Antigen-Specific Mammary Inflammation Depends on the Production of IL-17A and IFN-γ by Bovine CD4+ T Lymphocytes

Abstract

Intramammary infusion of the antigen used to sensitize cows by the systemic route induces a local inflammation associated with neutrophil recruitment. We hypothesize that this form of delayed type hypersensitivity, which may occur naturally during infections or could be induced intentionally by vaccination, can impact the outcome of mammary gland infections. We immunized cows with ovalbumin to identify immunological correlates of antigen-specific mammary inflammation. Intraluminal injection of ovalbumin induced a mastitis characterized by a prompt tissue reaction (increase in teat wall thickness) and an intense influx of leukocytes into milk of 10 responder cows out of 14 immunized animals. The magnitude of the local inflammatory reaction, assessed through milk leukocytosis, correlated with antibody titers, skin thickness test, and production of IL-17A and IFN-γ in a whole-blood antigen stimulation assay (WBA). The production of these two cytokines significantly correlated with the magnitude of the milk leukocytosis following the ovalbumin intramammary challenge. The IL-17A and IFN-γ production in the WBA was dependent on the presence of CD4+ cells in blood samples. In vitro stimulation of peripheral blood lymphocytes with ovalbumin followed by stimulation with PMA/ionomycin allowed the identification by flow cytometry of CD4+ T cells producing either IL-17A, IFN-γ, or both cytokines. The results indicate that the antigen-specific WBA, and specifically IL-17A and IFN-γ production by circulating CD4+ cells, can be used as a predictor of mammary hypersensitivity to protein antigens. This prompts further studies aiming at determining how Th17 and/or Th1 lymphocytes modulate the immune response of the mammary gland to infection.

Fig 1. Experimental scheme.

Two groups of 7 cows were immunized with ovalbumin (OVA) in two different adjuvant preparations. The immune response of the cows was monitored by taking blood samples on days 0, 30, 45, 60 and 75 days post-immunization. A skin test was performed 45 days post-immunization, and cows were challenged in one quarter with ovalbumin on day 60 post-immunization. The inflammatory response to the challenge was monitored for 4 days. M: Montanide ISA 61 VG.

Fig 2. Monitoring of the bodily temperature with an intravaginal sensor system.

Median values of body temperature of high responder and low-responder cows before and after the intramammary challenge with OVA at 48 h (red arrow). Minor ticks indicate 4-hour intervals.

Fig 3. Ultrasonographic assessment of the teat cistern volume following ovalbumin infusion.

(A) Graph representing the evolution of the teat cistern volume (mean values from 10 responders and 4 low responders, ± standard deviations) relative to T0 (before ovalbumin infusion) showing a significant reduction of the teat cistern volume 14 hours after ovallbumin infusion into teat cisterns of high-responder cows (*p<0.05); (B) Ultrasounds photographs representing an ovalbumin-infused teat before infusion; (1): the cistern is filled with liquid (milk) therefore weakly echogenic; (2) 14 hpi, there was a thickening and permeabilization of the wall of the teat (inflammation) reducing the cistern volume. The white dashed line represents the initial position of the cistern wall; (3) 96 hpi, the teat has regained its original appearance. The bar represents 1 cm.

Fig 4. Concentrations of cells (SCC, somatic cell count) in milk after intramammary infusion of ovalbumin.

(A) Peak SCC, i.e. highest cell concentrations reached at either 14 or 48 hours post-infusion in the infused quarter milk of cows of the Montanide (M) and the Montanide + Curdlan (M+C) groups. (B) Time-course of the influx of cells into infused quarter milk of the 14 cows distinguishing high-responder and low-responder cows. (C) Time course of cell influx into the infused quarter milk of only high-responder cows of each immunization group.

Fig 5. Concentrations of cytokines in milk samples of the 10 responder cows.

Time-course of concentration variation (median and interquartiles) of CXCL8 (A), IL-17A (B) and IFN- γ (C) in the milk of quarters infused with ovalbumin at 0 hpi.

Fig 6. Time-course of the humoral response following immunization with ovalbumin at days 0 and 30.

(A & B) Variations of antibody titers (IgG1 and IgG2) in serum samples of the responder cows (R) of the two immunization groups in the IgG1 and IgG2 subclasses. (C & D) Variation of antibody titers (IgG1 and IgG2) in serum samples of responder and low-responder cows (low-R).

Fig 7. Kinetics of increases in skin thickness following intradermal inoculation with ovalbumin of the 14 immunized cows.

(A) Increases in skinfold thickness calculated by subtracting the thickness value measured before inoculation from the values measured after injection. (B) Spearman correlation between Peak SCC and skin test values measured at 24 or 48 h post-inoculation, considering either the 14 immunized cows (All) or only the 10 responders (R). ns: not significant (p > 0.05).

Fig 8. Antigen-specific whole blood assay.

(A) Time-course of IL-17A and IFN-γ production by whole blood cultured with OVA. The blood of the 10 responder cows (taken 45 days after the first immunization) was cultured in the presence of OVA (10 μg/mL) for 1 to 4 days in 96-well microplates. For each cow triplicate wells were used for each incubation time. Results are median values and quartiles (Q1; Q3). (B) The effect of magnetic depletion of CD4+ cells on the production of IL-17A and IFN-γ in the antigen-specific whole blood assay. Results obtained by stimulating blood samples from responder cows 15 days after the booster immunization are expressed as the percentage of the production by CD4+-depleted PBMC relative to cytokine production by un processed PBMC (100%). The viability of cells was not altered by the magnetic cell separation.

Fig 9. Time-course of the IL-17A and IFN-γ production in the antigen-specific whole blood assay following immunization and correlation with Peak SCC.

Concentrations of IL-17A (A) or IFN-γ (B) after 3 days of culture with ovalbumin of blood samples taken before and after immunization at days 0 and 30 (median values and interquartiles) distinguishing the antigen-specific responses of responders and low-responder cows to the intramammary antigenic challenge. (B and C) Correlations (Spearman’s rank test) between peak SCC and IL-17A concentrations or IFN-γ concentrations yielded by the whole blood assay performed 45 days after the first immunization.

Fig 10. Production of IL-10 in the OVA-specific WBA did not correlate with mASR.

(A) IL-10 concentrations in the WBA performed at 45 days on the 10 responders (High R) and the 4 non responders (Low R). (B) Comparison of the milk leukocytosis of the 5 highest producers of IL-10 with the 5 lowest IL-10 producers among the 10 high-responder cows. Results are medians and interquartiles (Q1; Q3). Medians are significantly different at 14 hpi (Mann and Whitney, p = 0.028).

Fig 11. Persistence of reactivity to ovalbumin.

Concentrations of IL-17A and IFN-γ yielded by the whole blood assay performed at different times after immunization with ovalbumin. Results are the median values (and interquartiles) from the 8 responder cows still available 10 months post-immunization.

Fig 12. Intracellular expression of IL-17A and IFN-γ by CD4+ T lymphocytes.

PBMC were isolated one month after ovalbumin booster injection, stimulated in vitro with ovalbumin for 3 days, rested for 2 days and finally stimulated with PMA/ionomycin for 5 h with Brefeldin A for the last 3 hours. Cells were labeled for surface CD4 and intracellular IL-17A and IFN-γ. The numbers in the plots indicate the percentages of labeled cells in comparison to the isotype control. (A) Production of viable CD4+ T lymphoblasts after culture of PBMC with (OVA) or without (NS) ovalbumin. Left panels depict PBMC from a responder cow, right panels PBMC from a low-responder. (B) PBMC from two responder cows (R1 & R2) were labeled for surface CD4 and intracellular IL-17A or IFN-γ, showing CD4+ and CD4- IL-17A- and IFN-γ-producing cells. (C) labeling of CD4+ cells with anti-IL-17A and anti-IFN-γ antibodies, showing single-producing and double-producing cells. D) Double labeling of CD4+ cells from two low-responders (R3 and R4). Percentages of labeled cells are indicated in the quadrants. Results are from a representative experiment.

Fig 13. Principal component analysis of the inflammatory and immune parameters variability measured from the 14 immunized cows.

The two components retained explained respectively 53.5% (p < 0.001) a,d 16.8% (p = 0.16) of the total inertia. (A) The ellipses representing the LR (Low-responders) and R (Responders) groups are clearly separated (p = 0.004). (B) The 1.5 standard deviation-inertia ellipses of the Montanide (red ellipse and cow numbers) and the Montanide + curdlan (blue ellipse and cow numbers) overlap extensively. The between-group test was not significant (p = 0.58).