DESI-MSI-guided exploration of metabolic-phenotypic relationships reveals a correlation between PI 38:3 and proliferating cells in clear cell renal cell carcinoma via single-section co-registration of multimodal imaging.

A workflow has been evaluated that utilizes a single tissue section to obtain spatially co-registered, molecular, and phenotypical information suitable for AI-enabled image analysis. Desorption electrospray ionization mass spectrometry imaging (DESI-MSI) was used to obtain molecular information followed by conventional histological staining and immunolabelling. The impact of varying DESI-MSI conditions (e.g., heated transfer line (HTL) temperature, scan rate, acquisition time) on the detection of small molecules and lipids as well as on tissue integrity crucial for integration into typical clinical pathology workflows was assessed in human kidney. Increasing the heated transfer line temperature from 150 to 450 °C resulted in a 1.8-fold enhancement in lipid signal at a scan rate of 10 scans/s, while preserving histological features. Moreover, increasing the acquisition speed to 30 scans/s yielded superior lipid signal when compared to 10 scans/s at 150 °C. Tissue morphology and protein epitopes remained intact allowing full histological assessment and further multiplex phenotyping by immunofluorescence (mIF) and immunohistochemistry (mIHC) of the same section. The successful integration of the workflow incorporating DESI-MSI, H&E, and immunolabelling on a single tissue section revealed an accumulation of ascorbic acid in regions of focal chronic inflammatory cell infiltrate within non-cancerous kidney tissue. Additionally, a strong positive correlation between PI 38:3 and proliferating cells was observed in clear cell renal cell carcinoma (ccRCC) showing the utility of this approach in uncovering molecular associations in disease pathology.

A phase I open-label, dose-escalation study of NUC-3373, a targeted thymidylate synthase inhibitor, in patients with advanced cancer (NuTide:301).

PURPOSE: 5-fluorouracil (5-FU) is inefficiently converted to the active anti-cancer metabolite, fluorodeoxyuridine-monophosphate (FUDR-MP), is associated with dose-limiting toxicities and challenging administration schedules. NUC-3373 is a phosphoramidate nucleotide analog of fluorodeoxyuridine (FUDR) designed to overcome these limitations and replace fluoropyrimidines such as 5-FU. PATIENTS AND METHODS: NUC-3373 was administered as monotherapy to patients with advanced solid tumors refractory to standard therapy via intravenous infusion either on Days 1, 8, 15 and 22 (Part 1) or on Days 1 and 15 (Part 2) of 28-day cycles until disease progression or unacceptable toxicity. Primary objectives were maximum tolerated dose (MTD) and recommended Phase II dose (RP2D) and schedule of NUC-3373. Secondary objectives included pharmacokinetics (PK), and anti-tumor activity. RESULTS: Fifty-nine patients received weekly NUC-3373 in 9 cohorts in Part 1 (n = 43) and 3 alternate-weekly dosing cohorts in Part 2 (n = 16). They had received a median of 3 prior lines of treatment (range: 0-11) and 74% were exposed to prior fluoropyrimidines. Four experienced dose-limiting toxicities: two Grade (G) 3 transaminitis; one G2 headache; and one G3 transient hypotension. Commonest treatment-related G3 adverse event of raised transaminases occurred in < 10% of patients. NUC-3373 showed a favorable PK profile, with dose-proportionality and a prolonged half-life compared to 5-FU. A best overall response of stable disease was observed, with prolonged progression-free survival. CONCLUSION: NUC-3373 was well-tolerated in a heavily pre-treated solid tumor patient population, including those who had relapsed on prior 5-FU. The MTD and RP2D was defined as 2500 mg/m2 NUC-3373 weekly. NUC-3373 is currently in combination treatment studies. TRIAL REGISTRATION: Clinicaltrials.gov registry number NCT02723240. Trial registered on 8th December 2015. https://clinicaltrials.gov/study/NCT02723240 .

Automated reporting of cervical biopsies using artificial intelligence.

When detected at an early stage, the 5-year survival rate for people with invasive cervical cancer is 92%. Being aware of signs and symptoms of cervical cancer and early detection greatly improve the chances of successful treatment. We have developed an Artificial Intelligence (AI) algorithm, trained and evaluated on cervical biopsies for automated reporting of digital diagnostics. The aim is to increase overall efficiency of pathological diagnosis and to have the performance tuned to high sensitivity for malignant cases. Having a tool for triage/identifying cancer and high grade lesions may potentially reduce reporting time by identifying areas of interest in a slide for the pathologist and therefore improving efficiency. We trained and validated our algorithm on 1738 cervical WSIs with one WSI per patient. On the independent test set of 811 WSIs, we achieved 93.4% malignant sensitivity for classifying slides. Recognising a WSI, with our algorithm, takes approximately 1.5 minutes on the NVIDIA Tesla V100 GPU. Whole slide images of different formats (TIFF, iSyntax, and CZI) can be processed using this code, and it is easily extendable to other formats.

Deficiency of the paternally-expressed imprinted Peg3 gene in mice has sexually dimorphic consequences for offspring communication and social behaviour.

Introduction: Imprinted genes are expressed from one parental allele as a consequence of epigenetic processes initiated in the germline. Consequently, their ability to influence phenotype depends on their parent-of-origin. Recent research suggests that the sex of the individual expressing the imprinted gene is also important. We have previously reported that genetically wildtype (WT) dams carrying and caring for pups mutant for PEG3 exhibit anxiety-like behaviours and their mutant pups show a reduction in ultrasonic vocalisation when separated from their mothers. Sex-specificity was not examined. Methods: WT female mice were mated with WT, heterozygous Peg3-/+ or homozygous Peg3-/- studs to generate all WT (control), 50:50 mixed or 100% mutant litters, respectively, followed by behavioural assessment of both dams and their pups. Results: We reproduced our original finding that WT dams carrying and caring for 100% mutant litters exhibit postpartum anxiety-like symptoms and delayed pup retrieval. Additionally, these WT dams were found to allocate less time to pup-directed care behaviours relative to controls. Male Peg3-deficient pups demonstrated significantly reduced vocalisation with a more subtle communication deficit in females. Postweaning, male mutants exhibited deficits across a number of key social behaviours as did WT males sharing their environment with mutants. Only modest variations in social behaviour were detected in experimental females. Discussion: We have experimentally demonstrated that Peg3 deficiency confined to the offspring causes anxiety in mouse mothers and atypical behaviour including deficits in communication in their male offspring. A male-specific reduction in expression PEG3 in the fetally-derived placenta has previously been associated with maternal depression in human pregnancy. Maternal mood disorders such as depression and anxiety are associated with delays in language development and neuroatypical behaviour more common in sons. Peg3 deficiency could drive the association of maternal and offspring behavioural disorders reported in humans.

The burden of psychological trauma and post-traumatic stress disorder among adults with congenital heart disease: PTSD in ACHD.

Psychological trauma, symptoms of post-traumatic stress disorder (PTSD), and mental health conditions are common in adult congenital heart disease (ACHD). There is a gap in research examining PTSD in ACHD using the current Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5) criteria in assessing patient characteristics and experiences with trauma-focused treatment. Surveys were offered to outpatients over a 6-month enrollment period to be completed by way of a QR code on their personal smart phone. Patient-reported items include a detailed medical and psychosocial history, the Oslo social support scale, adverse childhood experiences survey, and the PTSD checklist for DSM-5. Of 158 patients (77% moderate or complex heart disease) who provided complete data, a provisional diagnosis of PTSD was found in 48 patients (30%) using a PTSD checklist for DSM-5 cut-off score of ≥31. A positive PTSD screen was associated with younger age, nonwhite race, presence of heart failure, lower New York Heart Association functional class, lower linear quality of life score, lower Oslo social support scale score, an insecure caregiver relation, period of unemployment, emergency department visits, medication nonadherence, and coexisting mental health disorders. Complexity of heart disease and number of surgical and/or catheter-based interventions were not associated with PTSD, although having undergone no cardiac surgeries until adulthood (aged ≥18 years) was associated with a lower prevalence of PTSD. Those who screened positive for PTSD were more likely to report multiple traumatic events, including noncardiac traumatic events. Only 14 of 48 patients (29%) reported a known diagnosis of PTSD, although 44 patients (92%) reported having ever seen a mental health provider. A total of 18 patients (38%) reported currently having a mental health provider. A total of 30 patients (62%) had heard of at least 1 evidence-based trauma-informed therapy, and 14 (29%) had tried at least 1. In conclusion, using the DSM-5 criteria, we observed a high prevalence of potential PTSD in ACHD associated with several novel cardiac and psychosocial patient factors. Future longitudinal studies will be necessary to establish causality. Few patients with ACHD have been formally diagnosed with PTSD or have experience with evidence-based trauma-informed therapies.

Magnifying Networks for Histopathological Images with Billions of Pixels.

Amongst the other benefits conferred by the shift from traditional to digital pathology is the potential to use machine learning for diagnosis, prognosis, and personalization. A major challenge in the realization of this potential emerges from the extremely large size of digitized images, which are often in excess of 100,000 × 100,000 pixels. In this paper, we tackle this challenge head-on by diverging from the existing approaches in the literature-which rely on the splitting of the original images into small patches-and introducing magnifying networks (MagNets). By using an attention mechanism, MagNets identify the regions of the gigapixel image that benefit from an analysis on a finer scale. This process is repeated, resulting in an attention-driven coarse-to-fine analysis of only a small portion of the information contained in the original whole-slide images. Importantly, this is achieved using minimal ground truth annotation, namely, using only global, slide-level labels. The results from our tests on the publicly available Camelyon16 and Camelyon17 datasets demonstrate the effectiveness of MagNets-as well as the proposed optimization framework-in the task of whole-slide image classification. Importantly, MagNets process at least five times fewer patches from each whole-slide image than any of the existing end-to-end approaches.

Snapshots of the Reaction Coordinate of a Thermophilic 2'-Deoxyribonucleoside/ribonucleoside Transferase.

Nucleosides are ubiquitous to life and are required for the synthesis of DNA, RNA, and other molecules crucial for cell survival. Despite the notoriously difficult organic synthesis of nucleosides, 2'-deoxynucleoside analogues can interfere with natural DNA replication and repair and are successfully employed as anticancer, antiviral, and antimicrobial compounds. Nucleoside 2'-deoxyribosyltransferase (dNDT) enzymes catalyze transglycosylation via a covalent 2'-deoxyribosylated enzyme intermediate with retention of configuration, having applications in the biocatalytic synthesis of 2'-deoxynucleoside analogues in a single step. Here, we characterize the structure and function of a thermophilic dNDT, the protein from Chroococcidiopsis thermalis (CtNDT). We combined enzyme kinetics with structural and biophysical studies to dissect mechanistic features in the reaction coordinate, leading to product formation. Bell-shaped pH-rate profiles demonstrate activity in a broad pH range of 5.5-9.5, with two very distinct pKa values. A pronounced viscosity effect on the turnover rate indicates a diffusional step, likely product (nucleobase1) release, to be rate-limiting. Temperature studies revealed an extremely curved profile, suggesting a large negative activation heat capacity. We trapped a 2'-fluoro-2'-deoxyarabinosyl-enzyme intermediate by mass spectrometry and determined high-resolution structures of the protein in its unliganded, substrate-bound, ribosylated, 2'-difluoro-2'-deoxyribosylated, and in complex with probable transition-state analogues. We reveal key features underlying (2'-deoxy)ribonucleoside selection, as CtNDT can also use ribonucleosides as substrates, albeit with a lower efficiency. Ribonucleosides are the building blocks of RNA and other key intracellular metabolites participating in energy and metabolism, expanding the scope of use of CtNDT in biocatalysis.

Human 2'-Deoxynucleoside 5'-Phosphate N-Hydrolase 1: The Catalytic Roles of Tyr24 and Asp80.

The human enzyme 2'-deoxynucleoside 5'-phosphate N-hydrolase 1 (HsDNPH1) catalyses the hydrolysis of 5-hydroxymethyl-2'-deoxyuridine 5'-phosphate to generate 5-hydroxymethyluracil and 2-deoxyribose-5-phosphate via a covalent 5-phospho-2-deoxyribosylated enzyme intermediate. HsDNPH1 is a promising target for inhibitor development towards anticancer drugs. Here, site-directed mutagenesis of conserved active-site residues, followed by HPLC analysis of the reaction and steady-state kinetics are employed to reveal the importance of each of these residues in catalysis, and the reaction pH-dependence is perturbed by each mutation. Solvent deuterium isotope effects indicate no rate-limiting proton transfers. Crystal structures of D80N-HsDNPH1 in unliganded and substrate-bound states, and of unliganded D80A- and Y24F-HsDNPH1 offer atomic level insights into substrate binding and catalysis. The results reveal a network of hydrogen bonds involving the substrate and the E104-Y24-D80 catalytic triad and are consistent with a proposed mechanism whereby D80 is important for substrate positioning, for helping modulate E104 nucleophilicity, and as the general acid in the first half-reaction. Y24 positions E104 for catalysis and prevents a catalytically disruptive close contact between E104 and D80.

Considerations for Gender-Affirming Hormonal and Surgical Care Among Transgender and Gender Diverse Adolescents and Adults With Congenital Heart Disease.

BACKGROUND: Transgender and gender diverse (TGD) individuals and long-term survivors with adult congenital heart disease (ACHD) are both growing populations with specialized needs. No studies assess temporal trends or evaluate the care of TGD individuals with ACHD. METHODS AND RESULTS: Meetings between congenital cardiology and gender-affirming care specialists identified unique considerations in TGD individuals with ACHD. A retrospective chart review was then performed to describe patient factors and outpatient trends in those with an ACHD diagnosis undergoing gender-affirming hormonal or surgical care (GAHT/S) at 1 adult and 1 pediatric tertiary care center. Thirty-three TGD individuals with ACHD were identified, 21 with a history of GAHT/S. Fourteen (66%) had moderate or complex ACHD, 8 (38%) identified as transgender male, 9 (43%) transgender female, and 4 (19%) other gender identities. Three had undergone gender-affirming surgery. There were zero occurrences of the composite end point of unplanned hospitalization or thrombotic event over 71.1 person-years of gender-affirming care. Median age at first gender-affirming appointment was 16.8 years [interquartile range 14.8-21.5]. The most common treatment modification was changing estradiol administration from oral to transdermal to reduce thrombotic risk (n=3). An increasing trend was observed from zero TGD patients with ACHD attending a gender diversity appointment in 2012 to 14 patients in 2022. CONCLUSIONS: There is a growing population of TGD patients with ACHD and unique medical and psychosocial needs. Future studies must fully evaluate the reassuring safety profile observed in this small cohort. We share 10 actionable care considerations for providers with a goal of overseeing a safe and fulfilling gender transition across all TGD patients with ACHD.

Transcription Factor NFE2L1 Decreases in Glomerulonephropathies after Podocyte Damage.

Podocyte cellular injury and detachment from glomerular capillaries constitute a critical factor contributing to kidney disease. Notably, transcription factors are instrumental in maintaining podocyte differentiation and homeostasis. This study explores the hitherto uninvestigated expression of Nuclear Factor Erythroid 2-related Factor 1 (NFE2L1) in podocytes. We evaluated the podocyte expression of NFE2L1, Nuclear Factor Erythroid 2-related Factor 2 (NFE2L2), and NAD(P)H:quinone Oxidoreductase (NQO1) in 127 human glomerular disease biopsies using multiplexed immunofluorescence and image analysis. We found that both NFE2L1 and NQO1 expressions were significantly diminished across all observed renal diseases. Furthermore, we exposed human immortalized podocytes and ex vivo kidney slices to Puromycin Aminonucleoside (PAN) and characterized the NFE2L1 protein isoform expression. PAN treatment led to a reduction in the nuclear expression of NFE2L1 in ex vivo kidney slices and podocytes.

Survivors of Pediatric Hematopoietic Stem Cell Transplantation Exhibit Progressive Diastolic Dysfunction Over Years of Follow-Up.

Patients who have undergone hematopoietic stem cell transplantation (HSCT) in childhood have a higher risk of diastolic heart failure (HF). The rate of progression of diastolic dysfunction in aging pediatric patients is unknown and is more difficult to assess in young patients secondary to changes in diastolic indices as they grow. HSCT recipients at our center were previously found to have decline in diastolic function indices at 1 year after HSCT. This study provides follow-up of this cohort, using age-normalized z-scores to assess whether the decline in diastolic function noted at 1-year post-HSCT persists, worsens, or improves over time. Patients age <21 years who underwent HSCT at Boston Children's Hospital/Dana-Farber Cancer Center between 2005 and 2008 with ≥3 surveillance echocardiograms, including 1 performed pre-HSCT, were included. Diastolic measures included mitral inflow (E/A ratio) and Doppler tissue imaging of left ventricular lateral wall (LV lateral e'), LV septal wall (septal e') and right ventricular free wall (RV e'). Systolic function was measured by LV ejection fraction (LVEF). Normalization by age was done using z-scores, and >±2 SD was defined as abnormal in linear modeling of diastolic dysfunction and systolic dysfunction over time. In a subset of patients with adequate post-HSCT images of the entire left atrium (LA), LA volume and LA strain analyses also were performed. The study cohort comprised 61 patients (41% female; median age at HSCT, 10.7 years; median follow-up, 7.4 years). Diastolic index z-scores declined by -.045/year for LV lateral e', -.06/year for LV septal e', and -.14/year for RV e' (P < .01). The E/A ratio z-score increased by .034/year (P = .028). Linear modeling demonstrated that LV lateral e' and LV septal e' would become abnormal at 25 and 20 years post-HSCT, respectively, whereas RV e' would become abnormal sooner, at 12.6 years. LVEF z-score declined by -.04/year (P < .01) and was estimated to become abnormal at 40 years post-HSCT. Exposure to total body irradiation (TBI) was associated with worsening diastolic indices, lower LVEF (P ≤ .002), and decreased LA reservoir strain (42.0% versus 45.0%; P = .016) and conduit strain (-31.5% versus -35.1%; P = .029), although there was significant overlap between TBI and anthracycline exposure. Treatment with anthracyclines even at low doses (median, 150 mg/m2) was associated with declining LVEF but not with changes in diastolic indices. Long-term survivors of childhood HSCT exhibit declines in both LV and RV diastolic function indices. These results inform the rate of progression of LV and RV diastolic dysfunction indices over time in long-term survivors of pediatric HSCT. A significant association was observed between TBI and diastolic dysfunction and a decline in LVEF. Treatment with anthracyclines even at low doses was associated with a mild decline in LVEF. Our results can inform a lifespan perspective on disease management in this population, encourage clinicians and patients to be vigilant in following guideline-directed surveillance echocardiography, and inform anticipatory responses by clinicians as patients transition from pediatric care to adult care.

Human 2'-Deoxynucleoside 5'-Phosphate N-Hydrolase 1: Mechanism of 2'-Deoxyuridine 5'-Monophosphate Hydrolysis.

The enzyme 2'-deoxynucleoside 5'-phosphate N-hydrolase 1 (DNPH1) catalyzes the N-ribosidic bond cleavage of 5-hydroxymethyl-2'-deoxyuridine 5'-monophosphate to generate 2-deoxyribose 5-phosphate and 5-hydroxymethyluracil. DNPH1 accepts other 2'-deoxynucleoside 5'-monophosphates as slow-reacting substrates. DNPH1 inhibition is a promising strategy to overcome resistance to and potentiate anticancer poly(ADP-ribose) polymerase inhibitors. We solved the crystal structure of unliganded human DNPH1 and took advantage of the slow reactivity of 2'-deoxyuridine 5'-monophosphate (dUMP) as a substrate to obtain a crystal structure of the DNPH1:dUMP Michaelis complex. In both structures, the carboxylate group of the catalytic Glu residue, proposed to act as a nucleophile in covalent catalysis, forms an apparent low-barrier hydrogen bond with the hydroxyl group of a conserved Tyr residue. The crystal structures are supported by functional data, with liquid chromatography-mass spectrometry analysis showing that DNPH1 incubation with dUMP leads to slow yet complete hydrolysis of the substrate. A direct UV-vis absorbance-based assay allowed characterization of DNPH1 kinetics at low dUMP concentrations. A bell-shaped pH-rate profile indicated that acid-base catalysis is operational and that for maximum kcat/KM, two groups with an average pKa of 6.4 must be deprotonated, while two groups with an average pKa of 8.2 must be protonated. A modestly inverse solvent viscosity effect rules out diffusional processes involved in dUMP binding to and possibly uracil release from the enzyme as rate limiting to kcat/KM. Solvent deuterium isotope effects on kcat/KM and kcat were inverse and unity, respectively. A reaction mechanism for dUMP hydrolysis is proposed.

Impact of Different Mammography Systems on Artificial Intelligence Performance in Breast Cancer Screening.

Artificial intelligence (AI) tools may assist breast screening mammography programs, but limited evidence supports their generalizability to new settings. This retrospective study used a 3-year dataset (April 1, 2016-March 31, 2019) from a U.K. regional screening program. The performance of a commercially available breast screening AI algorithm was assessed with a prespecified and site-specific decision threshold to evaluate whether its performance was transferable to a new clinical site. The dataset consisted of women (aged approximately 50-70 years) who attended routine screening, excluding self-referrals, those with complex physical requirements, those who had undergone a previous mastectomy, and those who underwent screening that had technical recalls or did not have the four standard image views. In total, 55 916 screening attendees (mean age, 60 years ± 6 [SD]) met the inclusion criteria. The prespecified threshold resulted in high recall rates (48.3%, 21 929 of 45 444), which reduced to 13.0% (5896 of 45 444) following threshold calibration, closer to the observed service level (5.0%, 2774 of 55 916). Recall rates also increased approximately threefold following a software upgrade on the mammography equipment, requiring per-software version thresholds. Using software-specific thresholds, the AI algorithm would have recalled 277 of 303 (91.4%) screen-detected cancers and 47 of 138 (34.1%) interval cancers. AI performance and thresholds should be validated for new clinical settings before deployment, while quality assurance systems should monitor AI performance for consistency. Keywords: Breast, Screening, Mammography, Computer Applications-Detection/Diagnosis, Neoplasms-Primary, Technology Assessment Supplemental material is available for this article. © RSNA, 2023.

Detection of malignancy in whole slide images of endometrial cancer biopsies using artificial intelligence.

In this study we use artificial intelligence (AI) to categorise endometrial biopsy whole slide images (WSI) from digital pathology as either "malignant", "other or benign" or "insufficient". An endometrial biopsy is a key step in diagnosis of endometrial cancer, biopsies are viewed and diagnosed by pathologists. Pathology is increasingly digitised, with slides viewed as images on screens rather than through the lens of a microscope. The availability of these images is driving automation via the application of AI. A model that classifies slides in the manner proposed would allow prioritisation of these slides for pathologist review and hence reduce time to diagnosis for patients with cancer. Previous studies using AI on endometrial biopsies have examined slightly different tasks, for example using images alongside genomic data to differentiate between cancer subtypes. We took 2909 slides with "malignant" and "other or benign" areas annotated by pathologists. A fully supervised convolutional neural network (CNN) model was trained to calculate the probability of a patch from the slide being "malignant" or "other or benign". Heatmaps of all the patches on each slide were then produced to show malignant areas. These heatmaps were used to train a slide classification model to give the final slide categorisation as either "malignant", "other or benign" or "insufficient". The final model was able to accurately classify 90% of all slides correctly and 97% of slides in the malignant class; this accuracy is good enough to allow prioritisation of pathologists' workload.

The Immunopeptidome from a Genomic Perspective: Establishing the Noncanonical Landscape of MHC Class I-Associated Peptides.

Tumor antigens can emerge through multiple mechanisms, including translation of noncoding genomic regions. This noncanonical category of tumor antigens has recently gained attention; however, our understanding of how they recur within and between cancer types is still in its infancy. Therefore, we developed a proteogenomic pipeline based on deep learning de novo mass spectrometry (MS) to enable the discovery of noncanonical MHC class I-associated peptides (ncMAP) from noncoding regions. Considering that the emergence of tumor antigens can also involve posttranslational modifications (PTM), we included an open search component in our pipeline. Leveraging the wealth of MS-based immunopeptidomics, we analyzed data from 26 MHC class I immunopeptidomic studies across 11 different cancer types. We validated the de novo identified ncMAPs, along with the most abundant PTMs, using spectral matching and controlled their FDR to 1%. The noncanonical presentation appeared to be 5 times enriched for the A03 HLA supertype, with a projected population coverage of 55%. The data reveal an atlas of 8,601 ncMAPs with varying levels of cancer selectivity and suggest 17 cancer-selective ncMAPs as attractive therapeutic targets according to a stringent cutoff. In summary, the combination of the open-source pipeline and the atlas of ncMAPs reported herein could facilitate the identification and screening of ncMAPs as targets for T-cell therapies or vaccine development.

The novel anti-cancer fluoropyrimidine NUC-3373 is a potent inhibitor of thymidylate synthase and an effective DNA-damaging agent.

INTRODUCTION: Fluoropyrimidines, principally 5-fluorouracil (5-FU), remain a key component of chemotherapy regimens for multiple cancer types, in particular colorectal and other gastrointestinal malignancies. To overcome key limitations and pharmacologic challenges that hinder the clinical utility of 5-FU, NUC-3373, a phosphoramidate transformation of 5-fluorodeoxyuridine, was designed to improve the efficacy and safety profile as well as the administration challenges associated with 5-FU. METHODS: Human colorectal cancer cell lines HCT116 and SW480 were treated with sub-IC50 doses of NUC-3373 or 5-FU. Intracellular activation was measured by LC-MS. Western blot was performed to determine binding of the active anti-cancer metabolite FdUMP to thymidylate synthase (TS) and DNA damage. RESULTS: We demonstrated that NUC-3373 generates more FdUMP than 5-FU, resulting in a more potent inhibition of TS, DNA misincorporation and subsequent cell cycle arrest and DNA damage in vitro. Unlike 5-FU, the thymineless death induced by NUC-3373 was rescued by the concurrent addition of exogenous thymidine. 5-FU cytotoxicity, however, was only reversed by supplementation with uridine, a treatment used to reduce 5-FU-induced toxicities in the clinic. This is in line with our findings that 5-FU generates FUTP which is incorporated into RNA, a mechanism known to underlie the myelosuppression and gastrointestinal inflammation associated with 5-FU. CONCLUSION: Taken together, these results highlight key differences between NUC-3373 and 5-FU that are driven by the anti-cancer metabolites generated. NUC-3373 is a potent inhibitor of TS that also causes DNA-directed damage. These data support the preliminary clinical evidence that suggest NUC-3373 has a favorable safety profile in patients.

Lymphocyte Classification from Hoechst Stained Slides with Deep Learning.

Multiplex immunofluorescence and immunohistochemistry benefit patients by allowing cancer pathologists to identify proteins expressed on the surface of cells. This enables cell classification, better understanding of the tumour microenvironment, and more accurate diagnoses, prognoses, and tailored immunotherapy based on the immune status of individual patients. However, these techniques are expensive. They are time consuming processes which require complex staining and imaging techniques by expert technicians. Hoechst staining is far cheaper and easier to perform, but is not typically used as it binds to DNA rather than to the proteins targeted by immunofluorescence techniques. In this work we show that through the use of deep learning it is possible to identify an immune cell subtype without immunofluorescence. We train a deep convolutional neural network to identify cells expressing the T lymphocyte marker CD3 from Hoechst 33342 stained tissue only. CD3 expressing cells are often used in key prognostic metrics such as assessment of immune cell infiltration, and by identifying them without the need for costly immunofluorescence, we present a promising new approach to cheaper prediction and improvement of patient outcomes. We also show that by using deep learning interpretability techniques, we can gain insight into the previously unknown morphological features which make this possible.

Use of High-Plex Data Reveals Novel Insights into the Tumour Microenvironment of Clear Cell Renal Cell Carcinoma.

Although immune checkpoint inhibitors (ICIs) have significantly improved the oncological outcomes, about one-third of patients affected by clear cell renal cell carcinoma (ccRCC) still experience recurrence. Current prognostic algorithms, such as the Leibovich score (LS), rely on morphological features manually assessed by pathologists and are therefore subject to bias. Moreover, these tools do not consider the heterogeneous molecular milieu present in the Tumour Microenvironment (TME), which may have prognostic value. We systematically developed a semi-automated method to investigate 62 markers and their combinations in 150 primary ccRCCs using Multiplex Immunofluorescence (mIF), NanoString GeoMx® Digital Spatial Profiling (DSP) and Artificial Intelligence (AI)-assisted image analysis in order to find novel prognostic signatures and investigate their spatial relationship. We found that coexpression of cancer stem cell (CSC) and epithelial-to-mesenchymal transition (EMT) markers such as OCT4 and ZEB1 are indicative of poor outcome. OCT4 and the immune markers CD8, CD34, and CD163 significantly stratified patients at intermediate LS. Furthermore, augmenting the LS with OCT4 and CD34 improved patient stratification by outcome. Our results support the hypothesis that combining molecular markers has prognostic value and can be integrated with morphological features to improve risk stratification and personalised therapy. To conclude, GeoMx® DSP and AI image analysis are complementary tools providing high multiplexing capability required to investigate the TME of ccRCC, while reducing observer bias.

Spatial Analysis of NQO1 in Non-Small Cell Lung Cancer Shows Its Expression Is Independent of NRF1 and NRF2 in the Tumor Microenvironment.

Nuclear factor erythroid 2-related factor 1 (NFE2L1, NRF1) and nuclear factor erythroid 2-related factor 2 (NFE2L2, NRF2) are distinct oxidative stress response transcription factors, both of which have been shown to perform cytoprotective functions, modulating cell stress response and homeostasis. NAD(P)H:quinone oxidoreductase (NQO1) is a mutual downstream antioxidant gene target that catalyzes the two-electron reduction of an array of substrates, protecting against reactive oxygen species (ROS) generation. NQO1 is upregulated in non-small cell lung cancer (NSCLC) and is proposed as a predictive biomarker and therapeutic target. Antioxidant protein expression of immune cells within the NSCLC tumor microenvironment (TME) remains undetermined and may affect immune cell effector functions and survival outcomes. Multiplex immunofluorescence was performed to examine the co-localization of NQO1, NRF1 and NRF2 within the tumor and TME of 162 chemotherapy-naïve, early-stage NSCLC patients treated by primary surgical resection. This study demonstrates that NQO1 protein expression is high in normal, tumor-adjacent tissue and that NQO1 expression varies depending on the cell type. Inter and intra-patient heterogenous NQO1 expression was observed in lung cancer. Co-expression analysis showed NQO1 is independent of NRF1 and NRF2 in tumors. Density-based co-expression analysis demonstrated NRF1 and NRF2 double-positive expression in cancer cells is associated with improved overall survival.

Weakly supervised learning and interpretability for endometrial whole slide image diagnosis.

Fully supervised learning for whole slide image-based diagnostic tasks in histopathology is problematic due to the requirement for costly and time-consuming manual annotation by experts. Weakly supervised learning that utilizes only slide-level labels during training is becoming more widespread as it relieves this burden, but has not yet been applied to endometrial whole slide images, in iSyntax format. In this work, we apply a weakly supervised learning algorithm to a real-world dataset of this type for the first time, with over 85% validation accuracy and over 87% test accuracy. We then employ interpretability methods including attention heatmapping, feature visualization, and a novel end-to-end saliency-mapping approach to identify distinct morphologies learned by the model and build an understanding of its behavior. These interpretability methods, alongside consultation with expert pathologists, allow us to make comparisons between machine-learned knowledge and consensus in the field. This work contributes to the state of the art by demonstrating a robust practical application of weakly supervised learning on a real-world digital pathology dataset and shows the importance of fine-grained interpretability to support understanding and evaluation of model performance in this high-stakes use case.