Paclitaxel compromises fundamental endothelial cell function and induces pro-inflammatory genes
We first examined the effect of PTX in primary human umbilical vein endothelial cells (ECs). ECs were treated with PTX for 18 hours and DMSO-treated cells served as controls. PTX significantly reduced cell viability in a dose-dependent manner, reaching half-maximal inhibitory concentration (IC50) at 50 nM (Fig. 1B). The effect of PTX on EC migration was examined using a scratch assay, where a monolayer of ECs was injured following treatment with PTX for 18 hours (Figs. 1C & 1D). The scar completely closed in the control group while the scar in the PTX-treated ECs remained unaltered over 48 hours, suggesting that PTX significantly suppresses EC viability and migration.
RNA sequencing was then performed using ECs treated with PTX over a range of concentrations centered on its IC50 (5, 50, and 500 nM). A one-way analysis of variance (ANOVA) was used to identify genes that are differentially regulated across PTX concentrations, and the 1,766 genes with the greatest significance (FDR q < 0.25) were divided into four groups by hierarchical clustering (Fig. 1E), including two large clusters of genes that were uniformly up- or down-regulated across all three concentrations of PTX. The tool Enrichr (see Methods) was then used to identify pathways and processes that are significantly overrepresented (adjusted p < 0.05) within each of these two clusters (Table 1). Cluster 1 (down-regulated by PTX) was enriched in genes associated with DNA repair, whereas cluster 4 (up-regulated by PTX) was enriched in genes associated with mitotic spindle formation (in accordance with the effect of PTX on microtubule assembly) as well as many inflammatory processes, including TNF-α signaling via NFkB and IL-2 and IL-6 signaling. We focused on several genes represented in these inflammatory gene sets, including cyclin-dependent kinases regulatory subunit 2 (CKS2), CD137 (TNFRSF9), and monocyte chemoattractant protein 1 (MCP-1/CCL2), due to their association with vascular diseases. CKS2 interacts with cyclin-dependent kinases to regulate the cell cycle and is associated with atherosclerosis6. CD137 mediates adhesion molecules on ECs and mediates EC dysfunction and pro-inflammatory cytokine response7,8. MCP-1 is a well-established pro-inflammatory cytokine associated with atherothrombotic diseases41,42. All these observations raised the possibility of inflammation as a contributory factor to the effects of PTX and formed the rationale for using dexamethasone (DEX), a prototypic anti-inflammatory agent, to abrogate these effects.
Dexamethasone abrogates transcriptional perturbations modulated by PTX in ECs
Next, we validated specific transcriptional perturbations at the protein level by treating ECs with titrated concentrations of PTX or DEX, with the hypothesis that DEX will revert changes in expression induced by PTX. Treatment with concentrations of PTX as low as 5 nM increased CD137 expression by 2.5-fold (P = 0.004) (Fig. 2A and 2C). DEX treatment alone had a minimal effect on CD137 levels, except for a marginal downregulation of CD137 at 100 uM DEX (P = 0.042); however, co-treatment with DEX prevented the induction of CD137 expression by PTX (Fig. 2B and 2D). Similarly, PTX upregulated CKS2 expression in ECs in a dose-dependent manner, which was prevented by co-treatment with DEX (Fig. 2E-2H). Immunoblot analysis also confirmed the observation by RNAseq that BMF expression is down-regulated by PTX (Supplementary Fig. 1A-1D), and showed that it was upregulated in a dose-dependent manner by DEX.
Dexamethasone prevents the induction of MCP-1 by PTX in vitro
The RNAseq analysis showed that the transcription of MCP-1 (CCL2) is upregulated by up to 2-fold in ECs by PTX in a dose-dependent manner. Accordingly, treatment with 5 nM PTX doubled the level of MCP-1 protein in ECs (Fig. 3A-3D). Furthermore, co-treatment of DEX suppressed this PTX-mediated upregulation of MCP-1 in the EC lysates. MCP-1 is a secreted protein and was measured in the media of ECs using multiplex cytokine analysis. Conditioned media obtained from ECs pre-treated with 5 nM or 50 nM PTX showed a significant increase in MCP-1 levels compared to vehicle-treated EC (Fig. 3E-3G). Interestingly, PTX treatment upregulated a host of pro-inflammatory cytokines in the media of ECs, including IFN-α and IL-6, which were downregulated by DEX in a dose-dependent manner. Collectively, these results validated transcriptional perturbations induced by PTX in ECs and supported further in vivo examination of DEX.
Peri-procedural treatment with DEX suppresses PTX-induced increase in MCP-1 levels
A group of C57BL/6 mice were randomized into four groups and administered 5 mg/kg PTX, 5 mg/kg DEX, or PTX + DEX intraperitoneally (IP). DMSO (vehicle) treated mice served as controls. Animals were harvested at either 20 minutes or 12 hours (Fig. 4A) to examine the immediate and delayed effects of PTX exposure. Serum levels of PTX were high in animals treated with PTX (average ± SEM of 6.8 ± 1.93 mmol/L) or PTX + DEX (4.99 ± 1.08 mmol/L). There was no significant difference in the PTX levels between these groups, and PTX was undetectable after 12 hours. Within 20 minutes of PTX injection, no significant increase in MCP-1 was detected; however, by 12 hours, PTX-injected mice showed a 3-fold increase in the levels of MCP-1 (P < 0.001) compared to vehicle-treated mice. This effect was entirely abrogated in mice treated with PTX + DEX (P < 0.001) (P < 0.001) (Fig. 4B).
We next examined whether PTX treatment altered MCP-1 levels in the aortic ECs of mice. The aorta of mice from different groups were stained and whole slide imaging was subjected to ImageJ analysis to quantitate the expression of protein using an integrated density (a composite of image intensity and the number of pixels normalized to the area). CD31 was used as an EC marker. There were no significant changes in MCP-1 expression at 20 minutes (Supplementary Fig. 2). However, at 12 hours, PTX-exposed mice showed a 2.3-fold (P < 0.001) increase in MCP-1 expression compared to control mice (Fig. 4C-4D), which was completely suppressed in the mice co-treated with PTX + DEX (P < 0.001). Similarly, no significant changes in CD137 levels were noted within 20 minutes of PTX treatment (Supplementary Fig. 3), but at 12 hours, a significant upregulation of CD137 was noted in the aortic ECs of PTX-treated mice (P = 0.018), which was absent in mice injected with PTX + DEX (P < 0.001) (Fig. 5A and 5B).
DEX suppresses the induction of pro-thrombotic mediators by PTX
Pro-inflammatory mediators such as MCP-1 and CD137 induce atherothrombotic factors through downstream mediators such as tissue factor (TF), E-selectin and VCAM-1, converting normal vascular endothelium from an anti-coagulant to a pro-coagulant state9,10. TF is the primary trigger of the extrinsic coagulation cascade, and its upregulation increases the risk of plaque rupture and cardiovascular events. We therefore evaluated whether PTX increased the expression of these downstream pro-thrombotic mediators in ECs (Fig. 5C).
At 20 minutes after PTX injection, there was no significant change in TF levels between the PTX and control groups (Supplementary Fig. 4). However, by 12 hours, TF expression was significantly increased in the PTX group compared to control (P < 0.001), and this increase was abrogated by co-administration of PTX + DEX (P < 0.001) (Fig. 5C and 5D). A similar pattern was observed with respect to VCAM-1 and E-selectin expression (Fig. 5C, 5E and 5F). Immunoblotting was also performed using aortic lysates (n = 5 mice per group), with GAPDH serving as loading control (Fig. 6; Supplemental Fig. 7). At 20 minutes following PTX treatment, there were no alterations in the expression of TF, VCAM-1 and E-selectin. However, in aortic lysates obtained 12 hours after PTX treatment, the expression of these proteins was upregulated (~ 40–50%) compared to the control mice and these upregulations were suppressed in the PTX + DEX-treated group (Fig. 6A-6I).