Poly I:C vaccination drives transient CXCL9 expression near B cell follicles in the lymph node through type-I and type-II interferon signaling

Abstract

Subunit vaccines drive immune cell-cell interactions in the lymph node (LN), yet it remains unclear how distinct adjuvants influence the chemokines responsible for this interaction in the tissue. Here, we tested the hypothesis that classic Th1-polarizing vaccines elicit a unique chemokine signature in the LN compared to other adjuvants. Polyinosinic:polycytidylic acid (Poly I:C) vaccination resulted in dynamic upregulation of CXCL9 that was localized in the interfollicular region, a response not observed after vaccination with alum or a combination of alum and poly I:C. Experiments using in vivo mouse models and live ex vivo LN slices revealed that poly I:C vaccination resulted in a type-I IFN response in the LN that led to the secretion of IFNγ, and type-I IFN and IFNγ were required for CXCL9 expression in this context. CXCL9 expression in the LN was correlated with an IgG2c antibody polarization after vaccination; however, genetic depletion of the receptor for CXCL9 did not prevent the development of this polarization. Additionally, we measured secretion of CXCL9 from ex vivo LN slices after stimulation with a variety of adjuvants and confirmed that adjuvants that induced IFNγ responses also promoted CXCL9 expression. Taken together, these results identify a CXCL9 signature in a suite of Th1-polarizing adjuvants and determined the pathway involved in driving CXCL9 in the LN, opening avenues to target this chemokine pathway in future vaccines.

Keywords: Antibody; Extrafollicular; Interferon gamma-induced protein 10 (IP-10); Monokine induced by gamma interferon (MIG); Polarization; Tissue slices.

Figure 1:. Poly I:C vaccination induced upregulation of CXCL9 around B cell follicles.

(a) Representative single channel images of CXCL9, CXCL12 and CXCL13, and composite image, from live LN slices collected 16 hours post vaccination and immunolabelled. (b-d) Mean grey value (MGV) of CXCL9, CXCL12, and CXCL13 immunolabeling, after vaccination with each adjuvant. Bars show mean ± standard deviation from N = 3 mice. Each dot represents the MGV averaged across one slice. Brown-Forsythe and Welch ANOVA with multiple comparisons. ****p < 0.0001. (e) Location of CXCL9 immunolabeling in the slices from the draining LN from mice 16 hours post vaccination with poly I: C. Bars show mean ± standard deviation of slices from N = 3 mice. Paired t-test. ***p < 0.001. (f) Concentrations of CXCL9 in the interstitial fluid of the draining lymph node, 16 hours post in vivo vaccination. Bars show mean ± standard deviation in tissues from N = 4 mice. Each dot represents two pooled and mechanically dissociated lymph nodes. The dotted line indicates LOD of 19.0 pg/mL.

Figure 2:. IFN-α and IFN-γ were required for CXCL9 expression after poly I:C vaccination.

(a-b) Secreted IFN-α (a) and IFN-γ (b) in the draining lymph node 12 hours post in vivo vaccination. Bars show mean ± standard deviation from N = 4 mice. Each dot represents two pooled and mechanically dissociated lymph nodes. Brown-Forsythe and Welch ANOVA with multiple comparisons. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Dotted line indicates LOD of 4.14 pg/mL in (a) and 40.1 pg/mL in (b). (c) Representative images of LN slices immunostained for CXCL9 and CXCL10, 16 hr post vaccination with poly I:C, in tissues from WT, IFNAR KO, and IFNGR KO mice. (d) Area of LN slices 16 hours post vaccination, measured by imaging. Two-way ANOVA with multiple comparisons. ****p < 0.0001. (e-f) Quantification of CXCL9 (e) and CXCL10 (f) immunofluorescence in LN slices. Two-way ANOVA with multiple comparisons. ****p < 0.0001. ns indicates p > 0.05.

Figure 3:. IFNα and IFNγ positive cells were detected in the interfollicular region of the lymph node after poly I:C stimulation ex vivo.

(a) Representative images of LN slices 16 hours post ex vivo poly I:C stimulation, with digitally zoomed in views of the IFNα and IFNγ channels. (b-c) Number of IFNα (b) or IFNγ (c) positive cells in each LN slice after ex vivo poly I:C stimulation. Bars show mean ± standard deviation from n = 8 slices from 3 mice. Each dot represents one slice. One-way ANOVA with multiple comparisons. **p < 0.01 ***p < 0.001. (d-e) Location of IFNα (d) and IFNγ (e) cells after ex vivo poly I:C stimulation and BFA treatment. IFZ indicates interfollicular zone and TCZ indicates T cell zone. Bars show mean ± standard deviation from n = 8 slices from 3 mice. Paired t-test. ****p < 0.0001. (f) Representative 40x single channel IFNα, IFNγ and B220 and composite images of LN slices 16 hours post ex vivo poly I:C stimulation.

Figure 4:. Type-I IFN signaling and IFNγ were sufficient for CXCL9 expression.

(a) Schematic of experimental design. Slices were cultured in media supplemented with cytokine for 16 hours, before the supernatant was collected for ELISA. (b-c) Secreted CXCL9 (a) or CXCL10 (b) by ex vivo LN slices after 16-hour stimulation with 0.2, 2 or 20 ng/mL IFNα, IFNγ, or IFNα and IFNγ. Bars show mean ± standard deviation from n = 8 slices from 4 mice. Brown-Forsythe and Welch ANOVA with comparisons against the unstimulated group. *p < 0.05, **p < 0.01, ***p < 0.001. (d-e) Secreted CXCL9 (d) or CXCL10 (e) by ex vivo LN slices from B6 or IFNgr −/− mice after 16-hour stimulation with 0.2, 2 or 20 ng/mL IFNα. Brown-Forsythe and Welch ANOVA comparisons against the unstimulated group **p < 0.01, ***p < 0.001, ****p < 0.0001

Figure 5:. Poly I:C vaccination drove a strongly IgG2c-polarized antibody response that did not require CXCR3.

(a) Timeline of experiment, in which C57BL/6 mice were subcutaneously vaccinated with OVA protein mixed with either alum, poly I:C (“PI”), a combination of alum and poly I:C, or PBS, and boosted after 4 weeks (dotted line). Blood was collected weekly for 6 weeks for antibody analysis. (b) OVA-specific IgG response in serum over time. Error bars indicate std dev; n=5 mice. (c-d) Quantification of OVA-specific IgG polarization at week 4 and week 6. Bars indicate mean and std dev.; n=5 mice. One-way ANOVA with multiple comparisons. *p < 0.05, **p < 0.01, ****p < 0.0001. (e) Timeline of experiment, in which C57BL/6 and CXCR3 KO mice were subcutaneously vaccinated with either alum, poly I:C or PBS with OVA protein. Blood was collected weekly for 3 weeks for antibody analysis. (f) OVA-specific IgG response in serum over time. Error bars indicate std. dev; n=4 mice. (g) Quantification of OVA-specific IgG polarization at week 3. Bars indicate mean and std dev; n=4 mice.

Fig 6:. Comparison of CXCL9 expression between different vaccine adjuvants.

(a-d) Secreted IFNα (a), IFNγ (b), CXCL9 (c), or CXCL10 (d) from LN slices 16 hours post stimulation 10 μg/mL of adjuvant. Kruskal-Wallis ANOVA. *p < 0.05, **p < 0.01, ***p < 0.001, ***p < 0.0001.

Figure 7.. Proposed pathway for CXCL9 expression in the lymph node after poly I:C vaccination.

Poly I:C (PI) vaccination initiates a type-I IFN response in the interfollicular zone. Next, IFNα signaling drives IFNγ secretion, and the IFNγ is required for CXCL9 secretion in the interfollicular zone of the lymph node. Other adjuvants, noted in grey, also induced IFNγ and CXCL9 secretion. IFZ: interfollicular zone. TCZ: T cell zone.