Effect of Pd2Spermine on Mice Brain-Liver Axis Metabolism Assessed by NMR Metabolomics

Abstract

Cisplatin (cDDP)-based chemotherapy is often limited by severe deleterious effects (nephrotoxicity, hepatotoxicity and neurotoxicity). The polynuclear palladium(II) compound Pd₂Spermine (Pd₂Spm) has emerged as a potential alternative drug, with favorable pharmacokinetic/pharmacodynamic properties. This paper reports on a Nuclear Magnetic Resonance metabolomics study to (i) characterize the response of mice brain and liver to Pd₂Spm, compared to cDDP, and (ii) correlate brain-liver metabolic variations. Multivariate and correlation analysis of the spectra of polar and lipophilic brain and liver extracts from an MDA-MB-231 cell-derived mouse model revealed a stronger impact of Pd₂Spm on brain metabolome, compared to cDDP. This was expressed by changes in amino acids, inosine, cholate, pantothenate, fatty acids, phospholipids, among other compounds. Liver was less affected than brain, with cDDP inducing more metabolite changes. Results suggest that neither drug induces neuronal damage or inflammation, and that Pd₂Spm seems to lead to enhanced brain anti-inflammatory and antioxidant mechanisms, regulation of brain bioactive metabolite pools and adaptability of cell membrane characteristics. The cDDP appears to induce higher extension of liver damage and an enhanced need for liver regeneration processes. This work demonstrates the usefulness of untargeted metabolomics in evaluating drug impact on multiple organs, while confirming Pd₂Spm as a promising replacement of cDDP.

Keywords: NMR; Pd2Spm; brain; cisplatin; liver; metabolomics; mice; palladium(II); platinum(II); spermine.

Figure A1

Experimental design workflow illustrating the animals’ exposure to the metal-drugs, cDDP (2 mg/kg/day) or Pd₂Spm (5 mg/kg/day), and vehicle solution (phosphate-buffered saline, PBS), followed by brain and liver excision and extracts preparation using a mechanical maceration in liquid nitrogen (N₂) and a biphasic separation with MeOH/H₂O/CDCl₃, and, ultimately, the NMR spectra acquisition and chemometric analysis. Figure created with BioRender.com [90].

Figure 1

Average 500 MHz ¹H Nuclear Magnetic Resonance (NMR) spectra of an aqueous profile of brain from a cell-derived xenograft (CDX) mouse model of TNBC exposed to (a) vehicle solution (phosphate-buffer saline, PBS), (b) cDDP (2 mg/kg/day), and (c) Pd₂Spm (5 mg/kg/day). * Cut-off of water suppression region (δ 4.6–5.1 ppm), not considered in the multivariate analysis. The arrows identify metabolic variations found with visual inspection of spectra from each treated group compared to controls (↑, increase; ↓, decrease). Abbreviations: 3-letter code for amino acids; 3-AIBA, 3-aminoisobutyrate; 3-HBA, 3-hydroxybutyrate; 3-HIBA, 3-hydroxyisobutyrate; Ado, adenosine; ADP, adenosine diphosphate; AMP, adenosine monophosphate; ATP, adenosine triphosphate; AXP, adenosine nucleotides, AMP, ADP and ATP; BCAAs, branched-chain amino acids (ile, leu and val); CA, cholate; Cho, choline; GABA, γ-aminobutyrate; Glyc., glycerol; GPC, glycerophosphocholine; GSH, glutathione (reduced); IMP, inosine mono-phosphate; Ino, inosine; m-Ino, myo-Inositol; Lac, lactate; NAA, N-acetyl-aspartate; NAD⁺/NADH, nicotinamide adenine dinucleotide (oxidized/reduced); NADPH nicotinamide adenine dinucleotide phosphate (reduced); PA, pantothenate; PC, phosphocholine; Suc, succinate; Tau, taurine; UDP-Glc/GlcA, UDP-Glucose/Glucuronate; UDP-GlcNAc, uridine diphosphate N-acetylglucosamine; UMP, uridine monophosphate.

Figure 2

Score scatter plots of PCA (right) and PLS-DA (left) models for ¹H NMR spectra of (a) aqueous and (b) lipophilic extracts of brain from CDX mice of TNBC, considering the pairwise analysis of the treated groups, with either cDDP or Pd₂Spm, compared to controls (controls, black triangles, n = 6 or 5 for aqueous or lipophilic extracts, respectively; cDDP-treated, blue diamonds, n = 8; Pd₂Spm-treated, red circles, n = 8). Validation parameters (R² and Q²) are indicated for each PLS-DA model, with Q² values > 0.5 highlighted in bold, indicating high predictive power and, thus, robust classes separation.

Figure 3

Heatmaps illustration of the effect size (ES) values of statistically significant variations in the (a) aqueous, and (b) lipophilic extracts of the brain of TNBC xenografts exposed to cDDP or Pd₂Spm in comparison to controls. Metabolic variations are colored from blue to red, representing an increasing ES scale from negative to positive values, respectively. Abbreviations: (a) as defined in Figure 1; (b) FAs, fatty acids; PLs, phospholipids; PTC, phosphatidylcholine; PTE, phosphatidylethanolamine; PUFAs, polyunsaturated fatty acids; TG, triacylglycerols; s, singlet; d, doublet; q, quartet; br, broad signal. ^† Tentative assignment. ^‡ Partial integration of resonance peak. * p-value < 5.0 × 10⁻²; ** p-value < 1.0 × 10⁻²; *** p-value < 1.0 × 10⁻³; **** p-value < 1.0 × 10⁻⁴.

Figure 4

Heatmaps colored according to the effect size (ES) of variations in the (a) aqueous, and (b) lipophilic extracts of liver of TNBC mice. Abbreviations: NADP⁺ nicotinamide adenine dinucleotide phosphate (oxidized); TMA, trimethylamine; other abbreviations as defined in Figure 1; s, singlet; d, doublet; br, broad signal. ^† Tentative assignment. ^‡ Partial integration of resonance peak. * p-value < 5.0 × 10⁻²; ** p-value < 1.0 × 10⁻²; *** p-value < 1.0 × 10⁻³; **** p-value < 1.0 × 10⁻⁴.

Figure 5

Spearman correlations between the significant metabolic variations within each tissue, namely (a) brain and (b) liver, for controls (left), the cDDP-treated group (middle), and the Pd₂Spm-treated group (right), considering an absolute correlation coefficient of 0.90. All p-values of correlations are comprehended between 0.05 and 0.001 and are then considered statistically relevant. Bold and underlined metabolites represent new correlations regarding those found in controls (left). Abbreviations: ALC, O-acetylcarnitine. Other abbreviations as defined in Figure 1 and Figure 4.

Figure 6

Spearman correlations between significant variations between liver and brain of (a) controls and groups treated with (b) cDDP or (c) Pd₂Spm, regarding an absolute correlation coefficient of 0.80. The correlations are statistically significant, exhibiting p-values between 0.05 and 0.01. Bold and underline metabolites represent new correlations regarding those found in controls. Abbreviations as defined in Figure 1, Figure 4 and Figure 5.