Human papillomavirus typing and DNA ploidy determination of squamous intraepithelial lesions in liquid-based cytologic samples

Authors


Abstract

BACKGROUND

Infection by high-risk human papillomavirus (HPV) plays a role in the evolution of cervical carcinoma. Cellular atypia and consecutive DNA content alterations in cytologic samples are consequences of a preexisting viral infection.

METHODS

We analyzed the frequency and association of HPV types and the presence of rare cells with abnormally high DNA content. We also evaluated whether these findings support the cytologic diagnosis in 112 routine cases with low and high-grade squamous intraepithelial lesions (LSIL/HSIL) when performed from liquid-based cytologic samples (ThinPrep). For DNA content measurements, laser scanning cytometry was applied and at least 10,000 cells were analyzed. HPV typing was performed by a direct sequencing approach using the consensus primers GP5+/GP6+ and MY09/MY11.

RESULTS

Of 112 SIL cases, 110 (98.2%) were HPV positive and 95 (84.8%) had a high-risk type HPV infection. Almost one-half of the cases (46 of 95, 48.4%) with a high-risk HPV infection presented aneuploid squamous cells with greater than 9c DNA content, whereas none of the low-risk HPV-positive or HPV-negative SIL cases showed any aneuploid cells in this range. Although 91.8% of the HSIL cases displayed greater than 9c aneuploid cells, only 7.9% of the LSIL cases were positive for such cells with abnormally high DNA content.

CONCLUSIONS

HPV typing and DNA measurements help in the objectivation of cytologic atypia and both can be performed efficiently from the same liquid-based cytologic samples. Cancer (Cancer Cytopathol) 2003;99:57–62. © 2003 American Cancer Society.

The evolution of cervical carcinoma is a complex multistep process. The strong association between early human papillomavirus (HPV) infection and the generation of precancerous lesions or manifest cancer is widely accepted.1–5 Oncogenic HPV types are responsible for cell cycle regulatory and replicative alterations that lead to malignant transformation and gross genetic instability (e.g., DNA aneuploidy).6 The effect of HPV-related biologic processes are also reflected at the cytologic level, with limited specificity. According to the The Bethesda System,7 squamous intraepithelial lesions (SIL) represent a clear premalignant condition with cellular changes that may lead to the development of cervical carcinoma. Although SIL reverts in a number of cases, the diagnosis of high-grade SIL (HSIL) always indicates continuous medical treatment. Therefore, objective criteria that define the progression of symptoms from low-grade (LSIL) to HSIL to cervical carcinoma, as well as provide reliable prognostic information, are needed. Currently, molecular biologic methods are used to demonstrate or exclude oncogenic types of HPV.8, 9 In addition, aneuploidy is frequently observed with atypia, which can be explained as an effect of HPV on replicating cells. Aneuploidy can be demonstrated in cervical cytological samples by DNA content measurements.10–12 Image analysis provides stoichiometric methods to measure total DNA content under microscopic conditions. Recently, laser scanning cytometry (LSC) was introduced for the slide-based automated analysis of a large number of cells.Whereas the occurrence of aneuploid cells with increasing cellular atypia seems to be reasonable, a clear correlation between DNA content abnormalities and HPV types has not been demonstrated.

The purpose of this study is to present the correlations between polymerase chain reaction (PCR)-based HPV typing and DNA aneuploid cells using LSC. We analysed samples with cytologic diagnoses of LSIL and HSIL as determined by the criteria of The Bethesda System. Liquid-based cytologic preparations were used for all methods analysed, including the Pap test, microscope-based LSC, and PCR-based HPV sequencing.

MATERIALS AND METHODS

Cytologic Diagnosis

In this study, 112 cases with a cytologic diagnosis of LSIL or HSIL according to The Bethesda System7 were analyzed blind and compared for DNA aneuploidy and HPV positivity. Routine liquid-based (CytoLyt, Cytyc, Boston, MA) cervical cytologic samples were used for all purposes. For routine cytology and DNA content measurements, cell monolayers (ThinPrep, Cytyc) were prepared according to the manufacturer's protocol. Cytologic screening was done following conventional Pap staining by trained cytologists. All suspicious slides were reevaluated by a pathologist and the diagnosis of LSIL and HSIL was made using criteria of The Bethesda System.

DNA Cytometry

DNA content measurements were performed on a second monolayer of the same material stained in a solution containing 50 μg/mL propidium iodide (PI) and 200 μg/mL RNase at 37 °C for 1 hour. The cells were covered with glycerol/PI (25 μg/mL) and a coverglass. At least 10,000 cells were measured using an LSC (CompuCyte, Boston, MA). Normal (2c) DNA content was determined according to the first peak intensity value of the DNA histogram, which contained normal leukocytes as verified in the microscope after repositioning. Coefficients of variation (CV) of the DNA histograms ranged between 4.0 and 7.5. PI-stained cells with elevated DNA content (> 5c and > 9c) were gated and individually evaluated in the microscope of the LSC following computer-assisted repositioning. Artifacts and occasional cell doublets/conglomerates were excluded from the analysis. Restaining of the slides with the Pap stain was also done in selected cases.

Isolated cells with a nonsuperficial cell morphology and a DNA content of greater than 9c were identified as “Rare cells with abnormally high DNA content,” according to the European Society for Analytical Cellular Pathology (ESACP) Consensus Report.13

Detection and Typing of HPV

Ten milliliters of the liquid-fixed cytologic sample was centrifuged at 2000g and the pellet was dissolved in 200 μL phosphate-buffered saline. DNA was extracted using the DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.

To detect HPV DNA, a two-tier PCR-direct sequencing (DS) method was performed according to Feoli-Fonseca et al.,14 with modifications. We used the general consensus primers GP5+/GP6+15 and the MY09/MY1116 primers for amplification of HPV DNA. Forty cycles of amplification were run with an initial denaturation at 95 °C for 5 minutes and for 30 seconds in each cycle. Temperatures of 31 °C and 58 °C were set for 30 seconds for annealing of the primer pairs GP5+/GP6+ and MY09/11, respectively. Extension was done at 72 °C for 30 seconds. A final extension step was done at 72 °C for 5 minutes. The integrity of human genomic DNA was verified by PCR amplification of the β-globin gene. This reaction served as a positive control. The amplification products of the two consensus primer pairs and the β-globin PCR were run on a 2% agarose gel and stained with ethidium bromide.

PCR products were purified using the High Pure PCR product purification kit (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's instructions. The sequence of one strand of the purified PCR fragments was determined with the BigDye Terminator sequencing kit (Applied Biosystems, Foster City, CA) using 3–5 pMol of GP5+ or MY09 as the sequencing primers. The results of the sequencing reactions were analyzed on an ABI Prism 310 automated sequencer (Applied Biosystems). The obtained sequences were compared with documented virus sequences available in the GenBank databank using the BLAST program (Blast, Pittsboro, NC).

RESULTS

One hundred twelve cervical cytologic samples were included in this study, 63 of which were identified as LSIL. The remaining 49 cases had the diagnosis of HSIL. PCR-DS found 110 samples (98.2%) to be positive for HPV. The frequency of the individual HPV types is presented in Figure 1. In samples with LSIL (n = 63), 48 (76.2%) were identified with high-risk HPV types, 3 (4.8%) with unknown-risk HPV types, and 9 (14.3%) with low-risk HPV types. The virus type was not classifiable in one case (1.6%). Two samples proved to be negative for the HPV genome (3.1%). In the samples diagnosed as HSIL (n = 49), 47 (95.9%) were identified with high-risk and 2 (4.1%) with unknown-risk types. None of the cases with HSIL presented with a low-risk HPV type.

Figure 1.

Distribution of human papillomavirus (HPV) types detected by polymerase chain reaction-based sequencing in 112 samples with the diagnosis of squamous intraepithelial lesions. Black bars, high-risk HPV; open bars, low-risk HPV; shaded bars, prognostically irrelevant (unknown risk, not classifiable).

Gross genomic changes characteristic for SIL were measured by determining cells with DNA content greater than 5c and 9c. Although 91 of 112 cases (81.2%) presented with cells with a greater than 5c DNA content, aneuploid cells with a greater than 9c DNA content occurred only in 50 of the 112 cases (44.6%) by LSC. A significant difference was observed when comparing events greater than 5c and 9c event in LSIL and HSIL. Of the 63 LSIL cases, 18 (28.6%) had no cells with a DNA content greater than 5c or 9c, whereas 45 cases (63.5%) presented with cells with a DNA content greater than 5c and only 5 cases (7.9%) presented with cells with a DNA content greater than 9c. In contrast, in the majority of HSIL cases, cells with a DNA content greater than 9c were demonstrated (45 of 49, 91.8%; Fig. 2).

Figure 2.

Frequency of cells with greater than 5c and 9c DNA content in 63 samples with low-grade squamous intraepithelial lesions (top) and 49 samples with high-grade squamous intraepithelial lesions (bottom).

The relation between HPV type and rare cells with an abnormally high DNA content was also addressed. Of the 95 high-risk HPV-infected cases, 46 (48.4%) contained cells with a DNA content greater than 9c compared with three of the five cases (60.0%) of unknown-risk HPV-infected type. Conversely, none of the low-risk HPV-infected or the HPV-negative cases displayed such aneuploid cells. The significant difference between the high-risk and the low-risk HPV type infected group was the presence of far more cases with aneuploid cells with a greater than 9c DNA content in the former group. DNA aneuploidy according to HPV type is displayed in Table 1.

Table 1. DNA Aneuploidy According to HPV Type
HPV typenFrequency of samples with 5cFrequency of samples with 9c
  1. HPV: human papillomavirus.

6100
11100
16423828
18631
31973
33875
34111
39110
42300
45330
51442
53440
56210
58442
59211
61433
62110
66651
68110
70331
73221
83110
Not classified111
Negative200
Total1129150

The relative frequency of cells with > 9c DNA content was slightly different between samples with different HPV types. Some of the high-risk HPV types were only poorly associated with such cells (HPV18, 31, 45, 66, 70) whereas the majority of HPV16 and 33-infected samples displayed rare cells with an abnormally high DNA content. In contrast, HPV61, enrolled as an unknown risk type, was associated with cells with a greater than 9c DNA content in three of four cases (Table 1).

The number of cells with > 5c and > 9c DNA content and their relation in the particular samples was also addressed. The frequency of these cells was generally low and ranged between 1 and 100 for cells with > 5c DNA content and between 1 and 19 for cells with > 9c DNA content after the analysis of 10,000 cells. Except for one case, which displayed one greater than 9c cell in the absence of any detectable events (i.e., cells with > 5c DNA content), the number of greater than 5c cells within a sample always exceeded the number of greater than 9c cells (Fig. 3).

Figure 3.

Correlation between cells with a greater than5c and a greater than 9c DNA content detected by laser scanning cytometry in 112 samples with squamous intraepithelial lesions. Except for one case with a single cell with a greater than 9c DNA content and no cells with a DNA content between 5c and 9c, the number of cells with a DNA content greater than 5c is systemically higher than the number of cells with a DNA content greater than 9c.

DISCUSSION

As with advanced cervical carcinoma, HSILs were reported to be strongly associated with high-risk HPV.17, 18 Persisting viral infection and the integration of the HPV genome are responsible for the early steps in squamous cell carcinogenesis.19 DNA hypermethylation,20 the disruption of the normal cell cycle, and chromosomal aberrations21, 22 triggered by the oncogenic virus may lead to nuclear DNA content alterations, the extent of which can be highly variable. In our study, we looked for correlations between HPV type and DNA content alterations. PCR-based HPV typing enabled a reliable demonstration of virtually all types of HPV. In the absence of obvious stem line DNA aneuploidy, which is characteristic for multiclonal precancerous processes, LSC favored the detection of individual rare cells with an abnormally high DNA content according to the definition of the ESACP consensus report on DNA image cytometry.13

With increased cytologic atypia (HSIL) and in the presence of oncogenic high-risk HPV, rare cells with a greater than 9c DNA content evolved. This was not evident in cases with low-risk HPV infection or in HPV-negative samples. The presence of these cells (with > 9c DNA content) was statistically distinctive between LSIL and HSIL.

Highly aneuploid lesions were strongly associated with HSIL and with HPV genome integration.23 However, polyploidization (i.e., the multiplication of the whole genomic DNA) in squamous cells of exophytic lesions (condyloma) is also a well described cytopathic phenomenon that is due to low-risk HPV infection.24 For this reason, the increase in nuclear DNA content to 4c and 8c could not be simply interpreted as an unfavorable finding with oncogenic potential. This had to be considered in this study. In our limited number of low-risk HPV-positive cases, various numbers of cells between 5c and 9c but no cells exceeding 9c DNA content were demonstrated. The most significant difference observed between low- and high-risk HPV-positive SIL samples was the presence of aneuploid cells (with > 9c DNA content) in the majority of the HSIL group. It is noteworthy that cells with greater than 9c aneuploidy were almost exclusively found together with significantly greater numbers of cells with DNA content between 5c and 9c (Fig. 3). This pattern suggests a continuous evolution of cells with > 9c aneuploidy through lower DNA contents by accumulating mitotic errors and progressive genetic alterations.

LSC showed that aneuploid cells with DNA content > 5c or 9c were only occasional events despite the high number of cells analyzed. The appearance of these cells, however, fulfilled the criterion of cellular atypia as demonstrated in the fluorescence microscope of the LSC following repositioning. This was further supported by the cytomorphology of the cells after restaining the slides with the Pap stain. In contrast to previously published data applying interactive DNA cytometry, the presented LSC approach measured all cells randomly placed in the search window, including mature superficial cells and reactive leucocytes. Without preselection on morphologic criteria, the DNA value of the analyzed cells proved to be highly variable and a distinct aneuploid DNA peak could not be seen in the majority of the cases. This is in agreement with recent flow cytometry data, demonstrating the lack of correlation between aneuploid DNA peaks and HPV infection in cervical intraepithelial lesions.25 In that study, however, single cells with abnormally high DNA content were not addressed, probably due to limitations of the method used.

The presence of dysplastic cells with variable DNA aneuploidy in the absence of a true aneuploid clone was characteristic for high-grade cervical lesions. Conversely, we should also point to the paradox, that the evolution of carcinomas with a highly aneuploid stem line is uncommon. One explanation for the presence of highly aneuploid cells with greater than 5c or 9c DNA content detectable in LSIL, HSIL, and even in atypical squamous cells of undetermined significance (ASCUS), may be that they reflect a terminal stage of the dysplastic procedure with limited growth potential. Induced proliferative capacity and the expansion of a transformed clone would be, according to this hypothesis, associated with a relatively low DNA load, possibly evolving from sequential DNA gains and losses.

Aneuploid cells with a greater than 9c DNA content, despite their unclear clonogenic potential, reflect cellular atypia and may be a useful biologic marker. This is further supported by the increase in the frequency of such cells from ASCUS through LSIL to HSIL. Compared with a collection of cases with the cytologic diagnosis ASCUS according to The Bethesda System, a significant increase in the high-risk HPV type was found for SIL (27.3% vs. 84.8%; unpublished data). Similarly, the number of aneuploid cells with a greater than 9c DNA content increased significantly from ASCUS (6.8 %) to SIL (LSIL, 7.9 %; HSIL, 91.8 %). The only reason for this increase was the change in the frequency of cells with a DNA content greater than 9c within the high-risk HPV-positive samples from 25.0% in the ASCUS group to 48.4% in the SIL group. Conversely, there was no measurable change in the other risk categories.

In summary, complex analysis of cervical lesions from liquid-fixed cytologic samples is highly informative. The determination of HPV risk category strongly supported the SIL cytologic finding. Cells with a greater than 9c DNA content were present in a significant portion of the high-risk HPV-positive cases. These cells, which can be demonstrated morphologically by LSC, represent most probably HSIL according to The Bethesda System.

Ancillary