T-cell immunity induced and reshaped by an anti-HPV immuno-oncotherapeutic lentiviral vector

Abstract

We recently developed an immuno-oncotherapy against human papillomavirus (HPV)-induced tumors based on a lentiviral vector encoding the Early E6 and E7 oncoproteins of HPV16 and HPV18 genotypes, namely "Lenti-HPV-07". The robust and long-lasting anti-tumor efficacy of Lenti-HPV-07 is dependent on CD8⁺ T-cell induction and remodeling of the tumor microenvironment. Here, we first established that anti-vector immunity induced by Lenti-HPV-07 prime has no impact on the efficacy of a homologous boost to amplify anti-HPV T-cell immunity. To longitudinally monitor the evolution of the T-cell repertoire generated after the prime, homologous or heterologous boost with Lenti-HPV-07, we tracked T-cell clonotypes by deep sequencing of T-Cell Receptor (TCR) variable β and α chain mRNA, applied to whole peripheral blood cells (PBL) and a T cell population specific of an immunodominant E7_HPV16 epitope. We observed a hyper-expansion of clonotypes post prime, accompanied by increased frequencies of HPV-07-specific T cells. Additionally, there was a notable diversification of clonotypes post boost in whole PBL, but not in the E7_HPV16-specific T cells. We then demonstrated that the effector functions of such Lenti-HPV-07-induced T cells synergize with anti-checkpoint inhibitory treatments by systemic administration of anti-TIM3 or anti-NKG2A monoclonal antibodies. While Lenti-HPV-07 is about to enter a Phase I/IIa clinical trial, these results will help better elucidate its mode of action in immunotherapy against established HPV-mediated malignancies.

Fig. 1. Anti-vector immunity and T-cell immunogenicity of Lenti-HPV-07 in a prime-boost regimen.

a Timeline of single injection or prime-boost immunization with 1 × 10⁹ TU of Lenti-HPV-07_ind or Lenti-HPV-07_nj, antibody and T-cell assays. b, c ED50 comparison of anti-VSV-G_ind neutralizing antibodies in sera from mice primed with Lenti-HPV-07_ind and boosted homologously with Lenti-HPV-07_ind or heterologously with Lenti-HPV-07_nj. ED50 against VSV-G_ind (b) or against VSV-G_nj (c) (n = 4/group). Statistical significance was determined using a repeated measures (RM) two-way ANOVA (*p < 0.05, ***p < 0.001, ****p < 0.001). IFN-γ ELISPOT responses of T splenocytes from the same individual mice were studied at the indicated time points after in vitro stimulation with (d) six peptide pools spanning the full sequence of VSV-G_ind, or (e) four peptide pools spanning the sequence of detoxified E6_HPV16, E7_HPV16, E6_HPV18 or E7_HPV18, as encoded by Lenti-HPV-07. f IFN-γresponses of T splenocytes of mice assessed at D14 after a mono-injection with 1 × 10⁹ TU of Ctrl lenti_ind, Lenti-HPV-07_ind or Lenti-HPV-07_ind heat-inactivated at 70 °C during 1 h (n = 6). Naive mice served as control (n = 2). Stimulation was performed using either the indicated lentiviral vectors or peptides both at 4 µg/ml. SFU spot forming unit. Statistical significance was determined by Mann–Whitney t-test (ns not significant, *p < 0.05, **p < 0.01, ****p < 0.0001).

Fig. 2. Dynamic evolution of TCRβ repertoire throughout a Lenti-HPV-07 prime-boost immunization regimen.

a Timeline of prime-boost immunizations with 1 × 10⁹ TU of Lenti-HPV-07_ind, Lenti-HPV-07_nj or Ctrl Lenti_ind and TCRβ sequencing, performed on RNA extracted from peripheral blood cells. b Estimation of the TCRβ repertoire diversity using “true diversity” index. c Relative abundance of clonotypes, defined as their frequencies (X), compared to the total number of reads, which allows to divide the clonotypes into Small (0 < X ≤ 10^–4), Medium (10^–4 < X ≤ 10^–3), Large (10^–3 < X ≤ 10^–2) and hyperexpanded (10^–2 < X ≤ 10⁰) categories. Relative abundance is the sum of the frequencies of all clonotypes from a given category. d Pie chart representing average proportions (mean) of Medium, Large and hyperexpanded categories in each group. e Follow up of the frequencies for the top 100 most abundant clonotypes, as defined at the post boost timepoint. f Classification of the top 100 most abundant clonotypes detected post boost, according to their expanding or contracting evolution since prime. New post boost clonotypes were defined from the comparison with top 100 post prime. Expanding or retracting/stabilizing clonotypes were defined as those having increasing or decreasing frequencies post boost respectively. g Frequencies of expanding or contracting clonotypes. Statistical significance was determined using a one-way ANOVA in (b) and, two-way ANOVA (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.001) in (c, f). Error bars on the histograms represent standard deviation.

Fig. 3. Vβ-Jβ rearrangements detected in Lenti-HPV-07-immunized mice.

Samples are those detailed in Fig. 2a. a Jβ- and Vβ-gene segment usage in all clonotypes from each sample. For each clonotype, the frequency of Jβ and Vβ usage was determined as detailed in Supplementary Figs. 2 and 3. Jβ and Vβ segment repertoire usage was compared among all mice at all timepoints according to a Jensen-Shannon Divergence model as detailed in Supplementary Figs. 2 and 3 and treated by Multidimensional Scaling (MDS) to represent repertoire usage gene proximity on a 2-dimension plot. b Chord diagrams representing the Jβ and Vβ association of in the top 100 most abundant clonotypes, as defined post boost. Chords are colored according to the Jβ-gene segment. c Heatmap repertoire overlap, showing Jβ and Vβ segment usage in association to CDR3 amino acid sequence of each clonotype, using Morisita’s overlap index. Ctrl Lenti (squares) and Lenti-HPV-07 (circles) groups, pre prime (pink), post prime (blue) and post boost (green) timepoints, and mouse numbers are indicated inside the symbols. Cluster dendrograms indicate the unweighted hierarchical proximities between samples. d Frequencies of the 15 clonotypes shared by all samples at each time point. e Frequencies of clonotypes which exclusively emerged in the Lenti-HPV-07_ind/nj-injected mice at each time point and their 3-to-6 sharing levels. Numbers of clonotypes per category are indicated on the right. Median of each violin box is indicated in red.

Fig. 4. Evolution of the TCRβ repertoire of the clonotypes specific to an immunodominant E7_HPV16 epitope after a single injection or a prime-boost Lenti-HPV-07_ind vaccination.

a Timeline of a mono-injection or prime-boost (i.m.) with 1 × 10⁹ TU of Lenti-HPV-07_ind. TCRβ sequencing was performed on bulk RNA extracted from CD8⁺ T splenocytes sorted by use of a PE-conjugated H-2D^b (RAHYNIVTF) dextramer (Supplementary Table 2 and Fig. 8). b Estimation of the TCRβ repertoire diversity using “true diversity” index. c Relative abundance of clonotypes, defined as their frequencies (X), compared to the total number of reads, which allows to divide the clonotypes into Small (0 < X ≤ 10^–4), Medium (10^–4 < X ≤ 10^–3), Large (10^–3 < X ≤ 10^–2) and hyperexpanded (10^–2 < X ≤ 10⁰) categories. Relative abundance is the sum of the frequencies of all clonotypes from a given category. d percentage of all reads of the top 10 most abundant clonotypes for each mouse. Each color corresponds to distinct clonotypes from the least abundant (top) to the most abundant (bottom). e Chord diagrams representing the Jβ and Vβ association of in the top 100 most abundant clonotypes. Chords are colored according to the Jβ-gene segment. f Network representing CDR3 sequence proximity between top 10 most abundant clonotypes of all post prime and post boost mice (n = 65 CDR3 sequences). An edge between CDR3 sequences was added according to their LD. LD = 1, large line; LD = 2, medium line; LD = 3, thin line. Groups represent a cluster of CDR3 sequence with an LD < 3. g Amino acids logo for each group of related CDR3 sequences. Statistical significance was determined using a one-way ANOVA in (b) and, two-way ANOVA (*p < 0.05, ****p < 0.001) in (c). Error bars on the histograms represent standard deviation.

Fig. 5. Synergistic effect of Lenti-HPV-07 and anti-TIM3 mAb in TC1 tumor regression.

a C57BL/6 mice were s.c. flank-engrafted with 1 × 10⁶ of TC1 cells. When the tumor volume reached an average of 130–140 mm³, mice were vaccinated with the optimal dose of 1 × 10⁹ TU/mouse of Lenti-HPV-07 (n = 7) or a Ctrl Lenti (n = 6). Tumor growth was monitored with electronic caliper. b Timeline of tumor engraftment and combinatory treatment with Lenti-HPV-07 and anti-TIM3 mAb. C57BL/6 mice (n = 11–12/group) were s.c. flank-engrafted with 1 × 10⁶ of TC1 cells. When the tumor volume reached an average of 130 mm³ (day 13), mice were randomized and vaccinated with the suboptimal dose of 1 × 10⁸ TU/mouse of Lenti-HPV-07 or a Ctrl Lenti. Mice were then treated with anti-TIM3 mAb (clone RMT3–23, Bioxcell) or Ig control (clone 2A3, Bioxcell) 2 or 3 times a week. A total of 8 injections was given from day 16 to day 36 (200 µg/injection). c Spaghetti plots of tumor growth. d Therapy response rates according to RECIST criteria as follows. Complete response is defined by the complete eradication of the tumor observed during the time of the experiment. Partial response is defined by at least two consecutive decreasing measures during the treatment. Progression-free survival (PFS) is defined in all mice to reflect the duration of the response; from the day that tumor reached their maximum size until the day that tumor exceeded the maximum size again. e PFS of mice bearing TC1 tumors and treated with Lenti-HPV-07 and TIM3 blocking antibody. f Survival curves of the animals shown in (c). Statistical significance was determined using a Log-rank Mantel–Cox test, *p = 0.0465, ****p < 0.0001) in (f). Mice were sacrificed when the size of the tumors reached 1500 mm³, in accordance with the defined humane endpoints. Error bars on the histograms represent standard deviation.

Fig. 6. Increased efficacy of Lenti-HPV-07 and anti-NKG2A mAb in TC1 tumor regression.

a Timeline of tumor engraftment and combinatory treatment with Lenti-HPV-07 and anti-NKG2A mAb. C57BL/6 mice (n = 12/group) were s.c. flank-engrafted with 1 × 10⁶ of TC1 cells. When the tumor volume reached an average of 120–140 mm³ (day 10), mice were randomized and vaccinated with the suboptimal dose of 1 × 10⁸ TU/mouse of Lenti-HPV-07. Mice were then treated 2-to-3 times a week with anti-NKG2A mAb (clone 20D5, Bioxcell) or Ig control (clone 2A3, Bioxcell). A total of 8 injections was given from day 14 to day 31 (200 µg/injection). b Spaghetti plots of tumor growth. c Therapy response rates according to RECIST criteria. d Progression-free survival time of mice bearing TC1 tumors treated with Lenti-HPV-07 and NKG2A blocking antibody. e Survival curves of the animals shown in (b). Statistical significance was determined using a Log-rank Mantel–Cox test, *p = 0.0134) in (e) or by unpaired t-test in (d). Mice were sacrificed when the size of the tumor volume reached 1500 mm³, in accordance with the defined humane endpoints. Error bars on the histograms represent standard deviation.