Human Bocavirus Infection of Permanent Cells Differentiated to Air-Liquid Interface Cultures Activates Transcription of Pathways Involved in Tumorigenesis

The parvoviral human bocavirus (HBoV) is a respiratory pathogen, able to persist in infected cells. The viral DNA has been identified in colorectal and lung tumors and thus it was postulated that the virus could be associated with tumorigenesis. This assumption was supported by the fact that in HBoV-infected patients and in an in vitro cell culture system, pro-cancerogenic and -fibrotic cytokines were expressed. In this work, it is shown by a whole transcriptome analysis that, also at the mRNA level, several pathways leading to neoplasia and tumorigenesis are significantly upregulated. In total, a set of 54 transcripts are specifically regulated by HBoV, of which the majority affects canonical pathways that may lead to tumor development if they become deregulated. Moreover, pathways leading to necrosis, apoptosis and cell death are downregulated, supporting the hypothesis that HBoV might contribute to the development of some kinds of cancer.


Introduction
Cancer is one of the leading causes of death worldwide. Meanwhile, several risk factors, such as environmental pollution, unhealthy lifestyle, inherited gene mutations, and chronic infections, have been identified, but for many cancer entities, the underlying processes of tumor development are not yet sufficiently understood. In this context, bocaviral DNA was detected in about 20% of colorectal and lung cancers [1], a finding that was meanwhile confirmed several times [2][3][4]. The human bocavirus (HBoV) is the second human pathogen parvovirus, which is related to the bovine parvovirus and the minute virus of canine (MVC). It is mainly associated with respiratory infections in all age groups (subtype 1) (reviewed by [5,6]), but may also display gastrointestinal symptoms (subtypes 2-4) (reviewed by [5]). After the identification of a covalently closed circular DNA [7] that most likely represents the form of a persisting genome, we and others have shown that HBoV is able to persist and induce long lasting infections in adults that result in a subclinical asymptomatic course [8][9][10][11] or that may go ahead with chronic symptoms like cough [12]. These data were confirmed by several findings from other groups [13][14][15] and led to the hypothesis that the virus may take a related pathophysiological route like the human hepatitis B virus by inducing fibrosis, followed by cancerogenesis in its target organs [16]. pathophysiological route like the human hepatitis B virus by inducing fibrosis, followed by cancerogenesis in its target organs [16].
Possible pathological mechanisms could be, on the one hand, a direct involvement in tumorigenesis by triggering cancerogenic pathways through hit-and-run mechanisms or, on the other hand, an indirect influence due to a chronic subclinical infection. Based on this assumption, a previous analysis of cytokine profiles in bronchoalveolar lavages (BALs) of HBoV-infected patients showed the expression of profibrotic and procancerogenic cytokines and chemokines, a result which was consistent with observations in CuFi-8 cell culture experiments [17]. Due to these findings, we identified pathways deregulated by HBoV infection, which are associated with the development of fibrosis and cancerogenesis. Therefore, we used the established CuFi-8 cell culture model [17][18][19][20][21] and analyzed transcription alterations, specifically induced by the HBoV infection, by whole transcriptome analyses with next-generation sequencing.

Results
The aim of the study was to identify signaling pathways that are triggered by the HBoV infection. Therefore, CuFi-8 cells were infected with HBoV or mock transfected as controls. Cells were harvested and subjected to RNA extraction five days after infection. The RNA was used for a whole transcriptome analysis based on next-generation sequencing analyses. A total number of 13,851 transcripts were identified, of which 208 were statistically significant regulated in HBoV-infected CuFi-8 cells (False Discovery rate FDR ≤ 0.05), compared to the mock infections ( Figure 1). HBoV-infected CuFi-8 cells. A total number of 208 genes out of 13,851 genes were regulated with statistically approved significance (False Discovery rate FDR ≤ 0.05). Of these 54 genes were HBoV-specific regulated. These 54 genes have been identified by comparison of mock-infected CuFi-1 and -5 transcriptomes with the mock-infected CuFi-8 transcriptome. Therefore, transcripts of HBoV-infected CuFi-8 cells, of which the RNA level in mockinfected CuFi-8 differs from mock-infected CuFi-1 and CuFi-5 cells, were excluded from the set of 208 transcripts. Of the 54 genes, 22 were upregulated, while 32 were downregulated. The experiments were separately performed in triplicate, based on three independent infections per cell line. HBoV-infected CuFi-8 cells. A total number of 208 genes out of 13,851 genes were regulated with statistically approved significance (False Discovery rate FDR ≤ 0.05). Of these 54 genes were HBoV-specific regulated. These 54 genes have been identified by comparison of mock-infected CuFi-1 and -5 transcriptomes with the mock-infected CuFi-8 transcriptome. Therefore, transcripts of HBoV-infected CuFi-8 cells, of which the RNA level in mock-infected CuFi-8 differs from mock-infected CuFi-1 and CuFi-5 cells, were excluded from the set of 208 transcripts. Of the 54 genes, 22 were upregulated, while 32 were downregulated. The experiments were separately performed in triplicate, based on three independent infections per cell line.  (1), apoptosis (2), cell death in general (3) and of pancreatic cancer cells (4) and tumor cell lines (5) in particular, as well as necrosis (6). (b) Influence of the HBoV-specific genes on the organization of connective tissues, including growth (1), proliferation (2) amongst others of fibroblasts (3), and the quantity of cells (4). (c) Prediction legend. Orange indicates an upregulation, and blue represents a downregulation of the respective pathway. Grey indicates that no pathway alterations based on single transcripts could be predicted. Brighter color indicates a weaker alteration, whereas darker color indicates a strong regulation.
In order to exclude any cell type-or host-specific effects, RNA was also isolated from mockinfected CuFi-1 and CuFi-5 cells. These cells are highly similar to CuFi-8 cells, but originate from different donors and do not productively support the replication of HBoV. All CuFi cells were immortalized by dual retroviral infection with HPV-16E6/E7-LXSN and hTERT-LXSN. In order to exclude general effects by this procedure, we decided to subtract the background and to focus on mechanisms that are not donor-specific. In CuFi-1 and CuFi-5, 1601 out of 14,861 transcripts were regulated with statistically approved significance compared to HBoV-negative CuFi-8 cells, of which 800 were downregulated and 801 were upregulated. Seven transcripts were only identified in CuFi-8 cells: membrane protein hyaluronidase 4 (HYAL4), LINC00920 (non-coding RNA), LOC102723568 and LOC102723342 (two RNAs of unknown function), the transcription factor POU2F2, the blood group antigen RHCE, and the intracellular protein TLDC2, which is related to the nuclear receptor family.
After these initial analyses, the transcriptomes of CuFi-1 and CuFi-5 were compared with the transcriptomes of mock-and HBoV-infected CuFi-8 cells. Out of the 208 transcripts regulated during HBoV infection in CuFi-8 cells, only 54 genes were equally expressed in all three non-infected CuFi cell lines ( Figure 1). Consequently, the expression of this remaining set of 54 genes (Table 1) has to be regulated specifically by the HBoV infection in CuFi-8 cells. In order to exclude any cell type-or host-specific effects, RNA was also isolated from mock-infected CuFi-1 and CuFi-5 cells. These cells are highly similar to CuFi-8 cells, but originate from different donors and do not productively support the replication of HBoV. All CuFi cells were immortalized by dual retroviral infection with HPV-16E6/E7-LXSN and hTERT-LXSN. In order to exclude general effects by this procedure, we decided to subtract the background and to focus on mechanisms that are not donor-specific. In CuFi-1 and CuFi-5, 1601 out of 14,861 transcripts were regulated with statistically approved significance compared to HBoV-negative CuFi-8 cells, of which 800 were downregulated and 801 were upregulated. Seven transcripts were only identified in CuFi-8 cells: membrane protein hyaluronidase 4 (HYAL4), LINC00920 (non-coding RNA), LOC102723568 and LOC102723342 (two RNAs of unknown function), the transcription factor POU2F2, the blood group antigen RHCE, and the intracellular protein TLDC2, which is related to the nuclear receptor family.
After these initial analyses, the transcriptomes of CuFi-1 and CuFi-5 were compared with the transcriptomes of mock-and HBoV-infected CuFi-8 cells. Out of the 208 transcripts regulated during HBoV infection in CuFi-8 cells, only 54 genes were equally expressed in all three non-infected CuFi cell lines ( Figure 1). Consequently, the expression of this remaining set of 54 genes (Table 1) has to be regulated specifically by the HBoV infection in CuFi-8 cells. In order to identify any regulated intracellular networks influenced by the HBoV infection, an extended bioinformatics analysis was performed using the Ingenuity Pathway Analysis (IPA) platform, an established internet-based analysis tool for the analysis, interpretation and mechanistic prediction of NGS-datasets [22].
The overall analysis of regulated transcripts reveals that the major regulated disease networks predicted by the IPA software include upregulation in these pathways that may push the HBoV-infected cell towards fibrosis and precancerous conditions (Figure 2b), while apoptotic and necrotic pathways are downregulated or inhibited (Figure 2a). The pathways, leading to increased phosphorylation of proteins, are also upregulated, as well as pathways leading to cell proliferation. For example, the increased expression of FN1, as well as the decreased expression of FOXO1, on the one hand, leads to a downregulation of apoptotic mechanisms, and on the other hand, to an increased proliferation ( Figure 2).
Besides, FN1, other transcripts like Tenascin C (TNC), and THBS1 involved in the extracellular matrix (ECM) metabolism have been identified as HBoV specifically regulated. In this context, we also analyzed immunohistochemically the expression of CEA in HBoV-positive tumors and cell cultures compared to HBoV-negative samples and observed that the CEA staining was intensive in HBoV-infected CuFi-8 cells and HBoV-positive lung tumor biopsies, whereas mock-infected CuFi-8 cultures as well as HBoV-negative lung tumors are CEA-negative ( Figure 3). In order to identify any regulated intracellular networks influenced by the HBoV infection, an extended bioinformatics analysis was performed using the Ingenuity Pathway Analysis (IPA) platform, an established internet-based analysis tool for the analysis, interpretation and mechanistic prediction of NGS-datasets [22].
The overall analysis of regulated transcripts reveals that the major regulated disease networks predicted by the IPA software include upregulation in these pathways that may push the HBoVinfected cell towards fibrosis and precancerous conditions (Figure 2b), while apoptotic and necrotic pathways are downregulated or inhibited (Figure 2a). The pathways, leading to increased phosphorylation of proteins, are also upregulated, as well as pathways leading to cell proliferation. For example, the increased expression of FN1, as well as the decreased expression of FOXO1, on the one hand, leads to a downregulation of apoptotic mechanisms, and on the other hand, to an increased proliferation ( Figure 2).
Besides, FN1, other transcripts like Tenascin C (TNC), and THBS1 involved in the extracellular matrix (ECM) metabolism have been identified as HBoV specifically regulated. In this context, we also analyzed immunohistochemically the expression of CEA in HBoV-positive tumors and cell cultures compared to HBoV-negative samples and observed that the CEA staining was intensive in HBoV-infected CuFi-8 cells and HBoV-positive lung tumor biopsies, whereas mock-infected CuFi-8 cultures as well as HBoV-negative lung tumors are CEA-negative ( Figure 3).   In order to identify molecular pathways, which may affect the progression to subsequent diseases due to the HBoV infection, we analyzed to what extend canonical pathways are affected by HBoV-dependent alterations of the expression profile. Accordingly, we refer to these disease pattern with a statistically significant p-value (p ≤ 0,01) predicted by the IPA core analyses that include 9 to 43 regulated transcripts, respectively, out of the 54 identified transcripts specifically regulated by the HBoV infection (Table 2). Table 2. Disease patterns predicted by IPA core analysis. Disease patterns with a statistically significant p-value (p ≤ 0.01) were taken into consideration. The analysis revealed that 40 out of the 54 identified transcripts, specifically regulated by the HBoV infection, are known to contribute to gastrointestinal cancer, whereas only 9 transcripts are associated with lung cancer. Numerals correspond to the ones in Table 1 In addition to the proposed pathways by IPA, we independently elaborated connections and intersections of the 54 identified target molecules based on published findings. These analyses reveal that, aside from the ECM, the majority of target molecules could be allocated to DNA damage response, transcriptional/translational regulation, integrin signaling, cell cycle control (Figure 4), and calcium signaling ( Table 1). As the production and exocytosis of mucins is dependent on the Ca 2+ concentration, we evaluated the mucin production in HBoV-positive and HBoV-negative cell cultures. Compared to CuFi-1 and CuFi-5 cells, CuFi-8 showed a higher production of acid mucins in general, accompanied by an increased expression of mucins after HBoV infection in CuFi-8 cells (Figure 3b).  Moreover, scanning electron microscopy (SEM) indeed revealed significant changes of the cell surface regarding the cilia and the appropriate glycocalix ( Figure 5). Moreover, scanning electron microscopy (SEM) indeed revealed significant changes of the cell surface regarding the cilia and the appropriate glycocalix ( Figure 5).

Discussion
Aim of this study was to identify signaling pathways that are changed by the infection of HBoV1 and that may contribute to the development of solid tumors, such as non-small cell lung cancer and colorectal cancer. These cancer entities have been associated with the occurrence of HBoV1 DNA as observed earlier [2][3][4]. Moreover, our group demonstrated that profibrotic and procancerogenic cytokines/chemokines are upregulated during HBoV infection in vitro and in vivo, thus suggesting a causative role of HBoV in tumorigenesis [17]. These findings support our preliminary model, in which we have postulated that HBoV induces tumorigenesis analogous to the hepatitis B virus [16].
In the present study, we used a whole transcriptome analysis to get more detailed information of HBoV-dependent cellular changes. As expected, the HBoV affects both transcription and translation by regulating activators and inhibitors of both processes. Moreover, our results confirm the observation of Deng et al. that replication of HBoV1 may be dependent on the cellular DNA damage repair system (DDR) [19,20]. The group showed that the kinases Ataxia telangiectasia mutated (ATM), ATR (ATM-and RAD3-related) as well as DNA-PKC (DNA-dependent protein kinase catalytic subunit) were activated in HBoV-infected cells, which in turn is crucial for genome amplification of HBoV1, but the exact mechanism remained unknown. Our transcriptome analyses revealed that some target proteins in HBoV-infected cells are associated with ATR. In this context, the kinase activity of ATR is necessary for the nuclear accumulation of NCK, an adaptor protein of which depletion promotes apoptosis [23] and which is a target of SOCS7. SOCS7 is responsible for the nuclear transport of NCK and also leads to an apoptotic phenotype if depleted. The fact that the septin-SOCS7-NCK axis intersects with the canonical DNA damage cascade downstream of ATM/ATR

Discussion
Aim of this study was to identify signaling pathways that are changed by the infection of HBoV1 and that may contribute to the development of solid tumors, such as non-small cell lung cancer and colorectal cancer. These cancer entities have been associated with the occurrence of HBoV1 DNA as observed earlier [2][3][4]. Moreover, our group demonstrated that profibrotic and procancerogenic cytokines/chemokines are upregulated during HBoV infection in vitro and in vivo, thus suggesting a causative role of HBoV in tumorigenesis [17]. These findings support our preliminary model, in which we have postulated that HBoV induces tumorigenesis analogous to the hepatitis B virus [16].
In the present study, we used a whole transcriptome analysis to get more detailed information of HBoV-dependent cellular changes. As expected, the HBoV affects both transcription and translation by regulating activators and inhibitors of both processes. Moreover, our results confirm the observation of Deng et al. that replication of HBoV1 may be dependent on the cellular DNA damage repair system (DDR) [19,20]. The group showed that the kinases Ataxia telangiectasia mutated (ATM), ATR (ATM-and RAD3-related) as well as DNA-PKC (DNA-dependent protein kinase catalytic subunit) were activated in HBoV-infected cells, which in turn is crucial for genome amplification of HBoV1, but the exact mechanism remained unknown. Our transcriptome analyses revealed that some target proteins in HBoV-infected cells are associated with ATR. In this context, the kinase activity of ATR is necessary for the nuclear accumulation of NCK, an adaptor protein of which depletion promotes apoptosis [23] and which is a target of SOCS7. SOCS7 is responsible for the nuclear transport of NCK and also leads to an apoptotic phenotype if depleted. The fact that the septin-SOCS7-NCK axis intersects with the canonical DNA damage cascade downstream of ATM/ATR has been already known since 2007 [24] and we observed, in our analysis, elevated RNA levels of SOCS7, which might prevent apoptosis. The protein AJUBA is part of the DDR repression, prevents apoptosis, and controls the switch between activation of ATR during S phase and the general ATR activation after extensive DNA damage [25]. As the amount of AJUBA RNA was also upregulated during HBoV infection; this may further prevent HBoV-positive cells from apoptosis. These findings are compatible with the fact that HBoV does not promote apoptosis, as shown by the group from Jianming Qiu [26] and, in concert with the independent confirmation of the involvement of the DDR response in HBoV DNA replication, support the significance of the NGS transcriptome analyses presented in this study. Another factor, which is involved in the DNA damage response, is AKAP12. AKAP12 becomes phosphorylated by ATR and in turn is necessary for scaffolding the PKA-mediated ATR phosphorylation at S435 [27]. This potential tumor suppressor, which is downregulated in HBoV-infected CuFi-8 cells, was shown to be also reduced on mRNA levels in human lung cancers [28,29].
Additionally, in the presented study, deregulated RNA levels of Hexim1 and NEAT1 were detected.
These are components of the HDP-RNP (HEXIM1-DNA-PK-paraspeckle components-ribonucleoprotein) complex that regulates DNA-mediated innate immune response [30]. The HDP-RNP complex also contains NONO/SFPQ/PSPC1, which dissociate upon stimulation with ISD (interferon stimulatory DNA) and then are available for the DDR [31,32]. Moreover, the lncRNA NEAT1 as an essential component of paraspeckle formation seems to be dependent on the degree of cell differentiation [33]. While human embryonic stem cells (hESCs) show neither any paraspeckles nor any NEAT1 expression, which is recovered after differentiation of the cells, one can speculate that HBoV by downregulation of NEAT1 might restore the state of hESCs regarding regulation of some genes.
Many of the genes deregulated by HBoV are known drivers in the development of fibrosis and tumorigenesis. Among them, we identified, besides others, transcripts encoding proteins of the ECM as upregulated.
On the one hand, the ECM displays a physical network, which is necessary for mechanical stability of tissues; on the other hand, it represents a dynamic entity, which transmits signals between the intra-and extracellular space. This means that the ECM not only influences tissue morphogenesis, homeostasis and remodeling, but also has influence on gene expression and consequently to proliferation, differentiation, and apoptosis. As HBoV regulated proteins involved in the assembly of the ECM, we identified THBS1, FN1, LAMC2 and TNC. All of them are known to influence cell adhesion, motility, and growth. TNC is a glycoprotein of the ECM that is regulated by the tissue microenvironment [34], of which increased expression is associated with different kinds of chronic airway inflammation, such as inflamed bronchi of smokers, interstitial pneumonia and bronchial asthma [35,36].
The fact that TNC is upregulated in HBoV-positive CuFi-8 cells further supports the hypothesis that HBoV may be involved in the development of fibrosis and tumors, as it was shown that TNC increases pro-collagen synthesis and activates a fibrotic response [37]. Moreover, TNC expression appears to be linked to lung metastasis in breast cancer [38] and leads to a downregulation of tropomyosin-1 and the Wnt inhibitor Dickkopf 1, what in turn results in the destabilization of actin stress fibers and an enhanced Wnt signaling [39]. Recent studies have shown that Wnt signaling, together with other signaling cascades, is coordinated at cilia. The fact that these pathways are often deregulated during malignant development, together with the frequent disappearance of cilia in transformed cells, suggests the possibility that deficient ciliary signaling may promote cancer [40]. SEM revealed defective cilia also in HBoV-infected CuFi-8 cells ( Figure 5). Taking into account that NEDD9, which is deregulated in this study, participates in the formation and disassembly of cilia and interacts with AurA at the centrosome, which is necessary for cellular progression through mitosis [41,42], it seems to be possible that HBoV may indirectly lead to cellular transformation due to reprogrammed cell-matrix signaling. The assumption that HBoV triggers tumor development by inducing fibrotic alterations of the connective tissue is supported by network predictions from the IPA core analyses, which show that target proteins are involved in connective tissue organization (Figure 2b). Moreover, the translational regulator LARP6, which specifically regulates type I collagen expression [43], is upregulated during HBoV infection. Stabilization of collagen mRNAs and promotion of their translation lead to an increased collagen expression, which is observed in fibrotic processes.
Finally, we observed that the HBoV infection is accompanied by an elevated CEA expression. It is already known that HRT-cells, which are known to support the replication of the canine parvovirus, are CEA-positive, and both the colorectal carcinoma and non small cell lung cancer (NSCLC) frequently go ahead with an increased CEA production [44][45][46]. Moreover, an increased CEA expression was also observed in idiopathic lung fibrosis, a clinical course that could occur in HBoV-positive patients [47].
One of the markers that were upregulated with the highest statistical significance is FN1. In colorectal cancers, in which HBoV has been detected in about 20%, fibronectin expression is frequently upregulated and goes ahead with an upregulation of the CEA, which in turn is used as a diagnostic marker [48]. As CEA staining is a common method in pathology for the immunohistochemical characterization of tumors, we applied this method to cell cultures infected with HBoV (Figure 3a). In fact, we observed an increase of CEA expression associated with HBoV. In addition, increased CEA expression may occur in concert with increased expression of mucins [49]. Thereby, mucin expression is regulated by the Wnt/β-catenin pathway [50], the latter being significantly disturbed by the HBoV infection, as shown in this study. Thus, it was obvious to analyze to which extent the HBoV infection also influences the mucin expression. Besides, an increase in Mucin15 RNA (ranking at place 320 of genes with altered expression within the whole transcriptome dataset of HBoV-positive CuFi-8 vs. HBoV-negative CuFi-8) was observed; the Alcian blue staining reveals an increase in mucin expression, as shown in Figure 3b, in turn supporting the hypothesis that HBoV contributes to a changed ECM, leading to a changed cell signaling.
The hypothesis that the CEA expression, which was also increased in our study, is associated with the HBoV infection is supported by the observation that non-HBoV-induced serious malformations in their natural hosts, as they preferably replicate in tissue with high-grade expression of CEA [45,46]. Further studies should, therefore, focus on the protein expression of the altered pathways identified in our study and should include a more differentiated kinetic analyses, which should include immediate early, early, and long-term transcription profiles to get a closer view on the putative HBoV-triggered cancerogenic mechanisms.

Cells and Infections
CuFi-8 cells were grown as air-liquid interface cultures and were infected as previously described [17,18]. Infections were performed in triplicate with a recombinant HBoV wildtype generated by transfecting the plasmid pIHBoV into HEK-293 cells and subsequent harvesting [18]. Mock infections were performed by inoculating the differentiated CuFi-8 cells with exhausted HEK-293 cell culture supernatants from mock transfections (i.e., transfection with transfection reagents but w/o plasmid) that were processed exactly as the transfection media from the recombinant HBoV production [18]. Ahead of infecting CuFi-8 cells, infectious HBoV particles were harvested by collecting the medium 5 days after transfection with pIHBoV and removal of cell debris by centrifugation at 2000× g. This supernatant was subject to DNAse digest and subsequent qPCR for quantification of packaged genome copies. Infections were performed with MOI1. At day five, the highest HBoV replication rat was reached [18], indicated by secretion of viral particles to the basal cell culture medium and the apical surface at this time point [18]. As further mock-infected controls, CuFi-1 and CuFi-5 cells (ATCC ® -CRL-4013 TM and ATCC ® -CRL-4016 TM ) were used for excluding transcripts that were regulated unspecifically.

RNA Extractions and Whole Transcriptome Analyses
RNA was extracted from HBoV-infected and mock-infected CuFi-8 cells, and from CuFi-1 and -5 cells by using the RNeasy Total RNA kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol five days after infection. Whole transcriptome analyses were performed by a sequencing services vendor (MWG Eurofins, Ebersberg, Germany), according to the vendor's protocol and strictly following the vendors' recommendation regarding RNA quantity and quality. The approach was a HiSeq 2500 Illumina sequencing, sequencing mode 1 × 125, software HiSeq Control Software 2.2.58 (MWG Eurofins, Ebersberg, Germany), RTA 1.18.64, bclfastq-1.84 using the HiSeq SBS Kit v4. The Q30 percentage for all libraries was >96%, mapping mismatches were ≤0.5%, the read coverage mean was between 17 and 27 for all libraries, and the normalized library sizes ranged between 22,174,885.9 and 35,458,600.4. The false discovery rates of genes that were statistically significant regulated was ≤0.05 in all cases.
The resulting files reporting the transcriptional fold-change in CuFi-8 cells were analyzed using the IPA (Qiagen, Hilden, Germany) [22].
The IPA procedure was performed using the standard parameters recommended for core analysis which is able to identify clinical conditions, biological pathways and network interactions that display an altered regulation during the HBoV infection. Only these pathways and clinical entities identified by the core analyses were included, in which a minimum of 9 genes out of 54 specific genes (see below) displayed a statistically significantly altered regulation. The nomenclature was completely taken from the IPA platform.

Immunohistochemical Analyses of CuFi air-Liquid-Interface Cultures
Differentiated CuFi-ALIs were formalin-fixed paraffin-embedded according to standard procedures. Subsequently, FFPE sections were deparaffinised, hydrated and stained with the appropriate method. In case of the ECM, acid mucins were stained with the standard Alcian blue (pH 2.5) method followed by PAS. For CEA staining, the EnVision TM FLEX+ detection system (Dako, Denmark) was used according to the manufacturer's recommendations. In brief, cellular peroxidase was blocked before incubation with the primary antibody. The antibody clone CEA II-7 from Dako (Dako, Glostrup, Denmark) was used in a dilution of 1:150 (v/v). The PT-link pretreatment was performed for 20 min at 97 • C with Envision Flex Target at low pH (nitrate buffer at pH 6.1). For visualization, the respective HRP-conjugated secondary antibody and DAB chromogen, which are also included in the EnVision TM FLEX+ kit, were utilized for 20 min. Mucins were visualized with the standard Alcian blue staining.

Scanning Electron Microscopy
SEM was used to analyze changes of the cellular surface of CuFi-8 ALIs after HBoV infection. The HAE cultures were fixed with 0.1 M cacodylate buffer containing 2.5% glutaraldehyd, 2% polyvinylpyrrolidone and 75 mM NaNO 2 for 12 h at 4 • C. Samples were washed in 0.1 M cacodylate buffer without glutaraldehyd and subsequently incubated in a solution containing 2% arginine-HCl, glycine, sucrose and sodium glutamate for 16 h at RT. The tissue cultures were rinsed in distilled water, followed by immersion in a mixture of each 2% tannic acid and guanidine-HCl for 8 h at RT. The samples were rinsed again in distilled water and incubated in a 2% OsO 4 solution for 8 h at RT. After three rinsing steps with distilled water the HAE cultures were dehydrated, dried in liquid CO 2 and finally examined with a Zeiss Sigma SEM (Zeiss, Oberkochen, Germany) using 2-5 kV acceleration voltages after sputtering with gold palladium.

Conclusions
In summary, although we must concede that not all HBoV1 patients develop cancer and not all tumors are positive for HBoV, there is an increased likelihood that the virus is involved in