Genome-scale metabolic models represent the entirety of metabolic reactions of an organism based on the annotation of the respective genome. These models commonly allow all reactions to proceed concurrently, disregarding the fact that at no point all proteins will be present in a cell. The metabolic reaction space can be constrained to a more physiological state using experimentally obtained information on enzyme abundances. However, high-quality, genome-wide protein measurements have been challenging and typically transcript abundances have been used as a surrogate for protein measurements. With recent developments in mass spectrometry-based proteomics, exemplified by SWATH-MS, the acquisition of highly quantitative proteome-wide data at reasonable throughput has come within reach. Here we present methodology to integrate such proteome-wide data into genome-scale models. We applied this methodology to study cellular changes in Enterococcus faecalis during adaptation to low pH. Our results indicate reduced proton production in the central metabolism and decreased membrane permeability for protons due to different membrane composition. We conclude that proteomic data constrain genome-scale models to a physiological state and, in return, genome-scale models are useful tools to contextualize proteomic data.

Arginine auxotrophy constitutes a weak point of several tumors, among them glioblastoma multiforme (GBM). Hence, those tumors are supposed to be sensitive for arginine-depleting substances, such as arginine deiminase (ADI). Here we elucidated the sensitivity of patient-individual GBM cell lines toward Streptococcus pyogenes-derived ADI. To improve therapy, ADI was combined with currently established and pre-clinical cytostatic drugs. Additionally, effectiveness of local ADI therapy was determined in xenopatients. Half of the GBM cell lines tested responded well toward ADI monotherapy. In those cell lines, viability decreased significantly (up to 50 %). Responding cell lines were subjected to combination therapy experiments to test if any additive or even synergistic effects may be achieved. Such promising results were obtained in 2/3 cases. In cell lines HROG02, HROG05 and HROG10, ADI and Palomid 529 combinations were most effective yielding more than 70 % killing after 2 rounds of treatment. Comparable boosted antitumoral effects were observed after adding chloroquine to ADI (>60% killing). Apoptosis, as well as cell cycle dysregulation were found to play a minor role. In some, but clearly not all cases, (epi-) genetic silencing of arginine synthesis pathway genes (argininosuccinate synthetase 1 and argininosuccinate lyase) explained obtained results. In vivo, ADI as well as the combination of ADI and SAHA efficiently controlled HROG05 xenograft growth, whereas adding Palomid 529 to ADI did not further increase the strong antitumoral effect of ADI. The cumulative in vitro and in vivo results proved ADI as a very promising candidate therapeutic, especially for development of adjuvant GBM combination treatments.

ObjectivesTotal joint arthroplasty is one of the most frequent and effective surgeries today. However, despite improved surgical techniques, a significant number of implant-associated infections still occur. Suitable in vitro models are needed to test potential approaches to prevent infection. In the present study, we aimed to establish an in vitro co-culture setup of human primary osteoblasts and S. epidermidis to model the onset of implant-associated infections, and to analyze antimicrobial implant surfaces and coatings.Materials and MethodsFor initial surface adhesion, human primary osteoblasts (hOB) were grown for 24 hours on test sample discs made of polystyrene, titanium alloy Ti6Al4V, bone cement PALACOS R®, and PALACOS R® loaded with antibiotics. Co-cultures were performed as a single-species infection on the osteoblasts with S. epidermidis (multiplicity of infection of 0.04), and were incubated for 2 and 7 days under aerobic conditions. Planktonic S. epidermidis was quantified by centrifugation and determination of colony-forming units (CFU). The quantification of biofilm-bound S. epidermidis on the test samples was performed by sonication and CFU counting. Quantification of adherent and vital primary osteoblasts on the test samples was performed by trypan-blue staining and counting. Scanning electron microscopy was used for evaluation of topography and composition of the species on the sample surfaces.ResultsAfter 2 days, we observed approximately 104 CFU/ml biofilm-bound S. epidermidis (103 CFU/ml initial population) on the antibiotics-loaded bone cement samples in the presence of hOB, while no bacteria were detected without hOB. No biofilm-bound bacteria were detectable after 7 days in either case. Similar levels of planktonic bacteria were observed on day 2 with and without hOB. After 7 days, about 105 CFU/ml planktonic bacteria were present, but only in the absence of hOB. Further, no bacteria were observed within the biofilm, while the number of hOB was decreased to 10% of its initial value compared to 150% in the mono-culture of hOB.ConclusionWe developed a co-culture setup that serves as a more comprehensive in vitro model for the onset of implant-associated infections and provides a test method for antimicrobial implant materials and coatings. We demonstrate that observations can be made that are unavailable from mono-culture experiments.