Cent. Eur. J. Biol.• 5(5) • 2010 • 554-571 DOI: 10.2478/s11535-010-0051-z

Central European Journal of Biology

Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

Review Article Július Rajčáni1,2,*, Marián Adamkov1,3, Jana Hybenova1, Jaroslav Jackuliak1, Marian Benčat1

1Alpha Medical a.s., 03601 Martin, Slovak Republic

2Institute of Virology, Slovak Academy of Sciences, 84505 Bratislava, Slovak Republic

3Institute of Histology and Embryology, Jessenius Faculty of Medicine, Comenius University, 03601 Martin, Slovak Republic Received 15 October 2009; Accepted 13 April 2010

Abstract: The p16/INK4A protein is a cellular regulatory polypeptide over-expressed in the presence of high levels of the Human Papillomavirus (HPV) coded E7 protein. This review outlines the use of p16 antigen staining in cervical biopsies as well as in PAP smears summarizing the corresponding literature and commenting the authors’ own experience. The p16 antigen is a reliable marker for dysplastic cells in CINII/CINIII (HSIL) lesions as viewed in cervical biopsies. When PAP smears were examined at large scale screening for p16 antigen- reactive and atypical cells, considerable variations could be found especially in ASCUS graded lesions. Therefore, the presence of p16-reactive atypical cells in PAP smears should be interpreted together with the cytological signs of dysplasia, such as the altered N/C ratio. In addition, women revealing p16-positive ASCUS/LSIL specimens should be examined for the presence of HPV DNA. Detection of HPV DNA alone, i.e. in the absence of cytological screening has a low predictive value, since the clearance of HPV may occur even in the absence of morphological alterations. Combined cytological as well as molecular follow up contributes to the efficiency of diagnostic and increases the probability of correct interpretation of the pre-cancerous lesions by non-invasive techniques. Keywords: Staining for p16/INK4 antigen • Human papillomavirus (HPV) DNA testing • Screening of pre-cancerous lesions • Cervical intraepithelial neoplasia (CIN) • Squamous intraepithelial lesion (SIL) © Versita Sp. z o.o.

Abbreviations HSIL - High grade SIL; ASC - Atypical squamous cell; L - Late (expressed) proteins; ASCUS - Atypical squamous cells of undetermined LCR - Long control region; significance; LR-HPV - Low risk HPV; CIN - Cervical intraepithelial neoplasia; LSIL - Low grade SIL; CPE - Cytopathic effect; NIL - No intraepithelial lesion; vDNA - Viral DNA; N/C ratio - Nuclear to cytoplasm ratio; E - Early (expressed) proteins; nt - Nucleotide; EGF - Epidermal growth factor; ORF - Open reading frame; GP - General primer; PAP - Papanicolaou (staining); H.E. - Haematoxylin eosin (staining); PDGF - Platelet derived growth factor; HPV - Human papillomavirus; PCR - Polymerase chain reaction; HC - Hybrid capture; SIL - Squamous intraepithelial lesion; HR-HPV - High risk HPV; SSC - .

* E-mail: [email protected] 554 Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

1. Host cell transformation by human and/or with severe dysplastic changes, were designated as high risk (HR-HPV). In contrast, genotypes causing papillomaviruses benign lesions and only exceptionally related to cancer Human papillomaviruses (HPV) infecting the female were designated as low risk. genital tract represent the main cause of cervical The HPV virions are small (55 nm in size) non- dysplasias [1,2] and subsequently developing enveloped capsids. As result of productive replication, squamous cell carcinomas (SSC). In addition to HPV, they form aggregates within nuclei of squamous the Papillomaviridae family encompasses viruses epithelium cells that show koilocytosis [6]. HPV does infecting cattle (BPV), monkeys (RhPV), horses (EcPV) not grow outside of squamous epithelium cells, i.e. it and other mammals. The human (HPV) isolates were is extremely difficult to isolate the virus in conventional grouped into 118 genotypes based on the similarity human cell cultures [7]. The purified virions reveal and/or variations of their genomes. The genotype was icosaedral symmetry, being composed of 72 subunits defined by at least 10% dissimilarity of the capsid formed by 2 structural proteins (L1 and L2), which are polypeptide (L1) sequence. Based on the comparison expressed at late intervals of the productive replication of L1 sequences of all genotypes, 96 of them fulfilling cycle. The viral double stranded DNA (vDNA) has the above mentioned criterion were included into the about 8 kilobase pairs (kbp) and encodes 7 or 8 non- novel classification, while the rest are regarded either structural polypeptides, designated E1-E8 (E=early). for candidate genotypes and/or for subtypes (2-10% The transcription of viral mRNA is under control of nucleotide dissimilarity). According to novel classification a 1.1 kbp vDNA region, which is rich in promoters [3,4], the Papillomaviridae family has 13 genera. The (LCR=long control region). The oncogenic virus coded clinically important genotypes that cause lesions at polypeptides E5, E6 and E7 induce cell proliferation and/ mucosal membranes and/or on the skin belong to or may drive cells into permanent division by different 12 out of 15 species within the Alphapapillomavirus mechanisms. Their open reading frames (ORF) are genus (Table 1). Noteworthy, the recently accepted L1 located close to LCR, which has a great impact for host phylogenetic tree-based classification may not reflect cell transformation resulting from the integration of HPV the serological relationships among the L1 proteins. genome in the course of latency [8-11]. Because the No Alphapapillomavirus common or species-specific HPV genome does not encode its own DNA polymerase, epitopes could be determined in cross ELISA tests its productive replication depends on the division of host against recombinant L1 fusion proteins from 15 HPV epithelium cells. The vDNA molecules are copied by the genotypes representing 6 species [5]. In clinical practice, help of E1 and E2 proteins within dividing cells of the the genotypes most probably associated with cancer basal and/or suprabasal layers of squamous epithelium.

Disease Frequent association Rare association

Skin infection:

Deep plantar 1, 2 4, 63

Common warts 2, 1 4, 26, 27, 29, 41, 57, 65 and 77

Butcher’s warts 7, 2 1, 3, 4, 10 and 28 Epidermodysplasia verruciformis (ER)* 2, 3, 5, 8, 9, 10, 12, 14, 15 and 17 19, 20, 21, 22, 23, 24, 25, 36, 37, 38, 47 and 70 Skin carcinoma associated with ER 5 and 8 Anogenital infection:

Condyloma acuminatum 6, 11 30, 42, 43, 44, 45, 51, 54, 55, 70 Intraepitelal dysplasia (16, 18, 74, 86) 6, 11, 26, 31, 33, 35, 39, 42, 43, 44, 45, 51, 6**, 11**, 16, 18 (CIN I a CIN II/III) 52, 53, 56, 58, 66, 73 and 82 Carcinoma in situ; invasive carcinoma 16, 18, 31, 45 33, 35, 39, 51, 52, 56, 58, 59, 66, 67, 68, 73 and 82

Recurrent laryngeal 6, 11

Conjuctival papilloma 6, 11 and 16

Table 1. High risk (in bold) and low risk HPV genotypes according to their frequency.

* congenital skin lesion with high sensitivity to HPV **(CIN I/LSIL, CIN II or non-invasive CINIII/HSIL)

555 J. Rajčáni et al.

During differentiation of the stratified epithelium, in a proportion of ASCUS graded smears (sometimes productive (vegetative) virus replication continues within characterized by ASC-H cells) and in LSIL smears, the spinous and/or surface epithelium cells, in which but later on becomes clearly prevalent in HSIL grade the L1/L2 and E4 proteins are abundantly expressed smears and, of course, in . Women (Table 2). The E4 protein causes dramatic changes showing the prevalence of integrated HPV DNA were in the cytokeratin network altering the cytoplasmic almost 10 years older than those with a predominating environment of spinous cells, which contributes to the episomal HPV DNA pattern, which points to a higher risk HPV induced cytopathic effect and koilocytosis [12-14]. of HPV infection in women aged over 35 years [20]. These changes accelerate the release of virions Three non-structural HPV coded oncoproteins are assembled due to self-aggregation of the L1/L2 protein involved in host cell proliferation and immortalization complexes. In contrast, the few vDNA copies harbored (E5, E6 and E7) causing continuous host cell division in suprabasal and transitory cells during latency get rather than direct malignant transformation (Table 3). distributed into dividing daughter cells by means of E2 The E5 protein enhances the sensitivity of HPV carrier protein [15,16]. In such a way, the latent HPV genome cells to external proliferative stimuli, such as the is continuously maintained in the basal and suprabasal epidermal growth factor [21]. It also binds to growth squamous epithelium layer despite the absence of factor receptors, for example the platelet growth productive virus formation. It should be mentioned factor receptor (PDGFR), and activates the signal that during long-term latency, the integration of HPV transmission in a ligand independent manner [22,23]. DNA may associate with linearization of the persisting Because the integrated vDNA always encompasses vDNA molecule. At integration, the circular HPV DNA is the ORFs of E6 and E7 oncoproteins as well as the interrupted, usually within the E2 gene ORF [17]. Thus, closely positioned LCR sequence, the continuing the HPV genome persists within transformed host cells and increased expression of E6/E7 proteins seriously in two different forms: as non-integrated (episomal) affects the regulation of host cell division [24,25]. circularized full-length vDNA and, less frequently, as Noteworthy, E6 polypeptide interacts with the pivotal a linear and integrated vDNA sequence with the E2 cell division inhibiting p53 protein [26], while the E7 ORF partially deleted (preferentially at nt 3362-3443) polypeptide binds the retinoblastoma (Rb) protein and [18,19]. The mixed (episomal as well as integrated) cyclin inhibitory proteins p27 and p21 [27]. The HPV vDNA distribution pattern seems the most prevalent coded oncoproteins E6/E7, when present in significant physical state of HPV16 DNA found in ASCUS graded amounts, continuously drive host cells from phase G1 smears (atypical cells of unknown significance), but can to phase S. Noteworthy, the process of cancerogenesis also be detected in cervical scrapings still not revealing is more complex and may be related, for example, to pathological changes. This indicates that HPV infection increased number of c-myc gene copies in HPV DNA may not always cause dysplasia of the squamous transformed cells, a phenomenon proportional to the epithelium cells. The prevalence of the integrated HPV grade of dysplasia seen in the corresponding lesion DNA sequence over episomal molecules then appears [28]. The aim of this paper is to consider the role of p16/

Protein MW Properties and function

Binds to the regulatory long control region (LCR/URR)1 upstream from the E6/E7 ORF promoters. Forms a heterodimer E1 68-76 kDa with E2, associates with H1 histones and cyclins, especially with cyclin E

Activates viral mRNA transcription, binds to the LCR/URR sequence, associates with the E1 protein; acts as cofactor of vDNA replication; operates at distribution of newly copied vDNA molecules during cell division and contributes to E2 40-58 kDa the maintenance of latency. Suppresses the expression of E6/E7 proteins and interacts with the L2 capsid polypeptide (especially at productive virus replication). Induces apoptosis

E3 10-17 kDa Function unknown

Associates with the L1 capsid polypeptide and facilitates virion formation (at productive replication); dissociates E4 cytokeratins in the medium and upper spinous layers

Table 2. Basic functions of the HPV coded early (non-structural) polypeptides involved in replication and latency.

1Upstream Regulatory Region

556 Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

Protein MW Properties and function

Increases and prolongs the activity of receptors interacting with external growth factors such as EGF1 and/or PDGF2 by binding to their cytoplasmic domain. Inhibits the acidification of endosomes, binds to ATPase within the vacuolar E5 10 kDa membrane and increases the stability of the engulfed EGF/EGFR3 complexes. Activates certain cellular transcription factors such AP1* and signaling proteins c-Jun4 and c-Fos

Binds to p53 (cellular anti-onc protein) and enhances its degradation by the ubiquitin system (most efficiently acts E6 E6 16-18 kDa protein encoded by the HPV-16 and/or HPV-18 genotypes

Binds to retinoblastoma (Rb) protein and the functionally related p107 and 130 proteins, which regulate cellular transcription (Figure 1). Binds to the signaling protein c-Jun and inhibits the expression of cyclin dependent kinases. E7 10-14 kDa Activates cyclins A and E and the transcription factor AP1. Activates the histone deacetylase (HDAC) and increases cellular transcription

Table 3. Basic functions of the HPV coded early (non-structural) proteins.

1EGF = epidermal growth factor; 2PDGF = platelet derived growth factor; 3EGF receptor; *AP1 = activator protein 1; 4c-Jun = cellular ju-nana.

INK4A polypeptide over-expression under conditions of increased E7 protein production and failure of its cell division regulating function.

2. Overexpression of the p16/INK4A polypeptide in HPV transformed cells

The p16/INK4A (inhibitor kinase) polypeptide is a cellular regulatory protein, which inhibits the cyclin dependent kinases (especially cdk4 and cdk6) associated with cyclin D. As shown in Figure 1, the activated kinases (cdk4 and/or cdk6) phosphorylate the retinoblastoma phosphoprotein (pRb) complex, which in turn, releases the transcription factor E2F, which in the non-dividing Figure 1. Cyclin dependent kinases (4 and/or 6) interact with cells is bound in its inactive form. Under physiological cyclin D (cell division regulator) and phosphorylate conditions, the free Rb protein liberated from the disrupted the retinoblastoma phosphoprotein (pRb). The complex has a feedback effect on p16 expression phosphorylated pRb becomes activated (ppRb) and releases the transcription factor transcription activator [29]. Since the E2 polypeptide binds to LCR (Table 3), E2F, which in turn, is transported to nucleus. Under the presence of E2 in cells carrying the episomal HPV physiologic conditions, the p16/INK4 polypeptide (which is an inhibitory kinase) inhibits cdk4/6 activity. The high DNA may efficiently control the transcription of mRNA level of free ppRb has a feedback effect on INK4/p16 encoding the E6/E7 oncoproteins [30]. In contrast, the synthesis. Because the E7 oncoprotein binds the pRb protein in dysplastic and/or HPV-transformed cells, the cells carrying an integrated HPV genome would stop availability of the latter decreases in them, while the producing the E2 polypeptide due to disruption of the E2 expression of p16 increases. ORF. Therefore, E2 but also E1 proteins (both are closely related to the maintenance of long term latency) may be thoroughly reviewed [35]. The mild forms are usually missing in HPV transformed cells, which eliminates the referred to as low grade SIL (LSIL), whereas the down regulation of E6/E7 oncoprotein production. Taken progressed forms are referred to as high grade SIL together, in HPV transformed cells, the expression of (HSIL). The SIL classification had been originally p16/INK4A protein increases proportionally to elevated introduced for vaginal/cervical smears [36], but later levels of the virus coded E7 polypeptide, since both arise on it became widely known as the in the absence of E2 [31-34]. and adopted for histological scoring [37]. However, The histopathology of precancerous cervical long before introduction of the latter scoring system, squamous intraepithelial lesions (SIL) has been pathologists developed their own nomenclature,

557 J. Rajčáni et al.

which is until now frequently used for cervical biopsy grading. In this system, the precancerous changes of exocervical squamous epithelium are designated as cervical intraepithelial neoplasia (CIN). The changes at CIN are associated with the appearance of altered (dysplastic) cells still showing a diploid nuclear pattern [38]. Therefore, the CIN I stage cannot be regarded as truly neoplastic. Dysplastic cells are characterized by loss of polarity, crowding, overlapping, disorganization and anisocytosis. At the cytological level, the dysplastic cells show altered nuclear to cytoplasm (N/C) ratio as well as wrinkling and thickening of nuclear membrane. For reasons discussed below, it seems useful to group any CIN I grade changes into the entity of LSIL [39]. Figure 2. Positive zonal staining of p16 within the parabasal layer In mild forms designated CIN I/LSIL, the dysplastic showing dysplastic cells in CIN I/LSIL (magn. x120). cells as a rule, occupy the parabasal epithelium layer. Being located in the lower third of cervical squamous epithelium, dysplastic cells form a continuous zone easily distinguishable by p16 antigen staining (Figure 2). Table 4 shows that parabasal staining of p16 antigen in mild grade CIN I/LSIL dysplasia might be positive in up to 81% of specimens. This result probably represents the highest limit as compared to the findings of others. Intensive parabasal staining for p16 antigen was described in not more than 47% of CIN I cases, but continuous positive staining of lower intensity was reported in over 70% of CIN I cases [40,41]. Summing up the results of antigen p16 staining in the biopsy sections graded CIN I/LSIL, Yildiz et al. [42] recognized Figure 3. Non-specific scattered p16 staining of proliferating continuous parabasal staining of high intensity, squamous cells within the lower and upper spinous layers as well as in some fibroblasts in subepithelial continuous parabasal staining of lower intensity and connective tissue. This lesion was classified as CIN I/ scattered staining of dispersed squamous epithelium LSIL at H.E. staining (magn. x140). cells. The latter pattern of p16 reactivity (Figure 3) cannot be regarded for an indicator of dysplasia, since it shows Sample number Patient number Histological p16 positive spinous cells at non-typical localization. diagnosis Scattered distribution of p16 expressing cells points at Total per patient Total p16 positive (%) their transient proliferation rather than at the association CIN I/LSIL 202 2.5 80 65 (81.3%) with HPV latency and E7 overproduction. If dysplastic cells occupy at least one half (or two CIN II/HSIL 113 2.4 47 45 (95.7%)* thirds) of the original squamous epithelium (but not the CIN III/HSIL 67 2.5 27 27 (100%) whole epithelial layer), the appropriate designation is HSIL total 180 2.5 74 72 (97.2%) CIN II. In CIN II/HSIL, the p16 staining may be helpful SIL for defining the precise thickness of dysplastic changes (dysplasia) 382 2.5 154 137 (88.9%) (Figure 4). At the stage of CIN III/HSIL, dysplastic cells total replace the original epithelium in full, so that the p16 Table 4. Results of p16 antigen staining in cervical biopsies. staining becomes diffuse (Figures 5, 6). In both, CIN II as * in 2 cases no dysplastic epithelium could be found in repeatedly cut well as CIN III, the p16 reactivity is a helpful indicator of parallel sections the extent of dysplasia. As reported [43], medium (CIN II) and/or severe (CIN III) dysplasia was accompanied with the presence of p16 antigen at 95% (in 45 out cytoplasm of dysplastic cells. If the basal membrane of 47) and/or at 100% (in each of 27) of specimens, is not yet disrupted, the diffusely positive p16 antigen respectively (Table 4). In CIN III/HSIL, intensive staining staining is a hallmark distinguishing carcinoma in situ of p16 antigen can be seen within nuclei as well as the (Figure 7) from microinvasive carcinoma. In clearly

558 Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

invasive cervical squamous carcinomas that penetrate sensitivity for the detection of glandular dysplasias and/ the basal membrane, the neoplastic cells also show or endocervical carcinomas originating from cylindrical overwhelming positive stain (Figure 8). Though the p16 epithelium cells [44]. staining is a useful marker for dysplastic squamous Along with the discovery of vaginal HPV infection, cells and/or spinous carcinomas, it may be of reduced the term condyloma acuminatum was introduced to describe the predominantly benign vaginal/cervical lesions characterized by viral cytopathic changes () and/or hyperplasia of the basal and parabasal epithelium cells [45-47]. Nevertheless, it has turned out that it is impossible to differentiate, using classical histology, between mild cervical lesions caused by LR-HPV types (such as HPV 6 or 11) and/or the early stages in the lesions caused by HR-HPV [48]. Many early HPV-related CIN I/LSIL lesions of the stratified epithelium include mild nuclear atypia associated with extensive perinuclear “halo” and cytoplasmic cavitation (koilocytosis). These changes occur at the spinous and surface layers along with the mild dysplasia and/ or hyperplasia of basal and suprabasal layers. In such Figure 4. Zonal p16 staining of dysplastic cells, which occupies nearly the half of squamous epithelium indicating the cases, p16 staining helps to recognize dysplastic cells transition to CIN II/HSIL; other sections from the same located within the parabasal layer from the p16 negative block which stained with H.E. were classified as CIN I/ LSIL (magn. x480). koliocytes in the medium and/or upper spinous layers

Figure 5. Diffuse staining of p16 antigen encompasses the whole Figure 7. Diffuse staining for p16 antigen in dysplastic cells which dysplastic epithelium in CIN III/HSIL. Dysplastic cells form a protruding structure still surrounded by the intact of squamous origin surround a group of intact cervical basement membrane; classified as CIN III /HSIL (magn. glands, which epithelium is negative for p16 (magn. x300). x140).

Figure 6. Positive p16 staining of the nuclei and cytoplasm of Figure 8. Undifferentiated cervical squamous cell carcinoma at low dysplastic cells in CIN III/HSIL classified as carcinoma power view shows diffuse staining of p16 demonstrating in situ (magn. x500). the area of dysplastic cells and the area of the invasive growth of neoplastic cells (magn. x56).

559 J. Rajčáni et al.

undergoing cavitation and vacuolization (Figure 9). (i.e. over the course of many years), depending on the Noteworthy, in situ hybridization shows the presence extent of vDNA integration (which is higher in HR-HPV of abundant HPV DNA predominantly in the nuclei of infection), disruption of E2 gene and loss of E2 protein productively infected p16 negative koilocytes, even production along with additional cofactors facilitate though dysplastic cells as well as cancer cells also are continuous division of carrier cells (immortalization) vDNA positive [49]. The positive rate of HPV 16 genome resulting in the development of carcinoma. presence as detected by the less sensitive in situ Immunohistochemical detection of p16 antigen in hybridization technique increased from mild to severe dysplastic cervical cells and/or in cervical carcinoma dysplasias reaching just about 50% in CIN III/HSIL. cells is less tedious than that of E7 protein and/or mRNA [54]. Despite the high statistical relevance of p16 detection in bioptic materials graded as HSIL, at the stage of LSIL some pitfalls may hamper the distinction between reactive (proliferative) changes and the initial dysplastic lesions. As already mentioned, an increased p16 expression unrelated HPV infection may occur in single epithelium cells for reasons other than E7 over expression. The expression of CDKN2A gene (encoding p16 protein) due to proliferative stimuli is activated by alternative pathways. As shown in Figure 3, dispersed non dysplastic spinous epithelium cells as well as proliferating fibroblasts below the epithelium layer may occasionally express p16 protein to the extent that they Figure 9. Positive p16 staining of dysplastic cells in the parabasal layer of squamous epithelium in CIN I/LSIL. Vacuolated are rendered stainable with the commercially available cells and numerous koilocytes within the spinous layer monoclonal antibody (CINTec histology kit). Tissue are negative (magn. x140). fixation, variations in pre- and post-staining procedures, differences in the antibody reactivity and dilution would Because the early HPV-related CIN I/LSIL grade also influence the staining results, especially in CIN I/ lesions do not represent true neoplasia, but reactive LSIL cases which show faint p16 reactivity. changes and/or beginning of immortalization due to HPV A frequently used immunohistochemical marker in infection, several authors prefer the SIL terminology. In intraepithelial cervical dysplasia (CIN/SIL) is the Ki-67 fact, only a small proportion of CIN I cases progresses protein that occurs in the nuclei of proliferating and/ quickly to grade CIN II and/or CIN III [50]. The CIN I/ or repeatedly dividing (immortalized) cells, including LSIL state frequently develops into subacute but still dysplastic cells [55]. The decreased availability of the transient HPV infection, which heals spontaneously functioning Rb protein due to the over-expression with a probability of about 60% [51]. In association with of E7 polypeptide, not only leads to p16 over- the accumulation of HPV virion-specific T lymphocytes expression, but also explains the higher expression below the epithelium layer, many koilocytes undergo Ki-67 protein [56]. To achieve precise grading of LSIL apoptosis. The T cell lymphoproliferative response is versus HSIL lesions following Ki-67 antigen staining, specific for E6/E7 antigens and accelerates the HPV 16 Kruse et al. [57] suggested to count the number of clearance [52]. As noted by others, 67% of the HR-HPV positive nuclei per 100 μm epithelium thickness starting positive cases may show regression of LSIL lesions and from the parabasal zone (stratification index). The clears without transition to HSIL. If the virus persists for authors in question found a positive correlation (of 83%) at least 12 months in the non-productive (latent) state, with the original histological classification based on H.E. the probability of progression to CIN II/CIN III within staining, but also noticed a tendency to worse grading next 18 months, i.e. by 30 months of the observation at the primary evaluation of the H.E. staining results. period, has increased to about 23% [53]. Even when the Namely, in two cases (3%) Ki-67 staining showed virus-producing squamous cells are slowly eliminated, lesions grade CIN I rather than lesions grade CIN II as the episomal full length vDNA still persists for a long described in H.E. stained sections. In nine cases (14%), time. During non-productive latency, the virus carrier the original diagnosis for H.E. stained sections was cells double at each division and vDNA gets copied and CIN III, but the Ki-67 staining based score was CIN II. A redistributed into daughter cells. The expression of E1/ similar tendency to worse grading if evaluating the H.E. E2 polypeptides continues at moderate levels, while the stained sections only, i.e. not regarding to p16 staining, intensity of E7 expression still remains low. Later on is noted in Table 5. According to this experience, the

560 Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

repeated reading of parallel sections stained for p16 Geisinger et al. [62], the main criterion for the clear antigen, showed an excellent agreement for HSIL in 60 definition of SIL is an increased size of nucleus (<3 out of 72 cases (83%). The remaining 12 HSIL cases fold for ASCUS, > 3 fold in LSIL), increased intensity of (17%) originally classified in H.E. stained sections were chromatin staining and its altered internal structure (finely later evaluated as LSIL at viewing the p16 stained granular chromatin structure and slight hyperchromasia sections. Nevertheless, such comparisons suffer from in ASCUS, coarse chromatin and definitely visible possible imprecision, since the repeatedly stained hyperchromasia in LSIL and/or HSIL). The diagnosis sections may not fully correspond to the previously cut ASC gains prediction value, only if based on the presence block level. Taken together, staining of parallel sections of a few HSIL-like cells (so called ASC-H), pointing at for Ki-67 and p16 antigens confirmed the usefulness of progression to LSIL and/or HSIL from early beginning. both markers [58,59]. Similarly to p16 antigen detection, Nevertheless, the ASC-H category seems problematic, the Ki-67 protein also shows non-specific expression in since the morphologic difference in the appearance of proliferating cells in at least in 10% of cases [60]. The a single metaplastic and/or neoplastic cell is not always zonal distribution of p16 reactivity fits better with the clear and opens the way to false positive as well as false classical grading into CIN I (dysplasia encompasses negative diagnoses [63-65]. Abati et al. [66] stressed that less than one third of stratified epithelium), CIN II nuclear hyperchromasia and irregular nuclear membrane (dysplastic cells encompassing more the half or up to contours count for the most reliable diagnostic feature of two thirds of squamous epithelium) and CIN III (nearly so-called pre-neoplastic atypia (possibly corresponding diffuse distribution of dysplastic cells encompassing to the dysplastic cells in histological sections). In most the whole epithelium). Therefore, staining of p16 uncertain cases, a short course of local estrogen cream antigen has recently emerged as a preferred diagnostic therapy followed by repeated PAP tests was suggested marker suitable for quick demonstration of the extent of as helpful [67]. More recently, improved cytological dysplasia. tests were introduced, such as liquid based cytology (Liqui-PREP), which resulted in a significant decrease Degree of dysplasia as of unsatisfactory smears and detected more atypical seen at p16 staining squamous cells and LSIL lesions due to enhanced Histological diagnosis (at HE staining) cell visualization in comparison with conventional PAP CIN I CIN II CIN III smears [68]. Several reports claimed the relative value of p16 Erosion (negative control) 4 0 0 staining in ASCUS smears, which may be positive in

CIN I/LSIL 61 3 1 60% of cases, a finding designated p16 reactive ASC. The proportion of p16 negative smears is usually the CIN II/HSIL 6 36 3 highest in ASCUS (40%) and the lowest in HSIL (5%) CIN III/HSIL 6 4 17 specimens [69,70]. According to Shin et al. [71], p16 HSIL total 12 60/72 (83.3%)* antigen was found in 66.7% of ASCUS and 70% of LSIL SIL (dysplasia) total 77 64/141 (45.4%)** cases, confirming that detection of p16/INK4 protein can be used as adjunct test especially at liquid-based Table 5. Comparative analysis of p16 and HE stain in parallel sections. cytology. However, several authors noticed a high degree of false p16 reactivity within otherwise negative * full agreement found in HSIL cases (per cent) ** number of HSIL from total SIL smears, especially in those containing atrophic cells [72]. The latter authors stressed that in such smears The diagnostic significance of p16 staining of the number of p16 reactive atypical cells per total cell cervical smears in comparison with HPV DNA testing number of cells may not represent a significant feature. will be discussed below. Here, the morphological This situation is unlikely for biopsy sections, where the criteria for ASC (atypical squamous cells) will be confluent distribution of p16 positive cells within the mentioned versus reactive squamous cell proliferation parabasal layer is essential for interpretation of the (no dysplastic epithelium cells as in LSIL and/or HSIL staining result. In smears the most important scoring cases). The Bethesda Committee (1991) definitions and criterion for each particular p16 reactive cell still remains classification were later improved and widely accepted the evaluation of nuclear changes, such as the altered [61]. Comparisons among laboratories showed that N/C ratio. According to our experience, a relatively the diagnosis of ASC-US (atypical squamous cells of high percentage of p16 reactive atypical cells could be undetermined significance) should not exceed 5-6% of found not only in LSIL, but also in smears classified the total number of smears investigated. As stressed by as ASCUS, if specimens were scored regardless

561 J. Rajčáni et al.

of the presence or absence of HPV DNA (Table 6). tested for HPV DNA presence, the p16 positive atypical There was little difference between the p16 reactivity cells in LSIL graded smears are more easily found as of atypical cells either in ASCUS and/or LSIL smears compared to the classical staining only (Figures 11 coming from women, in which the HPV DNA test was and 12). On the other hand, it became clear that p16 positive (Table 7, Figure 10). In the ASCUS smears reactive cells may be seen in the smears from HPV from women that were negative for HR HPV DNA in the negative patients, which were graded ASCUS and/ HC-2 test, the p16 reactive atypical cells still comprised or LSIL (Figure 13). As shown by the analysis of DNA a considerable proportion of investigated specimens, negative smears (Table 8), a relatively large proportion even though the use of a suitable staining kit reduced of false positive p16 reactive cells (about 37%) could be the number of false positive antigen-reactive cells still seen even when the ASCUS smears were stained (Table 8). Namely a significantly lower positive rate with the CYNTec histology kit. Therefore, we stress for p16 was noted in ASCUS smears, if the staining the need to test the alternatively taken specimens for procedure was performed with the CYNTec cytology HR-HPV DNA during the ASCUS/LSIL follow up. To kit (only about 40% of the tested smears showed p16 avoid confusing results of p16 staining especially in the reactive cells, Table 6). Our results with a relatively ASCUS smears (depending on the kit used) and due modest number of patients with DNA positive as well as to spontaneous healing of the HPV unrelated lesions, it p16 antigen positive smears showed that p16 staining seems reasonable to interpret the available PAP and/or of atypical squamous cells might be a helpful diagnostic p16 stained cytology along with the results of the HPV tool in the ASCUS/LSIL cases. If the patients were not DNA test.

p16 antigen HPV DNA test Cytological diagnosis Total positive negative positive negative ND

17 (8%) 92 (42%) ASCUS* 218 149 (68%) 69 (32%) 109 (50%) 15%*** 85%***

8 (10%) 27 (33%) ASCUS** 82 32 (39%) 50 (61%) 47 (57%) 23%*** 77%***

16 (26%) 28 (46%) LSIL 61 56 (92%) 5 (8%) 17 (28%) 36%*** 64%***

41 (11%) 147 (41%) Total 361 237 (66%) 124 (34%) 173 (48%) 22.5%*** 77.5%***

Table 6. Survey of p16 reactivity of atypical squamous cells in cervical smears.

* stained with the CYNTec histology kit ** stained with the CYNTec cytology kit (by 29% less p16 reactive cells) *** per cent from the total of HPV tests done

Cytological p16 antigen Cytological p16 antigen Total Total diagnosis positive negative diagnosis positive negative

ASCUS* 17 15 2 ASCUS* 92 67 (73%) 25 (27%)

ASCUS** 8 5 3 ASCUS** 27 10 (37%) 17 (63%)

ASCUS total*** 25* 20 (80%) 5 (20%) ASCUS total*** 119 77 (65.4%) 42 (35%)

LSIL 16 13 (81%) 3 (19%)**** LSIL 28 26 (93%) 2 (7%)

Total 41 33 (81.5%) 8 (19.5%) Total 147 103 44

Table 7. The p16 reactivity of atypical squamous cells in HPV DNA Table 8. Results of p16 antigen staining in HPV DNA negative positive smears. smears.

* stained with the CYNTec histology kit * stained with the CYNTec histology kit ** stained with the CYNTec cytology kit ** stained with the CYNTec cytology kit (by 36% less cases revealing *** sum of both ASCUS groups (compare Table 6) p16 reactive cells) **** improvement within an interval of 4-9 months *** sum of both ASCUS groups (compare Table 6)

562 Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

Figure 10. A group of atypical (dysplastic) squamous cells reactive Figure 11. Single atypical squamous cells reactive for p16 antigen for p16 antigen in a cervical smear classified LSIL, the in a cervical smear classified as LSIL (the N/C ratio is test for HR-HPV DNA was positive (magn. x 260). clearly altered)` HPV DNA was not tested (magn. x480).

Figure 12. Small atypical cells with enlarged hyperchromic Figure 13. A cluster of atypical metaplastic cells and single nuclei are reactive for p16 antigen in a cervical smear atypical squamous cells showing slightly enlarged classified as LSIL; the HPV DNA was not tested nuclei reactive for p16 antigen (magn. x300), no HR- (magn. x300). HPV DNA was found.

3. HPV DNA detection as compared conserved regions located either on L1 and/or E1 ORFs, to the scoring of morphological enabling detection of a wide spectrum of genotypes [75]. In resulted that the GPs annealed not only to ORFs lesions for which they had been designed, but also to some Since HPV is extremely difficult to culture [7], the most ORFs for which genotypes had not been sequenced feasible diagnostic approach is to prove the presence at that time (later on identified as HPV 13, 30, 31, 45 of HPV DNA by molecular techniques. In general, these and 51). The consensus or general primer GP5+/ are based on DNA/DNA hybridization (using labeled GP6+ based procedure became widely used, since it complementary DNA probes), on amplification of vDNA enabled the differentiation between several HR and LR by polymerase chain reaction (PCR) and/or on DNA/ HPV genotypes in a single assay [76,77]. Originally, the RNA hybridization (using complementary RNA probes) bands were visualized by Southern blot hybridization. followed by the amplification of labeled hybrid signal. The GP-PCR technique became further improved in Early approaches utilized various modifications of the order to detect more genotypes (at least 14 HR HPV vDNA to DNA probe hybridization tests, such as in situ along with the 6 frequent LR HPV genotypes). In hybridization and/or various blotting techniques [73]. To addition it was modified to visualize the reaction product improve the PCR method, multiple primers for L1 and/ by IEA [78]. For latter purposes, one of the primers was or E1 gene amplifications have been introduced, aiming biotinylated and the amplified mix was captured to the to amplify the most frequent genotypes (i.e. HPV 6, 11, wall of streptavidin-coated microwells. After incubation 16 and 18) in a single tube reaction [74]. The principle of micro-wells with the amplified product, following of multiple genotype amplification was further modified washing and denaturing, a hybridization cocktail of by using general primers (GPs), which flank the strongly digoxigenin-labeled internal oligonucleotide probes was

563 J. Rajčáni et al.

added to the wells in order to detect the bound hybrid predictive value of screening in the moderate as well using immunohistochemical reagents. The high risk as severe dysplasias to 77.7% and 96% respectively. types assessed this way were, as a rule, HPV 16, 18, The authors concluded that testing of HPV DNA by HC 31, 33, 39, 45, 51, 52, 56, 58, 59, 66 and 68, while the method is helpful for clinical diagnoses. Later on, the HC low risk ones were at least HPV 6, 11, 40, 42, 43 and 44. test was improved by introducing a second generation Further improvements of GPs allowed their annealing to HC protocol improving both the sensitivity as well as the the DNA of additional genotypes, such as HPV 26, 30, spectrum of RNA probes [85]. The recent HC2 based 53, 70, 73, 82 and 83, in order to increase the number HPV DNA test (Digene), which is used in our laboratory, of routinely detectable genotypes to 27 out of the 40 detects at least 14 HR genotypes (HPV 16, 18, 31, 33, possible mucosal human papillomavirus types [79]. 35, 35, 39, 45, 51, 52, 56, 58 and 68) at a sensitivity of Additional consensus primers, having been introduced 1-2 pg/ml (i.e. 1000 HPV DNA copies). This assay can for L1 ORF amplification, were MY09 and MY11. Their be also used for detection of at least 5 LR genotypes use was to be further modified to avoid synthesis of (HPV 6, 11, 42, 43 and 44). irreproducible fragments [80]. The latter primer set The HC2 test was found nearly as sensitive as (PGMY07/11) increased the proportion of detectable the GP5+/GP6+ primer directed PCR/EIA test. When multiple HPV infections enabling the detection of rare HPV DNA samples were collected for histological genotypes such as HPV 26, 35, 42, 45, 52, 54, 59, 66 diagnosis of severe dysplasia (CINII/CINIII/HSIL) and 73. The routine HR versus LR HPV testing may before conization and/or loop diathermy as well as 4 to be influenced by DNA extraction techniques, namely 9 months afterwards, a satisfactorily high substantial regarding manual versus automated extraction protocols agreement for both tests was noted, i.e. k=0.70 before [81]. To avoid the methodic variations, automated vDNA and k=0.72 after the treatment. This suggests that both extraction was recommended for HPV genotyping for tests are almost equal [86]. When in cervical smears both, the classical PCR (GP5/GP6 primers) as well the approved HC2 test and the nested PCR utilizing the as for the real time PCR based quantitative TaqMan MY/GP primers have been compared, some authors assay. An agreement was found when comparing both found a higher sensitivity (96.6% versus 58.6%) in techniques in nearly 300 specimens [82]. favor of the HC2 test [87]. Either the HC2 (Digene) Further modification of HPV detection in the test or the commercial PCR test were recommended direction of immunochemistry has resulted into an assay in addition to cytological testing in order to enhance omitting vDNA amplification, while introducing signal the probability of precise assessment as well as to amplification instead. In a latter assay, the denatured increase the screening efficiency among the clinically vDNA was hybridized under high stringency conditions healthy (symptomless) patients enrolled for screening to single stranded RNA probes either for LR genotypes [88]. Under the conditions of Saudi Arabia, the HC2 (at least 6, 11, 42, 43 and 44) and/or for HR genotypes detected HR HPV in 5 and LR-HPV in 1 patient only (at least 16, 18, 31, 33, 35, 45, 51, 52 and 56). The (6% of the total volunteers tested), while PCR detected RNA/DNA hybrid complex was then bound to micro- vDNA in 4 patients (4% of the total tested). When 148 wells (or tubes) coated with an alkaline phosphatase patients with mild dyskaryosis in their Pap smears conjugated monoclonal antibody, able to capture the (ASCUS/LSIL) were enrolled for HPV detection by specific RNA/DNA hybrid [83]. The reaction is then the HC2 test and by two modifications of the PCR test visualized by addition of the chemilumiscent substrate, (either PGMY09/PGMY011 primers and/or CPI/CPIIG in which emission light is amplified and measured in a primers), the HR-HPV by HC2 test was detected in luminometer; the results are expressed in relative light 32 out of 38 samples positive in either alternative test units (RLU). This method is referred to as hybrid capture (25.6% out of the total tested). The letter percentage (HC). At its beginning HC showed lower sensitivity, reflects how many patients were positive for at least one since only 50.000 vDNA copies could be detected as HPV DNA test, since HC2 results can remain negative compared to as few as 10-100 vDNA copies (about in cases of double infection and/or in cases of infection 100 fg HPV DNA) detectable by the more sensitive with unusual genotypes such as HPV 44, 53, 68 and 73 PCR. Recio et al. [84] used the first generation HC test types detectable by the PCR based tests only [89]. Mo for investigating the HPV DNA presence in patients with et al. [90] reported that the HC2 test as compared in ASCUS, LSIL, HSIL and carcinoma in situ smears, the 470 patients with the PCR-based Roche Amplicor HPV latter being used as relevant standard. Altogether 44% test, reached and agreement of 96.2%; only 18 cases of patients were tested positive, mainly for HR HPV were found discordant. It should be mentioned that the genotypes. The combination of LSIL/HSIL cytology and Amplicor HPV test identifies the PCR amplified vDNA HPV HC testing increased the sensitivity but not the by means of 13 HR-HPV genotype probes, and that the

564 Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

vDNA is being obtained from cervical cells collected (when using a cut off value of 2 pg/ml) as compared into a transport medium [91]. When compared with the with cytology, since the probability of spontaneous HC2 test, all HC2-positive, but PCR-negative cases regression was lower in the latter age group. In any were found to harbor HPV 53. On the other hand, the case, a combined and/or repeated cytology testing is HC2-negative but PCR-positive cases were found needed, if the HPV DNA is tested first. On the other multiple infected with HPV 6, 18 and 56. In patients hand, any HPV DNA test in women showing ASCUS/ revealing ASCUS smears both tests for HPV DNA LSIL smears was clinically useful [95]. As shown in showed a positive rate of 42.3%, while in patients over 4000 women, both the PAP cytology testing as showing LSIL smears the HPV DNA positive rate was well as the HPV DNA test (regardless of kind used) 66.3 and/or 66.8%, respectively, depending on the test were negative in 65% of subjects. Normal cytology, but performed (the PCR based test seemed in this case positive (at least one) DNA tests were found in 17% even less sensitive). In patients with HSIL smears, of patients. A similar proportion of women had positive the DNA was positive in both tests at the highest rate ASCUS/LSIL cytology (18%). The ASCUS/LSIL of 92.8%. In a smaller panel of 88 women the HC2 positive women when tested for HPV DNA fell into two (Digene) test was compared with the HPV consensus groups, namely those which were positive by at least primer kit (Argene), with the Roche Amplicor test and one DNA test (9%) and those who were DNA negative with the classical nested PCR (all amplifying the L1 (9%). The significantly higher proportion of CIN III/HSIL ORF); the letter test had been included as reference cases (58 out of 298 repeatedly investigated women, standard [92]. The authors in question claimed that the i.e. 19%) originated from the group of women, who had HC2 and the Amplicor kit results had been most closely been screened positive by PAP cytology as well as for related and that a complete agreement (positive in all HPV DNA. Patients who had been HPV DNA positive, tests) could be found in 75% of the specimens. They but were found negative by cytology, later developed tried to explain these disagreements by the relatively pre-invasive cancer in 19 cases only (8% out of the small virus load within the cervical fluids of chosen 238 repeatedly tested women). The ASCUS/LSIL patients. positive patients with HPV negative results developed CIN III/HSIL in 7 retested cases (the lowest positive rate of 3% within the retested group in question). Taken 4. The predictive value of combined together, the probability of developing precancerous HPV DNA and PAP smear testing (HSIL) lesions was high in women who were PAP and the development of cervical as well as DNA positive; a medium probability for carcinoma developing cancer was found in HPV DNA positive, but by cytology negative women. The lowest probability A large cohort of woman (about 25,000) with negative was noted in HPV negative but ASCUS/LSIL positive PAP smears (Thin-Prep imaging system) was enrolled cases. As stressed by Mandelblatt et al. [96], screening for HPV DNA testing over a period of 2 years [93]. The of HPV plus patients with PAP tests within two years highest proportion of HPV positives (average positive appears to save lives and is more reasonable than rate 1.9%) was found in women younger than 30 years performing alone. Another large cohort study and then sharply declined (from 2.8% to 1.7%). These in U.S. (performed from 2003 to 2005) found that data confirm the great diagnostic value of follow up women aged less than 30 who have ASCUS grade by PAP screening, even when the HPV DNA test had smears showed HPV positive rate at 53% [97]. On been done first. The New Technologies for Cervical the other hand, women older than 30 years with NIL Cancer Screening Study [94] compared the results of PAP test were HPV (HC2) positive at a rate of 9%. repeated screening with liquid-based cytology alone It comes from these results as well as our modest and the combined screening with the HC2 test (with the experience (presented in Tables 6-8) that women sensitivity of 1-2 pg/ml) and by cytology in a similarly receiving cervical screening by cytology should be large group of nearly 25.000 women. Those younger examined also by HC2 (or other suitable) HPV DNA than 35 years were retested within 1 year. Repeated test. Since ASCUS smear diagnosis is a clinical and testing showed that among women aged 25-34 years prognostic challenge, it should be combined with HPV the informative value of combined HPV and cytology testing and repeatedly investigated to show whether was higher, since there was a frequent regression of or not the transition to HSIL occurs [98]. Also in this even grade CIN II dysplasia developed in HPV positive follow up, a relatively low proportion (6.7%) of ASCUS women. In the age group over 35 years, there was a positive patients developed HSIL or cancer. Among the substantial gain for the informative value of HPV testing patients with HSIL smears, up to 98% was found HPV

565 J. Rajčáni et al.

DNA positive; theoretically such women harbor the 5. Conclusions integrated incomplete genome in the cervical tissue. In the study in question, it was concluded that the residual Overwhelming literature (over 2000 reports) confirmed specimens collected from routine cervical cytology in the strong association between HR-HPV infection and ASCUS cases, could provide additional information the development of the precancerous lesions of cervical about the HPV DNA status that is of substantial help epithelium (such as CINIII/HSIL, in situ carcinoma) and in identifying those patients, who are likely to develop their transition into cervical carcinoma. Out of various HSIL, if they later test positive for HPV DNA. screening approaches, most effective and less expensive The aim of this paper was to confirm whether p16/ is the cytological examination. Classical PAP staining INK4a antigen staining is a biomarker for transforming has been improved by introducing thin layer liquid HPV infections that could act as an adjunct to current cytology and by application of immunohistochemistry, cytological and histological assessment of cervical especially p16/INK4A antigen staining. However, the smears and biopsies. In addition to above presented detection of p16 antigen showed reliable results in considerations, most recent reports conclude that histological sections only. In contrast, a great proportion p16/INK4A seems to be a sensitive biomarker of (about 20-30%) of p16 antigen stained smears which the high grade cervical intraepithelial neoplasia and were graded ASCUS/LSIL, may yield uncertain results. cancer, especially in cervical biopsies [99]. According Therefore, the best strategy is to test for HPV DNA to our results, which are in agreement with others, in all women revealing ASCUS/LSIL graded smears, p16 staining of cervical smears may show uncertain which stained positive for p16 antigen. Several authors results, since false positive reactivity of the atypical recommend concomitant examination by HPV DNA squamous cells concerns about 20-25% of ASCUS test of the rest of cells submitted for smear preparation smears and depends on the kit used. Although there and/or repeated sampling for HSV DNA within a short is good evidence that p16/INK4A immunostaining interval thereafter (but not later than 4-6 months). The correlates with the severity of cytological/histological result of the HPV DNA test alone and/or if this test is abnormalities, the reproducibility is limited also due completed first, has a relatively low predictive value, to insufficiently standardized interpretation of p16 especially in women younger than 30-35 years, since staining results [100]. In future, a better consensus a large proportion of HPV-positive cases may become should to be reached for evaluation and interpretation DNA negative. If the HPV DNA is not cleared within 6-12 of the p16/INK4a antigen stain in cervical smears. The months following the first DNA test, preparation of the presence of p16 antigen in dysplastic versus atypical PAP smear along with p16 antigen staining is strongly squamaous cells needs to be assessed regarding recommended. The assessment of morphological the various clinical and morphological settings, status along with the clinical signs is essential for further each addressing specific diagnostic questions. The therapy and/or for the decision of biopsy examination. authors of this review are convinced that satisfactory The combined approach, i.e. cytology as well HPV DNA consensus was already achieved for the evaluation of testing can effectively assess the current state and histological lesions in biopsy material. The assessment predict the probability of transition to HSIL/CINIII grade of p16 staining results within the cell smears waits for dysplasia helping to avoid unnecessary radicalism in the additional large-scale studies. case of false positive diagnosis.

References

[1] Bonez W., Papillomaviruses, In: Richman V., [4] deVilliers E.M., Fauquet C.L., Brocker T., Whitley D., Whitley R., (Eds.), Clinical Virology, Bernard H.U., zur Hausen H., Classification of 2nd Ed., ASM Press, Washington, 2002, 557- papillomaviruses, Virology, 2004, 324, 17-27 596 [5] Rizk R.Z., Chjristensen N.D., Michael K.M., Miller M., [2] Howley P.A., Lowy D.R., Papillomaviruses, In: Sehr P., Waterboer T., et al., Reactivity pattern of 92 Knipe D.M., Howley P.M., (Eds.), Field´s Virology, monoclonal antibodies with 16 human papillomavirus 5th Ed., Walters Kluver/Lippincott, Wiliams and types, J. Gen. Virol., 2008, 89, 117-129 Wilkins, Philadelphia, 2007, 2299-2354 [6] Dourbar J., The papillomavirus life cycle, J. Clin. [3] Fauthquet V., Mayo C.M., Maniloff M.A.J., Virol., 2005, 32, S7-S15 Desselberger U., Ball L.A., Family Papillomaviridae, [7] Meyers C., Laimins L.A., In vitro systems for the Virus Taxonomy, 8th Report of the ICTV, Elsevier- study and propagation of human papillomaviruses, Academic Press, 2005, 239-249 Curr. Top. Microbiol. Immunol., 1994, 186, 199-215

566 Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

[8] Phelps W.C., Yee C.L., Munger K., Howley P.M., [19] White A., Livanos F.M., Tlsty T.D., Differential The human papillomavirus type 16 E7 gene disruption of genomic integrity and cell cycle encodes transactivation and transformation regulation in normal human fibroblasts by the HPV functions similar to those of adenovirus E1A, Cell, oncoproteins, Genes Dev., 1994, 8, 666-677 1988, 53, 539-547 [20] Kulmala S.M., Syrjanen S.M., Gyllensten U.B., [9] Jeon S., Lambert P.F., Integration of HPV 16 Shabalova I.P., Petrovichev N., Syrjanen K.J., et al., DNA into the human genome leads to increased Early integration of high copy HPV 16 detectable in stability of E6/E7 mRNAs: implications for cervical women with normal and low grade cervical cytology, carcinogenesis, Proc. Natl. Acad. Sci. USA, 1995, J. Clin. Pathol., 2006, 59, 513-517 92, 1654-1658 [21] Martin P., Vass W., Schiller J.T., The bovine [10] Cullen A., Reid R., Campion M., Lorincz A.T., papillomavirus E5 transforming protein can Analysis of the physical state of different human stimulate the transforming activity of EGF and papillomavirus DNA in epithelial and invasive CSF-1 receptors, Cell, 1989, 59, 21-23 cervical neoplasm, J. Virol., 1991, 65, 606-612 [22] Petti L., DiMayo D., Activation of the platelet derived [11] Hopman A.H., Smedts F., Dignef W., Ammelen M., growth factor receptor by the bovine papillomavirus Sonke G., Mravunac M., et al., Transition of high- E5 protein, E.M.B.O. J., 1991, 40, 845-855 grade cervical intraepithelial neoplasia in micro- [23] Petti L., DiMayo G., Stable association between invasive neoplasia to microinvasive carcinoma the bovine papillomavirus E5 transforming protein is characterized by integration of HPV 16/18 and and the b receptor for platelet derived growth factor numerical chromosome abnormalities, J. Pathol., in stably transformed and acutely transfected cells, 2004, 202, 23-33 J. Virol., 1994, 74, 3582-3592 [12] Dourbar J., Ely S., Sterling J., McLean C., [24] Dyson N., Howley P.M., Munger K., Harlow E., The Crawford L., Specific interaction between HPV human papillomavirus 16 E7 oncoprotein is able 16 E1-E4 and cytokeratins results in collapse of to bind the retinoblastoma gene product, Science, the epithelial cell intermediate filament network, 1989, 243, 934-937 Nature, 1991, 352, 824-827 [25] Werness B.A., Levine A.J., Howley P.M., [13] Roberts S., Ashmole I., Johnson G.D., Kreider J.W., Association of human papillomavirus types 16 and Gallimore P.H., Cutaneous and mucosal human 18 E6 proteins with p53, Science, 1990, 248, 76-79 papillomavirus E4 proteins form intermediate [26] Scheffner M., Huibregtse J.M., Vierstra R.D., filament likne structures in epithelial cells, Virology, Yoshihara M., Sumiri K., Kayiyama G., et al., 197, 1993, 176-187 The HPV-16 and E6-AP complex functions as [14] Rogel-Gaillard C., Pehau-Arnaudet G., Breitbund F., a ubiquitin-protein ligase in the ubiquitinations of Orth G., Cytopathic effect in human papillomavirus p53, Cell, 1993, 75, 495-499 type 1-induced inclusion: in vitro analysis of the [27] Hollingworth R.E., Lee W.H., Tumor suppression contribution of two forms of the viral E4 protein, J. genes: new prospects for cancer research, J. Natl. Invest. Dermatol., 1993, 101, 843-851 Cancer Inst., 1991, 83, 91-96 [15] Flores E.R., Lambert P.F., Evidence for switch in [28] Golijow C.D., Abba M.C., Mouron S.A., Gomez the mode of human papillomavirus DNA replication M.A., Doulot F.N., C-myc gene amplification during the viral life cycle, J. Virol., 1997, 71, 7167- detected in preinvasive intraepithelial lesions, Int. 7179 J. Gynecol. Cancer, 2001 11, 492-465 [16] Meyers C., Laimins L.A., In vitro systems for the [29] Li Y., Nichols M.A., Shay J.W., Xiong Y., study and propagation of human papillomaviruses, Transcriptional repression of the D type cyclin Curr. Top. Microbiol. Immunol., 1994, 186, 199-215 dependent kinase p16 by retinoblastoma [17] Choo K.-B., Pan C.-C., Han S.-H., Integration of susceptibility gene product pRb, Cancer Res., human papillomavirus type 16 into cellular DNA 1994, 54, 6078-6082 of cervical carcinoma. Preferential deletion of [30] Kalantari M., Karlsen F., Kritensen G., Holm R., the E2 gene and invariable retention of the long Hagmar B., Johansson B., et al., Disruption of the control region and the E6/E7 open reading frames, E1 and E2 reading frames of HPV16 in cervical Virology, 1987, 161, 259-261 carcinoma is associated with poor prognosis, Int. [18] Matsukura T., Kol S., Sugase M., Both episomal J. Gynecol. Pathol., 1998, 17, 146-153 and integrated forms of human papillomavirus [31] Klaes R., Friedrich T., Spitkovski D., Ridder R., type 16 are involved in invasive cervical cancers, Rudy W., Petri K., et al., Over expression of p16/ Virology, 1989, 172, 63-72 INK4A as a specific marker for dysplastic and

567 J. Rajčáni et al.

neoplastic epithelial cells of the uteri, Int. J. [43] Rajčáni J., Adamkov M., Hybenová J., Cancer, 2001, 92, 276-284 Moraveková E., Lauko L., Benčat M., Detection of [32] Murphy N., Ring M., Killalea A.G., Uhlmann V., regulatory p16/INK4A protein in dysplastic cerical O’Donovan M., Mulcahy F., Cervical dyskaryosis: squamous epithelium as a prognostic tool in CIN and cGIN in cervical biopsies and ThinPrep carcinoma prevention, CS Pathol. Soudn. Lékař., smears, J. Clin. Pathol., 2003, 56, 56-63 2009, 4, 101-107, (in Slovak). [33] Murphy N., Ring M., Heffron C.C., King B., [44] Mulvany N.J., Allen D.G., Wilson S.M., Diagnostic Killalea A.G., Hughes C., et al., p16/INK4a, CDC6 utility of p16INK4A: a reappraisal of its use in and MCM5: predictive biomarkers in cervical cervical biopsies, Pathology, 2008, 40, 335-344 preinvasive neoplasia and cervical cancer, J. Clin. [45] Meisels A,. Fortin R., Condylomatous lesions of Pathol., 2005, 58, 525-534 the cervix and vagina. I. Cytologic patterns, Acta [34] O´Neill C.J., McCluggage W.G., p16 expression in Cytol., 1976, 20, 505-809 the female genital tract and its value in diagnosis, [46] Meisels A., Fortin R., Roy M., Condylomatous Adv. Clin. Pathol., 2006, 13, 8-15 lesions of the cervix II. Cytologic, colposcopic and [35] Wright T.C., Kurman R.J., Ferenczy A., histopathologic study, Acta Cytol., 1977, 21, 379-390 Precancerous lesions of the cervix, In: Kurman R.J., [47] Purola E., Savia E., Cytology of gynecologic (Ed.), Blaustein´s Patology of the Female Genital condyloma acuminatum, Acta Cytol., 1977, 21, Tract, 5th Ed., Springer, New York/Berlin, 2001, 26-31 253-321 [48] Meisels A., Morin C., Casas-Cordero M., Human [36] National Cancer Institute (NCI) Workshop, The papillomavirus infection of the uterine cervix, Int. J. 1988 Bethesda system for reporting cervical/ Gynecol. Pathol., 1982, 1, 75-94 vaginal cytologic diagnoses, JAMA, 1989, 262, [49] Nagai N., Molecular biologic study on the 931-934 cancerogenesis of HPV uterine cervical cancer [37] Luff R.D., The Bethesda System for reporting and related lesions – analysis of HPV types 18, 18 cervical vaginal cytologic diagnoses, Report of the E6/E7 gene mRNA, Nippon Sanak Fujinka Gakkai 1991 Bethesda Workshop, Hum. Pathol., 1992, 23, Zasshi, 1990, 42, 823-833, (in Japanese) 719-721 [50] Koutsky I.A., Holmes K.K., Critchlow C.W., [38] Jacobs A., Kristensen P.B., Poulsen H.K., Flow Stevens C.E., Paavonen J., Beckmann A.M., et al., cytometric classification of biopsy specimens from A cohort study of the risk of cervical intraepithelial cervical intraepithelial neoplasia, Cytometry, 1983, neoplasia grade 2 and 3 in relation to papillomavirus 4, 166-169 infection, New Engl. J. Med., 1992, 327, 1272-1278 [39] Wright T.C., Kurman R.J., A critical review of the [51] Ostor A.G., Natural history of cervical intraepithelial morphologic classification systems of preinvasive neoplasia: a critical review, Int. J. Gynecol. Pathol., lesions of the cervix: the scientific basis of the 1993, 2, 186-192 paradigm, Papillomavirus Rep., 1994, 5, 175-181 [52] Kadish A.S., Ho G.Y.F., Burk R.D., Wang Y., [40] Dray M., Russel P., Wallman N., Leong A., Carter Romney S.L., Ledwige R., et al., Lymphoproliferative J., Cheerala B., p16(INK4a) as a complementary responses to human papillomavirus HPV type 16 marker of high grade intraepithelial lesions of the proteins E6/E7: outcome of HPV infection and uterine cervix. I. Experience with squamous lesions associated neoplasia, J. Natl. Cancer Inst., 1997, in 189 consecutive cervical biopsies, Pathology, 89, 1285-1291 2005, 37, 112-124 [53] Rodriguez A.C., Schiffman M., Herrero R., [41] Ikeda K., Tate G., Suzuki T., Mitsuya T., Coordinate Wacholder S., Hildesheim A., Castle P.E., et expression of cytokeratin 8 and cytokeratin al., Rapid clearance of human papillomavirus 17 immunohistochemical staining in cervical and implications for clinical focus on persistent intraepithelial neoplasia and cervical squamous infections, J. Natl. Cancer Inst., 2008, 100, 513- cell carcinoma: an immunohistochemical analysis 517 and review of the literature, Gynecologic Oncol., [54] Varnai A.D., Bollman M., Bankfalvi A., Speich N., 2008, 108, 598-602 Schmitt C., Griefingholt H., et al., Predictive [42] Yildiz I.Z., Usubutun A., Firat P., Kucukali T., testing of cervical pre-cancer by detecting human Efficiency of immunohistological p16 expression papillomavirus E6/E7 mRNA in cervical cytologies and HPV typing in cervical squamous intraepithelial up to high grade squamous intraepithelial lesions: lesion grading and review of p16 literature, Pathol. diagnostic and prognostic implications, Oncol. Res. Pract., 2007, 203, 445-449 Rep., 2008, 19, 457-465

568 Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

[55] Kruse A.J., Baak J.P.A., deBruin P.C., [66] Abati A., Jaffurs W., Wilder A.M., Squamous atypia Jiwa M., Snijders W.P., Fons G., et al. , Ki-67 in the atrophic cervical vaginal smear, Cancer immunoquantitation in cervical intraepithelial Cytopathol., 1998, 84, 218-225 neoplasia (CIN): a sensitive marker for grading, J. [67] Wright T.C., Cox J.T., Massad L.S., 2001 consensus Pathol., 2001, 196, 48-54 guidelines for the management of women with [56] Kruse A.J., Skaland I., Janssen E.A., Buhr- cervical cytological abnormalities, JAMA, 2002, Wildhagen S., Klos J., Arends M.J., et al., 287, 2120-2129 Quantitative molecular parameters to identify low [68] Cenda M.T., Demir N., Sezer O., Doganay L., and high risk early CIN lesions: role of markers Ortac R., Clinical results of the liquid cytology of proliferative activity and differentiation and Rb tool, liquid PREP, in comparison with conventional availability, Int. J. Gynecol. Pathol., 2004, 23, smears of detection of squamous abnormalities, 100-109 Asian Pac. J. Cancer Prev., 2009, 10, 399-402 [57] Kruse A.J., Baak J.P.A, Janssen E.A., Bol M.G., [69] Nieh S., Chen Su-F., Chu T.-Y., Lai H.-C.H., Kjellerod K.H., Fianne B., et al., Low and high risk Fu E., Expression of 16INK4A in Papanicolaou CIN I and II lesions: prospective predictive value of smears containing atypical squamous cells of grade, HPV and Ki-67, J. Pathol., 2003, 199, 462- undetermined significance from uterus cervix, 470 Gynec. Oncol., 2003, 91, 201-208 [58] Bean S.M., Eltoum I., Whitlow L., Chien D.C., [70] Guo M., Hu L., Baliga M., He Z., Hughson M.D., Immunohistochemical expression of p16 and Ki- The predictive value of p16INK4 and hybrid 67 correlates with degree of anal intraepithelial capture 2 human papillomavirus testing hot high neoplasia, Am. J. Surg. Pathol., 2007, 31, 555-561 grade cervical intraepithelial neoplasia, Am. J. Clin [59] Waltz A.E., Lachago J., Bose S., p16 and Ki- Pathol., 2004, 122, 894-901 67 immunostaining is a useful adjunct in the [71] Shin E.K., Lee S.R., Kim M.K., Kang E.J., Ju W., assessment of biopsies for HPV associated anal Lee Sh.N., et al., Immunocytochemical staining intraepithelial neoplasia, Am. J. Surg. Pathol., of p16ink4a protein as an adjunct test in equivocal 2006, 30, 795-801 liquid base cytology, Diagn. Cytopathol., 2008, 36, [60] Keating J.T., Cviko A., Rietdorf S., Quade M., 311-319 Sun D., Duensing S., et al., K-67, cyclin E and p16/ [72] Duncan L., Jacob S., Hubbard E., Evaluation of INK4 are complementary surrogate biomarkers for p16INK4a as a diagnostic tool in the triage of Pap human papillomavirus-related cervical neoplasia, smears demonstrating atypical squamous cells of Am. J. Surg Pathol., 2001, 25, 884-891 undetermined significance, Cancer Cytopathol., [61] Solomon D., Thomas P.A., Lenel J.C., Moriarti A., 2008, 114, 34-48 O’Connor D., Prey M., et al., The 2001 Bethesda [73] Rando F.F., Nucleic acid hybridization as a diagnostic system terminology for reporting results of cervical tool fot the detection of human papillomaviruses, cytology, JAMA, 2002, 287, 2114-2119 Adv. Exp. Med. Biol., 1990, 263, 89-109 [62] Geisinger K.R., Raab S.S., Stanley M.W., [74] Anceschi M.M., Falcinelli C., Pieretti M., Cosmi E.V., Silverman J.J., Abati A., In: Modern , Multiple primer pairs polymerase chain reaction for Churchill-Livingstone, New York, 2004, 83-196 the detection of human papillomavirus types, J. [63] Dressel D.M., Wilbour D.C., Atypical immature Virol. Meth., 1990, 28, 59-65 squamous metaplastic cells in cervical smears: [75] van den Brule A.J., Snijders P.J., Gordijn R.L., association with high grade squamous Becker O.P., Maijer C.J., Walboomers J.M., intraepithelial lesions and carcinoma of the cervix, General primer-mediated polymerase chain Acta Cytol., 1992, 36, 630-631 reaction permits the detection of sequenced and [64] Hatem F., Wilbur D.C., High-grade squamous still unsequenced human papillomavirus genotypes cervical lesions following negative Papanicoulaou in cervical scrapes and carcinomas, Int. J. Cancer, smears: false negative cervical cytology or rapid 1990, 45, 644-649 progression, Diagn. Cytopathol., 1995, 12, 135- [76] deRoda-Husman A.M., Walboomers J. M.M., van 141 den Brule A.J.C., Meijer C.J.L.M., Snijders P.J.F., [65] Gupta D.K., Komaromi-Hiller G., Raab S.S., The use of general primers GP5 and GP6 elongated Interobserver and intraobserver variability in at their 3` endswith adjacent highly conserved the cytologic diagnosis of normal and abnormal sequences improves human papillomavirus metaplastic squamous cells in Pap smears, Acta detection by polymerase chain reaction, J. Gen. Cytol., 2001, 45, 697-703 Virol., 1995, 76, 1057-1062

569 J. Rajčáni et al.

[77] Jacobs M.V., de Roda Husman A.M., van den assay for the detection of human papillomavirus Brule A.J.C., Snijders P.J.F., Walboomers J.M.M., in cervical scrapings, Med. J. Malaysia, 2007, 62, Group specific differentiation between high and 206-209 low risk human papillomavirus genotypes by [88] Gazaz F.B., Molecular testing of human general primer-mediated PCR and two cocktails papillomavirus in cervical specimens, Saudi Med. of oligonucleotide probes, J. Clin. Microbiol., 1995, J., 2007, 28, 1810-1818 33, 901-905 [89] Salimovic-Besic I., Evaluation of hybrid capture [78] Jacobs M.W., Snijders P.J.F., van den Brule 2 HPV DNA test and two variants of polymerase A.J.C., Helmhorst T.J.M., Meijer C.J.L.M, chain reaction (PCR-PGMY11/09 and PCR-CPI/ Walboomers J.M.M., A general primer GP5+/GP6+- CPIIG) according to HPV types, Med. Arh., 2007, mediated PCR enzyme immunoassay method for 61, 135-137 rapid detection of 14 high risk and 6 low risk human [90] Mo L.Z., Monier-Benoit S., Kantelip B., Petitiean A., papillomavirus genotypes in cervical scrapings, J. Riethmuller D., Pretet J.L., et al., Comparision of Clin. Microbiol., 1997, 35, 791-795 Amplicor and hybrid capture II assays for high risk [79] Schmidt M., Dondog B., Waterboer T., Pawlita M., HPV detection in normal and abnormal liquid based Homogenous amplification of general human cytology: use of immuno-LIPA for genotyping alpha papillomavirus by PCR using a novel assay to screen the discordant results, J. Clin. broad-spectrum GP5+ and GP6+ primers, J. Clin. Virol., 2008, 41, 104-110 Microbiol., 2008, 46, 1050-1059 [91] Monsenogo J., Bohbot J.M., Pollini G., Krawec C., [80] Gravitt P.E., Peyton C.L., Alessi T.Q., Wheeler C.M., Vincent C., Merinargues I., et al., Performance of Coutlee F., Hildesheim A., et al., Improved the Roche AMPLICOR human papillomavirus test amplification of genital human papillomaviruses, J. in prediction of cervical intraepithelial neoplasia Clin. Microbiol., 2000, 38, 357-361 (CIN) in women with abnormal PAP smear, Gynec. [81] Oevestad I.T., Janssen E.A., Beak J.P., The effect Oncol., 2005, 99, 160-168 of different DNA isolation methods on the outcome [92] Hantz S., Caly H., Decroisette E., Dutrop A., of high risk HPV testing, Indian J. Patol. Microbiol., Bakeland D., Pascal B., et al., Evaluation of 2007, 50, 733-739 accuracy of three assays for human papillomavirus [82] Broccolo F., Cocuzza C.E., Automated extraction detection and typing: hybrid capture 2, HPV and quantitation of oncogenic HPV genotypes consensus kit and Amplicor HPV, Pathol. Biol. from cervical samples by a real time PCR-based (Paris), 2008, 56, 29-35 system, J. Virol. Meth., 2008, 148, 48-57 [93] Bansai M., Austin R.N., Zhao C., High risk HPV DNA [83] Reid R., Lorincz A.T., Human papillomavirus tests, detected in less than 2% of over 25,000 cytology Bail. Clin. Obstet. Gynecol., 1995, 9, 65-103 negative imaged liquid based Pap test sample from [84] Recio F.O., Srivastava S., Wong C., Hempling R.E., women 30 and older, Gynecol. Oncol., 2009, 115, Eltabakh G.H., Piver M.S., The clinical value 257-261 of digene hybrid capture HPV DNA trstiny in a [94] Ronco G., Giorgi-Rossi P., Carozzi F., referral-based pc, Eur. J. Gynaecol. Oncol., 1998, Confortini M., Dalla Palma P., Deo Mistro A., 19, 203-208 et al. (44 contributors), Results and recruitment [85] Poljak M., Brencic A., Seme A., Vince A., Marin I.J., from a randomized controlled trial comparing Comparative evaluation of first and second human papillomavirus testing alone with generation of digene hybrid capture assays for conventional cytology as the primary cervical detection of human papillomaviruses associated cancer screening test, J. Natl. Cancer Inst., with high or intermediate risk for cervical cancer, J. 2008, 100, 492-501 Clin. Microbiol., 1999, 37, 796-797 [95] Kulasingan S.L., Hughes J.P., Kiviat N.B., Mao C., [86] Sonderlund-Strand A., Rymark P., Anderson P., Weiss N.S., Kuypers J.M., et al., Evaluation of Dillner J., Dillner L., Comparison between the human papillomavirus in primary screening for hybrid capture II test and a PCR-based human cervical abnormalities. Comparison of sensitivity, papillomavirus detection method for diagnosis and specificity and frequency referral, JAMA, 2002, posttreatment follow-up of cervical intraepithelial 288, 1749-1757 neoplasia, J. Clin. Microbiol., 2005, 3, 3260-3266 [96] Mandelblatt J.S., Lawrence W.F., Womack S.M., [87] Saini R., Shen T.H., Othamn N.H., Santhanan J., Jacobson D., Yi B., Hwang Y.T., et al., Benefits and Tang T.H., Evaluation of polymarease chain costs of using HPV testing to screen for cervical reaction (PCR) method and hybrid capture (HC2) cancer, JAMA, 2002, 287, 2372-2387

570 Diagnostic role of p16/INK4A protein in Human Papillomavirus (HPV) induced cervical dysplasia

[97] Datta S.D., Koutsky L.A., Ratelle S., Unger E.R., [100] Tsoumpou I., Arbyn M., Kyrgiou M., Wentzensen N., Shlay J., Weather B., et al., Human papillomavirus Koliopoulos G., Martin-Hirsch P., et al., The and cervical cytology in women screened for p16(INK4a) immunostaining in cytological and cervical cancer in the United States, 2003-2005, histological specimens from the uterine cervix: Ann. Intern. Med., 2008, 148, 557-559 a systematic review and meta-analysis, Cancer [98] Manos M.M., Kinney W.K., Hurley L.B., Treat. Rev., 2009, 35, 210-220 Sherman M.E., Shieh-Ngai J., Kurman R.J., et al., Identifying women with cervical neoplasia using human papillomavirus DNA testing and equivocal Papanicolaou results, JAMA, 1999, 281, 1605- 1610 [99] Haidopoulos D., Partsinevelos G.A., Vlachos G.D., Rodolakis A., Markaki S., Voulgaris Z., et al., The p16 INK4A is a strong biomarker for cervical intraepithelial neoplasia and invasive cervical carcinoma: a reappraisal, Reprod Sci., 2009, 16, 685-693

571