Impact of renal replacement therapy modality on coagulation and platelet function in critically ill patients: A prospective observational study.

BACKGROUND: Renal replacement therapy (RRT) may affect coagulation and platelet function in critically ill patients. However, the mechanism and the difference in the impact on coagulation between intermittent hemodialysis (iHD) and continuous renal replacement therapy (CRRT) remains unclear. This study aimed to investigate and compare the impact of iHD and CRRT on coagulation and platelet function. METHODS: Critically ill patients undergoing RRT were classified into the iHD group or the CRRT group. After the first blood sampling, patients underwent either a single session of hemodialysis or 48 h of CRRT, then a second blood sample was taken. Rotational thromboelastometry (ROTEM), platelet aggregometry and conventional coagulation tests were performed. The primary outcome was a change in extrinsically activated ROTEM (EXTEM) clotting time (CT). RESULTS: 60 dialysis sessions from 56 patients were finally included, with 30 dialysis sessions per group. EXTEM CT was prolonged significantly after dialysis in the iHD group (90 [74, 128] vs. 74 [61, 91], p < 0.001), but did not change in the CRRT group (94.4 ± 29.4 vs. 91.6 ± 22.9, p = 0.986). The platelet aggregation did not change after both iHD and CRRT. A change in EXTEM CT was significantly greater in the iHD group compared to the CRRT group (p = 0.006). The difference in the incidence of bleeding events was insignificant between the two groups (p = 0.301). CONCLUSIONS: EXTEM CT was significantly prolonged after iHD, but this change was not shown after CRRT. Platelet function was not affected by both dialysis modalities.

FR-MIL: Distribution Re-calibration based Multiple Instance Learning with Transformer for Whole Slide Image Classification.

In digital pathology, whole slide images (WSI) are crucial for cancer prognostication and treatment planning. WSI classification is generally addressed using multiple instance learning (MIL), alleviating the challenge of processing billions of pixels and curating rich annotations. Though recent MIL approaches leverage variants of the attention mechanism to learn better representations, they scarcely study the properties of the data distribution itself i.e., different staining and acquisition protocols resulting in intra-patch and inter-slide variations. In this work, we first introduce a distribution re-calibration strategy to shift the feature distribution of a WSI bag (instances) using the statistics of the max-instance (critical) feature. Second, we enforce class (bag) separation via a metric loss assuming that positive bags exhibit larger magnitudes than negatives. We also introduce a generative process leveraging Vector Quantization (VQ) for improved instance discrimination i.e., VQ helps model bag latent factors for improved classification. To model spatial and context information, a position encoding module (PEM) is employed with transformer-based pooling by multi-head self-attention (PMSA). Evaluation of popular WSI benchmark datasets reveals our approach improves over state-of-the-art MIL methods. Further, we validate the general applicability of our method on classic MIL benchmark tasks and for point cloud classification with limited points https://github.com/PhilipChicco/FRMIL.

Cobalt Oxide Nanoparticle Synthesis by Cell-Surface-Engineered Recombinant Escherichia coli and Potential Application for Anticancer Treatment.

Cell surface display engineering facilitated the development of a cobalt-binding hybrid Escherichia coli. OmpC served as the molecular anchor for showcasing the cobalt-binding peptides (CBPs), creating the structural model of the hybrid OmpC-CBPs (OmpC-CP, OmpC-CF). Subsequently, the recombinant peptide's cobalt adsorption and retrieval effectiveness were evaluated at various concentrations. When subjected to a pH of 7 and a concentration of 2 mM, OmpC-CF exhibited a significantly higher cobalt recovery rate (2183.87 mol/g DCW) than OmpC-CP. The strain with bioadsorbed cobalt underwent thermal treatment at varying temperatures (400 °C, 500 °C, 600 °C, and 700 °C) and morphological characterization of the thermally decomposed cobalt nanoparticle oxides using diverse spectroscopy techniques. The analysis showed that nanoparticles confined themselves to metal ions, and EDS mapping detected the presence of cobalt on the cell surface. Finally, the nanoparticles' anticancer potential was assessed by subjecting them to heating at 500 °C in a furnace; they demonstrated noteworthy cytotoxicity, as evidenced by IC50 values of 59 µg/mL. These findings suggest that these nanoparticles hold promise as potential anticancer agents. Overall, this study successfully engineered a recombinant E. coli capable of efficiently binding to cobalt, producing nanoparticles with anticancer properties. The results of this investigation could have significant implications for advancing novel cancer therapies.

MLPH is a novel adipogenic factor controlling redox homeostasis to inhibit lipid peroxidation in adipocytes.

Abnormal adipose tissue formation is associated with metabolic disorders such as obesity, diabetes, and liver and cardiovascular diseases. Thus, identifying the novel factors that control adipogenesis is crucial for understanding these conditions and developing targeted treatments. In this study, we identified the melanosome-related factor MLPH as a novel adipogenic factor. MLPH was induced during the adipogenesis of 3T3-L1 cells and human mesenchymal stem cells. Although MLPH did not affect lipid metabolism, such as lipogenesis or lipolysis, adipogenesis was severely impaired by MLPH depletion. We observed that MLPH prevented excess reactive oxygen species (ROS) accumulation and lipid peroxidation during adipogenesis and in mature adipocytes. In addition, increased MLPH expression was observed under cirrhotic conditions in liver cancer cells and its overexpression also reduced ROS and lipid peroxidation. Our findings demonstrate that MLPH is a novel adipogenic factor that maintains redox homeostasis by preventing lipid peroxidation and ROS accumulation, which could lead to metabolic diseases.

Engineering of itaconic acid pathway via co-localization of CadA and AcnA in recombinant Escherichia coli.

Itaconic acid is an excellent polymeric precursor with a wide range of industrial applications. The efficient production of itaconate from various renewable substrates was demonstrated by engineered Escherichia coli. However, limitation in the itaconic acid precursor supply was revealed by finding out the key intermediate of the tricarboxylic acid in the itaconic acid pathway. Efforts of enhancing the cis-aconitate flux and preserving the isocitrate pool to increase itaconic acid productivity are required. In this study, we introduce a synthetic protein scaffold system between CadA and AcnA to physically combine the two enzymes. Through the introduction of a synthetic protein scaffold, 2.1 g L-1 of itaconic acid was produced at pH 7 and 37 °C. By fermentation, 20.1 g L-1 for 48 h of itaconic acid was produced with a yield of 0.34 g g-1 glycerol. These results suggest that carbon flux was successfully increased itaconic acid productivity.

Application of the surface engineered recombinant Escherichia coli to the industrial battery waste solution for lithium recovery.

Escherichia coli were engineered to selectively adsorb and recover lithium from the environment by employing a bacterial cell surface display strategy. Lithium binding peptide (LBP1) was integrated into the Escherichia coli membrane protein OmpC. The effect of environmental conditions on the adsorption of lithium by a recombinant strain was evaluated, and lithium particles on the cellular surface were analyzed by FE-SEM and XRD. To elevate the lithium adsorption, dimeric, trimeric, and tetrameric repeats of the LBP1 peptide were constructed and displayed on the surface of E. coli. The constructed recombinant E. coli displaying the LBP1 trimer was applied to real industrial lithium battery wastewater to recover lithium.

Hexane Fraction of Adenophora triphylla var. japonica Root Extract Inhibits Angiogenesis and Endothelial Cell-Induced Erlotinib Resistance in Lung Cancer Cells.

The aim of this study was to investigate the anti-angiogenic effects of the hexane fraction of Adenophora triphylla var. japonica root extract (HAT) and its influence on the development of erlotinib resistance in human lung cancer cells. HAT significantly reduced the migration, invasion, and tube formation of human umbilical vein endothelial cells (HUVECs). The phosphorylation levels of vascular endothelial growth factor receptor 2 (VEGFR2) and its downstream molecules were decreased via HAT, indicating its anti-angiogenic potential in endothelial cells (ECs). A docking analysis demonstrated that ß-sitosterol and lupeol, representative components of HAT, exhibit a high affinity for binding to VEGFR2. In addition, conditioned media from HAT-pretreated H1299 human lung cancer cells attenuated cancer-cell-induced chemotaxis of HUVECs, which was attributed to the decreased expression of angiogenic and chemotactic factors in H1299 cells. Interestingly, co-culture of erlotinib-sensitive PC9 human lung cancer cells with HUVECs induced erlotinib resistance in PC9 cells. However, co-culture with HAT-pretreated HUVECs partially restored the sensitivity of PC9 cells to erlotinib. HAT inhibited the development of erlotinib resistance by attenuating hepatocyte growth factor (HGF) production by ECs. Taken together, our results demonstrate that HAT exerts its anticancer effects by regulating the crosstalk between ECs and lung cancer cells.

A Study on Wheel Member Condition Recognition Using 1D-CNN.

The condition of a railway vehicle's wheels is an essential factor for safe operation. However, the current inspection of railway vehicle wheels is limited to periodic major and minor maintenance, where physical anomalies such as vibrations and noise are visually checked by maintenance personnel and addressed after detection. As a result, there is a need for predictive technology concerning wheel conditions to prevent railway vehicle damage and potential accidents due to wheel defects. Insufficient predictive technology for railway vehicle's wheel conditions forms the background for this study. In this research, a real-time tire wear classification system for light-rail rubber tires was proposed to reduce operational costs, enhance safety, and prevent service delays. To perform real-time condition classification of rubber tires, operational data from railway vehicles, including temperature, pressure, and acceleration, were collected. These data were processed and analyzed to generate training data. A 1D-CNN model was employed to classify tire conditions, and it demonstrated exceptionally high performance with a 99.4% accuracy rate.

A Study on Wheel Member Condition Recognition Using Machine Learning (Support Vector Machine).

The wheels of railway vehicles are of paramount importance in relation to railroad operations and safety. Currently, the management of railway vehicle wheels is restricted to post-event inspections of the wheels whenever physical phenomena, such as abnormal vibrations and noise, occur during the operation of railway vehicles. To address this issue, this paper proposes a method for predicting abnormalities in railway wheels in advance and enhancing the learning and prediction performance of machine learning algorithms. Data were collected during the operation of Line 4 of the Busan Metro in South Korea by directly attaching sensors to the railway vehicles. Through the analysis of key factors in the collected data, factors that can be used for tire condition classification were derived. Additionally, through data distribution analysis and correlation analysis, factors for classifying tire conditions were identified. As a result, it was determined that the z-axis of acceleration has a significant impact, and machine learning techniques such as SVM (Linear Kernel, RBF Kernel) and Random Forest were utilized based on acceleration data to classify tire conditions into in-service and defective states. The SVM (Linear Kernel) yielded the highest recognition rate at 98.70%.

Regulation of NLGN3 and the Synaptic Rho-GEF Signaling Pathway by CDK5.

Neuroligins (NLGNs) are postsynaptic cell adhesion molecules that are involved in synapse assembly and function. The NLGN gene family consists of 5 genes (NLGN1-3, 4X, and 4Y). NLGN3 forms heterodimers with other NLGNs and is expressed at both excitatory and inhibitory synapses, although the distinct role at different synapses is not fully understood. Cyclin-dependent kinase 5 (Cdk5) is a proline-directed serine/threonine kinase that targets various neuronal substrates to impact neuronal migration, neurite outgrowth, synaptic transmission, and plasticity. Both NLGNs and their presynaptic binding partners neurexins are highly associated with neurodevelopmental disorders. The NLGN3 gene is on the X chromosome and variants in NLGN3 have been linked to the pathophysiology in neurodevelopmental disorders. To better understand the endogenous modulation of NLGN3, we generated an HA-tagged knock-in mouse. We found that Cdk5 associates with NLGN3 in vivo and phosphorylates NLGN3 on serine 725 (S725) in the knock-in mouse of either sex. The phosphorylation affects the NLGN3 association with Kalirin-7, a postsynaptic guanine nucleotide exchange factors for Rho GTPase family proteins. We further observed that the phosphorylation modulates NLGN3 surface expression and NLGN3-mediated synaptic currents in cultured rat neurons. Thus, we characterized NLGN3 as a novel Cdk5 substrate and revealed the functional consequences of NLGN3 S725 phosphorylation in neurons. Our study provides a novel molecular mechanism underlying Cdk5-mediated regulation of postsynaptic cell adhesion molecules.SIGNIFICANCE STATEMENT NLGN3 is involved in synapse assembly and function at both excitatory and inhibitory synapses and has been associated with the pathophysiology of neurodevelopmental disorders. Cdk5 has brain-specific activity and is involved in neuronal transmission, synapse function, and plasticity. Here, we characterize NLGN3 as a Cdk5 substrate for the first time and show that Cdk5-mediated phosphorylation regulates NLGN3 function. We demonstrate that NLGN3 S725 is a Cdk5 phosphorylation site, and reveal that the site is important for NLGN3 association with Kalirin-7, NLGN3 surface expression, and NLGN3-mediated synaptic transmission.

MicroRNA miR-214-5p induces senescence of microvascular endothelial cells by targeting the JAG1/Notch signaling pathway.

During cellular senescence, irreversible cell cycle arrest is accompanied by morphological and genetic alterations. MicroRNAs (miRNAs) play a critical role in regulating senescence by modulating the abundance of crucial senescence regulatory proteins. Therefore, to identify novel senescence-associated miRNAs, we analyzed differentially expressed miRNAs in microvascular endothelial cells (MVEC). Among the 80 differentially expressed miRNAs in replicative senescent MVECs, 16 miRNAs of unknown gene ontology were used in the senescence-associated ß-galactosidase assay. Thus, we identified miR-214-5p as having high senescence-inducing activity, inhibiting the proliferation and angiogenesis activity of MVECs. To reveal the senescence-regulating mechanism of miR-214-5p, we searched for target genes through sequence- and literature-based analysis. Molecular manipulation of miR-214-5p demonstrated that miR-214-5p regulated the expression and function of Jagged 1 (JAG1) in senescent MVECs. Silencing JAG1 or downstream genes of JAG1-Notch signaling, accelerated the senescence of MVECs. Additionally, ectopic overexpression of JAG1 reversed the senescence-inducing activity of miR-214-5p. In conclusion, we identified miR-214-5p as a senescence-associated miRNA. Targeting miR-214-5p may be a potential strategy to delay vascular aging and overcome the detrimental effects of senescence and age-related diseases.

Formation of Antifouling Brushes on Various Substrates Using a Melanin-Inspired Initiator Film.

In this study, we developed a substrate-independent initiator film that can undergo surface-initiated polymerization to form an antifouling brush. Inspired by the melanogenesis found in nature, we synthesized a tyrosine-conjugated bromide initiator (Tyr-Br) that contains phenolic amine groups as the dormant coating precursor and α-bromoisobutyryl groups as the initiator. The resultant Tyr-Br was stable under ambient air conditions and underwent melanin-like oxidation only in the presence of tyrosinase to form an initiator film on various substrates. Subsequently, an antifouling polymer brush was formed using air-tolerant activators regenerated by electron transfer for atom transfer radical polymerization (ARGET ATRP) of zwitterionic carboxybetaine. The entire surface coating procedure, including the initiator layer formation and ARGET ATRP, occurred under aqueous conditions and did not require organic solvents or chemical oxidants. Therefore, antifouling polymer brushes can be feasibly formed not only on experimentally preferred substrates (e.g., Au, SiO2, and TiO2) but also on polymeric substrates such as poly(ethylene terephthalate) (PET), cyclic olefin copolymer (COC), and nylon.

An Ethanol Extract of Perilla frutescens Leaves Suppresses Adrenergic Agonist-Induced Metastatic Ability of Cancer Cells by Inhibiting Src-Mediated EMT.

Previous studies have indicated that the adrenergic receptor signaling pathway plays a fundamental role in chronic stress-induced cancer metastasis. In this study, we investigated whether an ethanol extract of Perilla frutescens leaves (EPF) traditionally used to treat stress-related symptoms by moving Qi could regulate the adrenergic agonist-induced metastatic ability of cancer cells. Our results show that adrenergic agonists including norepinephrine (NE), epinephrine (E), and isoproterenol (ISO) increased migration and invasion of MDA-MB-231 human breast cancer cells and Hep3B human hepatocellular carcinoma cells. However, such increases were completely abrogated by EPF treatment. E/NE induced downregulation of E-cadherin and upregulation of N-cadherin, Snail, and Slug. Such effects were clearly reversed by pretreatment with EPF, suggesting that the antimetastatic activity of EPF could be related to epithelial-mesenchymal transition (EMT) regulation. EPF suppressed E/NE-stimulated Src phosphorylation. Inhibition of Src kinase activity with dasatinib completely suppressed the E/NE-induced EMT process. Transfecting MDA-MB-231 cells with constitutively activated Src (SrcY527F) diminished the antimigration effect of EPF. Taken together, our results demonstrate that EPF can suppress the adrenergic agonist-promoted metastatic ability of cancer cells by inhibiting Src-mediated EMT. This study provides basic evidence supporting the probable use of EPF to prevent metastasis in cancer patients, especially those under chronic stress.

Phytochrome B photobodies are comprised of phytochrome B and its primary and secondary interacting proteins.

Phytochrome B (phyB) is a plant photoreceptor that forms a membraneless organelle called a photobody. However, its constituents are not fully known. Here, we isolated phyB photobodies from Arabidopsis leaves using fluorescence-activated particle sorting and analyzed their components. We found that a photobody comprises ~1,500 phyB dimers along with other proteins that could be classified into two groups: The first includes proteins that directly interact with phyB and localize to the photobody when expressed in protoplasts, while the second includes proteins that interact with the first group proteins and require co-expression of a first-group protein to localize to the photobody. As an example of the second group, TOPLESS interacts with PHOTOPERIODIC CONTROL OF HYPOCOTYL 1 (PCH1) and localizes to the photobody when co-expressed with PCH1. Together, our results support that phyB photobodies include not only phyB and its primary interacting proteins but also its secondary interacting proteins.

Improving Performance of the Human Pupil Orbit Model (HPOM) Estimation Method for Eye-Gaze Tracking.

Eye-gaze direction-tracking technology is used in fields such as medicine, education, engineering, and gaming. Stability, accuracy, and precision of eye-gaze direction-tracking are demanded with simultaneous upgrades in response speed. In this study, a method is proposed to improve the speed with decreases in the system load and precision in the human pupil orbit model (HPOM) estimation method. The new method was proposed based on the phenomenon that the minor axis of the elliptical-deformed pupil always pointed toward the rotational center presented in various eye-gaze direction detection studies and HPOM estimation methods. Simulation experimental results confirmed that the speed was improved by at least 74 times by consuming less than 7 ms compared to the HPOM estimation. The accuracy of the eye's ocular rotational center point showed a maximum error of approximately 0.2 pixels on the x-axis and approximately 8 pixels on the y-axis. The precision of the proposed method was 0.0 pixels when the number of estimation samples (ES) was 7 or less, which showed results consistent with those of the HPOM estimation studies. However, the proposed method was judged to work conservatively against the allowable angle error (AAE), considering that the experiment was conducted under the worst conditions and the cost used to estimate the final model. Therefore, the proposed method could estimate HPOM with high accuracy and precision through AAE adjustment according to system performance and the usage environment.

A Study on Object Detection Performance of YOLOv4 for Autonomous Driving of Tram.

Recently, autonomous driving technology has been in the spotlight. However, autonomous driving is still in its infancy in the railway industry. In the case of railways, there are fewer control elements than autonomous driving of cars due to the characteristics of running on railways, but there is a disadvantage in that evasive maneuvers cannot be made in the event of a dangerous situation. In addition, when braking, it cannot be decelerated quickly for the weight of the body and the safety of the passengers. In the case of a tram, one of the railway systems, research has already been conducted on how to generate a profile that plans braking and acceleration as a base technology for autonomous driving, and to find the location coordinates of surrounding objects through object recognition. In pilot research about the tram's automated driving, YOLOv3 was used for object detection to find object coordinates. YOLOv3 is an artificial intelligence model that finds coordinates, sizes, and classes of objects in an image. YOLOv3 is the third upgrade of YOLO, which is one of the most famous object detection technologies based on CNN. YOLO's object detection performance is characterized by ordinary accuracy and fast speed. For this paper, we conducted a study to find out whether the object detection performance required for autonomous trams can be sufficiently implemented with the already developed object detection model. For this experiment, we used the YOLOv4 which is the fourth upgrade of YOLO.

ß2-Adrenergic Receptor Signaling Pathway Stimulates the Migration and Invasion of Cancer Cells via Src Activation.

Chronic stress has been reported to stimulate the release of catecholamines, including norepinephrine (NE) and epinephrine (E), which promote cancer progression by activating the adrenergic receptor (AR). Although previous studies showed that ß2-AR mediated chronic stress-induced tumor growth and metastasis, the underlying mechanism has not been fully explored. In this study, we aimed to investigate the molecular mechanism by which ß2-AR exerts a pro-metastatic function in hepatocarcinoma (HCC) cells and breast cancer (BC) cells. Our results showed that Hep3B human HCC cells and MDA-MB-231 human BC cells exhibited the highest ADRB2 expression among diverse HCC and BC cell lines. NE, E, and isoprenaline (ISO), adrenergic agonists commonly increased the migration and invasion of Hep3B cells and MDA-MB-231 cells. The phosphorylation level of Src was significantly increased by E/NE. Dasatinib, a Src kinase inhibitor, blocked E/NE-induced migration and invasion, indicating that AR agonists enhanced the mobility of cancer cells by activating Src. ADRB2 knockdown attenuated E/NE-induced Src phosphorylation, as well as the metastatic ability of cancer cells, suggesting the essential role of ß2-AR. Taken together, our results demonstrate that chronic stress-released catecholamines promoted the migration and invasion of HCC cells and BC cells via ß2-AR-mediated Src activation.

Antifouling Surface Coating on Various Substrates by Inducing Tyrosinase-Mediated Oxidation of a Tyrosine-Conjugated Sulfobetaine Derivative.

Inspired by the melanogenesis occurring in nature, we report tyrosinase-mediated antifouling surface coating by synthesizing a tyrosine-conjugated sulfobetaine derivative (Tyr-SB). Synthetic Tyr-SB contains zwitterionic sulfobetaine and tyrosine, whose phenolic amine group acts as a dormant coating precursor. In contrast to catecholamine derivatives, tyrosine derivatives are stable against auto-oxidation and are enzymatically oxidized only in the presence of tyrosinase to initiate melanin-like oxidation. When the surface of interest was applied during the course of Tyr-SB oxidation, a superhydrophilic poly(Tyr-SB) film was coated on the surfaces, thereby showing antifouling performance against proteins or adherent cells. Because the oxidation of Tyr-SB occurred under mild aqueous conditions (pH 6-7) without the use of any chemical oxidants, such as sodium periodate or ammonium persulfate, we anticipate that the coating method described herein will serve as a biocompatible tool in the field of biosensors, cell surface engineering, and medical devices, whose interfaces differ in chemistry.

Receptor interacting protein 1 knockdown induces cell death in liver cancer by suppressing STAT3/ATR activation in a p53-dependent manner.

The survival and death of eukaryotic cells are tightly controlled by a variety of proteins in response to the cellular environment. Receptor-interacting serine/threonine-protein kinase 1 (RIPK1) is a receptor-interacting Ser/Thr kinase that has recently been reported as an important regulator of cell survival, apoptosis, and necroptosis; however, its role in liver cancer remains unclear. In this study, we examined the effect of siRNA-mediated RIPK1 knockdown on the survival and death of liver cancer cells. Treatment with siRIPK1 decreased the growth rate of liver cancer cells and increased apoptotic, but not necrotic cell death, which was higher in wild-type p53 (wt-p53) cells than in mutant-type p53 (mt-p53) cells. In addition, RIPK1 knockdown increased p53 expression and G1 phase arrest in wt-p53 cells. Although suppressing p53 did not alter RIPK1 expression, it did attenuate siRIPK1-induced cell death. Interestingly, RIPK1 knockdown also increased the generation of reactive oxygen species and DNA damage by inhibiting signal transduced and activator of transcription 3 (STAT3) and ATM and RAD3-related (ATR) in wt-p53 cells but not in mt-p53 cells. Moreover, STAT3 or ATR inhibition in p53 mutant cells restored siRIPK1-mediated cell death. Together, the results of this study suggest that RIPK1 suppression induces apoptotic cell death by inhibiting the STAT3/ATR axis in a p53-dependent manner. Furthermore, these findings suggest that RIPK1, alone or in combination, may be a promising target for treating liver cancer.

Weakly supervised segmentation on neural compressed histopathology with self-equivariant regularization.

In digital pathology, segmentation is a fundamental task for the diagnosis and treatment of diseases. Existing fully supervised methods often require accurate pixel-level annotations that are both time-consuming and laborious to generate. Typical approaches first pre-process histology images into patches to meet memory constraints and later perform stitching for segmentation; at times leading to lower performance given the lack of global context. Since image level labels are cheaper to acquire, weakly supervised learning is a more practical alternative for training segmentation algorithms. In this work, we present a weakly supervised framework for histopathology segmentation using only image-level labels by refining class activation maps (CAM) with self-supervision. First, we compress gigapixel histology images with an unsupervised contrastive learning technique to retain high-level spatial context. Second, a network is trained on the compressed images to jointly predict image-labels and refine the initial CAMs via self-supervised losses. In particular, we achieve refinement via a pixel correlation module (PCM) that leverages self-attention between the initial CAM and the input to encourage fine-grained activations. Also, we introduce a feature masking technique that performs spatial dropout on the compressed input to suppress low confidence predictions. To effectively train our model, we propose a loss function that includes a classification objective with image-labels, self-supervised regularization and entropy minimization between the CAM predictions. Experimental results on two curated datasets show that our approach is comparable to fully-supervised methods and can outperform existing state-of-the-art patch-based methods. https://github.com/PhilipChicco/wsshisto.