Biomarkers of oxidant load and type-specific clearance of prevalent oncogenic human papillomavirus infection: Markers of immune response?

Abstract

Human papillomavirus (HPV) infection is the cause of cervical cancer. Increased production of reactive oxygen species (ROS) maybe the common mechanism through which HPV-cofactors (i.e. smoking and inflammation) influence duration of infections. Biomarkers of total oxidant load may serve as cumulative measures of ROS exposure due to these cofactors. Therefore, we conducted a study evaluating the association between biomarkers of oxidant load and duration of HPV infections, early HPV natural history events. Serum samples were obtained from 444 HPV-positive women in the Ludwig-McGill Cohort Study. Anti-5-hydroxymethyl-2’-deoxyuridine autoantibody (anti-HMdU aAb) and malondialdehyde (MDA) were measured at baseline. Cox-proportional hazard models were used to estimate the probability of clearing any HPV, oncogenic HPV, non-oncogenic HPV, and HPV-16 infections. Women with elevated MDA were significantly more likely to clear prevalent oncogenic HPV infections compared to those with lower MDA levels (Adjusted Hazard Ratio (AHR)=2.7; 95%CI=1.4–5.1). There did not appear to be an association between elevated MDA and clearance of incident oncogenic HPV infections. Similarly, women with elevated anti-HMdU aAb levels had higher rates of prevalent oncogenic HPV infection clearance (Quartile 3:AHR=2.2; 95%CI=1.2–4.4; Quartile 4:AHR=2.4; 95%CI=1.2–4.9). Higher levels of oxidant load biomarkers were associated with increased clearance of prevalent HPV infections. However, oxidant load biomarkers measured prior to incident infections were not associated, suggesting that the elevation of MDA and anti-HMdU aAb may reflect an ongoing effective immune response, such as increased innate immunity. More research focused on the immune responses to HPV and elevated markers of oxidant load is needed.

Introduction

Human papillomavirus (HPV) infection is the cause of cervical cancer. The prolonged presence of an HPV infection, (e.g. HPV persistence), has been demonstrated to be a key event in the progression from infection with oncogenic HPV types to precancerous lesions or cancer in women.^{1, 2} The determination of factors associated with the duration of HPV viral presence may shed light on neoplastic events in the progression to cancer and help identify women at higher risk for dysplasia and cervical cancer. Cofactors, such as antioxidant micronutrients (e.g., carotenoids and tocopherols), inflammation, and smoking have been associated with the progression of HPV infection to cervical dysplasia.^3–6 It has been proposed that increased production of reactive oxygen species (ROS) may be the common mechanism through which these cofactors influence the natural history of HPV infections.^{4, 7} A few prospective studies have demonstrated an association between the contributors of ROS (e.g., low antioxidant nutrient status, smoking, and inflammation) and longer duration of HPV infections (either increased HPV persistence or reduced HPV clearance).^3–5 Specifically, among the investigations of antioxidant nutrients,^{5, 8–13} lower levels of individual nutrients (e.g. lycopene^{11, 13, 14}, vitamin B₁₂,¹⁵ tocopherols,⁹ β-cryptoxanthin,^{10, 13} lutein/zeaxanthin,^{10, 13} α- and β-carotene¹³) have been associated with an increased risk of HPV persistence (or decreased clearance). However, only a few nutrients have been consistently statistically significant across studies (albeit results from these studies differed somewhat by exposure measurement and outcomes).⁵ Smoking has been associated with decreased risk of clearing oncogenic HPV infections in some,^16–18 but not all studies.^{19, 20} Cervical inflammation (e.g., cervicitis) has been associated with prolonged HPV infection²¹ and dysplasia.^22–24 It is commonly caused by sexually transmitted infections, such as Chlamydia trachomatis and Herpes simplex virus-2, or conditions such as bacterial vaginosis.²⁵ However, increased HPV clearance with C. trachomatis co-infection has also been reported.²⁶ It is possible that inconsistencies are due to the lack of simultaneous assessment of the different contributors of ROS. Therefore, in this study, we utilized biomarkers of cumulative ROS as alternative measures of ROS exposure.

While ROS cannot be directly measured, oxidative damage resulting from ROS exposure can be quantified and serves as a biomarker of such exposure. There are several such biomarkers including amino acid oxidation products, chemical modifications of protein following carbohydrate or lipid oxidation,²⁷ or oxidized DNA base derivative.^{28, 29} Elevated plasma malondialdehyde (MDA), a by-product of lipid oxidation that forms adducts with protein and DNA, has been reported in the progression of many diseases and is widely used as an index of overall peroxidation.²⁷ Another biomarker of ROS exposure is an autoantibody (aAb) that recognizes oxidized DNA base derivatives [i.e., 5-hydroxymethyl-2'-deoxyuridine (HMdU)] when present in a polymeric structure with bovine serum albumin (BSA).^28–34 The purpose of this study was to assess whether two biomarkers of oxidant load (MDA and anti-HMdU aAb), indirect measures of ROS exposure, are associated with type-specific HPV clearance and duration of HPV infections, among a cohort of women.

MATERIALS AND METHODS

Study Sample

Women included in the current analysis were a sub-cohort of participants from the Ludwig-McGill cohort study. The Ludwig-McGill cohort study was an HPV natural history study of low-income women attending a comprehensive maternal and child health maintenance program in São Paulo, Brazil.³⁵ Study design, clinical sampling, and HPV testing for the Ludwig-McGill cohort study have been previously described.³⁵ In brief, a total of 2528 women were recruited between 1993 and 1997 with pre-scheduled visits over a 5 year period. Participants were seen every 4 months in the first year and twice yearly thereafter. At each visit, study nurses conducted in-person interviews specific for the current visit and collected cervical specimens for Pap cytology and HPV testing. All participants signed an informed consent form before entering the study. The study protocol and informed consent form were approved by the institutional review boards and the ethical committees of participating institutions. Women who were enrolled during the first 2 years of the Ludwig-McGill study, not diagnosed with squamous intraepithelial lesions (SIL) at baseline, and tested HPV positive at least once within the first 3 years of follow-up were included in the current analysis (N=444).

Serum Sample Processing and Storage

All non-fasting blood samples were collected by venipuncture by a trained nurse at each clinical visit. The samples were centrifuged as soon as possible but no later than 6–8 hours of collection. Aliquots (1 ml) of serum were stored in 1.8 ml Nunc cryovials at −20°C in a non-frost free freezer until shipped for analyses.

Oxidant load markers

MDA and anti-HMdU aAb were measured in serum samples collected at the baseline visit. For MDA, a modified version of a published High Pressure Liquid Chromatography method was used to measure the thiobarbituric acid (TBA) adduct of MDA in 100 µl of serum.^{36, 37} Anti-HMdU aAbs were analyzed in sera in duplicate using ELISA, as previously reported by Frenkel and colleagues.²⁹ Mean aAb levels were calculated as A492/µl undiluted serum ± standard error after subtracting the non-specific binding to M-BSA. This assay had approximately 3% within day variability and <10% between day.^{29, 30} Baseline serum was exhausted for 29 women; therefore, no MDA and anti-HMdU aAb measures were available. Furthermore, 2 women had sufficient serum only for MDA analysis; thus, they are missing in the anti-HMdU aAb analysis.

HPV DNA Detection Method

All HPV analyses were performed as previously described³⁵ at the Ludwig Institute for Cancer Research, São Paulo, Brazil. Cervical cells were treated with 100 µg/mL proteinase K for 3 hours at 55°C, followed by organic extraction and ethanol precipitation of DNA. HPV DNA detection was conducted via polymerase chain reaction (PCR) amplification of a 450-bp segment in the L1 viral gene using consensus primers MY09/11.³⁸ Each assay included positive and negative controls and DNA quality was assessed using amplification of a 268-bp fragment of the β-globin gene. HPV typing was performed by dot-blot hybridization with individual oligonucleotide probes specific for 27 HPV genital types³⁹ (types 6/11, 16, 18, 26, 31, 33, 35, 39, 40, 42, 45, 51–59, 66, 68, 73, 82, 83, and 84). PCR amplification products were further tested by restriction fragment–length polymorphism (RFLP) analysis of the L1 fragment⁴⁰ to resolve ambiguous results from the dot-blot hybridization and to distinguish additional HPVs (32, 34, 44, 61, 62, 67, 69–72, and 89).

Statistical analysis

Time to clearance was defined as the time between the first positive test of the type-specific HPV infection and the subsequent first negative test. Clearance time was censored at the woman’s last visit if she did not clear the infection within 3-years of follow-up or the first of two consecutive visits with missing HPV results. Time to clearance was evaluated for all infections, either acquired during follow-up or present at baseline, and separately for incident and prevalent infections. All clearance events were determined on a type-specific basis and then grouped as any HPV, oncogenic infections (HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68) or non-oncogenic infections (6/11, 26, 32, 34, 40, 42, 44, 53, 54, 55, 57, 61, 62, 64, 66, 67, 69, 70, 71, 72, 73, 81, 82, 83, 84, 89).⁴¹ Analyses were conducted based on individual HPV infections; therefore, women with multiple infections contributed multiple outcome events. The majority of women contributed a single infection (n=245; 57%) and multiple infections were as follows: 2 infections (26%), 3 infections (11%), and >3 infections (6%). A total of 790 type-specific HPV infections were detected within 3- years of follow-up. Seventy-two infections from 16 women occurred on the last study visit and were excluded from the analysis. The association of HPV clearance with two oxidant load biomarkers (MDA and anti-HMdU aAb) was separately evaluated. The final sample size for the MDA level analysis was 399 women contributing 665 infections (333 oncogenic, 332 non-oncogenic and 72 HPV-16 infections), while 397 women contributing 662 infections (332 oncogenic, 330 non-oncogenic and 72 HPV-16 infections) were included in the anti-HMdU aAb analysis.

Oxidant load biomarkers were evaluated as continuous measures (log transformed for the skewed data) and categorized into quartiles. Differences in MDA and anti-HMdU level by baseline demographic and lifestyle characteristics were tested by the Wilcoxon rank-sum test or Kruskal-Wallis test. Median clearance time by oxidant load level was estimated using the Kaplan-Meier method and differences were tested using the log-rank test. We examined associations of baseline oxidant load status with type-specific HPV clearance during the first 3 years of follow-up using Cox proportional hazard models. A robust sandwich estimator was used for the covariance matrix to account for within-subject correlations.⁴² The variables in the final model were selected based on models run individually (using backward selection) for any HPV, oncogenic HPV, non-oncogenic HPV and HPV 16. Variables that were significant at 0.10 level in the individual models were then adjusted for in the final models. The factors retained in the final models were age, condom use, education, monthly income, menarche, lifetime number of sexual partners, oral contraceptive use, race, and smoking status. Adjusted hazard ratio (AHR) and 95% confidence interval were estimated for each biomarker. The proportional hazard assumption for each Cox Model was verified graphically with significance based on the Kolmogorov-type Supremum test using the PHREG procedure.⁴³ All statistical tests were two-sided and considered as statistically significant at the level of 0.05. All analyses were performed with SAS (SAS 9.2., SAS Institute).

RESULTS

Table 1 presents the demographic characteristics of women who tested positive for any HPV type at least once during the three years of follow up. The majority of subjects (73.4%) were between the ages of 21 and 40. There was no statistically significant difference in the mean oxidant load measures (MDA and HMdU aAb) by any participant characteristic examined except race/ethnicity and education. MDA levels were higher among white women and decreased with increasing education.

Table 1.

The median concentration of baseline oxidant load biomarkers by participant characteristics among HPV positive women (N=428)

		Baseline MDA (µmol/L)	Baseline HMdU aAb (A492/µl-serum)
	n (%)	Median (min-max)	Median (min-max)
Age (years)
≥ 20	45 (10.5)	8.0 (3.8–18.9)	13.2 (0.5–78.9)
21 – 30	171 (40.0)	7.6 (0.2–23.2)	12.2 (1.7–65.4)
31 – 40	143 (33.4)	8.4 (2.1–22.7)	11.4 (1.3–48.9)
> 40	69 (16.1)	7.1 (2.5–28.2)	11.3 (1.2–51.5)
Ethnicity
White	256 (60.0)	8.4 (0.2–28.2)*	11.1 (1.2–78.9)*
Non-White	171 (40.0)	7.4 (1.0–21.6)	12.7 (0.5–52.9)
Marital Status
Common Law	145 (33.9)	7.5 (0.9–28.2)	11.2 (0.5–65.4)
Divorced	40 (9.4)	7.7 (2.1–20.1)	11.8 (1.7–48.9)
Married	156 (36.5)	8.8 (1.0–27.5)	12.7 (1.3–60.3)
Single	80 (18.7)	8.0 (2.5–18.9)	10.7 (1.7–78.9)
Widow	7 (1.6)	5.8 (5.2–22.7)	6.1 (3.8–28.5)
Education
< Elementary	86 (20.1)	9.1 (2.5 – 28.2)*	10.3 (1.2 – 78.9)
Elementary	243 (56.9)	8.0 (0.2 – 23.2)	11.5 (0.5 – 65.4)
< High School	50 (11.7)	7.9 (2.5 – 16.2)	12.8 (2.2 – 51.1)
≥ High School	48 (11.2)	6.5 (3.0 – 27.5)	13.4 (1.7 – 67.4)
Monthly Income ($US)
< 250	104 (25.2)	7.9 (2.5–21.6)	11.0 (0.5–78.9)
250 – 450	109 (26.4)	7.7 (1.0–23.2)	12.5 (1.3–51.5)
451 – 725	91 (22.0)	8.0 (0.2–28.2)	12.3 (1.2–65.4)
> 725	109 (26.4)	8.3 (3.5–27.5)	11.4 (2.1–67.6)
Smoking status
Never	209 (49.0)	8.0 (1.0–27.5)	12.5 (1.2–78.9)
Former	71 (16.6)	7.7 (2.5–21.5)	10.8 (1.7–48.8)
Current	147 (34.4)	8.0 (0.2–28.2)	11.1 (0.5–60.3)
Alcohol use
Yes	311 (72.8)	7.9 (0.2–28.2)	11.9 (0.5–78.9)
No	116 (27.2)	8.8 (1.0–21.5)	11.5 (1.2–60.3)
Oral contraceptive use
Never	73 (17.1)	7.8 (2.5–27.5)	10.2 (0.5–78.9)
< 6 years	245 (57.4)	7.6 (1.5–28.2)	12.3 (1.2–67.6)
≥ 6 years	109 (25.5)	8.6 (0.2–23.2)	11.3 (1.3–60.3)
Total no. of pregnancies
0 – 1	82 (19.4)	8.0 (1.5–19.1)	11.9 (1.3–67.4)
2 – 3	178 (42.1)	7.9 (1.0–27.5)	11.2 (0.5–65.4)
4 – 6	124 (29.3)	8.4 (3.0–21.5)	11.4 (2.0–60.3)
7+	39 (9.2)	6.6 (0.2–28.2)	10.0 (1.2–35.6)
Age at first intercourse
≤ 15	132 (30.9)	8.4 (0.2–23.2)	11.2 (0.5–47.6)
16 – 17	117 (27.4)	7.9 (1.7–19.8)	12.8 (2.3–67.4)
18 – 19	90 (21.1)	7.4 (1.5–22.7)	11.2 (1.3–78.9)
≥ 20	88 (20.6)	8.0 (2.5–28.2)	11.5 (1.8–44.1)
Lifetime no. of sexual partners
0 – 1	151 (35.4)	8.5 (1.7–27.5)	12.0 (1.3–78.9)
2 – 3	165 (38.6)	7.7 (1.0–28.2)	11.9 (2.2–60.3)
≥ 4	111 (26.0)	7.9 (0.2–18.9)	11.2 (0.5–48.9)
Total no. of sexual partners during the last year
0 – 1	381 (90.3)	8.0 (0.2–28.2)	11.9 (0.5–78.9)
≥ 2	41 (9.7)	7.4 (3.5–16.1)	10.1 (3.5–49.7)
Age at menarche
0–11 years	92 (21.5)	8.3 (1.7–22.7)	13.7 (1.2–51.1)
12–19 years	336 (78.5)	8.0 (0.2–28.2)	11.4 (0.5–78.9)
Condom use
Always	18 (4.2)	8.0 (4.1–18.9)	10.0 (2.9–42.9)
Never/Occasionally	410 (95.8)	8.0 (0.2–18.2)	11.9 (0.5–78.9)

^*Significant difference at the 0.05 level based on Kruskal-Wallis or Wilcoxon Rank-sum test.

min=minimum, max=maximum, MDA= Malondialdehyde, HMdU aAb= Human serum anti-5-hydroxymethyl-2’-deoxyuridine autoantibody

We examined the median clearance time by quartile of oxidant load markers and type of infection (all infections, prevalent infections, incident infections) for any HPV, oncogenic HPV and non-oncogenic HPV (Table 2). Time to clearance of prevalent oncogenic HPV infection was significantly shorter among women with MDA levels in the highest quartile compared to women in the lowest quartile (4.1 months vs. 8.7 months; p=0.03; Table 2). No significant differences in time to clearance of any HPV type or non-oncogenic HPV infection were observed across quartiles of MDA for incident infections and incident/prevalent infections combined. For prevalent oncogenic HPV infections, the time to clearance was significantly longer among women with lower anti HMdU aAb levels (<11.5 A₄₉₂/µl-serum: ~8 months vs. ≥11.5 A₄₉₂/µl-serum: ~4 months; P=0.04); however no difference was observed for incident oncogenic HPV infections (Table 2). The time to clearance did not differ significantly by HMdU level for any HPV, non-oncogenic HPV or HPV 16.

Table 2.

Median clearance time (in months) of type specific HPV infection by quartile of oxidant biomarkers

		Any HPV		Oncogenic HPV		Non-oncogenic HPV
		No. Infections1	Median Clearance in Months2 (95% CI)	No. Infections	Median Clearance in Months (95% CI)	No. Infections	Median Clearance in Months (95% CI)
MDA (µmol/L)
All infections
<5.9		177	8.0 (6.0, 8.4)	85	9.9 (7.9, 11.9)	92	5.9 (4.8, 6.5)
5.9–8.0		173	8.0 (6.1, 9.6)	90	8.0 (6.0, 11.8)	83	8.0 (6.0, 9.6)
8.1–10.6		155	6.0 (4.9, 8.0)	75	6.3 (5.7, 8.3)	80	5.8 (4.3, 8.7)
>10.6		160	6.0 (5.5, 8.0)	83	8.0 (5.9, 9.8)	77	5.3 (4.2, 6.0)
P-value			0.2		0.2		0.2
Prevalent infections
<5.9		52	7.9 (4.2, 8.2)	24	8.7 (4.6, 18.5)	28	4.7 (4.0, 8.0)
5.9–8.0		51	8.0 (4.5, 8.4)	29	6.9 (4.4, 12.0)	22	8.0 (4.1, 8.4)
8.1–10.6		44	8.0 (4.3, 8.8)	20	8.0 (4.0, 11.9)	24	6.4 (4.2, 11.0)
>10.6		55	4.1 (4.0, 7.9)	27	4.1 (4.0, 8.0)	28	4.1 (4.0, 8.0)
P-value			0.1		0.03		0.4
Incident infections
<5.9		125	8.0 (6.0, 11.0)	61	11.5 (6.2, 12.0)	64	6.0 (5.0, 8.4)
5.9–8.0		122	8.0 (6.2, 11.5)	61	9.3 (6.0, 12.0)	61	8.0 (6.0, 11.4)
8.1–10.6		111	5.9 (4.9, 7.9)	55	6.1 (5.7, 10.8)	56	5.8 (4.1, 10.7)
>10.6		105	6.7 (5.9, 11.1)	56	10.7 (6.2, 12.1)	49	5.7 (4.4, 6.8)
P-value			0.4		0.3		0.5
HMdU aAb (A492/µl-serum)
All infections
<6.9		164	6.0 (5.9, 8.2)	72	8.5 (5.9, 12.0)	92	5.8 (4.4, 7.9)
6.9–11.5		169	8.0 (6.1, 11.0)	92	8.8 (7.0, 11.9)	77	6.1 (4.5, 9.0)
11.6–18.9		170	6.1 (5.9, 8.0)	91	6.2 (5.8, 9.3)	79	6.0 (5.2, 8.0)
>18.9		159	7.9 (5.9, 9.3)	77	7.9 (5.7, 11.9)	82	7.0 (5.5, 9.6)
P-value			0.2		0.2		0.8
Prevalent infections
<6.9		48	8.1 (4.4, 8.4)	21	8.3 (4.4, 12.0)	27	8.0 (4.1, 8.3)
6.9–11.5		54	8.0 (4.5, 8.4)	28	8.6 (8.0, 12.6)	26	4.5 (4.0, 8.0)
11.6–18.9		53	4.5 (4.1, 8.0)	27	4.5 (4.1, 8.0)	26	4.8 (4.0, 8.3)
>18.9		46	4.8 (4.1, 8.0)	24	4.2 (4.0, 7.9)	22	6.6 (4.2, 9.6)
P-value			0.2		0.04		0.6
Incident infections
<6.9		116	6.0 (5.9, 10.3)	51	10.8 (5.9, 12.0)	65	5.8 (4.4, 6.0)
6.9–11.5		115	8.0 (6.0, 11.9)	64	9.0 (6.0, 11.9)	51	8.0 (5.8, 12.0)
11.6–18.9		117	6.3 (6.0, 9.3)	64	7.1 (6.0, 11.5)	53	6.0 (5.3, 8.4)
>18.9		113	8.5 (6.1, 11.8)	53	11.7 (6.7, 12.9)	60	7.0 (5.5, 11.5)
P-value			0.4		0.3		0.7

Note. p values are bolded if p<0.05

^1.Number of type-specific HPV infections detected were 665 any-type infections; 333 oncogenic, and 332 non-oncogenic.

^2.Median clearance time in months is estimated based on the number of type-specific infections, not individual women.

MDA= Malondialdehyde, HMdU aAb= Human serum anti-5-hydroxymethyl-2'-deoxyuridine autoantibody, CI=Confidence Interval

The likelihood of HPV clearance by oxidant load status (continuous measure or categorized into quartiles) and type of infection (all infections, prevalent infections, and incident infections) was evaluated (Table 3). Continuous measures of MDA were significantly associated with oncogenic HPV clearance (AHR=1.3, 95% CI=1.0–1.8) when prevalent and incident infections were combined. A significant association was also observed for prevalent oncogenic HPV infections (Continuous MDA: AHR=2.7, 95% CI=1.4–5.1), but not for incident oncogenic HPV infections. A similar pattern was observed when MDA was categorized into quartiles (All oncogenic infections: p-trend<0.001; prevalent oncogenic infections: p-trend<0.001). Women with MDA levels in the highest quartile were 2.7 times more likely to clear a prevalent oncogenic infection compared to women with MDA in the lowest quartile. There was a significant linear trend in the association across increasing quartiles of MDA and clearance of any prevalent HPV infection (p-trend=0.03). There did not appear to be an association between the clearance of incident HPV infections and MDA level, nor was there an association between MDA level and non-oncogenic HPV clearance.

Table 3.

The association between baseline oxidant load biomarkers and type-specific HPV clearance

	Any HPV1			Oncogenic HPV			Non-Oncogenic HPV
	All infections	Prevalent infections	Incident infections	All infections	Prevalent infections	Incident infections	All infections	Prevalent infections	Incident infections
	AHR2(95% CI)			AHR (95% CI)			AHR (95% CI)
MDA (µmol/L)
N	645	195	450	323	96	227	322	99	223
Continuous3 MDA	1.1 (0.9, 1.3)	1.3 (0.9, 2.0)	1.0 (0.9, 1.3)	1.3 (1.0, 1.8)	2.7 (1.4, 5.1)	1.1 (0.8, 1.5)	1.1 (0.8, 1.4)	0.9 (0.6, 1.6)	1.1 (0.8, 1.4)
Quartiles MDA
<5.9	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref
5.9–8.0	0.9 (0.7, 1.2)	0.8 (0.6, 1.3)	0.9 (0.7, 1.2)	1.2 (0.9, 1.7)	0.9 (0.4, 2.0)	1.1 (0.7, 1.6)	0.8 (0.6, 1.1)	0.6 (0.3, 1.1)	0.8 (0.6, 1.2)
8.1–10.6	1.2 (0.9, 1.5)	1.1 (0.7, 1.6)	1.2 (0.9, 1.6)	1.6 (1.1, 2.2)	1.3 (0.4, 3.7)	1.5 (1.0, 2.4)	0.9 (0.7, 1.3)	0.6 (0.3, 1.0)	1.0 (0.7, 1.5)
>10.6	1.1 (0.9, 1.4)	1.6 (1.0, 2.5)	1.0 (0.7, 1.3)	1.5 (1.1, 2.2)	2.7 (1.2, 6.0)	1.2 (0.8, 1.9)	1.0 (0.8, 1.4)	0.8 (0.5, 1.5)	1.0 (0.7, 1.5)
p-trend	0.1	0.03	0.5	<0.001	<0.001	0.2	0.7	0.6	0.8
HMdU aAb (A492/µl-serum)
N	642	194	448	322	96	226	320	98	222
Continuous3 HMdU aAb	1.0 (0.9, 1.2)	1.2 (1.0, 1.4)	1.0 (0.9, 1.1)	1.1 (0.9, 1.3)	1.6 (1.2, 2.1)	1.0 (0.8, 1.2)	1.0 (0.9, 1.1)	1.0 (0.7, 1.2)	1.0 (0.8, 1.2)
Quartiles HMdU aAb
<6.9	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref
6.9–11.5	0.9 (0.7, 1.2)	1.1 (0.7, 1.6)	0.9 (0.6, 1.2)	0.9 (0.6, 1.3)	1.1 (0.6, 2.0)	0.8 (0.5, 1.4)	1.1 (0.7, 1.5)	1.4 (0.7, 2.5)	0.8 (0.5, 1.3)
11.6–18.9	1.2 (0.9, 1.5)	1.6 (1.1, 2.4)	1.1 (0.8, 1.4)	1.3 (1.0, 1.8)	2.2 (1.2, 4.4)	1.1 (0.8, 1.7)	1.1 (0.8, 1.5)	1.5 (0.8, 2.6)	0.9 (0.6, 1.4)
>18.9	1.0 (0.8, 1.3)	1.3 (0.8, 2.0)	0.9 (0.7, 1.2)	1.1 (0.8, 1.6)	2.4 (1.2, 4.9)	0.8 (0.5, 1.2)	1.0 (0.7, 1.4)	0.8 (0.4, 1.6)	1.0 (0.7, 1.5)
p-trend	0.6	0.06	0.7	0.2	0.001	0.7	0.9	0.7	1

Note. AHR and 95% CI are bolded if the 95% CI excludes 1.00

¹Infections, not women, were used as the unit of analyses for each outcome.

²All Cox models were adjusted for age, condom use, education, monthly income, menarche, lifetime number of sexual partners, oral contraceptive use, race, and smoking status.

³Continuous measures of MDA and HMdU aAb were log transformed.

MDA= Malondialdehyde, CI=Confidence Interval, HMdU aAb= Human serum anti-5-hydroxymethyl-2’-deoxyuridine autoantibody

Continuous measures of anti-HMdU aAbs were significantly associated with clearance of prevalent HPV infections (Any HPV type: AHR=1.2, 95% CI=1.0–1.4 and oncogenic: AHR=1.6, 95% CI=1.2–2.1, respectively). Women with baseline anti-HMdU aAb levels in the third and fourth quartiles were significantly more likely to clear prevalent oncogenic HPV infections (AHR=2.2, 95% CI=1.2–4.4 and AHR=2.4, 95% CI=1.2–4.9, respectively, p-trend=0.001). No association between incident HPV clearance and anti-HMdU aAb level was detected.

The time to clearance of incident HPV-16 infection was significantly longer among women with lower MDA levels compared to those with higher levels (<8 µmol/L: 11.8 months vs. ≥8 µmol/L: 6.5 months; P=0.04) (Table 4). For every log-unit increase in MDA, women were 2.7 times more likely to clear an HPV-16 infection (95% CI: 1.3–6.0) when both incident and prevalent infections were considered. Similar results were observed when MDA was dichotomized at the median (AHR=2.1, 95% CI=1.2–3.7). However, when analyzed separately, there was no association between incident HPV-16 clearance and MDA. Due to the small number of prevalent infections, no estimates could be generated for risk of clearing prevalent infections. There was a small increased risk of clearing an HPV 16 infection with increasing HMdU aAb; however, this association was not maintained in the dichotomous analysis or for incident HPV-16 clearance.

Table 4.

HPV 16 median clearance time (in months) and risk of clearance by oxidant load biomarkers

	All infections			Prevalent infections			Incident infections
	No.1	Median Clearance (95% CI)2	AHR (95% CI)3	No.	Median Clearance (95% CI)	No.	Median Clearance (95% CI)	AHR (95% CI)
MDA (µmol/L)
<8.00	35	12.0 (6.9, 16.7)4	Ref	10	15. 9 (4.5, -)	25	11.8 (6.0, 30.5)5	Ref
≥ 8.00	37	6.7 (6.0, 12.0)	2.1 (1.2, 3.7)	5	11.9 (4.0, -)	32	6.5 (6.0, 12.0)	1.4 (0.7, 2.8)
Continuous	72	-	2.7 (1.3, 6.0)	15	-	57	-	1.3 (0.6, 2.8)
HMdU aAb (A492/µl-serum)
<11.48	33	11.9 (6.3, 15.2)	Ref	9	16.5 (8.1, 26.4)	24	6.8 (5.9, 12.1)	Ref
≥11.48	39	7.1 (6.1, 12.0)	1.4 (0.8, 2.7)	6	9.4 (4.0, -)	33	7.1 (6.1, 12.2)	1.2 (0.6, 2.4)
Continuous	72	-	1.4 (0.9, 2.2)	15	-	57	-	1.4 (0.9, 2.4)

^1.Number of type-specific HPV 16 infections was 72.

^2.Median clearance time in months is estimated based on the number of type-specific infections, not individual women.

^3.Cox models were adjusted for age, condom use, education, monthly income, menarche, lifetime number of sexual partners, oral contraceptive use, race, and smoking status. Adjusted Hazard Ratios were non-estimatible for prevanent HPV 16 infections.

^4,5Log-Rank Test of medidan clearance time, (4) p=0.08, (5) p<0.05.

MDA= Malondialdehyde, HMdU aAb= Human serum anti-5-hydroxymethyl-2’-deoxyuridine autoantibody, CI=Confidence Interval.

Discussion

Women with elevated oxidant load biomarkers concurrent with HPV infection at baseline (e.g. baseline biomarkers and prevalent HPV at baseline) were more likely to clear their existing infections compared to women with lower biomarker levels. This association was restricted to oncogenic type infections and possibly HPV-16 infections. However, oxidant load biomarker levels measured prior to an infection (e.g. baseline biomarker and incident HPV infection acquired during the follow-up) were not associated with HPV clearance. Overall, our findings suggest that MDA and anti-HMdU aAb are markers of an active infection and possibly immune response to infection. The higher the biomarker level, the more likely the infection will be to clear.

The active clearance of viral infections via innate immunity (Natural Killer Cells) and cell mediated immune (CMI) response (Cytotoxic T- cells) may result in oxidative stress. This increase in oxidative stress may be due to production of ROS byproducts from NADPH oxidase⁴⁴ and increased respiratory bursts. While HPV has been shown to successfully evade the immune system through down regulation of interferon responsive pathways,⁴⁵ in a majority of women, an HPV infection is cleared through natural immune responses. Recent data indicate that T-cell immunity^46–48 and natural killer cell activity,⁴⁹ in particular, are important for the loss of HPV infection and regression of cervical dysplasia. HPV-specific CMI responses have also been demonstrated as important factors in clearing HPV-16 infections.^{46–48, 50} Specifically, anti-HPV E6 CMI responses have been significant in the immediate control of HPV.^{46, 48} Therefore, HPV clearance due to active immune responses may result in increased production of ROS and subsequently ROS damage biomarkers, such as MDA and anti-HMdU aAbs. The lack of an association between incident HPV clearance is probably due the timing of biomarkers assessment and not a lack of active CMI to incident infections. Assessment of oxidant load markers at the first HPV positive visit is needed to determine the true association between oxidant load makers and clearance of incident HPV infections. It remains unclear why significant finding in this study were limited to oncogenic HPV. There is limited information related to the extent of the immune response to individual HPV types or differences by oncogenicity. More research focused on the immunogenicity of HPV, immune responses, and elevated markers of oxidant load is needed to identify possible type-specific immune modulators.

As with any observational study, there are strengths and limitations that need to be considered when interpreting our findings. We utilized a comprehensive approach in evaluating the putative associations of biomarkers of oxidant load with HPV clearance by examining biomarkers across the range of values (log-transformed continuous measures), testing for a threshold effect (across quartiles), and examining several HPV outcomes (any type infection, oncogenic, non-oncogenic and HPV-16 infections). Furthermore, we evaluated differences in the association between prevalent and incident HPV infections. This study was nested within the Ludwig-McGill Cohort study, which had a relatively large sample size, multiple HPV measurements, and a long follow-up period. Thus, we had relatively sufficient power to adequately test our hypothesis that measures of oxidant load status were associated with duration of HPV infection (e.g. HPV clearance). Stratified analysis of prevalent and incident HPV infections was limited by small sample size and resulted in some non-estimable HR and confidence intervals for prevalent HPV 16 infections. Furthermore, the true duration of prevalent HPV infections was unknown due to left censoring of the data. We utilized a stringent definition of type-specific HPV clearance and considered multiple clearance events by using viral infection as the unit of measure. The proportional hazards assumption in the Cox models were met for all analyses except for MDA and incident any-type and non-oncogenic HPV infection. All Cox regression analyses adjusted for the potential within-subject correlation due to multiple infections that were present simultaneously in the same women, although it was not possible to make the same adjustment when we estimated the median clearance time.

Similar to other biological markers, the values of MDA and anti-HMdU aAbs may not have reflected the absolute biomarker value due to losses that might have occurred during specimen processing, storage and/or extraction; however, this loss would be similar across all samples and not differ by HPV status. Therefore, the associations observed within this study should be valid with potentially a lower magnitude of the associations due to methodological errors. Finally, MDA and anti-HMdU aAbs are indirect measures of ROS and may not represent the total underlying ROS exposure. However, ROS cannot be measured directly; thus only indirect measures of ROS damage, such as MDA and anti-HMdU aAbs, can be utilized.^27–29

In conclusion, we observed that higher MDA was associated with increased oncogenic type HPV and HPV-16 clearance; an association that was restricted to prevalent infections. Consistent with MDA, higher anti-HMdU aAb levels were associated with increased clearance of prevalent oncogenic HPV infections. Our observations may reflect a link between oxidant load biomarker level and active immune response to HPV infection. These data add to a growing literature suggesting that ROS are involved in the HPV immune response.

Novelty.

This is the first study to evaluate biomarkers of oxidant load and early events in the natural history of HPV. Higher MDA and anti-HMdU aAbs were associated with increased oncogenic type HPV and HPV-16 clearance, HPV types known to cause cancer in women.

Impact.

Oxidant load biomarker level may be a marker of an active, on-going immune response to HPV infection. As such elevated MDA and anti-HMdU aAb may indicate women with HPV infections that are more likely to be cleared.

Abbreviations Used

ROS

reactive oxygen species

HPV

human Papillomavirus

Anti-HMdUaAb

Anti-5-hydroxymethyl-2’-deoxyuridine autoantibody

MDA

malondialdehyde

AHR

adjusted hazard ratio

aAb

autoantibody

HMdU

5-hydroxymethyl-2'-deoxyuridine

BSA

bovine serum albumin

TBA

thiobarbituric acid

ELISA

Enzyme-linked immunosorbent assay

PCR

polymerase chain reaction

RFLP

restriction fragment length polymorphism

CMI

cell mediated immune

NADPH

nicotinamide adenine dinucleotide phosphate