According to Bonnez [31].

**Genus Properties Species Genotype(s) Clinical significance**

116 Human Papillomavirus - Research in a Global Perspective

E5 ORF is HPV45, HPV59, different HPV68, HPV70,

*7* HPV18, HPV39, HR genotypes, mucosal lesions

*9* HPV16, HPV31, HR genotypes, mucosal lesions

*11* HPV34, HPV73 (dysplasia and carcinoma) *13* HPV54 LR genotype, mucosal lesions *15* HPV71 LR genotype, mucosal lesions

2 HPV9, HPV15, Commonly associated with EV,

HPV23,HPV37, immunosuppression,

3 HPV 49, HPV75, LR genotypes, benign cutaneous

4 Candidate types 60 Pre-malignant cutaneous 5 and 88 Pre-malignant cutaneous

2 HPV48 Cutaneous lesions 3 HPV50 Cutaneous lesions 4 HPV60 Cutaneous lesions

HPV38,HPV80 most frequently LR genotypes

HPV95 homogenous inclusion bodies

HPV43 mucosa

HPV33, HPV35, HPV52, HPV58,

HPV13, HPV44,

HPV67

Betapapillomavirus Latent 1 HPV5, HPV8, Mainly benign cutaneous lesions

HPV93,

HPV76

Gammapapillomavirus 1 HPV4, HPV65 Cutaneous lesions, intracytoplasmic

infection HPV12, HPV14 possible in HPV19, HPV20 general HPV21, HPV25 population HPV36, HPV 47,

*8* HPV7, HPV40, LR genotypes, butcher warts, skin and

*10* HPV6, HPV11, LR genotypes, rarely verrucous carcinoma

HPV73, HPV74 HR genotypes, mucosal (cervical) lesions

HPV17, HPV22, mucosal lesions and detection of vDNA in

**Table 6.** Clinical relevance of high-risk (in bold) and low-risk HPV genotypes according to their frequency.

The HPV virions are small non-enveloped capsids 55 nm in size and of icosaedral symmetry. They are composed of 72 subunits formed by 2 structural proteins (L1 and L2, L = late), which are synthesized at the late intervals of productive replication cycle. The viral double-stranded DNA (vDNA) has about 8 kilobase pairs (kbp); it encodes 7 or 8 non-structural polypeptides, designated E1–E8 (E = early). The transcription of viral mRNA is directed clockwise under the control of two promoters, namely the early promoter (P97) and the late promoter (P670) sequence. In between the initiation codon for the transcription of E1 polypeptide ORF (open reading frame) and the stop codon for the L2 capsid protein ORF, an approximately 1.1 kbp long control repeat (LCR) is situated that contains the origin of vDNA replication as well as several binding sites for the binding of viral E2 and/or E1 regulatory proteins (**Table 7**). In addition to the motifs for down-regulation of the productive replication cycle (on the basis of a feedback mechanism), the LCR (also referred to as the URR, upstream regulatory region) contains enhancer sites for attachment of cellular transcription cofactors promoting the binding of cellular RNA polymerase to accomplish viral mRNA synthesis.


**Table 7.** Basic functions of HPV-coded early (non-structural) polypeptides involved in the replication and latency.



1 Epidermal growth factor 2 Platelet-derived growth factor, \* Activator protein 1, \* Cellular ju-na-na \*\*\* Cyclindependent kinas.

**Table 8.** Basic functions of the HPV-coded early (non-structural) oncoproteins.

The HPV virions are small non-enveloped capsids 55 nm in size and of icosaedral symmetry. They are composed of 72 subunits formed by 2 structural proteins (L1 and L2, L = late), which are synthesized at the late intervals of productive replication cycle. The viral double-stranded DNA (vDNA) has about 8 kilobase pairs (kbp); it encodes 7 or 8 non-structural polypeptides, designated E1–E8 (E = early). The transcription of viral mRNA is directed clockwise under the control of two promoters, namely the early promoter (P97) and the late promoter (P670) sequence. In between the initiation codon for the transcription of E1 polypeptide ORF (open reading frame) and the stop codon for the L2 capsid protein ORF, an approximately 1.1 kbp long control repeat (LCR) is situated that contains the origin of vDNA replication as well as several binding sites for the binding of viral E2 and/or E1 regulatory proteins (**Table 7**). In addition to the motifs for down-regulation of the productive replication cycle (on the basis of a feedback mechanism), the LCR (also referred to as the URR, upstream regulatory region) contains enhancer sites for attachment of cellular transcription cofactors promoting the binding of

> Binds to the regulatory long control repeat (LCR/URR) upstream from the E6/E7 ORF promoters. Forms a heterodimer with E2, associates with H1 histones and with cyclins, especially with cyclin E.

capsid polypeptide (especially at productive virus replication). Induces apoptosis.

E4 Associates with the L1 capsid polypeptide and facilitates virion formation (during productive

**Table 7.** Basic functions of HPV-coded early (non-structural) polypeptides involved in the replication and latency.

E5 10 kDa Increases and prolongs the activities of receptors interacting with external growth factors such

signal transmission mediating proteins such as c-Jun\*\* and/or c-Fos.

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 (also in latency). Suppresses the expression of E6/E7 proteins and interacts with the L2

replication); co-localizes with cytokeratins being produced in the medium and upper spinous layers.

as EGF and/or PDGF2 by binding to their cytoplasmic domain. Inhibits the acidification of endosomes, binds to ATPase within the membranes of vacuoles and increases the stability of the engulfed EGF/EGFR complexes. Activates cellular transcription factors such as AP1\*

Binds to the pivotal cell division regulator p53 (cellular anti-onc protein), enhances its degradation by ubiqitination (most efficiently acting are the E2 from HPV16 and/or HPV18

Binds to the retinoblastoma (Rb) protein (its p107 and p130 forms) , which regulate cellular DNA transcription via binding or release of the transcription initiation factor E2F (in response to phosphorylation of the Rb associated *cdk* complex). Activates mainly cyclins A and E,

and

cellular RNA polymerase to accomplish viral mRNA synthesis.

**Protein MW Properties and function**

118 Human Papillomavirus - Research in a Global Perspective

Function unknown.

Long control repeat/upstream regulatory region.

**Properties and function** 

genotypes).

E1 68–76

E2 40–58

E3 10–17

1

kDa

kDa

kDa

**Protein Molecular weight**

E6 16–18 kDa

E7 10–14 kDa

The E2 polypeptide was initially described as a transcriptional activator [26] capable to initiate viral transcription through the E2 recognition elements located within early promoter (for HPV16, it is the above mentioned P97, while alternatively, for HPV18, it is the P105). As described below in more detail, tumorigenesis is mediated by an integrated HPV DNA fragment, encompassing the E6 and E7 genes exerting transformation activity (**Table 8**). During pro‐ ductive (vegetative) virus replication in human squamous epithelial cells, the essential activities displayed either by the HPV type 16 P97 or by the HPV type 18 P105 promoters can be repressed by the full-length E2 polypeptide, which binds to one of the four E2 binding sites upstream of either P97 or P105 [27]. The E2 protein also facilitates long-term persistence of HPV genome in host cells by episomal maintenance providing a mechanism ensuring the distribu‐ tion of viral genome within the dividing epithelium and its segregation into daughter cells [28, 29]. This can be achieved due to the association of E2 with the mitotic spindles, when it interacts with condensed mitotic chromatin and thereby ensures that the viral genome (which dupli‐ cates within the dividing host cell during cellular DNA synthesis) gets attached to nuclear envelope at anaphase and reforms during telophase. The duplication of viral genome during mitosis is accomplished by means of the cellular DNA polymerase, which becomes modified by another early HPV-coded protein, the El polypeptide. The E1 is required for both the initiation and elongation of viral DNA synthesis being accomplished by the cellular DNA polymerase complex by means of its ATPase and DNA helicase activities. The E2 can complex with E1 to strengthen its affinity for binding to the origin of vDNA replication. The papillo‐ mavirus E2 protein has several well-characterized regulatory functions affecting viral tran‐ scription, viral DNA replication as well as long-term plasmid maintenance. In contrast to the duplication of HPV DNA, which is the main strategy to ensure long-term persistence of viral genome, in the course of acute virus replications, the necessary point is to generate many genome copies which will be packaged into virions. Why that process occurs only in the terminally differentiated cells of the upper squamous epithelium layer and in the benign vegetative vDNA replication, is not known. The switch in question may involve the presence or absence of cellular cofactors expressed in the differentiating keratinocytes, i.e. only within the cytokeratin forming upper squamous epithelium layer. As described later, these cells, if previously proliferating papilloma cells. The mechanisms regulating the switch from plasmid maintenance to vDNA replication not known. This switch may envolve the presence or absence of cellular cofactors expressed in the differentiating keratinocytes, thus is occurs only within the cytokeratin forming upper squamous epithelium layer. As described later, these cells, if previously infected within the lower squamous epithelium layer, then show typical morphol‐

ogy, the so-called koilocytes. Clearly, the relatively increased levels of HPV proteins such as E1 or E2 (or their modifications) may change the appearance of terminally differentiating keratinocytes. One might anticipate that the vegetative DNA replication occurs bi-direction‐ ally, through a theta structure intermediate or by a rolling circle mode, which is the principle of vDNA replication in general. Finally, the virion assembly must take place in the nuclei of terminally differentiated keratinocytes, which also contributes to koilocyte formation. The nascent capsids might randomly attach to the HPV DNA, and are further stabilized by the formation of disulfide bonds between conserved cysteines on adjacent L1 monomers acquiring resistance to proteolytic digestion. Taken together, as result of productive replication, the virions form large aggregates within the nuclei of infected keratin forming upper squamous epithelium cells, which then regularly show koilocytosis [32], but rarely reveal cytoplasmic inclusion bodies (see **Table 5** for details).

The newly produced HPV virions do not appear outside of squamous epithelium cells, and its productive replication is closely bound to the suprabasal, mainly medium and upper slayer of dividing keratinocytes. At infection, the HPV virions preferentially bind to heparan sulfate proteoglycans (HSPGs) on the basement membrane or to the basal stem cells, which may be exposed to environmental influence at sites of epithelial trauma or permeabilization. The differentiating squamous epithelium cells cannot become infected. To achieve selective infection of basal differentiating squamous epithelium cells, the initiation of infection prefer‐ entially requires attachment to the basal stem cells layered at the basal membrane of stratified squamous epithelium [33–35]. The basement membrane-bound virion undergoes a conforma‐ tional change of its L2 capsid protein that exposes a highly conserved N-terminal peptide motif to cleavage by furin or the closely related pro-protein convertase [36]. There is a remarkably long delay of 1–3 days between the capsid cell surface binding and its penetration resulting in the onset of viral genome transcription. Itself the internalization of capsids as starts from the attachment to cell surface until the uncoating process begins lasts at least 2–4 h and is very asynchronous. The endocytosis as well as the consequent transmembrane trafficking of HPV capsids is not fully understood. Penetration of engulfed capsids may be initiated either from the acidified late endosomes, in which the L2 conformation change takes place, or by clathrindependent uptake [37]. The penetrated virions are transported from the internal membrane surface by involutions called vilopodia, namely from their leading edge to the central cell body via actin-directed retrograde flow [38]. Classical observations testify that there is extremely difficult to isolate and propagate any HPV types in conventional human cell cultures [39].

During the last decade, several details of HPV induced continuous host cell proliferation have been elucidated. It became clear that three non-structural HPV-coded oncoproteins, namely E5, E6, and E7 (**Table 8**) are involved in host cell immortalization, which later on, under the conditions of continuous host cell division results in malignant transformation. It should be mentioned that the transformation as such is a multistep process, launched by the virus-coded oncoproteins as suggested the "hit and run" hypothesis. The E5 protein enhances the sensi‐ tivity of HPV carrier cells to external proliferative stimuli, such as the epidermal growth factor [40]. It also binds to the growth factor receptors, for example, to the platelet growth factor receptor (PDGFR), and activates the signal transmission in a ligand-independent manner [41, 42].

ogy, the so-called koilocytes. Clearly, the relatively increased levels of HPV proteins such as E1 or E2 (or their modifications) may change the appearance of terminally differentiating keratinocytes. One might anticipate that the vegetative DNA replication occurs bi-direction‐ ally, through a theta structure intermediate or by a rolling circle mode, which is the principle of vDNA replication in general. Finally, the virion assembly must take place in the nuclei of terminally differentiated keratinocytes, which also contributes to koilocyte formation. The nascent capsids might randomly attach to the HPV DNA, and are further stabilized by the formation of disulfide bonds between conserved cysteines on adjacent L1 monomers acquiring resistance to proteolytic digestion. Taken together, as result of productive replication, the virions form large aggregates within the nuclei of infected keratin forming upper squamous epithelium cells, which then regularly show koilocytosis [32], but rarely reveal cytoplasmic

The newly produced HPV virions do not appear outside of squamous epithelium cells, and its productive replication is closely bound to the suprabasal, mainly medium and upper slayer of dividing keratinocytes. At infection, the HPV virions preferentially bind to heparan sulfate proteoglycans (HSPGs) on the basement membrane or to the basal stem cells, which may be exposed to environmental influence at sites of epithelial trauma or permeabilization. The differentiating squamous epithelium cells cannot become infected. To achieve selective infection of basal differentiating squamous epithelium cells, the initiation of infection prefer‐ entially requires attachment to the basal stem cells layered at the basal membrane of stratified squamous epithelium [33–35]. The basement membrane-bound virion undergoes a conforma‐ tional change of its L2 capsid protein that exposes a highly conserved N-terminal peptide motif to cleavage by furin or the closely related pro-protein convertase [36]. There is a remarkably long delay of 1–3 days between the capsid cell surface binding and its penetration resulting in the onset of viral genome transcription. Itself the internalization of capsids as starts from the attachment to cell surface until the uncoating process begins lasts at least 2–4 h and is very asynchronous. The endocytosis as well as the consequent transmembrane trafficking of HPV capsids is not fully understood. Penetration of engulfed capsids may be initiated either from the acidified late endosomes, in which the L2 conformation change takes place, or by clathrindependent uptake [37]. The penetrated virions are transported from the internal membrane surface by involutions called vilopodia, namely from their leading edge to the central cell body via actin-directed retrograde flow [38]. Classical observations testify that there is extremely difficult to isolate and propagate any HPV types in conventional human cell cultures [39].

During the last decade, several details of HPV induced continuous host cell proliferation have been elucidated. It became clear that three non-structural HPV-coded oncoproteins, namely E5, E6, and E7 (**Table 8**) are involved in host cell immortalization, which later on, under the conditions of continuous host cell division results in malignant transformation. It should be mentioned that the transformation as such is a multistep process, launched by the virus-coded oncoproteins as suggested the "hit and run" hypothesis. The E5 protein enhances the sensi‐ tivity of HPV carrier cells to external proliferative stimuli, such as the epidermal growth factor [40]. It also binds to the growth factor receptors, for example, to the platelet growth factor

inclusion bodies (see **Table 5** for details).

120 Human Papillomavirus - Research in a Global Perspective

Because the integrated vDNA regularly encompasses just the ORFs of E6 and E7 oncoproteins along with the closely positioned LCR sequence, a continuing and increased expression of E6/ E7 proteins occurs which seriously affects the regulation of host cell division in direction of its down-regulation [43, 44]. The E6 polypeptide binds the pivotal cell division inhibiting p53 protein [45], while E7 polypeptide binds the retinoblastoma (Rb) protein and the cyclin inhibitory proteins p27 and p21 [46]. The HPV-coded oncoproteins E6/E7, when expressed in significant amounts, continuously drive the host cell from the phase G1 to phase S (synthesis), in which the replication of cell DNA proceeds in an unlimited manner. Noteworthy that immortalization and the process of cancerogenesis are not the same, since the latter is much more complex being related to several chromosomal alterations and to increased number of *c-myc* gene copies (probably arisen due to repeated host cell DNA sequence transpositions), a finding even proportional to the grade of HPV driven dysplasia [47]. In low-grade dysplasias (such as CIN I/LSIL), only a restricted number of the HPV oncogenes is expressed.

Nevertheless, E6 and E7, which have several other activities (**Table 8**), appear to be the main drivers for the progression to high-grade dysplasia and later on to cancer, by orchestrating a series of pathogenic changes. Both are transcribed from the same major early promoters located within the LCR region of their genome. As a rule, the expression of the episomal (not yet integrated) HPV genome is restricted to E6 and E7 polypeptides, which are present in the parabasal, lower layer of squamous epithelium. Here, the viral genome undergoes a regulated duplication along with the host cell proliferation, which is under physiological conditions regulated by corresponding stimuli. The E6 protein is expressed at substantially lower levels than E7, and in early LSIL lesions, since the level of E7 may be more limited than that of E6. Noteworthy, both genes are expressed from a single promoter (latency associated and different from the above-mentioned vegetative ones); their transcripts are alternatively spliced by a post-translation mechanism determining as well as regulating their relative level within the carrier host cell. Integration of the HPV DNA fragment into cellular DNA via non-homologous recombination represents a key change towards immortalization that appears to stabilize the high expression of E6/E7, which, in turn, becomes frequently associated with more severe lesions. Integration may still not occur in great majority of CIN I/LSIL lesions, but is present by an increased rate in HSIL/CIN II and/or CIN III lesions, and is the most frequent in preinvasive cancer (Ca *in situ*). As a rule, whoever, the integrated E6/E7 ORF containing the HPV DNA fragment is at high probability found in the CIN III lesions. The frequency of viral DNA integration may vary with the HPV genotype, being more frequent for the high-risk (HR) genotypes such as HPV18 and/or HPV16.

The integration of vDNA and/or its fragment occurs due to interruption of its sequence, which can appear at many sites throughout the genome, but is found preferentially at fragile genomic sites, which undergo nicking and cutting in association recombination and/or translocation events of the host cell genome. In certain cases, practically the rest of whole genome may be deleted, so that only the E6 and E7 ORFs remain intact and ready for the transcription along with the nearby located LCR sequence, containing the crucial promoter and enhancer signals lying upstream of the integration site [48]. In a given *in vitro* immobilized keratinocyte culture, the integration process usually involves only one locus or a few loci. The E6/E7 ORF tran‐ scription slowly increases due the loss of the feedback block provided by the viral E1 and E2 proteins, which ORFs had been either deleted or at least disrupted. This situation permits the constant expression of high levels of the undesired E6 and E7 mRNA molecules [49]. Consistent with the multistep nature of tumorigenesis, cervical cancers may show additional cytogenetic alterations as compared with adjacent high-grade dysplastic and/or carcinoma *in situ* lesions.

Low-grade dysplasia may be caused by infection with either low-risk (LR) or high-risk (HR) HPV. Persistent (i.e., long-term) infection with a HR HPV type, which occurs in a minority of infected women, is the most important risk factor for developing CIN III or pre-invasive cancer (CIN III+). However, the magnitude of the risk depends not only on the given HPV type, but even more on a HR variant within the given type (compare **Table 6**). In practical terms, HPV persistence usually means that the same HPV genotype can be recovered from at least two or more subsequently taken genital samples obtained over a period of 4–12 months. Persistent infection may clear spontaneously, but less likely it does so in the course of longer duration. Taken together, only some persistent infections progress to CIN III, and subsequently to invasive cancer, but the HR HPV16 infections so such outcome much more likely than other HPV types. The distinct biologic effects of HR E6 and E7 may present at least a partial explanation of the differences in the likelihood of low-grade dysplasia progression, which presence may not be consistently associated with the same probability of the progression to high-grade lesion. Thus, most genital HPV infections are self-limited, and the majority would clear within 12 months.

Taken together, the HPV genome persists within transformed host cells in two different forms: as a non-integrated (episomal) circularized full-length vDNA and, less frequently, as a linear and integrated vDNA sequence. There should be mentioned that during long-term latency, the integration of HPV DNA may occur due to the linearization of the persisting circular vDNA molecule. At integration, the HPV DNA chain gets either interrupted or partially deleted, preferentially at nt 3362–3443 of the E2 ORF [50, 51], usually within the E2 gene ORF [52]. The mixed (episomal as well as integrated) vDNA distribution pattern seems the most prevalent physical state of HPV16 DNA found in ASCUS-graded smears (atypical cells of unknown significance), but can also be detected in cervical scrapings which do not reveal dysplastic changes. This indicates that HPV infection may not always cause dysplasia of the squamous epithelium cells. The prevalence of the integrated HPV DNA sequence over the episomal molecules then appears in a proportion of ASCUS-graded smears (sometimes characterized by ASC-H cells) and in a proportion of LSIL smears, but later on becomes clearly prevalent in the HSIL-grade smears and, of course, in cervical cancer. Women with the prevalence of integrated HPV DNA were almost 10 years older than those with a predominating episomal HPV DNA pattern, which points to a higher risk of HPV infection in women aged over 35 years [53].

## **3. Overexpression of the p16/INK4A regulatory polypeptide in dysplastic and/or proliferating cells**

lying upstream of the integration site [48]. In a given *in vitro* immobilized keratinocyte culture, the integration process usually involves only one locus or a few loci. The E6/E7 ORF tran‐ scription slowly increases due the loss of the feedback block provided by the viral E1 and E2 proteins, which ORFs had been either deleted or at least disrupted. This situation permits the constant expression of high levels of the undesired E6 and E7 mRNA molecules [49]. Consistent with the multistep nature of tumorigenesis, cervical cancers may show additional cytogenetic alterations as compared with adjacent high-grade dysplastic and/or carcinoma *in situ* lesions.

Low-grade dysplasia may be caused by infection with either low-risk (LR) or high-risk (HR) HPV. Persistent (i.e., long-term) infection with a HR HPV type, which occurs in a minority of infected women, is the most important risk factor for developing CIN III or pre-invasive cancer (CIN III+). However, the magnitude of the risk depends not only on the given HPV type, but even more on a HR variant within the given type (compare **Table 6**). In practical terms, HPV persistence usually means that the same HPV genotype can be recovered from at least two or more subsequently taken genital samples obtained over a period of 4–12 months. Persistent infection may clear spontaneously, but less likely it does so in the course of longer duration. Taken together, only some persistent infections progress to CIN III, and subsequently to invasive cancer, but the HR HPV16 infections so such outcome much more likely than other HPV types. The distinct biologic effects of HR E6 and E7 may present at least a partial explanation of the differences in the likelihood of low-grade dysplasia progression, which presence may not be consistently associated with the same probability of the progression to high-grade lesion. Thus, most genital HPV infections are self-limited, and the majority would

Taken together, the HPV genome persists within transformed host cells in two different forms: as a non-integrated (episomal) circularized full-length vDNA and, less frequently, as a linear and integrated vDNA sequence. There should be mentioned that during long-term latency, the integration of HPV DNA may occur due to the linearization of the persisting circular vDNA molecule. At integration, the HPV DNA chain gets either interrupted or partially deleted, preferentially at nt 3362–3443 of the E2 ORF [50, 51], usually within the E2 gene ORF [52]. The mixed (episomal as well as integrated) vDNA distribution pattern seems the most prevalent physical state of HPV16 DNA found in ASCUS-graded smears (atypical cells of unknown significance), but can also be detected in cervical scrapings which do not reveal dysplastic changes. This indicates that HPV infection may not always cause dysplasia of the squamous epithelium cells. The prevalence of the integrated HPV DNA sequence over the episomal molecules then appears in a proportion of ASCUS-graded smears (sometimes characterized by ASC-H cells) and in a proportion of LSIL smears, but later on becomes clearly prevalent in the HSIL-grade smears and, of course, in cervical cancer. Women with the prevalence of integrated HPV DNA were almost 10 years older than those with a predominating episomal HPV DNA pattern, which points to a higher risk of HPV infection in women aged over 35

clear within 12 months.

122 Human Papillomavirus - Research in a Global Perspective

years [53].

The aim of our further considerations was to assess the role of p16/INK4A (inhibitor kinase) protein, which is, as a rule, overexpressed in cells revealing increased E7 polypeptide produc‐ tion. The p16/INK4A (inhibitor kinase) polypeptide is a cellular regulatory protein, which inhibits the cyclin-dependent kinases (especially cdk4 and cdk6) associated with cyclins D and/ or E. These kinases, if activated, phosphorylate the retinoblastoma phosphoprotein (pRb) complex, which in turn, releases the transcription factor E2F, which is bound in inactive form in non-dividing cells. Under physiological conditions (in normal cells), the Rb protein liberated from the disrupted complex has a feedback effect on p16 expression [47]. Since the E2 poly‐ peptide binds to LCR, the presence of E2 in cells carrying the episomal (i.e., non-integrated) HPV DNA may efficiently control the transcription of mRNA encoding the E6/E7 oncoproteins [54]. In contrast, in cells which carry the integrated HPV genome, the E2 polypeptide produc‐ tion stops, since the E2 ORF becomes disrupted. Therefore, the E2 but also the E1 proteins (both

**Figure 1.** Different patterns of positive p16 staining in CIN I/LSIL. Upper line: continuous staining of p16 antigen of low (A) and/or high intensity (B) confined to the lower spinous layer and/or to the transit amplifying cells, which are situated in parabasal location just adjacent to the p16 negative basal stem cells. No dysplasia can be seen. Line below: staining of p16 antigen in a thin layer of dysplastic cells at parabasal area showing signs of HPV infection such as koi‐ locytes (in the left, C). Bottom in the right (D): progressed dysplasia of the basal layer in part involving the stem cells, which cannot be clearly recognized.

closely involved in the maintenance of long-term latency) may be missing in dysplastic and/or HPV-transformed cells. Unlike to transformed cells, their production becomes downregulated in the late phase of vegetative (productive) virus replication cycle. It can be stated that in HPV transformed cells, the expression of p16/INK4A protein increases proportionally to elevated levels of the increasing expression of the virus-coded E7 polypeptide, since both rise in the absence of E2 protein [55–58]. The p16/INK4 mRNA more stable and is present in higher levels in the cells in which the HPV DNA sequence had been integrated [59]. The quantification real-time–polymerase reaction (qRT PCR) is useful to identify the levels of transcripts encoding the p16 polypeptide in cervical smears of patients obtained for diagnostic purpose, an approach which is more tedious as the antigen staining, but is similarly sensitive and occasionally may yield more confident results. The p16 mRNA was present in 30% of LGSL but in 75% of HGSIL cases reaching a rate of 85.7% in squamous cell carcinoma cases [60].

The Bethesda scoring system, originally destined for vaginal/cervical smears [1988/1989], has been later adopted for the cervical biopsies at histologic examination. This has happened regardless to the fact that pathologists already had their own nomenclature, which is still in use for cervical biopsy grading widely known as cervical intraepithelial neoplasia (CIN). The dysplastic cells showing a diploid nuclear pattern are characterized by the loss of polarity, crowding, overlapping disorganization, and anisocytosis [61]. Therefore, the stage of CIN I dysplasia cannot be regarded for a truly neoplastic process. At cytological level, the dysplastic cells show altered nuclear-to-cytoplasm (N/C) ratio as well as wrinkling and thickening of nuclear membrane (**Table 4**) so that any CIN I grade changes may correspond to the entity of LSIL [62]. In mild forms of CIN I/LSIL, the dysplastic cells occupy the parabasal layer only but form a continuous zone within the lower third of cervical squamous epithelium, in the socalled lower squamous layer along with the transit amplifying cells, which early dysplasia can be better recognized by the p16 antigen staining (**Figure 1**). Summing up the results of p16 antigen staining in biopsy sections graded CIN I/LSIL, Yildiz et al. [63] could distinguish the continuous parabasal staining of higher intensity (**Figure 1B**), from parabasal staining of lower intensity (**Figure 1A**). In addition, they described the scattered form of positive p16 staining of single and/or small groups of upper squamous epithelium cells, which do not correspond to the suprabasal layer of lower squamous epithelium or the transit epithelium cells (**Figure 2D**). The production of p16 protein to an extent stainable with the commercially available antip16 monoclonal antibody (for example the CINTec histology kit) has been attributed to the stimulation of CDKN2A gene (encoding the p16 protein) by the stimuli activated via alterna‐ tive receptor pathways regulating the transient reactive cell proliferation. This situation should be strongly distinguished from dysplasia, where the cell growth regulation undergoes some kind of dysregulation (for example, by the viral oncogenic polypeptide E7) that causes the switch to an autonomic growth due to repeated cell divisions. Thus, the increased p16 expression in single epithelium cells in a location where the E7 polypeptide may not be present may be related to HPV infection. Noteworthy, older textbooks before the introduction of p16 staining [64] referred to the histological picture of chronic cervicitis characterized by with hyperemia and round cell (lymphocyte) infiltration of the underlying connective tissue in addition to the reactivity and dilution would also influence the p16 antigen staining results ranging from clearly negative to false positive for above-mentioned reasons (**Figure 3**). During the last years, several big companies introduced automatic staining procedures, which on one hand standardized the staining intensity as well as its color, but on other hand, new staining variations appeared between different laboratories. Therefore, in cases suspicious for CIN I/ LSIL, or if thep16 staining reveals a faint reactivity or shows a distribution other than parabasal, or does not correspond to a dysplastic cells area as seen in a parallel HE stained section, the

closely involved in the maintenance of long-term latency) may be missing in dysplastic and/or HPV-transformed cells. Unlike to transformed cells, their production becomes downregulated in the late phase of vegetative (productive) virus replication cycle. It can be stated that in HPV transformed cells, the expression of p16/INK4A protein increases proportionally to elevated levels of the increasing expression of the virus-coded E7 polypeptide, since both rise in the absence of E2 protein [55–58]. The p16/INK4 mRNA more stable and is present in higher levels in the cells in which the HPV DNA sequence had been integrated [59]. The quantification real-time–polymerase reaction (qRT PCR) is useful to identify the levels of transcripts encoding the p16 polypeptide in cervical smears of patients obtained for diagnostic purpose, an approach which is more tedious as the antigen staining, but is similarly sensitive and occasionally may yield more confident results. The p16 mRNA was present in 30% of LGSL but in 75% of HGSIL cases reaching a rate of 85.7% in squamous cell carcinoma cases [60].

124 Human Papillomavirus - Research in a Global Perspective

The Bethesda scoring system, originally destined for vaginal/cervical smears [1988/1989], has been later adopted for the cervical biopsies at histologic examination. This has happened regardless to the fact that pathologists already had their own nomenclature, which is still in use for cervical biopsy grading widely known as cervical intraepithelial neoplasia (CIN). The dysplastic cells showing a diploid nuclear pattern are characterized by the loss of polarity, crowding, overlapping disorganization, and anisocytosis [61]. Therefore, the stage of CIN I dysplasia cannot be regarded for a truly neoplastic process. At cytological level, the dysplastic cells show altered nuclear-to-cytoplasm (N/C) ratio as well as wrinkling and thickening of nuclear membrane (**Table 4**) so that any CIN I grade changes may correspond to the entity of LSIL [62]. In mild forms of CIN I/LSIL, the dysplastic cells occupy the parabasal layer only but form a continuous zone within the lower third of cervical squamous epithelium, in the socalled lower squamous layer along with the transit amplifying cells, which early dysplasia can be better recognized by the p16 antigen staining (**Figure 1**). Summing up the results of p16 antigen staining in biopsy sections graded CIN I/LSIL, Yildiz et al. [63] could distinguish the continuous parabasal staining of higher intensity (**Figure 1B**), from parabasal staining of lower intensity (**Figure 1A**). In addition, they described the scattered form of positive p16 staining of single and/or small groups of upper squamous epithelium cells, which do not correspond to the suprabasal layer of lower squamous epithelium or the transit epithelium cells (**Figure 2D**). The production of p16 protein to an extent stainable with the commercially available antip16 monoclonal antibody (for example the CINTec histology kit) has been attributed to the stimulation of CDKN2A gene (encoding the p16 protein) by the stimuli activated via alterna‐ tive receptor pathways regulating the transient reactive cell proliferation. This situation should be strongly distinguished from dysplasia, where the cell growth regulation undergoes some kind of dysregulation (for example, by the viral oncogenic polypeptide E7) that causes the switch to an autonomic growth due to repeated cell divisions. Thus, the increased p16 expression in single epithelium cells in a location where the E7 polypeptide may not be present may be related to HPV infection. Noteworthy, older textbooks before the introduction of p16 staining [64] referred to the histological picture of chronic cervicitis characterized by with hyperemia and round cell (lymphocyte) infiltration of the underlying connective tissue in addition to the reactivity and dilution would also influence the p16 antigen staining results ranging from clearly negative to false positive for above-mentioned reasons (**Figure 3**). During

**Figure 2.** Metaplasia and/or focal dysplasia suspicious for CIN I /LSIL. Line above (A and B; C in the middle): transi‐ tion of squamous epithelium into a metaplastic area. The cells positive for p16 antigen correspond to the foci of meta‐ plasia (confluent staining), but single p16 positive squamous epithelium cells can be seen as well (A). The area of metaplasia is rich of scattered Ki-67 positive nuclei, while the regularly lined positive nuclei belong to the basal stem cells (C). The line below depicts a reactive proliferation of squamous epithelium (D and E) growing into the cervical gland at the squamocolumnar junction. The positive p16 staining is not precisely parabasal, but corresponds to an ini‐ tial focus of dysplasia within the lower and/or medium squamous layers (shown by arrow at HE staining). The other‐ wise dispersed p16 antigen positive squamous cells may be unrelated to HPV infection.

prognostic value of positive p16 staining should be interpreted along with the outcome of the HPV DNA test.

**Figure 3.** Examples of either negative (A, upper line) and/or midzonal (lower line, B and C) p16 antigen staining. Both examples show non-parabasal location of the p16 positive cells in the upper (rather than in the lower) squamous layer. The staining is of nearly confluent (B) or focal (C) distribution and often corresponds to the localization of koilocytes. Therefore, it is suggestive to accompany cases of productive (vegetative) HPV replication. Such pattern of p16 antigen staining may be suspicious of low-grade lesion (CIN I/LSIL), but should be combined with the HPV DNA test in order to obtain a precise diagnosis.

If dysplastic cells occupy at least one half (exceeding one third) of the original squamous epithelium thickness (but not the whole epithelium layer), the appropriate designation is CIN II. The p16 antigen staining is therefore useful to meet the diagnosis of CIN II/HSIL, since it allows to estimate the precise thickness of dysplasia involving the squamous epithelium (**Figure 4**). In contrast, at the stage of CIN III/HSIL, the dysplastic cells can fully replace the original epithelium (**Figure 5**). At progressed stage, the intensive staining of p16 polypeptide can be found within the nuclei of dysplastic cells as well as in their cytoplasm. The dysplastic cells in question show enlarged nuclei of ovoid shape, which are not equal in size. Since the presence of p16 antigen is a hallmark for distinguishing the immortalized and/or dysplastic cells, at first glance, the site of outgrowth and/or proliferation of squamous epithelium into cervical glands can be detected. The decreased availability of functioning Rb protein due to overexpression of E7 polypeptide, not only leads to increased p16 production, but also explains the higher frequency of nuclei positive for Ki-67 protein, a marker of cell division [65]. For more precise grading of the SIL lesions by the Ki-67 antigen staining, Kruse et al. [66] suggested to count the number of positive nuclei per 100 μm epithelium thickness starting from the parabasal zone (stratification index). The authors in question found a satisfactory correlation of 83% with the dysplastic squamous epithelium cells as seen in HE stained sections, but they also noticed that parallel sections may not fully correspond to the previously cut block level. Furthermore, as seen at our example of the CIN III/HSIL pattern shown on **Figure 5C** and **5D**, while the p16 antigen staining involves the whole dysplastic cell layer, in the Ki-67 stained section the calculation of stratification index may be less convenient. Taken together, staining of parallel sections for Ki-67 and p16 antigens confirmed the usefulness of both markers, even though the p16 marker is more suitable for practical reasons. On the other hand, the frequent zonal distribution of p16 reactivity fits well with the extent of dysplasia CIN I lesion encom‐ passing less than one third of stratified epithelium. In the CIN II-graded lesion, the dysplastic cells encounter the half or nearly two thirds of squamous epithelium (**Figure 4**), while in the CIN III lesion a diffuse distribution of dysplastic cells involves the whole thickness of the original squamous epithelium layer. Therefore, at least in biopsies, the staining for p16 antigen alone has recently emerged as a reliable diagnostic marker suitable for quick orientation and relevant demonstration of the extent dysplasia as well as for the search of potentially invasive growth of squamous cell carcinoma cells penetrating the basement membrane (**Figure 6**).

prognostic value of positive p16 staining should be interpreted along with the outcome of the

**Figure 3.** Examples of either negative (A, upper line) and/or midzonal (lower line, B and C) p16 antigen staining. Both examples show non-parabasal location of the p16 positive cells in the upper (rather than in the lower) squamous layer. The staining is of nearly confluent (B) or focal (C) distribution and often corresponds to the localization of koilocytes. Therefore, it is suggestive to accompany cases of productive (vegetative) HPV replication. Such pattern of p16 antigen staining may be suspicious of low-grade lesion (CIN I/LSIL), but should be combined with the HPV DNA test in order

If dysplastic cells occupy at least one half (exceeding one third) of the original squamous epithelium thickness (but not the whole epithelium layer), the appropriate designation is CIN II. The p16 antigen staining is therefore useful to meet the diagnosis of CIN II/HSIL, since it allows to estimate the precise thickness of dysplasia involving the squamous epithelium (**Figure 4**). In contrast, at the stage of CIN III/HSIL, the dysplastic cells can fully replace the original epithelium (**Figure 5**). At progressed stage, the intensive staining of p16 polypeptide can be found within the nuclei of dysplastic cells as well as in their cytoplasm. The dysplastic cells in question show enlarged nuclei of ovoid shape, which are not equal in size. Since the presence of p16 antigen is a hallmark for distinguishing the immortalized and/or dysplastic

HPV DNA test.

126 Human Papillomavirus - Research in a Global Perspective

to obtain a precise diagnosis.

**Figure 4.** Examples of different CIN II lesions and of the corresponding p16 antigen staining. Upper line (A and B): confluent staining of p16 antigen at transition from stage CIN I (in the left half of the Figure) into stage CIN II (in the right half of same Figure 1A). Figure 1B: The p16 positive dysplastic cells involve the lower and the medium squamous layers but still not the whole epithelium; numerous koilocytes can be seen in the in-close vicinity of the p16 antigen positive dysplastic epithelium cells. Bottom line (C and D): Extensive dysplasia involves the lower and medium squa‐ mous cell layers, but is still absent at thin superficial layer of flat granular cells. The distribution of p16 antigen staining

corresponds to transition from CIN II into CIN III (the area enlarged at D is pointed by arrow). As shown in detail (D), a few koilocytes may be still found.

**Figure 5.** High degree of dysplasia (CIN III/HSIL) which occupies the entire epithelium layer and carcinoma *in situ*. Line above: Transition of squamous epithelium into dysplasia as shown by HE (B) and by p16 antigen staining (A). Middle line: The CIN III/HSIL as seen by p16 and Ki-67 antigen staining (C and D). While the p16 positive dysplastic cells occupy the whole epithelium, the nuclei of dividing cells are either round shaped and of regular size (basal stem cells closely at the basement membrane) or show an irregular ovoid shape of varying in size (dispersed at whole dys‐ plastic layer but not each cell is positive). Bottom line: Carcinoma *in situ* proliferates into the surrounding connective tissue (cells p16 positive) and into a cervical gland replacing the cylindric epithelium but still keeping the basement membrane intact (E). The entirely dysplastic epithelium is positive for the p16 antigen; it shows loss of squamous dif‐ ferentiation along with extensive proliferative growth, but with the basement membrane preserved (F, no invasion at the arrow area).

**Figure 6.** Carcinoma *in situ* (A) versus the invasive growth of spinocellular carcinoma (B) as related to the expression of p16 polypeptide. Dysplastic cells, forming the carcinoma *in situ* (A) as well as the non-differentiated malignant cells of invasive spinocellular carcinoma (B), are strongly positive for p16 and showing cytoplasmic as well as nuclear stain‐ ing. After disrupting the basement membrane, groups of carcinoma cells grow not only within the surrounding con‐ nective tissue, but invade some capillaries and/or small venules as well as lymphatic vessels (C, arrow). At low power view, the origin of invasion can be recognized (D, arrow).

Nevertheless, any such comparisons may suffer from possible imprecision, since the repeat‐ edly stained a tendency to worse grading at the primary evaluation of HE results. Namely, 3% of Ki-67 stained sections were graded CIN I rather than CIN II as described in HE stained sections. Alternatively, 14% cases graded CIN III in the HE stained sections were interpreted CIN II according to the Ki-67 staining. A similar tendency to worse grading from the evalua‐ tion, the HE stained sections only was found by us in p16 staining [14]. According to our experience, the viewing of parallel sections stained for p16 as well as Ki-67 antigen markers is of great help at confirming the extent of dysplasia seen in CIN II and/or CIN III HSIL-graded sections. In addition, it contributes to the correct evaluation of CIN I/LSIL lesions among which the reactive and other non-neoplastic staining patterns should be distinguished [67].

corresponds to transition from CIN II into CIN III (the area enlarged at D is pointed by arrow). As shown in detail (D),

**Figure 5.** High degree of dysplasia (CIN III/HSIL) which occupies the entire epithelium layer and carcinoma *in situ*. Line above: Transition of squamous epithelium into dysplasia as shown by HE (B) and by p16 antigen staining (A). Middle line: The CIN III/HSIL as seen by p16 and Ki-67 antigen staining (C and D). While the p16 positive dysplastic cells occupy the whole epithelium, the nuclei of dividing cells are either round shaped and of regular size (basal stem cells closely at the basement membrane) or show an irregular ovoid shape of varying in size (dispersed at whole dys‐ plastic layer but not each cell is positive). Bottom line: Carcinoma *in situ* proliferates into the surrounding connective tissue (cells p16 positive) and into a cervical gland replacing the cylindric epithelium but still keeping the basement membrane intact (E). The entirely dysplastic epithelium is positive for the p16 antigen; it shows loss of squamous dif‐ ferentiation along with extensive proliferative growth, but with the basement membrane preserved (F, no invasion at

**Figure 6.** Carcinoma *in situ* (A) versus the invasive growth of spinocellular carcinoma (B) as related to the expression of p16 polypeptide. Dysplastic cells, forming the carcinoma *in situ* (A) as well as the non-differentiated malignant cells of invasive spinocellular carcinoma (B), are strongly positive for p16 and showing cytoplasmic as well as nuclear stain‐ ing. After disrupting the basement membrane, groups of carcinoma cells grow not only within the surrounding con‐ nective tissue, but invade some capillaries and/or small venules as well as lymphatic vessels (C, arrow). At low power

view, the origin of invasion can be recognized (D, arrow).

a few koilocytes may be still found.

128 Human Papillomavirus - Research in a Global Perspective

the arrow area).

According to Dray et al. [68], the CIN II/HSIL- and/or CIN III/HSIL-graded lesions were found in 40.8% of cervical biopsies, while 14.3% showed the mild CIN I/LSIL lesion; the rest of 45% revealed a range of non-dysplastic (inflammatory or reactive) changes. In the latter group, the focal and weaker midzonal or superficial p16/INK4A immunostaining, suggestive of episomal HPV infection, was noted in 10% of biopsies. As shown in our statistics concerning biopsies in the year 2015 examined at the Pathology Diagnostic Center in Martin (Pathology 2, bottom line in **Table 11**), the definitely positive CIN I/LSIL lesions were seen in 36% of biopsies (90 out of 250). These lesions showed parabasal dysplasia within the lower third of squamous epithelium as well as koilocytes among adjacent squamous cells rarely involving the upper squamous layer (**Figure 1**). A proportion of suspicious CIN I/LSIL cases revealed focal dysplasia which, as a rule, correlated with the non-continuous p16 antigen positivity of squamous epithelium (**Figure 2**). The CIN II- and/or CIN III-grade HSIL lesion was detected in 37% of biopsies (89/250). The former showed dysplasia along with positive p16 staining involving about the half of squamous epithelium (**Figure 4**), while by the latter, the entire squamous epithelium layer was strongly positive for the p16 antigen (**Figure 5**). In CIN II cases involving more than one third of squamous epithelium, in the lower as well as in upper squamous layers occasional koilocytes found (**Figure 4B** and **4D**). In clearly CIN III-graded cases, the p16 positive dysplastic cells not only have replaced the whole squamous epithelium but also showed extensive proliferative growth either into the cervical glands or into the underlying connective tissue still with the basement membrane preserved (**Figure 5E** and **5F**). Nevertheless, similar proliferative growth might be occasionally found in CIN II/LSIL-graded cases (**Figure 4A**–**4C**), but very rarely, even in the CIN I/LSIL focal lesion. (**Figure 2D** and **2E**). In general, the increasing incidence of the combined p16/Ki-67 staining indicates more severe lesions: It may be positive by 26.8% of normal histology (missing dysplasia) cases, by 46.5% in CIN I histology, by 82.8% of CIN II, and/or by 92.8% of CIN III-graded histology [69]. In our hands, squamous epithelium showing no dysplasia was found in 55 out of 250 cases (20%) in biopsies of the Pathology Department 2, but only 10% of biopsies in the Pathology Department 1 (**Table 11**). The findings in the absence of dysplasia fall into at least 2 groups (**Figure 3**): (1) p16 negative cases showing neither dysplasia nor p16 antigen presence, and (2) cases of positive p16 immunostaining in the non-dysplastic areas of squamous epithelium, especially in its midzonal location, which does clearly differ from the parabasal distribution of p16 antigen. The p16 positive staining in the absence of dysplasia should be always examined for HPV DNA presence, and otherwise, the diagnosis can be regarded either for incomplete or at least imprecise from the point of view future prognosis. Although there is good evidence that p16/INK4A immunostaining correlates with the severity of cytological/histological abnormal‐ ities, the reproducibility might be limited also because of the insufficiently standardized interpretation of p16 staining results [70].

## **4. The diagnostic value of cervical smears in carcinoma prevention**

As already mentioned in Introduction, due to examination of cervical smears in Slovakia during the last decade, the absolute number of cervical carcinoma cases as well as the morbidity rate has considerably decreased. Preventive cytological examination, which started in our country from 2006, and became especially frequent in the last 5–6 years (compare **Table 1**) in part explains this favorable development. As shown in **Table 4**, in just one out of 10 diagnostic centers serving for the population of approximately 2 million Slovak women in the age from 15 to 60 years, over 37,000 PA smears were enrolled from about 34,000 women were examined.

**Figure 7.** ASCUS- and/or LSIL-graded smears handled by the conventional PA method and stained for p16 antigen (A, B) in comparison with the appearance of atypical and/or abnormal squamous cells as seen after Liqui-PREP staining (C, D, E, and F). The upper line in the left depicts a group of atypical squamous epithelium cells with enlarged nuclei (A) in comparison with a group of abnormal cells-graded LSIL (B), both stained for p16 antigen. The middle line in the right (D) shows inflammation and metaplasia with a group of abnormal epithelium cells scored LSIL (pointed by ar‐ row), while in the middle line left a group of atypical cells with enlarged nuclei can be seen (ASCUS) some of which being suggestive for LSIL (C) . The smear in the right, bottom line shows atypical cells along with the presence of koi‐ locytes (F). In the bottom line left (E), one of the Ki-67 positive abnormal squamous cells shows faint p16 antigen stain‐ ing in the cytoplasm (simultaneously stained for p16 as well as Ki-67 antigens).

A positive rate of 4.6% experienced in our particular screening is in accord with the data previously reported from the US [16]. Despite of the success of mass screening based on the relatively simple PAP technique, improved cytological tests such as liquid-based cytology (Liqui-PREP) were introduced to achieve more precise reading. The new approach resulted in a significant decrease of low-quality samples [71], when allowing to identify and distinguish the atypical epithelium cells allowing to identify and better distinguish the atypical epithelium cells at their better visualization, which may be of great advantage especially for the recogni‐ tion of LSIL cases (**Figure 7**).

least imprecise from the point of view future prognosis. Although there is good evidence that p16/INK4A immunostaining correlates with the severity of cytological/histological abnormal‐ ities, the reproducibility might be limited also because of the insufficiently standardized

As already mentioned in Introduction, due to examination of cervical smears in Slovakia during the last decade, the absolute number of cervical carcinoma cases as well as the morbidity rate has considerably decreased. Preventive cytological examination, which started in our country from 2006, and became especially frequent in the last 5–6 years (compare **Table 1**) in part explains this favorable development. As shown in **Table 4**, in just one out of 10 diagnostic centers serving for the population of approximately 2 million Slovak women in the age from 15 to 60 years, over 37,000 PA smears were enrolled from about 34,000 women were examined.

**Figure 7.** ASCUS- and/or LSIL-graded smears handled by the conventional PA method and stained for p16 antigen (A, B) in comparison with the appearance of atypical and/or abnormal squamous cells as seen after Liqui-PREP staining (C, D, E, and F). The upper line in the left depicts a group of atypical squamous epithelium cells with enlarged nuclei (A) in comparison with a group of abnormal cells-graded LSIL (B), both stained for p16 antigen. The middle line in the right (D) shows inflammation and metaplasia with a group of abnormal epithelium cells scored LSIL (pointed by ar‐ row), while in the middle line left a group of atypical cells with enlarged nuclei can be seen (ASCUS) some of which being suggestive for LSIL (C) . The smear in the right, bottom line shows atypical cells along with the presence of koi‐ locytes (F). In the bottom line left (E), one of the Ki-67 positive abnormal squamous cells shows faint p16 antigen stain‐

ing in the cytoplasm (simultaneously stained for p16 as well as Ki-67 antigens).

**4. The diagnostic value of cervical smears in carcinoma prevention**

interpretation of p16 staining results [70].

130 Human Papillomavirus - Research in a Global Perspective

**Figure 8.** Examples of HSIL-graded smears: Above, in the left (A): The conventional PAP staining is of relatively good quality achieved using the CYNTec staining kit for p16 antigen detection (positive). The rest (B, C, and D) of smears were handled by liquid-based cytology and shows either groups or single dysplastic cells-graded HSIL (B, C) and/or HSIL+ (at D, a single dysplastic cell with an extremely large hyperchromatic nucleus is shown suggestive of carcinoma *in situ*). The bottom line (E and F) shows groups of distorted (shrunken) dysplastic epitelium cells stained with the anti/p16/anti-Ki-67 MoAb mix (nuclei dark purple, cytoplasm contains many p16 antigen brown granules).

As mentioned above, the Bethesda Committee classification [72] and the suggested definitions were later on slightly improved reaching the state which has become widely accepted [73]. Comparisons among laboratories showed that the diagnosis of ASCUS (atypical squamous cells of undetermined significance) may be often used just to avoid clear-cut decision making. Therefore, the principle has been be that this diagnostic category should not exceed 5–6% of the total number of smears investigated. As stressed by Geisinger et al. [74], the main criterion for the clear definition of SIL is the increased size of nucleus (<3 fold for ASCUS, >3 fold in LSIL), increased intensity of chromatin staining and its altered internal structure (finely granular chromatin structure and slight hyperchromasia in ASCUS, coarse chromatin and definitely visible hyperchromasia in LSIL reports claimed the relative value of p16 staining in ASCUS smears, which may be positive in cases designated as p16 reactive ASC. The proportion of p16 negative smears was reported for the highest in ASCUS (40%) and the lowest in HSIL (5%) specimens [75, 76]. According to Grapsa et al. [77], the p16 staining is weak in LSIL (compare **Figure 7A**), but strong in HSIL cases (**Figure 8A** and **8B**). Thus, the faint p16 staining argues against the diagnosis of LSIL, while p16 overexpression along with high levels of p53 favors the process of malignant transformation of the atypical squamous epithelium cells.

According to Shin et al. [78], the p16 antigen was found in 66.7% of ASCUS and 70% of LSIL cases, confirming that detection of p16/INK4A protein can be used as adjunct test especially at liquid-based cytology. However, several authors noticed a high degree of false p16 reactivity within otherwise negative smears, especially in those containing atrophic cells [79]. They also stressed that in such smears, the number of single p16 reactive atypical cells per the total cell number of cells may not be significant. Such situation is less likely to appear in biopsy sections, where typical confluent distribution of p16 antigen positive cells at the parabasal epithelium layer is of essential help for correct interpretation of the result. In our previous paper [14], we focused our interest on unclear and/or ASCUS-graded smears, which were found HPV DNA positive by a probability of 8–10%, but have been p16 antigen positive with a higher probability ranging from 32 to 61%. The wide range of p16 antigen false positivity of atypical cells was found associated with the staining procedure itself, since the CYN-Tec cytology staining kit showed more clear-cut results, allowing to distinguish the proportion of really p16 antigen positive, but HPV unrelated atypical cells on hand, from artificially stained ones on other hand. It should be stressed that in the ASCUS-graded smears but also in a great proportion of LSIL scored smears the most important criterion still remains the evaluation of nuclear changes, such as the altered N/C ratio, which should be considered for each particular p16 reactive cell. The diagnosis of ASC gains some prediction value, if based on the presence of a few HSIL-like cells (the so-called ASC-H). The latter points at a fast progression to HSIL from the very beginning (**Figure 8**). Nevertheless, the ASC-H category was challenged by other investigators as being problematic, since the morphologic difference in the appearance of such single metaplastic and/or neoplastic cell is not always clear and opens the way to false positive as well as false negative diagnoses [80–82]. Nuclear hyperchromasia and irregular nuclear membrane contours count as the most reliable diagnostic feature in the so-called pre-neoplastic atypia. The presence of such cells in the smears possibly corresponds to the dysplastic cells in histological sections [83], which, however, cannot be present in the smear in the case of CIN I/ LSIL (compare **Figure 1**), but may be already easily collected in the cases of CIN II/HSIL (compare **Figure 4**). For better orientation in the most uncertain cases, a short course of local estrogen cream therapy followed by repeated PA tests has been suggested for helpful [84].

According to our experience, the relatively high percentage of p16 reactive atypical cells causing false positive LSIL grading appears in smears if handled by the conventional PA method. Therefore, the LBC should be preferred at repeated examination of patients which had been selected according to the classical Papanicolaou diagnostic system (when scored PA IIIa/IIIb roughly corresponding ASCUS and/or LSIL, as well as PA IV corresponding to HSIL and/or even PA V). The advantage of recently introduced improved diagnostic approach can be demonstrated on the results from the Cytology Laboratory (CL) of the St. Elisabeth Cancer

Institute (SECI) in Bratislava obtained within the years 2012–2013. The total number of gynecological cytology samples enrolled during the given period was 17,272, which corre‐ sponds to approximately 46% of samples examined at the Cytology Laboratories of Alpha medical presented in **Table 4**. While **Table 4** shows the positive rate of the screening from samples enrolled by practitioners, the Cancer Institute samples were submitted by specialists to whom the positively screened patients were sent for further care and/or treatment. There‐ fore, the diagnostic approach met was more complex combining the LBC technic (for example the Liqui-PREP kit) along with the detection of HPV DNA in nearly 53% of samples reflects the advantage of the interpretation of results, which have been achieved by comparison of both methods (**Table 9**). When considering the fact there was no logical need to request the HPV DNA test in the majority of negative cases enrolled (out of these 28% were tested for HPV only) then in becomes clear that the number of smears complete by DNA testing prevailed not only in essentially inevitable cases, such as ASCUS and LSIL, but by average the HPV DNA was tested by a proportion of 52.3% of all smears examined. When considering all the important facts mentioned, then the complete (dual examination) approach has been applied in 68% of ASCUS/LSIL cases (shown in bold type). It comes from **Table 10** that the HPV positive rates ranged from 27 to 33% in ASCUS and/or LSIL cases, respectively, indicating that the probability of further progression in these mild lesions could be quite low. In contrast, by the HSIL-graded smears the HPV DNA test was positive in 92% of cases. The same was true for the combination of positive HPV test with the p16 antigen staining. This was seen by 32% of LSIL-graded cases, in which the probability of progression was then relatively high, since as has been such transition really may occur by up to 15% of LSIL cases (citácia), which means that each second such double positive LSIL case might progress into HSIL. As our statistics concerns, this was really the case, but by a lower probability equal to one out of three LSIL cases. Byun et al. [85] conducted a comparative study including the p16 INK4a/Ki-67 double staining as well as the L1 capsid protein immunostaining along with human papillomavirus (HPV) DNA detection and typing in 56 ASC-H or LSIL-H cases diagnosed by LBC stained smears came to the conclusion that their approach was sufficient to predict CIN II+ and/or CIN III+ later on diagnoses at histological examinations of biopsies obtained from patients who underwent conization. Further interesting considerations are coming from the comparison of the biopsies subsequently resulting from the previous cytological diagnostic based their histological grading made at different Pathology Departments. While in the Diagnostic Center at the Pathology in Martin revealed that the probability of negative and/or non-neoplastic findings (including false positive and/or reactive staining) was approximately 22%. In contrast, the probability of the occurrence of the same result at the Pathology Department of the Oncology Institute in Bratislava was slightly below 10% only (**Table 11**). This difference in the negative scoring of CIN lesions in cervical biopsies reflects a better forecast based on a complex diagnostic approach, involving the use of LBC technology at repeated examination of cervical smears with contemporary HPV DNA tests along with the staining of smears for at least two serological markers. We believe that the cytological diagnosis should reflect a better degree of cooperation between clinicians and laboratory workers. As seen according to the presented algorithm, the suggested minimum of 2 or 3 colposcopic sessions may be satisfactory for precise diagnosis (**Figure 9**). In positive cases (i.e., those graded over PA III according to

ASCUS smears, which may be positive in cases designated as p16 reactive ASC. The proportion of p16 negative smears was reported for the highest in ASCUS (40%) and the lowest in HSIL (5%) specimens [75, 76]. According to Grapsa et al. [77], the p16 staining is weak in LSIL (compare **Figure 7A**), but strong in HSIL cases (**Figure 8A** and **8B**). Thus, the faint p16 staining argues against the diagnosis of LSIL, while p16 overexpression along with high levels of p53 favors the process of malignant transformation of the atypical squamous epithelium cells.

132 Human Papillomavirus - Research in a Global Perspective

According to Shin et al. [78], the p16 antigen was found in 66.7% of ASCUS and 70% of LSIL cases, confirming that detection of p16/INK4A protein can be used as adjunct test especially at liquid-based cytology. However, several authors noticed a high degree of false p16 reactivity within otherwise negative smears, especially in those containing atrophic cells [79]. They also stressed that in such smears, the number of single p16 reactive atypical cells per the total cell number of cells may not be significant. Such situation is less likely to appear in biopsy sections, where typical confluent distribution of p16 antigen positive cells at the parabasal epithelium layer is of essential help for correct interpretation of the result. In our previous paper [14], we focused our interest on unclear and/or ASCUS-graded smears, which were found HPV DNA positive by a probability of 8–10%, but have been p16 antigen positive with a higher probability ranging from 32 to 61%. The wide range of p16 antigen false positivity of atypical cells was found associated with the staining procedure itself, since the CYN-Tec cytology staining kit showed more clear-cut results, allowing to distinguish the proportion of really p16 antigen positive, but HPV unrelated atypical cells on hand, from artificially stained ones on other hand. It should be stressed that in the ASCUS-graded smears but also in a great proportion of LSIL scored smears the most important criterion still remains the evaluation of nuclear changes, such as the altered N/C ratio, which should be considered for each particular p16 reactive cell. The diagnosis of ASC gains some prediction value, if based on the presence of a few HSIL-like cells (the so-called ASC-H). The latter points at a fast progression to HSIL from the very beginning (**Figure 8**). Nevertheless, the ASC-H category was challenged by other investigators as being problematic, since the morphologic difference in the appearance of such single metaplastic and/or neoplastic cell is not always clear and opens the way to false positive as well as false negative diagnoses [80–82]. Nuclear hyperchromasia and irregular nuclear membrane contours count as the most reliable diagnostic feature in the so-called pre-neoplastic atypia. The presence of such cells in the smears possibly corresponds to the dysplastic cells in histological sections [83], which, however, cannot be present in the smear in the case of CIN I/ LSIL (compare **Figure 1**), but may be already easily collected in the cases of CIN II/HSIL (compare **Figure 4**). For better orientation in the most uncertain cases, a short course of local estrogen cream therapy followed by repeated PA tests has been suggested for helpful [84].

According to our experience, the relatively high percentage of p16 reactive atypical cells causing false positive LSIL grading appears in smears if handled by the conventional PA method. Therefore, the LBC should be preferred at repeated examination of patients which had been selected according to the classical Papanicolaou diagnostic system (when scored PA IIIa/IIIb roughly corresponding ASCUS and/or LSIL, as well as PA IV corresponding to HSIL and/or even PA V). The advantage of recently introduced improved diagnostic approach can be demonstrated on the results from the Cytology Laboratory (CL) of the St. Elisabeth Cancer conventional Papanicolaou system), the repeated smear taken should be handled by the LBCbased technic and stained for at least two markers (p16 and Ki-67 and tested HPV DNA presence.


*Abbreviations*: LBC—liquid-based cytology; HPV—human papilloma virus; ASCUS—atypical squamous cells of undetermined significance; ASC-H—Atypical squamous cells-cannot exclude high-grade squamous intraepithelial lesion; LGSIL—Low-grade squamous intraepithelial lesion; HGSIL—High-grade squamous intraepithelial lesion; SCa —squamous cell carcinoma.

\* The CL is a part of the Pathology Department of the St. Elisabeth Cancer Institute, which also is a teaching center of Slovak Postgraduate University in Bratislava

\*\* Negative.

\*\*\* HPV DNA not tested.

**Table 9.** LBC examinations of cervical smears during the period of 2012–2013 at the CL of the St. Elisabeth Cancer Institute, Bratislava.


LBC—liquid-based cytology; HPV—human papilloma virus; ASCUS—atypical squamous cells of undetermined significance; ASC-H—Atypical squamous cells-cannot exclude high-grade squamous intraepithelial lesion; LSIL— Low-grade squamous intraepithelial lesion; HGSIL—High-grade squamous intraepithelial lesion; SCa—squamous cell carcinoma.

\* Calculated from the total completed examinations.

\*\* Scored as non-suspicious, see also in **Table 9**.

**Table 10.** The completed LBC-based cervical smears according to staining procedures and correlated with the HPV DNA test result.


LBC—liquid-based cytology; HPV—human papilloma virus; ASCUS—atypical squamous cells of undetermined significance; ASC-H—Atypical squamous cells-cannot exclude high-grade squamous intraepithelial lesion; LGSIL— Low-grade squamous intraepithelial lesion; HGSIL—High-grade squamous intraepithelial lesion; SCa—squamous cell carcinoma; CIN—cervical intraepithelial neoplasia and its grading.

\* Concerns the biopsies examined at CL of St. Elisabeth Cancer Institute.

conventional Papanicolaou system), the repeated smear taken should be handled by the LBCbased technic and stained for at least two markers (p16 and Ki-67 and tested HPV DNA

**Cytology Completed Not completed Total examined**

*Abbreviations*: LBC—liquid-based cytology; HPV—human papilloma virus; ASCUS—atypical squamous cells of undetermined significance; ASC-H—Atypical squamous cells-cannot exclude high-grade squamous intraepithelial lesion; LGSIL—Low-grade squamous intraepithelial lesion; HGSIL—High-grade squamous intraepithelial lesion; SCa

\* The CL is a part of the Pathology Department of the St. Elisabeth Cancer Institute, which also is a teaching center of

**Table 9.** LBC examinations of cervical smears during the period of 2012–2013 at the CL of the St. Elisabeth Cancer

**Ki-67 positive** **p16+/ Ki-67+** **HPV DNA test**

**p16+/HPV+**

**Positive alone**

**markers**

**positive**

Negative\* 63 9.9 0 0 0 3 0

ASC-H 5 0.8 2 2 2 2 2

SCa 0 0 0 0 0 0 0

ASCUS 139 21.9 87 (63%) 7 4 38 (27%) 21 (15%)

LSIL 304 47.9 268 (88%) 66 114 99 (33%) 97 (32%) HSIL 124 19.5 120 (96%) 106 103 114 (92%) 112 (90%)

Total 635 100% 477 181 223 256 (40%) 232 (37%)

LBC—liquid-based cytology; HPV—human papilloma virus; ASCUS—atypical squamous cells of undetermined significance; ASC-H—Atypical squamous cells-cannot exclude high-grade squamous intraepithelial lesion; LSIL— Low-grade squamous intraepithelial lesion; HGSIL—High-grade squamous intraepithelial lesion; SCa—squamous cell

**Table 10.** The completed LBC-based cervical smears according to staining procedures and correlated with the HPV

Non-suspicious 63 (28%) 225 288 ASCUS **139 (59%)** 96 (41%) 235 ASC-H **5 (63%)** 3 8 LSIL **304 (68%)** 143 (32%) 447 HSIL 124 (53%) 108 (47%) 232 SCa 0 4 4 **Total 635 (52.3%) 579 1214**

presence.

—squamous cell carcinoma.

\*\*\* HPV DNA not tested.

Institute, Bratislava.

\*\* Negative.

**Cytological diagnosis**

carcinoma. \*

DNA test result.

 Calculated from the total completed examinations. \*\* Scored as non-suspicious, see also in **Table 9**.

Slovak Postgraduate University in Bratislava

134 Human Papillomavirus - Research in a Global Perspective

**Number Antigenic**

**Per cent\* p16**

\*\* Includes the reactive not-dysplastic lesions, such as false positive p16 staining as well as the p16 negative (normal) squamous epithelium.

\*\*\* The data concern the Diagnostic Center of Pathology Ltd, Alpha medical, Martin.
