Biology subjects

Lauffenburger, D. A.

Publications and source records attributed to Lauffenburger, D. A..

36 records · Page 2Linked to original sources

bioRxiv2022

An affinity for brainstem microglia in pediatric high grade gliomas of brainstem origin

BackgroundHigh-grade gliomas (HGG) in children have a devastating prognosis and occur in a remarkable spatiotemporal pattern. Diffuse midline gliomas (DMG), including diffuse intrinsic pontine gliomas (DIPG), typically occur in mid-childhood, while cortical HGGs are more frequent in older children and adults. The mechanisms behind this pattern are not clear. MethodsWe used mouse organotypic slice cultures and glial cell cultures to test the impact of the microenvironment on human DIPG cells. Comparing the expression between brainstem and cortical microglia identified differentially expressed secreted proteins. The impact of some of these proteins on DIPGs was tested. ResultsDIPGs, pediatric HGGs of brainstem origin, survive and divide more in organotypic slice cultures originating in the brainstem as compared to the cortex. Moreover, brainstem microglia are better able to support tumors of brainstem origin. A comparison between the two microglial populations revealed differentially expressed genes. One such gene, Interleukin-33 (IL33), is highly expressed in the pons of young mice and its DIPG receptor is upregulated in this context. Consistent with this observation, the expression levels of IL33 and its receptor, IL1RL1, are higher in DIPG biopsies compared to low grade cortical gliomas. Furthermore, IL33 can enhance proliferation and clonability of HGGs of brainstem origin, while blocking IL33 in brainstem organotypic slice cultures reduced the proliferation of these tumor cells. ConclusionsCrosstalk between DIPGs and the brainstem microenvironment, in particular microglia, through IL33 and other secreted factors, modulates spatiotemporal patterning of this HGG and could prove to be an important future therapeutic target. Key PointsO_LIDIPGs divide more in the brainstem environment compared to the cortex. C_LIO_LIIL33 expression in brainstem and brainstem microglia is higher than in the cortex of young mice. C_LIO_LIIL33 propagates DIPG tumorigenesis, and its inhibition decreases DIPG proliferation. C_LI Importance of the StudyPediatric gliomas can arise in the supratentorial cerebral cortex with a median diagnosis age of 13 years, while diffuse midline gliomas (DMGs), present a median diagnosis age of 7 years, and appear most frequently in the pons. DMGs have a devastating prognosis and are currently untreatable. We tested the contribution of the brainstem microenvironment to the observed spatiotemporal pattern of pediatric HGGs. Our results support a model in which the brainstem microenvironment, particularly microglia, offers more supportive conditions for DIPG cells. By using the differences in gene expression between brainstem and cortical microglia we were able to identify factors that may contribute to the DIPG preference toward the brainstem. In this study we focused on the factor IL33, and provided evidence to support its contribution to the propagation of DIPGs in this specific anatomical location.

cancer biology↗

bioRxiv2022

Biphasic JNK-Erk Signaling Separates Induction and Maintenance of Cell Senescence after DNA Damage

Genotoxic stress in mammalian cells, including that caused by anti-cancer chemotherapy, can induce temporary cell cycle arrest, DNA damage-induced senescence (DDIS) or apoptotic cell death. Despite obvious clinical importance, it is unclear how the signals emerging from DNA damage are integrated together with other cellular signaling pathways monitoring the cells environment and/or internal state to control these different cell fates. Here, using a combination of single cell-based signaling measurements and tensor PLSR/PCA computational approaches, we show that the JNK and Erk MAPK signaling pathways regulate the initiation of senescence through the transcription factor AP-1 at early times after extrinsic DNA damage, and the Senescence Associated Secretory Phenotype, a hallmark of DDIS, at late times after damage. These results identify a time-based separation of function for the same signaling pathways beyond the classic DNA damage response that control the cell senescence decision and modulate the tumor microenvironment following genotoxic stress, and reveal a fundamental similarity between signaling mechanisms responsible for oncogene-induced senescence and senescence caused by extrinsic DNA damaging agents.

systems biology↗

bioRxiv2021

Artificial neural networks enable genome-scale simulations of intracellular signaling

Mammalian cells adapt their functional state in response to external signals in form of ligands that bind receptors on the cell-surface. Mechanistically, this involves signal-processing through a complex network of molecular interactions that govern transcription factor (TF) activity patterns. Computer simulations of the information flow through this network could help predict cellular responses in health and disease. Here we develop a recurrent neural network constrained by prior knowledge of the signaling network with ligand concentrations as input, TF activity as output and signaling molecules as hidden nodes. Simulations are assumed to reach steady state, and we regularize the parameters to enforce this. Using synthetic data, we train models that generalize to unseen data and predict the effects of gene knockouts. We also fit models to a small experimental data set from literature and confirm the predictions using cross validation. This demonstrates the feasibility of simulating intracellular signaling at the genome-scale.

systems biology↗

bioRxiv2021

Subtle immunological differences in mRNA-1273 and BNT162b2 COVID-19 vaccine induced Fc-functional profiles

The successful development of several COVID-19 vaccines has substantially reduced morbidity and mortality in regions of the world where the vaccines have been deployed. However, in the wake of the emergence of viral variants, able to evade vaccine induced neutralizing antibodies, real world vaccine efficacy has begun to show differences across the mRNA platforms, suggesting that subtle variation in immune responses induced by the BNT162b2 and mRNA1273 vaccines may provide differential protection. Given our emerging appreciation for the importance of additional antibody functions, beyond neutralization, here we profiled the postboost binding and functional capacity of the humoral response induced by the BNT162b2 and mRNA-1273 in a cohort of hospital staff. Both vaccines induced robust humoral immune responses to WT SARS-CoV-2 and VOCs. However, differences emerged across epitopespecific responses, with higher RBD- and NTD-specific IgA, as well as functional antibodies (ADNP and ADNK) in mRNA-1273 vaccine recipients. Additionally, RBD-specific antibody depletion highlighted the different roles of non-RBD-specific antibody effector function induced across the mRNA vaccines, providing novel insights into potential differences in protective immunity generated across these vaccines in the setting of newly emerging VOCs.

immunology↗

bioRxiv2021

CARPOOL: A library-based platform to rapidly identify next generation chimeric antigen receptors

CD19-targeted CAR therapies have successfully treated B cell leukemias and lymphomas, but many responders later relapse or experience toxicities. CAR intracellular domains (ICDs) are key to converting antigen recognition into anti-tumor effector functions. Despite the many possible immune signaling domain combinations that could be included in CARs, almost all CARs currently rely upon CD3{zeta}, CD28, and/or 4-1BB signaling. To explore the signaling potential of CAR ICDs, we generated a library of 700,000 CD19 CAR molecules with diverse signaling domains and developed a high throughput screening platform to enable optimization of CAR signaling for anti-tumor functions. Our strategy identifies CARs with novel signaling domain combinations that elicit distinct T cell behaviors from a clinically available CAR, including enhanced proliferation and persistence, lower exhaustion, potent cytotoxicity in an in vitro tumor rechallenge condition, and comparable tumor control in vivo. This approach is readily adaptable to numerous disease models, cell types, and selection conditions, making it a promising tool for rapidly improving adoptive cell therapies and expanding their utility to new disease indications.

immunology↗

bioRxiv2021

Early cross-coronavirus reactive signatures of protective humoral immunity against COVID-19

The introduction of vaccines has inspired new hope in the battle against SARS-CoV-2. However, the emergence of viral variants, in the absence of potent antivirals, has left the world struggling with the uncertain nature of this disease. Antibodies currently represent the strongest correlate of immunity against COVID-19, thus we profiled the earliest humoral signatures in a large cohort of severe and asymptomatic COVID-19 individuals. While a SARS-CoV-2-specific immune response evolved rapidly in survivors of COVID-19, non-survivors exhibited blunted and delayed humoral immune evolution, particularly with respect to S2-specific antibody evolution. Given the conservation of S2 across {beta}-coronaviruses, we found the early development of SARS-CoV-2-specific immunity occurred in tandem with pre-existing common {beta}-coronavirus OC43 humoral immunity in survivors, which was selectively also expanded in individuals that develop paucisymptomatic infection. These data point to the importance of cross-coronavirus immunity as a correlate of protection against COVID-19.

immunology↗

bioRxiv2021

Collateral Responses to Classical Cytotoxic Chemotherapies are Heterogeneous and Sensitivities are Sparse

Chemotherapy resistance is a major obstacle to curing cancer patients. Combination drug regimens have shown promise as a method to overcome resistance; however, to date only some cancers have been cured with this method. Collateral sensitivity - the phenomenon whereby resistance to one drug is co-occurrent with sensitivity to a second drug - has been gaining traction as a promising new concept to guide rational design of combination regimens. Here we survey collateral responses to acquisition of resistance to four classical chemotherapy agents. Although collateral sensitivities have been documented for antibiotics and targeted cancer therapies, we did not observe collateral sensitivities to any of the cytotoxic agents we studied. Interestingly, we did observe heterogeneity in the phenotypic response to acquisition of resistance to each drug, suggesting the existence of multiple different states of resistance for each drug. Surprisingly, this phenotypic heterogeneity was unrelated to transcriptomic heterogeneity in the resistant cell lines. These features of phenotypic and transcriptomic heterogeneity must be taken into account in future studies of treated tumor subclones and in design of chemotherapy combinations.

cancer biology↗

bioRxiv2021

COVID-19 mRNA vaccines drive differential Fc-functional profiles in pregnant, lactating, and non-pregnant women

Significant immunological changes occur throughout pregnancy to tolerize the mother and allow growth of the fetal graft. However, additional local and systemic immunological adaptations also occur, allowing the maternal immune system to continue to protect the dyad against foreign invaders both during pregnancy and after birth through lactation. This fine balance of tolerance and immunity, along with physiological and hormonal changes, contribute to increased susceptibility to particular infections in pregnancy, including more severe COVID-19 disease. Whether these changes also make pregnant women less responsive to vaccination or induce altered immune responses to vaccination remains incompletely understood. To holistically define potential changes in vaccine response during pregnancy and lactation, we deeply profiled the humoral vaccine response in a group of pregnant and lactating women and non-pregnant age-matched controls. Vaccine-specific titers were comparable, albeit slightly lower, between pregnant and lactating women, compared to non-pregnant controls. Among pregnant women, we found higher antibody titers and functions in those vaccinated with the Moderna vaccine. FcR-binding and antibody effector functions were induced with delayed kinetics in both pregnant and lactating women compared to non-pregnant women. Antibody boosting resulted in high FcR-binding titers in breastmilk. These data point to an immune resistance to generate highly inflammatory antibodies during pregnancy and lactation, and a critical need to follow prime/boost timelines in this vulnerable population to ensure full immunity is attained.

immunology↗

bioRxiv2021

Collaboration between the Fab and Fc contribute to maximal protection against SARS-CoV-2 in nonhuman primates following NVX-CoV2373 subunit vaccine with Matrix-M™ vaccination

Recently approved vaccines have already shown remarkable protection in limiting SARS-CoV-2 associated disease. However, immunologic mechanism(s) of protection, as well as how boosting alters immunity to wildtype and newly emerging strains, remain incompletely understood. Here we deeply profiled the humoral immune response in a cohort of non-human primates immunized with a stable recombinant full-length SARS-CoV-2 spike (S) glycoprotein (NVX-CoV2373) at two dose levels, administered as a single or two-dose regimen with a saponin-based adjuvant Matrix-M. While antigen dose had some effect on Fc-effector profiles, both antigen dose and boosting significantly altered overall titers, neutralization and Fc-effector profiles, driving unique vaccine-induced antibody fingerprints. Combined differences in antibody effector functions and neutralization were strongly associated with distinct levels of protection in the upper and lower respiratory tract, pointing to the presence of combined, but distinct, compartment-specific neutralization and Fc-mechanisms as key determinants of protective immunity against infection. Moreover, NVX-CoV2373 elicited antibodies functionally target emerging SARS-CoV-2 variants, collectively pointing to the critical collaborative role for Fab and Fc in driving maximal protection against SARS-CoV-2. Collectively, the data presented here suggest that a single dose may prevent disease, but that two doses may be essential to block further transmission of SARS-CoV-2 and emerging variants. HighlightsO_LINVX-CoV2373 subunit vaccine elicits receptor blocking, virus neutralizing antibodies, and Fc-effector functional antibodies. C_LIO_LIThe vaccine protects against respiratory tract infection and virus shedding in non-human primates (NHPs). C_LIO_LIBoth neutralizing and Fc-effector functions contribute to protection, potentially through different mechanisms in the upper and lower respiratory tract. C_LIO_LIBoth macaque and human vaccine-induced antibodies exhibit altered Fc-receptor binding to emerging mutants. C_LI

immunology↗

bioRxiv2020

Quantitative phosphoproteomics uncovers dysregulated kinase networks in Alzheimer's disease

Alzheimers disease (AD) is a form of dementia characterized by amyloid-{beta} plaques and Tau neurofibrillary tangles that progressively disrupt neural circuits in the brain. The signaling networks underlying the pathological changes in AD are poorly characterized at the level phosphoproteome. Using mass spectrometry, we performed a combined analysis of the tyrosine, serine, and threonine phosphoproteome, and proteome of temporal cortex tissue from AD patients and aged matched controls. We identified several co-correlated peptide modules that were associated with varying levels of phospho-Tau, oligodendrocyte, astrocyte, microglia, and neuronal pathologies in AD patients. We observed phosphorylation sites on kinases targeting Tau as well as other novel signaling factors that were correlated with these peptide modules. Finally, we used a data-driven statistical modeling approach to identify individual peptides and co-correlated signaling networks that were predictive of AD histopathologies. Together, these results build a map of pathology-associated phosphorylation signaling events occurring in AD.

neuroscience↗

bioRxiv2020

Multiplexed relative and absolute quantitative immunopeptidomics reveals MHC I repertoire alterations induced by CDK4/6 inhibition

Peptides bound to class I major histocompatibility complexes (MHC) play a critical role in immune cell recognition and can trigger an antitumor immune response in cancer. Surface MHC levels can be modulated by anticancer agents, altering immunity. However, understanding the peptide repertoires response to treatment remains challenging and is limited by quantitative mass spectrometry-based strategies lacking robust normalization controls. We describe a novel approach that leverages recombinant heavy isotope-coded peptide MHCs (hipMHCs) and multiplex isotope tagging to quantify peptide repertoire alterations using low sample input. HipMHCs improve quantitative accuracy of peptide repertoire changes by normalizing for variation across analyses and enable absolute quantification using internal calibrants to determine copies per cell of MHC antigens, which can inform immunotherapy design. Applying this platform in melanoma to profile the immunopeptidome response to CDK4/6 inhibition and interferon gamma, known modulators of antigen presentation, we uncovered treatment-specific alterations, connecting the intracellular response to extracellular immune presentation.

immunology↗

bioRxiv2019

An Inter-Species Translation Model Implicates Integrin Signaling in Infliximab-Resistant Colonic Crohn’s Disease

Anti-TNF therapy resistance is a major clinical challenge in Crohns Disease (CD), partly due to insufficient understanding of disease-site, protein-level mechanisms of CD and anti-TNF treatment resistance. Although some proteomics data from CD mouse models exists, data type and phenotype discrepancies contribute to confounding attempts to translate between preclinical animal models of disease and human clinical cohorts. To meet this important challenge, we develop and demonstrate here an approach called Translatable Components Regression (TransComp-R) to overcome inter-species and trans-omic discrepancies between CD mouse models and human subjects. TransComp-R combines CD mouse model proteomic data with patient pre-treatment transcriptomic data to identify molecular features discernable in the mouse data predictive of patient response to anti-TNF therapy. Interrogating the TransComp-R models predominantly revealed upregulated integrin pathway signaling via collagen-binding integrin ITGA1 in anti-TNF resistant colonic CD (cCD) patients. Toward validation, we performed single-cell RNA sequencing on biopsies from a cCD patient and analyzed publicly available immune cell proteomics data to characterize the immune and intestinal cell types contributing to anti-TNF resistance. We found that ITGA1 is indeed expressed in colonic T-cell populations and that interactions between collagen-binding integrins on T-cells and colonic cell types expressing secreted collagens are associated with anti-TNF therapy resistance. Biologically, TransComp-R linked previously disparate observations about collagen and ITGA1 signaling to a potential therapeutic avenue for overcoming anti-TNF therapy resistance in cCD. Methodologically, TransComp-R provides a flexible, generalizable framework for addressing inter-species, inter-omic, and inter-phenotypic discrepancies between animal models and patients to deliver translationally relevant biological insights.\n\nOne Sentence SummaryBrubaker et al. implicate dysregulated collagen-binding integrin signaling in resistance to anti-TNF therapy in Crohns Disease by developing a mouse-proteomic to human-transcriptomic translation model and confirm the associated inter-cellular signaling network using single-cell RNA sequencing.

systems biology↗

bioRxiv2019

Coagulopathy signature precedes and predicts severity of end-organ heat stroke pathology in a mouse model

Heat stroke is a life-threatening condition characterized by loss of thermoregulation and severe elevation of core body temperature, which can cause organ failure and damage to the central nervous system. While no definitive test exists to measure heat stroke severity, immune challenge is known to increase heat stroke risk, although the mechanism of this increased risk is unclear. In this study, we used a mouse model of classic heat stroke to test the effect of immune challenge on pathology. Employing multivariate supervised machine learning to identify patterns of molecular and cellular markers associated with heat stroke, we found that prior viral infection simulated with poly I:C injection resulted in heat stroke presenting with high levels of factors indicating coagulopathy. Despite a decreased number of platelets in the blood, platelets are large and non-uniform in size, suggesting younger, more active platelets. Levels of D-dimer and soluble thrombomodulin were increased in more severe heat stroke, and in cases presenting with the highest level of organ damage markers D-dimer levels dropped, indicating potential fibrinolysis-resistant thrombosis. Genes corresponding to immune response, coagulation, hypoxia, and vessel repair were up-regulated in kidneys of heat-challenged animals, and these increases correlated with both viral treatment and distal organ damage while appearing before discernible tissue damage to the kidney itself. We conclude that heat stroke-induced coagulopathy may be a driving mechanistic force in heat stroke pathology, especially when exacerbated by prior infection, and that coagulation markers may serve as an accessible biomarker for heat stroke severity and therapeutic strategies.\n\nKey pointsO_LIA signature of pro-coagulation markers predicts circadian core body temperature and levels of organ damage in heat stroke\nC_LIO_LIChanges in coagulopathy-related gene expression are evidenced before histopathological organ damage\nC_LI

physiology↗

bioRxiv2019

Sporadic ERK pulses drive non-genetic resistance in drug-adapted BRAFV600E melanoma cells

Anti-cancer drugs commonly target signal transduction proteins activated by mutation. In patients with BRAFV600E melanoma, small molecule RAF and MEK kinase inhibitors cause dramatic but often transient tumor regression. Emerging evidence suggests that cancer cells adapting by non-genetic mechanisms constitute a reservoir for the development of drug-resistant tumors. Here, we show that few hours after exposure to RAF/MEK inhibitors, BRAFV600E melanomas undergo adaptive changes involving disruption of negative feedback and sporadic pulsatile reactivation of the MAPK pathway, so that MAPK activity is transiently high enough in some cells to drive proliferation. Quantitative proteomics and computational modeling show that pulsatile MAPK reactivation is possible due to the co-existence in cells of two MAPK cascades: one driven by BRAFV600E that is drug-sensitive and a second driven by receptors that is drug-resistant. Paradoxically, this may account both for the frequent emergence of drug resistance and for the tolerability of RAF/MEK therapy in patients.

cancer biology↗

bioRxiv2019

Robustness and applicability of functional genomics tools on scRNA-seq data

Many tools have been developed to extract functional and mechanistic insight from bulk transcriptome profiling data. With the advent of single-cell RNA sequencing (scRNA-seq), it is in principle possible to do such an analysis for single cells. However, scRNA-seq data has characteristics such as drop-out events, low library sizes and a comparatively large number of samples/cells. It is thus not clear if functional genomics tools established for bulk sequencing can be applied to scRNA-seq in a meaningful way. To address this question, we performed benchmark studies on in silico and in vitro single-cell RNA-seq data. We included the bulk-RNA tools PROGENy, GO enrichment and DoRothEA that estimate pathway and transcription factor (TF) activities, respectively, and compared them against the tools AUCell and metaVIPER, designed for scRNA-seq. For the in silico study we simulated single cells from TF/pathway perturbation bulk RNA-seq experiments. Our simulation strategy guarantees that the information of the original perturbation is preserved while resembling the characteristics of scRNA-seq data. We complemented the in silico data with in vitro scRNA-seq data upon CRISPR-mediated knock-out. Our benchmarks on both the simulated and real data revealed comparable performance to the original bulk data. Additionally, we showed that the TF and pathway activities preserve cell-type specific variability by analysing a mixture sample sequenced with 13 scRNA-seq different protocols. Our analyses suggest that bulk functional genomics tools can be applied to scRNA-seq data, outperforming dedicated single cell tools. Furthermore we provide a benchmark for further methods development by the community.

bioinformatics↗

bioRxiv2019

C. difficile-associated antibiotics prime the host for infection by a microbiome-independent mechanism.

The most clinically relevant risk factor for Clostridioides difficile-associated disease (CDAD) is recent antibiotic treatment. Though most broad-spectrum antibiotics significantly disrupt the structure of the gut microbiota, only particular ones increase CDAD risk, suggesting additional factors might increase the risk from certain antibiotics. Here we show that commensal-independent effects of antibiotics collectively prime an in vitro germ-free human gut for CDAD. We found a marked loss of mucosal barrier and immune function with CDAD-associated antibiotic pretreatment distinct from pretreatment with an antibiotic unassociated with CDAD, which did not reduce innate immune or mucosal barrier functions. Importantly, pretreatment with CDAD-associated antibiotics sensitized mucosal barriers to C. difficile toxin activity in primary cell-derived enteroid monolayers. These data implicate commensal-independent host changes in the increased risk of CDAD with specific antibiotics. Our findings are contrary to the previously held belief that antibiotics allow for CDAD solely through disruption of the microbiome. We anticipate this work to suggest potential avenues of research for host-directed treatment and preventive therapies for CDAD, and to impact human tissue culturing protocols.

molecular biology↗

bioRxiv2019

Pancreatic Stellate Cells Secrete Deoxycytidine Conferring Resistance to Gemcitabine in PDAC

Pancreatic ductal adenocarcinoma (PDAC) is a leading cause of cancer deaths in the United States. The deoxynucleoside analog gemcitabine is among the most effective therapies to treat PDAC; however, nearly all patients treated with gemcitabine either fail to respond or rapidly develop resistance. One hallmark of PDAC is a striking accumulation of stromal tissue surrounding the tumor, and this accumulation of stroma can contribute to therapy resistance. To better understand how stroma limits response to therapy, we investigated cell-extrinsic mechanisms of resistance to gemcitabine. We show that conditioned media from pancreatic stellate cells (PSC), as well as from other fibroblasts, protects PDAC cells from gemcitabine toxicity. We find that the PSC conditioned media protective effect is mediated by secretion of deoxycytidine, but not other deoxynucleosides, through equilibrative nucleoside transporters. Deoxycytidine inhibits the processing of gemcitabine in PDAC cells, thus reducing the effect of gemcitabine and other nucleoside analogs on cancer cells. Our results suggest that reducing deoxycytidine production in PSCs may increase the efficacy of nucleoside analog therapies. Additional InformationFunding: This project was funded in part by the NIH (NCI U54-217377), the MIT Center for Precision Cancer Medicine, and by the Koch Institute Support (core) Grant P30-CA14051 from the National Cancer Institute. S.D. was supported by the David H. Koch Fellowship in Cancer Research. A.N.L was a Robert Black Fellow of the Damon Runyon Cancer Research Foundation, DRG-2241-15 and was supported by a NIH Pathway to Independence Award (K99/R00), 1K99CA234221. M.T.H and M.G.V.H. acknowledges funding from the MIT Center for Precision Cancer Medicine and the Ludwig Center at MIT. M.G.V.H also acknowledges funding from the Lustgarten Foundation, SU2C, the MIT Center for Precision Cancer Medicine, the NCI, and an HHMI Faculty Scholar award. Competing interests: M.G.V.H. is a consultant and advisory board member for Agios Pharmaceuticals, Aeglea Biotherapeutics, and Auron Therapeutics.

cancer biology↗

bioRxiv2018

A framework for translation of genomic responses from mouse models to human inflammatory disease contexts

The high failure rate of therapeutics showing promise in mouse disease models to translate to patients is a pressing challenge in biomedical science. However, mouse models are a useful tool for evaluating mechanisms of disease and prioritizing novel therapeutic agents for clinical trials. Though retrospective studies have examined the fidelity of mouse models of inflammatory disease to their respective human in vivo conditions, approaches for prospective translation of insights from mouse models to patients remain relatively unexplored. Here, we develop a semi-supervised learning approach for prospective inference of disease-associated human in vivo differentially expressed genes and pathways from mouse model experiments. We examined 36 transcriptomic case studies where comparable phenotypes were available for mouse and human inflammatory diseases and assessed multiple computational approaches for inferring human in vivo biology from mouse model datasets. We found that a semi-supervised artificial neural network identified significantly more true human in vivo associations than interpreting mouse experiments directly (95% CI on F-score for mouse experiments [0.090, 0.175], neural network [0.278, 0.375], p = 0.00013). Our study shows that when prospectively evaluating biological associations in mouse studies, semi-supervised learning approaches combining mouse and human data for biological inference provides the most accurate assessment of human in vivo disease and therapeutic mechanisms. The task of translating insights from model systems to human disease contexts may therefore be better accomplished by the use of systems modeling driven approaches.\n\nAuthor SummaryComparison of genomic responses in mouse models and human disease contexts is not sufficient for addressing the challenge of prospective translation from mouse models to human disease contexts. Here, we address this challenge by developing a semi-supervised machine learning approach that combines supervised modeling of mouse experiment datasets with unsupervised modeling of human disease-context datasets to predict human in vivo differentially expressed genes and pathways as if the model system experiment had been run in the human cohort. A semi-supervised version of a feed forward artificial neural network was the most efficacious model for translating experimentally derived mouse molecule-phenotype associations to the human in vivo disease context. We find that computational generalization of signaling insights from mouse to human contexts substantially improves upon direct generalization of mouse experimental insights and argue that such approaches can facilitate more clinically impactful translation of insights from preclinical studies in model systems to patients.

systems biology↗