Monitoring of Vaccine-Specific Gamma Interferon Induction in Genital Mucosa of Mice by Real-Time Reverse Transcription-PCR

Abstract

Monitoring of T-cell responses in genital mucosa has remained a major challenge because of the absence of lymphoid aggregates and the low abundance of T cells. Here we have adapted to genital tissue a sensitive real-time reverse transcription-PCR (TaqMan) method to measure induction of gamma interferon (IFN-γ) mRNA transcription after 3 h of antigen-specific activation of CD8 T cells. For this purpose, we vaccinated C57BL/6 mice subcutaneously with human papillomavirus type 16 L1 virus-like particles and monitored the induction of CD8 T cells specific to the L1_165-173 H-2D^b-restricted epitope. Comparison of the responses induced in peripheral blood mononuclear cells and lymph nodes (LN) by L1-specific IFN-γ enzyme-linked immunospot assay and TaqMan determination of the relative increase in L1-specific IFN-γ mRNA induction normalized to the content of CD8b mRNA showed a significant correlation, despite the difference in the readouts. Most of the cervicovaginal tissues could be analyzed by the TaqMan method if normalization to glyceraldehyde-3-phosphate dehydrogenase mRNA was used and a significant L1-specific IFN-γ induction was found in one-third of the immunized mice. This local response did not correlate with the immune responses measured in the periphery, with the exception of the sacral LN, an LN draining the genital mucosa, where a significant correlation was found. Our data show that the TaqMan method is sensitive enough to detect antigen-specific CD8 T-cell responses in the genital mucosa of individual mice, and this may contribute to elaborate effective vaccines against genital pathogens.

The mucosal surfaces, such as those of the genital tract, are common entry sites for many sexually transmitted infectious agents like human immunodeficiency virus, herpes simplex virus, human papillomaviruses (HPV), or chlamydia. The generation of mucosa-specific cytotoxic T-cell responses thus plays a crucial role in the defense against infection or tissue dissemination of these pathogens. However, the monitoring of specific T-cell responses in the genital mucosa has been difficult because of the absence of lymphoid aggregates and the low abundance of T cells in such tissues. Antigen-specific gamma interferon (IFN-γ) enzyme-linked immunospot (ELISPOT) assays or chromium release cytotoxicity assays using genital tissues have been reported (4, 6, 7, 14), but they required animal pools or large surgical pieces together with in vitro antigen-specific restimulation/expansion steps, which may bias the magnitude and quality of the initial immune response. Alternatively, knowledge of antigen-specific T-cell receptor sequences has allowed direct ex vivo detection of CD8 T cells in mucosal tissues by reverse transcription (RT)-PCR (15), but this type of assay may not reflect the functional status of the cells. Recently, studies on the priming/trafficking of specific CD8 T cells in the mouse genital tract using flow cytometry analysis were reported (12, 20), but this required the in vivo transfer of T cells expressing a particular antigen-specific T-cell receptor into recipient mice, which cannot be easily translated to other antigen/pathogen systems. IFN-γ expression/secretion is an early and functional response of antigen-activated CD8 T cells, and here we have exploited the high sensitivity of real-time RT-PCR analysis using TaqMan chemistry to measure a 3-h antigen-driven IFN-γ induction in CD8 T cells isolated from the genital mucosa of individual mice. C57BL/6 mice were vaccinated subcutaneously (s.c.) with HPV type 16 (HPV16) virus-like particles (VLPs) as a model antigen, and induction of IFN-γ mRNA expression in response to stimulation with the H-2D^b-restricted L1_165-173 epitope (17) was measured after normalization to a noninducible molecule. We first analyzed samples (lymph node [LN] cells and peripheral blood mononuclear cells [PBMC]) in which the conventional IFN-γ ELISPOT assay technique could be run in parallel to ensure that the two methods yield similar measurements of the L1-specific CD8 T-cell response. The results obtained with cervicovaginal (CV) tissues by TaqMan were further compared to results obtained by IFN-γ ELISPOT assay of PBMC and LN from individual animals. Our findings illustrate the importance of determining the immune response in genital tissue, and not only in the periphery, in the design of efficient vaccine strategies to fight against genital pathogens.

MATERIALS AND METHODS

Immunization of mice.

Six- to 8-week-old female C57BL/6 mice (Iffa Credo; France) were used in all experiments, and the ethical directives of the Swiss veterinary authorities were followed. Baculovirus-derived purified HPV16 VLPs were administered s.c. into the back at the base of the tail (2, 16). A dose of 75 or 100 μg, depending on the VLP stock, was delivered in a 100-μl inoculum volume. For each experiment, the number of mice per group is mentioned in Results.

Preparation of LN and CV tissue suspensions.

Mice were sacrificed by inhalation of CO₂, and LN and genital tracts were harvested. Single-cell suspensions were obtained by pressing the LN onto a 70-μm filter (Falcon) with a syringe piston and subsequently passing the cells through a 40-μm filter (Falcon). Dissociated cells were resuspended in complete high-glucose Dulbecco's modified Eagle medium containing glutamax-1 and sodium pyruvate supplemented with 10 mM HEPES, 1× nonessential amino acids, 100 U of penicillin-streptomycin/ml, 10% fetal calf serum (FCS) (all from Invitrogen], and 20 μM 2-mercaptoethanol (Sigma).

For preparation of CV cells, the uterus horns were removed. The remaining cervix and vagina were minced and washed twice in extraction buffer (Hanks balanced salt solution and 10 mM dithiothreitol). Minced tissues were then digested with 0.5 mg/ml thermolysin (Roche) in extraction buffer for 45 min at 4°C under agitation and then filtered through 150-μm-pore-diameter nylon filters. Isolated cells were kept at 4°C, and the remaining tissues were digested with 1 mg/ml collagenase/dispase (Roche) in Iscove's modified Dulbecco's medium with glutamax-1 (Gibco, Invitrogen) supplemented with 20% FCS for 45 min at 37°C under agitation and filtered as described above. Cells isolated after both enzymatic digestions were pooled, layered onto FCS, and centrifuged for 20 min at 1,800 rpm without a brake. Pelleted cells were then digested with 2 mg/ml DNase (Sigma) for 30 min at 37°C under agitation and filtered through 50-μm-pore-diameter nylon filters. Cells were collected after centrifugation, and viable cells were counted by trypan blue exclusion. Typically, 2 × 10⁵ to 1 × 10⁶ cells could be isolated from the vagina/cervix, depending on the estrous cycle stage of the mice.

Preparation of PBMC.

A 150-μl volume of tail blood was collected in 250 μl of phosphate-buffered saline (PBS) containing liquemine (500 UI/ml). PBMC were purified on lympholyte M (Cedarlane CL 5035) gradients and resuspended in complete Dulbecco's modified Eagle medium after red blood cell lysis.

IFN-γ ELISPOT assay.

Multiscreen-HA 96-well plates (MAHA S4510; Millipore) were coated overnight at 4°C with an anti-IFN-γ monoclonal antibody (R4-6A2; Pharmingen) at a concentration of 10 μg/ml in PBS. Plates were then blocked for 2 h at 37°C with PBS-1% BSA. One hundred thousand cells per well were incubated in duplicate with 1 μg/ml L1_165-173 peptide (17) or medium alone (control wells) for 16 to 24 h in the 96-well ELISPOT assay plates. A biotinylated anti-IFN-γ monoclonal antibody (AN 18.03.C12) (21) was then added at a concentration of 2 μg/ml in PBS-1% BSA, and the plates were incubated for 2 h at 37°C. After each incubation step, the plates were washed three times with PBS-0.1% Tween 20 and three times with PBS. After 1 h of incubation with streptavidin-alkaline phosphatase conjugate (1/2,000 in PBS-0.1% Tween 20; Boehringer), the plates were developed with a solution of 5-bromo-4-chloro-3-indolylphosphate (BCIP)-nitroblue tetrazolium (Roche) until blue spots appeared. Tap water was used to stop the reaction, and the plates were dried in air overnight. Individual spots were counted under a dissecting microscope. L1-specific responses were defined for each individual mouse as the number of IFN-γ spots per 10⁶ cells in the L1-stimulated wells minus the number of IFN-γ spots per 10⁶ cells in the control wells. The threshold of significance was calculated as the mean plus 3 standard deviations (SD) of the L1-specific responses of five naïve mice and corresponded to 30 spots/10⁶ cells in LN samples and 70 spots/10⁶ cells in PBMC.

Real-time PCR (TaqMan).

Quantification of the cDNAs was performed with the TaqMan universal master mix (Applied Biosystems) and the following forward (fw) and reverse (rv) primers and probes at optimized concentrations: IFN-γ fw primer, 5′CAGCAACAGCAAGGCGAAA3′ (400 nM); IFN-γ rv primer, 5′CTGGACCTGTGGGTTGTTGAC3′ (400 nM); IFN-γ probe, 5′FAM-AGGATGCATTCATGAGTATTGCCAAGTTTGA3′-TAMRA (200 nM); CD8b fw primer, 5′AAGAAGCAATGCCCGTTCC3′ (400 nM); CD8b rv primer, 5′TGAGGGTGGTAAGGCTGCA3′ (400 nM); CD8b probe, 5′TET-CCCAGAGACCCAGAAGGGCCTGAC3′-TAMRA (200 nM); GAPDH (glyceraldehyde-3-phosphate dehydrogenase) fw primer, 5′GAACATCATCCCTGCATCC3′ (80 nM); GAPDH rv primer, 5′CCAGTGAGCTTCCCGTTCA3′ (80 nM); GAPDH probe, 5′TET-CTTGCCCACAGCCTTGGCAGC3′-TAMRA (100 nM).

Primers and probes were designed to span exon-intron junctions to prevent amplification of genomic DNA (8, 18). All TaqMan probes were labeled at the 3′ end with the quencher dye molecule TAMRA (6-carboxytetramethylrhodamine). The IFN-γ probe was labeled at the 5′ end with the reporter dye FAM (6-carboxyfluorescein), and the CD8b and GAPDH probes were labeled with TET (tetrachloro-6-carboxyfluorescein). Reaction mixtures had a total volume of 25 μl, and the thermal cycler parameters included 2 min at 50°C, 10 min at 95°C, and 45 cycles involving denaturation at 95°C for 15 s and annealing/extension at 60°C for 1 min. Real-time monitoring of fluorescence emission from cleavage of sequence-specific probes by the nuclease activity of Taq polymerase allowed definition of the cycle threshold (Ct) during the exponential phase of amplification. The Ct was defined as the cycle at which fluorescence exceeds 10 times the SD above the mean of the background of the early cycles. All real-time PCR assays were performed in duplicate, and the mean cDNA quantities were calculated. Standard curves were generated for each gene quantified with 10-fold dilutions of cloned cDNAs (see below). The reagents were found to have good PCR amplification efficiency as determined by the slope of the standard curves (between −3.3 and −3.7), and linear regression analysis showed that all standard curves had an R² value of >0.99. To determine the intra-assay variability of the real-time PCR, 10-fold serial dilutions of CD8b, IFN-γ, and GAPDH cDNAs (ranging from 10¹ to 10⁵ copies) were amplified in quadruplicate in the same run. For interassay variability, the same cDNA dilutions were analyzed in six different experiments performed over a 1-month period.

Cloning of IFN-γ, CD8b, and GAPDH cDNA amplicons and preparation of DNA samples for standard curves.

IFN-γ, CD8b, and GAPDH amplicons obtained with the respective fw and rv primers were cloned into the EcoRV site of pGEM-T-easy vectors (Promega, Madison, WI). The plasmid DNAs were linearized with NcoI, purified, and quantified by spectrophotometer and on agarose gels. The number of copies was calculated, and the DNAs were stored at 10⁷ copies/μl in 10 mM Tris (pH 8.5) with 20 ng/μl salmon sperm DNA as a carrier.

Determination of antigen-specific relative increase in IFN-γ by real-time RT-PCR.

Isolated cells (50,000 to 400,000/well) were incubated in round-bottom 96-well plates (Nuncatom 16332) in the presence or absence of the L1_165-173 peptide at 1 μg/ml for 3 h at 37°C in 5% CO₂. This peptide pulse duration appeared to be optimal, as after an additional 2 h of incubation, the IFN-γ mRNA amount was decreased more than threefold (data not shown). Cells were harvested, and RNA was extracted with an RNeasy mini-kit (Qiagen), eluted in 30 μl of H₂O, and kept at −20°C. Aliquots of 9 μl of purified RNA were reverse transcribed with random hexamers with the TaqMan RT reagents (P/N N808-0234; Applied Biosystems) in 23-μl final volumes. Aliquots of 5 μl of cDNAs were directly used in duplicate for the quantification of IFN-γ and CD8b, while cDNA was diluted 20-fold for GAPDH quantification. Results were considered only when the difference between the Ct obtained for the duplicates was less than 1. In case the difference exceeded 1, the RT-PCR was performed again with a new cDNA aliquot of the same sample. Normalization of samples was performed by dividing the amount of IFN-γ cDNA by the amount of CD8b or GAPDH cDNA, depending on the experiments. Relative increases in IFN-γ were calculated by dividing the normalized number of IFN-γ cDNA copies in L1-stimulated wells by those measured in medium-stimulated wells. If samples were stimulated with an unrelated HPV peptide instead of medium alone, the baseline normalized IFN-γ cDNA levels were similar (data not shown). Therefore, medium alone was used as a control in the experiments described here.

Statistical analysis.

Spearman correlation analysis with two-tailed P values was performed with Prism 4.00 for Windows (GraphPad Software, San Diego, CA).

RESULTS

Method overview.

The amplitude of the antigen-specific CD8 T-cell response obtained after vaccination was determined by the induction of IFN-γ mRNA expression that could be achieved after 3 h of ex vivo antigen-specific stimulation. A real-time RT-PCR was set up to determine the quantities of IFN-γ, CD8b, and GAPDH mRNAs by a two-step method that consisted of RT with random hexamers, followed by a real-time PCR with the specific fw and rv primers, as well as the labeled probes indicated in Materials and Methods. PCR amplification of cDNA plasmids as standards showed that the amplification efficiency was high though not the same for the three genes of interest, with a low level of variability among independent experiments, as shown by the mean slopes ± SD and respective mean R² ± SD of 15 independent runs (−3.492 ± 0.150 and 0.9961 ± 0.0028 for IFN-γ, −3.545 ± 0.096 and 0.9950 ± 0.0034 for CD8b, and −3.539 ± 0.115 and 0.9954 ± 0.0032 for GAPDH). The coefficients of variation of the intra-assay Ct values ranged from 0.5 to 2.2% for IFN-γ, 1.2 to 1.5% for CD8b, and 0.7 to 3.5% for GAPDH (Table 1). The interassay coefficients of variation ranged from 1.5 to 2.9% for IFN-γ, 1.3 to 3.6% for CD8b, and 1.5 to 3.1% for GAPDH and were therefore comparable for each cDNA of interest. L1_165-173-specific CD8 T-cell ex vivo stimulation was first examined in PBMC and cervical and lumbar LN and CV tissues from five naïve mice to assess the background levels of IFN-γ induction. The thresholds of significance were calculated as the mean relative increase in L1_165-173-specific IFN-γ induction plus 3 SD and were as follows: 2.8 (normalized to CD8b cDNA) and 3.9 (normalized to GAPDH cDNA) for CV tissue samples, 2.6 (normalized to CD8b cDNA) for PBMC samples, and 5.7 (normalized to CD8b cDNA) for LN samples. For PBMC and LN samples, the IFN-γ content was normalized to the content of CD8b as CD8 T cells are those that express L1_165-173-inducibleIFN-γ. However, normalization to GAPDH was also performed in the case of genital cell samples because the quantity of CD8 mRNA was too low in some cases to be reproducibly detectable (see below).

TABLE 1.

Intra- and interassay variations

cDNA and no. of copiesa	Intra-assay variation		Interassay variation
	Mean Ct (SD)	CVb (%)	Mean Ct (SD)	CV (%)
IFN-γ
105	24.20 (0.11)	0.47	24.38 (0.72)	2.90
104	27.74 (0.47)	1.68	27.78 (0.81)	2.91
103	31.13 (0.27)	0.88	31.19 (0.82)	2.61
102	34.80 (0.57)	1.63	34.89 (0.53)	1.53
101	37.82 (0.83)	2.21	37.98 (0.69)	1.80
CD8
105	20.91 (0.31)	1.50	21.22 (0.31)	1.46
104	24.65 (0.34)	1.40	24.78 (0.45)	1.84
103	28.02 (0.43)	1.54	28.64 (0.37)	1.31
102	31.55 (0.38)	1.22	31.93 (0.53)	1.65
101	36.15 (0.31)	0.85	35.76 (1.30)	3.62
GAPDH
105	21.12 (0.14)	0.66	21.44 (0.64)	2.90
104	24.76 (0.26)	1.07	25.17 (0.45)	1.80
103	28.32 (0.25)	0.87	28.70 (0.44)	1.54
102	31.49 (0.22)	0.68	32.18 (0.82)	2.55
101	35.14 (1.23)	3.50	36.64 (1.14)	3.10

^aInput number of standard cDNA copies.

^bCV, coefficient of variation.

Time course analysis of L1-specific CD8 T-cell response in PBMC after s.c. vaccination with HPV16 VLPs.

To set up a method that allows determination of the antigen-specific CD8 T-cell response in the genital mucosa of individual mice, we evaluated first the amplitude and kinetics of this response in PBMC by conventional IFN-γ ELISPOT assay. The aim was twofold: first to determine the peak of the CD8 T-cell response in PBMC in order to analyze the immune response in the genital mucosa at a time when a maximal number of CD8 T cells are migrating to effector sites and second to generate a great number of samples/data for comparing the conventional method to the new TaqMan method. Figure 1 shows the L1_165-173-specific IFN-γ ELISPOT assay results of a compilation of four experiments in which a total of 41 mice were vaccinated once with 75 or 100 μg of HPV16 VLPs s.c. at day 1, sampled on 2 consecutive days between day 4 and day 8, and finally sacrificed between day 7 and day 9. The L1-specific response peaks at day 7 and then tends to decrease over the next days, meaning that a maximal number of L1-reactive CD8 T cells circulates in the periphery at this time point and is therefore susceptible to reach peripheral tissues such as the genital tract.

FIG. 1.

Kinetics of L1-specific CD8 T-cell response in PBMC after s.c. vaccination with HPV16 VLPs. C57BL/6 mice were immunized with VLPs at day 1 and bled daily from day 4 through day 9. The L1_165-173-specific CD8 T-cell response was then analyzed in PBMC by ex vivo IFN-γ ELISPOT assay as detailed in Materials and Methods. Results are shown as the number of L1-specific IFN-γ spots/10⁶ PBMC. The horizontal bar represents the mean response of mice at each time point.

Correlation between the data obtained by the TaqMan method and the conventional IFN-γ ELISPOT assay in the periphery.

As a further step in the validation of the TaqMan method, we compared the immune responses induced in PBMC and LN by conventional ex vivo L1_165-173-specific IFN-γ ELISPOT assays and TaqMan determination of the relative increase in L1_165-173-specific IFN-γ mRNA normalized to the content of CD8b mRNA. PBMC samples of 21 mice s.c. immunized with 100 μg of HPV16 VLPs, for which the IFN-γ ELISPOT assay results are shown in Fig. 1, were also analyzed by the TaqMan method. Figure 2A shows the data for samples taken at different time points between days 7 and 9. Cells isolated from pools of specific LN (cervical, ilio-sacral, inguinal, mesenteric, and brachial) of the same mice were also analyzed by both methods (Fig. 2B). Despite the difference between the readouts (the number of IFN-γ secreting cells versus the increase in IFN-γ mRNA expression), the two methods showed a significant correlation (Spearman r = 0.68 and 0.77 for PBMC and LN, respectively; P < 0.0001), suggesting that the two methods yield similar measurements of antigen-specific CD8 T-cell responses. The TaqMan method may thus advantageously replace the classical IFN-γ ELISPOT assay for the investigation of vaccine-induced CD8 responses in individual CV tissue samples, where limited numbers of cells, which harbor a low immune cell content (<1% of T cells; data not shown), are available.

FIG. 2.

Correlation of ELISPOT assay and TaqMan results obtained with PBMC and LN cells. PBMC from individual VLP-immunized mice (A) and pooled LN cells from five VLP-immunized mice (B) were evaluated for L1_165-173-specific IFN-γ secretion by ELISPOT assay and for mRNA induction by TaqMan assay (see Materials and Methods for details). ELISPOT assay data (horizontal axis) were plotted against TaqMan data (vertical axis) for each sample. ELISPOT assay results are expressed as the number of L1-specific IFN-γ spots/10⁶ cells. TaqMan results are shown as relative increases in L1-specific IFN-γ, which were determined as the IFN-γ/CD8b ratio of cell samples incubated with L1_165-173 divided by the IFN-γ/CD8b ratio of cell samples incubated with medium alone.

Analysis of the L1_165-173-specific CD8 response in CV cells by the TaqMan method after s.c. vaccination of mice with HPV16 VLPs.

Analysis of L1-specific IFN-γ induction in CV tissue samples was performed by the TaqMan method with 20 mice that were immunized s.c. with 75 μg of HPV16 VLPs at day 1 and sacrificed at day 7, for which the IFN-γ ELISPOT assay results of PBMC samples are shown in Fig. 1. The amount of RNA recovered from CV cells allowed us to perform three independent RT-PCR runs for each sample and each gene of interest (IFN-γ, CD8b, and GAPDH). When relative increases in IFN-γ induction were calculated after normalization to CD8b, only 15 out of 20 samples could be analyzed because in the remaining 5 samples, the CD8b copy number was too low to be reproducibly detectable after three attempts (Fig. 3A). When normalization to GAPDH was performed, 18 samples could be analyzed because in the other 2 samples, the number of IFN-γ cDNA copies measured was also very low and thus not reliable. Five mice were found to have a relative increase in IFN-γ in the CV tissues above the significance threshold after normalization to both CD8b and GAPDH. An additional mouse, for which the CD8b copy number was too low to be reproducibly detectable, was also positive after normalization to GAPDH. In one mouse, the relative increase in IFN-γ was shown to be positive after normalization to CD8b (4.8-fold) and not after normalization to GAPDH (2.7-fold), and in another mouse, it was the opposite (1.6-fold when normalized to CD8b and 4.3-fold when normalized to GAPDH), but the values were close to the respective detection limits. The mean relative increases calculated according to both normalization methods were very similar. Moreover, there was a strong correlation between the relative increases calculated by the two methods (Fig. 3B). In conclusion, out of 20 mice, 18 samples of CV tissues could be analyzed by the TaqMan method after normalization to GAPDH and 7 samples showed an L1-specific induction of IFN-γ mRNA.

FIG. 3.

L1-specific IFN-γ mRNA induction in CV cells of mice vaccinated with HPV16 L1 VLPs. (A) CV cells of VLP-immunized mice were evaluated for L1_165-173-specific IFN-γ mRNA induction by TaqMan. The number of IFN-γ mRNA copies was normalized to the number of CD8b and GAPDH mRNA copies for each sample. Relative increases in L1-specific IFN-γ were then determined as the IFN-γ/CD8b or GAPDH mRNA ratio of cell samples incubated with L1_165-173 divided by the IFN-γ/CD8b or GAPDH mRNA ratio of cell samples incubated with medium alone. Dashed lines indicate respective thresholds of significance. (B) Relative increases in L1-specific IFN-γ obtained for individual animals after normalization to GAPDH (horizontal axis) were plotted against the relative increases obtained after normalization to CD8b (vertical axis).

Correlation of the L1-specific TaqMan response of CV cells with the IFN-γ ELISPOT assay response in the sacral LN.

In the same 20 mice in which we investigated the L1-specific IFN-γ response in CV cells by the TaqMan method, we analyzed the L1-specific IFN-γ response by ELISPOT assay in the cervical LN (distant from the site of immunization) and the LN of the ilio-sacral region, in addition to PBMC. According to Soderberg et al., these latter LN comprise the lumbar LN (two or three LN) and the sacral LN (23). This sacral LN is located just underneath the lumbar LN and has been described as a “mucosal” LN because, in contrast to the lumbar LN, for example, it is maintained in lymphotoxin-β knockout mice, which are known to harbor only mucosal LN (23). In addition, the high endothelial venules of the sacral LN express the mucosal addressin MAdCAM-1. All of the mice harbored L1_165-173-specific IFN-γ spots in the PBMC and lumbar LN (Fig. 4A). Interestingly, we found no correlation between the magnitude of the responses detected in these immune compartments by IFN-γ ELISPOT assay and the induction of IFN-γ mRNA measured in CV tissue by the TaqMan method (Fig. 5A). Indeed, the mice in which we could observe L1-specific IFN-γ mRNA induction in the genital cells were not the ones that presented the highest ELISPOT assay response in the periphery. In a subset of 10 mice, the L1-specific IFN-γ ELISPOT assay response was, in addition, examined in the cervical and sacral LN (Fig. 4B). As reported by Soderberg et al., we found a sacral LN in 7/10 mice; interestingly, there was a strong correlation (Spearman r = 0.93, P < 0.01) between the relative increases in L1-specific IFN-γ mRNA measured in CV cells and the number of L1-specific IFN-γ spots detected in the sacral LN, but not with the number of spots enumerated in the cervical LN (Fig. 5B). In conclusion, the L1-specific response measured by the TaqMan method in the CV tissue correlates only with the response detected in the sacral LN and not with the response measured in the PBMC or lumbar or cervical LN.

FIG. 4.

L1-specific IFN-γ ELISPOT assay for PBMC and LN of mice immunized with HPV16 VLPs. PBMC and lumbar LN cells of 20 VLP-immunized mice (A) and cervical and sacral LN of 10 VLP-immunized mice (B) were evaluated for L1_165-173-specific IFN-γ secretion by ELISPOT assay. Results are shown as the number of L1-specific IFN-γ spots/10⁶ cells. Dashed lines indicate respective thresholds of significance.

FIG. 5.

Correlation between IFN-γ ELISPOT assay results obtained with PBMC and LN cells of mice immunized with HPV16 VLPs and TaqMan results obtained with CV cells from the same mice. PBMC (A) and lumbar LN cells (B) from 20 VLP-immunized mice and cervical LN (CLN) (C) and sacral LN (D) cells from 10 VLP-immunized mice were evaluated for L1_165-173-induced IFN-γ secretion by ELISPOT assay. CV cells were analyzed for L1_165-173-induced IFN-γ mRNA expression by TaqMan. ELISPOT assay results of PBMC or LN cells (vertical axis) were plotted against the TaqMan results of CV cells from the same animals (horizontal axis).

DISCUSSION

Monitoring of vaccine-induced T cells in the genital tract has rarely been done. This is mainly due to the paucity of immune cells that can be recovered from the genital mucosa of individual mice and the relatively high number of cells necessary to perform the conventional assays generally used to detect activated vaccine-specific CD8 T cells. Here we were interested in exploiting a sensitive method to quantify the expression of IFN-γ mRNA as a measurement of CD8 T-cell activation in CV cell samples from individual mice. We chose the HPV16 VLP as a model antigen and first determined the kinetics of the CD8 T-cell response specific for a known CTL epitope of HPV16 L1, L1_165-173, in the PBMC of mice by the IFN-γ ELISPOT assay. We found that the L1_165-173-specific response peaked at day 7 after s.c. vaccination with VLPs, which is in accordance with the kinetics observed after infection with viruses such as lymphocytic choriomeningitis virus (10) and influenza virus (5, 9) or after vaccination with synthetic peptides (13). We also observed that the highest number of specific CD8 T cells was present at the same time point in the spleen; however, the response declined less rapidly in that organ (data not shown). Analysis of PBMC and LN samples by IFN-γ ELISPOT assay and real-time RT-PCR in parallel showed that despite the difference between the readouts (the number of IFN-γ-secreting cells after overnight stimulation versus the increase in IFN-γ mRNA expression after 3 h of stimulation), the data obtained by the two methods present a significant correlation. Interestingly, out of 32 PBMC samples analyzed, 23 were scored positive and 7 were negative by both methods and only 2 were scored negative with the ELISPOT assay but showed positive, albeit small (3- and 7.1-fold), relative increases in IFN-γ by real-time RT-PCR. Out of the 22 LN samples, 17 were scored positive and 3 were negative by both methods and only 2 were scored negative by the TaqMan method but harbored a significant, although low (67 and 63 spots/10⁶ cells), number of L1-specific IFN-γ-secreting cells in the ELISPOT assay. Therefore, 92.6% of the samples showed concordant results by both methods; the four remaining samples were near the threshold of significance of both techniques and appeared to be positive by either the TaqMan or the ELISPOT assay. Thus, as previously reported (25), the two assays appear to be comparable in sensitivity. In our experience, around 50 L1-specific spots/10⁶ PBMC or LN cells can be routinely measured with the IFN-γ ELISPOT assay when as few as 100,000 cells/well are used, which corresponds to a frequency of 1 specific CD8 T cell out of a total of 20,000 cells. This is indeed close to what we observed by TaqMan analysis; i.e., 2 cells from a CD8 T-cell clone specific for a defined peptide were detected among 100,000 unrelated cells not producing IFN-γ (corresponding to a frequency of 1/50,000; unpublished results). We have to keep in mind, however, that individual effector cells likely produce less IFN-γ upon specific stimulation than do T-cell clones; therefore, their detection frequency may be lower. PBMC and LN cell populations harbor >10% CD8 T cells, which is at least 1 log higher than the level in CV cells. Tenfold more CV cells/well may thus be needed in an IFN-γ ELISPOT assay to detect immune responses close to the threshold of significance. Such a number of cells is clearly not achievable from a single mouse and cannot be handled in 96-well plates without leading to high background levels (unpublished observations). In conclusion, the main advantage of the real-time RT-PCR assay resides in the capacity to analyze smaller samples than those necessary for the ELISPOT assay or tetramer or flow cytometry analysis. Handling of cells for TaqMan analysis takes longer and requires multiple steps (such as extraction of total RNA, RT, and real-time PCR), compared to the ELISPOT assay. The latter will therefore remain the method of choice for the analysis of larger cell samples such as those isolated from the blood, spleen, and LN, although a real-time RT-PCR technique has already been set up to measure vaccine-induced IFN-γ mRNA in PBMC from melanoma patients previously vaccinated with a synthetically modified peptide epitope from the melanoma self-antigen gp100 (11). Here, we measured the induction of IFN-γ mRNA in response to a particular peptide, L1_165-173, after vaccination with HPV16 VLPs as a model antigen. Rather than targeting the response to a particular CTL epitope, a more general pattern of activated CD8 T cells could be analyzed by using antigen-presenting cells that have been engineered to express entire protein antigens or overlapping peptides spanning these antigens during the in vitro 3-h stimulation step. This would bypass the need to identify immunodominant CTL epitopes and therefore to be limited to particular major histocompatibility complex class I types. Such stimulation protocols are currently used in clinical trials using the IFN-γ ELISPOT assay to evaluate the PBMC responses of patients and healthy donors vaccinated with potential therapeutic vaccines against HPV16 (1, 3, 22). Application of the real-time RT-PCR method described here to human samples is therefore feasible, especially when a small amount of material is available, as in the case of tissue biopsy. In such a situation, it could be necessary to perform an in vitro expansion of the biopsy-isolated cells before the 3-h antigen exposure to reach a detectable number of vaccine-activated cells.

The ability to investigate the L1-specific IFN-γ mRNA induction in CV cells of individual mice by TaqMan analysis allowed us to compare the responses elicited in different immune compartments after vaccination with VLPs. Our data clearly show that the L1-specific immune response measured in genital tract cells only correlates with the response measured in the sacral LN and not with that detected in the population of circulating lymphocytes, as well as in the distant cervical LN and the local lumbar LN. It has to be noted that there is an ambiguity concerning the murine LN nomenclature in the scientific literature (24). In particular, several publications refer to the iliac LN but without clearly describing which LN were indeed taken. Our results underline the existence of a compartmentalization of the immune response in the anatomically juxtaposed LN of the ilio-sacral region, and they suggest that the sacral LN may represent a surrogate of the genital tract when analyzing the T-cell response. Alternatively, there may be a correlation between the L1-specific immune responses measured in the genital tract and the sacral LN because this particular LN, in contrast to the other LN of the iliac region, does not drain the s.c. immunization site we used in our experiments. Thus, the L1-specific CD8 T cells detected in the lumbar LN comprise, in addition to genital lymphocytes, lymphocytes preferentially recirculating through these LN because they have been primed there. The immune response observed in the lumbar LN would therefore not uniquely reflect the response detected in the genital tract. Knowing that four or five LN are localized in the ilio-sacral region and that they differ in their mechanisms of development, as well as in the expression of addressins on their high endothelial venules (19), further studies are needed to address the difference in the behavior of the lymphocytes present in these LN and their role in providing immunity to the genital mucosa. Development of sensitive assays to monitor the induction of immune responses in the genital tract is of particular relevance in order to fight against infection with many sexually transmitted infectious agents, such as human immunodeficiency virus, herpes simplex virus, HPV, or chlamydia. Measurements of immunological parameters have almost always focused on the systemic immune system, while vaccines capable of protecting against those pathogens most likely need to induce long-term mucosal immune responses, as the mucosal surfaces are the most natural routes for their transmission. The real-time RT-PCR approach described here provides a practical method to analyze the CD8 immune responses elicited in the genital tract of mice after vaccination. It may contribute to the optimization of immunization regimens in preclinical studies to better understand the quantitative and qualitative aspects of the immune responses that are needed to achieve an effective clinical response. In addition, our results suggest that determining the immune response in the genital tissue (or the sacral LN as a potential surrogate of the genital mucosa) and not only in the periphery may help in designing efficient vaccine strategies to fight against genital pathogens.