Hybrid Models for Endoscopy Image Analysis for Early Detection of Gastrointestinal Diseases Based on Fused Features.

The gastrointestinal system contains the upper and lower gastrointestinal tracts. The main tasks of the gastrointestinal system are to break down food and convert it into essential elements that the body can benefit from and expel waste in the form of feces. If any organ is affected, it does not work well, which affects the body. Many gastrointestinal diseases, such as infections, ulcers, and benign and malignant tumors, threaten human life. Endoscopy techniques are the gold standard for detecting infected parts within the organs of the gastrointestinal tract. Endoscopy techniques produce videos that are converted into thousands of frames that show the disease's characteristics in only some frames. Therefore, this represents a challenge for doctors because it is a tedious task that requires time, effort, and experience. Computer-assisted automated diagnostic techniques help achieve effective diagnosis to help doctors identify the disease and give the patient the appropriate treatment. In this study, many efficient methodologies for analyzing endoscopy images for diagnosing gastrointestinal diseases were developed for the Kvasir dataset. The Kvasir dataset was classified by three pre-trained models: GoogLeNet, MobileNet, and DenseNet121. The images were optimized, and the gradient vector flow (GVF) algorithm was applied to segment the regions of interest (ROIs), isolating them from healthy regions and saving the endoscopy images as Kvasir-ROI. The Kvasir-ROI dataset was classified by the three pre-trained GoogLeNet, MobileNet, and DenseNet121 models. Hybrid methodologies (CNN-FFNN and CNN-XGBoost) were developed based on the GVF algorithm and achieved promising results for diagnosing disease based on endoscopy images of gastroenterology. The last methodology is based on fused CNN models and their classification by FFNN and XGBoost networks. The hybrid methodology based on the fused CNN features, called GoogLeNet-MobileNet-DenseNet121-XGBoost, achieved an AUC of 97.54%, accuracy of 97.25%, sensitivity of 96.86%, precision of 97.25%, and specificity of 99.48%.

Hybrid Techniques for the Diagnosis of Acute Lymphoblastic Leukemia Based on Fusion of CNN Features.

Acute lymphoblastic leukemia (ALL) is one of the deadliest forms of leukemia due to the bone marrow producing many white blood cells (WBC). ALL is one of the most common types of cancer in children and adults. Doctors determine the treatment of leukemia according to its stages and its spread in the body. Doctors rely on analyzing blood samples under a microscope. Pathologists face challenges, such as the similarity between infected and normal WBC in the early stages. Manual diagnosis is prone to errors, differences of opinion, and the lack of experienced pathologists compared to the number of patients. Thus, computer-assisted systems play an essential role in assisting pathologists in the early detection of ALL. In this study, systems with high efficiency and high accuracy were developed to analyze the images of C-NMC 2019 and ALL-IDB2 datasets. In all proposed systems, blood micrographs were improved and then fed to the active contour method to extract WBC-only regions for further analysis by three CNN models (DenseNet121, ResNet50, and MobileNet). The first strategy for analyzing ALL images of the two datasets is the hybrid technique of CNN-RF and CNN-XGBoost. DenseNet121, ResNet50, and MobileNet models extract deep feature maps. CNN models produce high features with redundant and non-significant features. So, CNN deep feature maps were fed to the Principal Component Analysis (PCA) method to select highly representative features and sent to RF and XGBoost classifiers for classification due to the high similarity between infected and normal WBC in early stages. Thus, the strategy for analyzing ALL images using serially fused features of CNN models. The deep feature maps of DenseNet121-ResNet50, ResNet50-MobileNet, DenseNet121-MobileNet, and DenseNet121-ResNet50-MobileNet were merged and then classified by RF classifiers and XGBoost. The RF classifier with fused features for DenseNet121-ResNet50-MobileNet reached an AUC of 99.1%, accuracy of 98.8%, sensitivity of 98.45%, precision of 98.7%, and specificity of 98.85% for the C-NMC 2019 dataset. With the ALL-IDB2 dataset, hybrid systems achieved 100% results for AUC, accuracy, sensitivity, precision, and specificity.

Development of a VLP-Based Vaccine Displaying an xCT Extracellular Domain for the Treatment of Metastatic Breast Cancer.

Metastatic breast cancer (MBC) is the leading cause of cancer death in women due to recurrence and resistance to conventional therapies. Thus, MBC represents an important unmet clinical need for new treatments. In this paper we generated a virus-like particle (VLP)-based vaccine (AX09) to inhibit de novo metastasis formation and ultimately prolong the survival of patients with MBC. To this aim, we engineered the bacteriophage MS2 VLP to display an extracellular loop of xCT, a promising therapeutic target involved in tumor progression and metastasis formation. Elevated levels of this protein are observed in a high percentage of invasive mammary ductal tumors including triple negative breast cancer (TNBC) and correlate with poor overall survival. Moreover, xCT expression is restricted to only a few normal cell types. Here, we tested AX09 in several MBC mouse models and showed that it was well-tolerated and elicited a strong antibody response against xCT. This antibody-based response resulted in the inhibition of xCT's function in vitro and reduced metastasis formation in vivo. Thus, AX09 represents a promising novel approach for MBC, and it is currently advancing to clinical development.

Cystic Fibrosis Gene Mutation Frequency Among a Group of Suspected Children in King Hussein Medical Center.

INTRODUCTION: Cystic fibrosis (CF) is a genetic multisystem disorder that affects mostly the lungs, but other organs such as liver, pancreas and intestine also affected. CF is inherited in an autosomal recessive manner and occurs in males and females equally. Cystic fibrosis Transmembrane Conductance Regulator (CFTR) mutations are classified into five classes. Class 1 (non-functional protein), class 2 (near-absence of mature CFTR protein at the apical cell membrane), class 3 (full-length CFTR protein incorporated into the cell membrane), class 4 (reduced conductance CFTR mutation), and class 5 (reduced amount of CFTR protein with normal function). Globally F508 mutation is the most common. AIM: The aim of this study was to determine the frequency of CFTR gene mutation in Jordanian populations attending a major hospital (KHMC). MATERIAL AND METHODS: This is a retrospective study was conducted on 777 sera samples for patients clinically suspected to have cystic fibrosis over a six year period 1/1/2013-1/10/2018. The patient's age range between 1year and 33 years, of which 59.2% (460) were male and 40.8% (317) female. Blood samples were analyzed at Princess Iman Centre for Research and Laboratory Sciences at King Hussein Medical Centre. The samples were tested for 34 mutations of CFTR gene using CF Strip Assay VIENNA LAB Diagnostics GmbH, Austria by polymerase chain reaction (PCR). RESULTS: A total of 777 patients samples were analyzed for cystic gene mutations. Twelve (12) mutations were identified. In 49 patients (6.3%) were heterozygous genotype mutant and 28 (3.6%) were homozygous. The most frequent mutation F508del was found in 32/77 (41.5%). 20 (25.9%) of them were heterozygous genotype mutant and 12 (15.6%) were homozygous genotype mutant. The second frequent mutation was N1303K with frequency rate 15.6% (12/77), 9 (11.7%) of them were heterozygous and 3 (3.9%) were homozygous. Regarding frequency of cystic fibrosis gene mutation depending on sex, 55.8% (43/77) of mutations were found in male, whereas 44.2% (34/77) in female. CONCLUSION: Our findings suggest that cystic fibrosis in Jordan is not a rare disease, and found that the most frequent CFTR gene mutation was F508del, which is in keeping with results from other Mediterranean countries.

HER3 and LINC00052 interplay promotes tumor growth in breast cancer.

Here we report that the lncRNA LINC00052 expression correlates positively with HER3/ErbB3 levels in breast cancer cells. Gene silencing of LINC00052 diminished both LINC00052 and HER3 expression and reduced cancer cell growth in vitro and in vivo. LINC00052 overexpression promoted cancer cell growth in vitro and in vivo and increased HER3-mediated downstream signaling. Importantly, neutralization of HER3 signaling with HER3 targeting monoclonal antibodies blocked LINC00052 mediated cancer cell proliferation in vitro and tumor growth in vivo, suggesting LINC00052 promoting cancer growth through HER3 signaling. Taken together, our results indicate that high LINC00052 levels predict activation of HER3-mediated signaling, and LINC00052 expression level may serve as a potential biomarker for HER3 targeted antibody cancer therapies.

Novel association of DJ-1 with HER3 potentiates HER3 activation and signaling in cancer.

HER3/ErbB3 has emerged as a new therapeutic target for cancer. Currently, more than a dozen anti-HER3 antibodies are in clinical trials for treatment of various cancers. However, limited understanding of the complex HER3 signaling in cancer and lack of established biomarkers have made it challenging to stratify cancer patients who can benefit from HER3 targeted therapies. In this study, we identified DJ-1/PARK7 (Parkinson Protein 7) as a novel interaction partner of HER3 and demonstrated the potential of DJ-1 as a biomarker for anti-HER3 cancer therapy. DJ-1 association with HER3 protects HER3 from ubiquitination and degradation through the proteasomal pathway in breast cancer cells. However, neuregulin 1 (NRG-1) mediated HER3 activation results in a reduced association of DJ-1 with HER3. DJ-1 shRNA knockdown in cancer cells resulted in decreased levels of HER3 and its downstream signaling through the PI3K/AKT and Ras/Raf/ERK pathways. DJ-1 shRNA knockdown cancer cells significantly reduced cell proliferation and migration in vitro and tumor growth in vivo. Conversely, overexpression of DJ-1 increased HER3 levels and promoted cancer cell proliferation in vitro and tumor growth in vivo. Notably, cancer cells with high DJ-1 expression showed more sensitivity than DJ-1 knockdown cells to anti-HER3 antibody inhibition. In addition, there was a significant co-expression of HER3 and DJ-1 in tumor tissues of breast cancer patients. Taken together, these results suggest that high DJ-1 expression in breast cancer cells predicts elevated HER3 signaling and may therefore serve as a biomarker for HER3 targeted antibody cancer therapies.

Fatal Granulomatous Amoebic Encephalitis Caused by Acanthamoeba in a Patient With Kidney Transplant: A Case Report.

Granulomatous amoebic encephalitis (GAE) due to Acanthamoeba is almost a uniformly fatal infection in immune-compromised hosts despite multidrug combination therapy. We report a case of GAE in a female who received a deceased donor kidney graft. She was treated with a combination of miltefosine, pentamidine, sulfadiazine, fluconazole, flucytosine, and azithromycin.

PRUNE2 is a human prostate cancer suppressor regulated by the intronic long noncoding RNA PCA3.

Prostate cancer antigen 3 (PCA3) is the most specific prostate cancer biomarker but its function remains unknown. Here we identify PRUNE2, a target protein-coding gene variant, which harbors the PCA3 locus, thereby classifying PCA3 as an antisense intronic long noncoding (lnc)RNA. We show that PCA3 controls PRUNE2 levels via a unique regulatory mechanism involving formation of a PRUNE2/PCA3 double-stranded RNA that undergoes adenosine deaminase acting on RNA (ADAR)-dependent adenosine-to-inosine RNA editing. PRUNE2 expression or silencing in prostate cancer cells decreased and increased cell proliferation, respectively. Moreover, PRUNE2 and PCA3 elicited opposite effects on tumor growth in immunodeficient tumor-bearing mice. Coregulation and RNA editing of PRUNE2 and PCA3 were confirmed in human prostate cancer specimens, supporting the medical relevance of our findings. These results establish PCA3 as a dominant-negative oncogene and PRUNE2 as an unrecognized tumor suppressor gene in human prostate cancer, and their regulatory axis represents a unique molecular target for diagnostic and therapeutic intervention.

Ligand-directed targeting of lymphatic vessels uncovers mechanistic insights in melanoma metastasis.

Metastasis is the most lethal step of cancer progression in patients with invasive melanoma. In most human cancers, including melanoma, tumor dissemination through the lymphatic vasculature provides a major route for tumor metastasis. Unfortunately, molecular mechanisms that facilitate interactions between melanoma cells and lymphatic vessels are unknown. Here, we developed an unbiased approach based on molecular mimicry to identify specific receptors that mediate lymphatic endothelial-melanoma cell interactions and metastasis. By screening combinatorial peptide libraries directly on afferent lymphatic vessels resected from melanoma patients during sentinel lymphatic mapping and lymph node biopsies, we identified a significant cohort of melanoma and lymphatic surface binding peptide sequences. The screening approach was designed so that lymphatic endothelium binding peptides mimic cell surface proteins on tumor cells. Therefore, relevant metastasis and lymphatic markers were biochemically identified, and a comprehensive molecular profile of the lymphatic endothelium during melanoma metastasis was generated. Our results identified expression of the phosphatase 2 regulatory subunit A, α-isoform (PPP2R1A) on the cell surfaces of both melanoma cells and lymphatic endothelial cells. Validation experiments showed that PPP2R1A is expressed on the cell surfaces of both melanoma and lymphatic endothelial cells in vitro as well as independent melanoma patient samples. More importantly, PPP2R1A-PPP2R1A homodimers occur at the cellular level to mediate cell-cell interactions at the lymphatic-tumor interface. Our results revealed that PPP2R1A is a new biomarker for melanoma metastasis and show, for the first time to our knowledge, an active interaction between the lymphatic vasculature and melanoma cells during tumor progression.

Reduced adenosine-to-inosine miR-455-5p editing promotes melanoma growth and metastasis.

Although recent studies have shown that adenosine-to-inosine (A-to-I) RNA editing occurs in microRNAs (miRNAs), its effects on tumour growth and metastasis are not well understood. We present evidence of CREB-mediated low expression of ADAR1 in metastatic melanoma cell lines and tumour specimens. Re-expression of ADAR1 resulted in the suppression of melanoma growth and metastasis in vivo. Consequently, we identified three miRNAs undergoing A-to-I editing in the weakly metastatic melanoma but not in strongly metastatic cell lines. One of these miRNAs, miR-455-5p, has two A-to-I RNA-editing sites. The biological function of edited miR-455-5p is different from that of the unedited form, as it recognizes a different set of genes. Indeed, wild-type miR-455-5p promotes melanoma metastasis through inhibition of the tumour suppressor gene CPEB1. Moreover, wild-type miR-455 enhances melanoma growth and metastasis in vivo, whereas the edited form inhibits these features. These results demonstrate a previously unrecognized role for RNA editing in melanoma progression.

Receptor tyrosine kinase EphA5 is a functional molecular target in human lung cancer.

Lung cancer is often refractory to radiotherapy, but molecular mechanisms of tumor resistance remain poorly defined. Here we show that the receptor tyrosine kinase EphA5 is specifically overexpressed in lung cancer and is involved in regulating cellular responses to genotoxic insult. In the absence of EphA5, lung cancer cells displayed a defective G1/S cell cycle checkpoint, were unable to resolve DNA damage, and became radiosensitive. Upon irradiation, EphA5 was transported into the nucleus where it interacted with activated ATM (ataxia-telangiectasia mutated) at sites of DNA repair. Finally, we demonstrate that a new monoclonal antibody against human EphA5 sensitized lung cancer cells and human lung cancer xenografts to radiotherapy and significantly prolonged survival, thus suggesting the likelihood of translational applications.

Cooperative effects of aminopeptidase N (CD13) expressed by nonmalignant and cancer cells within the tumor microenvironment.

Processes that promote cancer progression such as angiogenesis require a functional interplay between malignant and nonmalignant cells in the tumor microenvironment. The metalloprotease aminopeptidase N (APN; CD13) is often overexpressed in tumor cells and has been implicated in angiogenesis and cancer progression. Our previous studies of APN-null mice revealed impaired neoangiogenesis in model systems without cancer cells and suggested the hypothesis that APN expressed by nonmalignant cells might promote tumor growth. We tested this hypothesis by comparing the effects of APN deficiency in allografted malignant (tumor) and nonmalignant (host) cells on tumor growth and metastasis in APN-null mice. In two independent tumor graft models, APN activity in both the tumors and the host cells cooperate to promote tumor vascularization and growth. Loss of APN expression by the host and/or the malignant cells also impaired lung metastasis in experimental mouse models. Thus, cooperation in APN expression by both cancer cells and nonmalignant stromal cells within the tumor microenvironment promotes angiogenesis, tumor growth, and metastasis.

c-ABL modulates MAP kinases activation downstream of VEGFR-2 signaling by direct phosphorylation of the adaptor proteins GRB2 and NCK1.

Vascular Endothelial Growth Factor-A (VEGF-A) is a key molecule in normal and tumor angiogenesis. This study addresses the role of c-ABL as a novel downstream target of VEGF-A in primary Human Umbilical Vein Endothelial Cells (HUVEC). On the basis of immunoprecipitation experiments, in vitro kinase assay and RNA interference, we demonstrate that VEGF-A induces the c-ABL kinase activity through the VEGF Receptor-2/Phosphatidylinositol-3-Kinase pathway. By treating HUVEC with the specific tyrosine kinase inhibitor STI571 and over-expressing a dominant negative c-ABL mutant, we show that the VEGF-A-activated c-ABL reduces the amplitude of Mitogen-Activated Protein Kinases (ERK1/2, JNKs and p38) activation in a dose-dependent manner by a negative feedback mechanism. By analysis of the adaptor proteins NCK1 and GRB2 mutants we further show that the negative loop on p38 is mediated by c-ABL phosphorylation at tyrosine 105 of the adaptor protein NCK1, while the phosphorylation at tyrosine 209 of GRB2 down-modulates ERK1/2 and JNKs signaling. These findings suggest that c-ABL function is to establish a correct and tightly controlled response of endothelial cells to VEGF-A during the angiogenic process.

Endothelial cell adhesion to the extracellular matrix induces c-Src-dependent VEGFR-3 phosphorylation without the activation of the receptor intrinsic kinase activity.

RATIONALE: Integrins cooperate with growth factor receptors to promote downstream signaling for cell proliferation and migration. However, the mechanism of receptor activation is still unknown. OBJECTIVE: To analyze the mechanism of phosphorylation of the vascular endothelial growth factor receptor (VEGFR)-3 by cell adhesion. METHODS AND RESULTS: We show that VEGFR-3 phosphorylation, induced by cell attachment to the extracellular matrix, is independent from the intrinsic kinase activity of the receptor, as evidenced from phosphorylation cell adhesion experiments with a mutant kinase dead receptor or in the presence of the specific kinase inhibitor MAZ 51. Cell adhesion experiments in the presence of the c-Src inhibitor PP2 or in fibroblast triple knockout for c-Src, Yes, and Fyn (SYF) demonstrate that VEGFR-3 phosphorylation, induced by extracellular matrix, is mediated by c-Src. Kinase assays in vitro with recombinant c-Src show that VEGFR-3 is a direct c-Src target and mass spectrometry analysis identified the sites phosphorylated by c-Src as tyrosine 830, 833, 853, 1063, 1333, and 1337, demonstrating that integrin-mediated receptor phosphorylation induces a phosphorylation pattern that is distinct from that induced by growth factors. Furthermore, pull-down assays show that integrin-mediated VEGFR-3 phosphorylation activates the recruitment to the receptor of the adaptor proteins CRKI/II and SHC inducing activation of JNK. CONCLUSIONS: These data suggest that cell adhesion to extracellular matrix induces a downstream signaling using the tyrosine kinase receptor VEGFR-3 as scaffold.

From combinatorial peptide selection to drug prototype (I): targeting the vascular endothelial growth factor receptor pathway.

Inhibition of blood vessel formation is a viable therapeutic approach in angiogenesis-dependent diseases. We previously used a combinatorial screening on vascular endothelial growth factor (VEGF)-activated endothelial cells to select the sequence CPQPRPLC and showed that the motif Arg-Pro-Leu targets VEGF receptor-1 and neuropilin-1. Here, we evaluated and validated (D)(LPR), a derivative molecule with strong antiangiogenesis attributes. This prototype drug markedly inhibits neovascularization in three mouse models: Matrigel-based assay, functional human/murine blood vessel formation, and retinopathy of prematurity. In addition to its systemic activity, (D)(LPR) also inhibits retinal angiogenesis when administered in an eye-drop formulation. Finally, in preliminary studies, we have showed targeted drug activity in an experimental tumor-bearing mouse model. These results show that drugs targeting extracellular domains of VEGF receptors are active, affect signal transduction, and have potential for clinical application. On a larger context, this study illustrates the power of ligand-directed selection plus retro-inversion for rapid drug discovery and development.

An unrecognized extracellular function for an intracellular adapter protein released from the cytoplasm into the tumor microenvironment.

Mammalian cell membranes provide an interface between the intracellular and extracellular compartments. It is currently thought that cytoplasmic signaling adapter proteins play no functional role within the extracellular tumor environment. Here, by selecting combinatorial random peptide libraries in tumor-bearing mice, we uncovered a direct, specific, and functional interaction between CRKL, an adapter protein [with Src homology 2 (SH2)- and SH3-containing domains], and the plexin-semaphorin-integrin domain of beta(1) integrin in the extracellular milieu. Through assays in vitro, in cellulo, and in vivo, we show that this unconventional and as yet unrecognized protein-protein interaction between a regulatory integrin domain (rather than a ligand-binding one) and an intracellular adapter (acting outside of the cells) triggers an alternative integrin-mediated cascade for cell growth and survival. Based on these data, here we propose that a secreted form of the SH3/SH2 adaptor protein CRKL may act as a growth-promoting factor driving tumorigenesis and may lead to the development of cancer therapeutics targeting secreted CRKL.

Vascular endothelial growth factor receptor-3 activity is modulated by its association with caveolin-1 on endothelial membrane.

Vascular endothelial growth factor receptor-3 (VEGFR-3) is constitutively expressed in lymphatic vessels and transiently in endothelial cells of blood vessels during angiogenesis. Here we report that VEGFR-3 localizes in the caveolae membrane of endothelial cells and co-immunoprecipitates with caveolin-1. Caveolin-1 silencing or its depletion from the cell membrane by cholesterol increases VEGFR-3 autophosphorylation, suggesting that caveolin acts as a negative regulator of VEGFR-3 activity. Receptor activation induces caveolin-1 phosphorylation on tyrosine residues including tyrosine 14. Cell treatment with Src or Abl inhibitors PP2 or STI571, prior to receptor stimulation, affects caveolin-1 phosphorylation without affecting receptor autophosphorylation, suggesting that both Src and Abl are involved in VEGFR-3-dependent caveolin-1 phosphorylation. Caveolin-1 phosphorylation in Src/Fyn/Yes knockout cells demonstrated that Abl phosphorylates caveolin-1 independently from Src family members. These results suggest a functional interaction between VEGFR-3 and caveolin-1 to modulate endothelial cell activation during angiogenesis.

Direct recruitment of CRK and GRB2 to VEGFR-3 induces proliferation, migration, and survival of endothelial cells through the activation of ERK, AKT, and JNK pathways.

Vascular endothelial growth factor receptor-3 (VEGFR-3) plays a key role for the remodeling of the primary capillary plexus in the embryo and contributes to angiogenesis and lymphangiogenesis in the adult. However, VEGFR-3 signal transduction pathways remain to be elucidated. Here we investigated VEGFR-3 signaling in primary human umbilical vein endothelial cells (HUVECs) by the systematic mutation of the tyrosine residues potentially involved in VEGFR-3 signaling and identified the tyrosines critical for its function. Y1068 was shown to be essential for the kinase activity of the receptor. Y1063 signals the receptor-mediated survival by recruiting CRKI/II to the activated receptor, inducing a signaling cascade that, via mitogen-activated protein kinase kinase-4 (MKK4), activates c-Jun N-terminal kinase-1/2 (JNK1/2). Inhibition of JNK1/2 function either by specific peptide inhibitor JNKI1 or by RNA interference (RNAi) demonstrated that activation of JNK1/2 is required for a VEGFR-3-dependent prosurvival signaling. Y1230/Y1231 contributes, together with Y1337, to proliferation, migration, and survival of endothelial cells. Phospho-Y1230/Y1231 directly recruits growth factor receptor-bonus protein (GRB2) to the receptor, inducing the activation of both AKT and extracellular signal-related kinase 1/2 (ERK1/2) signaling. Finally, we observed that Y1063 and Y1230/Y1231 signaling converge to induce c-JUN expression, and RNAi experiments demonstrated that c-JUN is required for growth factor-induced prosurvival signaling in primary endothelial cells.