Duke University

NIH Research Projects · FY 2026 · 2026-04

Abstract Plasma cell (PCs) antibody production is at the core of humoral immunity and inherently critical to the success of almost all existing vaccines. Their ability to persist and secret antibodies long after pathogen exposure or immunization provides long term sometimes lifelong immune protection against future encounter of the pathogens. What are the molecular pathways dictating the lifespan of PCs during their generation and subsequent maintenance remain largely unknown. Equally unknown are the regulators that control the diversity and activity of PCs reside in various lymphoid and non-lymphoid tissues. Cell lineage specific gene targeting and tracing (GT&T) with the Cre/LoxP system has been proven vital to understand the biology and regulation of various types of cells, especially rare cell populations such as PCs. However, existing mouse models for PC selective GT&T have significant drawbacks due to leakiness in either closely related cell lineage such as germinal center B cells or functionally intertwined cell lineage such as germinal center T cells. To improve selectivity in genetic dissection of PCs, we searched for genes that display high level expression in PCs but absent in the rest of lymphoid lineages. We found a single gene that fulfills this criterion and thus a good candidate for developing new mouse models with high selectivity and efficiency to drive Cre/CreERT2 expression in PCs. We propose to develop new Cre and CreERT2 models for GT&T studies in PCs. The proposed mouse model will address the immediate need for genetic and molecular investigations aimed at understanding longevity and behaviors of PCs in various tissues in response to infection or immunization.

NIH Research Projects · FY 2026 · 2026-04

ABSTRACT Kinase signaling plays a pivotal role in regulating neuronal physiology, yet the mechanisms underlying kinase specificity, subcellular compartmentalization, and disease-associated dysfunction remain poorly understood. Dysregulated kinase activity is a key driver of neurodegenerative disorders, including Parkinson’s Disease (PD), where mutations such as LRRK2G2019S alter kinase function and neuronal survival. This project leverages innovative AI/ML algorithms, advanced CRISPR-based proteomics, and in vivo models to bridge existing gaps in our fragmentary knowledge of how kinases alter neuronal signaling and contribute to disease. By integrating a unique combination of predictive computational frameworks with experimental validation, we aim to uncover how kinases dynamically regulate proteomes in dopamine neurons and how disease-associated mutations alter allostery and rewire kinase-substrate interactions. These findings are expected to advance our understanding of kinase neurobiology, provide insights into neurological disease mechanisms, and pave the way for future therapeutic strategies targeting kinase signaling.

NIH Research Projects · FY 2026 · 2026-04

PROJECT SUMMARY/ABSTRACT Standard treatment of prostate cancer (PCa) with anti-androgen agents fails due to development of therapy resistance and castration-resistant prostate cancer (CRPC), a terminal disease. Enhancer of Zeste Homolog 2 (EZH2), a histone methyltransferase and catalytic subunit of Polycomb Repressive Complex 2 (PRC2), and Androgen Receptor (AR) are the two most crucial gene/chromatin regulators involved in the development and progression of advanced PCa, including CRPC. How these oncoproteins contribute to CRPC development and progression remains far from clear. Recent works by us and others show that the EZH2 regulome in CRPC goes well beyond its well-studied, canonical gene-repressive role. Specifically, EZH2 also binds AR and its constitutively active variant, AR-V7, and has a non-canonical function in activation of prostate oncogenes, which differs from the well-known PRC2:EZH2-driven canonical function related to repression of tumor- suppressive genes (TSGs). EZH2’s cryptic transactivation domain (EZH2TAD) and AR’s poly-glutamine (polyQ) and poly-glycine (polyG) motifs were identified to be important for the activation of oncogenes in AR+ PCa. To target both canonical and non-canonical oncogenic functions by EZH2, we generated EZH2 small-molecule degraders including MS177 and MS8815 using the Proteolysis Targeting Chimera (PROTAC) technology. Our extensive preliminary studies have demonstrated that MS177 effectively degrades both PRC2:EZH2 and non- PRC2 partners of EZH2 (e.g., AR/AR-V7), thus suppressing both activities of EZH2 in canonical and non-canonical oncogenic PCa cells. Importantly, our preliminary results also show that MS177 is superior to all available enzymatic inhibitors of EZH2 in cell line models of PCa including CRPC. Thus, we hypothesize that: (i) EZH2 drives a non-canonical program via EZH2TAD:ARPolyQ/G interaction for oncogene activation, which operate in parallel with the canonical PRC2:EZH2-driven repression of TSGs and that (ii) EZH2 PROTACs simultaneously repress both canonical (PRC2/EZH2) and non-canonical (EZH2/AR) oncogenic pathways, providing a novel and more effective therapeutic strategy for lethal PCa. Dissection of the mechanisms underlying EZH2-mediated oncogenesis and evaluation of the in vitro and in vivo efficacy of EZH2 PROTACs using various PCa preclinical models will have significant impact on improving treatments of lethal PCa. Towards this goal, we will further characterize such a new non-canonical oncogenic role of EZH2 in CRPC (Aim 1a) and define effects of EZH2 PROTACs on suppressing both canonical and non-canonical oncogenic activities of EZH2 (Aim 1b). We will also determine in vitro and in vivo therapeutic effects of EZH2 PROTACs by employing independent PCa models (Aim 2). Completion of the proposed research will not only provide novel mechanistic understanding of how advanced PCa develop, but also validate an innovative therapeutic strategy and generate novel lead compounds for the treatment of CRPC patients.

NIH Research Projects · FY 2026 · 2026-03

ABSTRACT Infants born extremely preterm commonly experience poor growth and nutrient accretion in the postnatal period, which is independently associated with adverse neurodevelopmental outcomes. We hypothesize that altered development of the microbiome contributes to postnatal growth failure in extremely preterm infants, and that modification of the microbiome may be an effective approach to prevent growth failure. This hypothesis is supported by our prior work that demonstrates that extremely preterm infants with postnatal growth failure have differences in the development of the intestinal microbiome compared to extremely preterm infants with appropriate postnatal growth trajectories, including lower bacterial richness, persistent dominance of Enterobacteriaceae, and a delayed pattern of maturation. Further, we determined that these differences in microbiome maturation correlate with elevations in circulating markers of lipolysis and fatty acid oxidation (ketones, fatty acids, glycerol, acylcarnitines) in infants with growth failure. In recent preliminary studies, we found that colonization of neonatal gnotobiotic mice with microbiotas from extremely preterm infants with growth failure induced postnatal growth impairment and a catabolic signature strikingly similar to that observed in preterm infants with growth failure, including elevations in ketones and markers of fatty acid oxidation. Further, we found that in neonatal mice colonized at birth with microbiomes from infants with growth failure, postnatal treatment with microbial consortia from preterm infants with appropriate growth prevented growth impairment. The overall objective of this proposal is to identify the mechanisms of the microbiome’s effects on postnatal growth and host lipid metabolism. Our rationale is that delineating the specific bacterial signals and host responses underlying these effects could uncover new strategies to prevent postnatal growth failure. In Aim 1, we will identify the mechanisms by which the microbiomes of extremely preterm infants with and without growth failure alter host lipid metabolism and postnatal growth. In Aim 2, we will identify mechanisms of modulating the microbiomes of extremely preterm infants with growth failure to prevent growth impairment. The outcomes of this work are expected to significantly advance the field by delineating how host responses to the microbiome contribute to postnatal growth failure, the microbial mediators of these effects, and mechanisms of microbiome modification to prevent postnatal growth impairment. These results have high potential to lead to new microbiome-targeted strategies to support postnatal growth and development of extremely preterm infants.

NIH Research Projects · FY 2026 · 2026-03

Hypertension impacts up to 40% of adults worldwide with disproportionate effects in low-income countries. Uncontrolled hypertension is a leading driver of cardiovascular and chronic kidney disease (CKD). Blood pressure remains uncontrolled in up to half of hypertensive patients, highlighting the need for novel therapies. The profound success of SGLT2 inhibitors in reducing cardiovascular events, particularly among patients with kidney disease, highlights the remarkable health benefits that can accrue from disrupting solute reabsorption in the proximal tubule of the kidney. The dominant sodium transporter in this segment, Na/H-exchanger isoform 3 (NHE3), is also expressed in the loop of Henle. Inflammatory mediators can stimulate NHE3 activity, and systemic inflammation among patients with hypertension and cardiovascular disease is widely prevalent. Nevertheless, systemic immunosuppression in these patients to reduce blood pressure or target organ damage yields unacceptable risks of infection. Thus, there is an urgent need to elucidate more selective immune- mediated pathways that function locally in the kidney to drive NHE3-dependent hypertension. Human variants in the gene encoding the ubiquitin editor A20 (TNFAIP3) associate with several autoimmune diseases. In leukocytes, A20 limits NF-B-dependent activation. However, its direct actions in cells of injured target organ tissues are not well understood. The objective of this proposal is to define novel mechanisms through which A20 regulates the hypertensive response via selective actions in the kidney. A20 suppresses the generation of several inflammatory cytokines that modulate renal sodium transport. These cytokines are upregulated in humans who have incipient hypertension, suggesting a possible local role for A20 in the tubule to modulate blood pressure and renal damage via local renal mechanisms. Using a preclinical mouse model of hypertension with reduced nephron mass, we have generated preliminary data showing that A20 in the kidney tubule can attenuate blood pressure elevation and limit renal abundance of the NHE3 sodium transporter. Our central hypothesis is that A20 in kidney tubular cells ameliorates hypertension by constraining renal NHE3 accumulation.

NIH Research Projects · FY 2026 · 2026-03

PROJECT SUMMARY/ABSTRACT Transposons, comprising nearly 50% of the human genome, are powerful agents of genomic instability and evolutionary change. Despite their prevalence, the mechanisms regulating transposon activation, their interactions with host systems, and their contributions to health and disease remain largely underexplored. Our research leverages innovative tools, including single-cell resolution reporters and genome-wide approaches, to uncover the mechanisms of transposon regulation and their impacts on host biology. By using Drosophila as a powerful genetic system, our previous work has revealed how transposons hijack host DNA repair machinery for replication and how programmed retrotransposon activation shapes host immunity during development. We have also identified novel silencing factors that protect germline and somatic tissues from retrotransposon mobilization, demonstrating the pivotal role of transposon control in genome stability. These findings have laid the groundwork for uncovering host-transposon dynamics and their evolutionary significance. Our future research will expand on these discoveries by exploring the mechanisms underlying transposon regulation, particularly how host factors influence retrotransposon activation and silencing. We aim to identify new regulatory factors through genome-wide screens and dissect their roles in maintaining genome integrity and preventing mobilization. Additionally, we will investigate how programmed activation of retrotransposons can confer adaptive advantages, such as enhancing immune responses or shaping host genome variation. By integrating cutting- edge genomic, genetic, and proteomic approaches, we aim to uncover fundamental insights into transposon biology and its implications for health and disease.

NIH Research Projects · FY 2026 · 2026-03

ABSTRACT Older women (age 65 years or older) experience a high burden of breast cancer incidence and mortality along with complex health issues that impact symptom burden, health-related quality of life (HRQoL), and treatment outcomes. Adjuvant endocrine therapy with an aromatase inhibitor (AI) is a crucial component of treatment for older women with hormone receptor positive (HR+) breast cancer. The adverse effects of AI therapy lead to impaired functioning, reduced HRQoL, and treatment discontinuation among older women. AI-associated arthralgia is a major concern, with more than 70% of older women experiencing moderate to severe AI- associated arthralgia that significantly interferes with daily functioning. Adverse effects of AI therapy also exacerbate the impacts of geriatric specific issues (e.g., low functional status), which are prevalent among older women with breast cancer. When geriatric specific issues and the adverse effects of AI therapy go unaddressed, older women with breast cancer suffer from progressive disability, excess treatment toxicity, and avoidable mortality. The proposed hierarchical 2 x 2 factorial randomized controlled trial (RCT) will test the separate and combined benefits of two interventions that have strong foundations in rigorous prior research: a clinic and provider-level Geriatric Oncology Assessment to Link with Support for AI therapy (GOAL-AI) intervention and a patient-level CBT-based coping skills training intervention for older women taking AI therapy (CST-AI). Guided by the Expanded Chronic Care Model and recommendations for improving the care of older adults with cancer, the GOAL-AI intervention integrates geriatric oncology training, brief geriatric assessment, and geriatric assessment-guided recommendations and referral resources. The CST-AI intervention teaches patients skills for managing AI-associated arthralgia and other symptoms, and strategies for engaging in health care to address geriatric specific issues. We will randomize 16 oncology practice sites to GOAL-AI or enhanced usual care (EUC), stratifying by practice site location (i.e., urban vs. rural). Within each site, participants will be randomly assigned to the CST-AI intervention or EUC. Randomization will establish four groups: 1) GOAL-AI + CST-AI, 2) GOAL-AI alone, 3) CST-AI alone, and 4) EUC (i.e., no intervention control group). Participating patients will be 452 older women (age ≥ 65 years) who are within 12 months of initiating adjuvant AI therapy for HR+ non- metastatic breast cancer. Study outcomes will be assessed at baseline, 6, and 12 months. We propose the following aims: 1) examine whether GOAL-AI + CST-AI, GOAL-AI alone, or CST-AI alone improve pain severity and interference, geriatric and AI-associated symptom burden, and HRQoL compared to EUC; 2) evaluate whether GOAL-AI + CST-AI, GOAL-AI alone, or CST-AI alone improve patient activation and self-efficacy compared to EUC and whether these variables at least partially mediate changes in study outcomes; and 3) examine GOAL-AI and CST-AI implementation-related factors using quantitative and qualitative approaches.

NIH Research Projects · FY 2026 · 2026-03

Abstract: Metalloenzymes frequently catalyze chemically challenging radical-mediated reactions necessary for the synthesis of the complex structures of natural products and cofactors. Their functional and mechanistic characterization has been a critical basis for the discovery and development of small molecule therapeutics, research probes, and biocatalysts. However, many metalloenzymes in natural product and cofactor biosynthetic pathways remain significantly under-explored because (1) characterization of metalloenzymes requires specialized expertise and techniques due to the unique reactivities and properties of their metallo- centers, (2) their substrates are frequently not readily available, and (3) their functions may not be apparent from their primary amino acid sequence or structural homologies. Consequently, many metalloenzymes remain under- or unexplored, leaving numerous mechanistic questions unanswered. The long-term goal of my group is to elucidate the mechanisms and functions of metalloenzymes in cofactor and natural product biosynthesis. This application combines two ongoing NIGMS R01 projects and focuses on O2-independent and dependent Fe enzymes: Radical S-adenosyl-L-methionine (rSAM) enzymes and non-heme iron-dependent oxygenases/oxidases. The proposed structural and mechanistic characterizations will reveal novel catalytic functions and mechanisms of (metallo)enzymes and address key questions in each enzyme family, including the mechanism of substrate-triggered radical initiation by rSAM enzymes and the mechanism of C-H or O-H activation by mono or dinuclear Fe cofactors. The results of these studies will uniquely bridge the knowledge gap between bioorganic and bioinorganic chemistry and broadly impact the fields of cofactor and natural product biosynthesis and metalloenzymology.

NIH Research Projects · FY 2026 · 2026-03

ABSTRACT Neurological studies greatly benefit from functional brain imaging to investigate brain activity and understand the underlying mechanisms of healthy and disordered behavior. Studying non-human primates (NHP) stands at the forefront of neurological research, offering unparalleled insights into complex brain activities and advanced cognitive and behavioral processes. Utilizing optical or ultrasound technologies, researchers have developed various brain imaging methods that are versatile for studying a range of awake and behaving applications. However, these tools have been restricted to only small animal models. Recently, photoacoustic tomography (PAT) has emerged as a promising non-invasive, label-free technique capable of mapping deep brain hemodynamic functions by acoustically probing the brain’s optical contrast. Yet, the efficacy of traditional photoacoustic imaging is compromised by the strong acoustic aberration induced by the NHP’s dense, curved, and thick skull. Furthermore, current PAT systems are ill-suited for awake NHP imaging, due to their unwieldy size and complex operation. In this proposal, we will transcend these limitations by developing a four-dimensional smart epidermal photoacoustic tomography (4D-SEPAT) technology, allowing for real-time, longitudinal, functional brain imaging in behaving NHPs. The proposed 4D-SEPAT technology will leverage state-of-the-art epidermal electronics, which combines soft ultrasound transducer array and high-precision shape mapping. Most importantly, the soft ultrasound transducer array, with integrated shape sensor and high-power laser source, can closely adapt to the contour of the NHP head, effectively mitigating the skull’s aberration effect. Powered by fast 3D image reconstruction, 4D-SEPAT will enable real-time transcranial brain imaging while maintaining its high sensitivity to hemodynamic functions. With full conformability to the head, 4D-SEPAT is insensitive to motion artifacts, can be longitudinally applied to behaving NHPs, and allow for deep-brain analysis of sensory and cognition. To achieve this objective, we will pursue the design, development, and validation of the proposed 4D- SEPAT system in Aim 1 and Aim 2, and demonstrate its imaging performance in awake rhesus macaques during visual-oculomotor behavior in Aim 3. Success of this 4D-SEPAT technology has the potential to revolutionize the way the brain is studied, diagnosed, and treated, by providing non-invasive, longitudinal, real-time mapping of deep brain functions.

NIH Research Projects · FY 2026 · 2026-03

ABSTRACT Home and community-based services are the most common support for people with Intellectual and/or Developmental Disability (I/DD) who otherwise would require facility-level care. Many state Medicaid waivers allow participants to choose to hire family as direct support personnel to meet their care needs at home and in the community rather than a professional aide. Ability to hire family increased during the Covid-19 pandemic through state policy changes, and yet we do not know how involving paid family affects the ability for people with I/DD to remain at home. The objectives of this study are to use North Carolina (NC) as a case to elucidate the experiences of waiver participants with I/DD (Innovations Waiver participants). With no national data fields systematically identifying self-direction status or who is paid for personal care, foundational state-level work is required to understand people with I/DD’s experiences with self-direction. First, we will describe prevalence and dynamics of paid family care using Medicaid administrative data over the past 9+ years, including patterns by self-direction or not, by individual and geographic factors and by era (pre-post Covid-19 (Aim 1). Second, qualitative approaches will center the voices of people with I/DD and their families to obtain their perspectives on what is gained and what is lost from self-directed care including paid family care (Aim 2). Specifically, photo elicitation, case study, and focus group interviews will examine the lived experience of accessing and receiving care through the Innovations Waiver according to 1) individuals with I/DD, 2) their parent/partner/guardian/unpaid family caregiver, 3) their paid family caregivers, 4) their paid aide, and 5) Innovations Waiver experts. Third, we will estimate the comparative effectiveness of paid family care versus paid aide care only on person-centered outcomes (e.g., home time, preventive care) and on potential harms (potentially harmful medications, injurious falls, mistreatment) (Aim 3). We hypothesize that waiver participants with paid family care will have better person-centered outcomes and no increase in harms compared to those with paid aides alone. Effects of self- direction will also be explored. By using a convergent parallel mixed-methods process we will integrate results to paint a full picture of the comparative effectiveness of paid family care and self-direction from childhood to older adulthood, including identification of any harms. The results of this 5-year R01 study will be immediately applicable to state Medicaid office benefit design and inform strategies to optimize quality of care and life for people living with I/DD from childhood throughout the lifespan. Results from the North Carolina case will also position us to pursue a national study, given knowledge gained along with emerging efforts to identify “self- direction” in national CMS data sets. Examining paid family care and self-direction’s effects across the lifespan aligns with NIA’s strategic goal to improve the health, well-being, and independence of adults as they age.

NIH Research Projects · FY 2026 · 2026-02

SUMMARY Copper (Cu) is an essential trace element and a catalytic cofactor for a variety of enzymes involved in cell growth, development, and stress resistance. Fungal systems have been instrumental in identifying mechanisms of Cu utilization and homeostasis in diverse eukaryotic cells. The experiments in this proposal will use a relevant fungal model system to explore Cu homeostasis in microbial physiology and pathogenesis. Cryptococcus neoformans is an opportunistic fungal pathogen that is responsible for >100,000 deaths annually, especially among patients with poorly treated HIV infection. This pathogenic microorganism is an outstanding model to study rapid cellular adaptations between high and low Cu states. This fungus has sophisticated transcriptional and post-transcriptional Cu regulatory mechanisms to adapt to Cu overload and Cu limitation in biologically relevant niches. C. neoformans first infects the host lung where it is engulfed by alveolar macrophages and experiences toxic Cu bombardment within the phagolysosome. In immunocompromised hosts, this fungus disseminates through the bloodstream to the central nervous system where it causes lethal disease, and where it encounters extreme Cu starvation. Failure to acquire Cu, or to adapt to low bioavailable brain Cu levels, results in defective fungal survival in the host. Therefore, during the infection cycle this microorganism must be able to rapidly transition from a Cu-resistant state during toxic Cu exposure to a Cu-foraging state during Cu limitation. The C. neoformans Cuf1 transcription factor is the primary regulator of the cellular response to both high and low Cu levels. We and our prior collaborators therefore characterized the Cuf1- dependent transcriptome and used this data to define conserved and novel mechanisms of eukaryotic copper biology. These studies identified important cellular responses to Cu limitation (increasing Cu import) as well as to Cu excess (inducing Cu detoxification processes). In the experiments outlined in this proposal, we will build upon this rigorous experimental foundation to explore new intersections between Cu homeostasis and microbial pathogenesis. These include (SA1) exploring our new observations regarding the roles of Cu homeostasis and Cu-containing enzymes in cell wall/cell membrane function; and (SA2) defining the impact of failed copper regulation on microbial pathogenesis. Together, these studies will advance new concepts for how microbial cells acquire the essential metal ion Cu, invoke adaptive responses to Cu toxicity, and regulate these processes to mediate interactions with the infected host.

NIH Research Projects · FY 2026 · 2026-02

Mitochondrial function is fundamental to health, as soberingly demonstrated by the wide range of debilitating pathologies and diseases faced by patients with genetically-based mitochondrial deficiencies. Chemical exposure can directly affect mitochondrial function in many different ways, and can also dramatically exacerbate underlying genetic conditions. Because mitochondrial toxicants are common among pollutants, understanding the detailed mechanisms by which these chemicals affect mitochondria, and their interactions with genetic differences, is a highly significant environmental health problem. Unfortunately, the molecular and cellular impacts of different, specific mechanisms of mitochondrial toxicity are poorly understood, and the interactions of such exposures with genetic deficiencies are even more obscure. These knowledge gaps hinder testing, regulatory, and personalized medicine efforts. We seek to understand 1) the biological pathways that respond to mitochondrial DNA damage, especially irreparable mitochondrial DNA damage; 2) the molecular and cellular consequences of different, specific kinds of stressor-mediated changes to mitochondrial integrity and function; 3) how the effects of mitochondrial toxicants change in the context of human mitochondrial disease genes. To address these knowledge gaps, we are employing a translational approach integrating a powerful mechanistic in vivo laboratory model, Caenorhabditis elegans, with cell culture experiments including cells from mitochondrial disease patients. We will continue to define the molecular and cellular outcomes of different kinds of mitochondrial toxicity, and to identify novel biological pathways that defend mitochondria and the mitochondrial genome. We will test how deficiencies in these genes, and genes known to cause mitochondrial disease, alter sensitivity to mitochondrial toxicants, in both C. elegans and patient-derived cells. We are developing powerful transgenic tools that allow us to measure key aspects of mitocondrial function and dysfunction, including energetics, morphology, and alterations to redox tone, in a cell-specific fashion and in vivo, in C. elegans. We are deriving iPSCs and differentiated, mitochondrial disease-relevant cells for testing gene-environment interactions. Given the prevalence of both mitochondrial gene variability and mitotoxicant exposure, understanding the detailed mechanisms by which mitotoxicants act, the impacts of the genetic deficiencies, and their interactive effects will be impactful for many people. Addressing these scientific challenges will be transformative for human health by providing mechanistic knowledge critical to understanding cellular and higher-level effects of mitochondrial toxicity, informing treatment and personalized intervention options, and developing appropriate testing and regulatory strategies.

NIH Research Projects · FY 2026 · 2026-02

ABSTRACT Plastic additives are widely used in consumer products, yet thousands of plastic additives remain uncharacterized for their potential to disrupt endocrine function - posing significant public health risks. This project aims to develop an integrated computational (Aim 1) and experimental (Aim 2) workflow to systematically predict and validate the endocrine-disrupting potential of plastic additives. In Aim 1, we will design novel machine learning models trained on publicly available datasets to predict AR and ERα modulating activity of plastic additives and then used to predict the potential effects of all plastic additives to select the most promising based on novelty and predictive uncertainty for further in vitro and in vivo testing. In Aim 2, we will validate our predictions through a multi-step experimental characterization approach using our in-house AR and ERα assays, followed by dose-response studies in AR- and ERα-responsive cell lines to measure target gene activation and cell proliferation. The top three plastic additives with the strongest in vitro effects will be further evaluated in vivo using mice to assess systemic hormonal changes caused by the plastic additives. This work will have a substantial positive societal impact by establishing a first-in-kind machine learning-assisted predictive toxicological model to pinpoint plastic additives of highest concern to induce adverse health effects as well as generate a large dataset of plastic additive effects on endocrine function. Taken together, this work can serve to provide policy guidance on plastic additives to ban or remove from products, with potentially beneficial health outcomes for billions of consumers.

NIH Research Projects · FY 2026 · 2026-02

Abstract There remains an urgent need for expanded understanding of the pathophysiology of invasive fungal infections (IFI), and improved tools for diagnosis and management of these devastating diseases. Invasive candidiasis alone causes significant morbidity and mortality, especially in solid organ transplant patients where up to 40% of these infections result in death. Given the unacceptably high mortality rates, it is a priority to identify patients who may need more aggressive antifungal treatment or surgical debridement. Conversely, earlier cessation of antifungal therapy in patients deemed at lower risk of infective sequelae also reduces risk of medication interaction (particularly between immunosuppressants and antifungal therapy), side-effects, and antifungal resistance. Despite this, there are limited tools to accurately diagnose and predict the severity and outcomes of these infections. Microbial pathogen-based measures are slow, poorly sensitive, reliant on a high volume of circulating pathogen, and poorly suited to define treatment response. We and others have developed compelling evidence that examination of transcriptional responses in circulating immune cells can provide solutions to these dilemmas. Utilization of host gene expression patterns as biomarkers is a promising approach with proven efficacy as a diagnostic and prognostic tool in a myriad of infectious and noninfectious syndromes. We have previously developed a gene expression signature that accurately diagnoses Candidemia with a high degree of accuracy (auROC 0.94). In this project we proposed to utilize this diagnostic signature, made up of canonical elements of the immune response to Candida spp, to define the longitudinal evolution of the host response to invasive candidiasis, in order to develop prognostic markers and correlates of treatment response. Serial blood samples of patients with a confirmed diagnosis of invasive candidiasis from three pre-existing biobanks will be utilized. We will perform bulk RNA- sequencing on whole blood samples, and expression patterns (or signatures) will be described and correlated with clinical outcomes, specifically time to resolution, disease progression and mortality. Results of these endeavors will support development of PCR-based noninvasive diagnostics that offer earlier, more directed (and thus more effective) therapy, and reduce the need for potentially harmful empiric antimicrobial agents, thus offering a true paradigm shift in the way invasive candida infections are managed in these high-risk populations.

NIH Research Projects · FY 2026 · 2026-02

Most healthy individuals are persistently infected with the human polyomavirus BK (BKPyV) without significant consequences, yet in immunocompromised hosts such as kidney transplant or bone marrow transplant recipients, BKPyV is associated with significant morbidity. In particular, kidney transplant patients with active BKPyV are at highest risk to develop BKPyV-associated nephropathy (BKVN), one of the leading causes of graft loss. There is currently no effective treatment to prevent or treat BKPyV-associated diseases and recent clinical trials have focused on immune-based therapies, including infusion of immunoglobulins and allogeneic BKPyV- specific T cells. These strategies have been successful in subsets of patients but have shown limited efficacy in others, suggesting other immune factors contribute to control of BKPyV. In particular, Natural killer (NK) cell- based immunotherapy has the potential to help treat BKPyV-associated diseases in immunosuppressed individuals. However, NK cell responses to BKPyV are poorly understood. Our published and ongoing work is providing new insights into NK cell responses to polyomaviruses. Specifically, our preliminary data now show robust NK cell responses to BKPyV, modulation of NK cell responses against polyomaviruses via receptors that are checkpoint targets for cancer immunotherapies, and clonal expansion of single NK cells with potent cytotoxic responses against BKPyV. Importantly, we also identified significant differences in NK cell phenotypic profiles between kidney transplant recipients who control BKPyV following reactivation and those who don’t, with control being associated with lower expression of terminal differentiation markers and higher expression of key activating receptors. In this proposal, we will test the overarching hypothesis that specific subsets of NK cells mediate potent antiviral responses against BKPyV and protect immunocompromised patients against BKPyV- associated diseases. We will test our hypotheses using longitudinal samples from kidney transplant recipients and heathy donors, BKPyV-infected human renal proximal tubule epithelial cells and kidney organoids, through three focused independent Aims: (i) Characterize NK cell antiviral responses to BKPyV, (ii) Determine how mutations in BKPyV antigens modulate NK cell antiviral function, and (iii) Define NK cell signatures that are associated with control of BKPyV in kidney transplant recipients. If successful, the results of these innovative studies will contribute new knowledge of human NK cell immune responses against BKPyV and guide the development of novel immunotherapeutic approaches aimed at harnessing NK cells to target BKPyV.

NIH Research Projects · FY 2026 · 2026-02

Abstract Up to 3.5 billion dollars is spent each year, in the US alone, to manage bladder infections and their recurrence. The intractability of these infections is attributable, in part, to the capacity of bladder bacteria to persist in the bladder long after infection resolution only to remerge as another flareup. In view of the fact that antibiotics are largely ineffective there is a dire need for alternate approaches to protect against these infections. We have found that a single bladder treatment of live attenuated Bacillus Calmette-Guerin (BCG), an FDA approved therapy for bladder cancer, along with a peptide capable of disrupting the superficial epithelium was highly efficacious in protecting against uropathogenic E.coli (UPEC) mediated infection. Since tumor fighting properties of BCG is associated with their persistence in the bladder and their capacity to recruit Th1 cells, the protection against UPEC is probably also mediated, at least in part, by Th1 cells. We hypothesize that we can prolong the duration of this protection by simultaneously inducing bladder recruitment of E.coli specific Th1 cells which typically have an indefinite life span. Thus, we believe that supplementing the BCG and epithelial disruptor peptide formulation with UPEC antigens would markedly prolong and enhance the protective capacity of BCG immunotherapy against UPEC bladder infections. The following specific aims are proposed: (i) Determine cellular targets of BCG uptake and persistence in the bladder, (ii) Identify the immunological basis of cross-protection induced by co-treatment with BCG bladder and epithelial disruptor peptide (iii) Evaluate the efficacy and durability of intravesical administration of an optimized formulation comprising of epithelial disruptor peptide/BCG/ UPEC antigens in combating recurrent bladder infections. It is expected that these studies will reveal a distinct, effective, and long-lasting strategy to combat intractable bladder infections. Since this strategy involves repurposing a currently utilized bladder therapy, its implementation could be relatively rapid.

NIH Research Projects · FY 2026 · 2026-02

PROJECT SUMMARY/ ABSTRACT Iron deficiency is the most common micronutrient deficiency worldwide, affecting 25% of the population. In older adults with heart failure (HF), its prevalence ranges from 40–50% and is associated with worse health status, functional limitations, and increased risks of hospitalizations and death. Once considered only a reversible cause of anemia, iron deficiency is now recognized as an independent comorbidity in HF, even without anemia. Intravenous (IV) iron has proven effective in improving health status and functional capacity in HF with reduced ejection fraction (HFrEF), and guidelines now recommend IV iron for iron-deficient HFrEF patients. However, patients with HF with preserved ejection fraction (HFpEF)—the fastest-growing HF phenotype, with a higher prevalence of iron deficiency—remain largely excluded from these studies. HFpEF affects over 3 million older adults in the U.S., contributing to poor health status, functional impairment, and hospitalizations. To address this critical knowledge gap, we propose the Iron Needed for Function and Undesirable Symptoms in Elderly with HFpEF (INFUSE-HFpEF) study, a double-blind, placebo-controlled randomized clinical trial of IV iron in 300 ambulatory older patients with HFpEF and iron deficiency. This collaboration among Duke University/Duke Clinical Research Institute (DCRI), Kaiser Permanente Northern California (KPNC), Mount Sinai Health System (MSHS), and St. Luke’s Mid America Heart Institute (MAHI) will evaluate the net clinical benefit of IV iron on patient-centered outcomes, including a composite of death, all- cause hospitalization, and health-related quality of life (QoL), as well as its impact on physical function. The INFUSE-HFpEF study incorporates three innovative approaches. First, it employs a patient-centered strategy by prioritizing patient-reported outcomes such as QoL alongside traditional endpoints, reflecting their growing importance in HFpEF trials and FDA endorsement for regulatory approval. Second, it adopts a pragmatic design with broad eligibility criteria and diverse representation, embedding the trial within routine care and leveraging electronic health records to address gaps in inclusion of older adults, women, and ethnically diverse populations. Third, it incorporates novel physical activity measures as secondary endpoints, using wearable- derived metrics such as step count and gait speed. These measures will be validated against established gold standards, including the 6-minute walk test (6-MWT) and cardiopulmonary exercise testing (CPET) parameters like peak VO₂, aligning with the FDA’s Digital Health Innovation Action Plan. INFUSE-HFpEF will advance understanding of iron deficiency in HFpEF, addressing NIA’s Strategic Goal C-3 to “develop interventions for treating…or mitigating the impact of age-related diseases and conditions.” This trial has the potential to transform care for iron-deficient HFpEF patients, bridging evidence gaps and improving patient-centered and clinical outcomes, while serving as a model for research in aging populations with complex multimorbidity.

NIH Research Projects · FY 2026 · 2026-02

ABSTRACT People with HIV-1 (PWH) on antiretroviral therapy (ART) are prone to experiencing chronic inflammation despite effective viral suppression. This sustained inflammation has been linked to an elevated risk of developing a variety of age-associated comorbidities, including chronic kidney disease (CKD). Preliminary data supporting this study demonstrates multiple distinct inflammatory endotypes among aging PWH on ART, several defined by levels of chemokine C-C motif ligand 2 (CCL2), a critical mediator and biomarker of kidney injury and disease. However, the precise cellular and molecular inflammatory immune endotypes in PWH that could lead to disease remain undefined. Moreover, how endotypes defined by circulating inflammatory markers impact organ- compartmentalized inflammation and the functional and molecular states of immune cells are unknown. The overall objective of this project is to define the early cellular and molecular signatures of inflammation associated with the progressive development of CKD in PWH on ART in both blood and urine. This will be achieved by comprehensively defining plasma inflammatory endotypes in a retrospective cohort of aging (50+ years) PWH on ART, sampled as they progressed from early to later stage kidney disease, with comparison to PWH on ART with normal renal function. In addition, systems immunology will be used to characterize the soluble and cellular inflammatory profiles of peripheral blood and urine in a prospective cohort of 200 aging PWH on ART. Urine, a readily accessible non-invasive biofluid, contains proteins and viable cells originating from the kidney that can serve as indicators of renal inflammation and overall kidney function. A combination of advanced machine learning approaches, clinical tests, and human kidneys-on-chips models will be applied to define the relationship between systemic, urinary, and renal cell inflammation and dysfunction in PWH on ART that are associated with onset of CKD. This project will test the hypothesis that specific inflammatory immune endotypes can be identified in PWH on ART that promote the activation and dysregulation of immune and kidney cells, contributing to the development of CKD. The hypothesis will be tested, and the overall objective achieved, with completion of three Specific Aims. Aim 1 will define plasma inflammatory endotypes of aging PWH on ART and identify signatures that predict CKD. Aim 2 will identify how plasma inflammatory endotypes impact the cellular, metabolic, and functional programs in blood and urine of aging PWH on ART. Finally, Aim 3 will determine the mechanisms of activation of renal inflammatory programs using kidneys-on-chips. This research will identify specific endotypes of inflammation associated with development pf CKD and uncover pathways driving chronic inflammation in the blood and urine that promote kidney injury and disease. This knowledge will enable the development of CKD risk prediction tools and of therapeutic interventions like CCL2 signaling inhibitors to reduce inflammation-related kidney disease and other comorbidities in this expanding population.

NIH Research Projects · FY 2025 · 2026-01

Project Summary/Abstract Triple-negative breast cancer is an aggressive subtype of breast cancer that tends to metastasize to distant organs early in the course of the disease. Metastasis can occur by cancer cells from the breast tumor directly entering the blood to travel to distant organs (hematogenous metastasis), or by cancer cells spreading from the breast tumor to nearby lymph nodes, and then leaving the lymph nodes to enter the blood and travel to distant organs (lymphatic metastasis). We are interested in understanding how lymphatic metastasis occurs. Cancer cells change their gene expression in order to promote metastasis. One way that cancer cells can regulate gene expression is through the action of noncoding RNAs. We found that the noncoding RNA Snord67 promotes lymphatic metastasis in a mouse model of triple-negative breast cancer. Snord67 is a small nucleolar RNA (snoRNA) that leads to the 2′-O-methylation of another noncoding RNA, U6 small nuclear RNA (snRNA). However, it is unknown whether Snord67 promotes metastasis by increasing the methylation of U6 snRNA or through other, unknown mechanisms. For example, some snoRNAs are known to promote the 2′-O-methylation of protein-coding messenger RNAs (mRNAs), which can lead to changes in mRNA processing, stability, or translation into proteins. The goal of this proposal is to determine whether Snord67 promotes lymphatic metastasis by increasing the methylation of U6 snRNA or by performing other functions, such as increasing the methylation of mRNAs. This proposal includes two aims: (1) Determine the functional importance of Snord67- guided 2′-O-methylation of U6 snRNA at C60 (canonical mechanism) in the promotion of lymphatic metastasis by Snord67, and (2) Determine the contribution of Snord67-guided 2′-O-methylation of target mRNAs (non- canonical mechanism) to the promotion of lymphatic metastasis by Snord67. These studies will help us better understand how triple-negative breast cancer spreads and could potentially lead to new approaches for treating or preventing breast cancer metastasis. The trainee will work in a highly productive multidisciplinary scientific environment and develop proficiency in a variety of experimental methods in RNA biology and cancer biology under the mentorship of physician–scientist experts.

NIH Research Projects · FY 2025 · 2026-01

Despite recent regional control efforts that aimed to eliminate malaria by 2022, reduction in malaria transmission has stalled in Central America and in many cases is on the rise. In the Darien and Guna Yala regions of Panama, an extreme increase in international migration combined with potential changes in malaria vectors and vector ecology has coincided with increasing cases detected in the country. The majority of new malaria cases are occurring in indigenous communities, particularly near areas where transiting migrants are housed. These events – migration through Panama, temporary housing in indigenous communities, introduction of new malaria vectors, and increasing malaria rates – are likely interrelated and require further investigation to explain these new dynamics of malaria transmission. The overall goal of this project is to support interdisciplinary training and research experience in vector-borne disease epidemiology, migration, and environmental health, including the roles of hydrometeorological variables and land-use and land-cover (LULC) change in influencing disease ecology. The applicant’s long-term career goal is to become an independent academic researcher with expertise in vector-borne disease and One Health, with specific expertise in quantifying the effects of hydrometeorology and LULC changes on disease development. The proposed fellowship will support mentored training in (1) the assessment of human behavior with case-control survey design, data collection, and processing, (2) the application of entomological collections, morphological and genetic identification, and data processing, and (3) the development of Bayesian spatial-temporal models to assess local malaria risk. The described training will include coursework, one-on-one mentoring, fieldwork, extensive application of statistical methods, participation in workshops/conferences, and the publication of a minimum of 3 first-authored manuscripts. Two research aims are proposed for this application: (1) evaluating the effect of temporary migrant housing on malaria transmission in local communities using a longitudinal case-control study design; and (2) conduct mosquito collections to characterize the composition and distribution of vector species within communities with and without transiting migrants. Both aims will employ Bayesian models and assess the role of hydrometeorological and LULC change variables. These aims will test the main hypothesis that migration and vector ecology are contributing to increased malaria transmission in Panama, and that environmental change is impacting this dynamic. The proposed research leverages two externally funded studies supporting malaria elimination in Panama, including a longitudinal study evaluating risk factors associated with elevated malaria risk in the Darien. The proposed study will be a subset of the parent project. Ultimately, the proposed projects will provide critical insight into this new malaria landscape in Panama and will help inform malaria control efforts. Further, they will provide the applicant with intersectional training and unique expertise for a career in One Health and vector-borne disease epidemiology.

NIH Research Projects · FY 2026 · 2026-01

Project Summary Alzheimer’s Disease (AD) is a progressive neurodegenerative disorder which currently affects nearly 7 million people nationwide. Although the causes of AD are still not completely understood, it is clear that a combination of genetic, environmental, and age-related factors contribute to disease development and progression. However, how these diverse factors coalesce to lead to a pathogenic state remains poorly understood. Moreover, the molecular changes that occur in distinct cell types in the brain in AD and how brain aging affects these changes are not fully known. Of relevance to these critical unanswered questions, recent studies have revealed the importance of RNA regulatory pathways during both brain aging and AD. In particular, chemical modifications to mRNA have emerged as important regulators of gene expression in the brain with links to AD in both human and animal studies. The most abundant internal mRNA modification is adenosine methylation (m6A), which is found in thousands of mRNAs in the brain and which plays critical roles in regulating mRNA processing and expression in cells. Dysregulation of m6A leads to defects in neurodevelopment, learning and memory, and synaptic plasticity, and abnormal expression of m6A writer, reader, and eraser proteins has been linked to AD in both humans and mice. Our recent studies have also shown that some AD-associated mRNAs undergo dynamic methylation in distinct cell types in the brain with age. However, the effects of this dynamic methylation on gene expression during aging are unknown. Moreover, how mRNA methylation is differentially regulated within distinct brain cell types during AD has not been explored. Here, we will address both of these gaps in knowledge by determining how age-dependent differential methylation in the brain influences the expression of AD-associated genes and identifying the m6A reader proteins that mediate m6A function with age (Aim 1). Then, we will determine for the first time how m6A is altered in the AD mouse brain at single-cell resolution (Aim 2). Finally, we will investigate the effects of m6A on gene expression in distinct cell types in the AD mouse brain during disease progression (Aim 3). Altogether, these studies will provide important new insights into the role of m6A in regulating cell type-specific gene expression during brain aging and in AD, and they will provide critical new datasets for the research community.

NIH Research Projects · FY 2026 · 2025-12

ABSTRACT The radiation tolerance of the gastrointestinal (GI) tract can be exceeded in radiation accidents or in the case of nuclear warfare, resulting in a lethal GI acute radiation syndrome (GI-ARS). There are no current FDA- approved medical countermeasures to mitigate GI-ARS; thus, there is an urgent need to identify key mechanisms of repair and regeneration in the irradiated intestinal epithelium. We have shown that the regeneration of the intestinal epithelium in response to severe radiation injury is mediated through revival stem cells (revSCs), which are generated through fetal-like reprogramming and identified by a marker gene Clusterin (Clu) (the mouse ortholog of human CLU). Remarkably, whole animal deletion of Clu significantly sensitizes mice to GI-ARS, while GI-ARS in mice is significantly mitigated via treatment with exogenous mouse Clu protein starting 24 hours post-irradiation. In addition, our preliminary data suggest that the cGAS-STING pathway acts as an upstream regulator of Clu in the irradiated intestinal epithelium. Treatment of mice with a STING agonist MSA-2 starting 24 hours after irradiation significantly improves the regeneration of irradiated small intestines. Thus, the overall goal of this proposal is to develop novel mitigators of the GI-ARS that promote the regeneration of irradiated intestinal epithelium by targeting the STING-CLU axis. In Aim 1, we will define the mechanisms by which CLU mitigates the development of GI-ARS. In Aim 2, we will dissect the STING-CLU axis in regulating the development of GI-ARS. Successful completion of the proposed study will demonstrate the key role of CLU in promoting the regeneration of the intestinal epithelium following acute radiation injury in mice and human intestinal organoids. We anticipate that our findings will significantly impact the successful development of GI-ARS mitigators that target the STING-CLU axis.

NSF Awards · FY 2025 · 2025-10

Large Language Models (LLMs) are increasingly deployed as the backbone of real-world applications such as Google Search with AI Overviews and Microsoft Bing Copilot. When data and code are not properly separated within an application, the latter (including AI applications) is vulnerable to cyber-attacks. This project's novelties are twofold: (1) conducting a systematic study to deepen the understanding of such threats, and (2) developing new defenses to mitigate such attacks. Its broader significance and importance lie in establishing foundational security principles for the rapidly growing ecosystem of AI applications, which are now widely deployed across diverse societal domains. Moreover, the released code and materials produced by this project will not only help secure real-world LLM-integrated applications but also serve as valuable educational resources for undergraduate and graduate courses, fostering the next generation of researchers and practitioners in this emerging security area. Security history shows that when data and instructions are not properly separated within a system, injection attacks can emerge—for example, SQL injection attacks in traditional software. Similarly, due to the lack of a clear boundary between instructions and data in prompts, LLM-integrated applications are inherently vulnerable to prompt injection attacks. To understand and mitigate such threats this project adopts a holistic approach comprising three interconnected research thrusts to systematically investigate the security vulnerabilities of LLM-integrated applications to prompt injection attacks and to develop new methods to prevent, detect, and attribute such attacks. The project will also open-source a platform that integrates our developed algorithms along with a comprehensive tutorial on prompt injection attacks and defenses. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

This Research Experiences for Undergraduates (REU) Site, Nanoscale Detectives, offers undergraduate students an immersive research experience investigating the fundamental structure, dynamics, and properties of hybrid perovskite materials — a class of materials with transformative potential in next-generation semiconductor technologies, including solar cells, photodetectors, and advanced computing. By engaging students from various academic disciplines in cutting-edge, hands-on research across the three academic institutions in making up North Carolina’s Research Triangle Area (NC State University, UNC Chapel Hill, and Duke), the program addresses the critical national need to cultivate a highly skilled STEM workforce equipped to tackle complex challenges in materials science, semiconductors and energy security. The REU Site promotes the progress of science through discovery-focused learning, while advancing national prosperity by preparing students for graduate education and careers in high-demand STEM fields. The program also recruits and trains students from community colleges as well as first-generation college students, in support of the nation’s commitment to expanding the STEM workforce in semiconductors and other strategic areas. The Nanoscale Detectives REU Site will host cohorts of twelve undergraduate students each summer for a ten-week research program at NC State University, UNC Chapel Hill, and Duke University. Participants will work closely with faculty and graduate student mentors on interdisciplinary projects that apply advanced synthesis, spectroscopy, microscopy, and computational modeling to elucidate the nanoscale structure and dynamic behavior of hybrid perovskite systems. The program includes a structured professional development curriculum, mentor training workshops, and outreach activities designed to enhance student skills in research communication, ethics, and career planning. Research outcomes will contribute new insights into the stability and performance of perovskite materials, with the potential to inform the design of more efficient and durable devices for renewable energy technologies. Program assessment will combine formative and summative evaluations, with anonymized data analyzed to inform continuous improvement and shared broadly with the undergraduate research community. Through these activities, the REU Site will advance fundamental understanding of hybrid perovskites and prepare the next generation of scientists and engineers to lead innovation in energy and materials science. This Site is supported in part by funds provided to the National Science Foundation by the Semiconductor Research Corporation. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

Non-technical Abstract: This project aims to revolutionize the discovery of new solid-state materials that can precisely control the mobility of ions and electrons, an essential step toward building the next generation of energy storage systems, neuromorphic computers, and smart sensors. By leveraging advanced artificial intelligence (AI), machine learning (ML), and automated synthesis tools, the team will develop a transformative approach to design solid-state ion conductors using multi-element doping, enabling materials tailored for next-generation energy and electronic systems. A central goal is to establish a new data-driven approach to achieve an optimal balance of ion and electron conductivities for targeted applications while ensuring material stability during operation, a task difficult to achieve using traditional trial-and-error techniques. The project will also provide hands-on research and training opportunities in AI-driven materials discovery, fostering collaboration among U.S. and Canadian universities, national laboratories, and industry partners. Technical Abstract: This research will develop and apply a closed-loop, data-driven framework to design and optimize multi-element co-doping strategies in alkali-ion conductors. By integrating AI/ML-accelerated property prediction, high-throughput computational modeling, autonomous synthesis, and in-situ characterization, this project will systematically investigate how co-doping influences ionic transport, electronic structure, and lattice stability across bulk phases, grain boundaries, and interfaces. A fast, iterative inner loop will enable the screening of thousands of dopant combinations, while a slower outer loop will focus on extracting mechanistic insights and ensuring scalability, feeding knowledge back into the predictive models. Target systems include sodium- and lithium-ion based oxides and halides, where varying the balance of ionic and electronic conduction is critical for applications ranging from batteries to neuromorphic computing. The project will generate foundational design rules for tuning transport properties through co-doping, creating new pathways for energy-efficient materials innovation. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.