Exposure-related, global alterations in innate and adaptive immunity; a consideration for re-use of non-human primates in research

Abstract

Background: Non-human primates (NHPs) play an important role in biomedical research, where they are often being re-used in multiple research studies over the course of their life-time. Researchers employ various study-specific screening criteria to reduce potential variables associated with subsequent re-use of NHPs. However, criteria set for NHP re-assignments largely neglect the impact of previous exposures on overall biology. Since the immune system is a key determinant of overall biological outcome, an altered biological state could be predicted by monitoring global changes in the immune profile. We postulate that every different exposure or a condition can generate a unique global immune profile in NHPs.

Methods: Changes in the global immune profile were evaluated in three different groups of rhesus macaques previously enrolled in dengue or malaria vaccine studies over six months after their last exposure. Naïve animals served as the baseline. Fresh blood samples were stained with various immune cell surface markers and analyzed by multi-color flow-cytometry to study immune cell dynamics in the peripheral blood. Serum cytokine profile in the pre-exposed animals were analyzed by mesoscale assay using a customized U-PLEX NHP biomarker panel of 12 cytokines/chemokines.

Results: Pre-exposed macaques showed altered dynamics in circulating cytokines and certain innate and adaptive immune cell subsets such as monocytes, HLA-DR+NKT cells, B cells and T cells. Some of these changes were transient, while some lasted for more than six months. Each group seemed to develop a global immune profile unique to their particular exposure.

Conclusion: Our data strongly suggest that re-used NHPs should be evaluated for long-term, overall immunological changes and randomly assigned to new studies to avoid study bias.

Keywords: Cytokines; Global immune profile; Infectious diseases; Innate and adaptive immunity; Non-human primates (NHPs); Re-use of NHPs in research; Trained immunity.

Figure 1. Monocyte dynamics in peripheral blood.

(A) CD45+CD14+ monocyte frequencies in naïve animals (n = 9) and rhesus macaques received DPIV/TDENV-LAV vaccination/DENV-2 challenge (n = 10) (B) Frequency and (C) cell counts of HLA-DR+CD14+ monocytes in naïve animals (n = 9) and the D/Ad-PfCA vaccinated animals (n = 5). Scatter dot plots show values of individual NHPs with mean ± standard deviation. Asterisks (*) denote significant differences between naïve and the experimental groups at 0.05 alpha level.

Figure 2. HLA-DR+NKT cells dynamics in peripheral blood.

(A) Frequency and (B) counts of HLA-DR+NKT cells in naïve animals (n = 9) and gp96-Ig-PfCA vaccinated animals (n = 5). (C) Frequency and (D) counts of HLA-DR+NKT cells in naïve animals (n = 9) and D/Ad-PfCA vaccinated animals (n = 5). Scatter dot plots show values of individual NHPs with mean ± standard deviation. Asterisks (*) denote significant differences between naïve and the experimental groups at 0.05 alpha level.

Figure 3. Circulating concentrations of MCP-1, MIP1- α and IL-12/IL-23p40 cytokines.

(A) Naïve animals (n = 9) and rhesus macaques vaccinated with DPIV/TDENV-LAV followed by DENV-2 challenge (n = 10). (B) Naïve animals (n = 9) and gp96-Ig-PfCA vaccinated animals (n = 5) (C) naïve animals (n = 9) and D/Ad-PfCA vaccinated animals (n = 5). One outlier from the naïve group was removed during IL-12/IL-23p40 analysis (n = 8). Scatter dot plots show values of individual NHPs with mean ± standard deviation. Asterisks (*) denote significant differences between naïve and the experimental groups at 0.05 alpha level.

Figure 4. B cell dynamics in peripheral blood.

(A) Frequency and (B) counts of B cells in naïve animals (n = 9) and in rhesus macaques vaccinated with DPIV/TDENV-LAV followed by DENV-2 challenge (n = 10) (C) frequency and (D) counts of B cells in naïve animals (n = 9), gp96-Ig-PfCA vaccinated animals (n = 5) and D/Ad-PfCA vaccinated animals (n = 5). Data represent mean ± standard deviation. Asterisks (*) denote significant differences between naïve and the experimental groups at each time point (0.05 alpha level).

Figure 5. T cell dynamics in peripheral blood.

(A) Cell frequency of CD3+ gated CD8+ T cells (CD3+CD8+) and double negative T cells (CD3+CD8-CD4-) in naïve group (n = 9) and in DPIV/TDENV-LAV vaccinated animals at 1 month post-DENV-2 challenge (n = 10). (B) Frequency and (C) counts of CD3+ gated CD4+ T cells (CD3+CD4+) in naïve group (n = 9) and D/Ad-PfCA vaccinated animals (n = 5). Scatter dot plots show values of individual NHPs with mean ± standard deviation. Asterisks (*) denote significant differences between naïve and the experimental groups at 0.05 alpha level.

Figure 6. Dynamics of HLA-DR expressing, activated T cells in peripheral blood.

(A) Frequency and (B) cell counts of activated T cell subsets expressing HLA-DR in naïve (n = 9) and gp96-Ig-PfCA vaccinated (n = 5) animals. (C) Frequency of activated CD3+T cells expressing HLA-DR (CD45+CD3+HLA-DR+) in naïve animals (n = 9) and D/Ad-PfCA vaccinated animals (n = 5). (D) Frequency of CD3+ gated, CD4+ T cells expressing HLA-DR (CD3+CD4+HLA-DR+) in naïve animals (n = 9) and D/Ad-PfCA vaccinated animals (n = 5). Bar graphs show mean and the standard deviation. Scatter dot plots show values of individual NHPs with mean ± standard deviation. Asterisks (*) denote significant differences between naïve and the experimental groups at 0.05 alpha level.

Figure 7. Correlation analysis and hierarchical clustering of immune cell subsets of naïve and pre-exposed animals.

Correlograms show the correlation between cell counts (cell/µL) of circulating immune cell subsets, within each of the following groups; (A) Naïve (n = 9) (B) gp96-Ig-PfCA vaccinated (n = 25) and (C) D/Ad-PfCA vaccinated (n = 25) animals. Data from naïve animals represent only a single time point. For the two vaccinated groups, samples were pooled from the 5 time points post-last exposure (n = 5 × 5); 6 days, 20 days, 2.5 months, 4 months and 6 months to capture the immune cell dynamics over 6 months. DENV group was not included in the correlation analysis, due to unavailability of cell count data for all the time points. Positive correlations are displayed in blue and negative correlations in red color. Color intensity and the size of the circle are proportional to the correlation coefficients (r). Dendograms display hierarchal relationship between immune cell subsets produced by correlation analysis. Agglomerative hierarchical clustering integrated with complete linkage method (distance—Euclidean, Pvclust) was used to build the immune cell clusters. Each leaf of the dendrogram corresponds to one immune cell subset. Closely associated cell subsets are fused into branches. Similarity of the association between two immune cell subsets decreases as height of the fusion increases along the vertical axis. Values at branches are AU p-values (approximately unbiased (AU) probability values (%) by multiscale bootstrap resampling). Clusters with AU > =95% are considered to be strongly supported by data and highlighted by the red rectangles. We used cell counts calculated based on the cell frequencies of helper T (CD45+CD4+) and cytotoxic T cells (CD45+CD8+) directly gated on parent CD45+ leukocyte population, instead of sub-gating on CD3+ T cell population for correlation analysis and hierarchical clustering (Fig. S1B).