Global profiling of mRNA extracted from cell lines either transfected with HPV16 or isolated from HPV-positive lesions has been valuable for the identification of differentially expressed genes in response to viral gene expression. The questions posed in different studies vary. Profiling has been conducted on primary keratinocytes mainly immortalized with HPV16 - either with E6, E7, E5 or full genomes - or on HPV16-positive cell lines originally established from biopsy specimens. One of the earliest studies showed that differential expression of several cellular transcripts varied dependent on the expression of E7 alone or of E6 plus E7. It also identified several transcripts that were regulated in an opposing manner by E7 alone and E6 plus E7, e.g. vimentin, p21, damage-specific DNA binding protein p48 subunit, Nag-1 and IGFBP6. Other transcripts were potentiated by E6/E7 expression such as Cox-1 [27 Duffy CL, Phillips SL, Klingelhutz AJ. Microarray analysis identifies differentiation-associated genes regulated by human papillomavirus type 16 E6 Virol 2003; 314(1 ): 196-205.]. In a more recent study, E7 expression modified IFI44, IFI44L and TP53INP2 - transcripts, which were also found to be downregulated in HPV16-positive HaCaT cells [28 Boccardo E, Manzini Baldi CV, Carvalho AF, et al. Expression of human papillomavirus type 16 E7 oncoprotein alters keratinocytes expression profile in response to tumor necrosis factor Carcinogen 2010; 31(3 ): 521-31., 29 Kaczkowski B, Rossing M, Andersen DK, et al. Integrative analyses reveal novel strategies in HPV11,-16 and -45 early infection Sci Rep [Internet] Nat Pub Group 2012; 2: 515. Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fc gi?dbfrom=pubmed&id= 22808421&retmode=ref&cmd=prlinks]. The annotation of all genes discussed in the manuscript is included as Supplementary Table 1. The annotation includes gene description, function, involvement in biological processes and cellular localization.

Several studies have used cells transfected with the full HPV16 genome, thus allowing analysis of cellular changes related to the gene expression of more viral proteins. As expected, large numbers of transcripts were found to be differentially regulated by the expression of more viral genes. One study focused on the identification of hTERT responsive genes following chromosome 6 fusion into HPV16-immortalized keratinocytes. This led to identification of a different set of differentially expressed genes compared to the E7- and E6/E7-mediated modifications [30 de Wilde J, Wilting SM, Meijer CJ, et al. Gene expression profiling to identify markers associated with deregulated hTERT in HPV-transformed keratinocytes and cervical cancer Int J Cancer 2008; 122(4 ): 877-8.]. The study identified 32 candidate genes differentially regulated in response to the expression of hTERT, amongst which a large number of genes involved in signal transduction and cell proliferation. Among the upregulated genes were IGF2BP1 and regulators of G-protein signalling (RGS5) and the WNT pathway (WISP2), whereas the downregulated transcripts included HCCA2 and the DNA damage repair protein XRCC1.

One study explored the effect of TNF-alpha on HPV-immortalized cells and identified KCNK6, PRSS11 and ANKRD11 as being differentially regulated in HPV16-positive cells [1 Hausen zur H. Human papillomaviruses and their possible role in squamous cell carcinomas Curr Top Microbiol Immunol 1977; 78: 1-30., 31 Termini L, Boccardo E, Esteves GH, et al. Characterization of global transcription profile of normal and HPV-immortalized keratinocytes and their response to TNF treatment BMC Med Genom 2008; 1: 29.]. The most recent study where the HPV16 genome was transfected into HaCaT cells identified 338 differentially expressed genes (fold change between -3.8 and 3.5). The most upregulated genes were ANKRD1, PSGs and HSPB8 and among the most downregulated genes were MMPs, SERPINB3 and SERPINB4 [2 Bosch FX, Burchell AN, Schiffman M, et al. Epidemiology and natural history of human papillomavirus infections and type-specific implications in cervical neoplasia Vaccine 2008; 26(Suppl 10 ): K1-16., 29 Kaczkowski B, Rossing M, Andersen DK, et al. Integrative analyses reveal novel strategies in HPV11,-16 and -45 early infection Sci Rep [Internet] Nat Pub Group 2012; 2: 515. Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fc gi?dbfrom=pubmed&id= 22808421&retmode=ref&cmd=prlinks] (Fig. 1).

Fig. (1) Graphical representation of the genes that were differentially expressed in at least two cell culture studies. In five of these studies, in vitro cell cultures were transfected with genomes of different HPV types: Thomas_HPV11 [37 Thomas JT, Oh ST, Terhune SS, Laimins LA. Cellular changes induced by low-risk human papillomavirus type 11 in keratinocytes that stably maintain viral episomes J Virol 2001; 75(16 ): 7564-1.], normal human keratinocytes transfected with HPV11; Kaczkowski_HPV11, Kaczkowski_HPV16 and Kaczkowski_HPV45 [29 Kaczkowski B, Rossing M, Andersen DK, et al. Integrative analyses reveal novel strategies in HPV11,-16 and -45 early infection Sci Rep [Internet] Nat Pub Group 2012; 2: 515. Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fc gi?dbfrom=pubmed&id= 22808421&retmode=ref&cmd=prlinks], HaCaT cells transfected with HPV11, HPV16 and HPV45, respectively; Chang_HPV31 [53 Chang YE, Laimins LA. Microarray analysis identifies interferon-inducible genes and Stat-1 as major transcriptional targets of human papillomavirus type 31 J Virol 2000; 74(9 ): 4174-82.], normal human keratinocytes from foreskin tissue transfected with wild type HPV31 genome. W12 was obtained from a low-grade cervical lesion and was naturally infected with HPV16 [34 Alazawi W, Pett M, Arch B, et al. Changes in cervical keratinocyte gene expression associated with integration of human papillomavirus 16 Cancer Res 2002; 62(23 ): 6959-5.].The upregulated genes are depicted in orange, downregulated in green, and not differentially expressed in white. Interestingly, oncogene ABL2 was found upregulated in high-risk types HPV16, HPV31 and HPV45. The annotations of the genes are included as Supplementary Table 1.

Fig. (2) Graphical representation of the differentially expressed genes from 14 comparisons reported in 12 studies. Genes differentially expressed in at least three comparisons are shown. The comparisons represent genes differentially expressed in benign and malignant growth caused by HPV infection. The signatures represent the differentially expressed genes as follows: Santegoets_Condylomata and Santegoets_VIN [51 Santegoets LAM, van Baars R, Terlou A, et al. Different DNA damage and cell cycle checkpoint control in low- and high-risk human papillomavirus infections of the vulva Int J Cancer 2012; 130(12 ): 2874-85.], condylomata and vulval intraepithelial neoplasia (VIN) versus normal tissue; Rosty_HPV16_E7_Cervix [46 Rosty C, Sheffer M, Tsafrir D, et al. Identification of a proliferation gene cluster associated with HPV E6/E7 expression level and viral DNA load in invasive cervical carcinoma Oncogene 2005; 24(47 ): 7094-104.], genes correlated to HPV16 E7 expression in cervical cancer; Santin_Cervix [45 Santin AD, Zhan F, Bignotti E, et al. Gene expression profiles of primary HPV16- and HPV18-infected early stage cervical cancers and normal cervical epithelium: identification of novel candidate molecular markers for cervical cancer diagnosis and therapy Virol 2005; 331(2 ): 269-91.] and Carlson_Cervix [38 Carlson MW, Iyer VR, Marcotte EM. Quantitative gene expression assessment identifies appropriate cell line models for individual cervical cancer pathways BMC Genom Bio Med Central Ltd 2007; 8(1 ): 117.], cervical cancer versus normal cervical keratinocytes; Schlecht_HNSC [49 Schlecht NF, Burk RD, Adrien L, et al. Gene expression profiles in HPV-infected head and neck cancer J Pathol 2007; 213(3 ): 283-93.] and Slebos_HNSC [47 Slebos RJC, Yi Y, Ely K, et al. Gene expression differences associated with human papillomavirus status in head and neck squamous cell carcinoma Clin Cancer Res 2006; 12(3 Pt 1 ): 701-9.], HPV-positive versus HPV-negative head-and-neck squamous cell carcinoma; Martinez_oropharyngeal [48 Martinez I, Wang J, Hobson KF, Ferris RL, Khan SA. Identification of differentially expressed genes in HPV-positive and HPV-negative oropharyngeal squamous cell carcinomas Eur J Cancer 2007; 43(2 ): 415-32.], HPV-positive versus HPV-negative oropharyngeal squamous cell carcinoma; Pyeon_HPV_positive_cancers [50 Pyeon D, Newton MA, Lambert PF, et al. Fundamental Differences in Cell Cycle Deregulation in Human Papillomavirus-Positive and Human Papillomavirus-Negative Head/Neck and Cervical Cancers Cancer Res 2007; 67(10 ): 4605-19.], HPV-positive head-and-neck cancer (HNC) and cervical cancer versus HPV-negative HNC; DeVoti_Papillomas [52 DeVoti JA, Rosenthal DW, Wu R, Abramson AL, STEINBERG BM, Bonagura VR. Immune dysregulation and tumor-associated gene changes in recurrent respiratory papillomatosis: a paired microarray analysis Mol Med 2008; 14(9-10 ): 608-17.], respiratory papillomas versus adjacent tissue; Thomas_HPV11 [37 Thomas JT, Oh ST, Terhune SS, Laimins LA. Cellular changes induced by low-risk human papillomavirus type 11 in keratinocytes that stably maintain viral episomes J Virol 2001; 75(16 ): 7564-1.], normal human keratinocytes transfected with HPV11; Kaczkowski_HPV11, Kaczkowski_HPV16 and Kaczkowski_HPV45 [29 Kaczkowski B, Rossing M, Andersen DK, et al. Integrative analyses reveal novel strategies in HPV11,-16 and -45 early infection Sci Rep [Internet] Nat Pub Group 2012; 2: 515. Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fc gi?dbfrom=pubmed&id= 22808421&retmode=ref&cmd=prlinks], HaCaT cells transfected with HPV11, HPV16 and HPV45, respectively; Alazawi_W12_HPV16 [34 Alazawi W, Pett M, Arch B, et al. Changes in cervical keratinocyte gene expression associated with integration of human papillomavirus 16 Cancer Res 2002; 62(23 ): 6959-5.], signature from cell line generated from a low-grade cervical lesion naturally infected with HPV16; Chang_HPV31 [38 Carlson MW, Iyer VR, Marcotte EM. Quantitative gene expression assessment identifies appropriate cell line models for individual cervical cancer pathways BMC Genom Bio Med Central Ltd 2007; 8(1 ): 117., 53 Chang YE, Laimins LA. Microarray analysis identifies interferon-inducible genes and Stat-1 as major transcriptional targets of human papillomavirus type 31 J Virol 2000; 74(9 ): 4174-82.], normal human keratinocytes from foreskin tissue transfected with wild type HPV31 genome. The upregulated genes are depicted in orange, downregulated in green, and not differentially expressed in white. The annotations of the genes are included as Supplementary Table 1.

Fig. (3) Depiction of the genes that were differentially expressed in at least two studies based on samples from genital lesions. The signatures represent the differentially expressed genes as follows: Santegoets_VIN [51 Santegoets LAM, van Baars R, Terlou A, et al. Different DNA damage and cell cycle checkpoint control in low- and high-risk human papillomavirus infections of the vulva Int J Cancer 2012; 130(12 ): 2874-85.], vulval intraepithelial neoplasia (VIN) versus normal tissue; Rosty_HPV16_E7_Cervix [46 Rosty C, Sheffer M, Tsafrir D, et al. Identification of a proliferation gene cluster associated with HPV E6/E7 expression level and viral DNA load in invasive cervical carcinoma Oncogene 2005; 24(47 ): 7094-104.], genes correlated to HPV16 E7 expression in cervical cancer; Santin_Cervix [45 Santin AD, Zhan F, Bignotti E, et al. Gene expression profiles of primary HPV16- and HPV18-infected early stage cervical cancers and normal cervical epithelium: identification of novel candidate molecular markers for cervical cancer diagnosis and therapy Virol 2005; 331(2 ): 269-91.] and Carlson_Cervix [38 Carlson MW, Iyer VR, Marcotte EM. Quantitative gene expression assessment identifies appropriate cell line models for individual cervical cancer pathways BMC Genom Bio Med Central Ltd 2007; 8(1 ): 117.], cervical cancer versus normal cervical keratinocytes. The upregulated genes are depicted in orange, downregulated in green, and not differentially expressed in white. The annotations of the genes are included as Supplementary Table 1.

Fig. (4) Graphical representation of the genes that were differentially expressed in at least two studies reporting the differential expression between HPV-positive versus HPV-negative head-and-neck cancers. The signatures represent the differentially expressed genes as follows: Schlecht_HNSC [49 Schlecht NF, Burk RD, Adrien L, et al. Gene expression profiles in HPV-infected head and neck cancer J Pathol 2007; 213(3 ): 283-93.] and Slebos_HNSC [47 Slebos RJC, Yi Y, Ely K, et al. Gene expression differences associated with human papillomavirus status in head and neck squamous cell carcinoma Clin Cancer Res 2006; 12(3 Pt 1 ): 701-9.], HPV-positive versus HPV-negative head-and-neck squamous cell carcinoma; Martinez_oropharyngeal [48 Martinez I, Wang J, Hobson KF, Ferris RL, Khan SA. Identification of differentially expressed genes in HPV-positive and HPV-negative oropharyngeal squamous cell carcinomas Eur J Cancer 2007; 43(2 ): 415-32.], HPV-positive versus HPV-negative oropharyngeal squamous cell carcinoma; Pyeon_HPV_positive_cancers [50 Pyeon D, Newton MA, Lambert PF, et al. Fundamental Differences in Cell Cycle Deregulation in Human Papillomavirus-Positive and Human Papillomavirus-Negative Head/Neck and Cervical Cancers Cancer Res 2007; 67(10 ): 4605-19.], HPV-positive head-and- neck cancer (HNC) and cervical cancer versus HPV-negative HNC. The upregulated genes are depicted in orange, downregulated in green, and not differentially expressed in white. The annotations of the genes are included as Supplementary Table 1.

Fig. (5) Venn diagrams illustrating the overlap in differentially expressed genes reported in: A) studies based on cervical cancer tissues. The signatures reported by Rosty et al. [46 Rosty C, Sheffer M, Tsafrir D, et al. Identification of a proliferation gene cluster associated with HPV E6/E7 expression level and viral DNA load in invasive cervical carcinoma Oncogene 2005; 24(47 ): 7094-104.] and Santin et al. [45 Santin AD, Zhan F, Bignotti E, et al. Gene expression profiles of primary HPV16- and HPV18-infected early stage cervical cancers and normal cervical epithelium: identification of novel candidate molecular markers for cervical cancer diagnosis and therapy Virol 2005; 331(2 ): 269-91., 47 Slebos RJC, Yi Y, Ely K, et al. Gene expression differences associated with human papillomavirus status in head and neck squamous cell carcinoma Clin Cancer Res 2006; 12(3 Pt 1 ): 701-9.] show significant overlap. This is in contrast to the list of genes published by Carlson et al. [38 Carlson MW, Iyer VR, Marcotte EM. Quantitative gene expression assessment identifies appropriate cell line models for individual cervical cancer pathways BMC Genom Bio Med Central Ltd 2007; 8(1 ): 117.] that shows only negligible overlap with the other studies. The genes present in all three studies are RFC2, TOP2A, PTTG1, UBE2C (upregulated) and SERPINF1 (downregulated) (also seen in Fig. 3). B) Studies of low-risk HPV. Santegoets et al. [51 Santegoets LAM, van Baars R, Terlou A, et al. Different DNA damage and cell cycle checkpoint control in low- and high-risk human papillomavirus infections of the vulva Int J Cancer 2012; 130(12 ): 2874-85.] and DeVoti et al. [52 DeVoti JA, Rosenthal DW, Wu R, Abramson AL, STEINBERG BM, Bonagura VR. Immune dysregulation and tumor-associated gene changes in recurrent respiratory papillomatosis: a paired microarray analysis Mol Med 2008; 14(9-10 ): 608-17.] reported the differentially expressed genes from condyloma and papilloma samples, respectively. Thomas et al. [37 Thomas JT, Oh ST, Terhune SS, Laimins LA. Cellular changes induced by low-risk human papillomavirus type 11 in keratinocytes that stably maintain viral episomes J Virol 2001; 75(16 ): 7564-1.] reported the differential expression based on a cell culture model in which normal human keratinocytes were transfected with the HPV11 genome. The Venn diagram demonstrates a notable overlap between the tissue-based studies, especially concerning downregulated genes. However, little overlap can be seen between the tissue-based studies and the HPV11 cell culture model (Thomas_HPV11). C) Studies reporting the differential expression between HPV-positive versus HPV-negative head-and-neck cancers. These studies show very little overlap of the signatures. The two genes upregulated in all the studies are SYCP2 and CDKN2C (also seen in Fig. 4).

Several cell lines have been established from patient biopsy material. The HPV16-positive cell lines W12, CaSki and SiHa were originally isolated from a low-grade cervical squamous intraepithelial lesion and cervical cancer. W12 cells carry approximately 100 episomal HPV16 genomes [3 Du J, Nasman A, Carlson JW, Ramqvist T, Dalianis T. Prevalence of human papillomavirus (HPV) types in cervical cancer 2003- 2008 in Stockholm, Sweden, before public HPV vaccination Acta Oncol 2011; 50(8 ): 1215-9.-11 van Houten VM, Snijders PJ, van den Brekel MW, Kummer JA, Meijer CJ, van LB, et al. Biological evidence that human papillomaviruses are etiologically involved in a subgroup of head and neck squamous cell carcinomas Int J Cancer 2001; 93(2 ): 232-5., 32 Stanley MA, Browne HM, Appleby M, Minson AC. Properties of a non-tumorigenic human cervical keratinocyte cell line Int J Cancer 1989; 43(4 ): 672-.]. SiHa cells have only two genome copies and CaSki cells carry up to 600 genome copies [12 Chow LT, Broker TR, Steinberg BM. The natural history of human papillomavirus infections of the mucosal epithelia APMIS 2010; 118(6-7 ): 422-9.-15 Munger K, Basile JR, Duensing S, et al. Biological activities and molecular targets of the human papillomavirus E7 oncoprotein Oncogene 2001; 20(54 ): 7888-98., 33 Baker CC, Phelps WC, Lindgren V, Braun MJ, Gonda MA, Howley PM. Structural and transcriptional analysis of human papillomavirus type 16 sequences in cervical carcinoma cell lines J Virol 1987; 61(4 ): 962-71.].The W12 cell line has been analysed for differentially regulated transcripts following integration of the HPV genome in late passages and the analysis unveiled 85 differentially expressed genes [2 Bosch FX, Burchell AN, Schiffman M, et al. Epidemiology and natural history of human papillomavirus infections and type-specific implications in cervical neoplasia Vaccine 2008; 26(Suppl 10 ): K1-16., 34 Alazawi W, Pett M, Arch B, et al. Changes in cervical keratinocyte gene expression associated with integration of human papillomavirus 16 Cancer Res 2002; 62(23 ): 6959-5.]. The investigators found that the interferon-stimulated transcripts from the genes IFI44, MX1 and MX2 were upregulated and several serine proteases as well as HSP70 transcripts were downregulated (Fig. 1). The profiling of SiHa and CaSki RNA identified, among others, differentially regulated transcripts from cell cycle regulatory genes such as MCMs, p16, CDCs, CYCLINs and TOPO2a [4 Lajer CB, von BC. The role of human papillomavirus in head and neck cancer APMIS 2010; 118(6-7 ): 510-9., 35 Martin CM, Astbury K, McEvoy L, O'Toole S, Sheils O, O'Leary JJ. Gene expression profiling in cervical cancer: identification of novel markers for disease diagnosis and therapy Meth Mol Biol 2009; 511: 333-59.]. To better understand the global changes caused by HPV in tissue culture cell models, profiling data from cells transfected with malignant or benign virus genomes have been compared [16 Syrjanen S, Lodi G, I von B, et al. Human papillomaviruses in oral carcinoma and oral potentially malignant disorders: a systematic review Oral Dis 2011; 17(Suppl 1 ): 58-72., 29 Kaczkowski B, Rossing M, Andersen DK, et al. Integrative analyses reveal novel strategies in HPV11,-16 and -45 early infection Sci Rep [Internet] Nat Pub Group 2012; 2: 515. Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fc gi?dbfrom=pubmed&id= 22808421&retmode=ref&cmd=prlinks, 36 Chang YE, Laimins LA. Interferon-inducible genes are major targets of human papillomavirus type 31: insights from microarray analysis Dis Mark 2001; 17(3 ): 139-42., 37 Thomas JT, Oh ST, Terhune SS, Laimins LA. Cellular changes induced by low-risk human papillomavirus type 11 in keratinocytes that stably maintain viral episomes J Virol 2001; 75(16 ): 7564-1.]. Interestingly, the model in HaCaT cells revealed upregulation of a series of PSGs and ANKRD1 independent of the HPV type - genes that have not previously been discussed in relation to the life cycle of HPV (Fig. 1).

Other studies have focused on the analysis of differentially regulated transcripts in cell lines grown both as monolayers and as organotypic cultures. The organotypic cultures were found to mimic the cancers better than the monolayer cells [8 Shaw R, Robinson M. The increasing clinical relevance of human papillomavirus type 16 (HPV-16) infection in oropharyngeal cancer Br J Oral Maxillofac Surg 2011; 49(6 ): 423-9., 38 Carlson MW, Iyer VR, Marcotte EM. Quantitative gene expression assessment identifies appropriate cell line models for individual cervical cancer pathways BMC Genom Bio Med Central Ltd 2007; 8(1 ): 117., 39 Garner-Hamrick PA, Fostel JM, Chien W-M, et al. Global effects of human papillomavirus type 18 E6/E7 in an organotypic keratinocyte culture system J Virol 2004; 78(17 ): 9041-50.].

Not surprisingly, the global profiles obtained in the different studies vary not only because of different experimental models and growth conditions, but also because of the different technologies employed. Although different studies have identified similar differentially expressed genes (IFI44, MX1 and CYR61), it has not been possible to reach any consensus based on the model systems described in the present section (see also Fig. 1).

The mRNA profiles published in one study on vulval intraepithelial neoplasia (VIN), three studies on cervical SCCs, four studies on HNSCCs and six studies on cell lines were combined and the profiling ‘landscape’ is shown in Fig. (2). Interestingly, there are areas of strong similarity among the signatures from the cervical and vulval lesions, while profiles from HNSCCs show patches of overlap but to a much lesser extent.

Among the cervical and vulval specimens analysed by Rosty et al. [46 Rosty C, Sheffer M, Tsafrir D, et al. Identification of a proliferation gene cluster associated with HPV E6/E7 expression level and viral DNA load in invasive cervical carcinoma Oncogene 2005; 24(47 ): 7094-104.], Santin et al. [45 Santin AD, Zhan F, Bignotti E, et al. Gene expression profiles of primary HPV16- and HPV18-infected early stage cervical cancers and normal cervical epithelium: identification of novel candidate molecular markers for cervical cancer diagnosis and therapy Virol 2005; 331(2 ): 269-91.] and Santegoets et al. [51 Santegoets LAM, van Baars R, Terlou A, et al. Different DNA damage and cell cycle checkpoint control in low- and high-risk human papillomavirus infections of the vulva Int J Cancer 2012; 130(12 ): 2874-85.], the regulated genes are rather consistent. However, the differentially expressed genes published by Carlson et al. [38 Carlson MW, Iyer VR, Marcotte EM. Quantitative gene expression assessment identifies appropriate cell line models for individual cervical cancer pathways BMC Genom Bio Med Central Ltd 2007; 8(1 ): 117.] show only minor areas of overlap and some of the up- and downregulated genes match mostly genes regulated in the VIN specimens (e.g. upregulated CDC25A, AURKB and CDCs; downregulated ANK2, TIMP2 and TGFBR3) (Figs. 2, 3, 5A).

The comparison between the HPV6-positive condyloma specimens collected by Santegoets et al. [51 Santegoets LAM, van Baars R, Terlou A, et al. Different DNA damage and cell cycle checkpoint control in low- and high-risk human papillomavirus infections of the vulva Int J Cancer 2012; 130(12 ): 2874-85.] and the HPV6/11-positive laryngeal papillomas studied by DeVoti et al. [52 DeVoti JA, Rosenthal DW, Wu R, Abramson AL, STEINBERG BM, Bonagura VR. Immune dysregulation and tumor-associated gene changes in recurrent respiratory papillomatosis: a paired microarray analysis Mol Med 2008; 14(9-10 ): 608-17.] shows an overlap of 41 transcripts, whereas there is negligible overlap between the results from the HPV11 cell model of Thomas et al. [37 Thomas JT, Oh ST, Terhune SS, Laimins LA. Cellular changes induced by low-risk human papillomavirus type 11 in keratinocytes that stably maintain viral episomes J Virol 2001; 75(16 ): 7564-1.] and the two tissue studies (Fig. 5B). This observation emphasizes the difficulties of comparing results obtained from tissue samples and cell culture models.

Assessing the overlap of aberrantly expressed genes in the HNSCC studies, there is a reasonable agreement between the findings of Slebos et al. [47 Slebos RJC, Yi Y, Ely K, et al. Gene expression differences associated with human papillomavirus status in head and neck squamous cell carcinoma Clin Cancer Res 2006; 12(3 Pt 1 ): 701-9.] and Pyeon et al. [50 Pyeon D, Newton MA, Lambert PF, et al. Fundamental Differences in Cell Cycle Deregulation in Human Papillomavirus-Positive and Human Papillomavirus-Negative Head/Neck and Cervical Cancers Cancer Res 2007; 67(10 ): 4605-19.]. Schlecht et al. [49 Schlecht NF, Burk RD, Adrien L, et al. Gene expression profiles in HPV-infected head and neck cancer J Pathol 2007; 213(3 ): 283-93.] also observed the upregulation of CDC7, SYCP2 and CDKN2C. Among the downregulated genes, KLK10 and KLK8 should be noted. (The genes are listed in Fig. (4). See also Fig. (5C) for overlap between the results from the studies by Schlecht et al. [49 Schlecht NF, Burk RD, Adrien L, et al. Gene expression profiles in HPV-infected head and neck cancer J Pathol 2007; 213(3 ): 283-93.], Slebos et al. [47 Slebos RJC, Yi Y, Ely K, et al. Gene expression differences associated with human papillomavirus status in head and neck squamous cell carcinoma Clin Cancer Res 2006; 12(3 Pt 1 ): 701-9.] and Martinez et al. [48 Martinez I, Wang J, Hobson KF, Ferris RL, Khan SA. Identification of differentially expressed genes in HPV-positive and HPV-negative oropharyngeal squamous cell carcinomas Eur J Cancer 2007; 43(2 ): 415-32.].

Cultured HPV-positive cells carrying full genomes of HPV11, HPV16 or HPV31 were compared (Fig. 2). Both upregulated (SLP1, IFI27, SAT1) and downregulated (SLC7A5, FLNA, VARS, MCM2, INCENP) genes show consistency in the two studies of cell cultures transfected with the circular genome of HPV11, even though Kaczkowski et al. used HaCaT cells [29 Kaczkowski B, Rossing M, Andersen DK, et al. Integrative analyses reveal novel strategies in HPV11,-16 and -45 early infection Sci Rep [Internet] Nat Pub Group 2012; 2: 515. Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fc gi?dbfrom=pubmed&id= 22808421&retmode=ref&cmd=prlinks] and Thomas et al. used human foreskin keratinocytes [37 Thomas JT, Oh ST, Terhune SS, Laimins LA. Cellular changes induced by low-risk human papillomavirus type 11 in keratinocytes that stably maintain viral episomes J Virol 2001; 75(16 ): 7564-1.]. The HPV16 and HPV11 cell cultures show overlapping genes to some extent, but also unique HPV type-related transcripts. It should be noted that IFI44 and DDX60 are inversely regulated by the malignant and benign HPV. The IFI44 and PI3 genes are also inversely regulated by the HPV16-positive HaCaT cells and HPV16-positive W12 cells. Furthermore, these two genes are the only genes in common between the two HPV16 gene profiles (Fig. 1). This might be due to the integration of the virus genome in W12 cells, which were isolated from cancer tissue (Fig. 1). There are very few regulated genes in common between the cell lines carrying the related HPV types 16 and 31. Only the oncogene ABL2 is upregulated in both cell cultures [29 Kaczkowski B, Rossing M, Andersen DK, et al. Integrative analyses reveal novel strategies in HPV11,-16 and -45 early infection Sci Rep [Internet] Nat Pub Group 2012; 2: 515. Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fc gi?dbfrom=pubmed&id= 22808421&retmode=ref&cmd=prlinks, 37 Thomas JT, Oh ST, Terhune SS, Laimins LA. Cellular changes induced by low-risk human papillomavirus type 11 in keratinocytes that stably maintain viral episomes J Virol 2001; 75(16 ): 7564-1., 53 Chang YE, Laimins LA. Microarray analysis identifies interferon-inducible genes and Stat-1 as major transcriptional targets of human papillomavirus type 31 J Virol 2000; 74(9 ): 4174-82.] (Fig. 1). The observed differences in cellular gene profiles revealed by the comparison of different studies are most likely caused by the use of different host cells. The study by Kaczkowski et al. [29 Kaczkowski B, Rossing M, Andersen DK, et al. Integrative analyses reveal novel strategies in HPV11,-16 and -45 early infection Sci Rep [Internet] Nat Pub Group 2012; 2: 515. Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fc gi?dbfrom=pubmed&id= 22808421&retmode=ref&cmd=prlinks] were, as mentioned above, conducted with immortalized skin epithelial cells (HaCaT), whereas the other studies on HPV16-, HPV31- and HPV11-positive cells were performed using human cervical or foreskin epithelial cells. As discussed in a previous section of this review, a series of PSG as well as ANKRD1 were upregulated in HaCaT cells positive for benign HPV11 or malignant HPV types. Why this is only observed in HaCaT cells could be a topic for further investigation.

Matching of the lists of differentially expressed genes was performed using the official HUGO gene symbols. The matching poses a challenge due to updates and ambiguity of gene names listed in the literature. Additionally, the microarray platforms used in the studies were often based and mapped to older versions of the human genome, and due to a growing body of information concerning the genome, the lists of differentially expressed genes sometimes use outdated gene names. Furthermore, the ID of some probe sets can map to several genes, and since the mapping can be done differently in different studies, the match will not be found by exact matching of gene names. On the other hand, non-exact matching is almost impossible to automate, as it can be very difficult to decide if different gene variants or members of the same gene family represent the same up- or downregulated entity. Therefore, the matching used in this review is very strict. Related genes or gene variants are not presented as a match. This can lead to the bias that results presented by different studies look less coherent than they actually are.

In addition, the statistical analysis of differentially expressed genes is optimized to reduce the number of false positives. This results in strict significance thresholds that many differentially expressed genes do not meet, especially when the number of replicates in the experiment is low. Therefore, only genes showing the strongest differential expression may be detected by multiple or all studies.

The approach that different studies fill the gaps in differential expression analysis can provide a way to balance the number of false positive and false negative genes. For example, we believe that genes reported to be differentially expressed by two independent studies can be safely assumed to be affected in HPV infection or HPV-caused cancers. Therefore, the differential expression in all studies of similar scope should not be considered as a requirement. However, those genes are, of course, of interest due to their very strong signal of differential expression (Fig. 2).

For the analysis of miRNA profiles in HPV16-positive cultured cells, the cell lines SiHa and CaSki have often been employed. We focus on studies where the profiling results are available and related to expression of the HPV oncogenes, with the exclusion of studies designed to analyse the mRNA targets in depth for individual miRNAs as well as studies where the HPV status of the cell lines was not specified.

Martinez et al. conducted a global profiling of miRNAs in cervical cancer-derived SiHa and CaSki cells and showed that human miR-203 and miR-34a were upregulated and miR-218 was downregulated [55 Martinez I, Gardiner AS, Board KF, Monzon FA, Edwards RP, Khan SA. Human papillomavirus type 16 reduces the expression of microRNA-218 in cervical carcinoma cells Oncogene 2008; 27(18 ): 2575-82.]. However, the regulation of miR-34a has been shown to be inversely regulated in other studies - it was downregulated as a result of the expression of the high-risk E6 oncoprotein and the degradation of p53, a known regulator of miR-34a transcription [56 Chang TC, Wentzel EA, Kent OA, et al. Transactivation of miR- 34a by p53 broadly influences gene expression and promotes apoptosis Mol Cell 2007; 26(5 ): 745-52.-58 Raver-Shapira N, Marciano E, Meiri E, et al. Transcriptional activation of miR-34a contributes to p53-mediated apoptosis Mol Cell 2007; 26(5 ): 731-43.]. In addition, miR-21 was upregulated, as demonstrated in several HPV16-transformed cell lines, whereas miR-143 and miR-145 were downregulated [55 Martinez I, Gardiner AS, Board KF, Monzon FA, Edwards RP, Khan SA. Human papillomavirus type 16 reduces the expression of microRNA-218 in cervical carcinoma cells Oncogene 2008; 27(18 ): 2575-82., 59 Lui WO, Pourmand N, Patterson BK, Fire A. Patterns of known and novel small RNAs in human cervical cancer Cancer Res 2007; 67(13 ): 6031-43., 60 Wang X, Tang S, Le S-Y, et al. Aberrant expression of oncogenic and tumor-suppressive microRNAs in cervical cancer is required for cancer cell growth PLoS One 2008; 3(7 ): e2557.] (Table 1).

Table 1.

The Most Differentially Expressed or Promoted hsa-miRNAs in Cervical and Head-and-Neck Cancer from Cell Cultures

Table 2a.

The Most Differentially Expressed or Promoted hsa-miRNAs in Cervical Cancer from Tissues

Table 2b.

The Most Differentially Expressed or Promoted hsa-miRNAs in Head-and-Neck Cancer from Tissues

MicroRNA profiles have also been analysed in HPV16 E5-positive keratinocytes, where miR-146a was found to be upregulated and miR-203 downregulated [61 Greco D, Kivi N, Qian K, Leivonen SK, Auvinen P, Auvinen E. Human papillomavirus 16 E5 modulates the expression of host microRNAs PLoS One 2011; 6(7 ): e21646.]. The authors of the study proposed suppressor of cytokine signalling-3 (SOCS-3) as the target of miR-203 and its involvement in attenuation of the inflammatory response [61 Greco D, Kivi N, Qian K, Leivonen SK, Auvinen P, Auvinen E. Human papillomavirus 16 E5 modulates the expression of host microRNAs PLoS One 2011; 6(7 ): e21646., 62 Sonkoly E, Wei T, Janson PC, et al. MicroRNAs: novel regulators involved in the pathogenesis of psoriasis? PLoS One 2007; 2(7 ): e610.]. Furthermore, miR-324-5p was constantly repressed in E5-expressing cells, whereas its putative target, E-cadherin, was increased at the protein level. These data are interesting as the function of E5 in cell transformation is still not fully understood [63 Crusius K, Rodriguez I, Alonso A. The human papillomavirus type 16 E5 protein modulates ERK1/2 and p38 MAP kinase activation by an EGFR-Independent process in stressed human keratinocytes Virus Genes 2000; 20(1 ): 65-9.], and one additional mechanism by which HPV16 E5 might contribute to the carcinogenic process in cervical epithelial cells could be controlled by miR-324-5p [61 Greco D, Kivi N, Qian K, Leivonen SK, Auvinen P, Auvinen E. Human papillomavirus 16 E5 modulates the expression of host microRNAs PLoS One 2011; 6(7 ): e21646.].

In a recent study, a model for examining the HPV-mediated changes in cellular miRNA profiles was established in HaCaT cells. Cells positive for HPV were analysed and differentially expressed miRNAs shared among the different virus-positive cells were ascertained [64 Dreher A, Rossing M, Kaczkowski B, et al. Differential expression of cellular microRNAs in HPV 11, -16, and -45 transfected cells Biochemical and Biophysical Research Communications [Internet] Elsevier Inc 2011; 412(1 ): 20-5. Available from: http://eutils. ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&id=2178 2796&retmode=ref&cmd=prlinks]. In cells positive for HPV16, only three miRNAs unique for HPV16 were regulated - miR-181c and miR-486-5p were upregulated and miR-1293 was downregulated (Table 1). For the present review, it is also of interest to mention that 16 miRNAs were shared with the benign HPV11. Among these, miR-125a-5p was the most upregulated and miR-576-3p was the most downregulated [64 Dreher A, Rossing M, Kaczkowski B, et al. Differential expression of cellular microRNAs in HPV 11, -16, and -45 transfected cells Biochemical and Biophysical Research Communications [Internet] Elsevier Inc 2011; 412(1 ): 20-5. Available from: http://eutils. ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&id=2178 2796&retmode=ref&cmd=prlinks, 65 Dreher A, Rossing M, Kaczkowski B, Nielsen FC, Norrild B. Differential expression of cellular microRNAs in HPV-11 transfected cells An analysis by three different array platforms and qRT-PCR. Biochemical and Biophysical Research Communications [Internet]. Elsevier Inc Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&id=21078297&retmode=ref&cmd=prlinks 2010; 403(3-4 ): 357-62.]. Cell lines established from HPV-positive and -negative head-and-neck cancer tissue were also analysed for changes in the miRNA profile [66 Wald AI, Hoskins EE, Wells SI, Ferris RL, Khan SA. Alteration of microRNA profiles in squamous cell carcinoma of the head and neck cell lines by human papillomavirus Head Neck 2011; 33(4 ): 504-12.]. It was found that 11 miRNAs were differentially regulated in HPV-positive HNSCC cells when compared to the profiles obtained from HPV-negative HNSCC cells or normal oral keratinocytes. Three miRNAs were upregulated (miR-363, miR-33 and miR-497) and eight were downregulated, with the two most downregulated being miR-155 and miR-181a (Table 1). The authors of this study did not find miR-21 to be differentially expressed in HPV-positive cell lines, although it was abundant in the cells. The upregulation of miR-363 is interesting because it belongs to the cluster of oncogenic miRNAs, the miR-17-92 family, which might have evolved through a series of mutations [67 Ventura A, Young AG, Winslow MM, et al. Targeted deletion reveals essential and overlapping functionZs of the miR-17 through 92 family of miRNA clusters Cell 2008; 132(5 ): 875-6.]. It is therefore likely that miR-363 is involved in dysfunction of the cell cycle, analogous to the function of miR-25 which is also a member of the miR-17-92 family [68 Kim YK, Yu J, Han TS, et al. Functional links between clustered microRNAs: suppression of cell-cycle inhibitors by microRNA clusters in gastric cancer Nucleic Acids Res 2009; 37(5 ): 1672-81.]. It is worth noting that miR-155 as well as miR-181a were upregulated in other studies on head-and-neck cells, but the cell lines were not analysed for the presence of HPV [69 Chang SS, Jiang WW, Smith I, et al. MicroRNA alterations in head and neck squamous cell carcinoma Int J Cancer 2008; 123(12 ): 2791-7.-71 Wong TS, Liu XB, Wong BY, Ng RW, Yuen AP, Wei WI. Mature miR-184 as Potential Oncogenic microRNA of Squamous Cell Carcinoma of Tongue Clin Cancer Res 2008; 14(9 ): 2588-92.]. The inconsistency among the different studies most likely reflects different HPV status and different origin of the cell cultures.

The profiles of differentially expressed miRNAs observed in raft cultures established from HPV16-positive cells were analysed in a few studies. Raft cultures allow differentiation of the epithelial cells. It was shown that miR-34a and miR-203 were both downregulated [72 Melar-New M, Laimins LA. Human papillomaviruses modulate expression of microRNA 203 upon epithelial differentiation to control levels of p63 proteins J Virol 2010; 84(10 ): 5212-21., 73 Wang X, Wang HK, McCoy JP, et al. Oncogenic HPV infection interrupts the expression of tumor-suppressive miR-34a through viral oncoprotein E6 RNA 2009; 15(4 ): 637-47.]. How-ever, miR-203 is normally upregulated by differentiation of normal human keratinocytes [72 Melar-New M, Laimins LA. Human papillomaviruses modulate expression of microRNA 203 upon epithelial differentiation to control levels of p63 proteins J Virol 2010; 84(10 ): 5212-21.]. Supporting results were reported in raft cultures of HPV18 and HPV31 [72 Melar-New M, Laimins LA. Human papillomaviruses modulate expression of microRNA 203 upon epithelial differentiation to control levels of p63 proteins J Virol 2010; 84(10 ): 5212-21., 73 Wang X, Wang HK, McCoy JP, et al. Oncogenic HPV infection interrupts the expression of tumor-suppressive miR-34a through viral oncoprotein E6 RNA 2009; 15(4 ): 637-47.].

Many studies have focused on the miRNA profiling of tumour tissues with the purpose of identifying unique diagnostic markers for specific cancer types. The hope has also been to find new pathways and mechanisms involved in the progression to cancer and to identify possible drug targets. In cancers related to infection with high-risk HPV, several changes in the miRNA profiles have been observed and the specific findings in cervical tissue and in oral-pharyngeal tissue are discussed separately.

Extensive expression profiling studies have been performed for both mRNAs and miRNAs. However, little is known concerning the regulatory dependencies between miRNAs and their mRNA targets in HPV infections and HPV-caused cancers. Unveiling of the regulation is hampered by the fact that a single miRNA can target multiple mRNAs and a single mRNA can be targeted by multiple miRNAs. Additionally, miRNAs function as fine-tuners or modulators rather than binary switches of gene expression.

In the cited studies of this review, there is a large number of changes in mRNA expression that are not due to miRNA regulation, e.g. the activity of viral proteins on cellular transcription factors and signalling pathways. In addition, multiple miRNAs are affected during HPV infection, many of them having multiple mRNA targets. Finally, the causality is difficult to establish. First, miRNAs can be induced by viral activity. Secondly, there can be a cellular response to the changes caused by the virus.

The array studies published to date for HPV have addressed the regulation between miRNAs and mRNAs on a one-to-one basis only [56 Chang TC, Wentzel EA, Kent OA, et al. Transactivation of miR- 34a by p53 broadly influences gene expression and promotes apoptosis Mol Cell 2007; 26(5 ): 745-52., 75 Chan JA, Krichevsky AM, Kosik KS. MicroRNA-21 is an antiapoptotic factor in human glioblastoma cells Cancer Res 2005; 65(14 ): 6029-33., 91 Martin MM, Lee EJ, Buckenberger JA, Schmittgen TD, Elton TS. MicroRNA-155 regulates human angiotensin II type 1 receptor expression in fibroblasts J Biol Chem 2006; 281(27 ): 18277-84.]. In order to elucidate the regulatory dependencies in these complex settings, one needs a substantial number of paired expression profiles of miRNAs and mRNAs and an extensive computational analysis. Such an analysis has been reported for hepatitis C virus [92 Peng X, Li Y, Walters K-A, et al. Computational identification of hepatitis C virus associated microRNA-mRNA regulatory modules in human livers BMC Genomics 2009; 10: 373.]. However, no studies of this type have presently been employed in relation to HPV infections. Such studies could further add to the understanding of the importance of miRNA targeting as a regulatory tool used by the virus to modulate the host cell gene expression to the benefit of the virus.

The variation observed in different studies appearing to use similar experimental or clinical conditions should be considered with caution. First, single genes have multiple functions and can be involved in many pathways and cellular processes. If one finds that a HPV type leads to differentially expressed genes which turn out to be similar to those found in another study, this does not guarantee that the same pathway or cellular process is affected by the HPV type in both studies.

Secondly, different HPV types in different biological conditions can affect the same pathway, but by affecting different genes in the cascade of genes involved. Therefore, different genes observed to be differentially expressed might influence the same pathway.

We believe that future comparisons of the outcomes of different studies should be carried out on the level of affected pathways and cellular processes rather than by comparing lists of differentially expressed genes. Standard differential expression analysis does not provide the answers to which pathways and cellular processes are affected. Additionally, long lists of genes are very difficult for the reader to comprehend.

Often, arbitrary sets of genes are randomly selected from a long list of data and interpreted as the biological findings while the remainder of the genes on the list are ignored. Hence, in our opinion, more advanced bioinformatic analysis is required to overcome the limitations of such an approach. In our recent study, we integrated available knowledge of protein-protein interactions, molecular signatures, cellular pathways and transcription factor binding sites [29 Kaczkowski B, Rossing M, Andersen DK, et al. Integrative analyses reveal novel strategies in HPV11,-16 and -45 early infection Sci Rep [Internet] Nat Pub Group 2012; 2: 515. Available from: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fc gi?dbfrom=pubmed&id= 22808421&retmode=ref&cmd=prlinks]. The outcomes of such analyses add more information to which pathways and networks of genes are affected by the virus. The results are therefore more informative to the reader and make the comparison of different studies easier. We believe that such an approach can facilitate significant biological findings and an understanding of the biology behind the observed differential expression.