Mucosal but Not Parenteral Immunization with Purified Human Papillomavirus Type 16 Virus-Like Particles Induces Neutralizing Titers of Antibodies throughout the Estrous Cycle of Mice

Abstract

We have recently shown that nasal immunization of anesthetized mice with human papillomavirus type 16 (HPV16) virus-like particles (VLPs) is highly effective at inducing both neutralizing immunoglobulin A (IgA) and IgG in genital secretions, while parenteral immunization induced only neutralizing IgG. Our data also demonstrated that both isotypes are similarly neutralizing according to an in vitro pseudotyped neutralization assay. However, it is known that various amounts of IgA and IgG are produced in genital secretions along the estrous cycle. Therefore, we have investigated how this variation influences the amount of HPV16 neutralizing antibodies induced after immunization with VLPs. We have compared parenteral and nasal protocols of vaccination with daily samplings of genital secretions of mice. Enzyme-linked immunosorbent assay analysis showed that total IgA and IgG inversely varied along the estrous cycle, with the largest amounts of IgA in proestrus-estrus and the largest amount of IgG in diestrus. This resulted in HPV16 neutralizing titers of IgG only being achieved during diestrus upon parenteral immunization. In contrast, nasal vaccination induced neutralizing titers of IgA plus IgG throughout the estrous cycle, as confirmed by in vitro pseudotyped neutralization assays. Our data suggest that mucosal immunization might be more efficient than parenteral immunization at inducing continuous protection of the female genital tract.

The “high-risk” human papillomavirus (HPV) types, most commonly type 16 (HPV16), are etiologically linked to over 90% of cervical cancers (4). Cervical cancer is the second leading cause of cancer deaths in women worldwide, encouraging the development of a prophylactic vaccine to prevent genital infection by these viruses. Vaccine development has been hindered by the difficulty of virus propagation in culture and the lack of animal models for the genital mucosatropic HPV types (24). However, expression of the papillomavirus major capsid protein L1 in mammalian, insect, and yeast cells or bacteria has been shown to generate virus-like particles (VLPs) (15, 16, 18, 20, 26, 34, 36). Parenteral injection of VLPs elicits high titers of serum neutralizing antibodies and protection from experimental challenge with infectious virus in animal papillomavirus models (7, 19, 34, 36, 41). Protection from experimental infection with cottontail rabbit papillomavirus and canine oral papillomavirus by passive transfer of immunoglobulin G (IgG) from immunized to naive animals has been demonstrated for rabbits (7) and dogs (41), respectively, indicating that cell-mediated effector immune responses are not required for protection. However, to intercept genital mucosal HPV, neutralizing antibodies must act at mucosal surfaces. Protection of various mucosae is primarily mediated by secretory IgA (sIgA), which is only induced upon antigen delivery to mucosa-associated lymphoid tissues (21). However, in addition to the locally produced sIgA, other Igs, mainly IgG, are also present in genital secretions, where they are generally thought to result from transudation through the vaginal epithelium (6, 25, 29, 44). Delivery of antigen through the nasal route of immunization has been shown to be the most effective method of inducing both sIgA and IgG in genital secretions of mice (2, 10, 12, 14, 17, 26, 28, 39), monkeys (35), and women (3). In contrast, immunization by parenteral routes (subcutaneous, intramuscular, or intraperitoneal) only leads to specific IgG and no sIgA in genital secretions both for HPV VLPs (2, 23) and for other antigens (5, 14, 27, 29, 43). Recently, we demonstrated that both nasal and parenteral vaccinations with purified HPV16 VLPs induce HPV16 neutralizing antibodies in genital secretions of mice (2). In these experiments, the neutralizing activities of specific IgA and IgG were similar, suggesting that IgG alone could be sufficient to protect the genital tract. However, the amounts of IgA and IgG in genital secretions are influenced by sex hormones and thus vary along the estrous cycle in both rodents (30, 46, 47) and women (5, 22, 37, 40, 42, 45). Here we have analyzed with mice how these variations influence the outcome of parenteral or intranasal vaccinations with purified HPV16 VLPs. Our results suggest that induction of both antibody classes may be necessary to achieve continuous protection of the female genital tract.

Ig content in vaginal washes of immunized mice varies along their estrous cycle.

Twelve anesthetized female BALB/c mice were immunized intranasally with three weekly doses of 5 μg of HPV16 VLP and 5 μg of cholera toxin as described previously (2). This mode of nasal vaccination allows inhalation of about one-third of the inoculum (2), but for simplicity, it will be referred as “nasal immunization” hereafter. In parallel, a total of 15 mice were immunized parenterally by the subcutaneous or the intraperitoneal route with multiple doses (three to four) of 1 or 5 μg of HPV16 VLPs. Blood was obtained from all mice 2 to 3 weeks after the last immunization, and then vaginal washes were taken daily for 5 consecutive days as described previously (17). Portions of the vaginal washes were used to determine the stage of the estrous cycle (1). Total and HPV16 VLP specific antibody contents in serum and vaginal washes were determined by enzyme-linked immunosorbent assay (ELISA) as described previously (17, 26). A total of 134 vaginal washes were analyzed: 38 were found in estrus, 33 in metestrus, 48 in diestrus, and 15 in proestrus. The mean total IgA and IgG contents at the different stages of the estrous cycle are shown in Fig. 1. The largest amount of IgG and the smallest amount of IgA occurred during diestrus, in agreement with data reported by Gallichan and Rosenthal (13). This inverse correlation in the amount of IgA and IgG was similarly observed in mice immunized parenterally or nasally (data not shown). Overall, the amount of IgA was higher than the amount of IgG, except during diestrus. This is particularly striking when the IgG/IgA ratios are calculated for each sample, thus avoiding the variations introduced by the sampling method (data not shown). During estrus, there is about 10 times more IgA than IgG in vaginal washes, and a positive IgG/IgA ratio is only found during diestrus (2 times more IgG than IgA). We therefore determined how these variations influenced the specific antibody response induced by vaccination.

FIG. 1.

Igs in vaginal washes of immunized mice along the estrous cycle. The mean total IgG and IgA in vaginal washes collected from mice at different stages of the estrous cycle are shown. E, estrus; M, metestrus; D, diestrus; P, proestrus.

Variations in anti-HPV16 VLP antibodies in vaginal washes of mice immunized intranasally or parenterally along the estrous cycle.

In contrast to mucosal immunization, parenteral immunization can only induce antibodies of the IgG isotype in vaginal washes (2, 5, 14, 23, 27, 29, 43). Here we have compared two groups of mice immunized parenterally or intranasally with purified HPV16 VLPs (see above). For technical reasons, we have used cholera toxin for nasal immunization to increase the titers of the systemic (5× to 10×) and mucosal (2× to 5×) antibody responses (2). Clearly such an adjuvant is toxic in humans, but it is likely that a similar effect is achievable with either higher doses of VLPs or detoxified mucosal adjuvants (8, 11). This allowed us to use low VLP nasal doses (5 μg) in mice to induce specific titers of antibodies in serum and secretions (IgG) similar to parenteral immunization and high enough to allow neutralization assays to be performed. The mean titers of anti-HPV16 VLP IgG in serum were indeed similar in the two groups of mice, with mean titers of 205,000 for the mice immunized parenterally and 225,000 for the mice immunized intranasally. The titers of HPV16 VLP specific IgA and/or IgG induced by parenteral or intranasal immunization in the vaginal washes were grouped according to their estrous stage. The mean titers are shown in Fig. 2. The variation among specific antibody titers along the estrous cycle was similar to the variation observed when total IgG or IgA was considered (Fig. 1), although, as expected, no or barely detectable specific IgA was induced in mice immunized parenterally. The specific IgG titers in the vaginal washes were very similar in the two groups of mice, with a maximal mean titer of 150 during diestrus. This probably reflected the serum specific IgG titers, which were also similar in these mice (205,000 and 225,000). This is in agreement with the hypothesis that specific IgG in vaginal washes is derived mainly from the serum by transudation, a phenomenon that varies during the estrous cycle, probably along with changes in the permeability of the genital mucosa, while IgA is produced locally (6, 25, 29, 44).

FIG. 2.

Anti-HPV16 VLP antibodies in vaginal washes of mice immunized intranasally (A) or parenterally (B) with purified HPV16 VLPs. The mean anti-HPV16 VLP IgG and IgA titers, expressed as the reciprocal of the highest dilutions that yielded an optical density at 492 nm four times that of preimmune samples, in vaginal washes collected from mice at different stages of the estrous cycle are shown. E, estrus; M, metestrus; D, diestrus; P, proestrus.

To confirm this, we have determined the number of total or specific antibody-secreting cells (ASCs) in the genital tract by enzyme-linked immunospot assay (9, 38). Briefly, single-cell suspensions of uterine horns and vagina-cervix were obtained by enzymatic digestions (thermolysin at 0.25 mg/ml, followed by collagenase-dispase at 1 mg/ml and DNase at 2 mg/ml), followed by filtration through 50-μm-pore-diameter nylon filters. Maxisorb plates (NUNC) were coated with 70 ng of purified HPV16 VLPs or 100 ng of rabbit anti-mouse Igs (17, 26), and spots representing individual ASCs were visualized by alkaline phosphatase-conjugated secondary antibodies revealed with 5-bromo-4 chloro-3-indolyphosphate (BCIP) substrate. ELISPOT analysis of total Ig-secreting cells (SCs) performed with 11 mice in diestrus and 4 mice in estrus revealed that, in general, more ASCs were found in uterus than in vagina-cervix (Fig. 3). This is in agreement with the reported higher numbers of plasma IgA cells in the uterine horn and body (30, 31). However, we detected more IgA SCs during diestrus than during estrus, while the opposite was reported when plasma IgA cells were considered (32). The higher level of secretory component reported during estrus in rodents (46) may finally account for the large amount of IgA measured in secretions during estrus. Unexpectedly, high numbers of IgG SCs were found in both uterus and vagina-cervix, while IgA SCs appeared to be less abundant. However, high numbers of HPV16 VLP specific ASCs were only found after intranasal immunization both in uterus and in vagina-cervix. These ASCs were of the IgA isotype and represented between 2 and 15% of the total IgA SCs detected (Table 1), while few if any specific IgG SCs were detected after intranasal (Table 1) or parenteral immunization (below 0.1% of the total IgG SCs [data not shown]). Altogether, our data suggest that in contrast to the specific IgA, only a few specific IgGs might be produced locally in the genital tract, while the majority are derived from the serum.

FIG. 3.

Ig SCs in the genital tract of immunized mice sacrificed at estrus or diestrus. The mean numbers of IgA or IgG SCs/10⁶ cells are indicated separately for uterus and vagina-cervix.

TABLE 1.

HPV16 VLP specific ASCs after intranasal immunization with purified HPV16 VLPs

Sample	Anti-HPV16 VLP SCs
	IgA		IgG
	No./106 total cells	% of total IgA SCs	No./106 total cells	% of total IgG SCs
1
Uterus	15	2.3	0	0
Vagina-cervix	7	7	3	0.1
2
Uterus	220	6	4	0.1
Vagina-cervix	180	8.2	10	0.5
3
Uterus	112	4.3	0	0
Vagina-cervix	20	3.3	0	0
6
Uterus	265	13.2	3	0.1
Vagina-cervix	60	15	8	0.3
8
Uterus	26	3.5	3	0.1
Vagina-cervix	30	15	3	0.1

Neutralization efficacy of the vaginal washes collected along the estrous cycle of mice immunized intranasally or parenterally.

We had previously shown that anti-HPV16 VLP antibodies at titers of 64 and 98, respectively, for IgA plus IgG and IgG alone resulted in 50% neutralization in the HPV16 pseudotyped in vitro neutralization assay (2, 33). However, the mean anti-HPV16 VLP IgG titers in vaginal washes after parenteral immunization only reached titers over 100 when the samples were collected during diestrus (Fig. 2), suggesting that only at that time might the vaginal washes be fully neutralizing. In contrast, although similar specific IgG titers were found after intranasal immunization, titers of anti-HPV16 VLP IgA over 100 are found at all stages of the estrous cycle, suggesting that full neutralization could occur at all times. To confirm this, vaginal washes were pooled from mice at same estrous stage after either intranasal or parenteral immunization and then analyzed in the pseudotyped in vitro neutralization assay (33). To perform the neutralization assays, the vaginal washes must be diluted with pseudovirions. Therefore, in order to achieve relatively high final titers of specific antibodies in the neutralization assays, we have pooled vaginal washes harboring the highest titers of specific antibodies. This resulted in different volumes and titers of specific antibodies in the pooled secretions pending the estrous stage and the route of immunization. The final anti-HPV16 VLP IgA or IgG titers achieved in the neutralization assays with the resulting neutralization efficacies are shown in Table 2. Relatively high anti-HPV16 VLP titers (84 to 182) could be achieved with three pools of vaginal washes (diestrus, metestrus, and proestrus) collected from mice immunized intranasally, resulting in almost full neutralization in the assay. The lowest neutralization efficacy (71%) obtained with vaginal washes collected in estrus reflected the lowest final titers of IgA (a titer of 61) achieved in this pool of sample and also the lowest mean titer of specific IgA plus IgG (a titer of 172) measured after intranasal immunization (Fig. 2). In contrast, only with the samples collected during diestrus and metestrus after parenteral immunization could we achieve titers of specific IgG (titers of 63 and 41, respectively) that resulted in partial neutralization (67 and 61%), while the two other vaginal wash pools were not neutralizing. Overall, as we had previously reported (2), the results obtained with in vitro neutralization faithfully reflected the titers of anti-HPV16 antibodies measured by ELISA. This confirmed that only intranasally immunized mice harbor titers of specific antibodies in vaginal washes which could fully neutralize HPV16 all along the estrous cycle.

TABLE 2.

Neutralization efficacy of vaginal washes pooled from mice immunized intranasally (tubes 1 to 4) or parenterally (tubes 5 to 8) with purified HPV16 VLPs

Tube (type of sample)a	Final sample dilution	Anti-HPV16 VLP titer		% Neutralization
		IgA	IgG
1 (E, nasal)	2.6 × 10−1	61	0	71
	1.5 × 10−1	35	0	33
	5.7 × 10−2	13	0	5
2 (D, nasal)	3.2 × 10−1	182	23	95
	1.9 × 10−1	52	14	50
	7.4 × 10−2	20	5	16
3 (M, nasal)	3.4 × 10−1	106	15	93
	2.1 × 10−1	62	9	80
	7.9 × 10−2	25	4	35
4 (P, nasal)	2.6 × 10−1	84	10	93
	1.5 × 10−1	48	6	73
	5.4 × 10−2	18	2	30
5 (E, parenteral)	3.8 × 10−1	0	5	9
	1.7 × 10−1	0	2	17
6 (D, parenteral)	4.5 × 10−1	0	63	67
	2.2 × 10−1	0	30	20
7 (M, parenteral)	3.7 × 10−1	0	41	61
	1.7 × 10−1	0	18	29
8 (P, parenteral)	3.4 × 10−1	0	22	10
	1.5 × 10−1	0	9	11

^aVaginal washes were pooled according to the route of immunization (intranasal or parenteral) and the estrous stage: E, estrus; D, diestrus; M, metestrus; and P, proestrus. 

Whether the induction of sIgA in addition to IgG is important to protect a woman’s genital tract against HPV infections is not yet clear. Variation in Igs along the menstrual cycle has also been observed in women (5, 22, 37, 40, 42, 45), although the data reported with different sampling techniques, genital fluids, and periods of the menstrual cycle sampled do not permit a definitive conclusion to be drawn. However, our data from mice strongly suggest that a mucosal route of immunization with the additional induction of specific sIgA may overcome the variations in IgG content which also occur along the menstrual cycle in women.