Biology subjects

Feng, J.

Publications and source records attributed to Feng, J..

At least 253 records · Page 14Linked to original sources

bioRxiv2019

Predicting Alzheimer’s disease progression using deep recurrent neural networks

Early identification of individuals at risk of developing Alzheimers disease (AD) dementia is important for developing disease-modifying therapies. In this study, given multimodal AD markers and clinical diagnosis of an individual from one or more timepoints, we seek to predict the clinical diagnosis, cognition and ventricular volume of the individual for every month (indefinitely) into the future. We proposed and applied a minimal recurrent neural network (minimalRNN) model to data from The Alzheimers Disease Prediction Of Longitudinal Evolution (TADPOLE) challenge, comprising longitudinal data of 1677 participants (Marinescu et al. 2018) from the Alzheimers Disease Neuroimaging Initiative (ADNI). We compared the performance of the minimalRNN model and four baseline algorithms up to 6 years into the future. Most previous work on predicting AD progression ignore the issue of missing data, which is a prevalent issue in longitudinal data. Here, we explored three different strategies to handle missing data. Two of the strategies treated the missing data as a "preprocessing" issue, by imputing the missing data using the previous timepoint ("forward filling") or linear interpolation ("linear filling). The third strategy utilized the minimalRNN model itself to fill in the missing data both during training and testing ("model filling"). Our analyses suggest that the minimalRNN with "model filling" compared favorably with baseline algorithms, including support vector machine/regression, linear state space (LSS) model, and long short-term memory (LSTM) model. Importantly, although the training procedure utilized longitudinal data, we found that the trained minimalRNN model exhibited similar performance, when using only 1 input timepoint or 4 input timepoints, suggesting that our approach might work well with just cross-sectional data. An earlier version of our approach was ranked 5th (out of 53 entries) in the TADPOLE challenge in 2019. The current approach is ranked 2nd out of 63 entries as of June 3rd, 2020.

neuroscience↗

bioRxiv2019

Dynamic brain network changes reflect neuroplasticity: molecular and cognitive evidence

Resting-state functional brain networks demonstrate dynamic changes on the scale of seconds. However, their genetic mechanisms and profound cognitive relevance remain less explored. We identified 459 Bonferroni-corrected genes, by associating temporal variability of regional functional connectivity patterns with Allen Brain gene expression profiles across the whole brain. These genes are partially verified in developing human brain gene expression in the BrainSpan Atlas, and are found to be involved in the enrichment of short- and long-term plasticity processes. The former process depends on synaptic plasticity, involving ion transmembrane transport, action potential propagation, and modulation. The latter process depends on structural plasticity, including axonal genesis, development, and guidance. Results from a longitudinal cognitive training study further revealed that baseline variability of the hippocampal network predicted cognitive ability changes after three months of training. Our genetic association results suggest that the short-term plasticity processes may account for the rapid changes of regional functional connectivity, while the underlying long-term plasticity processes explain why temporal variability can predict long-term learning outcomes. To our knowledge, this is the first demonstration that measuring the dynamic brain network can lead to a non-invasive quantification of neuroplasticity in humans.

neuroscience↗

bioRxiv2019

Ranking of Major Classes of Antibiotics for Activity against Stationary Phase Pseudomonas aeruginosa and Identification of Clinafloxacin + Cefuroxime + Gentamicin Drug Combination that Eradicates Persistent P. aeruginosa Infection in a Murine Cystic Fibrosis Model

Pseudomonas aeruginosa can cause serious persistent infections such as ventilator-associated pneumonia, sepsis, biofilm-related infections as in cystic fibrosis (CF) patients. Although CF lung infections can be treated with antibiotics, full clearance is difficult due to P. aeruginosa persistence. While antibiotic activity against growing P. aeruginosa is well documented, their activity against the non-growing persisters enriched in stationary phase cultures has not been well studied. Here, we systematically evaluated and ranked the six major classes of antibiotics, cell wall and cell membrane inhibitors, protein synthesis inhibitors, DNA synthesis inhibitors, RNA synthesis inhibitors, sulfa drugs, and nitrofurantoin, for their activity against both growing and persister forms of P. aeruginosa using colony forming count (CFU) and SYBR Green I/Propidium Iodide (PI) viability assay. Among the six major classes of antibiotics, cell wall and cell membrane inhibitors (Cefuroxime and Colistin), DNA synthesis inhibitors (Clinafloxacin) and sulfa drugs (Sulfamethoxazole) had good activity against stationary phase cells. In contrast, protein synthesis inhibitors (Gentamicin), RNA synthesis inhibitor (Rifampicin) and Nitrofurantoin had relatively poor activity against the stationary phase P. aeruginosa but relatively high activity against log phase P. aeruginosa. Clinafloxacin is the only single drug that could completely kill all (109 CFU) stationary phase cells in a 4 day drug exposure. The Cefuroxime + Gentamicin+ Clinafloxacin combination could kill all biofilm bacteria in 2 days whereas the clinically used drug combination Cefuroxime + Gentamicin + Colistin only partially killed the biofilm bacteria with 103 CFU remaining. In a murine persistent CF lung infection model, only Cefuroxime + Gentamicin+ Clinafloxacin cleared all bacteria in the infected lungs, whereas Clinafloxacin alone, or Cefuroxime + Clinafloxacin, or the current recommended drug combination Cefuroxime + Gentamicin, all failed to completely clear the bacterial load in the lungs. The complete sterilization of the bacterial load is a property of Clinafloxacin combination, as Cefuroxime + Gentamicin+ Levofloxacin combination was unable to clear the bacterial load in the lungs. Our findings demonstrate the importance of persister drug clinafloxacin, offer new therapeutic approaches for more effective treatment of persistent P. aeruginosa infections, and may have implications for treating other persistent infections.

microbiology↗

bioRxiv2019

A Drug Combination Approach Targeting Both Growing Bacteria and Dormant Persisters Eradicate Persistent Staphylococcus aureus Biofilm Infection

Staphylococcus aureus can cause a variety of infections, many of which involve biofilm infections. Inside biofilms, growing and non-growing bacteria such as persisters co-exist, making it challenging to completely eradicate a persistent and recurrent infection with current treatments. Despite the clinical relevance, most of the current antibiotic treatments mainly kill the growing bacteria and have poor activity against non-growing persister bacteria and thus have limited effect on treating persistent infections including biofilm infections. We previously proposed a Yin-Yang model using a drug combination approach targeting both growing bacteria and persister bacteria for more effective clearance of persistent infections. Here, as a proof of principle, we showed that combining drugs that have high activity against growing forms, such as vancomycin or meropenem, with drugs that have robust anti-persister activity, such as clinafloxacin and oritavancin, could completely eradicate S. aureus biofilm bacteria in vitro. In contrast, single or two drugs including the current treatment for persistent S. aureus infection doxycycline plus rifampin failed to kill all biofilm bacteria in vitro. We then developed a chronic persistent skin infection mouse model with biofilm-seeded bacterial inocula demonstrating that biofilm bacteria caused more severe and persistent skin lesions than log phase S. aureus bacteria. More importantly, we found that the drug combination which eradicated biofilm bacteria in vitro is more efficacious than current treatments and completely eradicated S. aureus biofilm infection in mice. The complete eradication of biofilm bacteria is attributed to the unique high anti-persister activity of clinafloxacin, which could not be replaced by other fluoroquinolones such as moxifloxacin, levofloxacin or ciprofloxacin. Our study is the first to demonstrate that the combination of meropenem, daptomycin, plus clinafloxacin completely cleared the persistent infection, healed the lesions, and had less inflammation, while mice treated with doxycycline plus rifampin, the current clinically recommended treatment for chronic tissue infection, failed to do so. We also compared our persister drug combination with other approaches for treating persistent infections including gentamicin+fructose and ADEP4+rifampin in the S. aureus biofilm infection mouse model. Neither gentamicin+fructose nor ADEP4+rifampin could eradicate or cure the persistent biofilm infection in mice. In contrast, our drug combination regimen with persister drug clinafloxacin plus meropenem and daptomycin completely eradicated and cured the persistent biofilm infection in 7 days. An unexpected observation is that ADEP4 treatment group developed worsened skin lesions and caused more extensive pathology than the untreated control mice. Our study demonstrates an important treatment principle for persistent infections by targeting both growing and non-growing heterogeneous bacterial populations utilizing persister drugs for more effective eradication of persistent and biofilm infections. Our findings may have implications for improved treatment of many other persistent infections in general.

microbiology↗

bioRxiv2019

Efficient Long-Distance Transport in Upper Fluid of Bacterial Swarms

Flagellated bacteria move collectively in a swirling pattern on agar surfaces immersed in a thin layer of viscous \"swarm fluid\", but the role of this fluid in mediating the cooperation of the bacterial population is not well understood. Herein, we use gold nanorods (AuNRs) as single particle tracers to explore the spatiotemporal structure of the swarm fluid. We observed that individual AuNRs are transported in a plane of ~2 m above the motile cells. They can travel for long distances (>700 m) in a 2D plane at high speed (often >50 m2/s) without interferences from bacterial movements. The particles are apparently lifted up and transported by collective mixing of the small vortices around bacteria during localized clustering and de-clustering of the motile cells, exhibiting superdiffusive and non-Gaussian characteristics with alternating large-step jumps and confined lingering. Their motions are consistent with the Levy walk (LW) model, revealing efficient transport flows above swarms. These flows provide obstacle-free highways for long-range material transportations, shed light on how swarming bacteria perform population-level communications, and reveal the essential role of the fluid phase on the emergence of large-scale synergy. This approach is promising for probing complex fluid dynamics and transports in other collective systems.

microbiology↗

bioRxiv2019

Evaluation of Natural and Botanical Medicines for Activity against Growing and Non-growing Forms of B. burgdorferi

Lyme disease is the most common vector-borne disease in the US. Although the current recommended Lyme antibiotic treatment can cure the majority of Lyme disease patients, about 10-20% patients continue to suffer from persisting symptoms. There have been various anecdotal reports on the use of herbal extracts for treating patients with persisting symptoms with varying degree of improvements. However, it is unclear whether the effect of the herb products is due to their direct antimicrobial activity or their effect on host immune system. In the present study, we investigated the antimicrobial effects of 12 commonly used botanical medicines and 3 other natural antimicrobial agents for potential anti-Borrelia burgdorferi activity in vitro. Primary criteria for selecting compounds for the present study included agents that had shown significant anti-borrelial effects in previous studies, have favorable safety profiles, and can be absorbed systemically. Among them, 9 natural product extracts at 1% were found to have good activity against the stationary phase B. burgdorferi culture compared to the control antibiotics doxycycline and cefuroxime. These active herbs include Cryptolepis sanguinolenta, Juglans nigra (Black walnut), Polygonum cuspidatum (Japanese knotweed), Artemesia annua (Sweet wormwood), Uncaria tomentosa (Cats claw), Cistus incanus, and Scutellaria baicalensis (Chinese skullcap). In contrast, Stevia rebaudiana, Andrographis paniculata, Grapefruit seed extract, colloidal silver, monolaurin, and antimicrobial peptide LL37 had little or no activity against stationary phase B. burgdorferi. The minimum inhibitory concentration (MIC) values of Artemesia annua, Juglans nigra, and Uncaria tomentosa were quite high for growing B. burgdorferi, despite their strong activity against the non-growing stationary phase B. burgdorferi cells. On the other hand, the top two active herbs, Cryptolepis sanguinolenta and Polygonum cuspidatum, showed strong activity against both growing B. burgdorferi (MIC=0.03%-0.06% and 0.25%-0.5% respectively) and non-growing stationary phase B. burgdorferi. In subculture studies, only 1% Cryptolepis sanguinolenta extract caused complete eradication, while current Lyme antibiotics doxycycline and cefuroxime and other active herbs including Polygonum cuspidatum, Artemesia annua, Juglans nigra and Uncaria tomentosa could not eradicate B. burgdorferi stationary phase cells as many spirochetes were visible after 21-day subculture. Further studies are needed to identify the active ingredients of the effective herbs and evaluate their combinations for more effective eradication of B. burgdorferi in vitro and in vivo. The implications of these findings for more effective treatment of persistent Lyme disease are discussed.

microbiology↗

bioRxiv2019

iRhom2 serves as a facilitator in obesity by enhancing adipose inflammation and insulin resistance

Chronic inflammation of adipose tissues contributes to obesity-triggered insulin resistance. Unfortunately, the potential molecular mechanisms regarding obesity associated systemic inflammation and metabolic disorder remain complicated. Here we display that inactive rhomboid-like protein 2 (iRhom2) is increased in mice fat with adipose inflammation. After 16 weeks on a high fat diet (HFD), obesity, chronic inflammation in adipose tissues and insulin resistance are markedly mitigated in iRhom2-knockout (iRhom2 KO) mice, but exaggerated in iRhom2-overactivated mice. The adverse impressions of iRhom2 on adipose inflammation and associated pathologies are determined in db/db mice. Also, we further exhibit that in response to HFD, iRhom2 KO mice and mice with deletion only in myeloid cells showed less severe adipose inflammation and insulin resistance than the control groups. Conversely, transplantation of bone marrow cells from normal mice to iRhom2 KO mice unleashed the severity of systemic inflammation and metabolic dysfunction after HFD ingestion. In conclusion, we identify iRhom2 as a key regulator that promotes obesity-associated metabolic disorder. Loss of iRhom2 from macrophages in adipose tissues inhibited the inflammation and insulin resistance. iRhom2 might be a therapeutic target for obesity-induced metabolic dysfunction.\n\nSignificanceIncreased inactive rhomboid-like protein 2 signaling has recently been shown to trigger inflammation-associated activation of innate immune responses. Herein we investigate that this signal also plays a crucial role in obesity-triggered adipose tissue inflammation infiltration and metabolic disorder, beyond the well-known assignment in innate immune supervision. Also, we have reported the iRhom2 as a key promoter in regulating metabolic function, which enhances obesity-stimulated inflammation and systemic insulin resistance by up regulation of macrophages pro-inflammatory activation. Our current study indicates that targeting the iRhom2 signaling in adipose tissues could possibly be an efficient strategy to mitigating obesity-associated systemic inflammation and metabolic dysfunction.

pathology↗

bioRxiv2019

Printability study of metal ion crosslinked PEG-catechol based inks

Inspired by reversible networks present in nature, we have explored the printability of catechol functionalized polyethylene glycol (PEG) based inks with metal-coordination crosslinking. Material formulations containing Al3+, Fe3+ or V3+ as crosslinking ions were tested. The printability and shape fidelity were dependent on the ink composition (metal ion type, pH, PEG molecular weight) and printing parameters (extrusion pressure and printing speed). The relaxation time, recovery rate and viscosity of the inks were analyzed in rheology studies and correlated with thermodynamic and ligand exchange kinetic constants of the dynamic bonds and the printing performance (i.e. shape fidelity of the printed structures). The relevance of the relaxation time and ligand exchange kinetics for printability was demonstrated. Cells seeded on the crosslinked materials were viable, indicating the potential of the formulations to be used as inks for cell encapsulation. The proposed dynamic ink design offers significant flexibility for 3D (bio)printing, and enables straightforward adjustment of the printable formulation to meet application-specific needs.

biochemistry↗

bioRxiv2019

Estimation of cell lineage trees by maximum-likelihood phylogenetics

CRISPR technology has enabled large-scale cell lineage tracing for complex multicellular organisms by mutating synthetic genomic barcodes during organismal development. However, these sophisticated biological tools currently use ad-hoc and outmoded computational methods to reconstruct the cell lineage tree from the mutated barcodes. Because these methods are agnostic to the biological mechanism, they are unable to take full advantage of the datas structure. We propose a statistical model for the mutation process and develop a procedure to estimate the tree topology, branch lengths, and mutation parameters by iteratively applying penalized maximum likelihood estimation. In contrast to existing techniques, our method estimates time along each branch, rather than number of mutation events, thus providing a detailed account of tissue-type differentiation. Via simulations, we demonstrate that our method is substantially more accurate than existing approaches. Our reconstructed trees also better recapitulate known aspects of zebrafish development and reproduce similar results across fish replicates.

bioinformatics↗

bioRxiv2019

Augmented Interval List: a novel data structure for efficient genomic interval search

MotivationGenomic data is frequently stored as segments or intervals. Because this data type is so common, interval-based comparisons are fundamental to genomic analysis. As the volume of available genomic data grows, developing efficient and scalable methods for searching interval data is necessary.\n\nResultsWe present a new data structure, the augmented interval list (AIList), to enumerate intersections between a query interval q and an interval set R. An AIList is constructed by first sorting R as a list by the interval start coordinate, then decomposing it into a few approximately flattened components (sublists), and then augmenting each sublist with the running maximum interval end. The query time for AIList is O(log2N + n + m), where n is the number of overlaps between R and q, N is the number of intervals in the set R, and m is the average number of extra comparisons required to find the n overlaps. Tested on real genomic interval datasets, AIList code runs 5 - 18 times faster than standard high-performance code based on augmented interval-trees (AITree), nested containment lists (NCList), or R-trees (BEDTools). For large datasets, the memory-usage for AIList is 4% - 60% of other methods. The AIList data structure, therefore, provides a significantly improved fundamental operation for highly scalable genomic data analysis.\n\nAvailabilityAn implementation of the AIList data structure with both construction and search algorithms is available at code.databio.org/AIList.

bioinformatics↗

bioRxiv2019

Identification of Genes Regulating Cell Death in Staphylococcus aureus

Staphylococcus aureus is an opportunistic pathogen that causes acute and chronic infections. Due to S. aureus s highly resistant and persistent nature, it is paramount to identify better drug targets in order to eradicate S. aureus infections. Despite the efforts in understanding bacterial cell death, the genes and pathways of S. aureus cell death remain elusive. Here, we performed a genome-wide screen using a transposon mutant library to study the genetic mechanisms involved in S. aureus cell death. Using a precisely controlled heat-ramp and acetic acid exposure assays, mutations in 27 core genes (hsdR1, hslO, nsaS, sspA, folD, mfd, vraF, kdpB, USA300HOU_2684, 0868, 0369, 0420, 1154, 0142, 0930, 2590, 0997, 2559, 0044, 2004, 1209, 0152, 2455, 0154, 2386, 0232, 0350 involved in transporters, transcription, metabolism, peptidases, kinases, transferases, SOS response, nucleic acid and protein synthesis) caused the bacteria to be more death-resistant. In addition, we identified mutations in core 10 genes (capA, gltT, mnhG1,USA300HOU_1780, 2496, 0200, 2029, 0336, 0329, 2386, involved in transporters, metabolism, transcription, cell wall synthesis) from heat-ramp and acetic acid that caused the bacteria to be more death-sensitive or with defect in persistence. Interestingly, death-resistant mutants were more virulent than the parental strain USA300 and caused increased mortality in a Caenorhabditis elegans infection model. Conversely, death-sensitive mutants were less persistent and formed less persister cells upon exposure to different classes of antibiotics. These findings provide new insights into the mechanisms of S. aureus cell death and offer new therapeutic targets for developing more effective treatments caused by S. aureus.

microbiology↗

bioRxiv2018

The allotetraploid origin and asymmetrical genome evolution of common carp Cyprinus carpio

Common carp (Cyprinus carpio) is an allotetraploid Cyprinid species derived from recent whole genome duplication and provides an excellent model system for studying polyploid genome evolution in vertebrates. To explore the origins and consequences of tetraploidy in C. carpio, we generated three chromosome-level new reference genomes of C. carpio and compared them to the related diploid Cyprinid genome sequences. We identified a progenitor-like diploid Barbinae lineage by analysing the phylogenetic relationship of the homoeologous genes of C. carpio and their orthologues in closely related diploid Cyprinids. We then characterized the allotetraploid origin of C. carpio and divided its genome into two homoeologous subgenomes that are marked by a distinct genome similarity to their diploid progenitor. On the basis of the divergence rates of homoeologous genes and transposable elements in two subgenomes, we estimated that the two diploid progenitor species diverged approximately 23 million years ago (Mya) and merged to form the allotetraploid C. carpio approximately 12.4 Mya, which likely correlated with environmental upheavals caused by the extensive uplift of the Qinghai-Tibetan Plateau. No large-scale gene losses or rediploidization were observed in the two subgenomes. Instead, we found extensive homoeologous gene expression bias across twelve surveyed tissues, which indicates that subgenome B is dominant in homoeologous expression. DNA methylation analysis suggested that CG methylation in promoter regions plays an important role in altering the expression of these homoeologous genes in allotetraploid C. carpio. This study provides an essential genome resource and insights for extending further investigation on the evolutionary consequences of vertebrate polyploidy.

genomics↗

bioRxiv2018

Temporal modulation of host aerobic glycolysis determines the outcome of M. marinum infection

Macrophages are the first-line host defense where the invading Mycobacterium tuberculosis (Mtb) encounters. It has been recently reported that host aerobic glycolysis was elevated post the infection by a couple of virulent mycobacterial species. However, whether this metabolic transition is required for host defense against intracellular pathogens and the underlying mechanisms remain to be further investigated. By analyzing carbon metabolism, we found that macrophages infected by M. marinum, a surrogate mycobacterial specie to Mtb, showed a strong elevation of glycolysis. Next, three glycolysis inhibitors were examined for their ability to inhibit mycobacterial proliferation inside RAW264.7, a murine macrophage-like cell line. Among them, a glucose analog, 2-deoxyglucose (2-DG) displayed a protective effect on assisting host to resist mycobacterial infection, which was further validated in zebrafish-infection model. The phagocytosis of M. marinum was significantly decreased in macrophages pre-treated with 2-DG at concentrations of 0.5 and 1 mM, at which no inhibitory effect was posed on M. marinum growth in vitro. Moreover, 2-DG pre-treatment exerted a significant protective effect on zebrafish larvae to limit the proliferation of M. marinum, and such effect was correlated to tumor necrosis factor alpha (TNF-). On the contrary, the 2-DG treatment post infection did not restrain proliferation of M. marinum in WT zebrafish, and even accelerated bacterial replication in TNF--/- zebrafish. Together, modulation of glycolysis prior to infection boosts host immunity against M. marinum infection, indicating a potential intervention strategy to control mycobacterial infection.\n\nAuthor SummaryAs an intracellular pathogen, Mtb exploits multiple strategies to invade and hijack macrophages for its own advantages. Accordingly, recent investigations have shown that Mtb infection is accompanied with an alteration of host glucose metabolism. Macrophage and zebrafish infection models of M. marinum, facilitating our understanding towards mycobacterial pathogenesis, were applied in this study. We found that the pre-treatment of macrophages with a glucose analog, 2-DG, inhibited aerobic glycolysis and made host cells more inert to phagocytose the bud. In infected zebrafish larvae, bacterial load inside host pretreated with 2-DG remains at a significantly lower level compared to the untreated group. These findings imply that the modulation of host glycolysis regulates the fate of M. marinum infection, and indicate a promising metabolic target in TB intervention.

microbiology↗

bioRxiv2018

Whole-tree heartwood inducing technique for Dalbergia odorifera

The dried heartwood of Dalbergia odorifera is a very important traditional Chinese medicine. In nature, D. odorifera heartwood forms slowly over several decades, and excessive harvesting has caused large-scale reductions of its wild populations. To date, there have been no studies on artificial methods to induce heartwood formation in trees that have not already begun producing heartwood. We have published a patent for a whole-tree heartwood inducing technique for D. odorifera (the \"Dodor-Wit\" method); here, we describe and analyze its efficacy. Trees in four plantations in China were treated using this method. After two years, in 2016 and 2017, they were harvested and analyzed. Average heartwood yield per tree was 3,161.95 g. Heartwood induction rate (dry heartwood weight to total dry trunk weight after peeling) was 21.60%. Average volatile oil content per tree was 1.42%. Alcohol-soluble extract of the induced heartwood was 15.77%, exceeding the Chinese Pharmacopoeia standard. This is the first report on the Dodor-Wit method and heartwood induction in D. odorifera trees aged 5-8 years. The heartwood induced by Dodor-Wit satisfies the requirements of the Chinese Pharmacopoeia. The Dodor-Wit method has an important role in solving the extreme scarcity of Chinese medicinal herbs.

plant biology↗

bioRxiv2018

Proteomic Analyses of Morphological Variants of Borrelia burgdorferi Shed New Light on Persistence Mechanisms: Implications for Pathogenesis, Diagnosis and Treatment

Borrelia burgdorferi causes Lyme disease, which is the most common vector borne disease in the United States and Europe. Although 2-4 week antibiotic treatment for Lyme disease is effective in the majority of cases, about 10-20% patients suffer from prolonged post-treatment Lyme disease syndrome (PTLDS). While the mechanisms of PTLDS are unclear, persisting organisms not killed by current Lyme antibiotics has been suggested as a possible explanation. B. burgdorferi can spontaneously develop different morphological variant forms under stress or in stationary phase with increased persistence to antibiotics. To shed light on the possible mechanisms by which these variant forms develop persistence, here, we isolated three B. burgdorferi forms, log phase spirochetal form, stationary phase planktonic form, and stationary phase aggregated biofilm-like microcolony form. We showed that the two separated stationary phase forms especially microcolony form have more persistence to antibiotics than the log phase spirochetal form. Then, we performed mass spectrometry (MS/MS) analysis to determine the proteomic profiles of the three different forms to reveal the mechanisms of persistence in B. burgdorferi. We identified 1023 proteins in the three B. burgdorferi forms, with 642 proteins (63%) differentially expressed. Compared with the log phase spirochetal form of B. burgdorferi, a total of 143 proteins were upregulated in both stationary phase planktonic form and microcolony form. Among these common upregulated proteins, 90 proteins had predicted functions and were mapped to different pathways involved in infection and virulence, DNA repair, heat shock, transport, sporulation, cell envelope and metabolism, many of which are consistent with persister mechanisms in other bacteria. A particularly interesting observation is that infection and virulence related proteins are highly up-regulated in stationary phase planktonic form and microcolony form compared with log phase spirochetal form. These findings shed new light on the mechanisms of B. burgdorferi persistence and offer novel targets for developing more effective diagnostics, vaccines and treatments.

microbiology↗

bioRxiv2018

Infection with Persister Forms of Staphylococcus aureus Causes a Chronic Persistent Skin Infection with More Severe Lesion that Takes Longer to Heal and is not Eradicated by the Current Recommended Treatment in Mice

Staphylococcus aureus is an opportunistic pathogen that can cause persistent infections clinically. Treatment for chronic S. aureus infections ranges from at least one week to several months and such infections are prone to relapse likely due to the presence of persistent forms of bacteria such as persister cells. Persister cells, which are bacterial cells that become dormant under stress conditions, can be isolated in vitro but their clinical significance in in vivo infections are largely unclear. Here, we evaluated S. aureus persistent forms using stationary phase cultures and biofilm bacteria (enriched in persisters) in comparison with log phase cultures in terms of their ability to cause disease in a mouse skin infection model. Surprisingly, we found that infection of mice with stationary phase cultures and biofilm bacteria produced a more severe chronic skin infection with more pronounced lesions which took longer to heal than log phase (actively growing) cultures. After two week infection, the bacterial load and skin tissue pathology, as determined by hyperplasia, immune cell infiltration, and crust/lesion formation, of mice infected with the more persistent forms (e.g. stationary phase bacteria and biofilm bacteria) were greater than mice infected with log phase bacteria. Using our persistent infection mouse model, we showed that the clinically recommended treatment for recurrent S. aureus skin infection, doxycycline + rifampin, was not effective in eradicating the bacteria in the treatment study, despite reducing lesion sizes and pathology in infected mice. Analogous findings were also observed in a Caenorhabditis elegans model, where S.aureus stationary phase cultures caused a greater mortality than log phase culture as early as two days post-infection. Thus, we established a new model for chronic persistent infections using persister bacteria that could serve as a relevant model to evaluate therapeutic options for persistent infections in general. Our findings connect persisters with persistent infections, have implications for understanding disease pathogenesis, and are likely to be broadly valid for other pathogens.

microbiology↗

bioRxiv2018

Do Deep Neural Networks Outperform Kernel Regression for Functional Connectivity Prediction of Behavior?

There is significant interest in the development and application of deep neural networks (DNNs) to neuroimaging data. A growing literature suggests that DNNs outperform their classical counterparts in a variety of neuroimaging applications, yet there are few direct comparisons of relative utility. Here, we compared the performance of three DNN architectures and a classical machine learning algorithm (kernel regression) in predicting individual phenotypes from whole-brain resting-state functional connectivity (RSFC) patterns. One of the DNNs was a generic fully-connected feedforward neural network, while the other two DNNs were recently published approaches specifically designed to exploit the structure of connectome data. By using a combined sample of almost 10,000 participants from the Human Connectome Project (HCP) and UK Biobank, we showed that the three DNNs and kernel regression achieved similar performance across a wide range of behavioral and demographic measures. Furthermore, the generic feedforward neural network exhibited similar performance to the two state-of-the-art connectome-specific DNNs. When predicting fluid intelligence in the UK Biobank, performance of all algorithms dramatically improved when sample size increased from 100 to 1000 subjects. Improvement was smaller, but still significant, when sample size increased from 1000 to 5000 subjects. Importantly, kernel regression was competitive across all sample sizes. Overall, our study suggests that kernel regression is as effective as DNNs for RSFC-based behavioral prediction, while incurring significantly lower computational costs. Therefore, kernel regression might serve as a useful baseline algorithm for future studies.

neuroscience↗

bioRxiv2018

IL-25 induces beige fat to improve metabolic homeostasis via macrophage and innervation

Beige fat dissipates energy and functions as a defense against cold and obesity, but the underlying mechanisms remain unclear. We found that the signaling of interleukin (IL)-25 including its cognate receptor, IL-17 receptor B (IL-17RB), increased in adipose tissue upon cold and {beta}3-adrenoceptor agonist stimulation. IL-25 induced the browning effect in white adipose tissue (WAT) by releasing IL-4, 13 and promoting alternative activation of macrophages to regulate innervation, which characterized as tyrosine hydroxylase (TH) up-regulation to produce more catecholamine including norepinephrine. Blockade of IL-4R and depletion of macrophages with clodronate-loaded liposomes in vivo significantly impaired the browning of WAT. Obese mice administered with IL-25 were protected from obesity on a high-fat diet and the subsequent metabolic disorders, and the process involved the uncoupling protein 1 (UCP1)-mediated thermogenesis. In conclusion, the activation of IL-25 signaling on beige fat might play a therapeutic potential for obesity and its associated metabolic disorders.

biochemistry↗