Genome-scale reconstructions of the mammalian secretory pathway predict metabolic costs and limitations of protein secretion

Machine learning-based investigation of the cancer protein secretory pathway

Deregulation of the protein secretory pathway (PSP) is linked to many hallmarks of cancer, such as promoting tissue invasion and modulating cell-cell signaling. The collection of secreted proteins processed by the PSP, known as the secretome, is often studied due to its potential as a reservoir of tumor biomarkers. However, there has been less focus on the protein components of the secretory machinery itself. We therefore investigated the expression changes in secretory pathway components across many different cancer types. Specifically, we implemented a dual approach involving differential expression analysis and machine learning to identify PSP genes whose expression was associated with key tumor characteristics: mutation of p53, cancer status, and tumor stage. Eight different machine learning algorithms were included in the analysis to enable comparison between methods and to focus on signals that were robust to algorithm type. The machine learning approach was validated by identifying PSP genes known to be regulated by p53, and even outperformed the differential expression analysis approach. Among the different analysis methods and cancer types, the kinesin family members KIF20A and KIF23 were consistently among the top genes associated with malignant transformation or tumor stage. However, unlike most cancer types which exhibited elevated KIF20A expression that remained relatively constant across tumor stages, renal carcinomas displayed a more gradual increase that continued with increasing disease severity. Collectively, our study demonstrates the complementary nature of a combined differential expression and machine learning approach for analyzing gene expression data, and highlights key PSP components relevant to features of tumor pathophysiology that may constitute potential therapeutic targets.

The secretory pathway is a series of intracellular compartments and enzymes that process and export proteins from the cell to its surrounding environment. Dysfunction of the secretory pathway is associated with many diseases, including cancer, and therefore constitutes a potential target for novel therapeutic strategies. The large number of interacting components that comprise the secretory pathway pose a challenge when attempting to identify where the dysfunction originates or how to restore healthy function. To improve our understanding of how the secretory pathway is changed within tumors, we used gene expression data from normal tissue and tumor samples from thousands of individuals which included many different types of cancers. The data was analyzed using different machine learning algorithms which we trained to predict sample characteristics, such as disease severity. This training quantified the relative degree to which each gene was associated with the tumor characteristic, allowing us to predict which secretory pathway components were important for processes such as tumor progression—both within specific cancer types and across many different cancer types. The machine learning-based approach demonstrated excellent performance compared to traditional gene expression analysis methods and identified several secretory pathway components with strong evidence of involvement in tumor development.

Big-Data Glycomics: Tools to Connect Glycan Biosynthesis to Extracellular Communication

Characteristically, cells must sense and respond to environmental cues. Despite the importance of cell-cell communication, our understanding remains limited and often lacks glycans. Glycans decorate proteins and cell membranes at the cell-environment interface, and modulate intercellular communication, from development to pathogenesis. Providing further challenges, glycan biosynthesis and cellular behavior are co-regulating systems. Here, we discuss how glycosylation contributes to extracellular responses and signaling. We further organize approaches for disentangling the roles of glycans in multicellular interactions using newly available datasets and tools, including glycan biosynthesis models, omics datasets, and systems-level analyses. Thus, emerging tools in big data analytics and systems biology are facilitating novel insights on glycans and their relationship with multicellular behavior.

Systematic evaluation of parameters for genome-scale metabolic models of cultured mammalian cells

Genome-scale metabolic models describe cellular metabolism with mechanistic detail. Given their high complexity, such models need to be parameterized correctly to yield accurate predictions and avoid overfitting. Effective parameterization has been well-studied for microbial models, but it remains unclear for higher eukaryotes, including mammalian cells. To address this, we enumerated model parameters that describe key features of cultured mammalian cells - including cellular composition, bioprocess performance metrics, mammalian-specific pathways, and biological assumptions behind model formulation approaches. We tested these parameters by building thousands of metabolic models and evaluating their ability to predict the growth rates of a panel of phenotypically diverse Chinese Hamster Ovary cell clones. We found the following considerations to be most critical for accurate parameterization: (1) cells limit metabolic activity to maintain homeostasis, (2) cell morphology and viability change dynamically during a growth curve, and (3) cellular biomass has a particular macromolecular composition. Depending on parameterization, models predicted different metabolic phenotypes, including contrasting mechanisms of nutrient utilization and energy generation, leading to varying accuracies of growth rate predictions. Notably, accurate parameter values broadly agreed with experimental measurements. These insights will guide future investigations of mammalian metabolism.

Tracing production instability in a clonally derived CHO cell line using single-cell transcriptomics

A variety of mechanisms including transcriptional silencing, gene copy loss, and increased susceptibility to cellular stress have been associated with a sudden or gradual loss of monoclonal antibody (mAb) production in Chinese hamster ovary (CHO) cell lines. In this study, we utilized single-cell RNA-seq (scRNA-seq) to study a clonally derived CHO cell line that underwent production instability leading to a dramatic reduction of the levels of mAb produced. From the scRNA-seq data, we identified subclusters associated with variations in the mAb transgenes and observed that heavy chain gene expression was significantly lower than that of the light chain across the population. Using trajectory inference, the evolution of the cell line was reconstructed and was found to correlate with a reduction in heavy and light chain gene expression. Genes encoding for proteins involved in the response to oxidative stress and apoptosis were found to increase in expression as cells progressed along the trajectory. Future studies of CHO cell lines using this technology have the potential to dramatically enhance our understanding of the characteristics underpinning efficient manufacturing performance as well as product quality.

In situ detection of protein interactions for recombinant therapeutic enzymes

Despite their therapeutic potential, many protein drugs remain inaccessible to patients since they are difficult to secrete. Each recombinant protein has unique physicochemical properties and requires different machinery for proper folding, assembly, and posttranslational modifications (PTMs). Here we aimed to identify the machinery supporting recombinant protein secretion by measuring the protein-protein interaction (PPI) networks of four different recombinant proteins (SERPINA1, SERPINC1, SERPING1, and SeAP) with various PTMs and structural motifs using the proximity-dependent biotin identification (BioID) method. We identified PPIs associated with specific features of the secreted proteins using a Bayesian statistical model and found proteins involved in protein folding, disulfide bond formation, and N-glycosylation were positively correlated with the corresponding features of the four model proteins. Among others, oxidative folding enzymes showed the strongest association with disulfide bond formation, supporting their critical roles in proper folding and maintaining the ER stability. Knockdown of disulfide-isomerase PDIA4, a measured interactor with significance for SERPINC1 but not SERPINA1, led to the decreased secretion of SERPINC1, which relies on its extensive disulfide bonds, compared to SERPINA1, which has no disulfide bonds. Proximity-dependent labeling successfully identified the transient interactions supporting synthesis of secreted recombinant proteins and refined our understanding of key molecular mechanisms of the secretory pathway during recombinant protein production.

Translational landscape and protein biogenesis demands of the early secretory pathway in Komagataella phaffii

Background

Eukaryotes use distinct networks of biogenesis factors to synthesize, fold, monitor, traffic, and secrete proteins. During heterologous expression, saturation of any of these networks may bottleneck titer and yield. To understand the flux through various routes into the early secretory pathway, we quantified the global and membrane-associated translatomes of Komagataella phaffii.

Results

By coupling Ribo-seq with long-read mRNA sequencing, we generated a new annotation of protein-encoding genes. By using Ribo-seq with subcellular fractionation, we quantified demands on co- and posttranslational translocation pathways. During exponential growth in rich media, protein components of the cell-wall represent the greatest number of nascent chains entering the ER. Transcripts encoding the transmembrane protein PMA1 sequester more ribosomes at the ER membrane than any others. Comparison to Saccharomyces cerevisiae reveals conservation in the resources allocated by gene ontology, but variation in the diversity of gene products entering the secretory pathway.

Conclusion

A subset of host proteins, particularly cell-wall components, impose the greatest biosynthetic demands in the early secretory pathway. These proteins are potential targets in strain engineering aimed at alleviating bottlenecks during heterologous protein production.

Clavulanic Acid Production by Streptomyces clavuligerus: Insights from Systems Biology, Strain Engineering, and Downstream Processing

Clavulanic acid (CA) is an irreversible β-lactamase enzyme inhibitor with a weak antibacterial activity produced by Streptomyces clavuligerus (S. clavuligerus). CA is typically co-formulated with broad-spectrum β‑lactam antibiotics such as amoxicillin, conferring them high potential to treat diseases caused by bacteria that possess β‑lactam resistance. The clinical importance of CA and the complexity of the production process motivate improvements from an interdisciplinary standpoint by integrating metabolic engineering strategies and knowledge on metabolic and regulatory events through systems biology and multi-omics approaches. In the large-scale bioprocessing, optimization of culture conditions, bioreactor design, agitation regime, as well as advances in CA separation and purification are required to improve the cost structure associated to CA production. This review presents the recent insights in CA production by S. clavuligerus, emphasizing on systems biology approaches, strain engineering, and downstream processing.

Error propagation in constraint-based modeling of Chinese hamster ovary cells

Chinese hamster ovary (CHO) cells are the most popular mammalian cell factories for the production of glycosylated biopharmaceuticals. To further increase titer and productivity and ensure product quality, rational system-level engineering strategies based on constraint-based metabolic modeling, such as flux balance analysis (FBA), have gained strong interest. However, the quality of FBA predictions depends on the accuracy of the experimental input data, especially on the exchange rates of extracellular metabolites. Yet, it is not standard practice to devote sufficient attention to the accurate determination of these rates. In this work, we investigated to what degree the sampling frequency during a batch culture and the measurement errors of metabolite concentrations influence the accuracy of the calculated exchange rates and further, how this error then propagates into FBA predictions of growth rates. We determined that accurate measurements of essential amino acids with low uptake rates are crucial for the accuracy of FBA predictions, followed by a sufficient number of analyzed time points. We observed that the measured difference in growth rates of two cell lines can only be reliably predicted when both high measurement accuracy and sampling frequency are ensured.

Automated engineering of synthetic metabolic pathways for efficient biomanufacturing

Metabolic engineering involves the engineering and optimization of processes from single-cell to fermentation in order to increase production of valuable chemicals for health, food, energy, materials and others. A systems approach to metabolic engineering has gained traction in recent years thanks to advances in strain engineering, leading to an accelerated scaling from rapid prototyping to industrial production. Metabolic engineering is nowadays on track towards a truly manufacturing technology, with reduced times from conception to production enabled by automated protocols for DNA assembly of metabolic pathways in engineered producer strains. In this review, we discuss how the success of the metabolic engineering pipeline often relies on retrobiosynthetic protocols able to identify promising production routes and dynamic regulation strategies through automated biodesign algorithms, which are subsequently assembled as embedded integrated genetic circuits in the host strain. Those approaches are orchestrated by an experimental design strategy that provides optimal scheduling planning of the DNA assembly, rapid prototyping and, ultimately, brings forward an accelerated Design-Build-Test-Learn cycle and the overall optimization of the biomanufacturing process. Achieving such a vision will address the increasingly compelling demand in our society for delivering valuable biomolecules in an affordable, inclusive and sustainable bioeconomy.

The era of big data: Genome-scale modelling meets machine learning

With omics data being generated at an unprecedented rate, genome-scale modelling has become pivotal in its organisation and analysis. However, machine learning methods have been gaining ground in cases where knowledge is insufficient to represent the mechanisms underlying such data or as a means for data curation prior to attempting mechanistic modelling. We discuss the latest advances in genome-scale modelling and the development of optimisation algorithms for network and error reduction, intracellular constraining and applications to strain design. We further review applications of supervised and unsupervised machine learning methods to omics datasets from microbial and mammalian cell systems and present efforts to harness the potential of both modelling approaches through hybrid modelling.

Key Challenges in Designing CHO Chassis Platforms

Following the success of and the high demand for recombinant protein-based therapeutics during the last 25 years, the pharmaceutical industry has invested significantly in the development of novel treatments based on biologics. Mammalian cells are the major production systems for these complex biopharmaceuticals, with Chinese hamster ovary (CHO) cell lines as the most important players. Over the years, various engineering strategies and modeling approaches have been used to improve microbial production platforms, such as bacteria and yeasts, as well as to create pre-optimized chassis host strains. However, the complexity of mammalian cells curtailed the optimization of these host cells by metabolic engineering. Most of the improvements of titer and productivity were achieved by media optimization and large-scale screening of producer clones. The advances made in recent years now open the door to again consider the potential application of systems biology approaches and metabolic engineering also to CHO. The availability of a reference genome sequence, genome-scale metabolic models and the growing number of various “omics” datasets can help overcome the complexity of CHO cells and support design strategies to boost their production performance. Modular design approaches applied to engineer industrially relevant cell lines have evolved to reduce the time and effort needed for the generation of new producer cells and to allow the achievement of desired product titers and quality. Nevertheless, important steps to enable the design of a chassis platform similar to those in use in the microbial world are still missing. In this review, we highlight the importance of mammalian cellular platforms for the production of biopharmaceuticals and compare them to microbial platforms, with an emphasis on describing novel approaches and discussing still open questions that need to be resolved to reach the objective of designing enhanced modular chassis CHO cell lines.

Generating therapeutic monoclonal antibodies to complex multi-spanning membrane targets: Overcoming the antigen challenge and enabling discovery strategies

Complex integral membrane proteins, which are embedded in the cell surface lipid bilayer by multiple transmembrane spanning helices, encompass families of proteins which are important target classes for drug discovery. These protein families include G protein-coupled receptors, ion channels and transporters. Although these proteins have typically been targeted by small molecule drugs and peptides, the high specificity of monoclonal antibodies offers a significant opportunity to selectively modulate these target proteins. However, it remains the case that isolation of antibodies with desired pharmacological function(s) has proven difficult due to technical challenges in preparing membrane protein antigens suitable to support antibody drug discovery. In this review recent progress in defining strategies for generation of membrane protein antigens is outlined. We also highlight antibody isolation strategies which have generated antibodies which bind the membrane protein and modulate the protein function.

Multiplex secretome engineering enhances recombinant protein production and purity

Host cell proteins (HCPs) are process-related impurities generated during biotherapeutic protein production. HCPs can be problematic if they pose a significant metabolic demand, degrade product quality, or contaminate the final product. Here, we present an effort to create a "clean" Chinese hamster ovary (CHO) cell by disrupting multiple genes to eliminate HCPs. Using a model of CHO cell protein secretion, we predict that the elimination of unnecessary HCPs could have a non-negligible impact on protein production. We analyze the HCP content of 6-protein, 11-protein, and 14-protein knockout clones. These cell lines exhibit a substantial reduction in total HCP content (40%-70%). We also observe higher productivity and improved growth characteristics in specific clones. The reduced HCP content facilitates purification of a monoclonal antibody. Thus, substantial improvements can be made in protein titer and purity through large-scale HCP deletion, providing an avenue to increased quality and affordability of high-value biopharmaceuticals.

Integration of Time-Series Transcriptomic Data with Genome-Scale CHO Metabolic Models for mAb Engineering

Chinese hamster ovary (CHO) cells are the most commonly used cell lines in biopharmaceutical manufacturing. Genome-scale metabolic models have become a valuable tool to study cellular metabolism. Despite the presence of reference global genome-scale CHO model, context-specific metabolic models may still be required for specific cell lines (for example, CHO-K1, CHO-S, and CHO-DG44), and for specific process conditions. Many integration algorithms have been available to reconstruct specific genome-scale models. These methods are mainly based on integrating omics data (i.e., transcriptomics, proteomics, and metabolomics) into reference genome-scale models. In the present study, we aimed to investigate the impact of time points of transcriptomics integration on the genome-scale CHO model by assessing the prediction of growth rates with each reconstructed model. We also evaluated the feasibility of applying extracted models to different cell lines (generated from the same parental cell line). Our findings illustrate that gene expression at various stages of culture slightly impacts the reconstructed models. However, the prediction capability is robust enough on cell growth prediction not only across different growth phases but also in expansion to other cell lines.