NMR-Based Assay for the Ex Vivo Determination of Soluble CD73 Activity in Serum.

Extracellular adenosine, produced through the activity of ecto-5'-nucleotidase CD73, elicits potent immunosuppressive effects, and its upregulation in tumor cells as well as in stromal and immune cell subsets within the tumor microenvironment is hypothesized to represent an important resistance mechanism to current cancer immunotherapies. Soluble CD73 (sCD73) enzymatic activity measured in patient serum or plasma at a baseline is reported to have prognostic as well as predictive relevance, with higher sCD73 activity associating with poor overall and progression-free survival in melanoma patients undergoing anti-PD1 monoclonal antibody treatment. Here, we report a novel NMR-based method that measures the ex-vivo kinetics of sCD73 activity with high specificity and reproducibility and is suitable for future high-throughput implementation. Unlike the existing assays, this method has the advantage of directly and simultaneously measuring the concentration of both the CD73 substrate and product with minimal sample manipulation or special reagents. We establish the utility of the assay for measuring the activity of sCD73 in human serum and show a strong linear correlation between sCD73 protein levels and enzyme activity. Together with our finding that sCD73 appears to be the predominant activity for the generation of adenosine in human blood, our results demonstrate a link between activity and protein levels that will inform future clinical application.

Discovery and Validation of Pyridoxic Acid and Homovanillic Acid as Novel Endogenous Plasma Biomarkers of Organic Anion Transporter (OAT) 1 and OAT3 in Cynomolgus Monkeys.

Perturbation of organic anion transporter (OAT) 1- and OAT3-mediated transport can alter the exposure, efficacy, and safety of drugs. Although there have been reports of the endogenous biomarkers for OAT1/3, none of these have all of the characteristics required for a clinical useful biomarker. Cynomolgus monkeys were treated with intravenous probenecid (PROB) at a dose of 40 mg/kg in this study. As expected, PROB increased the area under the plasma concentration-time curve (AUC) of coadministered furosemide, a known substrate of OAT1 and OAT3, by 4.1-fold, consistent with the values reported in humans (3.1- to 3.7-fold). Of the 233 plasma metabolites analyzed using a liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based metabolomics method, 29 metabolites, including pyridoxic acid (PDA) and homovanillic acid (HVA), were significantly increased after either 1 or 3 hours in plasma from the monkeys pretreated with PROB compared with the treated animals. The plasma of animals was then subjected to targeted LC-MS/MS analysis, which confirmed that the PDA and HVA AUCs increased by approximately 2- to 3-fold by PROB pretreatments. PROB also increased the plasma concentrations of hexadecanedioic acid (HDA) and tetradecanedioic acid (TDA), although the increases were not statistically significant. Moreover, transporter profiling assessed using stable cell lines constitutively expressing transporters demonstrated that PDA and HVA are substrates for human OAT1, OAT3, OAT2 (HVA), and OAT4 (PDA), but not OCT2, MATE1, MATE2K, OATP1B1, OATP1B3, and sodium taurocholate cotransporting polypeptide. Collectively, these findings suggest that PDA and HVA might serve as blood-based endogenous probes of cynomolgus monkey OAT1 and OAT3, and investigation of PDA and HVA as circulating endogenous biomarkers of human OAT1 and OAT3 function is warranted.

NMR in drug design.

The use of NMR as a tool to determine 3 dimensional protein solution structures, once a darling of the pharmaceutical industry, has largely given way to study of the interaction of prospective drugs with macromolecular targets. Many of these approaches involve ligand-centered studies, which have the advantage of speed and efficiency, but there are also many approaches that take directly from our learnings in macromolecular NMR and provide greater structural detail yet are still optimized for rapid turn-around of information. In the evolution of NMR in the pharmaceutical industry, the unique strengths of NMR to provide dynamic and atomic level information continue to be exploited to discover and design new drugs. Numerous methods have been developed over the past two decades that fall into the categories of fragment-based pre-lead discovery, ligand binding studies and qualitative structural screening.

The utility of stable isotope labeled (SIL) analogues in the bioanalysis of endogenous compounds by LC-MS applied to the study of bile acids in a metabolomics assay.

The growing field of biomarker bioanalysis by liquid chromatography mass spectrometry (LC-MS) is challenged with the selection of suitable matrices to construct relevant and valid calibration curves resulting in not only precise but also accurate data. Because surrogate matrices are often employed with the associated concerns about the accuracy of the obtained data, here we present an assay using surrogate analytes in naive biological matrices. This approach is illustrated with the analysis of endogenous bile acids (e-BAs) in serum and plasma using stable isotope-labeled (SIL) analogues as calibration standards to address the matrix concerns. Several deuterated BAs (d-BAs) were used as standards representing respectively grouped e-BAs with structural similarity allowing for the simultaneous bioanalysis of 16 e-BA. The utility of this LC-MS assay employing d-BAs is demonstrated with the analysis of samples resultant of a controlled metabolomics study where a cohort of rats was fed/fasted to investigate the change of e-BAs dependent on food consumption and fasting time.

Branched chain amino acid metabolism profiles in progressive human nonalcoholic fatty liver disease.

Nonalcoholic fatty liver disease (NAFLD) is a globally widespread disease of increasing clinical significance. The pathological progression of the disease from simple steatosis to nonalcoholic steatohepatitis (NASH) has been well defined, however, the contribution of altered branched chain amino acid metabolomic profiles to the progression of NAFLD is not known. The three BCAAs: leucine, isoleucine and valine are known to mediate activation of several important hepatic metabolic signaling pathways ranging from insulin signaling to glucose regulation. The purpose of this study is to profile changes in hepatic BCAA metabolite levels with transcriptomic changes in the progression of human NAFLD to discover novel mechanisms of disease progression. Metabolomic and transcriptomic data sets representing the spectrum of human NAFLD (normal, steatosis, NASH fatty, and NASH not fatty livers) were utilized for this study. During the transition from steatosis to NASH, increases in the levels of leucine (127% of normal), isoleucine (139%), and valine (147%) were observed. Carnitine metabolites also exhibited significantly elevated profiles in NASH fatty and NASH not fatty samples and included propionyl, hexanoyl, lauryl, acetyl and butyryl carnitine. Amino acid and BCAA metabolism gene sets were significantly enriched among downregulated genes during NASH. These cumulative alterations in BCAA metabolite and amino acid metabolism gene profiles represent adaptive physiological responses to disease-induced hepatic stress in NASH patients.

Metabolomics in the pharmaceutical industry.

Metabolomics has roots in the pharmaceutical industry that go back nearly three decades. Initially focused on applications in toxicology and disease pathology, more recent academic and commercial efforts have helped advance metabolomics as a tool to reveal the molecular basis of biological processes and pharmacological responses to drugs. This article will discuss areas where metabolomic technologies and applications are poised to have the greatest impact in the discovery and development of pharmaceuticals.

1H NMR-based lipidomics of rodent fur: species-specific lipid profiles and SCD1 inhibitor-related dermal toxicity.

A method is described that allows noninvasive identification and quantitative assessment of lipid classes present in sebaceous excretions in rodents. The method relies on direct high-field proton NMR analysis of common group lipid protons in deuterated organic solvent extracts of fur. Extracts from as little as 15 mg of fur from rat, mouse, and hamster provided acceptable results on a 600 MHz NMR equipped with a cryogenically cooled proton-observe probe. In rats, sex- and age-related differences in lipid composition are larger than differences in fur collected from various body regions within an individual and much larger than interanimal differences in age- and sex-matched specimens. The utility of this method to noninvasively monitor drug-induced sebaceous gland atrophy in rodents is demonstrated in rats dosed with a stearoyl-CoA desaturase 1 (SCD1) inhibitor. In this model, a 35% reduction in sebum lipids, extracted from fur, was observed. Finally, structural elucidation of cholesta-7,24-dien-3ß-ol ester as the most prominent, previously unidentified sebum sterol ester in male Syrian hamsters is described. The utility of this method for drug and cosmetic safety and efficacy assessment is discussed.

Characterization of hepatocellular carcinoma related genes and metabolites in human nonalcoholic fatty liver disease.

BACKGROUND: The worldwide prevalences of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) are estimated to range from 30 to 40 % and 5-17 %, respectively. Hepatocellular carcinoma (HCC) is primarily caused by hepatitis B infection, but retrospective data suggest that 4-29 % of NASH cases will progress to HCC. Currently the connection between NASH and HCC is unclear. AIMS: The purpose of this study was to identify changes in expression of HCC-related genes and metabolite profiles in NAFLD progression. METHODS: Transcriptomic and metabolomic datasets from human liver tissue representing NAFLD progression (normal, steatosis, NASH) were utilized and compared to published data for HCC. RESULTS: Genes involved in Wnt signaling were downregulated in NASH but have been reported to be upregulated in HCC. Extracellular matrix/angiogenesis genes were upregulated in NASH, similar to reports in HCC. Iron homeostasis is known to be perturbed in HCC and we observed downregulation of genes in this pathway. In the metabolomics analysis of hepatic NAFLD samples, several changes were opposite to what has been reported in plasma of HCC patients (lysine, phenylalanine, citrulline, creatine, creatinine, glycodeoxycholic acid, inosine, and alpha-ketoglutarate). In contrast, multiple acyl-lyso-phosphatidylcholine metabolites were downregulated in NASH livers, consistent with observations in HCC patient plasma. CONCLUSIONS: These data indicate an overlap in the pathogenesis of NAFLD and HCC where several classes of HCC related genes and metabolites are altered in NAFLD. Importantly, Wnt signaling and several metabolites are different, thus implicating these genes and metabolites as mediators in the transition from NASH to HCC.

Decreased hepatotoxic bile acid composition and altered synthesis in progressive human nonalcoholic fatty liver disease.

Bile acids (BAs) have many physiological roles and exhibit both toxic and protective influences within the liver. Alterations in the BA profile may be the result of disease induced liver injury. Nonalcoholic fatty liver disease (NAFLD) is a prevalent form of chronic liver disease characterized by the pathophysiological progression from simple steatosis to nonalcoholic steatohepatitis (NASH). The hypothesis of this study is that the 'classical' (neutral) and 'alternative' (acidic) BA synthesis pathways are altered together with hepatic BA composition during progression of human NAFLD. This study employed the use of transcriptomic and metabolomic assays to study the hepatic toxicologic BA profile in progressive human NAFLD. Individual human liver samples diagnosed as normal, steatosis, and NASH were utilized in the assays. The transcriptomic analysis of 70 BA genes revealed an enrichment of downregulated BA metabolism and transcription factor/receptor genes in livers diagnosed as NASH. Increased mRNA expression of BAAT and CYP7B1 was observed in contrast to decreased CYP8B1 expression in NASH samples. The BA metabolomic profile of NASH livers exhibited an increase in taurine together with elevated levels of conjugated BA species, taurocholic acid (TCA) and taurodeoxycholic acid (TDCA). Conversely, cholic acid (CA) and glycodeoxycholic acid (GDCA) were decreased in NASH liver. These findings reveal a potential shift toward the alternative pathway of BA synthesis during NASH, mediated by increased mRNA and protein expression of CYP7B1. Overall, the transcriptomic changes of BA synthesis pathway enzymes together with altered hepatic BA composition signify an attempt by the liver to reduce hepatotoxicity during disease progression to NASH.

Intra- and interlaboratory reproducibility of ultra performance liquid chromatography-time-of-flight mass spectrometry for urinary metabolic profiling.

Liquid chromatography coupled to mass spectrometry (LC-MS) is a major platform in metabolic profiling but has not yet been comprehensively assessed as to its repeatability and reproducibility across multiple spectrometers and laboratories. Here we report results of a large interlaboratory reproducibility study of ultra performance (UP) LC-MS of human urine. A total of 14 stable isotope labeled standard compounds were spiked into a pooled human urine sample, which was subject to a 2- to 16-fold dilution series and run by UPLC coupled to time-of-flight MS at three different laboratories all using the same platform. In each lab, identical samples were run in two phases, separated by at least 1 week, to assess between-day reproducibility. Overall, platform reproducibility was good with median mass accuracies below 12 ppm, median retention time drifts of less than 0.73 s and coefficients of variation of intensity of less than 18% across laboratories and ionization modes. We found that the intensity response was highly linear within each run, with a median R(2) of 0.95 and 0.93 in positive and negative ionization modes. Between-day reproducibility was also high with a mean R(2) of 0.93 for a linear relationship between the intensities of ions recorded in the two phases across the laboratories and modes. Most importantly, between-lab reproducibility was excellent with median R(2) values of 0.96 and 0.98 for positive and negative ionization modes, respectively, across all pairs of laboratories. Interestingly, the three laboratories observed different amounts of adduct formation, but this did not appear to be related to reproducibility observed in each laboratory. These studies show that UPLC-MS is fit for the purpose of targeted urinary metabolite analysis but that care must be taken to optimize laboratory systems for quantitative detection due to variable adduct formation over many compound classes.

NMR-based metabolomics of mammalian cell and tissue cultures.

NMR spectroscopy was used to evaluate growth media and the cellular metabolome in two systems of interest to biomedical research. The first of these was a Chinese hamster ovary cell line engineered to express a recombinant protein. Here, NMR spectroscopy and a quantum mechanical total line shape analysis were utilized to quantify 30 metabolites such as amino acids, Krebs cycle intermediates, activated sugars, cofactors, and others in both media and cell extracts. The impact of bioreactor scale and addition of anti-apoptotic agents to the media on the extracellular and intracellular metabolome indicated changes in metabolic pathways of energy utilization. These results shed light into culture parameters that can be manipulated to optimize growth and protein production. Second, metabolomic analysis was performed on the superfusion media in a common model used for drug metabolism and toxicology studies, in vitro liver slices. In this study, it is demonstrated that two of the 48 standard media components, choline and histidine are depleted at a faster rate than many other nutrients. Augmenting the starting media with extra choline and histidine improves the long-term liver slice viability as measured by higher tissues levels of lactate dehydrogenase (LDH), glutathione and ATP, as well as lower LDH levels in the media at time points out to 94 h after initiation of incubation. In both models, media components and cellular metabolites are measured over time and correlated with currently accepted endpoint measures.

Metabolomic and transcriptomic changes induced by overnight (16 h) fasting in male and female Sprague-Dawley rats.

The overnight (16-h) fast is one of the most common experimental manipulations performed in rodent studies. Despite its ubiquitous employment, a comprehensive evaluation of metabolomic and transcriptomic sequelae of fasting in conjunction with routine clinical pathology evaluation has not been undertaken. This study assessed the impact of a 16-h fast on urine and serum metabolic profiles, transcript profiles of liver, psoas muscle, and jejunum as well as on routine laboratory clinical pathology parameters. Fasting rats had an approximate 12% relative weight decrease compared to ad libitum fed animals, and urine volume was significantly increased. Fasting had no effect on hematology parameters, though several changes were evident in serum and urine clinical chemistry data. In general, metabolic changes in biofluids were modest in magnitude but broad in extent, with a majority of measured urinary metabolites and from 1/3 to 1/2 of monitored serum metabolites significantly affected. Increases in fatty acids and bile acids dominated the upregulated metabolites. Downregulated serum metabolites were dominated by diet-derived and/or gut-microflora derived metabolites. Major transcriptional changes included genes with roles in fatty acid, carbohydrate, cholesterol, and bile acid metabolism indicating decreased activity in glycolytic pathways and a shift toward increased utilization of fatty acids. Typically, several genes within these metabolic pathways, including key rate limiting genes, changed simultaneously, and those changes were frequently correlative to changes in clinical pathology parameters or metabolomic data. Importantly, up- or down-regulation of a variety of cytochrome P450s, transporters, and transferases was evident. Taken together, these data indicate profound consequences of fasting on systemic biochemistry and raise the potential for unanticipated interactions, particularly when metabolomic or transcriptomic data are primary end points.

Advances in mass spectrometry applied to pharmaceutical metabolomics.

Metabolomics, also referred to in the literature as metabonomics, is a relatively new systems biology tool for drug discovery and development and is increasingly being used to obtain a detailed picture of a drug's effect on the body. Metabolomics is the qualitative assessment and relative or absolute quantitative measurement of the endogenous metabolome, defined as the complement of all native small molecules (metabolites less than 1,500 Da). A metabolomics study frequently involves the comparative analysis of sample sets from a normal state and a perturbed state, where the perturbation can be of any nature, such as genetic knockout, administration of a drug, or change in diet or lifestyle. Advances in mass spectrometry (MS) technologies including direct introduction or in-line chromatographic separation modes, ionization techniques, mass analyzers, and detection methods have provided powerful tools to assess the molecular changes in the metabolome. This review focuses on advances in MS pertaining to the analytical data generation for the main metabolomics methods, namely, fingerprinting, nontargeted, and targeted approaches, as they are applied to pharmaceutical drug discovery and development.

Ultra performance liquid chromatography-mass spectrometry profiling of bile acid metabolites in biofluids: application to experimental toxicology studies.

We have developed an ultra performance liquid chromatography-mass spectrometry (UPLC-MS(E)) method to measure bile acids (BAs) reproducibly and reliably in biological fluids and have applied this approach for indications of hepatic damage in experimental toxicity studies. BAs were extracted from serum using methanol, and an Acquity HSS column coupled to a Q-ToF mass spectrometer was used to separate and identify 25 individual BAs within 5 min. Employing a gradient elution of water and acetonitrile over 21 min enabled the detection of a wide range of endogenous metabolites, including the BAs. The utilization of MS(E) allowed for characteristic fragmentation information to be obtained in a single analytical run, easily distinguishing glycine and taurine BA conjugates. The proportions of these conjugates were altered markedly in an experimental toxic state induced by galactosamine exposure in rats. Principally, taurine-conjugated BAs were greatly elevated ( approximately 50-fold from control levels), and were highly correlated to liver damage severity as assessed by histopathological scoring (r = 0.83), indicating their potential as a sensitive measure of hepatic damage. The UPLC-MS approach to BA analysis offers a sensitive and reproducible tool that will be of great value in exploring both markers and mechanisms of hepatotoxicity and can readily be extended to clinical studies of liver damage.

Mechanistic aspects and novel biomarkers of responder and non-responder phenotypes in galactosamine-induced hepatitis.

The amino sugar galactosamine (galN) induces alterations in the hepatic uridine nucleotide pool and has been widely used as a model of human viral hepatitis. Histopathological and clinical chemistry analyses of a cohort of rats following administration of galN revealed extreme interindividual variability in the extent of the toxic response which enabled classification of 'responder' and 'non-responder' phenotypes. An integrative metabolic profiling approach was applied to characterize biomarkers of exposure to galN in urine, serum, feces and liver from responders and non-responders. The presence of N-acetylglucosamine and galN in the urine correlated with the occurrence and extent of toxic response. Conversely, the novel identification of galN-pyrazines in the feces of non-responders and their virtual absence in the feces of responders suggests an alternative means of distribution and metabolism of galN in non-responders. The absence of the UDP-hexosamines in the liver of non-responders further supports differential metabolism of galN and suggests an ability of non-responders to avoid UDP-glucose depletion. An observed disturbance of gut microbial derived metabolites in the urine and feces of non-responders may suggest a role of the microflora in reducing the effective dose of galN. This systems level metabonomic approach has provided new mechanistic insights into differential response to galN and is widely applicable to the study of interindividual variation in metabolism for any xenobiotic intervention.

Normalization strategies for metabonomic analysis of urine samples.

Unlike plasma and most biological fluids which have solute concentrations that are tightly controlled, urine volume can vary widely based upon water consumption and other physiological factors. As a result, the concentrations of endogenous metabolites in urine vary widely and normalizing for these effects is necessary. Normalization approaches that utilized urine volume, osmolality, creatinine concentration, and components that are common to all samples ("total useful MS signal") were compared in order to determine which strategies could be successfully used to differentiate between dose groups based upon the complete endogenous metabolite profile. Variability observed in LC/MS results obtained from targeted and non-targeted metabonomic analyses was highly dependent on the strategy used for normalization. We therefore recommend the use of two different normalization techniques in order to facilitate detection of statistically significant changes in the endogenous metabolite profile when working with urine samples.

Chemical shift calibration of 1H MAS NMR liver tissue spectra exemplified using a study of glycine protection of galactosamine toxicity.

High-resolution (1)H magic angle spinning (MAS) NMR spectroscopy is a useful tool for analysing intact tissues as a component of metabonomic studies. The effect of referencing MAS NMR spectra to the chemical shifts of glucose or to that or trimethylsilylpropionic acid on the resultant multivariate statistical models have been investigated. It is shown that referencing to known chemical shifts of either alpha-glucose or beta-glucose in (1)H MAS NMR-based metabolic data of intact liver tissues is preferred. This has been exemplified in studies of galactosamine toxicity in the rat where co-administration of glycine ameliorates the toxic response. This approach leads to better aligned sets of spectra and reduces the inter-sample variability in multivariate statistical models. If glucose is not present in the tissue under study, then a number of alternative internal reference chemical shifts are presented. Finally, the chemical shift difference between that of the anomeric H1 proton of alpha-glucose and residual water is confirmed as a suitable internal temperature calibration method.

Pharmacokinetics of 40 kDa PEG in rodents using high-field NMR spectroscopy.

Conjugation of macromolecular drugs to polyethylene glycol (PEG) improves their therapeutic potential by reducing their rate of degradation, thereby extending the drugs half life. As a substantial component of the drug, it is necessary to measure the pharmacokinetic (PK) characteristics of PEG in vivo. A quantitative NMR-based method was developed and successfully applied to measuring double-branched polyethylene glycol 40 kDa (PEG40) in serum samples, enabling determination of PK parameters of PEG40 in preclinical species. NMR is ideal for measuring such polymers because a single, sharp peak is obtained for all the equivalent methylene protons; this amplifies the signal and makes the method insensitive to polymeric heterogeneity. High field NMR (600 MHz) with proton-observe cryoprobe technology allowed for analysis of samples in 300 nM range. Mice received 50 mg/kg of PEG40 intravenously (IV) and serum samples were collected at regular intervals for up to 72 h after dosing. The serum samples were analyzed for PEG40 using the NMR method and PK parameters were calculated using non-compartmental analysis. The volume of distribution was determined to be 0.17 L/kg for IV dosing, indicating limited distribution to interstitial space. A low clearance and observed half life of 18 h is consistent with previous reports on the PK properties of a variety of different PEG molecules ranging from 3 kDa to 190 kDa using 125I-labeled PEG in mice. The current NMR technique is easy to implement and does not require labeling of the PEG. Additionally, this is the first report, to our knowledge, of NMR spectroscopy application to PK profiling in serum.

Metabolomic profiling distinction of human nonalcoholic fatty liver disease progression from a common rat model.

OBJECTIVE: Characteristic pathological changes define the progression of steatosis to nonalcoholic steatohepatitis (NASH) and are correlated to metabolic pathways. A common rodent model of NASH is the methionine and choline deficient (MCD) diet. The objective of this study was to perform full metabolomic analyses on liver samples to determine which pathways are altered most pronouncedly in this condition in humans, and to compare these changes to rodent models of nonalcoholic fatty liver disease (NAFLD). METHODS: A principal component analysis for all 91 metabolites measured indicated that metabolome perturbation is greater and less varied for humans than for rodents. RESULTS: Metabolome changes in human and rat NAFLD were greatest for the amino acid and bile acid metabolite families (e.g., asparagine, citrulline, gamma-aminobutyric acid, lysine); although, in many cases, the trends were reversed when compared between species (cholic acid, betaine). CONCLUSIONS: Overall, these results indicate that metabolites of specific pathways may be useful biomarkers for NASH progression, although these markers may not correspond to rodent NASH models. The MCD model may be useful when studying certain end points of NASH; however, the metabolomics results indicate important differences between humans and rodents in the biochemical pathogenesis of the disease.

Understanding and Controlling Sialylation in a CHO Fc-Fusion Process.

A Chinese hamster ovary (CHO) bioprocess, where the product is a sialylated Fc-fusion protein, was operated at pilot and manufacturing scale and significant variation of sialylation level was observed. In order to more tightly control glycosylation profiles, we sought to identify the cause of variability. Untargeted metabolomics and transcriptomics methods were applied to select samples from the large scale runs. Lower sialylation was correlated with elevated mannose levels, a shift in glucose metabolism, and increased oxidative stress response. Using a 5-L scale model operated with a reduced dissolved oxygen set point, we were able to reproduce the phenotypic profiles observed at manufacturing scale including lower sialylation, higher lactate and lower ammonia levels. Targeted transcriptomics and metabolomics confirmed that reduced oxygen levels resulted in increased mannose levels, a shift towards glycolysis, and increased oxidative stress response similar to the manufacturing scale. Finally, we propose a biological mechanism linking large scale operation and sialylation variation. Oxidative stress results from gas transfer limitations at large scale and the presence of oxygen dead-zones inducing upregulation of glycolysis and mannose biosynthesis, and downregulation of hexosamine biosynthesis and acetyl-CoA formation. The lower flux through the hexosamine pathway and reduced intracellular pools of acetyl-CoA led to reduced formation of N-acetylglucosamine and N-acetylneuraminic acid, both key building blocks of N-glycan structures. This study reports for the first time a link between oxidative stress and mammalian protein sialyation. In this study, process, analytical, metabolomic, and transcriptomic data at manufacturing, pilot, and laboratory scales were taken together to develop a systems level understanding of the process and identify oxygen limitation as the root cause of glycosylation variability.