Structure-activity relationship study of vitamin k derivatives yields highly potent neuroprotective agents

Abstract

Historically known for its role in blood coagulation and bone formation, vitamin K (VK) has begun to emerge as an important nutrient for brain function. While VK involvement in the brain has not been fully explored, it is well-known that oxidative stress plays a critical role in neurodegenerative diseases. It was recently reported that VK protects neurons and oligodendrocytes from oxidative injury and rescues Drosophila from mitochondrial defects associated with Parkinson's disease. In this study, we take a chemical approach to define the optimal and minimum pharmacophore responsible for the neuroprotective effects of VK. In doing so, we have developed a series of potent VK analogues with favorable drug characteristics that provide full protection at nanomolar concentrations in a well-defined model of neuronal oxidative stress. Additionally, we have characterized key cellular responses and biomarkers consistent with the compounds' ability to rescue cells from oxidative stress induced cell death.

Figure 1

A. Structures of VK₁, VK₂, and VK₃. VK₃ is a pro-vitamin, and UBIAD1 converts VK₃ or cleaved VK₁ into VK₂ in situ through geranylgeranylation. Defects in UBIAD1 have been shown to be a dominant enhancer of Parkinson’s related PINK1 mutations. B. Experimental flow. Synthetic approach and selective criteria used to generate more potent and non-toxic VK analogs.

Figure 2

HT22 cells treated for 8 hrs with 10 mM glutamate. A. Depletion of total cellular GSH occurs in HT22 cells treated with glutamate. Co-treatment with VK₂, 2j, or 2q (500 nM) does not prevent GSH depletion. Nec-1 (50 μM), Ideb (5 μM), and Trolox (25 μM) also did not prevent GSH depletion. B. Free radical accumulation measured using Rho123. Co-treatment with 250 nM 2j and 2q prevent the accumulation of free radicals in response to glutamate treatment, with VK₂ being less effective. One-way ANOVA with Bonferroni’s posttest was used to compare mean intensities. Drug treatments were all significantly less than glutamate treatment, with 2j and 2q treatments being statistically similar to control, p<.01. C. Free radical accumulation is visualized with CM-H₂DCFDA. Results are consistent with those found with Rho123.

Figure 3

A. Free radical scavenging capacity determined by monitoring the disappearance of the optical absorbance of the stable free radical DPPH. Known free radical scavengers vitamin C ( ) and Trolox (●) used as controls. VK₂ (■), 2j ( ), and 2q ( ) did not show direct antioxidant capacity. All compounds tested at 20 μM B. Expression of antioxidant response genes. Significant cellular antioxidant responses are elicited in HT22 cells after 8 hours of glutamate treatment with significant increase in HO-1 and NQO-1 gene expression. VK₂, 2q, and 2j significantly decreased HO-1 expression but did not affect NQO-1 expression. One-way ANOVA with Bonferroni’s posttest was used to compare mean levels (n = 3), p<.01.

Figure 4

A. Glutamate treatment increases superoxide generation within mitochondria. MitoTracker DR (green) stains for active mitochondria and MitoSOX (red) localizes in the mitochondria and selectively reacts with superoxide. Co-localization of MitoTracker and MitoSox supports that the superoxide is likely generated by mitochondria. B. Mitochondrial morphology. Mitochondria under normal cellular conditional exhibits a complex network morphology (teal arrows). Under glutamate injury, mitochondrial fragmentation occurs (red arrows) and VK₂ treatment maintains normal mitochondrial morphology.

Figure 5

A. Western blot analysis of HT22 cells treated with 10 mM glutamate for 16 hrs. Glutamate treatment causes a dramatic increase in the lower band of PGAM5, as well as a decrease in phosphorylation of Drp1 at residue Ser 637. VK₂, 2q, and 2j prevent PGAM5 cleavage and activation and subsequent dephosphorylation of Drp1. B. Quantification of p-Drp1 (Ser 637) relative to Drp1. Densitometric analysis confirms that there is a significant decrease in phosphorylation of Drp1 at residue Ser 637 with 10 mM glutamate treatment for 16 hrs, and the phosphorylation state is maintained by co-treatment with 500 nM VK₂ and compounds 2q or 2j at 250 and 125 nM. Co-treatment with 250 nM VK₂ was less effective. One-way ANOVA with Bonferroni’s posttest was used to compare mean levels (n = 3), p<.01.