Source: COLORADO STATE UNIVERSITY submitted to NRP
NOVEL IDENTIFICATION OF EXISTING AND EMERGING ANTIMICROBIAL DRUG RESISTANCE MARKERS FOR DEVELOPMENT OF A CLINICALLY-APPLICABLE BIOSTATISTICAL MODEL
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
Reporting Frequency
Annual
Accession No.
1004894
Grant No.
(N/A)
Cumulative Award Amt.
(N/A)
Proposal No.
(N/A)
Multistate No.
(N/A)
Project Start Date
Oct 1, 2014
Project End Date
Sep 30, 2017
Grant Year
(N/A)
Program Code
[(N/A)]- (N/A)
Recipient Organization
COLORADO STATE UNIVERSITY
(N/A)
FORT COLLINS,CO 80523
Performing Department
Clinical Science
Non Technical Summary
Antimicrobial resistance (AMR) in bacteria of animal origin is a major global public health threat. Tools for detecting phenotypic resistance patterns are limited and require advanced molecular methods to reveal genetic associations with the AMR patterns. The aim of our proposal is to identify cellular biomarkers associated with mechanisms of AMR in Salmonella using novel metabolic and proteomic profiling techniques and investigate the diversity of these markers among established genetic patterns of resistance. Previous research performed by the PI discovered various sizes of class I integrons associated with AMR patterns. However, in some cases several of these AMR patterns were still present despite the absence of class I integron genes. This highlights a need to further investigate AMR mechanisms, which could be occurring at the cellular level.Metabolomics and Proteomics are the two "omic" approaches that will produce metabolite and protein profiles, respectively and would help differentiate AMR coded integron genes. Metabolic profiling will be used to develop a bio-statistical model to predict the presence of integrons. This work could have an immediate impact on human health and lead to advancements in drug discoveries as well as diagnostic tools to screen large numbers of samples. The successful completion of the proposed work will revolutionize how we address major global health problems associated with emerging and existing antimicrobial drug resistance.
Animal Health Component
(N/A)
Research Effort Categories
Basic
75%
Applied
(N/A)
Developmental
25%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
31140101040100%
Knowledge Area
311 - Animal Diseases;

Subject Of Investigation
4010 - Bacteria;

Field Of Science
1040 - Molecular biology;
Goals / Objectives
This novel approach will develop a bio-statistical model that will identify biomarker profiles predictive of antimicrobial resistant gene coded class I integrons. This approach will be applicable epidemiologically to assess the antimicrobial resistance of large numbers of samples.
Project Methods
Efforts: To identify and characterize metabolite and protein biomarkers which differentiate AMR and susceptible strains of Salmonella will be grown from previously cultured and identified from pooled fecal samples from four swine farms in Illinois. AMR pattern were carried out using NARMS (National Antimicrobial Resistance Monitoring System) panels. Non- targeted Protein profiling will be performed using 12 isolates (small sample size due to expense of the method), proportionally selected from each integron as well as no integron category. Metabolite profiling will be performed on 100 Salmonella isolates, again proportionally selected from each category.Proteomic profiling approach: Extracted proteins will be precipitated, washed and re-suspended in urea and ProteaseMAX™Surfactant to maximize solubility of hydrophobic proteins. Proteins will be reduced, alkylated and digested with trypsin using standard methods. The resulting peptides will be desalted and analyzed by nano- scale high pressure liquid chromatography coupled with tandem mass spectrometry. The MS/MS spectra will be searched against the most recent Uniprot protein database using both the Mascot and Sequest database search engines. A decoy database approach will be employed to calculate a false discovery rate. Search results for each sample will be imported and combined using probabilistic protein identification algorithms implemented in Scaffold software. Importantly, data analysis will be performed separately for each isolate. Relative quantitation will be determined using the label free spectral counting approach.Metabolomic profiling approach: UPLC-MS (Ultra Performance Liquid Chromatography-Mass Spectrometry) analysis will be performed on a Waters Xevo G2- TOF MS coupled with a Waters Acquity UPLC Separation will be performed on a UPLC T3 reverse phase column and data will be collected in MSE mode (alternating low and high collision energy) (Plumb et al., 2006). For GC-MS (Gas Chromatography-Mass Spectrometry) analysis, cell extracts will be dried and derivatized using a standard protocol. GC-MS data will be acquired on a Thermo Scientific Trace-ISQ GC-MS system with separation using a 30m TG-5MS column. Data from both UPLC-MS and GC-MS acquisitions will be processed using XCMS (open source software) for peak detection, retention time alignment, and normalization. Principal component analysis will be performed to visualize differences between integron groups. Univariate (ANOVA) and multivariate statistical analyses will be performed to determine features in the data that vary significantly between strains. Metabolite annotation of GC-MS data will be performed by grouping molecular features into peak groups using AMDIS software and screening spectra against spectral libraries. Annotation of UPLC-MS data will be performed by unbiased grouping of molecular features into spectra based on correlational clustering across the dataset and screening spectra against the spectral libraries. For this pilot investigation we will select 100 isolates for testing.Statistical analysis: Chi-square analysis will be used to identify significant associations between metabolic/protein markers and AMR coded integrons (p<0.05).Aim 2. To develop a bio-statistical model that will identify biomarker profiles predictive of AMR gene coded class I integrons: To identify potential biomarkers, multivariate analytical approaches will be applied to the mass spectrometry data on mass and retention time of the metabolites. Once a PLS-DA model is calculated and validated, it can be used for prediction of class membership for unknown samples. Hierarchical cluster analysis will be applied to determine clusters of metabolites that are common in AMR coded class I integrons. To determine the principal components (PCs) that are significantly associated with the class I integrons, an initial step of univariate logistic regression analysis followed by multivariable analysis will be performed to determine PCs that have significant associations with presence of different integron sizes. The metabolites highly predictive of presence of integron sizes will be determined and depicted in terms of odds ratios with their 95% confidence limits or a p-value of <0.05. Biostatistics will not be performed for the proteomic data due to the small sample size.

Progress 10/01/14 to 09/30/17

Outputs
Target Audience:Public health Diagnosticians Epidemiologists Changes/Problems:Analysis of metabolomics and proteomics data What opportunities for training and professional development has the project provided?MPH Capstone project Undergraduate student training in laboratory methods. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?Publish manuscript. Present in conference.

Impacts
What was accomplished under these goals? Poster and oral presentation for the capstone project of an MPH student. Abstract submitted for INFECTIOUS DISEASE DIAGNOSTICS FOR THE 21ST CENTURY symposium held on July 8-10, 2019 in Fort Collins, CO

Publications

  • Type: Conference Papers and Presentations Status: Submitted Year Published: 2019 Citation: INFECTIOUS DISEASE DIAGNOSTICS FOR THE 21ST CENTURY, Fort Collins, CO from July 8-10, 2019


Progress 10/01/15 to 09/30/16

Outputs
Target Audience:Metabolomics and Proteomics scientists Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?CVMBS Research day poster presentation by Sean Montgomery (MS student of MIP department) 01.30.2016 How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?Scientific publication in a peer review journal.

Impacts
What was accomplished under these goals? Integron size groups that were found in dairy cattle were 1000, 1000+1200 and 1800 bp; in swine were 1000, 1000+1200; in poultry were 1000 and in humans were 1000, 1200, 1000+1200, 1000+1200+1600, and 1800. 48% of bovine samples, 83% of swine, 11.5% of poultry and 32% of human samples carried at least one integron group, 100% of them resistant to at least one antibiotic. The majority represented 1000 bp in humans and poultry, 1000+1200 in dairy cattle and swine. The most common AMR pattern among all resistant samples (23/126%=18.25%) was ampicillin, amoxicillin-clavulanate, streptomycin, sulfisoxazole, tetracyclines, chloramphenicol and florfenicol (coded as AMCAMS10SSSTECFFC). Hence, these samples are chosen for performing 'omic' profiling. All of those samples carried 1000 and 1200 bp integrons, 5 from human, 8 from bovine and 10 from swine. The samples that did not show resistance to any antibiotics did not carry any integrons, indicating that integron testing is highly specific in detecting AMR. However, there were 35% of the dairy samples not carrying integrons but showed resistance to at least one antibiotic, 83% of swine, 35% of poultry samples and 50% of human samples not carrying any integron showed resistance to at least one antibiotic. Thereby, integron testing provides high confidence in the positive result (high predictive value positive) in detecting AMR. The proteomic analysis results from 9 isolates (3 human, 3 bovine, 3 porcine) grown in ACSSuT-treated broth and no-drug broth (groups coded as FD and ND, respectively), produced 1631 confidently identified proteins (confident identification signifies a presence in at least two out of three replicates and a spectral count of 2). A total of 822 proteins significantly varied in abundance between the 2 groups, with 366 proteins significantly upregulated in the presence of the ACSSuT drug panel (Figure 2), and 456 proteins significantly upregulated in the absence of the ACSSuT drug treatment. Results of ACSSuT-exposed upregulated protein samples were organized according to cellular pathways via the Kyoto Encyclopedia of Genes and Genomes. The metabolic pathways, ribosomal pathways and biosynthesis of secondary metabolites were the top 3 cellular pathways identified to upregulate proteins with the ACSSuT drug stress.

Publications


    Progress 10/01/14 to 09/30/15

    Outputs
    Target Audience:Scientists, graduate and undergraduate students in Veterinary Medicine and Biomedical Sciences Changes/Problems:Specific integron size was selected depicting a particular pattern of AMR to begin answering the question on what metabolites and proteins are expressed in samples carrying those patterns. An experimental approach was undertaken where 9 samples were grown with and without antibiotics to be able to identify the markers among 3 host species. The proteomic method is very expensive and hence the availablebudget was utilizedto answer an important scientific question to begin with. What opportunities for training and professional development has the project provided?The project has involved 2 undergraduate students and 1 graduate student from MIP department. It was a very good opportunity for them to learn the omics techniques in terms of sample prep and submission. How have the results been disseminated to communities of interest?Proteomic part of the project was presented on a poster for the CVMBS research day 2016. What do you plan to do during the next reporting period to accomplish the goals?Metabolomics results are yet to be received from the PMF laboratory. Once those results are received, all the data will be compiled to model statistically the associations between the metabolite and protein profiles and the antimicrobial resistance in the Salmonella isolates carrying1000+1200 bp integrons.

    Impacts
    What was accomplished under these goals? Total number of Salmonella enterica ser Typhimurium cultures retrieved from culture banks of various laboratories were 183 representing dairy cows, swine, poultry and humans. Among the 183 samples tested against 16 antimicrobial drugs for resistance, 69% of were resistant to at least one antibiotic. The number of samples resistant to at least 5 antibiotics were 42%, mostly representing humans followed by bovine and then swine; birds showed the least number of AMRs. Integron size groups that were found in dairy cattle were 1000, 1000+1200 and 1800 bp; in swine were 1000, 1000+1200; in poultry were 1000 and in humans were 1000, 1200, 1000+1200, 1000+1200+1600, and 1800. Forty eight percent of bovine samples, 83% of swine, 11.5% of poultry and 32% of human samples carried at least one integron group, 100% of them resistant to at least one antibiotic. The majority represented 1000 bp in humans and poultry, 1000+1200 in dairy cattle and swine. The most common AMR pattern among all resistant samples (23/126%=18.25%) was ampicillin, amoxicillin-clavulanate, streptomycin, sulfisoxazole, tetracyclines, chloramphenicol and florfenicol (coded as AMCAMS10SSSTECFFC). Hence, these samples are chosen for performing 'omic' profiling. All of those samples carried 1000 and 1200 bp integrons, 5 from human, 8 from bovine and 10 from swine (Table 4). The samples that did not show resistance to any antibiotics did not carry any integrons, indicating that integron testing is highly specific in detecting AMR. However, there were 35% of the dairy samples not carrying integrons but showed resistance to at least one antibiotic, 83% of swine, 35% of poultry samples and 50% of human samples not carrying any integron showed resistance to at least one antibiotic. Thereby, integron testing provides high confidence in the positive result (high predictive value positive) in detecting AMR. Proteomic analysis: The proteomic analysis results from 9 isolates (3 human, 3 bovine, 3 porcine) grown in ACSSuT-treated broth and no-drug broth (groups coded as FD and ND, respectively), produced 1631 confidently identified proteins (confident identification signifies a presence in at least two out of three replicates and a spectral count of 2). A total of 822 proteins significantly varied in abundance between the 2 groups, with 366 proteins significantly upregulated in the presence of the ACSSuT drug panel (Figure 2), and 456 proteins significantly upregulated in the absence of the ACSSuT drug treatment. Results of ACSSuT-exposed upregulated protein samples were organized according to cellular pathways via the Kyoto Encyclopedia of Genes and Genomes (Table 5). The metabolic pathways, ribosomal pathways and biosynthesis of secondary metabolites were the top 3 cellular pathways identified to upregulate proteins with the ACSSuT drug stress. Table 5. KEGG Pathway Enrichment Analysis for upregulated proteins under ACSSuT drug stress Cellular Pathways/Processes Protein Count1 p-value2 Metabolic pathways 93 1.01E-04 Ribosome 49 3.73E-45 Biosynthesis of secondary metabolites 38 4.21E-01 Purine metabolism 19 1.17E-03 Aminoacyl-tRNA biosynthesis 17 1.03E-10 Pyrimidine metabolism 15 2.39E-03 Protein export 11 2.03E-06 Bacterial secretion system 11 8.53E-02 Fatty acid biosynthesis 10 7.71E-08 Fatty acid metabolism 10 1.18E-04 RNA degradation 8 8.13E-04 Biotin metabolism 6 5.55E-02 RNA polymerase 4 4.99E-03 1The number of enriched proteins upregulated during exposure to the ACSSuT drug panel, organized by their cellular pathway/process in the cell. 2Bonferroni corrected. Host Specific Differences: The total number of proteins upregulated within the drug group (FD) were 142 in humans, 80 in bovine and 32 in porcine samples. When the 3 host species were compared, there were 91 unique proteins in humans, 39 in bovine and 33 in porcine in the drug group.

    Publications