David B. Moody - US grants

Experimental animals (monauralized guinea pigs and chinchillas) will be trained by operant conditioning procedures with food reinforcement on one or more of several protocols for hearing evaluation. In addition to pure-tone audiometry, measures of differential thresholds for frequency and intensity, psychophysical turning curves, and threshold functions in high-pass masking noise will be determined. When normal behavioral baselines are stable following training, the normal inner ear of each animal will be lesioned by exposure to low-frequency noise, application of a cryoprobe to the bony wall of the cochlea, or surgical destruction of a portion of the organ of Corti. These treatments all result in damage to the apical-most portion of the organ of Corti that is primarily responsible for detection and discrimination of low- frequency sounds. Following these lesioning procedures, hearing assessment by behavioral means will be continued until sufficient data has been obtained on changes in the measures of hearing and until no further changes are noted for a least 30 days. Animals will then be sacrificed and their temporal bones taken for microdissection and analysis of th cochlea by light and electron microscopy. It is hoped that the relation between the pathology observed in the cochlea and the changes in the bahavioral measures will increase our understanding of the function of the apical region of the cochlea in hearing. These studies are particularly important in understanding the apparent ability of the auditory system to compensate for damage to the apex of the cochlea. This compensation can occur to such an extent that, even with total loss of apical receptors, hearing impairment for low frequencies is slight relative to the profound deafness that results from total hair-cell loss in the base. On goal of this research is a thorough description of the role played by higher-frequency receptors in detection of low frequencies under conditions of total loss of apical hair cells. The studies are also important in view of demonstrations that partial damage can exist in apical regions that is not detected by routine audiometric threshold tests. Developing test that could better detect such damage is an additional goal of this research.

The molecular basis of processing and presenting peptide antigens by MHC class I and class II proteins to T cells is now well understood. A third pathway of presentation of lipid antigens by CD1 proteins to T Cells has recently been described. The non-MHC encoded CD1 bears structural homology to both MHC Class I and Class II. T cell recognition of CD1 restricted lipid antigens functionally parallels that of T cell recognition of Class Il restricted peptides in many ways: 1) T cells show fine specificity for closely related antigens 2) Antigen is recognized only in the context of the appropriate "restricting" CD1 isotype 3) A chloroquine sensitiVe processing step is required for recognition. CD1 restriction of alpha-beta CD8+, alpha-beta "double negative" and gamma-delta T cell populations has now been demonstrated. The CD1 restricted antigens are nonpeptide lipid and glycolipid antigens including mycobacterial cell wall derived mycolic acid and lipoarabinomannan (LAM). The study of molecular models of a new biochemical class of T cell antigens broadly supports clinical research of autoimmune and infectious diseases. The candidate will define the molecular events underlying lipid antigen presentation and recognition. Heterogeneous preparations of mycolic acid will be purified into homogenous species and tested in T cell proliferation assays to define the contributions of specific chemical structures (R groups, chain length) to antigenicity. The complete chemical structure of a mycolic acid species will be determined with gas chromatography and mass spectroscopy (GC-MS). The candidate will radiolabel the lipid antigens and demonstrate that they bind to CD1 by immunoprecipitating CD1/lipid antigen complexes. The structure of the processed, CD1 bound antigen will be determined with conventional GC-MS and with Fourier Transform Ion Cyclotron Resonance Mass Spectroscopy (FTICR-MS). The candidate will identify candidate subcellular compartments of the macrophage for the loading of lipid antigen onto CD1 by defining compartments of colocalization using immunofluorescence and electron microscopy. Using cellular fractionation, cellular subcompartments will be directly examined for the presence of CD1/LAM complexes. The candidate will clone and transfect the TCR from LAM and mycolic acid specific T cell clones into the TCR- cell line, Jurkat. The structural elements of the TCR with functional importance for antigen recognition will be defined by TCR alpha or beta chain substitutions and CDR3 point mutations. The relevance of lipid T cell antigens outside of mycobacterial immunity will be explored by making T cell lines specific for Lipopolysaccharide and lipoteichoic acid, glycolipids of the gram negative and gram positive bacterial cell walls that have chemical and biological similarities with LAM. These studies are included as part of a training program in basic immunology for the candidate for this Mentored Clinical Scientist Award. The program will build on the candidate's previous experience, teaching laboratory techniques in molecular biology, cell biology and cellular immunology. Concurrently, it will provide strong theoretical background in basic immunology. With this training the candidate will develop the skills necessary to establish himself as an independent investigator in immunology. At the end of this training program, the applicant intends to obtain a faculty position and pursue an academic career in basic research.

This proposal consists of 4 Specific Aims that will investigate the perceptual consequences of selected cochlear and stimulus manipulations. The overall goal of the proposed research is to document the extent to which changes in hearing function, as measured psychophysically in behaving animals, can be related to changes as the physiological status induced or measured in the auditory system. The first Aim will investigate the perceptual consequences of electrical stimulation of the round window, as also studied in Projects 3 and 6. This project will determine the extent to which the acoustic component generated by such stimulation is perceived as a high-fidelity representation of the input waveform. The second Aim will examine the perceptual consequences of infusion of neuroactive compounds into the fluids of the inner ear; in particular, of substances that are agonists or antagonists for glutamate receptors which also are studied in Project 4. For example, in appropriate doses, AMPA, a glutamate receptor agonist will induce swelling of afferent nerve endings, and will produce a change in sensitivity which can be reversed when the drug infusion is ceased. In the proposed studies, osmotic pumps will be used to infuse the substances into the scala tympani of awake animals that have been trained as psychophysical observers, and measures of changes in auditory function will be observed as the drug regimen is varied. The third Aim will examine the effects of efferent inactivation in a subject trained to attend to differences in multiple dimensions between a variety of auditory stimuli. The psychophysical task and stimulus contrasts will be optimized for demonstrating differential efferent effects, and reversible efferent inactivation techniques will enable multiple manipulations in each subject. For the first 3 Aims, an important feature of the research is documentation of the relationship between changes in function and structural changes determined from morphological evaluation. The fourth Aim will examine changes in the detection and localization of sound sources as a function of various spatial configurations of signal source and noise masker. These studies complement physiological studies of spatial hearing being carried out in Project 1.

The technical services core at Kresge Hearing Research Institute consists of three service units: the electronics shop; the machine shop; and computing services. Each of these units provides essential services to the research programs of the Institute that make it possible to accomplish research goals more efficiently and inexpensively than would be possible under other arrangements. Each of the units has made it possible to develop specialized instrumentation and applications that are not commercially available, and to maintain existing lab equipment and facilities. The staff of these units have a long history of working together to assist the investigators of the Institute in accomplishing their research goals. Because of their familiarity with auditory research, the staff members are able to proceed from the conceptual description of what is required, through the fabrication and implementation stages of product creation without the need for detailed engineering specifications such as might be required to obtain custom designed products from outside suppliers. The on-site facilities have proven to be extremely valuable in allowing interactions between core staff and investigators to make modifications to devices and applications to optimize them for the research application.

CD1 proteins (CD1a, CD1b, CD1c, CD1d) present lipid antigens for specific interactions with the T cell antigen receptor (TCR). This represents a fundamentally new cellular pathway leading to T cell activation and significantly expands the range of antigens recognized by T cells. The applicant's preliminary results define the first structure of a CD1c-presented antigen to be mannosyl phosphodolichol (MPD), a member of a class of long chain isoprenoid lipids that are present in all cellular organisms. Preliminary studies indicated that the TCR and CD1c mediated human T cell responses to semi-synthetic analogs of both foreign (mycobacterial) and self (human) MPD, thus defining a lipid autoantigen for alphabeta T cells. A trimolecular model of this recognition predicts that CD1 presents amphipathic glycolipids by sequestering the lipid within the hydrophobic groove of CD1, resulting in presentation of the carbohydrate moiety of the antigen to the TCR. The proposed studies will test this model by preparing isoprenoid glycolipids that differ in glycosylation, saturation branching and length of the lipid. Antigen analogs will be tested in plasmon resonance, scintillation proximity and cell-based assays to determine the role of antigen structure in the separate processes of CD1c binding and T cell activation. In particular, analogs will e synthesize to define the structures that distinguish CD1c ligands from those presented by CD1b and CD1d and to determine the molecular basis of discrimination of self from foreign MPDs. We propose a new model of glycolipid autoimmunity by which human autoreactive T cells recognize self isoprenoid glycolipids bound to the CD1c protein. These studies will provide a basic understanding of the molecular events underlying CD1c-presentation of lipids to the TCR and the cellular basis of presentation of a lipid autoantigen.

DESCRIPTION (provided by the applicant): Mycobacterium tuberculosis has infected one third of all humans, resulting in 2-3 million deaths annually and increasing rates of coinfection with the human immunodeficiency virus (HIV). The main impediment to eradication of tuberculosis relates to the ability of M. tuberculosis to persist long term intracellularly within host tissues. Cellular immune responses, particularly activation of Th1 T cells, are crucial for killing intracellular mycobacteria and successfully resolving tuberculosis infections. Although most studies have evaluated mycobacterial protein antigen for activation of T cells, it is now known that group 1 CD1 molecules (CD1a, CD1b, CD1c) mediate T cell activation by mycobacterial glycolipids, including two classes of glycolipid antigen discovered by the applicant's group, glucose monomycolate (GMM) and mannosyl phosphodolichol (MPD). Preliminary studies indicate that CD1-restricted T cells that recognize MPD and GMM are detectable in the peripheral blood of human subjects infected with M. tuberculosis, but not naive controls, indicating that M tuberculosis infection generates glycolipid-specific T cell responses in vivo. T cells use clonally variable T cell receptors to specifically recognize several mycobacterial glycolipids without crossreactivity. T cell recognition of mycobacterial glycolipids is generally specific for the carbohydrate structure of the antigens, including a product of glycosylation reactions that is produced during intracellular growth within host tissues. Now the applicant proposes to measure polyclonal T cell responses from naive and M tuberculosis infected humans to the major classes of purified mycobacterial glycolipids typical of extracellular and intracellular growth. We will use purified CD1-presented antigens as well as mycobacterial glycolipids that are specifically upregulated during intracellular growth to measure human lymphocyte responses during the first year after infection. Antigen-specific lymphocytes will be detected using proliferation assays, antibody-capture cytokine ELISA (elispot) and glycolipid-loaded CD1 -tetramers. Glycolipid-specific T cells will be characterized with regard to restriction by CD1 -isoforms, dependence on prior infection and expression of cell surface markers of immunological memory. These studies will determine whether human infection by M tuberculosis generally results in acquired T cell responses that are specific for mycobacterial glycolipids expressed during intracellular growth. Determination of the immunodominant glycolipid targets of the human T cell response during natural tuberculosis infection will provide crucial information for development of CD1-presented glycolipids as immunomodulatory agents, including vaccines.

DESCRIPTION (provided by applicant): CD1 proteins and glycolipid antigens activate human T cells during the natural course of human infectious and autoimmune diseases such as tuberculosis infection and autoimmune diabetes. The proposed functions of CD1-restricted T cells in tumor immunity, infection and autoimmunity are based on the premise that infection or other cellular perturbations lead to the presentation of microbial or altered self antigens that differ in structure from in the larger pool of self glycolipids that normally comprise cell membranes. However, the particular chemical features of natural glycolipids that control their loading onto CD1 proteins and recognition by T cells are not known. The first known antigens presented by CD1c have recently been described by investigator to be an evolutionarily conserved family of polyisoprenoid glycolipids including mannosyl phosphodolichol (MPD). These antigens were originally isolated from the cell walls of pathogenic mycobacteria, but they have strong structural homology to human MPD, self glycolipid found in all human cells. Preliminary data now demonstrate that MPD, which is similar or identical in structure to human MPD activates CD1c-restricted T cells, including T cells that were previously thought to be autoreactive to CD1c Furthermore, human tissues that are rich in dolichyl glycolipids are infiltrated by CD1c-expressing antigen presenting cells and CD1 c-restricted, MPD-dependent T cells during the normal course of autoimmune thyroiditis. Therefore, the investigator hypothesizes that CD1c-MPD complexes are a molecular target of autoimmune T cell responses in vivo. This will be tested by eluting endogenous glycolipids from cellular CD1 c proteins and measuring the ability of purified dolichyl glycolipids to bind to CD1c in vitro. The ability of mammalian cells to generate antigenic CD1 c-MPD complexes from endogenous glycolipids will be assessed by T cell recognition of cells that are genetically engineered to produce altered MPD. The precursor frequency of antigen specific lymphocytes from human patients with autoimmune thyroiditis will be measured using intracellular cytokine stains and CD1c tetramers. This will test the hypothesis that CD1c and MPD reactive T cells are activated during a human autoimmune disease that targets dolichol-rich tissues. These studies will define the molecular features of natural self and foreign dolichyl glycolipids that control the separate processes of binding to CD1c and activation of T cells.

[unreadable] DESCRIPTION (provided by applicant): Project summary. M. tuberculosis succeeds as a pathogen based on its ability to survive within the endosomal network of infected cells and persist in host tissues for decades. Microscopic examination of mycobacteria growing in tissues shows that they undergo a substantial transformation of their lipid-laden cell walls during growth in cells. Recent studies have identified two mycobacterial lipid synthesis systems that contribute to the remodeling process. The applicant's laboratory has discovered a new class of mycobacterial lipids produced by polyketide synthase 12 (Pks12) (Nature 404, p. 884; J Exp Med 200, p. 1559). Also, sigma factor L (sigL) regulates the expression of polyketide synthase 7 (Pks7) and polyketide synthase 10 (PkslO) (J Bact., in press). Both systems are determinative for the outcomes of M. tuberculosis infection in vivo, but the structures and functions of lipid intermediates controlled by these newly discovered systems are not known. We now propose to measure the effects of pks7 and pkslO on infection in vivo and to identify global changes in cell wall lipids during adaptation of M. tuberculosis to growth in lung tissue and myeloid cells. These studies use targeted gene deletion and global analysis of cell wall structure with a newly developed experimental system for rapid and nearly comprehensive monitoring of cell wall lipids. This system uses orthogonal liquid chromatography directly coupled to high throughput mass spectrometry with software-assisted analysis to detect the small number of altered lipids from among the thousands of molecular ions screened. The complete structures of lipids altered by these stimuli will be identified using collision-induced dissociation mass spectrometry, Fourier transform ion cyclotron mass spectrometry and nuclear magnetic resonance. These experiments combine targeted and global lipid detection to uncover systems used by M. tuberculosis to adapt to growth conditions in the host. Relevance: M. tuberculosis has infected 1.7 billion humans and kills 2 million patients annually, yet there is no simple and effective treatment for tuberculosis. Experts have argued that development of a drug or vaccine that kills all bacteria in the early phase of infection could eliminate tuberculosis in developed countries. Identification of the genes, enzymes and cell wall lipids that mediate adaptation to growth in tissues provides the information necessary to achieve this goal. [unreadable] [unreadable] [unreadable]

M. tuberculosis has infected 1.8 billion humans and accounts for 2 to 3 million deaths annually based on its ability to persist as an intracellular pathogen in human tissues. The host response to mycobacterial infection depends crucially on T cells, which until recently, were thought to be activated solely by peptide antigens bound to MHC class I and II proteins. The discovery of CD1 antigen presenting molecules shows how Langerhans cells and dendritic cells can activate T cells by presenting lipid antigens, including glucose monomycolate (Science 278, p. 283), mannosyl phosphomycoketides (Nature 404, p.884) and mycobactin- like lipopeptides (Science 303, p. 527). This proposal aims to use high performance liquid chromatography to isolate immunodominant antigens from the M. tuberculosis and M. leprae cell walls and then determine their structures using collision induced dissociation mass spectrometry, nuclear magnetic resonance and X- ray crystallography. Because mycobacteria remodel their cell walls and alter their antigen profiles during adaptation to intracellular growth, antigen discovery efforts will focus on pathogenic mycobacteria isolated directly from mammalian tissues and genetically modified M. tuberculosis that are deficient in the enzymes necessary for iron-scavenging from host tissues. The immunogenicity of each lipid antigen will be investigated by ex vivo analysis of T cell precursor frequencies in tuberculosis patients as measured by cytokine-capture ELISA, cell surface cytokine-capture immunofluorescence and staining with lipid-loaded CD1 tetramers. By measuring memory responses and the complexity of lipid antigen specificities in the CD1-restricted T cell repertoire, these studies will provide insight into the basic question of whether CD1 functions in the innate or acquired immune responses in humans. In addition, identification of the precise molecular structures of antigenic lipids offers the prospect of fundamentally novel, MHC-unrestricted immunodulatory drugs and vaccines against leprosy, tuberculosis and multi-drug resistant tuberculosis.

The ability of M. tuberculosis to acquire resistance to first and second line drugs is emerging as a worldwide problem that threatens to undermine TB control efforts in high burden settings. Highly successful multi-drug resistant (MDR) and extensively drug resistant (XDR) strains have been identified that are readily transmitted between hosts. These strains may result from genetic alterations that lead to a hyper-mutable state which enables rapid acquisition of diverse mechanisms of drug inactivation, from compensatory mutations that restore reproductive fitness, or from genetic lesions that lead to loss of uptake of drugs into the bacterium. This project will investigate these hypotheses using genetic and metabolomic analyses of susceptible, MDR and XDR isolates from human patients. Using well-characterized M. tuberculosis (MTB) isolates obtained from existing archives as well as prospectively collected specimens, we will develop and validate laboratory tools to experimentally measure the mutability of strains of M. tuberculosis and use these tools to assess this characteristic in clinical drug sensitive and resistant strains. We will also compare the neutral evolution rates in MDR and drug-sensitive isolates to test for hyper-mutation in these M. tuberculosis populations. Finally we will assess changes in the M. tuberculosis cell wall structure that may correlate with drug resistance and virulence, using a new liquid chromatography-mass spectrometry system for simultaneously measuring thousands of lipid species.

[unreadable] DESCRIPTION (provided by applicant): Building on decades of work showing how T cell receptors recognize peptide antigens bound to MHC proteins, the discovery of the CD1 antigen presentation system shows that T cells also respond to changes in cellular lipids. During the first phase of this proposal, the applicant has discovered phosphoisoprenoid antigens presented by CD1c proteins, determined the enzymatic basis of their biosynthesis and found that they activate T cells derived from the lesions of human patients with autoimmune thyroiditis. These studies also identified two previously unknown classes of lipopeptide antigens presented by CD1c and CD1a, which use fattyl acyl chains to anchor to CD1 proteins, leading to presentation of their peptide moieties to T cell receptors. Recent advances in the synthesis of lipid antigens, along with crystal structures showing the precise dimensions and shapes of CD1 antigen binding grooves, now allow detailed investigation of the molecular interactions among CD1, lipid antigens and T cell receptors. The continuation proposal uses dolichyl glycolipids and acylpeptides as model antigens to determine the mechanisms by which CD1a and CD1c present a carbohydrate moiety and a peptide moiety to human T cell receptors in vitro and during the course of autoimmune thyroiditis. Using synthetic antigen analogs, we will determine the influence of carbohydrate linkage, lipid size and peptide length for binding to recombinant CD1 proteins. CD1-lipid complexes will be tested for their ability to bind to and activate T cell receptors in vitro. Mass spectrometric analysis of lipids eluted from cellular CD1 proteins will identify the molecular structures of endogenous lipids loaded onto CD1 proteins in cells. Last, existing and newly discovered lipid antigens will be tested for their ability to be recognized by polyclonal T cells from the thyroid glands and peripheral blood of human patients with Graves' disease and Hashimoto's thyroiditis. Relevance to public health. These studies provide a new model for understanding the molecular targets of T cells in a common human autoimmune disease, which is highly organ-specific. Further, determination of the specific size and chemical characteristics of lipids that bind to CD1 and activate T cells provides the necessary information for design of immunomodulatory lipids that can be applied as drugs to increase, decrease or polarize the functions of human T cells in vivo. [unreadable] [unreadable] [unreadable]

? DESCRIPTION (provided by applicant): For many years, human ?? T cells were thought solely to recognize peptides bound to MHC proteins. Early stages of work on this parent grant provided broad evidence that non-polymorphic CD1 proteins (CD1a, CD1b, CD1c) display mycobacterial lipids to human ?? T cells. After generating working tetramers comprised of human CD1a, CD1b and CD1c proteins, our published data show that polyclonal, lipid-specific T cells exist as cell populations in tuberculosis (TB) patients and possess effector functions (TNF and IFN-?) that are known to protect the human host. Further, CD1c tetramers detected polyclonal ?? T cells recognizing mycobacterial mycoketides. CD1b tetramers detected a previously unknown human T cell type: germline-encoded mycolylreactive (GEM) T cells. GEM T cells have a degree of T cell receptor (TCR) conservation, which is equivalent to NKT cells or MAIT cells, but GEM T cells function to recognize CD1b. Building on these new reagents and recent success in solving the crystal structure of CD1a-lipid-TCR, the renewal will carry out detailed biophysical analyses of ternary TCR-CD1-mycobacterial lipid interactions involving both ?? and ?? TCRs. Turning to translational questions in TB disease, we will collaborate with Socios En Salud to determine the immunodominant mycobacterial lipids stimulating human response in latent and active TB. Over the past 3 years, we validated guinea pig CD1b and CD1c tetramers and anti-CD1 antibodies to create a tractable small animal model of in vivo T cell response to CD1. Now we propose to measure the duration and magnitude of response to lipid antigens and determine whether CD1 presented lipids can protect against M. tuberculosis infection. Finally, based on the discovery of GEM T cells, we propose to detect networks of interdonor conserved TCRs in the humans. These studies move beyond NKT cells and MAIT cells to detect stereotyped human T cell response to CD1a, CD1b and CD1c bound to mycobacterial lipids. Such studies seek to reveal a currently unappreciated structure of a public human T cell response. Such a system would change general views of the structure of the human T cell repertoire and the particular TCRs identified might be used for immunodiagnosis of TB infection.

The Metabolomics Core (B) is comprised of a Biosafety Level 3 suite for handling infectious samples, massively parallel detection of mycobacterial metabolites using Time of Flight mass spectrometry (Agilent Accurate Mass ToF 6230), specialized resources for identifying known mycobacterial compounds, and analytical capabilities to discover previously unknown compounds (Agilent Accurate Mass 6520, QTof, Thermo LXQ Advantage 2 Dimensional Ion Trap with MSn). The integrated technologies were assembled by the Moody laboratory and supported a Biomarker Discover Initiative of the Broad Institute, NIH U19 and ROl Projects. Dataflow involves receipt of mycobacteria (Project 3) or patient samples (Project 4), from which total metabolites are extracted and sterilized in organic solvents. Tissue or mycobacterial extracts enter into a liquid chromatography-mass spectrometry system, which was specifically designed to broadly detect the highly diverse and hydrophobic compounds in mycobacteria. In the first phase of whole organism analysis, the platform rapidly detects triplicate intensity values for ~10,000 distinct compounds in each sample. By aligning large datasets derived from different patients, clinical isolates or genetically engineered bacteria, in-house-designed software pipeline identifies all compounds that are changed at statistically significant levels. In a second, targeted phase, all changed compounds are ranked by biological or quantitative criteria to define compounds of interest, whose structures are solved by comparing their masses to the literature (MycoMass) and in-house (MycoMap) databases or are solved through collisional mass spectrometry. This system has discovered previously unknown compounds, identified strain-specific mycobacterial biomarkers in vitro and from tissues and identified lipids changed after gene deletion. This overview describes expansion of the substantial existing core facilities, including a new generation of high accuracy mass spectrometry and expansion of mycobacterial databases, as well as use of the Core to discover biomarkers in drug-resistant or latent mycobacteria or biomarkers of infection.

The absence of an effective vaccine for tuberculosis means that TB control relies on the early diagnosis and effective treatment of infectious cases, which is compromised by the relatively low sensitivity and specificity of standard diagnostic tools. Because TB infection most often results in a chronic asymptomatic state, prevention of disease by targeting those who are infected, but not yet ill, has been difficult to implement in high burden settings where more than half of the population is TB infected. The long duration of treatment necessary to achieve high cure rates and the emergence and spread of drug resistant organisms have further undermined the potential impact of national TB control programs. Our proposed plan responds to these research priorities and grows out of a series of recent research findings from our own groups and others that suggest an innovative interdisciplinary approach to the discovery of basic mechanisms through. Our proposed project begins with the identification and longitudinal follow-up of patients diagnosed with active TB and their household contacts. Patients that progress to active TB disease (progressors) are followed for disease outcomes, including relapse, and household contacts are followed for evidence of TB infection and disease. This design and our extensive longitudinal follow up capabilities will allow us to identify and characterize TB index cases and their exposed household contacts through careful clinical and epidemiologic studies, human genomics (by exome sequencing) human genetics (by exome chip), transcriptomics, and metabolomics. We have established Cores in Human subjects, Bio-informatics, and Metabolomics that will work in parallel to identify targets including pathways linking human metabolism and immune response, T cells involved in Mtb response, pathogen determinants of drug resistance and pathogen-shed markers of clinical TB phenotypes. Each project includes validation of these targets in the guinea pig model. Based on our results, we will then go on to test specific interventions in the animal model, focusing in particular on pharmacologic agents that alter human metabolic and immune responses.

The large and growing problem of drug resistance of Mtb in humans requires new approaches to determine the mechanism of multi-drug resistance and markers of the resistant state in humans. Using 116 drug resistant Mtb isolates, we used whole genome sequencing to identify Mtb mutations stastically associated to drug resistance during natural infection of humans. This unbiased and broad approach identified all of the known genes that promote drug resistance and identified 11 genes of known function with no known connection to drug resistance. Three of these genes (ppsA, polyketide synthase 12 and polyketide synthase 3) produce polykefides with inter-related functions in outer membrane integrity of Mtb. An independent, mass spectromety analysis of molecules changed in a strain evolving drug resistance in vivo identified mannosyl phosphomycoketide, the product of polyeketide synthase 12, as being inversely correlated with resistance. Based on independent genetic and metabolomic discovery of altered polyketide synthesis function in drug resistant strains, we will test a new model in which altered polykefide function controls cell wall integrity and drug resistance. Using patient derived strains from Lima, Peru, we seek independent validation of altered polykefide biosynthetic genes in human populations. In the laboratory, we will test whether polykefide synthase genes and the lipids they produce influence molecular remodeling of the cell wall, Mtb permeability changes and drug resistance. To determine if genes showing strong statistical associations are actually causal of the drug resistant state or altered fitness, we will test measure the cell wall components and an in vivo fitness of polyketide deficient and drug-resistanct Mtb. Finally, taking advantage of whole organism screens that identify the particular bacterial metabolites changed in polyketide-deficient or drug resistant states, we will identify the changed molecules to give insight to the mechanism of resistance and to identify biomarkers of the drug resistant state.

? DESCRIPTION (provided by applicant): Role of tuberculosinyl metabolites in M. tuberculosis virulence Abstract Among all types of mycobacteria, Mycobacterium tuberculosis is the world's most prevalent and deadly pathogen. To find the particular molecules that enable its pathogenicity, we compared all detectable lipids in M. tuberculosis to those in a non-pathogenic vaccine strain (BCG). This subtractive screen identified a previously type of molecule called tuberculosinyl adenosine (TbAd) as well as the genes (Rv3377c, Rv3378c) that produce it. Although TbAd was overlooked in a century of tuberculosis research, our preliminary data show that TbAd is one of the most abundant lipids and is produced as six structurally related subfamilies. Published data show that M. tuberculosis is unique among mycobacteria in its ability to survive within the phagosomes of infected macrophages, based in its blockade of acid-mediated bacterial killing by macrophages. Our preliminary data show that TbAd is sufficient to block acidification of macrophage phagosomes. Therefore, we posit that TbAd and related tuberculosinyl metabolites are the long sought molecules that influence M. tuberculosis' unique ability to escape intracellular death. Here we propose to discover new tuberculosinyl metabolites in M. tuberculosis and determine their natural structures. Using synthetic TbAd and human macrophages, we will determine the cellular mechanism by which TbAd selectively inhibits phagosome acidification, while still allowing M. tuberculosis to be taken up into its phagosomal niche. Using M. tuberculosis lacking the biosynthetic enzymes (Rv 3378c, Rv3377c), we will measure the influence of TbAd in the outcome of natural infections. Last, we will test tuberculosinyl metabolites as targets for new diagnostic tests for human tuberculosis. Several features suggest that TbAd and TbAd-specific antibodies could be a highly specific chemical marker or infection. TbAd is abundantly secreted by M. tuberculosis, but is lacking in other pathogens that mimic tuberculosis. Therefore, we will detect TbAd and TbAd specific antibodies in serum and urine from a well-characterized patient cohort in Lima, Peru and in mice.

ADMINISTRATIVE CORE A Co-leaders: D. Branch Moody (Brigham and Women's Hospital) and Kyu Rhee (Weill Cornell Medicine) ABSTRACT The Administrative Core A is led by D. Branch Moody (contact Principal Investigator) and Kyu Rhee (co- Principal Investigator), who have broad prior administrative experience in managing multi-disciplinary collaborative projects. The Core leaders will communicate with NIH program officers, the External Scientific Advisory Group, the Core Administrator and the Data Manager to meet administrative, financial and scientific goals of the tuberculosis research unit (TBRU). Core A will coordinate and monitor scientific progress toward defined milestones, while encouraging interdisciplinary collaboration within the TBRU and among TBRU groups that form the TBRU Network. To accomplish this goal, Core A will implement a program of scientific communication that involves yearly national meetings at the NIH, monthly TBRU-wide scientific meetings, as well as weekly administrative and affinity group meetings. The Data Manager will implement the data management plan, tracking large scale metabolomics, lipidomics and genetics datasets, and coordinating data submissions and interaction with NIAID-supported Bioinformatics Centers. The experienced Administrative Core Associate will be responsible for the daily administration and fiscal management of the proposed TBRU, including monitoring of subcontracts and expenditures, and managing annual reporting requirements. Core leaders will support scientific training through individual development plans for trainees and internal programs to support the development of non-tenured investigators working in the tuberculosis field, including a collaborative projects program and an internship program for on-site cross-training in new techniques available through the TBRU. Core A has an Administrative Plan that guides interactions among members of the TBRU and provides specific timelines for scientific progress and outline processes to resolve problems.

Project Summary - CD1 Presentation of Self Lipids to Human T cells ? D. Branch Moody T cells play a central role in the pathogenesis of human autoimmune diseases, including allergy and drug hypersensitivity syndromes. For decades studies of T cell autoantigens have emphasized peptide antigens bound to MHC proteins. This proposal investigates the basis by which self lipid antigens bind to human CD1 antigen presenting molecules, leading to autoreactive T cell responses. This competing renewal builds on an existing program of development that has invented human CD1a and CD1b tetramers and used them to discover previously unknown lipid autoantigens for T cells, including squalene and phosphatidylglycerol. To provide a general basis for lipid epitope mapping, we will use mass spectrometry to broadly identify many hundreds of self lipids that bind to CD1 proteins in human cells. Using tetramers, we will identify and clone autoreactive T cell receptors and measure their binding to CD1-self lipid complexes. Building on recently published studies showing that CD1a autoreactive T cells are increased in patients with allergy syndromes, we will measure CD1a-reactive T cell responses in human cohorts. Finally, we will investigate a new mechanism by which CD1a proteins present common skin irritants from clinical patch tests to T cells, bypassing mechanisms that relate to peptide haptenization. Overall, these studies will extend the spectrum of natural T cell autoantigens to include self lipids and investigate their roles in common allergic diseases.

ABSTRACT - Metabolic determinants of Mtb virulence, vulnerability and variation Mycobacterium tuberculosis (Mtb) has emerged as the world's most deadly pathogen based in large part on the highly unusual biological and chemical properties of its cell envelope. Comprised of a distinctive hydrophobic outer mycolate membrane, anchored to an underlying complex of polysaccharide and peptidoglycan polymers, the Mtb envelope serves as both the primary interface with, and barrier to, the human host. In human tuberculosis (TB) disease the Mtb envelope mediates a years-long standoff, and serves as the barrier to all anti-mycobacterial drugs. Yet, knowledge of its native composition, variation and regulation of drug entry remains fragmentary. This team of applicants has created new genetic and metabolomic tools to comprehensively dissect and analyze the metabolite and lipid components of the Mtb envelope on an organism-wide basis across a large set of clinical isolates. Moreover, this TBRU proposes to provide the first descriptions of cell envelope variation among isolates from human patients and identify key determinants of its virulence and barrier to drug action that could inform the development of better diagnostics and therapeutics. Structures of new molecules will first be determined using synthetic chemistry and mass spectrometry. The genes encoding these metabolites will then be identified and functionally validated using new genome-scale CRISPR interference technologies, assays for penetration into the cell envelope, and genetically defined mouse models of in vivo growth. Using mass spectrometry, we will solve the structures of up to 250 surface barrier lipids and more than 41 gene-lipid pairs that dominate in cell envelope variation among patients. Patient-derived Mtb strains will be obtained from clinical samples collected at our field sites in Masiphumelele, South Africa, where we will implement clinically relevant technology for detection of live Mtb in exhaled (non- coughed) human bioaerosols. Studies of barrier function place special emphasis on rifampicin as a model compound due to its clinical importance as a frontline drug and role as a defining element of drug resistant TB. The ability to analyze patient urine and serum has further resulted in the discovery of new biomarkers of disease activity and response to drug therapy, motivating linked translational efforts to advance the development of non-sputum based, real time point-of-care diagnostic tests. This highly interactive group of scientists thus seeks to provide better drugs and diagnostic tests, as well as a deep and durable scientific foundation for understanding of the Mtb envelope, especially the particular genes and molecules that control active remodeling, drug action and human host response.

Profiling and Mapping Core B Project Leaders: Kyu Rhee and D. Branch Moody Co-investigators: Jacob Mayfield ABSTRACT - Profiling and Mapping Core B Modern infectious disease research relies fundamentally on genomic maps of the major pathogens. With the recent development of whole-organism metabolomics profiling tools, comprehensively mapping the M. tuberculosis (Mtb) metabolome is now feasible. Core B supports mass spectrometry-based profiling of Mtb metabolites to support Projects 1, 2 and 3 that focus on discovery, drugs and diagnostics, respectively. The Core will carry out whole-organism lipidomic and metabolomic profiling, archive datasets and use modern bioinformatic methods to generate metabo-lipidomic maps of Mtb. Maps are comprised of a structured listing of all metabolite subclasses, a list of all named metabolites within each class, and links of all named metabolites to mass values and other biological variables. We previously generated the MycoMap Database, which is the largest existing map of 183 mycobacterial metabolite classes linked to ~195,000 accurate mass values. We will now expand the database by adding newly discovered lipids generated in Projects 1, 2 and 3. Supporting Projects 1 and 2, we will solve the Mycolate Outer Membrane Lipid Map, which describes compounds located at the host-pathogen barrier and involved in drug penetration. Supporting Projects 1 and 3, the core will measure metabolite variation among Mtb strains from South African patients to generate the Human Mtb Strain Lipid Variation Database. This database will describe the prevalence and variation of lipids among strains circulating in human communities to identify lipids under control by host selection and to support diagnostics development.

Project 1. Study of M. tuberculosis under human host selection to identify virulence and barrier lipids Project Leader: D. Branch Moody Coinvestigators: Kyu Rhee, Jacob Mayfield Collaborating Investigators: Adriaan Minnaard (Core C), Jeremy Rock (Core D), Clare Smith (Core E) ABSTRACT Comparative genomics has served as a dominant paradigm for tracking the tuberculosis (TB) epidemic, understanding Mycobacterium tuberculosis (Mtb) virulence and developing new drugs and diagnostics. Mycobacterial metabolism, in contrast, has been viewed as invariant feature of all clinical Mtb strains. Through comparative metabolomic profiling of ~10,000 lipids among 84 patient-derived Mtb strains, we discovered that Mtb?s pathognomonic lipid envelope shows identifiable patterns of variance among strains circulating among human populations. To determine the impact of phenotypic diversity within the infecting bacterial population, we will map cell wall lipid variation among 140 Mtb strains among TB patients from Masiphulemele, South Africa. The resulting lipid map will describe variations in lipid composition among Mtb strains transmitting in community. From a biological perspective, Mtb?s lipid envelope forms the primary interface with the host and is therefore a direct and ongoing biochemical target of evolutionary selection. This project aims to reveal the previously undescribed chemical diversity and lipid products that have arisen as a consequence of host- and drug-derived clinical pressure. Using organism wide lipid profiling and genome wide sequencing, we have identified 42 lipid-gene pairs that dominate in Mtb strain variance, as well as 1150 lipid species overexpressed in virulent Mtb and 250 lipids selectively expressed at the host interface. Preliminary data support our ability to then link these lipids to specific bacterial genes, even when prior to knowledge of the metabolite?s structure or a gene?s function is lacking. CRISPR interference strategies will then establish causal linkages between genes of unknown function and newly discovered lipids. We will further test lipid deficient strains in collaborative cross mice to reveal specific roles of newly identified lipids in Mtb virulence. These discovery studies will identify biologically important lipids that determine key outcomes in virulence, the host interface and Mtb survival in vivo, supporting new approaches for tuberculosis diagnosis and treatment.