SEARCH

SEARCH BY CITATION

Keywords:

  • Betapapillomavirus;
  • case–control study;
  • cutaneous squamous cell carcinoma;
  • organ transplant recipient

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure
  9. References

We examined the association between betapapillomavirus (betaPV) infection and cutaneous squamous cell carcinoma (SCC) in organ transplant recipients. A total of 210 organ transplant recipients with previous SCC and 394 controls without skin cancer were included. The presence of 25 betaPV types in plucked eyebrow hairs was determined using a human papillomavirus (HPV) DNA genotyping assay, and antibodies for the 15 most prevalent betaPV types were detected using multiplex serology. We used multivariate logistic regression models to estimate associations between various measures of betaPV infection and SCC. BetaPV DNA was highly prevalent (>94%) with multiple types frequently detected in both groups. We found a significant association between SCC and the concordant detection of both antibodies and DNA for at least one betaPV type (adjusted OR 1.6; 95% CI 1.1;2.5). A borderline-significant association with SCC was found for HPV36 (adjusted OR 2.4; CI 1.0;5.4), with similar associations for HPV5, HPV9 and HPV24. These data provide further evidence of an association between betaPV infection and SCC in organ transplant recipients. Confirmation of a betaPV profile predictive of risk for SCC may pave the way for clinically relevant pretransplant HPV screening and the development of preventive and therapeutic HPV vaccination.


Abbreviations: 
BetaPV

betapapillomavirus

CI

confidence interval

ELISA

enzyme linked immunoabsorbant assay

EV

epidermodysplasia verruciformis

HPV

human papillomavirus

MFI

median fluorescence intensity

OR

odds ratio

OTR

organ transplant recipient

PV

papillomavirus

SCC

squamous cell carcinoma

SD

standard deviation

UV

ultraviolet

VLP

virus-like particle

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure
  9. References

Cutaneous squamous cell carcinoma (SCC) is the most common malignancy in organ transplant recipients (OTR), with an approximately 100-fold increased incidence compared to the general population (1,2). The strength of association between solid organ transplantation and SCC parallels that seen for other virally associated posttransplant cancers such as posttransplant lymphoproliferative disease (Epstein–Barr virus), Kaposi sarcoma (human herpes virus 8) and hepatocellular cancer (hepatitis B virus). In contrast, the incidence of the majority of common human cancers, including lung, colon, breast and prostate, is not substantially greater than in the general population (3–6). Whether the greatly increased incidence of cutaneous SCC in OTR could reflect a viral etiology, specifically human papillomaviruses (HPV), has been intensely investigated over the years, but remains uncertain.

The first evidence that HPV types from the betapapillomaviruses (betaPV) genus may play a role in SCC came from observations of the rare genodermatosis, epidermodysplasia verruciformis (EV). People with EV are predisposed to HPV infection and develop SCC harboring specific betaPV types (notably HPV5, HPV8, HPV20) on UV exposed sites (7,8). BetaPV DNA has been detected in SCC both from OTR and immunocompetent people (9,10, 11, 12), together with expression of betaPV genes (13,14), but usually with viral loads of less than one copy per cell (15). Functional data support a potential role for betaPV as cocarcinogens with UV, but with mechanisms likely to differ fundamentally to those established in anogenital cancer (16). Specifically, the virus fails to integrate into the host genome and the viral oncoproteins E6 and E7 do not target p53 and pRb, with the exception of HPV38 (17–21). BetaPV abrogation of UV-induced apoptosis (22,23) and interference with DNA repair and cell-cycle arrest (24–26) may contribute to skin carcinogenesis. Additionally, betaPV E7 promotes keratinocyte invasion (27,28) and spontaneous skin carcinogenesis is a feature of HPV8 transgenic mice (29–31).

Detection of HPV DNA has been associated with SCC and its precursor lesions, actinic keratoses (AK), in case–control studies (32–34), although no specific betaPV types prevail (34). However, HPV is ubiquitous (35–37) with betaPV DNA highly prevalent in normal skin (38,39) and hair follicles the likely reservoir (40,41). In addition, infections are frequently multiple, especially in people who are immunosuppressed, and prevalence increases with duration of transplantation (41,42), posing epidemiological challenges to establishing a hierarchy of risk for different betaPV types (43). The presence of betaPV DNA is not necessarily accompanied by a detectable antibody response and precisely what determines the induction and intensity of seroresponses is unknown (44), although seropositivity increases with age until middle age (45). Detection of antibodies to betaPV-8 and -38 using VLP ELISA methods is an independent risk factor for immunocompetent SCC in some studies (46–48), but this is not a consistent finding and differing methodologies limit the comparability of these studies (reviewed in 49).

The ubiquity of betaPV DNA in normal skin and hairs suggests that carriage of betaPV DNA may often represent a latent viral infection. In contrast, HPV antibodies to viral capsid proteins imply that there was an infection at some point in the past that was sufficient to evoke an immune response (42). Thus, DNA positivity is more likely to be clinically relevant in the presence of a concordant antibody response. We aimed to contribute to this knowledge base by systematically examining measures of betaPV DNA and antibody prevalence using state-of-the methodologies in the largest case–control study in OTR yet reported. We provide further evidence for an association between betaPV and SCC risk in people who are immunocompromised.

Material and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure
  9. References

Study population

This hospital-based case–control study was designed to examine a possible association between cutaneous SCC and betaPV infection. Patients with solid organ transplants who had been transplanted at least 2 years were recruited in the following hospitals; Leiden University Medical Center, Leiden, The Netherlands; Barts and the London NHS Trust, London, UK; Hospices Civils de Lyon, France; Ospedali Riuniti di Bergamo, Bergamo, Italy; and Ospedale Civile Maggiore, Verona, Italy. Recruitment of OTR took place between 2003 and 2005 as part of a EU Commission collaborative research grant (QLK2-CT-2002-0117).

Cases had a history of histologically confirmed posttransplant primary cutaneous SCC (excluding in situ carcinoma) and were recruited from outpatient dermatology and transplant clinics. Controls had no history of skin cancer and were selected from the same clinics. We attempted to match to cases by sex, age and time since transplantation (2–7. 8–12, 13–17, 18–22, 23 and more years) in a 2:1 ratio. Patients with brown or black skin (Fitzpatrick skin type V and VI) were excluded, as were patients with only liver, lungs andpancreas transplants. The local medical research ethics committees of the hospitals in the four countries approved the study design. All participants gave their written informed consent.

Questionnaire

Standardized questionnaires were used to gather information from participants in each country, including history of keratotic skin lesions; hair and eye color; ability to tan and sun reactivity (Fitzpatrick skin phototype); amount of spring and summertime weekday and weekend sun exposure during the longest held occupation; recreational sun exposure and number of painful sunburns before the age of 20. A full description of the collection of data and counting of skin lesions is available (50).

Physical examination

Trained dermatologists conducted full skin examinations on all participants and biopsied suspected skin cancers for histological diagnosis. Common warts and palmoplantar viral warts were counted separately and all other hyperkeratotic skin lesions (AK, seborrhoeic warts, flat warts and hyperkeratotic papillomas) were combined because of the difficulty of reliably distinguishing such hyperkeratotic skin lesions in OTR.

BetaPV detection and genotyping

The presence of betaPV DNA in eyebrow hairs was determined (36,41); 8–10 hairs were plucked from each participant and stored at −70°C until processing. DNA extraction was by the QIAamp DNA Mini Kit (Qiagen GmbH, Hilden, Germany) and genotyping was carried out with the skin (beta) HPV prototype research assay (Diassay BV, Rijswijk, The Netherlands) as described previously (36,41). This assay includes probes to 25 betaPV types including betaPV species 1 (HPV 5, 8, 12, 14, 19, 20, 21, 24, 25, 36, 47, 93), betaPV species 2 (HPV 9, 15, 17, 22, 23, 37, 38, 80), betaPV species 3 (HPV 49, 75, 76), betaPV species 4 (HPV 92) and betaPV species 5 (HPV 96). All DNA analyses were performed blinded with respect to case or control status.

BetaPV serology

Sera were stored at −20°C and shipped on dry ice to the German Cancer Research Center (DKFZ) in Heidelberg, Germany. Serum analyses included antibodies to the L1 proteins of the betaPV species 1 (HPV types 5, 8, 20, 24, 36, 93), betaPV species 2 (HPV 9, 15, 17, 23, 38), betaPV species 3 (HPV 49, 75, 76) and betaPV species 4 (HPV 92). In addition, the VP1 proteins of the human polyomaviruses BK and JC were included as control antigens. Samples were analyzed simultaneously for all antigens by multiplex serology, with investigators blinded to the clinical status of the samples, as previously described (45,51). Study sera were analyzed once on three consecutive assay days and the glutathione-casein coupled bead sets were loaded with their respective antigens in one batch. A quality control panel (QC) of 94 sera was included each day resulting in three QC data sets to determine inter-day variation. Correct antigen loading was verified by using 28 reference sera with known HPV antibody patterns from two earlier studies. Every day, binding of glutathione S-transferase (GST)–L1–tag fusion proteins to glutathione–casein-coated beads was quantified by antitag monoclonal antibody. Antitag median fluorescence intensity (MFI) values for the first day varied less than twofold (range 7773–14904 MFI) and the interantigen coefficient of variation (CV) was 17.2%, indicating similar full length L1 fusion protein density for the different HPV types. Antigen-specific ratios of antitag MFI for the second (mean ± SD, 0.99 ± 0.19) and third (0.99 ± 0.11) day revealed stable antigen binding to the beads throughout the 3 assay days. Cut-off values to define seropositivity for all antigens were set to 200 MFI as described previously (45). Two recent studies have shown that changing the cut off used to determine positivity does not substantially alter the OR, nor the proportion of people who remained serostable for most HPV types (52,53).

Statistical analysis

We calculated the odds ratios (OR) and 95% confidence intervals (95% CI) associated with various measures of betaPV infection using multivariable logistic regression analysis in SPSS software version 16 for Windows. Our approach was to fit minimally adjusted models. Potential confounding variables that did not alter the odds ratio or significantly improve the fit of the model were not included. Given the exploratory nature of our study, and given that we were not testing the ‘universal null hypothesis’, we followed the advice of Rothman (54) and Savitz & Olshan (55) in electing not to perform adjustment for multiple testing.

Exposure variables

We categorized participants according to whether or not betaPV DNA for any of 25 viruses was detected in plucked hairs and then according to the number of betaPV types detected. Similarly, participants were categorized depending on whether or not antibodies to any of 15 betaPV were detected and on the number of viruses to which antibodies were detected. We defined a new variable ‘concordant measures of infection’ by classifying participants according to whether or not they were both antibody and DNA positive for the same betaPV type. We performed type-specific analyses, and also created a summary concordance variable in which we classified participants into ‘antibody negative, irrespective of DNA status’, ‘antibody positive but with no types for which DNA was also found’ and ‘antibody positive with at least one type concordant for DNA’.

Possible confounding variables

Types of transplants were grouped into ‘renal’ and ‘other’, which included heart, combined kidney–pancreas, kidney–liver, and kidney–heart. We did not include lung, liver or pancreas only transplant recipients because the participating centers had limited access to these patients. Skin phototype was classified as ‘olive’ and ‘medium/fair’ based on questions about tanning ability, sun reactivity and Fitzpatrick skin type (50). The number of hours of work and recreational sun exposure were grouped into low (0–3 hours/day) and high (4 or more hours/day) exposure.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure
  9. References

Characteristics of the study populations (Table 1)

Table 1.  Baseline characteristics of the organ-transplant recipients and associations with squamous cell carcinoma
 Controls (N = 394) N (%)Squamous cell carcinoma (N = 210) N (%)
  1. Numbers do not always add up to the total number because of some missing values.

  2. Aza, azathioprine; MMF, mycophenolate mofetil; CyA, ciclosporin A; Tac, tacrolimus.

Sex
 Women115 (29)49 (23)
 Men279 (71)161 (77)
  (p = 0.123)
Type of organ
 Kidney334 (85)169 (80)
 Kidney and pancreas37 (9)31 (15)
 Heart23 (6)10 (5)
  (p = 0.128)
Immunosuppressive therapy
 Aza in any combination184 (47)114 (54)
 MMF in any combination117 (30)44 (21)
 CyA or Tac without Aza or MMF93 (23)52 (25)
  (p = 0.061)
Age at transplantation (years)
 Median40.145.3
 25–75%30.6–50.633.9–53.8
  (p = 0.004)
Age at investigation (years)
 Median54.160.5
 25–75%46.1–61.954.1–66.6
  (p < 0.0001)
Total follow-up time (years)
  2–7114 (29)35 (17)
  8–12109 (28)51 (24)
 13–1778 (20)43 (21)
 18–2250 (13)41 (20)
 23 or more41 (10)38 (18)
  (p = 0.001)
Skin phototype
Dark/olive196 (50)79 (38)
Medium135 (34)90 (43)
Fair63 (16)41 (19)
  (p = 0.017)
Average daily sun exposure
 1–3 hours233 (59)122 (58)
 4 or more hours161 (41)88 (42)
  (p = 0.804)
Sunburns before the age of 20 years
 0215 (55)90 (43)
 1–4124 (31)84 (40)
 5 or more55 (14)26 (17)
  (p = 0.023)
Keratotic skin lesions
 0132 (34)22 (11)
 1–9200 (51)108 (52)
 10–4925 (6)24 (11)
 50 and more34 (9)54 (26)
  (p < 0.0001)
Common viral warts
 0214 (55)107 (51)
 1–49161 (41)72 (35)
 50 and more16 (4)29 (14)
  (p < 0.0001)

A total of 210 OTR with a history of SCC and 394 controls without skin cancer were recruited. The majority (83%) were renal allograft recipients and there were more men (73%). We experienced some difficulty in matching by age and time since transplantation, so younger controls were recruited in some cases. The cases were older at time of investigation (p < 0.0001) and were also older at transplantation (p = 0.004), and the time period after transplantation was longer in cases than in controls (p = 0.001) as detailed in Table 1.

Associations of patient characteristics with SCC (Table 1)

All 210 cases (35%) had had at least one SCC (range 1–45, mean 3.6, median 1). Of these 210 cases, 85 (41%) also had at least one BCC (range 1–27, mean 3.4, median 2). Skin phototype (olive, medium, fair) and painful sunburns before the age of 20 were associated with an increased risk of SCC, as were keratotic skin lesions and common warts (Table 1) (50). Average daily sun exposure was not significantly associated with SCC in this study population (Table 1).

Prevalence of betaPV in OTR

The prevalence of betaPV-DNA was 94% in OTR without a history of skin cancer and 97% in those with SCC (Table 2). Antibodies to betaPV were present in 53% of controls and 62% of cases (Table 2). The prevalence of individual betaPV infections is presented in Figure 1. HPV 23, 24, 36, 38, 5 and 15 were the most prevalent betaPV DNA types detected in eyebrow hairs, with a prevalence of more than 30% in controls (Figure 1). Over 20% of controls were antibody positive to HPV 8, 15, 38, 49 and 17 (Figure 2) and approximately 10% of controls were concordant for DNA and antibodies to HPV 8, 15 and 38 (Figure 3).

Table 2.  Association between markers of betapapillomavirus infection and squamous cell carcinoma
 Controls N = 394 N (%)Squamous cell carcinoma N = 210 N (%)OR* (95% CI)
  1. Numbers do not add up to the total number because of some missing values.

  2. *The odds ratio is adjusted for sex, age, years after transplantation and country.

Presence of betaPV DNA
 Negative (reference)24 (6)6 (3)1
 Positive361 (94)200 (97) 2.1 (0.78;5.5)
 0 types (reference)24 (6)6 (3)1
 1–4 types165 (43)69 (33)1.8 (0.67;5.0)
 5 or more types196 (51)131 (64)2.3 (0.84;6.1)
  P for trend  p = 0.082
Presence of antibodies to betaPV
 Negative (reference)181 (47) 79 (38)1
 Positive205 (53)129 (62)1.4 (0.95;2.0)
 0 types (reference)181 (47)79 (38)1
 1–3 types128 (33)73 (35)1.4 (0.89;2.1)
 4 or more types77 (20)56 (27)1.4 (0.88;2.3)
  P for trend  p = 0.114
Antibody responses to betaPV in the absence or presence of concordant betaPV DNA
 Antibody negative regardless of DNA178 (47)77 (38)1
 Antibody positive without concordant DNA80 (21)32 (16)0.99 (0.58;1.7) 
 Antibody positive with concordant DNA119 (32)95 (46)1.6 (1.1;2.5) 
image

Figure 1. Association of betaPV DNA and cutaneous SCC in OTR. Percentage of cases and controls with detectable DNA in plucked eyebrow hairs: (a) for the 15 betaPV types for whom serology was also available, (b) for the 10 additional betaPV types without serological analyses.

Download figure to PowerPoint

image

Figure 2. Association of betaPV antibodies and cutaneous SCC in OTR. Percentage of cases and controls with detectable antibodies to 15 betaPV types using multiplex serology.

Download figure to PowerPoint

image

Figure 3. Association of betaPV antibody in the presence of concordant DNA and cutaneous SCC in OTR. Percentage of cases and controls with detectable antibodies and detectable DNA to the same betaPV type.

Download figure to PowerPoint

BetaPV infection and association with SCC

Association between markers of betaPV infection and SCC is displayed in Tables 2 and 3. We found a moderate association between the presence of any HPV DNA in plucked eyebrow hairs and SCC, although this was not statistically significant (adjusted OR 2.1; 95% CI 0.78;5.5). There was no evidence of increasing risk with an increasing number of HPV types present (p for trend 0.082). We found a weak association between antibody positivity for any betaPV type and SCC (adjusted OR 1.4; 95% CI 0.95;2.0). Those who were both antibody and DNA positive for at least one concordant betaPV type were at a significantly greater risk of SCC (adjusted OR 1.6; 95% CI 1.1;2.5). In contrast, this association was not apparent if DNA and antibodies were both present, but for different betaPV types (adjusted OR 0.99; 95% CI 0.58;1.7).

Table 3.  Association between markers of betapapillomavirus infection and squamous cell carcinoma (type specific)
Presence of betaPV DNA and/or antibody response (type specific)Controls N = 377 N (%)Squamous cell carcinoma N = 204 N (%)OR* (95% CI)
  1. *Adjusted for sex, age at transplantation, number of years after transplantation and country.

Beta 1 types
HPV5
  DNA and antibody both negative245 (65)112 (55)1
  DNA positive and antibody negative99 (26)59 (29)1.1 (0.75;1.7)
  DNA negative and antibody positive16 (4)14 (7)2.2 (0.95;5.0)
  DNA and antibody both positive18 (5)19 (9)2.0 (0.95;4.3)
HPV8
  DNA and antibody both negative225 (60)101 (50)1
  DNA positive and antibody negative54 (14)37 (18)1.6 (0.96;2.7)
  DNA negative and antibody positive58 (15)35 (17)1.3 (0.74;2.1)
  DNA and antibody both positive40 (11)31 (15)1.4 (0.80;2.5)
HPV20
  DNA and antibody both negative265 (70)130 (64)1
  DNA positive and antibody negative57 (15)40 (20)1.2 (0.73;2.0)
  DNA negative and antibody positive37 (10)25 (12)1.2 (0.66;2.2)
  DNA and antibody both positive18 (5)9 (4)0.77 (0.31;1.9)
HPV24
  DNA and antibody both negative223 (59)92 (45)1
  DNA positive and antibody negative105 (28)74 (36)1.4 (0.89;2.1)
  DNA negative and antibody positive31 (8)17 (9)1.2 (0.59;2.4)
  DNA and antibody both positive18 (5)21 (10)2.0 (0.94;4.1)
HPV36
  DNA and antibody both negative218 (58)103 (51)1
  DNA positive and antibody negative125 (33)74 (36)1.1 (0.73;1.7)
  DNA negative and antibody positive21 (6)10 (5)0.98 (0.41;2.3)
  DNA and antibody both positive13 (2)17 (8)2.4 (1.0;5.4)
HPV93
  DNA and antibody both negative272 (72)144 (71)1
  DNA positive and antibody negative87 (23)49 (24)0.93 (0.60;1.4)
  DNA negative and antibody positive14 (4)6 (3)0.80 (0.29;2.2)
  DNA and antibody both positive4 (1)5 (2)1.7 (0.42;7.0)
Beta 2 types
HPV9
  DNA and antibody both negative251 (67)115 (56)1
  DNA positive and antibody negative72 (19)45 (22)1.1 (0.70;1.8)
  DNA negative and antibody positive39 (10)24 (12)1.2 (0.65;2.2)
  DNA and antibody both positive15 (4)20 (10)2.1 (0.98;4.6)
HPV15
  DNA and antibody both negative195 (52)93 (46)1
  DNA positive and antibody negative81 (21)43 (21)1.1 (0.65;1.7)
  DNA negative and antibody positive64 (17)47 (23)1.5 (0.91;2.4)
  DNA and antibody both positive37 (10)21 (10)1.0 (0.55;2.0)
HPV17
  DNA and antibody both negative251 (67)120 (59)1
  DNA positive and antibody negative48 (13)27 (13)1.2 (0.68;2.2)
  DNA negative and antibody positive61 (16)46 (23)1.5 (0.91;2.4)
  DNA and antibody both positive17 (4)11 (5)1.2 (0.49;2.7)
HPV23
  DNA and antibody both negative187  (50)83  (41)1
  DNA positive and antibody negative147  (39)87  (42)1.1  (0.71;1.6)
  DNA negative and antibody positive19  (5)16  (8)1.6  (0.71;3.4)
  DNA and antibody both positive24  (6)18  (9)1.2  (0.58;2.4)
HPV38
  DNA and antibody both negative185  (49)75  (37)1
  DNA positive and antibody negative104  (28)64  (31)1.5  (0.95;2.3)
  DNA negative and antibody positive47  (12)33  (16)1.5  (0.84;2.6)
  DNA and antibody both positive41  (11)32  (16)1.5  (0.80;2.7)
Beta 3 types
HPV49
  DNA and antibody both negative264  (70)127  (62)1
  DNA positive and antibody negative29  (8)20  (10)1.5  (0.77;2.9)
  DNA negative and antibody positive75  (20)50  (25)1.2  (0.79;2.0)
  DNA and antibody both positive9  (2)7  (3)0.99  (0.33;2.9)
HPV75
  DNA and antibody both negative301  (80)148  (73)1
  DNA positive and antibody negative31  (8)24  (12)1.1  (0.58;2.0)
  DNA negative and antibody positive40  (11)27  (13)1.1  (0.59;1.9)
  DNA and antibody both positive5  (1)5  (2)1.8  (0.48;6.6)
HPV76
  DNA and antibody both negative300  (80)148  (73)1
  DNA positive and antibody negative33  (9)25  (12)1.5  (0.78;2.7)
  DNA negative and antibody positive35  (9)27  (13)1.4  (0.80;2.6)
  DNA and antibody both positive0  (2)4  (2)0.54  (0.15;2.0)
Beta 4 type
HPV92
  DNA and antibody both negative304  (80)145  (71)1
  DNA positive and antibody negative26  (7)24  (12)1.8  (0.94;3.4)
  DNA negative and antibody positive40  (11)29  (14)1.4  (0.83;2.5)
  DNA and antibody both positive7  (2)6  (3)1.4  (0.42;5.0)

Table 3 displays the concordant variable analysis individually for the 15 betaPV types for which both DNA and antibody results were available. Using concordant type-specific antibody and DNA positivity as the measure of infection, we found a borderline significant association with SCC for HPV36 (adjusted OR 2.4; CI 1.0;5.4) with similar associations seen for HPV9 and HPV24. For HPV5 there was a moderate association with SCC for positive serology (OR 2.2; CI 0.95;5.0), with no additional risk seen by adding concordant DNA (OR 2.0; CI 0.95;4.3). Similarly, the risk associated with HPV38 resulted from the presence of either DNA or antibodies without additional risk conferred by concordant positivity.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure
  9. References

This is the first study in which the association between SCC risk and both betaPV DNA and antibodies was assessed in immunocompromised persons. We found that those people who had both betaPV DNA and antibodies present for the same HPV type (concordant DNA and antibodies) were at significantly increased risk of SCC, even if individually these measures of HPV infection (DNA or serum antibodies) did not show a statistically significant association with SCC.

BetaPVs are highly prevalent in OTR populations; 94% of patients without skin cancers carried DNA in eyebrow hairs and 53% have betaPV antibodies, increasing to 97% and 62%, respectively, in those with a history of skin cancer. Similar spectra of betaPV DNA and seroresponses were found in cases and controls (Figures 1 and 2). Although the spectrum of betaPVs detected is not altered compared to the immunocompetent population (5,41), immunosuppression does appear to increase the risk of HPV infection. Previously, in a study involving this same population of OTR without skin cancer, we showed that immunosuppression was significantly associated with betaPV infection (adjusted OR 1.6; CI 1.1–2.5), as well as with multiple betaPV infection (adjusted OR 1.5; CI 1.1–1.9) (41). We also showed that, in contrast to the immunocompetent population, betaPV DNA is not significantly associated with age in OTR (41). In the normal population, betaPV seropositivity increases with age up until approximately the fourth and fifth decades (45). Since many OTR do not receive their transplant until they are middle-aged it is not surprising that they show no detectable increase in the prevalence of antibodies with increasing age.

This ubiquitous betaPV DNA carriage does not necessarily represent ‘active’ infection. A positive serological response may be a better measure of biologically relevant current or past infection. However, serological cross-reactivity between different HPV types may lead to misclassification of type-specific infections. There is also the theoretical argument for ‘reverse causality’ whereby a positive serological response could be the consequence of a cutaneous malignancy, or of a cancer-related cutaneous insult such as sunburn (44). A recent study in 807 healthy people from the Netherlands, Italy and Australia found overall HPV seroprevalence was similar across the three countries (50–57% for betaPV types, 40–48% for gammaPV types) despite significant differences in sunlight intensity (56), possibly arguing against reverse causality due to sunburn. We are currently undertaking the prospective serial testing of HPV antibody status from the time of transplantation on an OTR cohort recruited in parallel to this case–control study. When the longitudinal study matures in 2015, we will be in a position to address whether antibody response may develop as a result of cutaneous malignancy, though in principle this seems unlikely in OTR who are often profoundly immunosuppressed. Except for those who receive an organ transplant when young, we expect the majority of HPV antibodies to have been acquired prior to transplantation and, in support of this, an interim analysis of our prospective data finds no change in antibody status in the first 2 years following organ transplantation.

These difficulties with either HPV DNA or antibodies examined individually as markers of active infection, led us to combine the two variables and to classify people according to whether or not they had both DNA and antibodies present for specific HPV types. There were only 15 betaPV types for which we had both measures, but these included candidates implicated in malignancies in EV (HPV-5, -8, -17, -20), as well as according to previous epidemiology studies (HPV-5, -8, -36, -38) and functional studies (HPV-5, -8, -38). They also included the most prevalent betaPV types from DNA studies (HPV-23, -24, -36, -38) and the most seroprevalent betaPV types (HPV-8, -15, -17, -38, -49). Using this analytic approach, we found that the concordant presence of antibodies and DNA for at least one betaPV type was significantly associated with SCC. Antibody and DNA concordance conferred increased risk for HPV36 and borderline significant increased risk for HPV-9 and -24. It would seem from this that for some betaPV types concordance does result in an increased risk, but there are other types (e.g. HPV-5 and -38) where serology alone is enough to increase risk and adding concordant DNA does not further increase the risk. Given the relatively low levels of individual virus types, and the exploratory nature of these analyses, it will be important to confirm these associations in further studies.

Assessing the actual level of risk conferred by betaPV in skin cancer is complicated by the multiplicity of different betaPV types carried in any one person, especially when that person is immunosuppressed. It is possible that many betaPVs are ‘passengers’ rather than ‘drivers’ of the carcinogenic process within skin. Despite this complexity, we have shown that specific betaPVs such as HPV types 5, 9, 24 and 36 may confer additional risk. It is important for future studies to consider both DNA and antibody findings and to be adequately powered to assess whether specific betaPVs do indeed carry a clinically meaningful additional risk. At this time, we cannot use betaPV infection status as a predictive marker. However, if specific betaPVs are confirmed to play a synergistic role, then pretransplant screening and HPV vaccination at the time of transplantation might become a reality to help reduce the burden of posttransplant cSCC in OTR.

In this study we failed to fully match controls to cases for age and duration of transplantation. This reflects the difficulty in finding older OTR who have been transplanted for long periods without developing SCC. In a longitudinal cohort study undertaken at one of the study centers (London, UK) we showed that 74% of Caucasian OTR will develop one or more SCC by 30 years posttransplant, and that long-term transplant recipients without SCC typically have brown or black skin and would therefore have been excluded from this study (Harwood, 2010, unpublished data). This failure to match for age can be considered an issue of misclassification. Controls may have become cases had they been older or followed for longer; we may therefore have contaminated the control group with cases. This contamination would most likely have biased the RO toward the null.

In conclusion, in a hospital-based, case–control study in OTR, we have studied the association of betaPV DNA and betaPV serology with development of cutaneous SCC in 210 cases and 394 controls. We have shown a significant association between betaPV infection and SCC risk using concordant antibodies and DNA as a marker of infection. Using this measure, we found an association between HPV36 infection and cutaneous SCC in OTR and a trend toward positive associations for HPV5, -9, and -24. These positive associations may have clinical significance, but are not strong and first must be confirmed in prospective cohort studies that are adequately powered to examine individual betaPV types.

This study confirms a multiplicity of betaPV DNA and betaPV serological responses in OTR. We know from other epidemiological studies using the same methodologies that a given betaPV profile tends to persist within individual persons (37,53,57). Thus, it would be possible to detect a ‘high-risk profile’ from pretransplant screening. Our data suggest that detection of HPV antibodies from a blood test in addition to HPV DNA typing on 10 plucked eyebrow hairs would be appropriate for pretransplant screening. HPV vaccines currently in clinical use are preventative rather than therapeutic and would need to be given in childhood to successfully inhibit HPV infection. This is clearly not practical for most transplant situations. Therefore, to deliver an effective antiviral strategy pretransplantation, or very early posttransplant, one would need therapeutic vaccines or an effective antipapillomavirus treatment. HPV-specific antiviral agents are not currently available, but therapeutic anti-HPV vaccines are in development and would be especially appropriate for those persons for whom we can predict a high risk of multiple SCC, due to fair skin, sun-damage, older age and presence of precancerous skin lesions. However, development and deployment of vaccine strategies will only be realistic if a hierarchy of risk for individual betaPV in transplant-associated SCC has been established and this should be the focus of future studies.

Acknowledgement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure
  9. References

We thank the persons who participated in this study. These studies were funded by an EC grant (QLK2-CT-2002-01179). C.M.P and C.A.H are supported by Cancer Research UK. R.E.N. is funded by a NHMRC (Australia) Career Development Award. M.C.W.F was supported by a Clinical Fellowship from the Netherlands Organization for Health Research and Development (grant 907–00-150). T.W. was supported by the Peter und Traudl Engelhorn-Stiftung zur Förderung der Biotechnologie und Gentechnik. We thank J. Lindeman and Labo Bio-Medical Products B.V. (Rijswijk, The Netherlands) for providing the RHA strips.

Disclosure

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure
  9. References

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. Disclosure
  9. References