A Closed Computational-Experimental Loop Identifies Metabolic Collapse at the Root of Macrophage Dysfunction due to Zinc Dyshomeostasis

Abstract

Zinc plays a crucial role in immune regulation, the oxidative stress response, and epithelial barrier integrity, yet zinc's precise role in regulating metabolic and immunological functions in myeloid cells remains poorly understood. Here, we employ a systems biology approach using constraint-based modeling to elucidate the consequences of myeloid-specific loss of ZIP8 on macrophage metabolic function and antibacterial capabilities. We demonstrate that macrophage populations in the lung of ZIP8 knockout (Zip8KO) mice exhibit widespread metabolic disruption, spanning glycolysis, butanoate metabolism, amino acid metabolism, and mitochondrial function. Specifically, Zip8KO macrophages exhibit impaired nutrient uptake and dysregulated energy metabolism, which is exacerbated following Streptococcus pneumoniae infection. Genome-scale metabolic modeling and flux analysis revealed a paradoxical pattern of metabolic suppression prior to infection, followed by overcompensation post-infection, potentially driving immune dysfunction. Consistent with these predictions Zip8KO bone marrow-derived macrophages displayed increased ATP demand and disrupted mitochondrial energetics, compromising their ability to control infection. Importantly, we identified succinate, and kynurenic acid as metabolites capable of restoring immune responses and validated their ability to enhance bacterial clearance in Zip8KO BMDMs. Together, these findings establish ZIP8 as a central regulator of immune-metabolic homeostasis and suggest potential therapeutic avenues to restore immune function in settings of zinc deficiency.

Figure 1.

Schematic Work-Flow Model. Starting from the top left and working across and down, scRNA-seq analysis was conducted on lung tissue obtained from WT and Zip8KO mouse lung tissue before and after infection with S.pneumoniae. Macrophage-specific clusters were identified and then subject to genome-scale metabolic modeling via the ftINIT pipeline. These models were analyzed to identify metabolic capacities and uncover significant metabolic vulnerabilities. This revealed that ZIP8 loss was strongly associated with mitochondrial dysfunction which is associated with disruption of multiple essential metabolic processes, all of which have the potential to impair macrophage-mediated bacterial clearance.

Figure 2.. Disrupted Pathways in Zip8KO Models Pre- and Post-Infection.

(a) Scatterplot for pathway net flux distribution in WT and Zip8KO macrophage clusters both pre-and post infection show distinct pattern shifts. Variance magnitude was determined by calculating the fold-change of each comparison relative to the smallest observed variance (** FC >= 1, *** FC >= 1.1, and **** FC >= 1.3). (b & c) Heatmap and metabolic map of metabolic pathways that vary the most across all conditions- highlighting major shifts originating from all parts of metabolism.

Figure 3:. In silico Prediction Indicates Major Dysregulation in Amino Acid Metabolism.

(a) Comparison of amino acid utilization in WT and Zip8KO macrophage clusters before and after infection. Significant utilization shift was measured by fold-change of fluxes (unit: mmol/gDW/hr) **FC >= 0.5, *** FC >=1, **** FC >= 5. (b) Heatmap showing the Flux Sum of amino acids across three organelle compartments (cytosol, mitochondria, and lysosome). The shift in amino acid utilization and in silico concentration indicates major metabolic shifts occurring in Zip8KO macrophage clusters.

Figure 4.. Glycolytic and Mitochondrial Dysfunction as Drivers of Zip8KO Dysregulation.

(a) Box plot comparison of ATP production and demand in WT and Zip8KO macrophage clusters before and after infection, (b, c) Comparison of differences in the predicted pool size of reactive species and oxygen in the cytoplasm and mitochondria of WT and Zip8KO macrophage clusters, (d) Changes in net glycolytic flux (mmol/gDW/hr) in WT versus Zip8KO macrophage clusters, (e) Schematic diagram connecting the potential for metabolic changes to influence immune function. Significant change in fluxes is calculated by Fold Change (FC) where ** FC >= 0.5, *** FC >= 3, **** FC >= 5.

Figure 5:. Validation of Predicted Metabolic Dysfunction in Zip8KO BMDMs Before and After Infection.

WT and Zip8KO BMDMs were analyzed before and after S.pneumoniae infection for: (a) Intracellular ROS as determined by MitoSox staining; (b-g) the SeaHorse XF Mito Stress test; and (h-k) the XF Glycolysis Stress Test. As predicted by in silico analysis, Zip8KO BDMSs have significantly increased ROS production, oxygen demand, extracellular acidification, and glycolytic activity compared to WT BMDMs that is further exacerbated in response to infection. (*p < 0.05; **p < 0.01; ***p < 0.001. Data representative of a minimum of 3 independent experiments; Data were analyzed using GraphPad Prism version 10.5.0. Statistical significance was determined using one-way ANOVA followed by Sidak’s corrections for multiple comparisons between groups. Values are expressed as means ± SEM).

Figure 6.. In silico Identification and Validation of Host-Derived Metabolites that Restore Bacterial Clearance.

(a) Heatmap of leading in silico-host-derived metabolites via Flux shift analysis reveals kynureic acid, succinate, n-Acetyl-Cysteine, and 4-aminobutyrate as lead candidates. Pretreatment of WT and Zip8KO BMDMs with (b) kynureic acid and (c) succinate restored bacterial clearance in Zip8KO BMDMs similar to WT controls. Pretreatment with (d) N-acetyl-cysteine and (e) 4-aminobutyrate had no beneficial effect. (* p ≤ 0.05; ** p ≤ 0.01; Data representative of a minimum of 3 independent experiments; Data were analyzed using GraphPad Prism version 10.5.0. Statistical significance was determined using an unpaired t-test. Values are expressed as means ± SEM).

Figure 7.. Disruption in Mitochondrial Function Impacts the Overall Physiology, including Signaling Cascades.

Based on in silico analysis, derangement of multiple metabolomic pathways appeared in Zip8KO macrophage clusters compared to WT clusters before and after infection. Many pathways converged directly or indirectly upon the mTOR signaling pathway, a pathway that is vital for proper phagolysomal-mediated removal of bacteria. Red denotes molecules within the Zip8KO model whose activity is predicted to be significantly decreased compared to WT models. Green denotes molecules with increased activity.