Figure 1.
The biological impact of GA is significantly more consistent than that of generic.
Among probes with variability induced by activation, 4-fold more probes had significant variation by F-test across 11 generic-activated samples from 5 batches, when compared to the number of probes with significant variation by F-test across 34 GA-activated samples from 30 batches (A). Defining tolerance as the percentage of samples with expression levels falling within the range between the maximum and minimum expression levels induced by reference standard for that probe, for any given tolerance threshold the number of probes failing to meet this this threshold is displayed for both generic and GA (B), showing that in almost all cases more probes fail to meet tolerance following induction by generic.
Figure 2.
GA induces Tregs more effectively than generic.
(A) GA induces significantly higher expression of FoxP3 than generic. FoxP3 is a key marker of Tregs, and (B) another key Treg marker Gpr83 shows a similar pattern of expression. (C) Both FoxP3 and Gpr83 are low in the same samples as indicated by scatter plot, further strengthening the case that generic fails to induce a strong Treg response in some patients. (D) As further evidence of the difference in FoxP3 induction, GSEA analysis found a significantly stronger upregulation of FoxP3 target genes in GA-activated samples than in generic-activated samples. (E) GSEA analysis also found a significant enrichment of Treg-specific genes among the genes with higher expression in GA than in generic. NS = not significant.
Figure 3.
Cell-type specific differences in the biological impact of GA and generic.
The heat map depicts relative expression of specific genes in GA-activated samples and generic activated samples. Each of the rows within the Treg section represents a gene with a high cell-type specificity scores for Tregs, while each of the rows in the macrophages and monocyte sections represents genes with high cell-type specificity scores for each of those cell types. The associated gene lists appear as supplementary information. Overall, GA induces higher expression of Treg-associated genes than generic, while generic induces higher expression of macrophage and monocyte-associated genes than GA.
Figure 4.
The generic’s impact on monocytes may differ from GA’s impact.
(A) generic induces significantly higher expression of CD14 and TLR2, as determined by a Wilcoxon rank sum test and depicted as kernel density plots, which can be likened to a smoothed histogram. (B) CD14 and TLR2 expression are both unusually high in the same (mostly generic) samples. (C) FoxP3 expression is unusually low in the sample samples in which CD14 expression is unusually high, suggesting that the generic’s different impact on monocytes may be related to its different impact on Tregs and consistent with literature suggesting that monocytes play a role in GA-induced FoxP3 expression. (D) FoxP3 expression is unusually low in the sample samples in which IL1B expression is unusually high, suggesting that the generic’s different impact on monocytes may be related to the differences between LPS-activated monocytes and T-cell contact activated monocytes, which have been described in the literature as having opposite impacts on IL1B production. (E) GSEA analysis found a significant enrichment of monocyte and macrophage-specific genes among the genes with higher expression in generic than GA. NS = not significant.
Figure 5.
Flow chart of process for comparing a branded medicine to a generic, and model of key differences between GA and generic.
(A) Overview of the methods for analyzing gene expression data to compare the immunological impact of GA to that of generic. After processing, direct differences are identified by multiple parametric methods, non-parametric methods, as well as ANOVA-based pattern analysis, and variability analysis. The genes identified by these methods are analyzed using a variety of enrichment-based methods, which result in hypotheses that are then verified through additional methods. (B) The key hypotheses emerging from our studies involve the greater heterogeneity in the generic’s biological impact compared to GA’s, and the fact that GA appears to more effectively upregulate FoxP3 expression and promote tolerance-inducing Tregs, while generic appears to upregulated myeloid lineage cells such as monocytes and macrophages which may impair tolerance. Given these findings, it is reasonable to hypothesize that GA may suppress harmful cytotoxic cells more effectively than generic, and this hypothesis warrants further investigation.