Human papilloma virus detection and typing in 334 lung cancer patients

Abstract

Background. Unlike cervical, anogenital and oropharyngeal cancers, where high-risk human papillomavirus (hrHPV) has long been known to play a major role, a causative link between HPV and lung cancer has been investigated for decades with discrepant results.

Methods. Lung cancer patients eligible for surgical treatment were tested for the presence of HPV-DNA in excised, fresh frozen lung tumor tissue. Patients that tested positive were further examined for the presence of HPV-DNA in adjacent normal lung parenchyma. HPV detection and genotyping was performed using a polymerase chain reaction (PCR)-based approach and allowed the typing of 13 “high-risk”-HPV-types and 2 “low-risk”-HPV-types.

Results. Of the 334 tumor-DNA samples tested, 13 (3.9%) showed presence of HPV-DNA, of which 12 were of a high-risk HPV type (16, 33, 66). In those tested positive, HPV-DNA was not found in adjacent normal lung tissue. No correlation with smoking or EGFR/KRAS mutation status was seen, and only one of 84 squamous cell carcinomas was HPV-positive.

Conclusion. We conclude that HPV is rarely associated with lung cancer in a Northern European population and in those tested positive, more functional studies are required to determine the role HPV plays in lung cancer oncogenesis.

Lung cancer incidence in never-smokers is comparable to incidence rates of cervical carcinoma [1]. Although smoking is by far the risk factor most strongly linked with this disease, lung cancer among non-smokers is currently the seventh leading cause of cancer death [2]. Other factors, such as occupational exposure (asbestos, radon) [3,4], cooking fumes [5] and environmental pollution [6] have been associated with increased risk. In many cases, however, apparent causative factors are lacking.

Human papilloma virus (HPV) is shown to play an integral role in the etiology of cervical and other anogenital cancers [7], and is also associated with non-genital, especially oropharyngel, cancers [8,9]. Since first mentioned as a potential risk factor in lung cancer tumorigenesis by Syrjanen in 1979 [10], discrepant results concerning detection of HPV in lung cancer tumor tissue have been presented [11].

The prevalence pattern seems based on regional differences, with a higher frequency of HPV- positive cases reported in Asia, and lower in Western Europe and the US [12]. Differences in smoking habits, as well as other potential cultural differences, such as sexual preferences, might influence these results. A lack of consensus of detection methods, as well as methodological flaws, could also play a role in the large discrepancies reported, especially regarding conflicting results on the prevalence of HPV in areas within the same geographical region [11].

No larger study has previously tested for the presence of HPV in lung tumor tissue in a Northern European cohort of primary lung cancer patients. In this study, we wanted to perform nested PCR-based HPV genotyping in a large cohort of lung cancer patients with well-defined clinical and follow-up data. The presence of HPV-DNA was tested in both fresh frozen tumor tissue and surrounding normal tissue in those positive.

Methods

Study population

Three hundred thirty-four patients with newly diagnosed operable lung cancer were included in this study. Each participant signed a written consent form prior to inclusion. Patient inclusion started in March 2006 and ended in December 2011. The participation rate in the study was 53% (based on total number of patients operated on at the regional hospital of study recruitment), selected based on practicalities (study nurse resources). The diagnosis of lung cancer prior to surgery was made by use of standard diagnostic methods, and a histopathological diagnosis on resected specimen was independently reviewed by a pathologist associated with the study.

Patient and tumor characteristics are summarized in Table I.

Table I. Patient and tumor characteristics.

Variable	Female	Male	Total
Number of patients (%)	168 (50.3)	166 (49.7)	334
Age (years)
Median (range)	65.3 (40.2–84.1)	66.4 (33.9–84)	66.1 (33.9–84.1)
Histology
Adenocarcinoma	111	87	198
BAC	5	3	8
Squamous cell carcinoma	28	56	84
Large cell carcinoma	12	12	24
Carcinoid	9	6	15
Small cell carcinoma	2	2	4
Stage
Ia	51	41	92
Ib	42	51	93
IIa	34	34	68
IIb	8	17	25
IIIa	30	20	50
IIIb	–	1	1
IV	3	2	5
Smoking history
Never smoker	20	7	27
Current smoker	52	57	109
Former smoker	96	102	198
Cumulative smoking
(Pack years; mean, range)	25 (0–82)	32 (0–110)	27 (0–110)

The study was approved by the institutional review board and the Regional Ethics Committee.

Sample collection

Tumor and corresponding tissue from adjacent normal lung tissue was collected during lung cancer surgery at Oslo University Hospital, Rikshospitalet. Tissue samples were immediately frozen in liquid nitrogen and stored at −80°C until processing. Prior to HPV analysis, in over 30% of samples, frozen sections from the sampled tumor and normal tissue were prepared for Hematoxylin & Eosin- staining and reviewed by a pathologist for confirmation of representativity. TTF-1 staining and EGFR- and KRAS mutational analyses were performed according to local routine.

Detection and typing of HPV-DNA

Total DNA was extracted from sections of surgically excised fresh frozen lung tumor tissue as well as surrounding normal tissue using the Maxwell DNA purification method (Promega, Madison, WI, USA), following the manufacturer's instructions. The amount of DNA was measured with a UV spectrophotometer (Nanodrop, Thermo Scientific, Wilmington, DE, USA). The samples were HPV genotyped either using a nested multiplex PCR method as described by Sotlar and coworkers [13], or according to the method described by Lindh and coworkers [14]. Genotype specific primers allow the typing of 13 “high-risk”-HPV (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59 and 68) and 2 “low-risk”- HPV (6 and 11). Additional in-house assays used to detect HPV 68b and HPV 66 was included to meet WHO test panel requirements. Distilled water was used as negative controls. To facilitate the application of these type-specific nested PCR primers in a multiplex PCR assay, they were selected to amplify products of markedly different sizes. Three primer cocktails, each containing four to five type-specific primers, were used, along with Beta-actin as an internal template control. Clinical samples with known genotypes were included as positive controls. The PCR products from the nested PCR are detected by capillary electrophoresis (fragment analysis) on the ABI3130xl (Invitrogen, Carlsbad, CA, USA) according to standard protocol. Results were analyzed using GeneMapper version 4.0 (Applied Biosystems, Foster City, CA, USA) [14].

Results

Of the 334 DNA samples tested, 13 (3.9%) showed presence of HPV-DNA. Twelve tested positive for high-risk HPV (16, 33, 66), whereas one tested positive for HPV11 (low risk). The majority of HPV-positive cases were male (10 of 13) and adenocarcinomas (11 of 13). In patients who tested positive for HPV-DNA, adjacent normal lung parenchyma was also tested for the presence of HPV-DNA with negative results. One patient had a former or concurrent diagnosis of another primary cancer. Of the 334 patients, 92% were former or current smokers as were 11 of the 13 who tested positive for HPV-DNA. Two of the 13 were EGFR-mutated (one exon 19 deletion and one L858R exon 21 point mutation). Eleven of the 13 HPV-positive were KRAS-tested, four of these harbored mutation in either exon 12 or 13. Ten of 12 TTF-1-tested HPV-positive cases were TTF-1 positive, including the only case with a prior cancer (colon cancer).

Results are summarized in Tables I and II.

Table II. HVP status and patient characteristics in HPV-DNA positive.

Variable	Female	Male	Total
HPV positive (%)	3 (1.8)	10 (6)	13 (3.9)
HPV strain:
11	–	1	1
16	1	3	4
33	2	5	7
66	–	1	1
Age (years)
Median (range)	59 (55.5–64.7)	72.5 (43.8–79.1)	64.7 (43.8–79.1)
Histology
Adenocarcinoma	2	9	11
Squamous cell carcinoma	–	1	1
Adenosquamous carcinoma	1	–	1
Stage
Ia	–	1	1
Ib	1	6	7
IIa	2	1	3
IIb	–	1	1
IIIa	–	1	1
Smoking history
Never smoker	1	1	2
Current smoker	1	3	4
Former smoker	1	6	7
Cumulative smoking
(Pack years; mean, range)	47 (0–55)	26.5 (0–45)	27 (0–55)

Discussion

Of the 334 operated lung cancer patients in our study, less than 4% tested positive for HPV-DNA. In those tested positive, there was no clear correlation to smoking behavior, histology or mutational status. No HPV-DNA was found in adjacent non-cancerous tissue in HPV-positive cases. The low prevalence of HPV-positive cases supports the postulate that HPV does not play a significant role in the etiology of most lung cancers in Western populations.

Compared to the number of studies published on the link between HPV and cervical cancer, research on the subject of HPVs possible etiological role in lung cancer is modest. In currently available publications, there are large discrepancies regarding results of the prevalence of HPV in tumor tissue. Interpretation of results is further complicated by different tissue handling, detection methods, and in many cases; small sample sizes [12].

Three recent Western European studies with larger patient sample sizes (n = 218/223/399) support our conclusion of no association between HPV and lung cancer [15–17]. As in our study, they used a PCR-based approach for the detection and subtype-specification of HPV. Whereas we found a HPV detection rate of 3.9%, these studies had detection rates in the range of 0–1.8%. Smaller European studies have shown larger discrepancies in HPV prevalence in lung cancer tissue. In a systematic literature review by Klein et al from 2009, only three studies regarding the prevalence of HPV in lung cancer included over 100 cases, all with detection rates less than 3% [11]. No larger European study (> 50 samples) with unselected cases of lung cancer have shown HPV detection rates over 12.8% [18–20]. The reported mean incidence was 17% in all European studies, whereas in Asia it was higher; at 35.7% [11]. Recently, a North American study showed a 15% positivity rate in 208 early stage NSCLC samples [21], whereas a recent large Canadian study concluded with HPV-positivity being a rarity in NSCLC (five positives in 336 samples, but all five suspected being metastases) [22]. In our cohort of HPV-positive samples, only one patient had a history of previous cancer, but this tumor was TTF-1-positive, which confirms the lung origin. In fact, over 80% of HPV-positive tumors were TTF-1 positive.

It has been posited that different tissue handling and a variance in the sensitivity of various PCR-based approaches may explain part of these variances. We used fresh-frozen tissue in our study. Differences in DNA quality based on tumor tissue being fresh-frozen or paraffin embedded have been reported; when tissue is paraffin embedded, significant DNA degradation occurs which can impact PCR amplification of longer fragments [23]. Inappropriate PCR primers may also play a role in false-negative HPV detection rates, as part of the L1 and E2 gene might be lost during integration of the virus [24]. A standardization of detection methods could possibly contribute to more homogenous HPV genetic expression data. However, even studies that have used the same methods have failed to reproduce similar results [18] and suggest that ethnicity and geography could play a role in the variation.

In our cohort 27 (8%) of all cases were never-smokers, of whom two tested positive for HPV-DNA. The proportion of never-smokers among lung cancer patients vary greatly between regions and sex. As expected, the majority of never-smokers in our cohort were female (74%). Lung cancer incidence is higher among women than men who have never smoked [1]. Despite this fact, less than 20% of female lung cancer patients are never-smokers in the western world, while epidemiological studies show a larger proportion of never-smokers among Asian female lung cancer patients [25]. Based on the low proportion of female smokers and a high prevalence of HPV 16/18 in lung tumor tissue seen in many Asian studies, it may seem that HPV can play a role as a co-carcinogen in never-smoking lung cancer patients. Interestingly, in a recent Finnish case control study they found no association between lung cancer and HPV16 and HPV 18 seropositivity in smoking or non-smoking women [26].

Some authors have presented data on a high frequency of HPV-positivity in EGFR-inhibitor- responsive lung cancer patients in Asian patients, pointing to a potential correlation between HPV and activating mutations in the EGFR-gene [27]. Our data are too limited to draw any firm conclusion, nevertheless, two EGFR-mutated cases of the 13 HPV-positives in our cohort is close to the expected number from the general population. Also the KRAS-mutation frequency (4 of 11 tested, 30%) is as expected. Interestingly, the HPV frequency was significantly higher in adenocarcinomas than in squamous cell carcinomas (5.6% and 1.2%, respectively).

If suspected to play a role in the etiology of cancer development, one can argue that an oncogenic virus should only be present in tumor cells and not adjacent normal tissue. In the 13 cases tested positive in our cohort, HPV-DNA was indeed only present in the tumor. A few other studies have also shown presence of HPV exclusively in tumor tissue [28–30]. Although this finding may support an oncogenic role for the virus, we did not examine the expression of HVP E6 and E7 oncoproteins. Expression of these oncogenes seems essential for malignant transformation to occur. Analyses on mutational rate of p53, known to be lower in HPV-positive than in HPV-negative head and neck carcinomas, is similarly warranted in further studies [31].

The HPV-frequency in our cohort was unevenly distributed between the sexes; 1.8% in females versus 6% in males. In our case, the total number of HPV-DNA positive cases is too low to give much significance to the asymmetrical gender distribution. It is interesting to note, however, that our results mirror that of a recent US-based cross-sectional study by Gillison et al., where the incidence of oral HPV infection was about three times higher in men than women (10.1% vs. 3.6%) [32]. Information of HPV epidemiology worldwide is absent. A systematic literature review on prevalence of cervical HPV-DNA in women with normal cervical cytology worldwide estimates that around 291 million women worldwide are carriers of HPV-DNA [33]. The pattern of direct viral transmission in the etiology of cervical cancer has been well established, yet it is not clear how viral load is transferred to bronchial cells. Several pathways of transmission have been postulated. The presence of HPV in other sites than the genitoanal and oro-pharynx indicates other routes of transmission than through sexual contact. Hennig et al. showed a high prevalence of HPV-positive primary lung cancers (49%) in patients with previous high-grade cervical intraepithelial neoplasia (CIN III) [34]. An increased cumulative risk of secondary primary cancers in the lung in women with primary cancers in the genito-anal sites have also been found elsewhere [35,36]. Interestingly, as mentioned above, all five HPV-positive cases found in a large Canadian study had a previous history of HPV-associated squamous cell carcinoma of other organs [22]. Furthermore, similar findings were reported in a recent Dutch study, where all three HPV-positive cases from a cohort of 223 cases with cancer in the lung, had a previous history of either oropharyngeal or cervical cancer [17].

As patients selected for inclusion in this retrospective study were all considered operable, we cannot assume that our cohort of patients necessarily depicts an accurate representation of Norwegian lung cancer patients. Although 51 (15.3%) patients were stage III lung cancers and five patients stage IV (1.5%), an association between more aggressive lung cancers, detected at more advanced stages, and HPV infection is possible. However, 62% of all HPV-positive cases were in stage I, whereas this stage represented 55% of the total cohort, thus there was no selection of more advanced tumors in the HPV-positive group. Likewise, in a similar study by Coissard et al. where only a total of four of 218 cases tested positive for HPV virus, 64 cases (29%) were stage III (1 case stage IV) [16].

In conclusion, with the use of a well-tested and verified PCR-based approach to HPV-DNA detection, we found that HPV is rarely associated with lung cancer in a Northern European study population and in those tested positive, more functional studies are required to determine the exact role HPV plays in lung cancer oncogenesis.

Acknowledgments

The technical assistance provided by research nurse Ingjerd Solvoll is highly appreciated.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper. The study was funded by South-Eastern Norway Regional Health Authority and The Research Council of Norway.