Machine learning of cellular metabolic rewiring.

Metabolic rewiring allows cells to adapt their metabolism in response to evolving environmental conditions. Traditional metabolomics techniques, whether targeted or untargeted, often struggle to interpret these adaptive shifts. Here, we introduce MetaboLiteLearner, a lightweight machine learning framework that harnesses the detailed fragmentation patterns from electron ionization (EI) collected in scan mode during gas chromatography/mass spectrometry to predict changes in the metabolite composition of metabolically adapted cells. When tested on breast cancer cells with different preferences to metastasize to specific organs, MetaboLiteLearner predicted the impact of metabolic rewiring on metabolites withheld from the training dataset using only the EI spectra, without metabolite identification or pre-existing knowledge of metabolic networks. Despite its simplicity, the model learned captured shared and unique metabolomic shifts between brain- and lung-homing metastatic lineages, suggesting cellular adaptations associated with metastasis to specific organs. Integrating machine learning and metabolomics paves the way for new insights into complex cellular adaptations.

Assessing Engraftment Following Fecal Microbiota Transplant.

Fecal Microbiota Transplant (FMT) is an FDA approved treatment for recurrent Clostridium difficile infections, and is being explored for other clinical applications, from alleviating digestive and neurological disorders, to priming the microbiome for cancer treatment, and restoring microbiomes impacted by cancer treatment. Quantifying the extent of engraftment following an FMT is important in determining if a recipient didn't respond because the engrafted microbiome didn't produce the desired outcomes (a successful FMT, but negative treatment outcome), or the microbiome didn't engraft (an unsuccessful FMT and negative treatment outcome). The lack of a consistent methodology for quantifying FMT engraftment extent hinders the assessment of FMT success and its relation to clinical outcomes, and presents challenges for comparing FMT results and protocols across studies. Here we review 46 studies of FMT in humans and model organisms and group their approaches for assessing the extent to which an FMT engrafts into three criteria: 1) Chimeric Asymmetric Community Coalescence investigates microbiome shifts following FMT engraftment using methods such as alpha diversity comparisons, beta diversity comparisons, and microbiome source tracking. 2) Donated Microbiome Indicator Features tracks donated microbiome features (e.g., amplicon sequence variants or species of interest) as a signal of engraftment with methods such as differential abundance testing based on the current sample collection, or tracking changes in feature abundances that have been previously identified (e.g., from FMT or disease-relevant literature). 3) Temporal Stability examines how resistant post-FMT recipient's microbiomes are to reverting back to their baseline microbiome. Individually, these criteria each highlight a critical aspect of microbiome engraftment; investigated together, however, they provide a clearer assessment of microbiome engraftment. We discuss the pros and cons of each of these criteria, providing illustrative examples of their application. We also introduce key terminology and recommendations on how FMT studies can be analyzed for rigorous engraftment extent assessment.

Fatty acid oxidation fuels natural killer cell responses against infection and cancer.

Natural killer (NK) cells are a vital part of the innate immune system capable of rapidly clearing mutated or infected cells from the body and promoting an immune response. Here, we find that NK cells activated by viral infection or tumor challenge increase uptake of fatty acids and their expression of carnitine palmitoyltransferase I (CPT1A), a critical enzyme for long-chain fatty acid oxidation. Using a mouse model with an NK cell-specific deletion of CPT1A, combined with stable 13C isotope tracing, we observe reduced mitochondrial function and fatty acid-derived aspartate production in CPT1A-deficient NK cells. Furthermore, CPT1A-deficient NK cells show reduced proliferation after viral infection and diminished protection against cancer due to impaired actin cytoskeleton rearrangement. Together, our findings highlight that fatty acid oxidation promotes NK cell metabolic resilience, processes that can be optimized in NK cell-based immunotherapies.

Machine learning of cellular metabolic rewiring.

Metabolic rewiring allows cells to adapt their metabolism in response to evolving environmental conditions. Traditional metabolomics techniques, whether targeted or untargeted, often struggle to interpret these adaptive shifts. Here, we introduce MetaboLiteLearner, a machine learning framework that harnesses the detailed fragmentation patterns from electron ionization (EI) collected in scan mode during gas chromatography/mass spectrometry (GC/MS) to predict abundance changes in metabolically adapted cells. When tested on breast cancer cells with different preferences to metastasize to specific organs, MetaboLiteLearner predicted the impact of metabolic rewiring on metabolites withheld from the training dataset using only the EI spectra, without metabolite identification or pre-existing knowledge of metabolic networks. The model learned captures shared and unique metabolomic shifts between brain- and lung-homing metastatic lineages, suggesting potential organ-tailored cellular adaptations. Integrating machine learning and metabolomics paves the way for new insights into complex cellular adaptations.

The Ratio of Key Metabolic Transcripts Is a Predictive Biomarker of Breast Cancer Metastasis to the Lung.

Understanding the rewired metabolism underlying organ-specific metastasis in breast cancer could help identify strategies to improve the treatment and prevention of metastatic disease. Here, we used a systems biology approach to compare metabolic fluxes used by parental breast cancer cells and their brain- and lung-homing derivatives. Divergent lineages had distinct, heritable metabolic fluxes. Lung-homing cells maintained high glycolytic flux despite low levels of glycolytic intermediates, constitutively activating a pathway sink into lactate. This strong Warburg effect was associated with a high ratio of lactate dehydrogenase (LDH) to pyruvate dehydrogenase (PDH) expression, which correlated with lung metastasis in patients with breast cancer. Although feature classification models trained on clinical characteristics alone were unable to predict tropism, the LDH/PDH ratio was a significant predictor of metastasis to the lung but not to other organs, independent of other transcriptomic signatures. High lactate efflux was also a trait in lung-homing metastatic pancreatic cancer cells, suggesting that lactate production may be a convergent phenotype in lung metastasis. Together, these analyses highlight the essential role that metabolism plays in organ-specific cancer metastasis and identify a putative biomarker for predicting lung metastasis in patients with breast cancer. SIGNIFICANCE: Lung-homing metastatic breast cancer cells express an elevated ratio of lactate dehydrogenase to pyruvate dehydrogenase, indicating that ratios of specific metabolic gene transcripts have potential as metabolic biomarkers for predicting organ-specific metastasis.

High-resolution analyses of associations between medications, microbiome, and mortality in cancer patients.

Discerning the effect of pharmacological exposures on intestinal bacterial communities in cancer patients is challenging. Here, we deconvoluted the relationship between drug exposures and changes in microbial composition by developing and applying a new computational method, PARADIGM (parameters associated with dynamics of gut microbiota), to a large set of longitudinal fecal microbiome profiles with detailed medication-administration records from patients undergoing allogeneic hematopoietic cell transplantation. We observed that several non-antibiotic drugs, including laxatives, antiemetics, and opioids, are associated with increased Enterococcus relative abundance and decreased alpha diversity. Shotgun metagenomic sequencing further demonstrated subspecies competition, leading to increased dominant-strain genetic convergence during allo-HCT that is significantly associated with antibiotic exposures. We integrated drug-microbiome associations to predict clinical outcomes in two validation cohorts on the basis of drug exposures alone, suggesting that this approach can generate biologically and clinically relevant insights into how pharmacological exposures can perturb or preserve microbiota composition. The application of a computational method called PARADIGM to a large dataset of cancer patients' longitudinal fecal specimens and detailed daily medication records reveals associations between drug exposures and the intestinal microbiota that recapitulate in vitro findings and are also predictive of clinical outcomes.

The TaxUMAP atlas: Efficient display of large clinical microbiome data reveals ecological competition in protection against bacteremia.

Longitudinal microbiome data provide valuable insight into disease states and clinical responses, but they are challenging to mine and view collectively. To address these limitations, we present TaxUMAP, a taxonomically informed visualization for displaying microbiome states in large clinical microbiome datasets. We used TaxUMAP to chart a microbiome atlas of 1,870 patients with cancer during therapy-induced perturbations. Bacterial density and diversity were positively associated, but the trend was reversed in liquid stool. Low-diversity states (dominations) remained stable after antibiotic treatment, and diverse communities had a broader range of antimicrobial resistance genes than dominations. When examining microbiome states associated with risk for bacteremia, TaxUMAP revealed that certain Klebsiella species were associated with lower risk for bacteremia localize in a region of the atlas that is depleted in high-risk enterobacteria. This indicated a competitive interaction that was validated experimentally. Thus, TaxUMAP can chart comprehensive longitudinal microbiome datasets, enabling insights into microbiome effects on human health.

A compilation of fecal microbiome shotgun metagenomics from hematopoietic cell transplantation patients.

Hospitalized patients receiving hematopoietic cell transplants provide a unique opportunity to study the human gut microbiome. We previously compiled a large-scale longitudinal dataset of fecal microbiota and associated metadata, but we had limited that analysis to taxonomic composition of bacteria from 16S rRNA gene sequencing. Here we augment those data with shotgun metagenomics. The compilation amounts to a nested subset of 395 samples compiled from different studies at Memorial Sloan Kettering. Shotgun metagenomics describes the microbiome at the functional level, particularly in antimicrobial resistances and virulence factors. We provide accession numbers that link each sample to the paired-end sequencing files deposited in a public repository, which can be directly accessed by the online services of PATRIC to be analyzed without the users having to download or transfer the files. Then, we show how shotgun sequencing enables the assembly of genomes from metagenomic data. The new data, combined with the metadata published previously, enables new functional studies of the microbiomes of patients with cancer receiving bone marrow transplantation.

Optimal Strategy and Benefit of Pulsed Therapy Depend On Tumor Heterogeneity and Aggressiveness at Time of Treatment Initiation.

Therapeutic resistance is a fundamental obstacle in cancer treatment. Tumors that initially respond to treatment may have a preexisting resistant subclone or acquire resistance during treatment, making relapse theoretically inevitable. Here, we investigate treatment strategies that may delay relapse using mathematical modeling. We find that for a single-drug therapy, pulse treatment-short, elevated doses followed by a complete break from treatment-delays relapse compared with continuous treatment with the same total dose over a length of time. For tumors treated with more than one drug, continuous combination treatment is only sometimes better than sequential treatment, while pulsed combination treatment or simply alternating between the two therapies at defined intervals delays relapse the longest. These results are independent of the fitness cost or benefit of resistance, and are robust to noise. Machine-learning analysis of simulations shows that the initial tumor response and heterogeneity at the start of treatment suffice to determine the benefit of pulsed or alternating treatment strategies over continuous treatment. Analysis of eight tumor burden trajectories of breast cancer patients treated at Memorial Sloan Kettering Cancer Center shows the model can predict time to resistance using initial responses to treatment and estimated preexisting resistant populations. The model calculated that pulse treatment would delay relapse in all eight cases. Overall, our results support that pulsed treatments optimized by mathematical models could delay therapeutic resistance.

Optimizing the future: how mathematical models inform treatment schedules for cancer.

For decades, mathematical models have influenced how we schedule chemotherapeutics. More recently, mathematical models have leveraged lessons from ecology, evolution, and game theory to advance predictions of optimal treatment schedules, often in a personalized medicine manner. We discuss both established and emerging therapeutic strategies that deviate from canonical standard-of-care regimens, and how mathematical models have contributed to the design of such schedules. We first examine scheduling options for single therapies and review the advantages and disadvantages of various treatment plans. We then consider the challenge of scheduling multiple therapies, and review the mathematical and clinical support for various conflicting treatment schedules. Finally, we propose how a consilience of mathematical and clinical knowledge can best determine the optimal treatment schedules for patients.

Haematopoietic cell transplantation outcomes are linked to intestinal mycobiota dynamics and an expansion of Candida parapsilosis complex species.

Allogeneic haematopoietic cell transplantation (allo-HCT) induces profound shifts in the intestinal bacterial microbiota. The dynamics of intestinal fungi and their impact on clinical outcomes during allo-HCT are not fully understood. Here we combined parallel high-throughput fungal ITS1 amplicon sequencing, bacterial 16S amplicon sequencing and fungal cultures of 1,279 faecal samples from a cohort of 156 patients undergoing allo-HCT to reveal potential trans-kingdom dynamics and their association with patient outcomes. We saw that the overall density and the biodiversity of intestinal fungi were stable during allo-HCT but the species composition changed drastically from day to day. We identified a subset of patients with fungal dysbiosis defined by culture positivity (n = 53) and stable expansion of Candida parapsilosis complex species (n = 19). They presented with distinct trans-kingdom microbiota profiles, characterized by a decreased intestinal bacterial biomass. These patients had worse overall survival and higher transplant-related mortality independent of candidaemia. This expands our understanding of the clinical significance of the mycobiota and suggests that targeting fungal dysbiosis may help to improve long-term patient survival.

Cooperation between melanoma cell states promotes metastasis through heterotypic cluster formation.

Melanomas can have multiple coexisting cell states, including proliferative (PRO) versus invasive (INV) subpopulations that represent a "go or grow" trade-off; however, how these populations interact is poorly understood. Using a combination of zebrafish modeling and analysis of patient samples, we show that INV and PRO cells form spatially structured heterotypic clusters and cooperate in the seeding of metastasis, maintaining cell state heterogeneity. INV cells adhere tightly to each other and form clusters with a rim of PRO cells. Intravital imaging demonstrated cooperation in which INV cells facilitate dissemination of less metastatic PRO cells. We identified the TFAP2 neural crest transcription factor as a master regulator of clustering and PRO/INV states. Isolation of clusters from patients with metastatic melanoma revealed a subset with heterotypic PRO-INV clusters. Our data suggest a framework for the co-existence of these two divergent cell populations, in which heterotypic clusters promote metastasis via cell-cell cooperation.

Compilation of longitudinal microbiota data and hospitalome from hematopoietic cell transplantation patients.

The impact of the gut microbiota in human health is affected by several factors including its composition, drug administrations, therapeutic interventions and underlying diseases. Unfortunately, many human microbiota datasets available publicly were collected to study the impact of single variables, and typically consist of outpatients in cross-sectional studies, have small sample numbers and/or lack metadata to account for confounders. These limitations can complicate reusing the data for questions outside their original focus. Here, we provide comprehensive longitudinal patient dataset that overcomes those limitations: a collection of fecal microbiota compositions (>10,000 microbiota samples from >1,000 patients) and a rich description of the "hospitalome" experienced by the hosts, i.e., their drug exposures and other metadata from patients with cancer, hospitalized to receive allogeneic hematopoietic cell transplantation (allo-HCT) at a large cancer center in the United States. We present five examples of how to apply these data to address clinical and scientific questions on host-associated microbial communities.

The gut microbiota is associated with immune cell dynamics in humans.

The gut microbiota influences development1-3 and homeostasis4-7 of the mammalian immune system, and is associated with human inflammatory8 and immune diseases9,10 as well as responses to immunotherapy11-14. Nevertheless, our understanding of how gut bacteria modulate the immune system remains limited, particularly in humans, where the difficulty of direct experimentation makes inference challenging. Here we study hundreds of hospitalized-and closely monitored-patients with cancer receiving haematopoietic cell transplantation as they recover from chemotherapy and stem-cell engraftment. This aggressive treatment causes large shifts in both circulatory immune cell and microbiota populations, enabling the relationships between the two to be studied simultaneously. Analysis of observed daily changes in circulating neutrophil, lymphocyte and monocyte counts and more than 10,000 longitudinal microbiota samples revealed consistent associations between gut bacteria and immune cell dynamics. High-resolution clinical metadata and Bayesian inference allowed us to compare the effects of bacterial genera in relation to those of immunomodulatory medications, revealing a considerable influence of the gut microbiota-together and over time-on systemic immune cell dynamics. Our analysis establishes and quantifies the link between the gut microbiota and the human immune system, with implications for microbiota-driven modulation of immunity.

Microbiota as Predictor of Mortality in Allogeneic Hematopoietic-Cell Transplantation.

BACKGROUND: Relationships between microbiota composition and clinical outcomes after allogeneic hematopoietic-cell transplantation have been described in single-center studies. Geographic variations in the composition of human microbial communities and differences in clinical practices across institutions raise the question of whether these associations are generalizable. METHODS: The microbiota composition of fecal samples obtained from patients who were undergoing allogeneic hematopoietic-cell transplantation at four centers was profiled by means of 16S ribosomal RNA gene sequencing. In an observational study, we examined associations between microbiota diversity and mortality using Cox proportional-hazards analysis. For stratification of the cohorts into higher- and lower-diversity groups, the median diversity value that was observed at the study center in New York was used. In the analysis of independent cohorts, the New York center was cohort 1, and three centers in Germany, Japan, and North Carolina composed cohort 2. Cohort 1 and subgroups within it were analyzed for additional outcomes, including transplantation-related death. RESULTS: We profiled 8767 fecal samples obtained from 1362 patients undergoing allogeneic hematopoietic-cell transplantation at the four centers. We observed patterns of microbiota disruption characterized by loss of diversity and domination by single taxa. Higher diversity of intestinal microbiota was associated with a lower risk of death in independent cohorts (cohort 1: 104 deaths among 354 patients in the higher-diversity group vs. 136 deaths among 350 patients in the lower-diversity group; adjusted hazard ratio, 0.71; 95% confidence interval [CI], 0.55 to 0.92; cohort 2: 18 deaths among 87 patients in the higher-diversity group vs. 35 deaths among 92 patients in the lower-diversity group; adjusted hazard ratio, 0.49; 95% CI, 0.27 to 0.90). Subgroup analyses identified an association between lower intestinal diversity and higher risks of transplantation-related death and death attributable to graft-versus-host disease. Baseline samples obtained before transplantation already showed evidence of microbiome disruption, and lower diversity before transplantation was associated with poor survival. CONCLUSIONS: Patterns of microbiota disruption during allogeneic hematopoietic-cell transplantation were similar across transplantation centers and geographic locations; patterns were characterized by loss of diversity and domination by single taxa. Higher diversity of intestinal microbiota at the time of neutrophil engraftment was associated with lower mortality. (Funded by the National Cancer Institute and others.).

The Cancer Microbiome: Distinguishing Direct and Indirect Effects Requires a Systemic View.

The collection of microbes that live in and on the human body - the human microbiome - can impact on cancer initiation, progression, and response to therapy, including cancer immunotherapy. The mechanisms by which microbiomes impact on cancers can yield new diagnostics and treatments, but much remains unknown. The interactions between microbes, diet, host factors, drugs, and cell-cell interactions within the cancer itself likely involve intricate feedbacks, and no single component can explain all the behavior of the system. Understanding the role of host-associated microbial communities in cancer systems will require a multidisciplinary approach combining microbial ecology, immunology, cancer cell biology, and computational biology - a systems biology approach.

Regenerative lineages and immune-mediated pruning in lung cancer metastasis.

Developmental processes underlying normal tissue regeneration have been implicated in cancer, but the degree of their enactment during tumor progression and under the selective pressures of immune surveillance, remain unknown. Here we show that human primary lung adenocarcinomas are characterized by the emergence of regenerative cell types, typically seen in response to lung injury, and by striking infidelity among transcription factors specifying most alveolar and bronchial epithelial lineages. In contrast, metastases are enriched for key endoderm and lung-specifying transcription factors, SOX2 and SOX9, and recapitulate more primitive transcriptional programs spanning stem-like to regenerative pulmonary epithelial progenitor states. This developmental continuum mirrors the progressive stages of spontaneous outbreak from metastatic dormancy in a mouse model and exhibits SOX9-dependent resistance to natural killer cells. Loss of developmental stage-specific constraint in macrometastases triggered by natural killer cell depletion suggests a dynamic interplay between developmental plasticity and immune-mediated pruning during metastasis.

High Heterogeneity of Multidrug-Resistant Enterobacteriaceae Fecal Levels in Hospitalized Patients Is Partially Driven by Intravenous ß-Lactams.

Multidrug-resistant Enterobacteriaceae (MRE) colonize the intestine asymptomatically from where they can breach into the bloodstream and cause life-threatening infections, especially in heavily colonized patients. Despite the clinical relevance of MRE colonization levels, we know little about how they vary in hospitalized patients and the clinical factors that determine those levels. Here, we conducted one of the largest studies of MRE fecal levels by tracking longitudinally 133 acute leukemia patients and monitoring their MRE levels over time through extensive culturing. MRE were defined as Enterobacteriaceae species that acquired nonsusceptibility to ≥1 agent in ≥3 antimicrobial categories. In addition, due to the selective media used, the MRE had to be resistant to third-generation cephalosporins. MRE were detected in 60% of the patients, but their fecal levels varied considerably among patients and within the same patient (>6 and 4 orders of magnitude, respectively). Multivariate analysis of clinical metadata revealed an impact of intravenous beta-lactams (i.e., meropenem and piperacillin-tazobactam), which significantly diminished the fecal MRE levels in hospitalized patients. Consistent with a direct action of beta-lactams, we found an effect only when the patient was colonized with strains sensitive to the administered beta-lactam (P < 0.001) but not with nonsusceptible strains. We report previously unobserved inter- and intraindividual heterogeneity in MRE fecal levels, suggesting that quantitative surveillance is more informative than qualitative surveillance of hospitalized patients. In addition, our study highlights the relevance of incorporating antibiotic treatment and susceptibility data of gut-colonizing pathogens for future clinical studies and in clinical decision-making.

Systems-level analysis of NalD mutation, a recurrent driver of rapid drug resistance in acute Pseudomonas aeruginosa infection.

Pseudomonas aeruginosa, a main cause of human infection, can gain resistance to the antibiotic aztreonam through a mutation in NalD, a transcriptional repressor of cellular efflux. Here we combine computational analysis of clinical isolates, transcriptomics, metabolic modeling and experimental validation to find a strong association between NalD mutations and resistance to aztreonam-as well as resistance to other antibiotics-across P. aeruginosa isolated from different patients. A detailed analysis of one patient's timeline shows how this mutation can emerge in vivo and drive rapid evolution of resistance while the patient received cancer treatment, a bone marrow transplantation, and antibiotics up to the point of causing the patient's death. Transcriptomics analysis confirmed the primary mechanism of NalD action-a loss-of-function mutation that caused constitutive overexpression of the MexAB-OprM efflux system-which lead to aztreonam resistance but, surprisingly, had no fitness cost in the absence of the antibiotic. We constrained a genome-scale metabolic model using the transcriptomics data to investigate changes beyond the primary mechanism of resistance, including adaptations in major metabolic pathways and membrane transport concurrent with aztreonam resistance, which may explain the lack of a fitness cost. We propose that metabolic adaptations may allow resistance mutations to endure in the absence of antibiotics and could be targeted by future therapies against antibiotic resistant pathogens.