Search

Abstract Background Pathological complete response (pCR) is associated with favorable prognosis in patients with triple-negative breast cancer (TNBC). However, only 30–40% of TNBC patients treated with neoadjuvant chemotherapy (NAC) show pCR, while the remaining 60–70% show residual disease (RD). The role of the tumor microenvironment in NAC response in patients with TNBC remains unclear. In this study, we developed a machine learning-based two-step pipeline to distinguish between various histological components in hematoxylin and eosin (H&E)-stained whole slide images (WSIs) of TNBC tissue biopsies and to identify histological features that can predict NAC response. Methods H&E-stained WSIs of treatment-naïve biopsies from 85 patients (51 with pCR and 34 with RD) of the model development cohort and 79 patients (41 with pCR and 38 with RD) of the validation cohort were separated through a stratified eightfold cross-validation strategy for the first step and leave-one-out cross-validation strategy for the second step. A tile-level histology label prediction pipeline and four machine-learning classifiers were used to analyze 468,043 tiles of WSIs. The best-trained classifier used 55 texture features from each tile to produce a probability profile during testing. The predicted histology classes were used to generate a histology classification map of the spatial distributions of different tissue regions. A patient-level NAC response prediction pipeline was trained with features derived from paired histology classification maps. The top graph-based features capturing the relevant spatial information across the different histological classes were provided to the radial basis function kernel support vector machine (rbfSVM) classifier for NAC treatment response prediction. Results The tile-level prediction pipeline achieved 86.72% accuracy for histology class classification, while the patient-level pipeline achieved 83.53% NAC response (pCR vs. RD) prediction accuracy of the model development cohort. The model was validated with an independent cohort with tile histology validation accuracy of 83.59% and NAC prediction accuracy of 81.01%. The histological class pairs with the strongest NAC response predictive ability were tumor and tumor tumor-infiltrating lymphocytes for pCR and microvessel density and polyploid giant cancer cells for RD. Conclusion Our machine learning pipeline can robustly identify clinically relevant histological classes that predict NAC response in TNBC patients and may help guide patient selection for NAC treatment.

Antimicrobial resistance is now commonly observed in bacterial isolates from multiple settings, compromising the efficacy of current antimicrobial agents. Therefore, there is an urgent requirement for efficacious novel antimicrobials to be used as therapeutics, prophylactically or as preservatives. One promising source of novel antimicrobial chemicals is phytochemicals, which are secondary metabolites produced by plants for numerous purposes, including antimicrobial defence. In this report, we compare the bioactivity of a range of phytochemical compounds, testing their ability to directly inhibit growth or to potentiate other antimicrobials against Salmonella enterica Typhimurium, Pseudomonas aeruginosa, Listeria monocytogenes, and Staphylococcus aureus. We found that nine compounds displayed consistent bioactivity either as direct antimicrobials or as potentiators. Thymol at 0.5 mg/mL showed the greatest antimicrobial effect and significantly reduced the growth of all species, reducing viable cell populations by 66.8%, 43.2%, 29.5%, and 70.2% against S. enterica Typhimurium, S. aureus, P. aeruginosa, and L. monocytogenes, respectively. Selection of mutants with decreased susceptibility to thymol was possible for three of the pathogens, at a calculated rate of 3.77 × 10−8, and characterisation of S. enterica Typhimurium mutants showed a low-level MDR phenotype due to over-expression of the major efflux system AcrAB-TolC. These data show that phytochemicals can have strong antimicrobial activity, but emergence of resistance should be evaluated in any further development.

ABSTRACT Cephalosporins are important beta lactam antibiotics, but resistance can be mediated by various mechanisms including production of beta lactamase enzymes, changes in membrane permeability or active efflux. We used an evolution model to study how Salmonella adapts to subinhibitory concentrations of cefotaxime in planktonic and biofilm conditions and characterized the mechanisms underpinning this adaptation. We found that Salmonella rapidly adapts to subinhibitory concentrations of cefotaxime via selection of multiple mutations within the CA-domain region of EnvZ. We showed that changes in this domain affect the ATPase activity of the enzyme and in turn impact OmpC, OmpF porin expression and hence membrane permeability leading to increased tolerance to cefotaxime and low-level resistance to different classes of antibiotics. Adaptation to cefotaxime through EnvZ also resulted in a significant cost to biofilm formation due to downregulation of curli. We assessed the role of the mutations identified on the activity of EnvZ by genetic characterization, biochemistry and in silico analysis and confirmed that they are responsible for the observed phenotypes. We observed that sublethal cefotaxime exposure selected for heterogeneity in populations with only a subpopulation carrying mutations within EnvZ and being resistant to cefotaxime. Population structure and composition dynamically changed depending on the presence of the selection pressure, once selected, resistant subpopulations were maintained even in extended passage without drug. IMPORTANCE Understanding mechanisms of antibiotic resistance is crucial to guide how best to use antibiotics to minimize emergence of resistance. We used a laboratory evolution system to study how Salmonella responds to cefotaxime in both planktonic and biofilm conditions. In both contexts, we observed rapid selection of mutants within a single hot spot within envZ. The mutations selected altered EnvZ which in turn triggers changes in porin production at the outer membrane. Emergence of mutations within this region was repeatedly observed in parallel lineages in different conditions. We used a combination of genetics, biochemistry, phenotyping and structural analysis to understand the mechanisms. This data show that the changes we observe provide resistance to cefotaxime but come at a cost to biofilm formation and the fitness of mutants changes greatly depending on the presence or absence of a selective drug. Studying how resistance emerges can inform selective outcomes in the real world.