lymph node using a DNA probe of 458 bp (in positions: 4469 to 4927 of LN-HPyV6 genome) showing the presence of the HPyV6 DNA (green signal, central panel) inside human cells colocalizing with nuclei. Total nucleic acids were stained with 4′,6-diamidino-2-phenylindole (DAPI; blue signal, left panel). The merged image of DAPI staining and HPyV6 FISH is shown in the right panel. Abbreviations: BK-PyV, BK polyomavirus; FITC, fluorescein isothiocyanate; HPyV7, human polyomavirus 7; HPyV9, human polyomavirus 9; HPyV12, human polyomavirus 12; JC-PyV, JC polyomavirus; KI-PyV, KI polyomavirus; MC-PyV, Merkel cell polyomavirus; NJ-PyV, New Jersey polyomavirus; STL-PyV, STL polyomavirus; TS-PyV, trichodysplasia spinulosa-associated polyomavirus; WU-PyV, WU polyomavirus.

Open in new tabDownload slide

At first examination, the diagnosis of Kimura disease was proposed based on the very suggestive clinical presentation along with a new readout of the past pathological reports available, which all mentioned a lymphoid infiltrate with vascular hyperplasia. The diagnosis was confirmed by a new skin biopsy as well as by reviewing the previous skin biopsies, which showed the hallmarks of angiolymphoid hyperplasia with eosinophilia and also from Kimura disease with endothelial hyperplasia, “tombstone” endothelial cells, and lymphoid infiltrate. All biological examinations, including HIV serology, were normal. A cervical node was removed and showed a lymphoid hyperplasia with some follicles, without any vascular proliferation (Figure 1B). The patient was reassured and a symptomatic treatment of skin lesions by laser was proposed, which was of some benefit from the cosmetic point of view. The cervical node was processed immediately after surgery for research purposes, with the hypothesis that such a minimally aggressive tumoral skin disorder with endothelial and inflammatory components could be of microbial or viral origin on the model of bacillary angiomatosis or Kaposi sarcoma, respectively.

We used a shotgun metagenomic approach to determine whether microbial or viral pathogens could be present in the sample. The cervical node was processed within the next 2 hours, starting a protocol for enriching the sample in viral particles, as described in Popgeorgiev et al, and using a cesium chloride gradient and ultracentrifugation [13]. All manipulations were done in sterile conditions in a laboratory with biosafety level II. Total DNA was prepared from the enriched viral fraction and sequenced by shotgun sequencing in an Illumina MiSeq machine. We obtained 940 298 reads (deposited in the short read archive as SRP068830) and cleaned human sequences using Deconseq software [14], reducing the dataset to 32 003 reads. These sequences were then assembled into contigs using CLC software. Finally, we analyzed all cleaned single reads and contigs using BlastN against National Center for Biotechnology Information nt database. Sequence global alignments and neighbor-joining phylogenetic analyses were conducted in MEGA6, as described elsewhere [15].

Unexpectedly, we detected a high abundance (13% of clean sequences) of hits to human polyomavirus in the lymph node sample (Figure 1C). All hit sequences were assembled into a single contig of 4926 bp and classified as HPyV6, which covered the full genomes of all sequenced HPyV6 with an identity of >99%. (The genome sequence of the new lymph node-HPyV6 [LN-HPyV6] strain obtained in our study was deposited in GenBank as KU596573.)

The identity of the virus was also confirmed by phylogenetic analysis, which compared the new LN-HPyV6 genome (from our study) with representative genomes of all 13 known HPyVs (Figure 1D, left), showing a clear clustering with the reference HPyV6 genome. Further phylogenetic analysis showed that LN-HPyV6 clustered outside the 9 available HPyV6 genomes (previously sequenced by other laboratories in independent studies), indicating higher genome variability with respect to the other HPyV6s. However, this variability was represented by only 42 point mutations (across the 4926 bp genome) in comparison to the most different strain, which also had the highest number of mutations between any pair of HPyV6 genomes. Considering that all HPyV6s were obtained from different studies that analyzed different individuals from distant geographic regions (China, Australia, the United States, and now France in our report) and were processed by different laboratories, these results suggest that HPyV6 may maintain extremely low mutation rates at the population level, probably due to strong selective pressures.

A blood sample from the patient was also processed simultaneously by following the same procedures and using the same reagents from sample collection to DNA extraction. A PCR analysis using the primers developed by Schrama et al [12] was negative for the blood sample, indicating that reagents and samples were not contaminated with HPyV6 during manipulation (Figure 1E, lane 1). The presence of HPyV6 in the DNA extracted from lymph node was confirmed by PCR (Figure 1E, lane 2), indicating that there was no contamination during library preparation and sequencing. These results also indicate that viral proliferation was likely localized at the lymph nodes and did not spread into the blood circulation.

In order to confirm the presence and infectivity of HPyV6 in the biopsy, we performed a fluorescence in situ hybridization on thin sections of the lymph node using a DNA probe of 458 bp (the region covered by the probes was 4469–4927 using the primer pair HPyV6_4469Fw: TAGCACTTGTAGCACCAG and HPyV6–4907R: ATGGATCGCCTTTTAGCCAG) and the protocols described elsewhere [13]. We were able to detect clear intracellular HPyV6 DNA signals in several regions of the tissue, which colocalized with cell nuclei (Figure 1F). The low and dispersed number of infected cells suggests that HPyV6 has a low infection rate.

Here, we showed that a patient with a systemic inflammation of lymph nodes presented an evident infection with HPyV6. To date, this virus has been predominantly found in the skin of healthy individuals, and few studies have reported detection inside the body using PCR or other indirect methods [8, 10, 11, 16]. However, other polyomaviruses (JCPyV, BKPyV, WUPyV, KIPyV, MCPyV, and TSPyV) have been detected in tonsillar tissues, suggesting that lymphoid tissue may be a latency site for some HPyVs [7, 17, 18]. We discovered that HPyV6 infects internal tissues, particularly lymph nodes. Our hypothesis is that HPyV6s could be opportunistic viruses that replicate when the immune system is down. This observation raises the possibility that HPyV6s could have a significant role in the development of Kimura disease; however, further research is required to demonstrate such a link. Our work shows the potential of using high-throughput sequencing and metagenomic approaches to elucidate the putative causes of infectious diseases when diagnosis is unclear and for the discovery of new pathogens.

Notes

Financial support. This work was supported by a Management des talents fellowship from the A*MIDEX Foundation awarded to C. D.

Potential conflicts of interests. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References