IgG and IgM Seroreactivity Against Natural HPV16 Infection and HLA‐DRB1 and ‐DQB1 Polymorphism

TL;DR

This study explores how genetic variations in HLA-DRB1 and HLA-DQB1 affect immune responses to HPV16 infection, focusing on IgG and IgM antibody levels.

Contribution

The study identifies specific HLA polymorphisms linked to IgG antibody levels against HPV16, but not IgM or infection persistence.

Findings

HLA-DRB1*03:02 is associated with higher likelihood of HPV16 DNA detection.

HLA-DRB1*13:01 and HLA-DQB1*04:02 are linked to lower IgG antibody levels against HPV16.

IgM antibody levels and infection type are not influenced by HLA-DRB1 or HLA-DQB1 variants.

Abstract

Little is known about the factors associated with the natural humoral immune response against human papillomavirus (HPV) infection. The association between HLA‐DRB1 and ‐DQB1 polymorphism and (1) HPV16 DNA detection, (2) naturally developed humoral immune response (IgG and IgM) against HPV16 L1 and L2 capsid proteins, and (3) type of HPV16 infection (transient or persistent) was investigated. Data from 943 women participating in the Ludwig‐McGill cohort study was analyzed. HLA‐DRB1 and HLA‐DQB1 genotyping was done by PCR‐based methods. HPV DNA was assessed every 4 months during the first year of follow‐up, and HPV16 IgG and IgM antibody measurements were performed by ELISA in serum samples from the first and/or second visits. Associations were estimated by adjusted logistic regression models. HLA‐DRB1*03:02 was associated with HPV16 DNA positivity (OR = 2.49, 95% CI: 1.11–5.60). HLA‐DRB1*13:01 and HLA‐DQB1*04:02 were associated with low levels of HPV16 IgG antibodies (OR = 0.44, 95% CI:0.23–0.87; OR = 0.10, 95% CI: 0.02–0.55, respectively). Neither IgM HPV16 antibodies nor the type of HPV16 infection (transient or persistent) were associated with HLA‐DRB1 or ‐DQB1 variants. HLA‐DRB1 and ‐DQB1 polymorphism might be associated with different levels of IgG against HPV16 infection, but IgM production and the type of infection seem to occur independently of these variants.

Tables

Table 1. Characteristics of the Ludwig‐McGill cohort participants at baseline.

Characteristics at baseline	All participants (n = 2462)	Subset tested for HLA genotyping (‐DRB1 and ‐DQB1) (n = 948)	Subset tested for HLA genotyping (‐DRB1 and ‐DQB1) and HPV16 seroreactivity
			IgG (n = 943)	IgM (n = 476)
Age, year
n, mean (SD)	2461, 32.7 (8.8)	948, 30.9 (7.7)	943, 30.9 (7.7)	476, 31.0 (8.5)
n, median (IQR)	2461, 32.0 (26.0–39.0)	948, 30.0 (25.0–36.0)	943, 30.0 (25.0–36.0)	476, 30.0 (24.0–36.0)
Race, n (%)
White	1585 (64.4)	620 (65.4)	618 (65.5)	313 (65.8)
Non‐white	874 (35.5)	328 (34.6)	325 (34.5)	163 (34.2)
Marital status, n (%)
Single	252 (10.2)	101 (10.7)	101 (10.7)	71 (14.9)
Married	1179 (47.9)	473 (49.9)	472 (50.1)	224 (47.1)
Widowed	57 (2.3)	19 (2.0)	18 (1.9)	10 (2.1)
Separated	140 (5.7)	52 (5.5)	51 (5.4)	33 (6.9)
Cohabiting	832 (33.8)	303 (32.0)	301 (31.9)	138 (29.0)
Income quartiles, n (%)
First quartile	703 (28.6)	267 (28.2)	264 (28.0)	125 (26.3)
Second quartile	530 (21.5)	168 (17.7)	167 (17.7)	57 (12.0)
Third quartile	613 (24.9)	171 (18.0)	171 (18.1)	69 (14.5)
Fourth quartile	615 (25.0)	342 (36.1)	341 (36.2)	225 (47.3)
Education, n (%)
< Elementary	554 (22.5)	195 (20.6)	195 (20.7)	84 (17.7)
Elementary	1438 (58.4)	562 (59.3)	558 (59.2)	290 (60.9)
High School	397 (16.1)	164 (17.3)	163 (17.3)	86 (18.1)
College/University	70 (2.9)	26 (2.7)	26 (2.8)	15 (3.2)
Ever smoked, n (%)
Never	1168 (47.4)	464 (49.0)	461 (48.9)	228 (47.9)
Current	864 (35.1)	328 (34.6)	327 (34.7)	176 (37.0)
Former	429 (17.4)	156 (16.5)	155 (16.4)	72 (15.1)
Ever drank alcohol occasionally, n (%)
Never	852 (34.6)	305 (32.2)	303 (32.1)	150 (31.5)
Ever drinker	1601 (65.0)	642 (67.7)	639 (67.8)	326 (68.5)
Lifetime n o of sexual partners, n (%)
0–1	1,089 (44.2)	414 (43.7)	412 (43.7)	190 (39.9)
2–3	856 (34.8)	344 (36.3)	342 (36.3)	174 (36.6)
4+	515 (20.9)	189 (19.9)	188 (19.9)	112 (23.5)
Age at first sexual intercourse, year
n, mean (SD)	2,461, 17.9 (4.0)	948, 17.6 (3.6)	943, 17.6 (3.6)	476, 17.5 (3.8)
n, median (IQR)	2,461, 17.0 (15.0–20.0)	948, 17.0 (15.0–19.0)	943, 17.0 (15.0–19.0)	476, 17.0 (15.0–19.0)
Parity
n, mean (SD)	2,338, 2.3 (2.1)	888, 2.2 (1.9)	883, 2.2 (1.9)	437, 2.3 (2.0)
n, median (IQR)	2,338, 2.0 (1.0–3.0)	888, 2.0 (1.0– 3.0)	883, 2.0 (1.0–3.0)	437, 2.0 (1.0–3.0)
Ever used oral contraceptive, n (%)
Never	401 (16.3)	136 (14.4)	133 (14.1)	69 (14.5)
0–5 years	1,349 (54.8)	556 (58.7)	555 (58.9)	278 (58.4)
≥ 6 years	711 (28.9)	256 (27.0)	255 (27.0)	129 (27.1)
Ever had STD, n (%)
Never	1,881 (76.4)	732 (77.2)	729 (77.3)	361 (75.8)
Condyloma or venereal warts	108 (4.4)	53 (5.6)	53 (5.6)	36 (7.6)
Othersa	463 (18.8)	160 (16.9)	158 (16.8)	78 (16.4)
HPV DNA status, n (%)
HPV negative	2,026 (82.3)	745 (78.6)	741 (78.6)	330 (69.3)
Positive for low‐risk genotypes	156 (6.3)	76 (8.0)	75 (8.0)	60 (12.6)
Positive for HPV16	67 (2.7)	36 (3.8)	36 (3.8)	30 (6.3)
Positive for HPV31 or 35	37 (1.5)	19 (2.0)	19 (2.0)	13 (2.7)
Positive for HPV52 or 67 or 33 or 58	46 (1.9)	26 (2.7)	26 (2.8)	18 (3.8)
Positive for other high‐risk genotypes	107 (4.4)	46 (4.9)	46 (4.9)	25 (5.3)
Type of HPV16 infection, n (%)
Single	45 (1.8)	24 (2.5)	24 (2.6)	21 (4.4)
Multiple	22 (0.9)	12 (1.3)	12 (1.3)	9 (1.9)
Number of HPV genotypes detected with HPV16, n (%)
2	17 (77.3)	8 (66.7)	8 (66.7)	7 (77.8)
3+	5 (22.7)	4 (33.3)	4 (33.3)	2 (22.2)
Number of HPV genotypes detected, n (%)
0	2,026 (82.3)	745 (78.6)	741 (78.6)	330 (69.3)
1	336 (13.7)	163 (17.2)	162 (17.2)	123 (25.8)
2	63 (2.6)	32 (3.4)	32 (3.4)	19 (4.0)
3+	14 (0.6)	8 (0.9)	8 (0.9)	4 (0.8)
Cytology, n (%)
Negative	2,361 (95.9)	883 (93.1)	880 (93.3)	434 (91.2)
ASCUS/AGUS	43 (1.8)	22 (2.3)	21 (2.2)	15 (3.2)
LSIL	31 (1.3)	21 (2.2)	21 (2.2)	9 (1.9)
HSIL	21 (0.9)	21 (2.2)	20 (2.2)	18 (3.8)
HPV16 seroreactivity, NAR
n, mean (SD)	NA	NA	943, 0.9 (0.6)	476, 1.2 (0.4)
n, median (IQR)	NA	NA	943, 0.8 (0.5–1.2)	476, 1.1 (0.9–1.3)

Table 2. Logistic regression analyses with HPV16 positivity, IgG, and IgM HPV16 seroreactivity as the outcomes.

Allele	HPV16 positivitya (n = 948)		Model 1b				Model 2c				Model 3d
			IgG (n = 452)		IgM (n = 242)		IgG (n = 300)		IgM (n = 161)		IgG (n = 49)		IgM (n = 34)
	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
HLA‐DRB1
DRB1*01:01	0.89	0.35–2.29	1.50	0.74–3.06	1.27	0.44–3.68	1.42	0.61–3.32	1.71	0.41–7.08	NE	NE	NE	NE
DRB1*01:02	0.92	0.36–2.36	0.90	0.47–1.73	0.82	0.31–2.15	0.76	0.34–1.69	0.99	0.29–3.40	NE	NE	2.20	0.17–28.15
DRB1*03:01	0.73	0.34–1.57	1.36	0.82–2.25	1.80	0.83–3.90	1.25	0.66–2.37	1.95	0.79–4.81	1.07	0.10‐11.33	1.46	0.09–24.98
DRB1*03:02	2.49	1.11–5.60	0.85	0.39–1.88	1.05	0.40–2.75	1.45	0.52–4.03	1.49	0.42–5.31	0.26	0.04‐1.55	NE	NE
DRB1*07:01	0.90	0.50–1.65	0.78	0.51–1.19	1.37	0.75–2.49	0.74	0.44–1.26	1.19	0.57–2.48	0.62	0.12‐3.27	0.62	0.13–3.06
DRB1*11:01	1.34	0.68–2.66	0.88	0.50–1.53	0.57	0.27–1.19	1.11	0.57–2.17	0.46	0.19–1.08	1.96	0.20‐19.13	0.66	0.11–3.91
DRB1*13:01	0.59	0.23–1.51	0.61	0.35–1.07	1.63	0.77–3.47	0.44	0.23–0.87	1.76	0.63–4.96	0.89	0.08‐10.11	0.46	0.04–6.01
DRB1*13:02	1.19	0.50–2.87	1.81	0.93–3.55	0.90	0.38–2.13	1.92	0.85–4.35	0.45	0.13–1.60	NE	NE	1.04	0.06–19.91
DRB1*15:01	1.18	0.54–2.57	1.23	0.67–2.26	0.87	0.41–1.86	1.62	0.73–3.60	0.80	0.34–1.91	0.28	0.03‐2.27	0.41	0.08–2.16
DRB1*15:03	1.10	0.45–2.69	1.55	0.80–3.01	0.89	0.39–2.04	1.04	0.51–2.12	1.60	0.56–4.55	NE	NE	0.87	0.11–7.10
HLA‐DQB1
DQB1*02:01	0.71	0.41–1.22	0.87	0.60–1.27	1.52	0.89–2.57	0.86	0.54–1.37	1.36	0.72–2.56	0.67	0.15–2.98	0.48	0.09–2.50
DQB1*03:01	1.27	0.74–2.19	0.81	0.54–1.21	0.66	0.38–1.14	1.03	0.61–1.72	0.76	0.39–1.47	1.30	0.25–6.69	1.88	0.44–7.93
DQB1*03:02	0.79	0.35–1.77	0.72	0.42–1.24	1.08	0.51–2.29	0.58	0.29–1.17	0.72	0.28–1.87	NE	NE	0.32	0.02–4.36
DQB1*04:02	1.71	0.92–3.22	0.97	0.57–1.65	1.06	0.55–2.07	0.96	0.51–1.82	1.05	0.46–2.40	0.10	0.02–0.55	1.45	0.25–8.25
DQB1*05:01	0.88	0.48–1.60	1.24	0.81–1.90	1.56	0.84–2.91	1.12	0.68–1.87	1.66	0.74–3.74	6.35	0.66–61.15	1.02	0.16–6.54
DQB1*06:02	0.81	0.43–1.53	1.32	0.85–2.06	0.88	0.50–1.55	1.22	0.72–2.08	1.16	0.59–2.27	0.74	0.12–4.71	0.36	0.08–1.63
DQB1*06:03	0.74	0.29–1.90	0.69	0.37–1.29	1.13	0.51–2.50	0.57	0.27–1.19	1.12	0.37–3.38	NE	NE	1.87	0.14–24.80

Table 3. Multinomial logistic regression analysis with transient and persistent HPV16 infection as the outcomes.

Allele	Transient infectiona (n = 919)		Persistent infectiona (n = 919)
	OR	95% CI	OR	95% CI
HLA‐DRB1
DRB1*01:01	0.86	0.20–3.71	0.81	0.19–3.50
DRB1*01:02	1.40	0.41–4.78	0.88	0.20–3.77
DRB1*03:01	0.20	0.03–1.49	1.14	0.43–3.07
DRB1*03:02	3.01	0.98–9.29	2.25	0.64–7.88
DRB1*07:01	1.07	0.44–2.58	0.80	0.32–2.02
DRB1*11:01	1.13	0.38–3.35	2.18	0.90–5.30
DRB1*13:01	0.90	0.27–3.07	0.26	0.04–1.92
DRB1*13:02	1.45	0.42–4.95	0.91	0.21–3.91
DRB1*15:01	1.59	0.53–4.79	1.40	0.47–4.17
DRB1*15:03	1.25	0.35–4.38	0.86	0.19–3.79
HLA‐DQB1
DQB1*02:01	0.52	0.22–1.25	0.96	0.44–2.10
DQB1*03:01	1.07	0.46–2.50	1.60	0.73–3.51
DQB1*03:02	0.55	0.13–2.35	1.45	0.54–3.91
DQB1*04:02	1.44	0.53–3.91	1.80	0.71–4.56
DQB1*05:01	1.02	0.42–2.47	0.64	0.24–1.72
DQB1*06:02	1.11	0.45–2.70	0.72	0.27–1.93
DQB1*06:03	0.73	0.17–3.16	0.67	0.16–2.88

Full text

Introduction

1

Human papillomavirus (HPV) ranks as the most common sexually transmitted virus worldwide. More than 90% and 80% of sexually active men and women, respectively, will be infected by HPV in their lifetime [1]. About 90% of HPV infections are transient and clear within 12–24 months [2], but infections from high oncogenic risk HPV (HR‐HPV) genotypes that persist can raise the risk of cervical precancerous lesions, which if left untreated can progress to cancer [3]. More than 200 HPV genotypes have been identified and characterized [4]. However, HPV16 is the main oncogenic genotype, responsible for over 50% of cervical cancers worldwide [3], and is also the most common type found in normal cervix [5].

During primary HPV infection, the cellular immune response is activated mainly via local antigen‐presenting cells (APCs) of the infected epithelium [6]. These APCs present antigens via HLA class I and II molecules to CD8 and CD4 T lymphocytes, respectively. The CD4 T lymphocytes will promote the activation of CD8 T cytotoxic lymphocytes and B cell‐derived humoral immune response. The interaction between naïve B cells and CD4 T lymphocytes leads to B cell activation and antibody production [7]. IgM is the first isotype produced during primary infections [8], followed by IgG, the isotype with the highest plasma concentration [7]. However, naturally developed HPV16 IgG antibodies seem to provide only moderate protection against subsequent HPV16 DNA detection in young women (RR (95% CI) = 0.65 (0.55–0.74)), and not in women aged over 25 years (RR (95% CI) = 0.88 (0.73 –1.04) [9]. IgG seroreactivity to HPV16 has even been considered a marker of past or latent HPV16 infections among older women [10].

HLA classical class II molecules are coded by HLA‐DRA1, HLA‐DRB1, HLA‐DQA1, HLA‐DQB1, HLA‐DPA1, and HLA‐DPB1 genes, located on the short arm of chromosome 6 (6p21.3) [11]. Due to its high genetic variability, essential role in antigen presentation, and consequent involvement, in mediating the immune response, numerous studies have investigated the association between HLA class II loci and cervical cancer, mainly exploring HLA‐DRB1 and HLA‐DQB1 genes. The HLA‐DRB104, HLA‐DRB107, HLA‐DRB111, HLA‐DRB115, HLA‐DQB103*, and HLA‐DQB106:02* were previously shown to be associated with susceptibility to squamous cervical cancer (SCC) [12]. In contrast, HLA‐DRB113:01* has been associated with lower risk of cervical cancer [13].

The possible influence of HLA polymorphism on the humoral immune response to HPV16 infection remains poorly understood. Our objective was to investigate whether HLA class II alleles influence HPV16 infection outcome. More specifically, we analyzed the associations between HLA‐DRB1 and ‐DQB1 alleles and (1) HPV16 DNA detection, (2) naturally developed HPV16 IgG and IgM antibodies, and (3) type of HPV16 infections (transient or persistent).

Materials and Methods

2

Study Design

2.1

The Ludwig‐McGill cohort study is a prospective investigation of the natural history of HPV infection [14]. In brief, the cohort included 2462 women recruited from 1993 to 1997—before the implementation of HPV vaccination programs ‐ and followed up for 10 years. The women were recruited from a maternal and child health program for low‐income families at a public hospital in Sao Paulo, Brazil. To be included in this study, women had to be permanent residents in Sao Paulo, aged between 18 and 60 years old; had an intact uterus without previous reported treatment of cervical abnormalities in the past 6 months; no current indication for hysterectomy; no current pregnancy and no intention to become pregnant during the first year of follow‐up; and have reported no use of vaginal medication up to 2 days before data collection. In the first year, blood and cervical cell samples were collected every 4 months for serology testing and cytology and HPV DNA testing, respectively. During the following years, cervical specimens were collected 6 months apart, and HPV16 serology was tested annually. Sociodemographic and behavioral information—such as sexual and contraceptive characteristics ‐ were obtained by questionnaires at every visit. This study was approved by the institutional ethical and research review boards of the participating institutions in Canada and Brazil.

HPV Genotyping

2.2

DNA from ectocervical and endocervical cells was extracted and purified using previously described standard techniques [14]. Afterwards, HPV DNA positivity was tested using a PCR protocol which amplifies a L1 viral gene conserved fragment of 450 bp flanked by MY09/11 and PGMY09/11 primers [15, 16]. HPV genotyping was obtained by hybridization with individual oligonucleotide probes for high or intermediate oncogenic risk genotypes (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, 73, and 82) and low oncogenic risk genotypes (6, 11, 26, 32, 34, 40, 42, 44, 53, 54, 57, 61, 62, 64, 67, 69, 70, 71, 72, 81, 83, 84, and 89 [17]). In case of ambiguity, the results were confirmed by restriction fragment length polymorphism analysis of the L1‐amplified fragment using a set of restriction enzymes [14].

HPV Serology

2.3

Blood samples were centrifuged to separate the serum, where anti‐HPV16 IgM and IgG antibodies were quantified by ELISA protocol based on genotype‐specific L1 + L2 virus‐like particles (VLPs). Recombinant HPV16 virus‐like particles (VLPs) expressing L1 and L2 (in a baculovirus‐expression system) were kindly provided by Dr. John Schiller, National Institute of Health, United States. Normalized absorbance ratio (NAR) ‐ a standardized procedure used to reduce measurement errors in ELISAs ‐ was used to express seroreactivity, in which the mean blank‐subtracted optical densities (ODs) was divided by the equivalent value of the control serum pool included in the same plate in triplicate, using 1:10 and 1:50 dilutions [18]. HPV16 IgG seroreactivity was measured from samples collected in the first two visits (at 0 and 4 months), and IgM was measured at the first visit (baseline).

HLA Genotyping

2.4

HLA class II genotyping was performed by previously reported PCR‐based methods [19, 20]. Briefly, the second exon of HLA‐DRB1 and ‐DQB1 loci were amplified and visualized by electrophoresis 1.5% agarose gel colored with ethidium bromide. Dot‐blotted PCR products were hybridized with sequence‐specific 3′ biotin‐labeled oligonucleotide probes to differentiate HLA alleles. The hybridized probes were then detected by chemiluminescence with the ECL kit (Amersham Pharmacia Biotech) after incubation with streptavidin‐peroxidase conjugate. The complete genotypic data for all individuals genotyped at the HLA‐DRB1 and ‐DQB1 loci are available in Supporting Information S1, irrespective of inclusion in the present study.

Statistical Analyses

2.5

The analyses were restricted to HPV16 seroreactivity due to its relevance in cervical lesions and carcinogenesis. They were based on a subcohort of women tested for both HLA‐DRB1 and ‐DQB1 (n = 948) and also for HPV16 IgG serology (n = 943), which was enriched to include women with HPV outcomes as previously described [10, 20]. From the group of women tested for anti‐HPV16 IgG antibodies, 476 were also tested for anti‐HPV16 IgM antibodies.

The associations between HLA‐DRB1 and HLA‐DQB1 alleles and (1) HPV16 positivity, (2) anti‐HPV16 IgG and IgM seroreactivity, and (3) type of HPV16 infection (transient or persistent) were evaluated using logistic regression models, which estimated odds ratios (OR) and their respective 95% confidence intervals (CI). Only HLA‐DRB1 alleles with a frequency ≥ 5% and HLA‐DQB1 with a frequency ≥ 10% were analyzed. The association between HLA‐DRB1 and ‐DQB1 and HPV16 DNA positivity was restricted to those women who completed the first year of the study. The estimations were adjusted for race (white vs. non‐white) and age (continuous).

The association between HLA alleles and HPV16 IgG and IgM seroreactivity was evaluated using three models. The comparisons were based on NAR values: above versus below median (model 1) and upper tertile versus lower tertile (model 2). These models were restricted to women with the highest probability of having an HR‐HPV infection. The same cutoff of model 2 was used in model 3, restricting the analysis to women who tested positive for HPV16 during the first year of follow‐up, and the ORs were adjusted for race (white vs. non‐white) and age (continuous). For IgG, the highest NAR value obtained between serum dilutions 1:10 and 1:50 tested on the first and second visits were considered. Only the highest NAR value obtained after testing both serum dilutions (1:10 and 1:50) on the first visit was considered for IgM.

To select women who had the highest probability of having an HR‐HPV infection, a propensity score based on the probability of detecting HPV DNA during the first year was estimated, considering the following variables measured at baseline: age (continuous), race (categorical: white vs. non‐white, as inferred by skin color), marital status (categorical: single, married, widowed, separated, cohabiting), age at the first intercourse (continuous), lifetime sexual partners (continuous), and condom use (categorical: always, sometimes, never). Women above the median of the propensity score were considered to have a high risk of HPV infection.

Finally, the association between HLA‐DRB1 and ‐DQB1 alleles and the type of HPV16 infection (transient or persistent) was evaluated using multinominal logistic regression. In this analysis, women with persistent infection (defined as at least two positive results for HPV16 during the first year of follow‐up) and women with transient infections (defined as one positive result for HPV16 during the first year) were compared to the referent group defined as women who were consistently negative for HPV16 during the first year of follow‐up. The analyses were performed using STATA statistical software version 17 [21].

Results

3

Sociodemographic and behavioral characteristics from the entire Ludwig‐McGill cohort (n = 2462) and subsets of women tested for both HLA‐DRB1 and ‐DQB1 and for IgG seroreactivity (n = 943) or IgM seroreactivity data (n = 476) are described in Table 1. Overall, the distribution of characteristics was similar among groups, except that high‐risk HPV genotypes and the presence of lesions (HSIL) were slightly more common in the subsets.

Table 2 shows the associations between HLA alleles and HPV16 positivity, IgG, and IgM seroreactivity. An association between HLA‐DRB103:02* and HPV16 positivity (OR = 2.49, 95% CI = 1.11–5.60) was observed. HLA‐DRB113:01* and HLA‐DQB104:02* were associated with low levels of HPV16 IgG seroreactivity (OR = 0.44, 95% CI = 0.23–0.87; OR = 0.10, 95% CI = 0.02–0.55, respectively). No significant associations were observed between IgM seroreactivity and HLA‐DRB1 and ‐DQB1 alleles.

Associations between HLA‐DRB1 and ‐DQB1 polymorphism, and type of HPV16 infection are presented in Table 3. No significant associations were observed for transient or persistent HPV16 infections and the HLA alleles.

Discussion

4

HLA class II molecules may play a role in the antibody immune response against HPV infection, considering that T‐follicular helper (Tfh) cells, a subset of CD4 T cells, help antigen‐primed B lymphocytes differentiate into plasma cells and secrete antibodies [22]. The interaction between the HLA class II molecule–peptide complex and the T‐cell receptor is necessary for CD4 T‐cell activation and to build an effective humoral response. Therefore, it is reasonable to hypothesize that differences in HLA class II molecules caused by genetic polymorphism may affect the efficacy in binding and presenting HPV16 antigens to lymphocytes, affecting the humoral immune response.

Although several studies have investigated the associations between HLA alleles and the susceptibility to HPV‐associated diseases [23, 24, 25, 26, 27, 28, 29], the mechanism behind this potential relationship is still unknown. It was previously showed that patients with certain haplotypes are less likely to clear HPV infection [12]. Some studies have also analyzed the relationship between HLA polymorphism and the serological immune response to HPV infections. A genome‐wide association study (GWAS) identified a common genetic variant (rs9357152–A/G), located on chromosome 6p21.32 in the major histocompatibility complex (MHC) class II region nearby gene HLA‐DQB1, associated with HPV8 L1 seropositivity in Europeans [30]. Moreover, it was further observed in this same cohort that HLA‐DRB111:01* and ‐DQB103:01* were associated with HPV8 seropositivity, and HLA‐DRB107:01* and ‐DQA102:01* were associated with HPV77 seropositivity [31]. It is important to highlight that even though all HPVs infect cells in the basal layer of stratified epithelia, HPV8 and HPV77 are cutaneous genotypes and are not involved in cervical infections [32]. Additionally, a recent study showed that antibody levels against HPV16 L1 and E6 viral proteins measured in HPV+ oropharyngeal cancer (HPV( + )OPC) patients were affected by HLA variants [33]. The locus rs4713462 was associated with antibody levels against HPV16 E6, and an amino acid variation in HLA‐DRβ1 (71‐Glu) was associated with antibody levels against HPV16 L1, the viral capsid protein [33]. Despite these investigations, none of them analyzed the association between HLA variants and HPV seroreactivity in the context of cervical lesions.

In our study, associations between HLA‐DRB113:01* and HLA‐DQB104:02* and low levels of HPV16 IgG seroreactivity were observed, indicating a decreased likelihood of antibody response against HPV16 with these alleles. Contradictorily, HLA‐DRB113* was associated with a reduced risk of cervical cancer in different populations pooled in a meta‐analysis [13] while no association was found with HPV infection (transient or persistent infections with any genotypes) in the Ludwig‐McGill cohort [20]. Our findings also showed that neither transient nor persistent HPV16 infections were affected by HLA‐DRB1 and ‐DQB1 polymorphisms. Similarly, another cohort study showed no evidence of association between HPV persistence and HLA alleles, although only a few alleles (HLA‐B07*, DQB103*, DQB106:02*, DRB113*, and DRB115:01*) or haplotypes were considered [34]. A previous analysis from the Ludwig‐McGill cohort examining the risk of HPV persistence for all genotypes combined (not limited to HPV16) showed that HLA‐DRB103:01‐DQB102:01 haplotype was associated with lower risk of transient and persistent HPV infections, HLA‐DRB111:02‐DQB103:01 with a lower risk of HPV persistence and HLA‐DRB116:01‐DQB105:02 and HLA‐DRB108:07‐DQB104:02 with a higher risk of HPV persistence [19].

It should be recognized that the associations found in studies investigating HLA may be more likely to be observed by chance given the large number of statistical tests that are typically performed. However, it is important to note that our analysis was performed based on HLA alleles, not haplotypes, only considering HPV16, and including a large sample size. Thus, the results of our study apply to HPV16 and may not generalize to all HPV genotypes. Furthermore, our study included women with a mean age of 33 years. Expression of HLA molecules in the cell surface, as well as the serology levels may be different in these women compared to younger women. Indeed, studies have shown that natural antibodies developed against HPV16 have a moderate protective effect against subsequent HPV detection in younger women (under 25‐30 years), while it has been associated with an increased risk of HPV outcomes in older women [9, 10]. It is therefore possible that the effect of HLA observed in our study generalizes only to older women.

This study has several strengths. IgG and IgM antibodies against HPV were measured from a large sample of unvaccinated women, which provides insight into the development of natural infections. Also, testing was done using the most common method to test HPV serology in the literature, with all assays performed using the same technology at the same laboratory by two trained students. However, even though the study had a meticulous data collection protocol, the participants’ history of past HPV infection was unknown. It is impossible to determine if the antibody levels detected at that time point were from current/recent or past HPV infections. Another limitation of the study is that HPV16 IgG and IgM antibodies were not measured in all follow‐up visits, making it impossible to account for antibody dynamics over time. Finally, although there was an adjustment for potential confounders such as race and age, residual confounding is possible.

In conclusion, these results suggest that HLA‐DRB1 and ‐DQB1 polymorphism might be associated with different levels of IgG against HPV16 infection, but IgM production and the type of infection (transient or persistent) seem to occur independently of these variants. Further evaluations are required to elucidate the genetic background behind the differences in naturally developed antibody levels against HPV infections.

Author Contributions

Letícia Boslooper Gonçalves performed the statistical analysis, analyzed the data, and drafted the manuscript. Andrea Trevisan helped with the statistical analysis and revised the manuscript. Helen Trottier supervised the study, helped analyze the data, and critically revised the manuscript. Luisa L. Villa and Eduardo L. Franco contributed to the conceptualization, methodology, project administration, funding acquisition, and manuscript revision. Patrícia Savio de Araujo‐Souza conceived and supervised the study, analyzed the data, and critically revised the manuscript.

Ethics Statement

Eligible participants signed informed consent form. The study protocol was approved by the ethical review boards of the participating institutions in Canada and Brazil.

Conflicts of Interest

ELF's institution has received grants from Merck in support of his investigator‐initiated studies. He has also served as an occasional advisor for companies involved with HPV vaccines (Merck, GSK) and HPV diagnostics (Roche, BD). LLV has received institutional grants from Merck to support her investigator‐initiated studies. She is also an occasional advisor and speaker of Merck, Sharp and Dohme for HPV vaccines and HPV diagnostics (Roche, BD. Qiagen, Cepheid). HT has received occasional lecture fees from Merck. The other authors reported no conflict of interest.

Supporting information

Supporting material 1.