Determining potential immunomodulatory drug efficacy in sepsis using ELISpot

Abstract

This study evaluated the ability of ELISpot to identify potential immuno-modulatory drug therapies in sepsis. ELISpot was performed ex vivo on whole blood from septic patients and healthy controls. Innate and adaptive immunity were evaluated by production of TNF-α and IFN-γ, respectively. Drug efficacy was determined by their effects to modulate the both the number of cytokine-producing cells and amount of cytokine produced per cell. The corticosteroid dexamethasone was evaluated for its ability to down modulate TNF-α and IFN-γ production. The TLR7/8 agonist resiquimod (R848) and T cell stimulants IL-7 and anti-PD-1 mAb were tested for their ability to enhance immunity. LPS and resiquimod increased total TNF-α production in septic patients by 1,549% and 1,829%, respectively. Conversely, dexamethasone diminished the responses to LPS or resiquimod by 75% and 61%, respectively. IL-7, but not anti-PD-1 mAb markedly increased IFN-γ production in both healthy subjects (121%) and septic patients (82%). Dexamethasone also reduced anti-CD3/CD28 mAb stimulated IFN-γ production by 69%; while IL-7 ameliorated dexamethasone-induced suppression. IL-7 significantly enhanced lymphocyte function in over 90% of septic patients. ELISpot can reveal host immune response patterns and the effects of drugs to selectively down- or up-regulate patient immunity. Furthermore, the ability of ELISpot to detect the effect of specific immuno-modulatory drugs to independently regulate the innate and adaptive host response could enable precision-based immune drug therapies in sepsis.

Keywords: Adaptive immunity; Anti-PD-1; Checkpoint inhibitors; Corticosteroids; IL-7; Innate immunity.

Fig. 1

Adjuvant effect on IFN-γ production in conjunction with anti-CD3/CD28 mAb stimulation. Samples were stimulated with anti-CD3/28 mAb with and without either anti-PD-1 mAb (Nivolumab; NIVO) or IL-7. Anti-CD/CD28 mAb (CD3) increased IFN-γ SFU (upper panels), SS (upper-middle panels) and TWI (lower-middle panels) in samples from both Healthy donors (left, n = 46) and Septic patients (right, n = 63) as compared to unstimulated spontaneous (SPON) production. NIVO did not increase CD3 induced IFN-γ SFU, SS, or TWI, while IL-7 did increase CD3 induced IFN-γ SFU, SS, and TWI. SFU (Top panels) were normalized to the number of lymphocytes plated in the ELISpot well (Bottom panels, n = 34 healthy and n = 59 septic). Lymphocyte number per well was determined by multiplying the draw specific ALC (K/cumm) by 1000 (converting K/cumm to cells/cumm) then by the volume of blood added to the well (5µL). Pairwise statistical relationships of SFU/Lymphocyte remain largely the same as SFU. Red bars represent Median with Interquartile Ranges (IQR) *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by Friedman Test with multiple comparisons.

Fig. 2

Difference in relative IFN-γ response to IL-7 between healthy donors and septic patients. A. IL-sevenfold-changes were calculated for SFU, SS, and TWI for all samples; anti-CD3/CD28 mAb (CD3) + IL-7 values were divided by CD3 alone to yield fold-change. Samples from healthy donors displayed a higher responsiveness to IL-7 in the SFU (Left) and TWI (Right) measurements, based on fold-change, than did those from septic patients. B. Visualization of paired data points for CD3 alone or with IL-7 for samples from healthy donors (Left, n = 46) and septic patients (Right, n = 63). The bulk of samples show an increase of SFU and SS with IL-7 treatment. Red bars represent Median with IQR. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by (A.) Mann–Whitney or (B.) Wilcoxon tests.

Fig. 3

Septic patient samples produce greater IFN-γ SFU and SFU/Lymphocyte than Healthy donor samples. Comparison between Healthy donor (n = 46) and Septic patient (n = 63) whole blood samples shows septic blood samples have higher IFN-γ SFU (top) than healthy donor samples in response to anti-CD3/CD28 mAb (CD3) alone (left), CD3 + anti-PD-1 mAb (Nivolumab; NIVO) (middle), and CD3 + IL-7 (right). To account for high variability in cell counts among septic patients and between septic and healthy samples, SFU were normalized to number of lymphocytes plated in ELISpot wells (bottom, n = 34 healthy and n = 59 septic). When SFU were normalized to lymphocyte numbers, the differences between septic and healthy IFN-γ production became more apparent. As no significant differences were discerned between CD3 treated samples and CD3 + NIVO treated samples, the differences displayed in the “CD3 + NIVO” column between healthy and septic are a consequence of the greater responsiveness of septic samples to CD3 treatment. Red bars represent Median with IQR. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by Mann–Whitney test.

Fig. 4

Adjuvant effect of dexamethasone on TNF-α production in conjunction with resiquimod (R848) stimulation. Samples were stimulated with R848 with and without dexamethasone (DEX). R848 increased TNF-α SFU (upper panels), SS (upper-middle panels) and TWI (lower-middle panels) in samples from both Healthy donors (left, n = 46) and Septic patients (right, n = 63) as compared to unstimulated (spontaneous, SPON) production. DEX markedly abrogated R848 induced TNF-α production in SFU, SS, and TWI. SFU (Top panels) were normalized to the number of TNF-α producing cells (Cells) plated in the ELISpot well (Bottom panels, n = 34 healthy and n = 53 septic). The number of TNF-α producing cells per well was determined by multiplying the draw specific Absolute Neutrophil and Absolute Monocyte counts (ANC + AMC) in K/cumm by 1000 (converting K/cumm to cells/cumm) then by the volume of blood added to the well (5µL). Pairwise statistical relationships of SFU/Cell remain largely the same as SFU. Red bars represent Median with IQR. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by Friedman Test with multiple comparisons.

Fig. 5

Adjuvant effect of dexamethasone on TNF-α production in conjunction with Lipo-polysaccharide (LPS) stimulation. Samples were stimulated with LPS with and without dexamethasone (DEX). LPS increased TNF-α SFU (upper panels), SS (upper-middle panels) and TWI (lower-middle panels) in samples from both Healthy donors (left, n = 48) and Septic patients (right, n = 61) as compared to unstimulated (spontaneous, SPON) production. DEX markedly abrogated LPS induced TNF-α production in SFU, SS, and TWI. SFU (Top panels) were normalized to the number of TNF-α producing cells (Cells) plated in the ELISpot well (Bottom panels, n = 34 healthy and n = 53 septic). The number of TNF-α producing cells per well was determined by multiplying the draw specific Absolute Neutrophil and Absolute Monocyte counts (ANC + AMC) in K/cumm by 1000 (converting K/cumm to cells/cumm) then by the volume of blood added to the well (5µL). Pairwise statistical relationships of SFU/Cell remain largely the same as SFU. Red bars represent Median with IQR. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by Friedman Test with multiple comparisons.

Fig. 6

Comparison of responsiveness to dexamethasone between healthy donor and septic samples. Dexamethasone (DEX) was able to abrogate TNF-α production equally in healthy donor (n = 46) and septic patient (n = 63) samples for most parameters when looking at spontaneous TNF-α production (Left), LPS-induced TNF-α production (middle), or R848-induced TNF-α production (Right). DEX-based inhibition of LPS-induced TNF-α production was more pronounced in septic samples in the TWI measurement and R848-induced TNF-α production on the TWI measurement and on an SFU per cell basis (Bottom, n = 34 healthy and n = 53 septic) in septic patient than in healthy donor samples. Red bars represent Median with IQR. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by Mann–Whitney test.

Fig. 7

Effect of dexamethasone on IFN-γ production. IFN-γ SFU (Top), SS (Middle), and TWI (Bottom), are reduced when dexamethasone (DEX) is added to ELISpot. DEX abrogates anti-CD3/CD28 mAb (CD3) induced IFN-γ production. DEX-induced reduction of IFN-γ production is still present when co-cultured with anti-CD3/CD28 mAb and IL-7, though there is still an IL-7 effect present even with DEX co-treatment. Results were similar in healthy donor (n = 46) and septic patient (n = 63) samples. Red bars represent Median with IQR. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by Friedman Test with multiple comparisons.