Development and validation of a deep learning-based microsatellite instability predictor from prostate cancer whole-slide images.

Microsatellite instability-high (MSI-H) is a tumor-agnostic biomarker for immune checkpoint inhibitor therapy. However, MSI status is not routinely tested in prostate cancer, in part due to low prevalence and assay cost. As such, prediction of MSI status from hematoxylin and eosin (H&E) stained whole-slide images (WSIs) could identify prostate cancer patients most likely to benefit from confirmatory testing to evaluate their eligibility for immunotherapy and need for Lynch syndrome testing. Prostate biopsies and surgical resections from prostate cancer patients referred to our institution were analyzed. MSI status was determined by next-generation sequencing. Patients sequenced before a cutoff date formed an algorithm development set (n = 4015, MSI-H 1.8%) and a paired validation set (n = 173, MSI-H 19.7%) that consisted of two serial sections from each sample, one stained and scanned internally and the other at an external site. Patients sequenced after the cutoff date formed a temporally independent validation set (n = 1350, MSI-H 2.3%). Attention-based multiple instance learning models were trained to predict MSI-H from H&E WSIs. The predictor achieved area under the receiver operating characteristic curve values of 0.78 (95% CI [0.69-0.86]), 0.72 (95% CI [0.63-0.81]), and 0.72 (95% CI [0.62-0.82]) on the internally prepared, externally prepared, and temporal validation sets, respectively, showing effective predictability and generalization to both external staining/scanning processes and temporally independent samples. While MSI-H status is significantly correlated with Gleason score, the model remained predictive within each Gleason score subgroup.

Artificial intelligence-augmented histopathologic review using image analysis to optimize DNA yield from formalin-fixed paraffin-embedded slides.

To achieve minimum DNA input requirements for next-generation sequencing (NGS), pathologists visually estimate macrodissection and slide count decisions. Unfortunately, misestimation may cause tissue waste and increased laboratory costs. We developed an artificial intelligence (AI)-augmented smart pathology review system (SmartPath) to empower pathologists with quantitative metrics for accurately determining tissue extraction parameters. SmartPath uses two deep learning architectures, a U-Net based network for cell segmentation and a multi-field-of-view convolutional network for tumor area segmentation, to extract features from digitized H&E-stained formalin-fixed paraffin-embedded slides. From the segmented tumor area, SmartPath suggests a macrodissection area. To predict DNA yield per slide, the extracted features from within the macrodissection area are correlated with known DNA yields to fit a regularized linear model (R = 0.85). Then, a pathologist-defined target yield divided by the predicted DNA yield per slide gives the number of slides to scrape. Following model development, an internal validation trial was conducted within the Tempus Labs molecular sequencing laboratory. We evaluated our system on 501 clinical colorectal cancer slides, where half received SmartPath-augmented review and half traditional pathologist review. The SmartPath cohort had 25% more DNA yields within a desired target range of 100-2000 ng. The number of extraction attempts was statistically unchanged between cohorts. The SmartPath system recommended fewer slides to scrape for large tissue sections, saving tissue in these cases. Conversely, SmartPath recommended more slides to scrape for samples with scant tissue sections, especially those with degraded DNA, helping prevent costly re-extraction due to insufficient extraction yield. A statistical analysis was performed to measure the impact of covariates on the results, offering insights on how to improve future applications of SmartPath. With these improvements, AI-augmented histopathologic review has the potential to decrease tissue waste, sequencing time, and laboratory costs by optimizing DNA yields, especially for samples with scant tissue and/or degraded DNA.

Identification of a druggable protein-protein interaction site between mutant p53 and its stabilizing chaperone DNAJA1.

The TP53 gene is the most frequently mutated gene in human cancers, and the majority of TP53 mutations are missense mutations. As a result, these mutant p53 (mutp53) either directly lose wildtype p53 (wtp53) tumor suppressor function or exhibit a dominant negative effect over wtp53. In addition, some mutp53 have acquired new oncogenic function (gain of function). Therefore, targeting mutp53 for its degradation may serve as a promising strategy for cancer prevention and therapy. Based on our previous finding that farnesylated DNAJA1 is a crucial chaperone in maintaining mutp53 stabilization, and by using an in silico approach, we built 3D homology models of human DNAJA1 and mutp53R175H proteins, identified the interacting pocket in the DNAJA1-mutp53R175H complex, and found one critical druggable small molecule binding site in the DNAJA1 glycine/phenylalanine-rich region. We confirmed that the interacting pocket in the DNAJA1-mutp53R175H complex was crucial for stabilizing mutp53R175H using a site-directed mutagenesis approach. We further screened a drug-like library to identify a promising small molecule hit (GY1-22) against the interacting pocket in the DNAJA1-mutp53R175H complex. The GY1-22 compound displayed an effective activity against the DNAJA1-mutp53R175H complex. Treatment with GY1-22 significantly reduced mutp53 protein levels, enhanced Waf1p21 expression, suppressed cyclin D1 expression, and inhibited mutp53-driven pancreatic cancer growth both in vitro and in vivo. Together, our results indicate that the interacting pocket in the DNAJA1-mutp53R175H complex is critical for mutp53's stability and oncogenic function, and DNAJA1 is a robust therapeutic target for developing the efficient small molecule inhibitors against oncogenic mutp53.

Ontogenesis and Modulation of Intestinal Unesterified Cholesterol Sequestration in a Mouse Model of Niemann-Pick C1 Disease.

BACKGROUND: Mutations in the NPC1 gene result in sequestration of unesterified cholesterol (UC) and glycosphingolipids in most tissues leading to multi-organ disease, especially in the brain, liver, lungs, and spleen. Various data from NPC1-deficient mice suggest the small intestine (SI) is comparatively less affected, even in late stage disease. METHODS: Using the Npc1nih mouse model, we measured SI weights and total cholesterol (TC) levels in Npc1-/- versus Npc1+/+ mice as a function of age, and then after prolonged ezetimibe-induced inhibition of cholesterol absorption. Next, we determined intestinal levels of UC and esterified cholesterol (EC), and cholesterol synthesis rates in Npc1-/- and Npc1+/+ mice, with and without the cholesterol-esterifying enzyme SOAT2, following a once-only subcutaneous injection with 2-hydroxypropyl-ß-cyclodextrin (2HPßCD). RESULTS: By ~ 42 days of age, intestinal TC levels averaged ~ 2.1-fold more (mostly UC) in the Npc1-/- versus Npc1+/+ mice with no further increase thereafter. Chronic ezetimibe treatment lowered intestinal TC levels in the Npc1-/- mice by only ~ 16%. In Npc1-/- mice given 2HPßCD 24 h earlier, UC levels fell, EC levels increased (although less so in mice lacking SOAT2), and cholesterol synthesis was suppressed equally in the Npc1-/-:Soat2+/+ and Npc1-/-:Soat2-/- mice. CONCLUSIONS: The low and static levels of intestinal UC sequestration in Npc1-/- mice likely reflect the continual sloughing of cells from the mucosa. This sequestration is blunted by about the same extent following a single acute treatment with 2HPßCD as it is by a prolonged ezetimibe-induced block of cholesterol absorption.

Inhibition of mutant Kras and p53-driven pancreatic carcinogenesis by atorvastatin: Mainly via targeting of the farnesylated DNAJA1 in chaperoning mutant p53.

Recent studies have indicated that using statins to inhibit the mevalonate pathway induces mutant p53 degradation by impairing the interaction of mutant p53 with DnaJ subfamily A member 1 (DNAJA1). However, the role of the C-terminus of DNAJA1 with a CAAX box for farnesylation in the binding, folding, and translocation of client proteins such as mutant p53 is not known. In the present study, we used a genetically engineered mouse model of pancreatic carcinoma and showed that atorvastatin significantly increased animal survival and inhibited pancreatic carcinogenesis. There was a dramatic decrease in mutant p53 protein accumulation in the pancreatic acini, pancreas intraepithelial neoplasia lesions, and adenocarcinoma. Supplementation with farnesyl pyrophosphate, a substrate for protein farnesylation, rescued atorvastatin-induced mutant p53 degradation in pancreatic cancer cells. Tipifarnib, a farnesyltransferase inhibitor, mirrored atorvastatin's effects on mutant p53, degraded mutant p53 in a dose-dependent manner, and converted farnesylated DNAJA1 into unfarnesylated DNAJA1. Farnesyltransferase gene knockdown also significantly promoted mutant p53 degradation. Coimmunoprecipitation either by an anti-DNAJA1 or p53 antibody confirmed the direct interaction of mutant p53 and DNAJA1 and higher doses of atorvastatin treatments converted more farnesylated DNAJA1 into unfarnesylated DNAJA1 with much less mutant p53 pulled down by DNAJA1. Strikingly, C394S mutant DNAJA1, in which the cysteine of the CAAX box was mutated to serine, was no longer able to be farnesylated and lost the ability to maintain mutant p53 stabilization. Our results show that farnesylated DNAJA1 is a crucial chaperone in maintaining mutant p53 stabilization and targeting farnesylated DNAJA1 by atorvastatin will be critical for inhibiting p53 mutant cancer.

Epoxy-Oxylipins and Soluble Epoxide Hydrolase Metabolic Pathway as Targets for NSAID-Induced Gastroenteropathy and Inflammation-Associated Carcinogenesis.

Polyunsaturated fatty acids (PUFAs) including epoxide-modified ω-3 and ω-6 fatty acids are made via oxidation to create highly polarized carbon-oxygen bonds crucial to their function as signaling molecules. A critical PUFA, arachidonic acid (ARA), is metabolized to a diverse set of lipids signaling molecules through cyclooxygenase (COX), lipoxygenase (LOX), cytochrome P450 epoxygenase, or cytochrome P450 hydroxylase; however, the majority of ARA is metabolized into anti-inflammatory epoxides via cytochrome P450 enzymes. These short-lived epoxide lipids are rapidly metabolized or inactivated by the soluble epoxide hydrolase (sEH) into diol-containing products. sEH inhibition or knockout has been a practical approach to study the biology of the epoxide lipids, and has been shown to effectively treat inflammatory conditions in the preclinical models including gastrointestinal ulcers and colitis by shifting oxylipins to epoxide profiles, inhibiting inflammatory cell infiltration and activation, and enhancing epithelial cell defense via increased mucin production, thus providing further evidence for the role of sEH as a pro-inflammatory protein. Non-steroidal anti-inflammatory drugs (NSAIDs) with COX-inhibitor activity are among the most commonly used analgesics and have demonstrated applications in the management of cardiovascular disease and intriguingly cancer. Major side effects of NSAIDs however are gastrointestinal ulcers which frequently precludes their long-term application. In this review, we hope to bridge the gap between NSAID toxicity and sEH-mediated metabolic pathways to focus on the role of epoxy fatty acid metabolic pathway of PUFAs in NSAIDS-ulcer formation and healing as well as inflammation-related carcinogenesis. Specifically we address the potential application of sEH inhibition to enhance ulcer healing at the site of inflammation via their activity on altered lipid signaling, mitochondrial function, and diminished reactive oxygen species, and further discuss the significance of dual COX and sEH inhibitor in anti-inflammation and carcinogenesis.

Inhibition of Pancreatic Carcinoma Growth Through Enhancing ω-3 Epoxy Polyunsaturated Fatty Acid Profile by Inhibition of Soluble Epoxide Hydrolase.

BACKGROUND/AIM: Cytochrome P450 epoxygenase is a major enzyme involved in the metabolism of ω-3 polyunsaturated fatty acids (PUFAs) to produce biologically active ω-3 epoxy fatty acids (ω-3 epoxides). In general, all epoxy PUFAs including ω-3 epoxides are quickly metabolized/inactivated by soluble epoxide hydrolase (sEH) to form diol products. The aims of this study were to determine the effect and mechanism of fat-1 transgene, and ω-3 PUFA combined with sEH gene knockout or inhibitor on inhibiting pancreatic cancer and the related mechanisms involved. MATERIALS AND METHODS: PK03-mutant KrasG12D murine pancreatic carcinoma cells were inoculated into mouse models including fat-1, sEH-/- and C57BL/6J mice. The mice were fed with AIN-76A diet with or without ω-3 PUFA supplementation or treated with sEH inhibitor. In addition to tumor growth (tumor size and weight), cell proliferation, mutant Kras-mediated signaling, inflammatory reaction and angiogenesis were analyzed immunohisto-chemically and by western blot assay. ω-3 PUFA metabolism, particularly focusing on ω-3 epoxy fatty acids (ω-3 epoxides), was measured using a liquid chromatography with tandem mass spectrometry (LC-MS/MS) approach. RESULTS: Significant decreases of weight and size of the PK03 pancreatic carcinoma were observed in the fat-1 transgenic mice treated with sEH inhibitor compared to those of C57BL/6J control mice fed with AIN-76A diet (weight: 0.28±0.04 g vs. 0.58±0.06 g; size: 187.0±17.5 mm3 vs. 519.3±60.6 mm3). In a separate experiment, sEH-/- mice fed ω-3 PUFA supplement and C57BL/6J mice treated with sEH inhibitor and fed ω-3 PUFA supplement exhibited a significant reduction in the weight and size of the pancreatic carcinoma compared to C57BL/6J control mice (weight: 0.26±.26 g and 0.39±.39 g vs. 0.69±0.11 g, respectively; size: 274.2±36.2 mm3 and 296.4±99.8 mm3 vs. 612.6±117.8 mm3, respectively). Moreover, compared to the pancreatic tumors in C57BL/6J control mice, the tumors in fat-1 transgenic mice treated with sEH inhibitor showed a significant less inflammatory cell infiltrate (62.6±9.2/HPF (high power field) vs. 8.0±1.2/HPF), tumor cell proliferation (48.5±1.7% vs. 16.5±1.6%), and angiogenesis (micro-vessel density (MVD): 35.0±1.0 vs. 11.1±0.5) immunohistochemically, as well as significantly increased caspase-3 labeled apoptosis (0.44±0.06% vs. 0.69±0.06%, respectively). Using western blot approach, significant inhibition of mutant Kras-activated signals including phosphorylated Serine/threonine kinases (cRAF), Mitogen-activated protein kinase kinase (MEK), and extracellular signal-regulated kinase (ERK) were identified in pancreatic carcinoma of fat-1 transgenic mice treated with sEH inhibitor. Eicosanoic acid metabolic profiling of the serum specimens detected a significant increase of the ratios of epoxides to dihydroxy fatty acid (DiHDPE) for docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), and epoxides/dihydroxy octadecenoic acid (DiHOME) for arachidonic acid (ARA) and linoleic acid (LA), as well as a significant increase of epoxy metabolites of DHA, EPA, ARA and LA in fat-1 transgenic mice treated with a sEH inhibitor. CONCLUSION: ω-3 epoxy products from ω-3 PUFA metabolism play a crucial role in inhibiting pancreatic cancer growth, and use of ω-3 PUFAs combined with sEH inhibition is a strategy with high potential for pancreatic cancer treatment and prevention.

Niemann-Pick C1-deficient mice lacking sterol O-acyltransferase 2 have less hepatic cholesterol entrapment and improved liver function.

Cholesteryl esters are generated at multiple sites in the body by sterol O-acyltransferase (SOAT) 1 or SOAT2 in various cell types and lecithin cholesterol acyltransferase in plasma. Esterified cholesterol and triacylglycerol contained in lipoproteins cleared from the circulation via receptor-mediated or bulk-phase endocytosis are hydrolyzed by lysosomal acid lipase within the late endosomal/lysosomal (E/L) compartment. Then, through the successive actions of Niemann-Pick C (NPC) 2 and NPC 1, unesterified cholesterol (UC) is exported from the E/L compartment to the cytosol. Mutations in either NPC1 or NPC2 lead to continuing entrapment of UC in all organs, resulting in multisystem disease, which includes hepatic dysfunction and in some cases liver failure. These studies investigated primarily whether elimination of SOAT2 in NPC1-deficient mice impacted hepatic UC sequestration, inflammation, and transaminase activities. Measurements were made in 7-wk-old mice fed a low-cholesterol chow diet or one enriched with cholesterol starting 2 wk before study. In the chow-fed mice, NPC1:SOAT2 double knockouts, compared with their littermates lacking only NPC1, had 20% less liver mass, 28% lower hepatic UC concentrations, and plasma alanine aminotransferase and aspartate aminotransferase activities that were decreased by 48% and 36%, respectively. mRNA expression levels for several markers of inflammation were all significantly lower in the NPC1 mutants lacking SOAT2. The existence of a new class of potent and selective SOAT2 inhibitors provides an opportunity for exploring if suppression of this enzyme could potentially become an adjunctive therapy for liver disease in NPC1 deficiency. NEW & NOTEWORTHY In Niemann-Pick type C1 (NPC1) disease, the entrapment of unesterified cholesterol (UC) in the endosomal/lysosomal compartment of all cells causes multiorgan disease, including neurodegeneration, pulmonary dysfunction, and liver failure. Some of this sequestered UC entered cells initially in the esterified form. When sterol O-acyltransferase 2, a cholesterol esterifying enzyme present in enterocytes and hepatocytes, is eliminated in NPC1-deficient mice, there is a reduction in their hepatomegaly, hepatic UC content, and cellular injury.

Impact of physiological levels of chenodeoxycholic acid supplementation on intestinal and hepatic bile acid and cholesterol metabolism in Cyp7a1-deficient mice.

Mice deficient in cholesterol 7α-hydroxylase (Cyp7a1) have a diminished bile acid pool (BAP) and therefore represent a useful model for investigating the metabolic effects of restoring the pool with a specific BA. Previously we carried out such studies in Cyp7a1(-/-) mice fed physiological levels of cholic acid (CA) and achieved BAP restoration, along with an increased CA enrichment, at a dietary level of just 0.03% (w/w). Here we demonstrate that in Cyp7a1(-/-) mice fed chenodeoxycholic acid (CDCA) at a level of 0.06% (w/w), the BAP was restored to normal size and became substantially enriched with muricholic acid (MCA) (>70%), leaving the combined contribution of CA and CDCA to be <15%. This resulted in a partial to complete reversal of the main changes in cholesterol and BA metabolism associated with Cyp7a1 deficiency such as an elevated rate of intestinal sterol synthesis, an enhanced level of mRNA for Cyp8b1 in the liver, and depressed mRNA levels for Ibabp, Shp and Fgf15 in the distal small intestine. When Cyp7a1(-/-) and matching Cyp7a1(+/+) mice were fed a diet with added cholesterol (0.2%) (w/w), either alone, or also containing CDCA (0.06%) (w/w) or CA (0.03%) (w/w) for 18days, the hepatic total cholesterol concentrations (mg/g) in the Cyp7a1(-/-) mice were 26.9±3.7, 16.4±0.9 and 47.6±1.9, respectively, vs. 4.9±0.4, 5.0±0.7 and 6.4±1.9, respectively in the corresponding Cyp7a1(+/+) controls. These data affirm the importance of using moderate levels of dietary BA supplementation to elicit changes in hepatic cholesterol metabolism through shifts in BAP size and composition.

Carboxylesterases are uniquely expressed among tissues and regulated by nuclear hormone receptors in the mouse.

Carboxylesterases (CES) are a well recognized, yet incompletely characterized family of proteins that catalyze neutral lipid hydrolysis. Some CES have well-defined roles in xenobiotic clearance, pharmacologic prodrug activation, and narcotic detoxification. In addition, emerging evidence suggests other CES may have roles in lipid metabolism. Humans have six CES genes, whereas mice have 20 Ces genes grouped into five isoenzyme classes. Perhaps due to the high sequence similarity shared by the mouse Ces genes, the tissue-specific distribution of expression for these enzymes has not been fully addressed. Therefore, we performed studies to provide a comprehensive tissue distribution analysis of mouse Ces mRNAs. These data demonstrated that while the mouse Ces family 1 is highly expressed in liver and family 2 in intestine, many Ces genes have a wide and unique tissue distribution defined by relative mRNA levels. Furthermore, evaluating Ces gene expression in response to pharmacologic activation of lipid- and xenobiotic-sensing nuclear hormone receptors showed differential regulation. Finally, specific shifts in Ces gene expression were seen in peritoneal macrophages following lipopolysaccharide treatment and in a steatotic liver model induced by high-fat feeding, two model systems relevant to disease. Overall these data show that each mouse Ces gene has its own distinctive tissue expression pattern and suggest that some CES may have tissue-specific roles in lipid metabolism and xenobiotic clearance.

Delineation of biochemical, molecular, and physiological changes accompanying bile acid pool size restoration in Cyp7a1(-/-) mice fed low levels of cholic acid.

Cholesterol 7α-hydroxylase (CYP7A1) is the initiating and rate-limiting enzyme in the neutral pathway that converts cholesterol to primary bile acids (BA). CYP7A1-deficient (Cyp7a1(-/-)) mice have a depleted BA pool, diminished intestinal cholesterol absorption, accelerated fecal sterol loss, and increased intestinal cholesterol synthesis. To determine the molecular and physiological effects of restoring the BA pool in this model, adult female Cyp7a1(-/-) mice and matching Cyp7a1(+/+) controls were fed diets containing cholic acid (CA) at modest levels [0.015, 0.030, and 0.060% (wt/wt)] for 15-18 days. A level of just 0.03% provided a CA intake of ~12 µmol (4.8 mg) per day per 100 g body wt and was sufficient in the Cyp7a1(-/-) mice to normalize BA pool size, fecal BA excretion, fractional cholesterol absorption, and fecal sterol excretion but caused a significant rise in the cholesterol concentration in the small intestine and liver, as well as a marked inhibition of cholesterol synthesis in these organs. In parallel with these metabolic changes, there were marked shifts in intestinal and hepatic expression levels for many target genes of the BA sensor farnesoid X receptor, as well as genes involved in cholesterol transport, especially ATP-binding cassette (ABC) transporter A1 (ABCA1) and ABCG8. In Cyp7a1(+/+) mice, this level of CA supplementation did not significantly disrupt BA or cholesterol metabolism, except for an increase in fecal BA excretion and marginal changes in mRNA expression for some BA synthetic enzymes. These findings underscore the importance of using moderate dietary BA levels in studies with animal models.