Mammalian cells consume large amount of nutrients during growth and production. However, endogenous metabolic inefficiencies often prevent cells to fully utilize nutrients to support growth and protein production. Instead, significant fraction of fed nutrients is diverted into extracellular accumulation of waste by-products and metabolites, further inhibiting proliferation and protein synthesis. In this study, an LC-MS/MS based metabolomics pipeline was used to screen Chinese hamster ovary (CHO) extracellular metabolites. Six out of eight identified inhibitory metabolites, caused by the inefficient cell metabolism, were not previously studied in CHO cells: aconitic acid, 2-hydroxyisocaproic acid, methylsuccinic acid, cytidine monophosphate, trigonelline, and n-acetyl putrescine. When supplemented back into a fed-batch culture, significant reduction in cellular growth was observed in the presence of each metabolite and all the identified metabolites were shown to impact the glycosylation of a model secreted antibody, with seven of these also reducing CHO cellular productivity (titer) and all eight inhibiting the formation of mono-galactosylated biantennary (G1F) and biantennary galactosylated (G2F) N-glycans. These inhibitory metabolites further impact the metabolism of cells, leading to a significant reduction in CHO cellular growth and specific productivity in fed-batch culture (maximum reductions of 27.2% and 40.6% respectively). In-depth pathway analysis revealed that these metabolites are produced when cells utilize major energy sources such as glucose and select amino acids (tryptophan, arginine, isoleucine, and leucine) for growth, maintenance, and protein production. Furthermore, these novel inhibitory metabolites were observed to accumulate in multiple CHO cell lines (CHO-K1 and CHO-GS) as well as HEK293 cell line. This study provides a robust and holistic methodology to incorporate global metabolomic analysis into cell culture studies for elucidation and structural verification of novel metabolites that participate in key metabolic pathways to growth, production, and post-translational modification in biopharmaceutical production.

BACKGROUND AND OBJECTIVE: Based on previous experience of sorbent-mediated ticagrelor, dabigatran, and radiocontrast agent removal, we set out in this study to test the effect of two sorbents on the removal of edoxaban, a factor Xa antagonist direct oral anticoagulant. METHODS: We circulated 100 mL of edoxaban solution during six first-pass cycles through 40-mL sorbent columns (containing either CytoSorb in three passes or Porapak Q 50-80 mesh in the remaining three passes) during experiments using human plasma and 4% bovine serum albumin solution as drug vehicles. Drug concentration was measured by liquid chromatography-tandem mass spectrometry. RESULTS: Edoxaban concentration in two experiments performed with human plasma dropped from 276.8 to 2.7 ng/mL and undetectable concentrations, respectively, with CytoSorb or Porapak Q 50-80 mesh (p = 0.0031). The average edoxaban concentration decreased from 407 ng/mL +/- 216 ng/mL to 3.3 ng/mL +/- 7 ng/mL (p = 0.017), for a removal rate of 99% across all six samples of human plasma (two samples) and bovine serum albumin solution (four samples). In four out of the six adsorbed samples, the drug concentrations were undetectable. CONCLUSION: Sorbent-mediated technology may represent a viable pathway for edoxaban removal from human plasma or albumin solution.

Postpartum depression (PPD) is associated with abnormalities in resting-state functional connectivity (RSFC) but the underlying neurochemistry is unclear. We hypothesized that peripartum GABAergic neuroactive steroids (NAS) are related to cortical GABA concentrations and RSFC in PPD as compared to healthy comparison women (HCW). To test this, we measured RSFC with fMRI and GABA+/Creatine (Cr) concentrations with proton magnetic resonance spectroscopy ((1)H MRS) in the pregenual anterior cingulate (pgACC) and occipital cortices (OCC) and quantified peripartum plasma NAS. We examined between-group differences in RSFC and the relationship between cortical GABA+/Cr concentrations with RSFC. We investigated the relationship between NAS, RSFC and cortical GABA+/Cr concentrations. Within the default mode network (DMN) an area of the dorsomedial prefrontal cortex (DMPFC) had greater connectivity with the rest of the DMN in PPD (peak voxel: MNI coordinates (2, 58, 32), p = 0.002) and was correlated to depression scores (peak HAM-D17 voxel: MNI coordinates (0, 60, 34), p = 0.008). pgACC GABA+/Cr correlated positively with DMPFC RSFC in a region spanning the right anterior/posterior insula and right temporal pole (r = +0.661, p = 0.000). OCC GABA+/Cr correlated positively with regions spanning both amygdalae (right amygdala: r = +0.522, p = 0.000; left amygdala: r = +0.651, p = 0.000) as well as superior parietal areas. Plasma allopregnanolone was higher in PPD (p = 0.03) and positively correlated with intra DMPFC connectivity (r = +0.548, p = 0.000) but not GABA+/Cr. These results provide initial evidence that PPD is associated with altered DMN connectivity; cortical GABA+/Cr concentrations are associated with postpartum RSFC and allopregnanolone is associated with postpartum intra-DMPFC connectivity.