Identification of Three Small Molecules That Can Selectively Influence Cellular Manganese Levels in a Mouse Striatal Cell Model

Abstract

Manganese (Mn) is a biologically essential metal, critical as a cofactor for numerous enzymes such a glutamine synthetase and kinases such as ataxia-telangiectasia mutated (ATM). Similar to other essential metals such as iron and zinc, proper levels of Mn need to be achieved while simultaneously being careful to avoid excess levels of Mn that can be neurotoxic. A lifetime of occupational exposure to Mn can often lead to a Parkinsonian condition, also known as "manganism", characterized by impaired gait, muscle spasms, and tremors. Despite the importance of its regulation, the mechanisms underlying the transport and homeostasis of Mn are poorly understood. Rather than taking a protein or gene-targeted approach, our lab recently took a high-throughput-screening approach to identify 41 small molecules that could significantly increase or decrease intracellular Mn in a neuronal cell model. Here, we report characterization of these small molecules, which we refer to as the "Mn toolbox". We adapted a Fura-2-based assay for measuring Mn concentration and for measuring relative concentrations of other divalent metals: nickel, copper, cobalt, and zinc. Of these 41 small molecules, we report here the identification of three that selectively influence cellular Mn but do not influence the other divalent metals tested. The patterns of activity across divalent metals and the discovery of Mn-selective small molecules has potential pharmacological and scientific utility.

Keywords: divalent metals; manganese; metal homeostasis; metal transport; neurotoxicity; small molecules.

Figure 1

Small molecule modifiers of manganese have a tendency to lose their efficacy when exposed in culture media. A set of 37 small molecules identified to influence intracellular Mn in a high throughput screen (see reference 33), commonly referred to as the “Mn Toolbox”, were confirmed independently (A). A red box around each small molecule, identified using its assigned VU identity number (see reference 33), denotes non-significance as determined by Brown–Forsythe ANOVA and Welch’s ANOVA tests with Dunnett’s T3 multiple comparisons. The number of compounds that did not reach statistical significance increases when exposed in DMEM rather than HBSS (B). When incubated in media for 18 h (C), several small molecules had an impact on Mn when previously no statistically significant impact existed at 3 h (boxed in purple). Several small molecules changed directions (Mn-increaser to decreaser, or decreaser to increaser), which are denoted by a box in blue. Error bars represent standard deviation of three biological replicates.

Figure 2

Screening paradigm for small molecule activity against other divalent metals. A primary screen of 33 small molecules was completed, assaying each molecule for fluctuations in Fura-2 fluorescence after Co, Cu, Zn, and Ni exposure (see Supplementary Table S1). Significant deviations (by one-way ANOVA, Sidak’s multiple comparisons) were detected in 23 of the small molecules, leaving 10 that were not significantly different (and within two standard deviations from the vehicle mean; see Supplementary Table S2). These 10 were tested in the presence of lower concentrations of each metal, and it was confirmed that they could indeed influence intracellular Mn levels. Three molecules remained.

Figure 3

Cellular Fura-2 Divalent Metal Extraction Assay (CFXEA) screen of 33 small molecules from toolbox. Cells were pre-exposed with small molecule (5 μM), then the metal ((A): Cobalt, (B): Copper; (C): Nickel; (D): Zinc) was added. After 3 h, the cells were washed with PBS and lysed open with 0.1% Triton containing Fura-2. The presence of divalent metal quenches the Fura-2 signal in a concentration-dependent manner. We compared the Fura-2 signal of metals taken up by cells (Q7s) under vehicle conditions and determined whether the Fura-2 signal was significantly different when cells were co-exposed with a small molecule. Cells were exposed to a small molecule and metal in HBSS. The dotted line shows vehicle levels for a given metal concentration. A one-way ANOVA for each metal concentration tested was run, with post-hoc Dunnett’s T3 multiple comparisons test.

Figure 4

The profiles of the Mn-selective small molecules VU0025173, VU0035619, and VU0029414. Q7 cells were exposed in HBSS with DMSO vehicle or small molecules (5 μM) with a single divalent metal (Co 10 μM; Cu 10 μM; Zn 5 μM; Ni 20 μM; or Mn 125 μM) for 2 h at 37 °C. Following the exposure, cells were washed with PBS and lysed open with PBS + 0.1% Triton for CFXEA. (A) shows the %Fura-2 signal of CFXEA with means ± standard deviations. For ease of visualization a colored dashed line is shown at the %Fura-2 signal of the vehicle control for each metal. (B) shows the chemical structures of the molecules.

Figure 5

Hierarchical clustering analysis of small molecule activity for the 33 molecules screened reveals unique patterns of activity across the library. A direction (increase, decrease, no change) was assigned to each small molecule for each divalent metal based on the statistically significant results of the screening process. If no significant effect was detected, no change (marked with a “=”) was imputed. A dendrogram was then generated using a multivariate classical clustering method, a paired group algorithm, and a Euclidean similarity index. Of the 33 small molecules screened, the majority of molecules did not cluster together, demonstrating 24 distinct metal activity profiles. The assigned functional activity is listed in the table at the right of the dendrogramF and was based on the deductive logic described in the flow chart of Figure 2. Due to conflicting results in the screens, the directionality of one small molecule (VU0050661) could not be determined for Zn (marked in grey). Metal “increasers” are denoted with a “+” in a red square. Metal “decreasers” are denoted with a “-” in blue squares.