Real-time PCR-based system for simultaneous quantification of human papillomavirus types associated with high risk of cervical cancer

Abstract

We have previously shown that women with a high titer of human papillomavirus type 16 (HPV16) in cervical epithelial cells have an increased risk of developing cervical carcinoma in situ. In order to study the relationship between viral DNA amount and risk of cervical carcinoma for the HPV types most commonly found in cervical tumors, we developed a real-time PCR assay for the detection and quantification of HPV16, -18, -31, -33, -35, -39, -45, -52, -58, and -67. These HPV types are analyzed in two reaction tubes, allowing for independent quantification of three viral types, or groups of viral types, in each reaction. A separate reaction is used for estimating the number of a nuclear single-copy gene and is used to calculate the HPV copy number per genomic DNA equivalent in the sample. The system has a dynamic range from 10(2) to 10(7) HPV copies per assay and is applicable to both fresh clinical samples and DNA extracted from archival samples. Reconstitution experiments, made to mimic infections with several HPV types, shows that individual HPV types can be detected in a mixture as long as they represent 1 to 10% of the main type. The system was evaluated with respect to technical specificity and sensitivity, reproducibility, reagent stability, and sample preparation protocol and then used to analyze clinical samples. This homogeneous assay provides a fast and sensitive way for estimating the viral load of a series of the most frequent oncogenic HPV types in biopsies, as well as cervical smear samples.

FIG. 1.

Standard curves for the HPV16 (A) and HPV31 (B) assays. The threshold cycle (C_t) number is plotted against number of HPV genomes. The datum points represent the mean of 12 independent measurements. r² values: r²_HPV16 = 0.984; r²_HPV31 = 0.972.

FIG. 2.

Standard curves for HPV18 (solid line) and HPV45 (broken line) (A) and HPV33 group (i.e., HPV33, -52, -58, and -67) (B) assays. The threshold cycle (C_t) number is plotted against number of HPV genomes. The datum points represent the mean of 12 independent measurements. r² values: r²_HPV18 = 0.883; r²_HPV45 = 0.981; r²_HPV33 = 0.968; r²_HPV52 = 0.971; r²_HPV58 = 0.960; r²_HPV67 = 0.944.

FIG. 3.

Standard curves for HPV35 (A) and HPV39 (B) assays. The threshold cycle (C_t) number is plotted against the number of HPV genomes. The datum points represent the mean of 12 independent measurements. r² values: r²_HPV35 = 0.963; r²_HPV39 = 0.869.

FIG. 4.

Standard curves for the human gene (HMBS) assay. The threshold cycle (C_t) number is plotted against number of cell equivalents. The datum points represent the mean of 12 independent measurements. r²_HMBS = 0.996.

FIG. 5.

Analysis of individual HPV types in synthetic mixtures made to mimic mixed infections. Each bar represents the mean threshold cycle (C_t) of three independent measures. (A) Quantification of HPV31 in a background of HPV16. The 95% confidence interval is based on the C_t values obtained when no HPV16 is added to the reaction. When the ratio of HPV31 to HPV16 is less than 1:100, the estimate of HPV31 falls outside of the 95% confidence interval. (B) Quantification of HPV16 in a background of HPV31. The 95% confidence interval is based on the C_t value obtained when no HPV31 is added to the reaction. When the ratio of HPV16 to HPV31 is less than 1:10 the estimate of HPV16 falls outside of the 95% confidence interval.

FIG. 6.

DNA yields of three DNA-extraction protocols compared to that of organic extraction (protocol D). Protocols: A, freezing-boiling; B, Wizard kit; C, proteinase K digestion. The number of samples that did not result in measurable amounts of DNA by each protocol was as follows: protocol A, n = 4; protocol B, n = 1; protocol C, n = 4; and protocol D, n = 0. These datum points are excluded. The lines represent the linear regression of the comparisons (broken line, protocol D versus protocol A [r² = 0.877]; thin solid line, protocol D versus protocol B [r² = 0.902]; thick solid line, protocol D versus protocol C [r² = 0.736]).