Office of Research Affairs

Competition Abstracts

Poster 1

Characterizing Candida albicans morphogenesis regulation in the context of host macrophages

Hannah Wilson^1,2 and Michael Lorenz^1,2

Department of Microbiology and Molecular Genetics, McGovern Medical School, University of Texas Health Science Center, Houston, TX, USA¹

The University of Texas MD Anderson Cancer Center UTHealth Houston Graduate School of Biomedical Sciences, Houston, TX, USA²

The opportunistic pathogen Candida albicans is the leading cause of systemic fungal infections globally. The disseminated disease carries a fatality rate approaching 40-50%, despite treatment with antifungal drugs. The main predisposing factor for these infections is an impaired innate immune response. Thus, characterizing the interaction between C. albicans and innate immune cells is critical in order to understand the mechanisms by which C. albicans is able to disseminate in compromised hosts. In particular, the interaction between C. albicans and host macrophages is dynamic and complex. Once phagocytosed, C. albicans engages alternative metabolic pathways and rapidly adapts to the phagosomal environment. The fungus neutralizes the acidic phagosome and induces morphogenesis, generating hyphae that physically rupture and kill the macrophage. The molecular signals responsible for hyphal growth in this context have not yet been identified, although several models have been implicated. In this study, we test CO2, intracellular pH, and extracellular pH as potential inducers of morphogenesis within macrophages. Using live-imaging microscopy, we assessed the hyphal phenotypes of the C. albicans CO2-sensing mutants nce103 / (lacking the enzyme carbonic anhydrase) and a cyr1 point mutant (possessing a bicarbonate-insensitive adenylyl cyclase). These mutants generated hyphae after phagocytosis, suggesting that CO2-sensing is dispensable for this process. In order to ask if intracellular pH could be a regulator of morphogenesis, we generated a C. albicans strain expressing the pH probe pHluorin2 cytosolically. Ratiometric pH measurements were acquired over the course of hyphal morphogenesis both in vitro and inside of macrophages. The fungal cytosol remained near neutral in all cases, suggesting that intracellular pH changes do not regulate hyphal morphogenesis. Concurrently, in order to ask if extracellular pH sensing is important for this process, we tested a rim101/ mutant in co-culture and saw no hyphal defect. Thus, neither CO2 sensing nor pH fluctuations are required for the induction of morphogenesis in this context. Work is ongoing to reveal the signal(s) driving this process.

Poster 2

Elucidation Of The Molecular Signal for the Regulator Of Capsule Synthesis Stress Response

Hitt SJ ^1,2, Konovalova A ^1,2

Department of Microbiology and Molecular Genetics, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, 77030, USA¹.

The University of Texas MD Anderson UTHealth Graduate School of Biomedical Sciences, Houston, TX, 77030, USA².

Corresponding author: Anna Konovalova, Department of Microbiology and Molecular Genetics, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, 77030, USA anna.konovalova@uth.tmc.edu.

The Regulator of Capsule Synthesis (Rcs) envelope stress response is highly conserved in Enterobacteriaceae. Rcs is activated by multiple immune factors and antibiotics, which target the cell envelope. These include cationic antimicrobial peptides, that disrupt the outer membrane (OM) by targeting lipopolysaccharide (LPS), and lysozyme and β-lactam antibiotics that target the peptidoglycan (PG) cell wall. Rcs regulates expression of many genes to prevent or mitigate cell envelope damage, and as such, Rcs is essential for survival in the host, virulence, and antibiotic resistance. Despite its importance, how Rcs detects envelope damage remains unknown. Rcs is a signal transduction pathway consisting of six components, including the sensor protein RcsF. RcsF forms a complex with several outer membrane proteins (OMP), which allows RcsF to co-localize with LPS at the cell surface. My project aims to identify a molecular signal and mechanism of RcsF activation. Our hypothesis is that RcsF monitors LPS packing through direct interaction with LPS. Mutations that alter LPS charge and structure induce Rcs in a Mg²⁺-dependent manner. I showed that increased expression of eptA, which modifies LPS and strengthens lateral interactions in a cation-independent manner, also causes a reduction in Rcs signaling in an LPS biosynthesis mutant, providing strong evidence for LPS lateral interactions as a potential Rcs signal. Moreover, the addition of divalent cations during PG synthesis inhibition by antibiotics such as A22 and mecillinam also significantly reduces Rcs activity. My results support two initial conclusions. First, RcsF seems to monitor LPS packing at the OM and not the LPS structure itself. Second, stabilizing LPS packing alleviates Rcs signaling, not only when LPS is targeted but also when peptidoglycan biosynthesis is inhibited. Together, these findings suggest that disruptions to LPS packing may be a direct and universal signal for Rcs induction.

Acknowledgments.

This research is supported by the National Institute of General Medical Sciences R01GM133904 and the Welch Foundation Research Grant AU-1998.

Poster 3

Candida albicans biofilm development in catheter-associated urinary tract infections (CAUTIs)

Shane A. Cristy^1,2, Jennifer N. Walker^1,2,3, & Michael C. Lorenz^1,2

Dept. of Microbiology and Molecular Genetics, The University of Texas Health Science Center at Houston, TX¹; Dept. of Microbiology and Infectious Diseases, MD Anderson UTHealth Graduate School of Biomedical Sciences, Houston, TX²; Dept. of Epidemiology, Human Genetics, and Environmental Sciences (EGHES), School of Public Health, University of Texas Health Science Center at Houston, TX³

Abstract:   The fungal opportunistic pathogen Candida albicans is the second most common species isolated from biofilms that cause CAUTIs. These biofilms have the potential to occlude the catheter, which necessitates frequent catheter changes and reduces the quality of life, particularly for individuals that are chronically catheterized.  Additionally, these biofilms serve as a reservoir for pathogens that may promote symptomatic infection.  C. albicans biofilms have been extensively studied in oral, bloodstream, and in vitro models, identifying a complex transcriptional regulatory network, but despite the frequency of these infections, little is known about the molecular basis of C. albicans biofilm formation in the unique urinary environment.  The objective of this project is to determine molecular mechanisms intrinsic to C. albicans that promote biofilm growth in the CAUTI environment.  First, we measured biomass and structure in vitro to examine the biofilm forming potential of mutants lacking each of the major biofilm transcriptional regulators. We observed that when grown in artificial urine medium, mutants lacking ROB1 and TEC1 had substantially different phenotypes compared to growth in RPMI medium.  We then used RNA-seq on biofilms grown to early and late time points to gain an unbiased view of the total transcriptome.  This was done in a historical lab strain and two CAUTI clinical isolates of C. albicans. These data will give us an indication of the systems important for proliferation and biofilm growth in urinary catheters. We next intend to test mutant strains in a murine model that replicates clinical biofilms and takes into account the complexities of multispecies biofilms, the presence of host factors and immunity, and continuous nutrient exchange.  The completion of this study should elucidate the major determinants of C. albicans biofilm development in urinary catheters, thus identifying potential targets for anti-biofilm drugs or catheter coatings.

Poster 4

Targeting cancer stem cell plasticity to overcome colorectal cancer resistance and relapse

Shraddha Subramanian^1,2, Tressie Posey^1,2, Joan Jacob^1,2, and Kendra S. Carmon^1,2

Center for Translational Cancer Research, Brown Foundation Institute of Molecular Medicine, University of Texas Health Science Center, Houston, TX, USA¹

University of Texas MD Anderson Cancer Center UTHealth Houston Graduate School of Biomedical Sciences, Houston, TX, USA²

Despite therapeutic advancements, colorectal cancer (CRC) remains the second deadliest malignancy in the US. CRC relapse can be attributed to cancer stem cells (CSC), an immortal cell population thought to potentiate metastatic progression by exploiting its self-renewability and differentiation capacity. CSCs exhibit plasticity, whereby they alter their phenotype in response to environmental cues, which bolsters inherent drug resistance. Owing to these mystifying properties, CSCs are an attractive drug target. In the past, our group and others have attempted targeting CSCs using antibody-drug conjugates (ADCs) against the Leucine-rich repeat-containing G protein-coupled Receptor 5 (LGR5), a well-recognized CSC marker that is frequently upregulated in CRC tumors. Our previous attempts at eradicating CRC by targeting LGR5⁺ CSCs with an LGR5-directed ADC have resulted in incomplete tumor regression or relapse. Follow-up studies suggest that CRC cells evade this therapeutic insult by converting it into a LGR5^– state. Further, we recently showed that the LGR5^–, drug-resistant cells use the MET-STAT3 signaling cascade to likely bolster their invasive and metastatic potential. The long-term goals of this project include elucidating the intercellular players fueling LGR5⁺ CSC plasticity and developing a dual-targeting therapeutic strategy to overcome CSC-driven tumor heterogeneity. Mechanistically, I postulate that LGR5⁺ CSC plasticity and drug resistance depend on the feedback loop between LGR5 and MET. Therapeutically, I hypothesize that co-targeting LGR5 and MET will overcome CSC heterogeneity and abrogate drug-recalcitrant CRC tumors. Anti-MET monoclonal antibodies (mAbs) binding different MET structural domains were cloned, evaluated in vitro for internalization and binding affinity, and are currently being conjugated to cytotoxic drug payloads. Both have revealed to be promising candidates following evaluations for specificity, internalization, and ability to inhibit MET phosphorylation using fluorescence-based binding assays, immuno-cytochemistry and immunoblotting, respectively. My preliminary data also suggests that both candidates demonstrate high cell-killing efficacy using a secondary ADC conjugated with Pyrrolobenzodiazepine (PBD). In the future, I intend to conjugate the anti-MET mAb candidates to different drug payloads. The cytotoxic potential of these ADCs will be explored further in vitro and in vivo. Efficacy of MET-and LGR5-ADCs alone or in combination will also be evaluated.  Additionally, I will also probe the mechanistic relationship between MET-STAT3 and LGR5 on CRC plasticity through MET ablation and investigate its effect on downstream signaling. My proposed dual-targeted therapeutic modality will act as guided missiles that delivers cytotoxic agents to heterogenous CRC tumors. I believe that eliminating LGR5⁺ CSCs and their LGR5^– counterparts will be the linchpin of tumor eradication in CRC patients.

Poster 5

Prelimbic neurons differently encode changes in metabolic or threat states during cued food seeking

Xu O. Zhang¹, Claire E. Cho¹, Guillermo Aquino-Miranda¹, Yongzhe Wang², Nikita Elison-Watson¹, Fabricio H. Do Monte¹

¹Department of Neurobiology and Anatomy, The University of Texas Health Science Center, Houston, TX 77030, USA, ²Department of Population Science and Surgery, City of Hope, Duarte, CA 91010, USA

Flexibly adjusting foraging behavior based on metabolic needs and environmental threats is crucial for animal survival. However, the neural computations underlying the transition in cued food-seeking behaviors under distinct metabolic and threat states remain unclear. Neurons in the prelimbic (PL) subregion of the prefrontal cortex encode food-cue association. To test if PL responses to food cues vary according to animals’ metabolic states (hungry vs. satiated) or threat states (safe vs. threatened), we used a miniscope to record calcium transients from PL glutamatergic neurons in freely behaving rats. Food-restricted animals previously trained to press a lever for sucrose during an audiovisual cue were presented with 12 food cues before and after a sucrose ad libitum period to induce satiation, or a period of predator odor (cat saliva) exposure to induce innate defensive responses in a separate test. Rats showed a reduction in food seeking after transitioning to satiated or threatened states. We observed no differences in the averaged activity of all PL neurons following food-cue onset from hungry to satiated states, but an increase from safe to threatened states, suggesting that PL neurons are preferentially recruited during more salient internal states. Interestingly, tracking the activity of the same PL neurons across the session revealed that different subpopulations of neurons are recruited to respond to food cues during distinct internal states. Next, we used a support vector machine decoder to infer the internal state of the animals based on the post-cue activity of the entire PL population. We found that the decoding accuracy of metabolic states was similar to chance, whereas the decoding accuracy of threat states was near 95%. To explore the underlying circuit mechanism, we combined in vivo electrophysiological recording with optogenetics to photoidentify PL neurons that project to the paraventricular nucleus of the thalamus (PVT). Similar to PL glutamatergic neurons, different subpopulations of PL-PVT neurons are recruited to respond to food cues during distinct internal states. Finally, chemogenetic inhibition of PL-PVT neurons restored food approach behavior and reduced defensive behaviors under threatened states. Together, our results suggest that the recruitment of distinct neuronal subpopulations in PL when animals transition between metabolic or threat states may serve to adjust behavioral responses to environmental food cues.

Poster 6

A shared cortical language network for heteromodal speech production

Kathryn Snyder,^1,2 Kiefer Forseth,^1,2 Nitin Tandon^1,2,3

 ¹Vivian L. Smith Department of Neurosurgery, McGovern Medical School at UT Health Houston, Houston, TX USA

²Texas Institute for Restorative Neurotechnologies, The University of Texas Health Science Center at Houston, Houston, TX, USA

³Memorial Hermann Hospital, Texas Medical Center, Houston, TX, USA

Lexical access describes the process involved in the mapping between conceptual representations and phonology and is an integral component of speech production. While various brain regions have been hypothesized to support separable language processes, the complete cortical language network and its functional mapping to lexical access remain unclear. We used electrocorticography to identify cortical brain networks involved in lexical cognition in the context of heteromodal sensory input modalities with convergent design.

Data were obtained from epilepsy patients who underwent invasive electrophysiology. Recordings were acquired during three cued naming tasks using pictures (n=154; 16,229 electrodes), auditory descriptions (n=119; 12,347 electrodes), and orthographic descriptions (n=73; 8,716 electrodes). High gamma power (65-115 Hz) was used to measure cortical engagement, and electrode recording zones were defined on the cortical surface. Surface-based mixed-effects multilevel analysis was used to estimate group-level high gamma activity and compared across tasks to identify regions with significant activation.

Analyses showed left-lateralized heteromodal activity in the inferior frontal gyrus (IFG), the middle fusiform gyrus (mFus), and the intraparietal sulcus (IPS) following initial sensory processing. Activation of mFus peaked 150 ms before stimulus offset for auditory (15.2%, p < 10^-6) and orthographic (21.4%, p < 10^-6) naming and 350 ms after picture onset (39.0%, p < 10^-6) for visual naming. Activation of IPS peaked just prior to stimulus offset for auditory (38.1%, p = 7.7*10^-5) and orthographic (27.5%, p = 0.008) naming and 450 ms after picture onset (43.0%, p < 10^-6) for visual naming. Activation of IFG also peaked just prior to stimulus offset for auditory (30.1%, p < 10^-6) and orthographic (40.1%, p = 10^-6) naming and 500 ms after picture onset (36.4%, p < 10^-6) for visual naming. There was also heteromodal activity in the left posterior middle temporal gyrus (pMTG) for auditory and orthographic descriptions but not for pictures with peak activation occurring prior to the convergence of all tasks 250 ms before stimulus offset (auditory: 34.1%, p < 10^-6; orthographic: 25.5%, p = 0.0035).

These results reveal that a shared, heteromodal brain network consisting of IFG, mFus, and IPS supports lexical access. Furthermore, our findings also implicate the role of pMTG in phonological access.  Altogether, this work further characterizes the functional roles of key brain regions within language networks and provides important insights that are critical to the development of improved treatment methods for speech-related disorders.

Poster 7

Membrane Dynamics and Druggability of Small GTPases Rheb and RhoA

Chase Hutchins^1,2, Alemayehu Gorfe^1,2

1. Department of Integrative Biology and Pharmacology, McGovern Medical School, University of Texas Health Science Center, Houston, TX, USA
2. University of Texas MD Anderson Graduate School of Biomedical Sciences, Houston, TX, USA

Membrane proteins are an integral part of cell signaling and function, with many being anchored to the membrane by a lipid anchor motif. The quintessential example of these membrane signaling proteins are the small Guanine Tri-Phosphate (GTP) hydrolyzing proteins of the Ras superfamily of GTPases. All members of this family are anchored to the lipid membrane by a prenyl group at the C-Terminus Hypervariable Region (HVR). Our lab utilizes Molecular Dynamics Simulation to examine how the dynamics of these GTPases can effect function as well as how they can be therapeutically targeted to modulate that function. My project focuses on two members of this family: Rheb and RhoA. Rheb is an important modulator of mTOR, the master regulator of growth, and thus is an attractive target for restricting cell growth in disease states like cancer. RhoA is a signaling protein involved in cytoskeleton regulation and associated with malignancy and metastasis. Rheb is anchored to the lipid membrane by a farnesyl addition at the end of a long string-like region (referred to as the Hyper Variable Region) that contains a poly-basic domain. This is contrasted by RhoA, which is anchored by a geranylgeranyl group and does not contain a polybasic domain in its HVR. 20 µs long Molecular dynamics simulations of GTP/GDP-Rheb and RhoA have revealed very significant differences in both intrinsic membrane orientation dynamics, and membrane-protein interactions. Rheb is much more mobile on the membrane, both swinging and rolling to sample many conformations. RhoA samples fewer conformations due to its propensity to engage the membrane for long periods, and swings much less than Rheb.

Poster 8

Pattern Dynamics of Brain Waves affected by Alzheimer’s Disease

Clarissa Hoffman¹, Jingheng Cheng², Daoyun Ji², Yuri Dabaghian¹

Department of Neurology, The University of Texas McGovern Medical School at UT Health Houston, Houston, TX, USA¹

Department of Neurosciences, Baylor College of Medicine, Houston, TX, USA²

We focus on understanding the impact of Alzheimer’s Disease (AD) pathology at the neurocircuit level. The main physiological manifestation of circuit activity is the synchronized extracellular field, which gives rise to the recorded local field potential (LFP). These fields are widely studied using a variety of methods that primarily address time-localized (instantaneous) or time-averaged characteristics of LFPs. We hypothesize that AD-damaged hippocampal circuits produce abnormal LFP dynamics that clearly manifest at the level of wave patterns and waveforms. We propose an alternative approach that focuses on the morphologies of waveforms—the patterns of the brain waves over finite timescales. Specifically, we use two independent methods for quantifying the structural regularity and irregularity of brain waves and correlate the resulting “stochasticity scores” with behavior. The first, quantifies the pattern’s consistency with the underlying mean. The second, measures how “structured” or “orderly” (e.g., periodic-like or time-clustered) the pattern is. Our previous work in wild-type mice revealed a curious interrelationship between morphologies of θ-waves, γ-waves, and sharp wave events (SWE) and the animal’s speed and acceleration. We also noticed spatial clustering of waves with different morphology along the animal’s trajectory, reminiscent of hippocampal place fields. Based on these observations, we studied circuit activity of hippocampal networks in AD brains and found a number of alterations in LFP rhythmicity. For example, there is a loss of distinct quiescence and movement states in the AD model, indicating an ambiguity of physiological context. Furthermore, the coupling of waveform patterns with speed is weaker in the and SWE in AD, yet, the strength of coupling between -patterns and speed is strengthened revealing potential –  dysfunction. Additionally, the spatial selectivity of patterns is lost in AD, suggesting that damaged synaptic circuits compromise wave patterning and information exchange between brain regions. This failure to discriminate environmental position may be a potential cause of pattern dysfunction, and thus memory loss in AD. These differences in brain wave patterns can be used to better understand and potentially to detect circuit-level pathologies in AD brains. Overall, these results offer a novel perspective on studying the structure, the dynamics, and the functionality of the brain waves and will provide a deeper understanding of AD at a neurocircuit level.

Acknowledgements: NIH R01AG074226, NIH R01NS110806, NSF 1901338, NIH R01NS097764

Poster 9

Untangling the Complex Regulatory Networks Controlling Urease Activity Among Chronic Staphylococcus aureus Urinary Catheter Isolates

Jana Gomez^1,2, Jesus M Duran Ramirez^1,2, Jennifer N Walker¹

Department of Microbiology and Molecular Genetics, McGovern Medical School, University of Texas Health Science Center, Houston, TX, USA¹

The University of Texas MD Anderson Cancer Center UTHealth Houston Graduate School of Biomedical Sciences, Houston, TX, USA²

Catheter associated urinary tract infections are one of the most common hospital-associated infections and are caused by a broad range of uropathogens. Of these uropathogens, Staphylococcus aureus is particularly problematic as it frequently disseminates to bacteremia and commonly exhibits methicillin resistance, making it difficult to treat. Furthermore, S. aureus produces the enzyme urease, which promotes the formation of crystals on the catheter, resulting in catheter encrustations, further antibiotic recalcitrance, and chronic infection. Notably, our previous work demonstrated S. aureus isolates persistently colonizing urinary catheters exhibited temporal increases in urease activity, suggesting the enzyme contributes to chronicity. Thus, we aim to understand how S. aureus urease functions on a molecular level to develop better treatment or prevention strategies. Other reports suggest urease expression is controlled by the CodY, CcpA, and Agr regulators. Using bioinformatics analyses, we identified potential binding sites for CcpA and CodY in the upstream regulatory region of the urease operon, suggesting these proteins directly regulate urease activity. However, we found no binding site for AgrA upstream of the urease operon, suggesting that urease may be regulated indirectly by Agr. Furthermore, we found that a codY mutation exhibited significantly increased urease activity, while an agr mutant displayed only slightly higher activity compared to wildtype, suggesting this regulator may act as a repressor of urease activity. These data provide further insight into the molecular mechanism of urease production and activity in S. aureus. Further work will continue to investigate this mechanism and the transcriptional and translational changes that select for increased urease activity.

Poster 10

HAP40 is a conserved regulator of Huntingtin and a modulator of Huntington’s disease

Stephen Farmer¹, Yue Yu¹, Amanda Solbach¹, Shiyu Xu¹, Gang Li¹, Xin Ye¹, Dongsheng Chen¹, Zhihua Chen¹, Zhen Xu¹, Moretti Daniele², Sara Tambone², Alessandra Ceccacci², Licia Tomei², Lili Ye¹, Erin Furr Stimming³, George McAllister⁴, Deanna M. Marchionini⁴, Sheng Zhang¹

¹The Brown Foundation Institute of Molecular Medicine, University of Texas Health Science Center, Houston, TX, USA

²Department of Translational and Discovery Research, IRBM SPA, Pomezia (Rome), Italy

³Department of Neurology, HDSA Center of Excellence, University of Texas Health Science Center, Houston, TX, USA

⁴CHDI Management, CHDI Foundation, New York, NY, USA

Huntington’s disease (HD) is a devastating brain degenerative disorder caused by amplifications of the poly-glutamine tract (polyQ) in Huntingtin (HTT). Aberrations in HTT’s physiological functions due to polyQ expansion is one postulated factor in HD pathogenesis; however, the molecular function and regulation of HTT remain to be elucidated. Drosophila (fruit flies) is a powerful genetic model that allows convenient interrogation of evolutionarily-conserved pathways related to human diseases, including HD, given that HTT is conserved in fruit flies. Using a proteomics-based approach, we identified a novel 40kDa protein encoded by an uncharacterized fly gene CG8134 as a strong interactor of Drosophila HTT and further demonstrated it was the functional ortholog of human HAP40, an HTT-associated protein shown recently to modulate HTT’s conformation but with unclear physiologic and pathologic roles. Validation experiments in flies and human cells supported conserved physical and functional interactions of HAP40 with HTT across the evolutionarily distant species, and HTT interacts with endosomal membranes in a HAP40-depedent manner. Corroborating with these observations, molecular modeling revealed conserved structural architectures between human and Drosophila HTT/HAP40 complexes, in particular a solvent-exposed membrane-association loop on HAP40. Moreover, genetic interaction assays showed that loss of HAP40 causes similar phenotypes as HTT knockout including a small lysosomal-related granule phenotype, which provides the first in vivo evidence on HAP40’s regulation of HTT. At the molecular level, HAP40 strongly affected HTT’s protein stability, as depletion of HAP40 significantly reduced the levels of endogenous HTT while HAP40 overexpression markedly extended its half-life. Conversely, in HTT-deficient cells, the majority of HAP40 protein was degraded, potentially by the proteasome. Further, polyQ expansion did not significantly alter the affinity of the HTT/HAP40 complex based on computer modeling predictions and in vitro biochemical assays. Lastly, when tested in Drosophila models of HD, HAP40 modulated the neurodegenerative effects of full-length mutant HTT but not HTT exon 1 fragment. Taken together, our study uncovers a conserved regulatory mechanism of HTT by HAP40, and demonstrates that HAP40 is a conserved partner of HTT, governing HTT’s protein stability and normal physiological functions while also modulating the toxicity of mutant HTT, therefore can be a potential therapeutic target against HD.

Poster 11

Translational evaluation of a next-generation fluorescent somatostatin analog for intraoperative imaging

 Servando Hernandez Vargas¹, Solmaz AghaAmiri¹, Sukhen C. Ghosh¹, Michael P. Luciano², Luis C. Borbon³, Po Hien Ear³, James R. Howe³, Jennifer M. Bailey-Lundberg¹, Gregory D. Simonek¹, Daniel M. Halperin⁶, Hop S. Tran Cao⁷, Naruhiko Ikoma⁷, Martin J. Schnermann², and Ali Azhdarinia¹

¹The University of Texas Health Science Center at Houston, Houston, Texas 77054, USA.

²National Cancer Institute, Frederick, MD, 21702, USA.

³University of Iowa Carver College of Medicine, Iowa City, Iowa, 52242, USA.

⁴The University of Texas MD Anderson Cancer Center, Houston, TX, 77030, USA.

• These authors contributed equally.

Introduction: Fluorescence-guided surgery (FGS) is an imaging specialty capable of identifying tumors that would otherwise be undetected [1]. Patients with neuroendocrine tumors (NETs) are excellent candidates for tumor-targeted FGS since up to 45% of surgical cases have partial resection and most overexpress the somatostatin receptor subtype-2 (SSTR2). Here we introduce two key advances in the translational evaluation of an FGS approach for NETs. First, we developed a second-generation near-infrared fluorescent (NIRF) somatostatin analog, MMC(FNIR-Tag)-TOC, to enhance tumor specificity. Second, we examined agent performance using novel animal models and simulated a clinical workflow to bridge preoperative nuclear imaging with SSTR2-targeted FGS.

Methods: The SSTR2-targeting peptide, TOC, was conjugated to a multimodality chelator (MMC) [2] on solid-phase, and followed by FNIR-Tag (charge balanced dye [3]) conjugation via click chemistry. The fluorescent conjugate was radiolabeled with ⁶⁷Ga as previously described [2]. SSTR2-mediated binding was confirmed using established radioligand assays [2]. To determine the optimal dose and time, we injected increasing doses (2, 5 and 10 nmol) of dual-labeled agent into nude mice with NCI-H69 xenografts and performed in vivo imaging up to 24 h post-injection (p.i.) (n = 4/group). Key tissues were resected for ex vivo imaging and gamma counting. To simulate a patient selection protocol, we used an orthotopic pancreatic tumor model (BON1-SSTR2; n = 5) and performed surgical planning on day 0 via PET/CT scan using ⁶⁸Ga-DOTA-TOC (1 h p.i.; 200 µCi; 0.5 nmol). Two days later, we injected mice with 5 nmol of MMC(FNIR-Tag)-TOC, acquired in vivo fluorescence images 3 h p.i., and harvested tumors under white light. After direct visual inspection of the wound bed, we identified and harvested suspicious lesions for multiscale imaging and immunohistopathology. To assess imaging performance in a translational setting, we first used our novel patient-derived xenograft (PDX) models, NEC913 (SSTR2+) and NEC1452 (SSTR2-) [4], and performed NIRF imaging studies in vivo (n = 3) as described above. Finally, we examined specific binding in frozen sections from freshly resected pancreatic NETs, metastatic lesions (liver), and involved lymph nodes.

Results/Discussion: The dose-time finding study showed that agent uptake in tumor and nontumor tissues varied slightly as a function of dose or time, suggesting saturable tumor binding (1.5-2.0 %IA/g) and efficient clearance (<0.25 and <0.50 %IA/g in muscle and blood at 3 h, respectively). On average, fluorescent TBRs were >3.5 and >5.5 in pancreas and small intestine, respectively, with no differences (P > 0.05) associated with dose or time. From these experiments, we identified 5 nmol and 3 h as preferred parameters. Using a theranostic approach for “patient” selection with a FDA-approved radiopharmaceutical, we observed excellent correlation between nuclear and fluorescence imaging findings in the BON1-SSTR2 orthotopic model. Ex vivo imaging yielded a tumor-to-pancreas ratio of 17.7±9.3, suggesting strong potential for visual contrast in an intraoperative setting. Furthermore, we detected fluorescent focal points in the spleen and small intestine suspected to be metastatic disease. Analysis of tumor and suspicious lesions confirmed cancer status (H&E) and SSTR2-positive regions (IHC) that correlated with mesoscopic NIRF imaging. In the PDX imaging study, agent accumulation was only observed in SSTR2+ tumors, with minimum to no signal in tumors lacking the receptor (SSTR2+ tumor/SSTR2- tumor ratio >4). MMC(FNIR-Tag)-TOC staining of surgical biospecimens showed excellent co-localization with IHC.

Conclusions: Using multiple translational models and a combination of fluorescent and radioactive readouts, we observed excellent correlation between preoperative nuclear imaging, fluorescence-guided surgery, and histopathology, indicating strong potential of MMC(FNIR-Tag)-TOC for clinical applications.

References:

1. Hernot, S., et al., Lancet Oncol., (2019), DOI: 10.1016/S1470-2045(19)30317-1
2. Hernandez Vargas, S., et al., Clin. Cancer Res., (2019), DOI: 10.1158/1078-0432.CCR-18-3312
3. Luciano, M. P., et al., ACS Chem. Biol., (2019), DOI: 10.1021/acschembio.9b00122
4. Tran, C. G., et al., Cancers, (2022), DOI: 10.3390/cancers14081910

Poster 12

Prelimbic cortex neurons encode reward-cue information during an approach-avoidance conflict test in rats

Vicky Chuong¹, Guillermo Aquino-Miranda¹, Xu O. Zhang¹, Douglas S. Engelke¹, Fabricio H. Do Monte¹

Dept. of Neurobiology and Anatomy, The University of Texas Health Science Center at Houston, Houston, TX, USA¹

Neurons in the prelimbic (PL) subregion of the medial prefrontal cortex change their firing rates in response to fear- and reward-associated cues. PL activity is essential for the retrieval of both fear- and reward-related memories. However, it remains unknown how PL neurons respond during situations of conflict when reward cues occur together with fear cues or in a fear-inducing context. To explore this question, male adult Long-Evans rats with single-unit recording electrodes previously implanted in PL were exposed to an approach-food vs. avoid-predator odor conflict model and contextual fear memory was evaluated in the following day. Rats were initially trained to press a lever for sucrose during the presentation of audiovisual cues. During the predator odor conflict test, animals were exposed to three phases: (i) reward phase, only food cues presented, (ii) cat odor phase, only a fear-inducing cat odor presented, and (iii) conflict phase, food cues concomitantly presented with cat odor. Compared to reward phase, rats displayed increased defensive behaviors and reduced food-seeking responses during conflict. PL recordings (413 neurons from 29 rats) revealed changes in the spontaneous firing rate across the three phases with an increase in the percentage of neurons showing inhibition and a reduction in the percentage of neurons showing excitation during conflict. Tracking the activity of the same cells across the phases, we found that most PL neurons changed their firing rates in more than one phase, and ~70% of these cells responded in opposite directions, suggesting valence encoding. After aligning PL activity to the onset of the food cues, we observed a similar proportion of responsive neurons during the reward (~30%) and conflict (~35%) phases. Surprisingly, only a small fraction of these cells (~10%) responded in both phases, suggesting that distinct subpopulations of food cue responsive neurons are recruited during conflict. During the contextual fear test, clustering analysis of food-seeking responses revealed two distinct phenotypes. Whereas a subset of rats exhibited increased lever-press responses (risk-takers), another subset displayed reduced lever-press responses and increased avoidance behavior (risk-avoiders). When aligning the activity of PL neurons to the onset of food-cues, excitatory responses in risk-takers were maintained in the same cells across the entire contextual fear test. In contrast, excitatory food-cue responses in risk-avoiders were attenuated from the early to the late phase and vice versa, which may reflect modifications in the predictive value of the food cues across the session. Together, our results show that activity in PL neurons correlates with changes in behavioral choice during motivational conflict, suggesting that PL may be involved in risky-decision making.

Grant Support: NIH-R01-MH120136 and Rising STARs Award from University of Texas System.

Poster 13

Detection of chronic wasting disease prions in processed meats.

Rebeca Benavente¹, Francisca Bravo^1,2 Paulina Soto^1,2, J Hunter Reed³, Mitch Lockwood³, Rodrigo Morales^1,2

¹Department of Neurology, McGovern Medical School, University of Texas Health Science Center at Houston, Texas, USA; ²Universidad Bernardo O’Higgins. Santigo, Chile; ³Texas Parks and Wildlife Department, Texas, USA.

Background: The zoonotic potential of chronic wasting disease (CWD) remains unknown. Currently, there are no known natural cases of CWD transmission to humans but increasing evidence suggests that the host range of CWD is not confined only to cervid species. Alarmingly, recent experimental evidence suggests that certain CWD isolates can induce disease in non-human primates. While the CDC strongly recommends determining CWD status in animals prior to consumption, this practice is voluntary. Consequently, it is plausible that a proportion of the cervid meat entering the human food chain may be contaminated with CWD. Of additional concern is that traditional diagnostic techniques used to detect CWD have relatively low sensitivity and are only approved for use in tissues other than those typically ingested by humans.

Methods: In this study, we analyzed different processed meats derived from a pre-clinical, CWD-positive free-ranging elk. Products tested included filets, sausages, boneless steaks, burgers, ham steaks, seasoned chili meats, and spiced meats. CWD-prion presence in these products were assessed by PMCA using deer and elk substrates.

Results: Our results show positive prion detection in all products. To confirm the resilience of CWD-prions to traditional cooking methods, we grilled and boiled the meat products and evaluated them for any remnant PMCA seeding activity. Results confirmed the presence of CWD-prions in these meat products suggesting that infectious particles may still be available to people even after cooking.

Conclusion: Our results strongly suggest ongoing human exposure to CWD-prions and raise significant concerns of zoonotic transmission through ingestion of CWD contaminated meat products.

Poster 14

Prelimbic cortical neurons signal risk-taking vs. risk-avoiding behavior in morphine-exposed rats

Cana B. Quave¹, Andres M. Vasquez^1,2, Esha P. Bora¹, Chinenye L. Chidomere^1,3, Guillermo Aquino-Miranda¹, Douglas S. Engelke¹, Fabricio H. Do Monte¹

Department of Neurobiology & Anatomy, The University of Texas Health Science Center at Houston, Houston, TX, USA¹; Department of Neuroscience, Rice University, Houston, TX, USA²; Department of Biomedical Engineering, University of Houston, Houston, TX, USA³

Opioid use disorder is associated with impaired risk-related decision-making. However, it is unclear how repeated opioid exposure affects the brain to alter risk taking under motivational conflict. In a modified conditioned place preference protocol, rats injected with either saline or morphine were exposed to the side of the apparatus preferred least at baseline. Conditioning occurred over 10 alternating days (5 pairings in each side). Two days after conditioning, rats underwent a preference test immediately followed by a conflict test in which an aversive stimulus (cat saliva) was introduced in the side of the chamber previously paired with morphine injections. In the preference test, morphine-treated rats spent more time in the drug-paired side of the apparatus than did saline-treated rats. In the conflict test, saline group rats avoided the side of the apparatus containing cat odor. In contrast, rats in the morphine group continued to prefer the previously drug-paired side despite the presence of cat odor, demonstrating increased risk-taking behavior. K-means clustering uncovered two subsets of morphine-treated rats that exhibited either: i) enhanced place preference and persistent drug-seeking during conflict (risk-takers, RT), or ii) moderate place preference and suppressed drug-seeking during conflict (risk-avoiders, RA). Single-unit recordings from neurons in the prelimbic (PL) cortex, a region involved in decision-making and strategy shifting, revealed decreased firing rates upon acute morphine exposure. In contrast, on the final drug conditioning day, morphine failed to suppress neuronal firing rates, suggesting that PL neurons undergo adaptation to repeated morphine exposure. Recordings during the preference test identified distinct populations of PL neurons that were either excited or inhibited when rats entered the drug-paired side of the apparatus. Interestingly, while cells inhibited during paired side entries in the preference test showed no response to paired side entries during the conflict test in saline-treated and RA rats, these inhibitory entry responses persisted in RT rats. Additionally, RT rats showed a greater proportion of paired side entry-responsive neurons that altered their response type between the preference and conflict tests. Taken together, our results suggest that a loss of PL inhibition after opioid conditioning is associated with the formation of contextual reward memory. Furthermore, persistent inhibitory signaling and enhanced spatial remapping of the drug-associated context in PL during conflict may underlie increased risk-taking following opioid exposure.

Poster 15

Long non-coding RNA (lncRNA) expression is upregulated in microglia during regeneration of rods in zebrafish retina

Anna Naglis¹, Abirami Santhanam², Sean P. Marrelli³, John O’Brien²

Program in Biochemistry and Cell Biology, The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences (GSBS), UTHealth, Houston, Texas, USA¹

Department of Vision Science, College of Optometry, University of Houston, Houston, Texas, USA²

Department of Neurology, McGovern Medical School at UTHealth, Houston, TX, USA³

Microglia are resident immune cells of the retina that can change their function in a disease state. While their transition between states in disease has been documented, their state specific functional roles in regeneration and how this transition is regulated remain unclear. Our lab has established a transgenic P23H rhodopsin zebrafish model of Retinitis Pigmentosa with degeneration of rod photoreceptors followed by regeneration of new rods. We hypothesize that microglia undergo transcriptomic and phenotypic changes that might promote rod regeneration.

Analysis of whole retina single-cell transcriptomes from three P23H and WT zebrafish was performed using Seurat package. Unsupervised clustering was performed to identify microglia clusters. To visualize microglia/macrophages, retinal sections from adult P23H and WT fish were stained with antibody 7.4.C.4.

Immunostaining of retinal sections with 7.4.C.4 antibody revealed that microglia/macrophages in P23H retinas acquired a more ameboid shape compared to WT. We detected a higher number of microglia/macrophages in the outer nuclear layer (ONL), harboring photoreceptors, and subretinal space between photoreceptors and retinal pigmented epithelium. Some of the microglia acquire ameboid shape suggesting that they phagocytose dying rods.  Analysis of scRNA seq data identified three distinct microglia clusters in both WT and P23H. Gene expression changed in all clusters in P23H. Genes associated with cytokine/chemokine signaling (tnfrsf9b, traf1, ccl38.6) were upregulated in P23H microglia cluster 27, whereas expression of genes in pathways related to phagocytosis (p2ry12, lamp2) were increased in P23H microglia cluster 31. We found that expression of several lncRNAs, si:ch211-214p16.2, LOC103909107, LOC103910136, LOC103909099 and LOC110439915 was microglia cluster-specific. Furthermore, these were among the top differentially-expressed genes and were upregulated by 3.5 to 17.6 fold in P23H microglia compared to WT.

We conclude that microglia/macrophages in regenerating retinas of P23H fish undergo transcriptional and morphological changes and migrate towards rod photoreceptors in the ONL and subretinal space. We also have identified upregulation of lncRNAs expression in P23H microglia. LncRNAs regulate many cellular processes and we suggest that they might drive microglia/macrophage’s transcriptional change to promote regeneration of rods. Further knock-down studies will reveal specific roles of the lncRNAs in retinal regeneration.

Poster 16

Design and application of a tau seed amplification assay as a novel method for the screening of small molecule aggregation inhibitors

Damian Gorski^1,2, Fei Wang¹, Nicolas Mendez¹, Carla Barria Loaiza¹, Victor Banerjee¹, Danielle Harrison¹ and Claudio Soto^{1, 2}

Mitchell Center for Alzheimer’s Disease and Related Brain Disorders, Department of Neurology, McGovern Medical School, University of Texas, Houston, TX 77030, USA¹

The University of Texas MD Anderson Cancer Center UTHealth Houston Graduate School of Biomedical Sciences, Houston, TX, USA²

Alzheimer’s Disease (AD) is a progressive neurological disorder and represents the single leading cause of dementia worldwide. It is characterized pathologically by the aggregation and accumulation of amyloid-beta as extracellular plaques and tau as neurofibrillary tangles. These toxic aggregates are suspected to result in cell loss and a gradual cognitive impairment. Recent studies have demonstrated that the disease-associated forms of these proteins spread throughout interconnected regions of the brain in a prion-like manner. This hypothesis, often referred to as the seeding-nucleation model, suggests that the pathological proteins trigger the corruption of natively folded proteins, inducing their misfolding and incorporation into expanding amyloid fibrils. There is currently no cure for AD and recent therapeutic efforts have focused largely on amyloid-beta despite mounting evidence suggesting that tau pathology more closely correlates with loss of cognitive function. As such, preventing or inhibiting the aggregation of tau remains a promising therapeutic strategy.

Here we take advantage of the prion-like properties of pathological tau to develop a sensitive and specific in vitro seed amplification assay (SAA) for the detection of tau seeding activity in post-mortem human brain. Through the careful fine-tuning of SAA conditions, we are able to detect the presence of tau seeding activity up to a 100 million-fold dilution of post-mortem AD brain. Importantly, our assay detects no seeding activity upon challenge with healthy human control brain. To test our platforms capability as a tool for the identification of tau aggregation inhibitors we tested several known anti-amyloid compounds and discriminated several with the capacity to inhibit AD-brain induced tau aggregation. As an effort to explore potential candidates for drug repositioning we screened a multitude of compounds from a commercially available drug library and identified two compounds that nullify tau aggregation.

Our results suggest that the Tau SAA can serve as a robust high-throughput screening method to identify small molecules that inhibit or entirely prevent the aggregation of tau when challenged with a biologically relevant source of tau seeds . Furthermore, through the screening of FDA approved compounds, it may serve as a means for the identification of drugs suitable for repurposing.

Poster 17

P63+/Krt8+ Alveolar transitional cell dysfunction in hydrogels as a novel 3D in vitro model for pulmonary fibrosis

Sarah Shin^1,2, Tinne, Martens⁵, Scott Collum¹, René Girard¹, Andrew Peters¹, Danielle Wu^2,3 and Harry Karmouty-Quintana^1,4

¹Department of Biochemistry and Molecular Biology, McGovern Medical School; ² Department of Diagnostic and Biomedical Sciences at UTHealth Houston School of Dentistry; Department of Bioengineering, George R. Brown School of Engineering, Rice University, Houston, TX 77005; ⁴ Divisions of Critical Care, Pulmonary and Sleep Medicine, Department of Internal Medicine, McGovern Medical School, University of Texas Health Science Center at Houston, Houston, TX-77030, USA; Galapagos Pharma, Mechelen, Belgium⁵

Idiopathic pulmonary fibrosis (IPF) is a chronic respiratory disease that causes irreversible destruction of the blood-gas interface of the lung that ultimately causes respiratory failure. Though multiple cell types have been implicated, the pathogenesis of IPF is not fully understood. Recent studies have identified multiple new aberrant transitional cell populations. Additionally, there are well-documented changes in the extracellular matrix (ECM) of IPF lungs, such as aberrant collagen, elastic, and hyaluronic acid deposition. These dramatic changes not only affect the biochemical microenvironment of the alveoli, but studies increasingly demonstrate the important role of the biomechanical dynamics of the parenchyma on cell differentiation and proliferation.

This project aims to optimize physiologically relevant pro-fibrotic hydrogel conditions that induces primary P63+/Krt8+ alveolar transitional cells to acquire a fibrotic phenotype. Our lab has isolated P63+/Krt8+ epithelial cells from IPF and control patients and observed that they express secretory (SCGB1A1), alveolar (AXIN2), and progenitor (P63) markers. We will further characterize our transitional cells, focusing on expression of mechanosensors (ie. Piezo2). We will investigate whether the differentiation of these transitional cells is modulated by a pro-fibrotic gel environment to study their role in the pathogenesis of IPF.

To develop hydrogels, we collect rheometry data from primary tissue of human fibrotic and control lungs to measure the stiffness needed for our hydrogels. We will optimize the concentration of growth factors (ie. TGF-B) and fibroblasts, then culture our transitional cells under mechanically dynamic conditions, and investigate how our pro-fibrotic in vitro system affects transitional cell differentiation and expression of fibrotic markers. This will expand the model systems available for drug screening and to study the pathogenesis of fibrotic lung diseases.

Poster 18

Mutant smooth muscle cell (SMC) α-actin drives atherosclerosis via activation of heat shock factor 1 and cholesterol-driven SMC modulation

Pujun Guan, MM^1,a, Kaveeta Kaw, MD, PhD^1,a, Abhijnan Chattopadhyay, PhD^1,a, Jiyuan Chen, PhD¹, Suravi Majumder, PhD¹, Xue-yan Duan, PhD¹, Shuangtao Ma, PhD^1,2, Chen Zhang, PhD³, Callie S. Kwartler, PhD¹, Dianna M. Milewicz, MD, PhD^1*.

^a These authors contributed equally to this work; * corresponding author.

Affiliations:  ¹Division of Medical Genetics, Department of Internal Medicine, McGovern Medical School, The University of Texas Health Science Center at Houston; Houston, TX, USA. ²Current address: Department of Medicine, Michigan State University; Lansing, MI, USA.³Division of Cardiothoracic Surgery, Michael E. DeBakey Department of Surgery, Baylor College of Medicine, and Department of Cardiovascular Surgery, Texas Heart Institute; Houston, Texas, USA.

Background and aims: The variant p.Arg149Cys in ACTA2, which encodes smooth muscle cell-specific a-actin, predisposes to thoracic aortic disease and early onset coronary artery disease in individuals without cardiovascular risk factors. This study investigated how this variant drives increased atherosclerosis.

Methods: Apoe^-/- mice with and without the variant were fed a high-fat diet for 12 weeks, followed by evaluation of atherosclerotic plaque formation and single cell transcriptomics analysis. Smooth muscle cells explanted from Acta2^R149C/+ and wildtype ascending aortas were used to investigate atherosclerosis-associated SMC phenotypic modulation.

Results: Hyperlipidemic Acta2^R149C/+Apoe^-/- mice have a 2.5-fold increase in atherosclerotic plaque burden compared to Apoe^-/- mice with no differences in serum lipid levels. At the cellular level, misfolding of the R149C a-actin activates heat shock factor 1, which increases endogenous cholesterol biosynthesis and intracellular cholesterol levels through increased HMG-CoA reductase expression and activity. The increased cellular cholesterol in Acta2^R149C/+ smooth muscle cells induces endoplasmic reticulum stress and activates PERK-ATF4-KLF4 signaling to drive atherosclerosis-associated phenotypic modulation in the absence of exogenous cholesterol, while WT cells require higher levels of exogenous cholesterol to drive phenotypic modulation. Treatment with the HMG-CoA reductase inhibitor pravastatin successfully reverses the increased atherosclerotic plaque burden in Acta2^R149C/+Apoe^-/- mice.

Conclusions: These data establish a novel mechanism by which a pathogenic missense variant in a smooth muscle-specific contractile protein predisposes to atherosclerosis in individuals without hypercholesterolemia or other risk factors. The results emphasize the role of increased intracellular cholesterol levels in driving smooth muscle cell phenotypic modulation and atherosclerotic plaque burden.

Poster 19

Functional dynamics of prelimbic projections to the paraventricular nucleus of the thalamus

Dounya Jalloul, Guillermo Aquino-Miranda, Xu O. Zhang, Joseph Campbell, Fabricio H. Do Monte, Michael Beierlein

Department of Neurobiology and Anatomy, McGovern Medical School, University of Texas Health Science Center at Houston, Houston, TX 77030, USA

ABSTRACT

The paraventricular nucleus of the thalamus (PVT) is involved in the regulation of a myriad of biological responses, such as arousal, stress, fear, anxiety, and reward seeking. The PVT receives extensive projections from the prelimbic (PL) subregion of the medial prefrontal cortex that conveys information relevant to decision-making and goal-oriented behavior. Similar to other corticothalamic circuits, PL neurons located in deep cortical layers target PVT via two functionally distinct pathways: i) direct monosynaptic projections that lead to PVT excitation, and ii) indirect disynaptic projections via the anteroventral portion of the thalamic reticular nucleus (avTRN) that lead to feed-forward inhibition of PVT neurons. However, it is unknown how the intrinsic neuronal properties and the synaptic dynamics of these two pathways interact to control PVT neuronal output.

Using a combination of whole-cell recordings, neuronal tracing and optogenetics in acute brain slices of adult rats, we found that in the majority of PVT neurons, depolarizing current steps or brief trains of PL-evoked synaptic excitation leads to a rapid decay of action potential amplitudes followed by depolarization block, thereby strongly limiting PVT maximum firing rates. In agreement, performing single-unit recordings from PVT neurons in vivo, we found that for a subset of PVT cells, photoactivation of PL inputs at low frequencies led to sustained PVT firing, whereas higher stimulation frequencies led to a progressive reduction of spike amplitude in the same cells. Recordings in avTRN in vitro revealed that the majority of neurons fire in tonic mode and fail to display rebound bursting, due to the absence of low-threshold calcium spikes. Photoactivation of PL afferents in avTRN with brief stimulus trains led to large-amplitude postsynaptic responses that displayed short-term facilitation, which in current clamp led to precise and sustained action potential activity in avTRN. In turn, photoactivation of avTRN afferents in PVT evoked short-term synaptic facilitation, thereby generating reliable GABA_AR-mediated inhibition at different frequencies. For PVT neurons that displayed depolarization block, photoactivation of avTRN inputs could restore action potential activity.

Taken together, our findings suggest that monosynaptic inputs from PL cause prominent depolarization block in PVT neurons, and that highly reliable and frequency-invariant avTRN-mediated disynaptic inhibition is critical to maintain PVT neuronal output over a range of PL afferent activity.

Poster 20

Elucidating the structure-function of the ionotropic delta receptor in the presence of trans-synaptic partners:

Cuauhtemoc U. Gonzalez^1,2, Elisa Carrillo¹, Vladmir Berka¹, Vasanthi Jayaraman^1,2.
Department of Biochemistry and Molecular Biology, University of Texas Health Science Center at Houston, Houston, TX, USA;¹The University of Texas MD Anderson Cancer Center UTHealth Houston Graduate School of Biomedical Sciences, Houston, TX, USA²

The ionotropic glutamate receptors (iGluRs) are glutamate-gated cationic channels expressed in the brain. They are known for their involvement in learning and memory, behavior, cognition, and motor coordination. The delta receptor is a member of the iGluR family due to its structural similarity and sequence homology. This receptor has been demonstrated to play a role in synaptic long-term depression, learning, and memory and has been implicated in neurological disorders like autism and schizophrenia.  Although CryoEM structures show the formation of a pore-like structure like other iGluRs, the delta receptor does not demonstrate isolated ligand-gated currents like its family members, making it difficult to study its function. Our lab recently demonstrated the delta receptor is capable of glycine-gated activity when the extracellular domains come into closer proximity by using chemical linkers at the amino-terminal domain. Using FRET/FLIM, we further demonstrated that the trans-synaptic partners of the delta receptor, presynaptic Neurexin1β and soluble cerebellin, are capable of bringing the extracellular domains into closer proximity. Based on these findings, I developed a Neurexin1β dimer complex using two different coiled tags at the carboxyl-terminal domain and tagged them with separate purification tags for co-immunoprecipitation. Using this Neurexin1β dimer complex in conjunction with cerebellin, we are capable of inducing glycine-gated activity in single-channel patch recordings. Furthermore, using this Neurexin1β dimer complex, I have started investigating the conformational effects of Neurexin1β on the delta receptor using single-molecule FRET, and have begun to uncover the conformational landscape of the extracellular domains when in the presence and absence of glycine and the Neurexin1β /cerebellin dimer complex. The findings obtained from these studies will further elucidate the required structure and conformational changes for delta receptor activity, and reduce the challenges of targeting the ionotropic activity of the delta receptor.

Poster 21

Development of an Epiregulin Antibody Drug Conjugate for the Treatment of Colorectal Cancer

Joan Jacob^1,2, Yasuaki Anami^1,2, Solmaz AghaAmari^1,2, Sukhen Ghosh^1,2, Zhengdong Liang^1,2 , Yukimatsu Toh^1,2 , Ali Azhdarinia^1,2 , Qingyun Liu^1,2  , Kyoji Tsuchikama^1,2 Kendra S. Carmon^1,2

Center for Translational Cancer Research, Institute of Molecular Medicine, UTHealth Houston​, Houston, TX, USA¹
Texas Therapeutics Institute, Institute of Molecular Medicine, UTHealth Houston, Houston, TX, USA²

Colorectal cancer (CRC) is the second leading cause of cancer-related deaths in the U.S. This is largely due to metastasis, tumor relapse, and the vast differences in CRC tumors leading to treatment failure. Cancer stem cells (CSCs) are a small population of tumor-initiating cells thought to act as immortal seeds that form metastases and contribute to tumor relapse by transitioning between a differentiated (metastatic) and undifferentiated (tumor-initiating, immortal) state. Epiregulin (EREG) is a protein highly expressed in both states, as well as in treatment-resistant CRCs of various mutation statuses, providing a target for combating cellular plasticity in CRC. Therefore, we hypothesize that EREG can mediate CRC tumor growth and an EREG-targeted antibody-drug conjugate (ADC) can act as a guided missile to deliver cytotoxic drugs into EREG-expressing tumors to eliminate treatment-resistant CRCs. To this end, we generated EREG-targeted antibody-drug conjugates (ADCs) to act as guided missiles for the delivery of cytotoxic drugs into EREG-expressing tumors to overcome plasticity and resistance in CRC. Our EREG mAbs were shown to bind with high affinity to EREG tethered to the surface of CRC cells and internalize to the lysosome, which is important for processing the release of drug from the ADC. We then evaluated antibody tumor specificity in vivo with our highest affinity EREG mAb and a non-targeting isotype control antibody and found good tumor binding specificity and clearance. We conjugated each control and EREG antibody to various highly cytotoxic drugs via a cleavable peptide linker to develop an ADC. EREG ADCs were evaluated in vitro against a wide panel of CRC cell lines of various mutational statuses. EREG mAb alone did not result in significant CRC cell death. However, EREG ADCs exhibited potent cell-killing in EREG-expressing CRC cells, whereas minimal effects were observed with cADC. These preliminary results demonstrate the therapeutic potential of EREG ADCs for targeting plasticity and drug resistance in CRC.

Poster 22

Identifying putative hyphal regulators of Candida albicans

Dakota Archambault^1,2, Michael Lorenz^1,2

1. University of Texas MD Anderson Cancer Center UTHealth Houston Graduate School of Biomedical Sciences, Houston, TX
2. University of Texas Health Science Center, McGovern Medical School, Department of Microbiology and Molecular Genetics, Houston, TX

Candida albicans is an opportunistic fungal pathogen with limited treatment options and increasing prevalence in immunocompromised individuals. C. albicans is a normal resident of the human microbiota, but defects in the innate immune response can permit C. albicans to disseminate throughout the human body, leading to multisystem organ failure and carries a patient mortality rate of 40%. How C. albicans interacts with phagocytes is thus a relevant determinant for the development of infection due to the complex and dynamic relationship of C. albicans and host macrophages. Macrophages will phagocytose C. albicans, whereupon C. albicans will transition from a yeast cell to the hyphal form, piercing the macrophage and allowing for escape. Using transcriptomic data from several Candida species before and during phagocytosis, two sets of uncharacterized genes were identified as candidate virulence factors. The first set of genes was expressed at a significantly higher level in phagocytosed cells compared to controls and we have generated a mutant library to validate these genes for host-relevant phenotypes. I am screening this library in solid and liquid media conditions to investigate changes in morphogenesis compared to a parental strain and have found several candidate hyphal regulators, such as C2_04280W_A, C6_02450W_A, and CR_07170W_A. The second set of genes are more highly expressed in C. albicans compared to other less virulent species. This set was also enriched for hyphal-associated genes and includes an uncharacterized homolog of the yeast Dig2 morphogenetic regulator. In this study, I will test these sets of novel genes to find novel regulators of morphogenesis and virulence.

Poster 23

A Deep Learning Approach to Naturalistic Tunings

 Slapik M¹, Andrei A¹, Khan S¹, Dragoi V¹

Department of Neurobiology and Anatomy, McGovern Medical School¹

Corresponding Author: Mitchell Slapik, Department of Neurobiology and Anatomy, McGovern Medical School, 6431 Fannin St, Houston, TX 77030. email: mitchell.slapik@uth.tmc.edu.

Introduction: Visual cortex builds a visual representation of our world, starting from basic features like lines and edges and leading up to more complex features like shapes and objects. However, these tunings are generally examined separately: it remains unclear how they come together when many visual features are combined in naturalistic settings. Using new machine learning techniques, we investigate tunings to these multi-faceted stimuli and demonstrate the encoding of naturalistic features in early visual cortex. In the process, we better capture the functioning of visual cortex in everyday life.

Methods: We present visual stimuli to a macaque monkey on a computer monitor, and record from early visual cortex (V1) using a Plexon laminar electrode. Our machine learning algorithm was pioneered by Ponce et al. (2019) and consists of two components: an image generator and an optimizer. The image generator, a generative adversarial network or “GAN,” is trained on over a million natural images and can flexibly produce a wide range of stimuli that resemble textures, objects, and landscapes. Meanwhile, our optimizer, a co-variate matrix adaptation evolution strategy or “CMA-ES,” uses neural feedback to iteratively evolve these stimuli and maximize the firing rate of a target neuron in early visual cortex.

Results: Our results validate a new method of creating optimal stimuli in early visual cortex. By combining an image generator and an optimizer, we can consistently develop “optimal stimuli” that vastly outperform traditional stimuli such as oriented gratings or uniform colors. These optimal stimuli tend to incorporate preferred orientations and colors as measured by single-modality tasks, showing how these features optimally combine in naturalistic images. However, they also include suboptimal orientations and colors, as well as more complex features like textures and shapes not traditionally associated with early visual cortex. These findings reveal the diverse tunings of early visual cortex in response to naturalistic stimuli.

Discussion: In this study, we demonstrate the power of using machine learning and neural feedback in stimulus design. Optimal stimuli vastly outperform traditional, single-modality stimuli and incorporate a diverse range of visual features such as color, shape and texture. This supports an overarching view of visual cortex as encoding multi-faceted, naturalistic features rather than simple, single-modality features like orientation or color alone. Furthermore, it shows how neural feedback can be used to guide stimulus design and discover new properties of neural circuits.

Acknowledgements: Supported by a training fellowship from UTHealth Houston Center for Clinical and Translational Sciences TL1 Program (Grant No. TL1 TR003169).

Poster 24

The yeast endonuclease Rnt1 processes and degrades a wide range of messenger RNAs

Lee-Ann Notice^1,2 and Ambro van Hoof^1,2

Department of Microbiology and Molecular Genetics, University of Texas Health Science Center at Houston, Houston, TX 77030, USA¹

 The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences, Houston, TX 77030, USA²

RNA processing and degradation are critical for maintaining cellular homeostasis. These processes regulate the steady-state level of RNAs in the cell (quantity control) and ensure the destruction of aberrant mRNAs that, if translated, would be deleterious to the cell (quality control). Indeed, mutations in several RNA processing and degradation enzymes frequently lead to human diseases such as cancer, developmental disorders, and neurodegenerative diseases. Key enzymes that participate in the processing and degradation of RNAs are endoribonucleases that cleave transcripts internally, providing entry sites for the RNA-degrading exoribonucleases. However, despite their crucial involvement in maintaining normal cell physiology, endoribonucleases remain largely uncharacterized. Furthermore, although a large fraction of newly transcribed RNA is degraded in the nucleus, nuclear mRNA degradation mechanisms have not been well studied. This project aims to characterize the role of a nuclear endoribonuclease, Rnt1, in eukaryotic mRNA processing and degradation. Although Rnt1 is known to process many ncRNAs by recognizing a double-stranded RNA stem loop with an AGNN sequence, its role in mRNA processing is much less defined. To first identify Rnt1 cleavage sites throughout the yeast transcriptome, we employed an RNA sequencing approach along with rigorous bioinformatics in a strategy called comparative parallel analysis of RNA ends (comPARE). This analysis confirmed the vast majority of ncRNA processing sites for Rnt1 and revealed many previously unknown Rnt1 cleavage sites within protein-coding regions, indicating an underappreciated role for the enzyme in mRNA degradation. Investigation into the sequence- and structure-specific determinants of Rnt1 mRNA target selection further revealed that they closely mirror the established requirements for the enzyme’s selection of ncRNA targets. We anticipate that our findings will fill considerable gaps in the current understanding of nuclear mRNA processing and degradation, and will contribute to a clearer picture of endoribonuclease function in these essential gene expression-regulating processes.

Poster 25

Are Difficult to Read Patient Education Materials Problematic? A Readability Assessment of Online Patient Education Resources Related to Breast Implant Removal and Breast Implant Illness

Stephen Parlamas,¹ Praneet Paidisetty,¹ Wendy Chen^1,2

¹McGovern Medical School at UTHealth, Houston, Texas

²Division of Plastic and Reconstructive Surgery, UTHealth, Houston, Texas

Introduction:

Hispanic women pursue breast augmentation surgery more than any other minority group.¹ Additionally, breast implant removals increased by 34.4% in 2019 and by 8% in 2020, perhaps due to rising concerns for breast implant-associated anaplastic large cell lymphoma.² However, this trend may not be reflected in all racial segments. Patient education materials (PEMs) with advanced readability can limit Hispanic patients’ health literacy and understanding of treatment plans, increasing barriers to breast implant removal and worsening health outcomes. When compared to non-Hispanic Whites, Hispanics attain less high school and college education, have less household income, and face higher poverty rates.³ These factors negatively influence health literacy. Despite the National Institutes of Health recommending PEMs be written at the 6^th-7^th grade readability level, online PEMs are often difficult to read. This study assessed and compared the readability of Spanish and English online breast implant removal PEMs.

Methods:

The English and Spanish terms “breast implant removal” and “extracción de implantes mamarios” were queried on Google and the first fifty Uniform Resource Locators (URLs) for each search term were selected. Included resources were non-video, free articles. Each URL was categorized into private business (private clinics, independent practice), institutional (governmental agencies, medical schools/teaching hospitals, medical sources), and academic (journal article, book chapter) groups. To ensure analyses provided useful insights into resources used by patients rather than surgeons, academic articles were regarded as a separate entity and analyzed independently. Readability scores were generated using the Simple Measure of Gobbledygook (SMOG) and the Spanish SMOG scales.

Results:

PEMs published by private businesses comprised the majority (> 60%) of both English and Spanish articles. Overall, online breast implant removal PEMs surpassed recommended reading levels in both English (x̄ = 12.2, highschool senior) and Spanish (x̄ = 11.0, high school junior). Among PEMs in the private business category, those written in English were more difficult to read than Spanish (p < 0.00001). As a whole, English PEMs were similarly more difficult to read than Spanish PEMs (p < 0.00001, excluding academic). There were no significant differences in readability when comparing English and Spanish institutional PEMs, English private business and English institutional PEMs, or Spanish private business and Spanish institutional PEMs.

 Conclusion:

Our findings show that the readability of English and Spanish materials for breast implant removal exceeds recommended levels. This discrepancy may have ramifications for certain groups as breast implant removal is a procedure that can have significant cosmetic and psychological impacts.⁴ It is vital that these patients are provided with educational material appropriate to their literacy level in order for them to understand the procedure, its implications, and indications. Additionally, the significant proportion of PEMs authored by private businesses may be concerning due to an inherent self-serving bias. Online information for breast implant removals should be revised and/or standardized materials should be created by healthcare providers and national organizations in order to more fully educate the public and prospective patients prior to intervention.

Poster 26

Staphylococcus aureus Breast Implant Infection Isolates Display Recalcitrance to Antibiotic Pocket Irrigants in vivo Despite Exhibiting Susceptibility in vitro

Duran Ramirez, JM^1,2, Gomez, J¹, Hanson BM^2,3, Myckatyn, TM⁴, Walker, JN^1,2

1. Department of Microbiology and Molecular Genetics, UTHealth Houston
2. Department of Epidemiology, Human Genetics, and Environmental Sciences, School of Public Health UTHealth Houston
3. Center for Antimicrobial Resistance and Microbial Genomics, UTHealth Houston
4. Division of Plastic and Reconstructive Surgery, Washington University School of Medicine

Breast implant associated infection (BIAI) following reconstructive surgery post-mastectomy is common, with a prevalence of up to 35%. For these patients, BIAI results in increased patient morbidity, including swelling, pain, and tissue necrosis, as well as escalated healthcare costs. Furthermore, these infections can result in the delay of critical oncologic treatments, such as radiation and chemotherapy. Additionally, BIAIs require explantation of the infected protheses and additional broad spectrum antibiotic treatment for complete resolution of the infection. To reduce infection rates, surgeons have implemented additional prevention strategies, including flushing the surgical pocket with a triple antibiotic pocket irrigant (TAPI) before implant placement. TAPI usually consists of 50,000 U bacitracin, 1 g cefazolin, and 80 mg gentamicin diluted in 500 mL of saline. Despite these additional strategies, infections rates remain high.

This study examines the efficacy of TAPI against Staphylococcus aureus, one of the most common causes of BIAIs, both in vitro and in vivo.

The antimicrobial resistance pattern of S. aureus BIAI isolates (117 and 158) and a reference strain (JE2) were assessed via Minimal inhibitory concentration (MIC) assays using TAPI, as well as the individual antibiotics that make up the solution. Additionally, the recalcitrance of the biofilm formed by these strains to TAPI was assessed. Furthermore, one BIAI isolate (117) and JE2 were further characterized in a mouse BIAI model. Finally, whole genome sequencing was performed on the BIAI isolates to identify potential virulence and antimicrobial resistance mechanisms that may contribute to recalcitrance to TAPI.

MIC assays revealed S. aureus BIAI isolates were susceptible to gentamicin, cefazolin, and TAPI, and resistant to bacitracin. JE2 was the only strain that was also resistant to gentamicin. All strains formed biofilm under standard in vitro conditions and these biofilms resisted TAPI treatment. Notably, in the mouse BIAI model, TAPI significantly reduced JE2 infection on the implant and in the surrounding tissue at 1- and 7- day post infection (dpi). In contrast, infection with the BIAI isolate (117) persisted out to 14-dpi, despite TAPI treatment. To gain insights into how the BIAI isolates were resisting TAPI we sequenced the strains. These data revealed that the sequence types differed among the S. aureus strains, JE2 is a ST 8, 117 is a ST 39, and 158 is a ST 45. Additionally, JE2 was the only strain to encode the mecA gene, which provides resistance to -lactams, while the BIAI isolates carry tet-38 and mepA/R and encodes resistance to tetracyclines. Furthermore, while all three strains carried 57 known virulence genes, the BIAI isolates share 6 factors not found in JE2, including cap8H-K, sec, selL. While these factors are known virulence determinants, it is unclear how they affect biofilm formation or recalcitrance to antimicrobials.

This study indicates that S. aureus BIAI isolates encode unique mechanisms that allow them to persist despite the use of prophylactic antibiotic treatment, such as TAPI, and promote chronic infection.

Acknowledgements: UTHealth Start-up funds, Rising Star Award, Plastic Surgery Foundation Grant.

Poster 27

Effect of EYA2 Knockout in Bleomycin Induced Pulmonary Fibrosis

Rene Girard^1,2, Scott Collum¹, Sarah Shin^1,2, Andrew Peters¹, Weizhen Bi¹, Harry Karmouty-Quintana^1,2

¹Department of Biochemistry and Molecular Biology, McGovern Medical School, The University of Texas Health Science Center at Houston (UTHealth), Houston, TX, 77030, USA¹.

²The University of Texas MD Anderson UTHealth Graduate School of Biomedical Sciences, Houston, TX, 77030, USA².

Background: Idopathic pulmonary fibrosis is a fatal chronic disease which causes scarring of the lung. The progressive scarring of the lung causes respiratory failure which necessitates lung transplant. Sine Oculis Homeobox Homolog 1 (SIX1) is a transcription factor that has been shown to increase in pulmonary fibrosis, and when knocked out has a protective effect against fibrosis. Overexpression of SIX1 also worsens the fibrotic response to a bleomycin challenge. Eyes absent 2 (EYA2) is a transcriptional activator in mice and has been shown to be a coactivator for SIX1. EYA2 is increased in bleomycin challenged mice and in fibroblasts from fibrotic lungs. I hypothesize that the knockout of EYA2 will have a decreased fibrotic injury due to the lack of transcriptional activation of SIX1.

Methods: For this poster we gave female EYA2 Flox^+/+ (knockout) and EYA2 Flox^-/- (wild type) mice a bleomycin challenge (2.5 units/kg) with PBS controls through oropharyngeal aspiration. Lung function tests were performed using a FlexiVent FX system (SCIREQ Inc.) after 21 days from the challenge. Mice were anesthetized with an IP injection of 5% avertin (0.012 mL/g), then a 18-guage cannula was inserted into the trachea. Respiratory mechanics like respiratory system resistance (Rrs), dynamic compliance (Crs), elastance (Ers), airway resistance (Rn), tissue damping (G), tissue elastance (H), and static compliance (Cst) were performed on the mice. Lungs were lavaged with PBS to collect bronchoalveolar lavage (BAL) fluid was collected, and total cell counts were performed. Cytospins of the BAL were taken for cellular differentials. Right lobes were collected for RNA and protein, with the left lobe collected for paraffin embedding. Masson’s trichrome was performed on the slides and an Ashcroft score was used to quantify the pulmonary fibrosis present.

Results: Flexivent data showed that there was a significance difference in WT PBS vs WT Bleomcyin challenged mice in G, H, Ers, Cst, and Rrs. The results show that there was a fibrotic injury in the mice due to the bleomycin challenged mice having a significant difference in lung function when compared to PBS controls. There was significance between WT Bleo and KO Bleomycin in H (tissue elastance). KO PBS and KO Bleomycin had significant differences in Cst. However, there is no significance in the EYA2 KO when comparing the PBS vs Bleomycin challenge. Cell counts show a significant increase in the number of cells present in the bleomycin challenged mice when compared to PBS.

Conclusion: Overall, the experiments show that bleomycin caused a fibrotic injury but there is not enough significance between the KO mice to say that lack of EYA2 made a difference in the injury. The significance in the KO in H (tissue elastance) means that there was a difference in the energy conservation of the alveoli, meaning that the KO had a harder time getting back to baseline when stretched to breathe. The lack of significance overall could be due to differences in the animals and the challenge of bleomycin being hard to perform. Giving the bleomycin through the trachea presents a challenge due to unequal amounts of bleomycin being absorbed. The experiment was also only conducted on female mice due to availability, but the male mice may show that significance due to their predisposition to a higher injury from bleomycin.

Poster 28

SARS-CoV-2 restructures host chromatin architecture

Ruoyu Wang ^1,2,†, Joo-Hyung Lee ^1,†,  Jieun Kim ^3,4, Feng Xiong ¹, Lana Al Hasani ^1,2, Yuqiang Shi ¹, Erin N. Simpson ^1,2, Xiaoyu Zhu ¹, Yi-Ting Chen^1,2, Pooja Shivshankar ^1,3,4, Joanna Krakowiak ¹, Yanyu Wang ^3,4, David M. Gilbert ⁵, Xiaoyi Yuan ^3,4, Holger K. Eltzschig ^2,3,4, Wenbo Li ^1,2,*

Affiliations:

¹ Department of Biochemistry and Molecular Biology, McGovern Medical School, University of Texas Health Science Center, Houston, 77030, TX, USA

² Graduate School of Biomedical Sciences, University of Texas MD Anderson Cancer Center and UTHealth, Houston, 77030, TX, USA

³ Department of Anesthesiology, McGovern Medical School, University of Texas Health Science Center, Houston, 77030, TX, USA

⁴ Center for Perioperative Medicine, McGovern Medical School, University of Texas Health Science Center, Houston, 77030, TX, USA

⁵ Laboratory of Chromosome Replication and Epigenome Regulation, San Diego Biomedical Research Institute, San Diego, 92121, CA, USA

ABSTRACTS:

Some viruses restructure host chromatin, influencing gene expression, with implications for disease outcome. Whether this occurs for SARS-CoV-2, the virus causing Covid-19, is largely unknown. Here, we characterized the 3D genome and epigenome of human cells after SARS-CoV-2 infection, finding widespread host chromatin restructuring featuring widespread A compartmental weakening, A-B mixing, reduced intra-TAD contacts and decreased H3K27ac euchromatin modification levels. Such changes were not found following common-cold virus HCoV-OC43 infection. Intriguingly, the cohesin complex was significantly depleted from intra-TAD regions, indicating that SARS-CoV-2 disrupts cohesin loop extrusion. These altered 3D genome/epigenome structures correlated with transcriptional suppression of interferon response genes by the virus, while increased H3K4me3 was found in the promoters of pro-inflammatory genes highly induced during severe Covid-19. These findings show that SARS-CoV-2 acutely rewires host chromatin, facilitating future studies of the long-term epigenomic impacts of its infection.

Poster 29

Multi-Center Comparison of Head Shape Outcomes for Three Minimally Invasive Strip Craniectomy Techniques for Sagittal Craniosynostosis.

Imran Rizvi, BS^1,2; Rami R. Hallac, PhD^1,2; Emily L. Geisler, MD³; Jeyna K. Perez, MS^1,2; Bar Y. Ainuz, MD^1,2; Mark P. Pressler, MD^1,2; Sarah A. Jackson, MS^1,2; Gary B. Skolnick, BS⁴; Matthew D. Smyth, MD⁶; James R. Seaward, MD^1,2; Christopher A. Derderian, MD^1,2; Lisa R. David, MD⁵; Christopher M. Runyan, MD, PhD⁵; Kamlesh B. Patel, MD, MSc⁴; Alex A. Kane, MD^1,2

1. Department of Plastic Surgery, University of Texas Southwestern Medical Center, Dallas, Texas, USA.
2. Analytical Imaging and Modeling Center, Children’s Health, Dallas, Texas, USA.
3. Division of Plastic Surgery, University of Texas Medical Branch, Galveston, Texas, USA.
4. Division of Plastic and Reconstructive Surgery, Department of Surgery, Washington University School of Medicine in St. Louis, St. Louis, Missouri, USA.
5. Department of Plastic and Reconstructive Surgery, Wake Forest University School of Medicine, Winston-Salem, North Carolina, USA.
6. Department of Neurosurgery, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.

 Background: There is currently a literature gap comparing entire head shape outcome following correction of sagittal craniosynostosis. This multi-center study provides analysis of long-term results following three different endoscopic strip craniectomy techniques for correction of sagittal craniosynostosis: spring-assisted strip craniectomy, wide-strip craniectomy with biparietal and bitemporal “barrel stave” wedge osteotomies plus helmet orthotic, and narrow-strip craniectomy plus orthotic without barrel staves.

Methods: Preoperative and postoperative 3D stereophotogrammetric images were collected from patients who underwent craniosynostosis surgery. Procedures were divided among institutions as such: Spring-assisted strip craniectomies were performed at Atrium Health Wake Baptist Health; Narrow strip craniectomies were performed at St. Louis Children’s Hospital by one craniofacial surgeon, while wide-vertex craniectomies were performed at St. Louis Children’s Hospital prior to 2010, and then continued at Children’s Medical Center Dallas. Preoperative and postoperative 3D whole-head composite images were generated for each procedure to visually represent outcomes at final follow-up and compared to age matched normal controls.

Results: Patients in the spring-assisted strip craniectomy group showed normalization of frontal bossing and skull height compared to age-matched controls, whereas patients undergoing wide-strip craniectomy showed greater correction of occipital bulleting. Patients in the narrow-strip craniectomy cohort had intermediate results between these outcomes. Nested aggregate head shapes showed good correction of head shapes from all techniques.

Conclusions: This is a large retrospective multi-center study illustrating whole head shape outcomes from three different craniectomy procedures. Although each showed some difference in loci of primary correction, all three surgical methods demonstrated good correction of primary scaphocephalic deformity.

Poster 30

Pontocerebellar hypoplasia (PCH) linked mutations in the cap component Csl4 impair RNA exosome function.

Khondakar Sayef Ahammed^1;2, and Ambro van Hoof ^*;1;2

1 Department of Microbiology and Molecular Genetics, University of Texas Health Science Center at Houston.

2 MD Anderson Cancer Center UTHealth Houston Graduate School of Biomedical Sciences.

Pontocerebellar hypoplasia (PCH) is a group of mendelian diseases characterized by abnormal brain development and other neurodegenerative complications that often lead to early childhood death. Several classes of PCH1 have been linked to mutations in the evolutionarily conserved RNA exosome complex that facilitates the degradation and processing of cytoplasmic and nuclear RNA from the 3’ end. Along with multiple mutations in the exosome core components such as EXOSC3 (pch1b), EXOSC8 (pch1c), and EXOSC9 (pch1d), only a single patient with a EXOSC1 (Csl4 in yeast) mutation (p.Ser35Leu) has been reported with clinical features of pontocerebellar hypoplasia type 1 (pch1f).

Here, we report another infant with clinical features of PCH and dilated cardiomyopathy. The patient’s whole-exome sequencing revealed a novel homozygous missense variant c.547C>T (p.Arg183Trp) in the EXOSC1 gene. Functional studies in the budding yeast model complemented with the human EXOSC1 variants show a growth-defect phenotype and significantly reduced protein levels compared to wild-type. Both the orthologous amino-acid substitutions in yeast Csl4 (Ala58Leu and Arg286Trp) are lethal when introduced in the functionally active truncated domains of yeast Csl4, indicating functional redundancy between the C-terminal and N-terminal domains of Csl4. In line with the previously identified importance of the C-terminal domain of Csl4 in the cytoplasmic function of RNA exosome, the pathogenic Arg183Trp variant show defect in cytoplasmic RNA exosome function, including nonstop mRNA decay. On the other hand, the yeast Csl4 variants show impaired genetic interactions between the nuclear cofactors (Rrp6 and Mpp6) and RNA exosome core complex.

Thus, PCH-related mutation in EXOSC1/Csl4 causes defects in RNA exosome function likely due to compromised protein stability and impaired functional interaction of RNA exosome core with the cofactors. Moreover, the budding yeast model of Csl4 will provide an in-vivo tool to investigate the molecular consequences of PCH-related defects in the RNA exosome.

Poster 31

Cortical processing during visually guided foraging in freely moving non-human primates

A.G. McConnell¹, M. Franch¹, V. Dragoi¹

¹Dept. of Neurobiology and Anatomy, The University of Texas Health Science Center at Houston, Houston, TX, USA

Foraging is a pervasive and natural behavior that animals exhibit while searching for food. To make efficient foraging decisions, animals must integrate sensory data, memory, and predictions to continually evaluate the decision about the location and duration of foraging and how to move between food locations. Historically, foraging has been studied using head-fixed animals performing foraging tasks on a computer or observing freely moving animals without recording neural information, limiting our understanding of the neural underpinnings of sensory integration of environmental cues during foraging. To bridge this knowledge gap, we utilize a novel approach using high-yield wireless electrical recordings and wireless eye tracking to study the neural activity in the prefrontal cortex and primary visual cortex while animals engage in an ethologically relevant foraging task. In this task, animals could freely move between two food patches with a monitor to display a color reflecting the reward probability at the given patch. Over time, animals’ behavioral strategies changed in response to the changing certainty of the visual stimuli during foraging. Additionally, neural activity in the dorsolateral prefrontal cortex (dlPFC) and V4 showed sensitivity to reward dynamics, such as cue information or reward probability during foraging. Finally, the dlPFC predicted the animals’ decision to change foraging patches. Overall, this work will allow us to uncover neural computations that orchestrate complex decisions involving sensory integration, memory, and prediction in a freely moving foraging task.

Poster 32

Disaggregated blunt cardiac injury screening is more informative than the composite

Zachary Terrell¹, Robin Wright¹, Thaddeus Puzio^1,2, Michael Wandling^1,2, John Harvin^1,2, Bryan Cotton^1,2, Lillian Kao^1,2, Laura Moore^1,2, David Meyer^1,2

 ¹McGovern Medical School at UT Health Houston, Houston, TX USA

²Department of Trauma/Critical Care Surgery, Red Duke Trauma Institute, Memorial Hermann Hospital, Houston, TX USA

Background: In 2012, guidelines were published establishing screening criteria for blunt cardiac injury (BCI). However, these criteria (abnormal EKG or elevated troponin) may be overly broad, resulting in unnecessary testing. Additionally, the impact of EKG abnormalities and troponin elevation on clinical outcomes remains undefined.

Aims: This study aims to explore the prognostic value of EKG and troponin elevation in patients with BCI.

Methods: Five-year (2016-2020) retrospective study of BCI in consecutive sternal fracture patients surviving >24h. Injury mechanism, demographic, and outcome data were collected from the trauma registry. 12-lead electrocardiogram, cardiac enzyme, and echocardiogram data were also collected. Patients screened positive for BCI if they had an abnormal EKG or troponin within the first 24h. Patients were then dichotomized by BCI screening status, and baseline characteristics were compared. Multivariate logistic regression was used to determine the association between EKG changes, troponin elevation, and mortality. Zero-inflated and negative binomial generalized linear models were used to model the effect of BCI on hospital and ICU lengths of stay (LOS).

Results: Of 959 sternal fracture patients, 464 (48%) screened positive for BCI. Age, sex, race, and mechanism of injury were similar between groups. BCI patients had a higher ISS (22±14 vs 17±11, p<0.0001) and presented with a higher heart rate (98±17 vs 89±24 bpm; p<0.0001) and lower GCS (15[13-15] vs 15[14-15]; p=0.0003). They also experienced more ventilator days (0[0-2] vs 0[0-0]), longer ICU LOS (0[0-7] vs 0[0-2] days) and hospital LOS (7[3-17] vs 4[2-9] days), and greater mortality (40[9%] vs 10[2%]; all p<0.0001). After controlling for confounders, screening positive for BCI remained associated with increased mortality. However, only the disaggregated ST segment abnormalities were associated with this risk (OR 3.1, 95%CI 1.4-6.7, p=0.004); conduction abnormalities, arrhythmias, and other EKG abnormalities were not. Troponin elevation was not associated with an increased risk of mortality, but initial and 48h peak troponin were both associated with increased ICU LOS (RR 1.19, 95%CI 1.06-1.34, p=0.004) and hospital LOS (RR 1.14, 95%CI 1.06-1.23, p<0.0001) (FIGURE).

Conclusion: In patients with sternal fracture, nearly half screened positive for BCI. Increased mortality was primarily associated with ST segment abnormalities, while troponin elevations were associated with increased ICU and hospital LOS. Revised screening criteria, including a consideration of specific EKG abnormalities, may help physicians better focus treatment efforts.

Poster 33

Use of an Open Partial Laryngectomy Technique for a Massive Laryngeal Leiomyosarcoma in a 40-year-old Man with Functional Preservation Postoperatively

Emily M. Cao¹, Gabriela Cruz¹, David Z. Allen MD², Joshua J. Kain MD³

1. Center for Clinical Research & Evidence-Based Medicine, McGovern Medical School, The University of Texas Health Science Center, Houston, TX, USA
2. Department of Otorhinolaryngology–Head and Neck Surgery, McGovern Medical School, The University of Texas Health Science Center, Houston, TX, USA
3. Houston Methodist Hospital Department of Otolaryngology-Head & Neck Surgery, Houston, TX, USA

Laryngeal cancers are one of the most common head and neck malignancies and can lead to significant morbidity and mortality. While most commonly squamous cell carcinoma in cellular origin, there are multiple different pathologies that can arise. Treatment most commonly entails an aggressive surgical approach, most frequently by a total laryngectomy with adjunctive radiotherapy for bulky tumors and long-term follow-up. However, this approach leaves the patient with life-long debilitation in swallow and speech function, even with proper rehabilitation. We present a patient with an atypical presentation of a glottic mass for which final pathology revealed a leiomyosarcoma (LMS). We report the first case in the English literature of a large, bulky LMS that was removed entirely through a partial vertical laryngectomy and pharyngectomy with maintained swallow and speech function postoperatively.

Poster 34

An Evaluation of Quality Metric Inequity Between Socioeconomic Status- Maternal Drop in Hemoglobin After Vaginal Delivery: A Single Site Analysis

Maitreyi Narayan*, BS¹; Barbara Orlando, MD, PhD; Swaroop Gantela MD²; Kavitha Gopal, MS³;

Michelle Narat, MS³; Seena John, MSN, CVRNII -BC, FACCN, SSBBP⁴; Bela Patel, MD, CMQ, FCCP¹

• McGovern Medical School at UTHealth, Houston, TX
• UTHealth School of Biomedical Informatics, Houston, TX
• UTHealth Texas Center of Healthcare Innovation, Houston, TX
• Memorial Hermann Hospital Quality Institute, Houston, TX

 Background: Collecting and analyzing health equity metrics are an essential part of healthcare quality improvement. Vizient is a healthcare improvement company that reports institutional equity metric data. Quality metric analysis provides useful clinical data to fulfill reporting requirements for governmental services, state agencies, and to compare institutional performance with clinical benchmarks and/or other organizations.

Objective: Analyze and determine the etiology of the consistent inequity of a Vizient core metric at Memorial Hermann Hospital-Texas Medical Center (MHH-TMC): Hemoglobin drop after an unremarkable vaginal delivery and the discrepancy between patients of different socioeconomic status (SES). At our institution (MHH-TMC) high SES patients have a higher rate of hemoglobin drop than low SES patients after an unremarkable vaginal delivery.

Materials and Methods: MHH-TMC’s Vizient year 2021 vaginal deliveries (n=725) were separated into buckets based on SES and fallout rate (those patients with a drop in hemoglobin >1 gram per deciliter). These SES and fallout buckets were compared based on the following variables: rates of gestational hypertension, episiotomy, gestational diabetes, transfusion, manual perineal protection, substance use, Medicare severe diagnosis related groups, and difference in blood loss. Additionally, an analysis of SES reporting was done using the Vizient Vulnerability Index (VVI) tool, to assess if the method of reporting SES (insurance type vs census data) would impact the inequity between the two groups.

Results/Future Outlook: None of the clinical variables tested explained the inequity of this metric between SES at MHH-TMC. There was no statistically significant difference in blood loss between the two groups.  We did observe that there was a statistically significant difference in body mass index (BMI) and total blood volume (TBV) between the two groups, in that high SES patients had lower BMIs and TBVs than their lower SES counterparts which could contribute to the inequity due to hemodilution or hemoconcentration or a larger drop in hemoglobin with an absolute similar degree of blood loss in both groups. Additionally, when reporting SES by census data, not insurance type, the inequity of this metric disappeared. This could suggest that the inequity is not derived from the quality of care delivered, but rather the way SES is delineated.

Poster 35

How does your garden grow: assessing the landscape of kangaroo care for premature infants in the neonatal intensive care unit

Bentley HA¹, Sabroske E, DO, MBA¹, Tran TD¹, Li W, PhD²., Khan AK, MD¹, Ruggiero J, MD¹

1. Department of Neonatology, Division of Pediatrics, McGovern Medical School, University of Texas Health Science Center, Houston, TX, USA
2. Division of Clinical and Translational Sciences, McGovern Medical School, University of Texas Health Science Center, Houston, TX, USA

Background: Kangaroo Care (KC) is the act of holding an infant upright in direct skin-to-skin contact with the caregiver. Research shows that KC improves infant thermal regulation and cardiorespiratory stability, promotes early parent-infant bonding, lowers pain scores and infection rates, and improves survival. In the Children’s Memorial Hermann Hospital (CMHH) NICU, extremely premature (EP) infants born < 28 weeks receive humidity protocol (HP) for 10 days in addition to standard incubator care (SIC). HP serves to decrease evaporative water losses and improve thermoregulation across an immature skin barrier. Premature (PT) infants born between 28-34 weeks require SIC only. In the absence of sufficient data to describe humidity procedure, variation exists among NICUs in this country. We hypothesize that HP may be a barrier to delay KC in the CMHH NICU.

Methods: This single-site retrospective cross-sectional study evaluated 122 PT infants born between 08/2018 – 04/2021 at CMHH. A random number sequence was used to select our sample population from >2,500 PT infants. To account for variation, PT infants were stratified into two groups: 28 – 31 weeks gestation (younger PT infants) and 32-34 weeks gestation (older PT infants). The primary outcome variable was time to first KC. We also analyzed factors considered to be potential barriers to KC, including respiratory support, oxygen requirement, vasopressor use, intraventricular hemorrhage (IVH), umbilical catheter (UVC), central line (PICC), and peripheral arterial line. We compared the PT group to our previously studied EP infants to assess humidity as a potential barrier. Data were analyzed using descriptive statistics, Fisher’s exact test, and the non-parametric Kruskal-Wallis test.

Results: Median time to first KC in PT infants is 3.4 days (IQR 1.7 – 8.0). By subgroup, median time to first KC was 11.2 days (IQR 10.3 – 15.5) in EP infants, 5.1 days (IQR 2.5 – 9.4) in younger PT infants, and 2.4 days (IQR 1.1 – 4.5) in older PT infants. Analysis of potential barriers to KC showed that EP infants were more likely to be intubated, require vasopressor support, have evidence of IVH, and have a PICC present at the time of the first KC. In contrast, PT infants were more likely to have a UVC compared to EP infants. Younger PT infants were more likely to have a UVC or PICC than older PT infants.

Conclusions: Time to first KC decreases with increasing gestational age, correlating with the decreased need for humidity protocol. Therefore, use of humidity is the most common reason for delaying time to first KC in infants born prematurely. While humidity is essential for thermoregulation in premature infants, the duration of use and effects of interrupting humidity to promote kangaroo care have not been adequately studied. An RCT of the duration of humidity and interruption of humidity is warranted to avoid depriving infants of the benefits of kangaroo care.

Poster 36

Benefits of Social Emotional Learning Delivered in Community Based After School Programs

Sarah Chiang¹, Jasmine Zhou², Funlola Are¹, Ronald Acierno¹

Department of Psychiatry and Behavioral Sciences, McGovern Medical School, Houston, TX, USA¹

Department of Psychiatry, Seattle Children’s Hospital, Seattle, Washington, USA²

Social emotional learning (SEL) is the process of developing self-awareness, self-control, and interpersonal skills which are vital skills for children to learn and develop. Existing research has shown that youth with greater social emotional skills perform better academically, reporter greater friendships and limited research has suggested that these benefits may also carry over to reduce internalizing and externalizing behaviors in children. Previous studies of social emotional interventions have shown effectiveness in school settings, but there is limited research on SEL interventions outside of school settings. The present work sought to explore the effectiveness of SEL programs in after school settings by implementing an SEL intervention based on the Merrell’s Strong Start and Strong Kids curriculum.  and adapted for brevity and time. The goal of the intervention was to target youth in grades k-8^th who showed behavioral challenges and engage them in the intervention which focused on identifying emotions through body and social cues, improving problem solving skills, and learning coping skills. The program lasted for six weeks with one session occurring each week for a total of six sessions. Sessions lasted between 35-45 minutes and were conducted in groups composing of between six to eight students. The curricula were delivered in a community service organization and administered by clinical psychology doctoral fellows. 121 youth participated in the intervention (mean age = 8.43, SD = 2.37). Boys made up the majority of participants N= 67 and N= 53 girls. Students were identified by staff at the community organization and screened for participation utilizing a social emotional screener developed for the program and adapted from the Social, Academic, and Emotional Behavior Risk Screener (SAEBRS). Students who scored above the minimum cutoff and below the maximum cutoff were invited to participate in the group sessions. Prior to the first session, organization staff who knew the students well completed the Devereux Student Strengths Assessment, a standardized rating scale designed to assess social-emotional competence. This assessment was readministered as a post-test following group completion. Data analyses were performed in SPSS. Descriptive analyses indicated that the majority of youth screened were experiencing significant internalizing symptoms and emotion dysregulation challenges, less so than social deficits-as measured by the screener. Correlational analyses revealed that students whose staff members reported that they were experiencing the greatest internalizing symptoms also scored highest on pre-tests (indicative of greater social-emotional difficulties). Further, social skills difficulties were significantly correlated to internalizing problems, and emotion dysregulation (p> .05). Overall, pre-post revealed statistically significant improvement to social emotional skills. Notably, larger improvement was observed for participants who scored the highest on the pre-test.  These results show the effectiveness and benefit of implementing SEL programs in afterschool programs. This preliminary data could be used to advocate for more SEL programs, especially in underserved communities.

Poster 37

Developing of an Operator-independent Deep Segmentation Model to Measure Tumor Volume in MRIs for Glioblastoma

Tanjida Kabir M.S.^1*; Kang-Lin Hsieh Ph.D.^1*; Luis Nunez, M.D²; Yu Cai M.D.²; Yu-Chun Hsu Ph.D.¹; Juan C Rodriguez Quintero M.D.²; Octavio Arevalo M.D.³; Kangyi Zhao M.S.⁴; Jay-Jiguang Zhu M.D., Ph.D.⁴; Roy F Riascos M.D.²; Xiaoqian Jiang Ph.D.¹; Shayan Shams Ph.D.⁵ [* Equal Contribution]

1. Center Director of Secure Artificial Intelligence For Healthcare, University of Texas Health Science Center at Houston, TX
2. Department of Diagnostic and Interventional Imaging, The University of Texas Health Science Center at Houston, TX
3. Louisiana State University Health Shreveport, Department of Radiology, LA
4. Department of Neurosurgery, The University of Texas Health Science Center at Houston, TX
5. Department of Applied Data Science, San Jose State University, CA

Glioblastoma (GBM) is the most common malignant primary brain tumor. Despite advanced therapeutic advancements, the median overall survival for GBM patients is 11 months in real-world population in USA and 14.6 to 20.9 months for clinical trial patients. As complete tumor removal is impossible due to the infiltrative nature of the disease, maximum safe resection of enhancing tumors is associated with better survival. Therefore, measuring tumor volumes accurately on post-craniotomy and subsequent follow-up magnetic resonance imaging (MRI) is paramount for determining tumor status regarding treatment responses and clinical decision-making. MRI is the most common imaging evaluation method in GBM. Preoperative (preop) MRIs are crucial for initial diagnosis and surgical planning, while follow-up MRIs will help evaluate treatment response, GBM recurrence, and treatment-related side effects. However, radiologists manually segment tumor regions from normal brain tissue for follow-up evaluations, which is time-consuming, error-prone, and operator-dependent. Several deep-learning (DL) models have been developed utilizing preop MRIs, but their performances have yet to be evaluated on follow-up MRIs. We address this gap by investigating these DL models’ performances and their generalizability on independent preop and follow-up MRIs. We subsequently bridge the gap by developing the first deep segmentation model to estimate the tumor volume considering the brain and tumor structural variations utilizing the largest cohort with follow-up MRI. Ten state-of-the-art DL models were selected for performance evaluation on independent MRIs for segmenting tumor sub-regions/signals- Fluid attenuation inversion recovery (FLAIR) region (FR), enhancing tumor region (ER), and non-enhancing necrosis region (NENR). They were trained by 369 preop BraTS’20 MRIs and tested by 50 pairs of institutional preop and follow-up MRIs. Additionally, an encoder-decoder-based follow-up model was developed using 311 institutional follow-up MRIs. The encoder learns essential features from MRIs to identify tumor location and shape by consolidating domain-specific and problem-specific features to reduce the tumor’s irregularity variation. The decoder’s deterministic layers were replaced by Bayesian layers to quantify the uncertainty in the model’s prediction. Transfer learning and uncertainty information were utilized to identify complex regions in follow-up MRIs. The dice similarity coefficient (DSC) was used to evaluate model performance. BraTS’20 trained models’ performance decreased by 3-20% among the 10 DL models and showed a lack of generalizability (p-value<0.05) on independent MRIs. Performance reduction was more pronounced in follow-up MRIs compared to preop MRIs (13.05% in independent preop MRI and 19.04% in follow-up MRIs). The proposed follow-up model achieved higher performance than all preop-based models. The DSC scores are 0.78, 0.87, and 0.92 for FR, ER and NENR, respectively. Structural changes following surgery decrease the BraTS’20-based models’ performance on follow-up MRIs for tumor volume measurement. Therefore, a follow-up MRI-based model is essential for accurate and generalizable results. The developed model utilizes transfer learning and uncertainty information and incorporates post-operation brain structural changes. Therefore, it performs better than preop-based models and can measure tumors accurately and efficiently in an operator-independent manner.