Deep learning-based image analysis predicts PD-L1 status from H&E-stained histopathology images in breast cancer.

Programmed death ligand-1 (PD-L1) has been recently adopted for breast cancer as a predictive biomarker for immunotherapies. The cost, time, and variability of PD-L1 quantification by immunohistochemistry (IHC) are a challenge. In contrast, hematoxylin and eosin (H&E) is a robust staining used routinely for cancer diagnosis. Here, we show that PD-L1 expression can be predicted from H&E-stained images by employing state-of-the-art deep learning techniques. With the help of two expert pathologists and a designed annotation software, we construct a dataset to assess the feasibility of PD-L1 prediction from H&E in breast cancer. In a cohort of 3,376 patients, our system predicts the PD-L1 status in a high area under the curve (AUC) of 0.91 - 0.93. Our system is validated on two external datasets, including an independent clinical trial cohort, showing consistent prediction performance. Furthermore, the proposed system predicts which cases are prone to pathologists miss-interpretation, showing it can serve as a decision support and quality assurance system in clinical practice.

Advances on the impact of thermal processing on structure and antigenicity of chicken ovomucoid.

BACKGROUND: Ovomucoid (OVM) is the dominant allergen found in egg white. The heat-induced changes on chicken OVM structure and antigenic properties were assessed at acidic, neutral and alkaline pH values. RESULTS: The fluorescence spectroscopy measurements indicated changes in the conformation of OVM caused by both pH and thermal treatment. The OVM molecule exhibited higher exposure of hydrophobic residues at 7.0, as indicated by the synchronous spectra, intrinsic fluorescence and quenching experiments. When heating the protein at pH 9.5, the molecular structure appeared more compact. The antigenic properties of OVM, estimated through the enzyme-linked immunosorbent assay, appeared not to be sensitive to heat at pH 7.0 and 4.5. Single molecule level investigations indicated that the secondary and tertiary structure of OVM was affected by the thermal treatment. CONCLUSIONS: Experimental results indicated over 90% reduction of the antigenicity at pH 9.5 and temperature of 100 °C. Significant changes of the linear epitopes exposure and location of the conformational epitopes were highlighted after performing heating molecular dynamics simulations of OVM from 25 °C to 100 °C. © 2017 Society of Chemical Industry.

Regulation of cell proliferation by ERK and signal-dependent nuclear translocation of ERK is dependent on Tm5NM1-containing actin filaments.

ERK-regulated cell proliferation requires multiple phosphorylation events catalyzed first by MEK and then by casein kinase 2 (CK2), followed by interaction with importin7 and subsequent nuclear translocation of pERK. We report that genetic manipulation of a core component of the actin filaments of cancer cells, the tropomyosin Tm5NM1, regulates the proliferation of normal cells both in vitro and in vivo. Mouse embryo fibroblasts (MEFs) lacking Tm5NM1, which have reduced proliferative capacity, are insensitive to inhibition of ERK by peptide and small-molecule inhibitors, indicating that ERK is unable to regulate proliferation of these knockout (KO) cells. Treatment of wild-type MEFs with a CK2 inhibitor to block phosphorylation of the nuclear translocation signal in pERK resulted in greatly decreased cell proliferation and a significant reduction in the nuclear translocation of pERK. In contrast, Tm5NM1 KO MEFs, which show reduced nuclear translocation of pERK, were unaffected by inhibition of CK2. This suggested that it is nuclear translocation of CK2-phosphorylated pERK that regulates cell proliferation and this capacity is absent in Tm5NM1 KO cells. Proximity ligation assays confirmed a growth factor-stimulated interaction of pERK with Tm5NM1 and that the interaction of pERK with importin7 is greatly reduced in the Tm5NM1 KO cells.

Inhibition of tumor-associated αvß3 integrin regulates the angiogenic switch by enhancing expression of IGFBP-4 leading to reduced melanoma growth and angiogenesis in vivo.

A more complete understanding of the mechanisms that regulate the angiogenic switch, which contributes to the conversion of small dormant tumors to actively growing malignancies, is important for the development of more effective anti-angiogenic strategies for cancer therapy. While significant progress has been made in understanding the complex mechanisms by which integrin αvß3 expressed in endothelial cells governs angiogenesis, less is known concerning the ability of αvß3 expressed within the tumor cell compartment to modulate the angiogenic output of a tumor. Here we provide evidence that αvß3 expressed in melanoma cells may contribute to the suppression of IGFBP-4, an important negative regulator of IGF-1 signaling. Given the multiple context-dependent roles for αvß3 in angiogenesis and tumor progression, our novel findings provide additional molecular insight into how αvß3 may govern the angiogenic switch by a mechanism associated with a p38 MAPK and matrix metalloproteinases-dependent regulation of the endogenous angiogenesis inhibitor IGFBP-4.

A FAK-Cas-Rac-lamellipodin signaling module transduces extracellular matrix stiffness into mechanosensitive cell cycling.

Tissue and extracellular matrix (ECM) stiffness is transduced into intracellular stiffness, signaling, and changes in cellular behavior. Integrins and several of their associated focal adhesion proteins have been implicated in sensing ECM stiffness. We investigated how an initial sensing event is translated into intracellular stiffness and a biologically interpretable signal. We found that a pathway consisting of focal adhesion kinase (FAK), the adaptor protein p130Cas (Cas), and the guanosine triphosphatase Rac selectively transduced ECM stiffness into stable intracellular stiffness, increased the abundance of the cell cycle protein cyclin D1, and promoted S-phase entry. Rac-dependent intracellular stiffening involved its binding partner lamellipodin, a protein that transmits Rac signals to the cytoskeleton during cell migration. Our findings establish that mechanotransduction by a FAK-Cas-Rac-lamellipodin signaling module converts the external information encoded by ECM stiffness into stable intracellular stiffness and mechanosensitive cell cycling. Thus, lamellipodin is important not only in controlling cellular migration but also for regulating the cell cycle in response to mechanical signals.

Insulin-like growth factor binding protein-4 differentially inhibits growth factor-induced angiogenesis.

An in-depth understanding of the molecular and cellular complexity of angiogenesis continues to advance as new stimulators and inhibitors of blood vessel formation are uncovered. Gaining a more complete understanding of the response of blood vessels to both stimulatory and inhibitory molecules will likely contribute to more effective strategies to control pathological angiogenesis. Here, we provide evidence that endothelial cell interactions with structurally altered collagen type IV may suppress the expression of insulin-like growth factor binding protein-4 (IGFBP-4), a well documented inhibitor of the IGF-1/IGF-1R signaling axis. We report for the first time that IGFBP-4 differentially inhibits angiogenesis induced by distinct growth factor signaling pathways as IGFBP-4 inhibited FGF-2- and IGF-1-stimulated angiogenesis but failed to inhibit VEGF-induced angiogenesis. The resistance of VEGF-stimulated angiogenesis to IGFBP-4 inhibition appears to depend on sustained activation of p38 MAPK as blocking its activity restored the anti-angiogenic effects of IGFBP-4 on VEGF-induced blood vessel growth in vivo. These novel findings provide new insight into how blood vessels respond to endogenous inhibitors during angiogenesis stimulated by distinct growth factor signaling pathways.

Studying the effects of matrix stiffness on cellular function using acrylamide-based hydrogels.

Tissue stiffness is an important determinant of cellular function, and changes in tissue stiffness are commonly associated with fibrosis, cancer and cardiovascular disease. Traditional cell biological approaches to studying cellular function involve culturing cells on a rigid substratum (plastic dishes or glass coverslips) which cannot account for the effect of an elastic ECM or the variations in ECM stiffness between tissues. To model in vivo tissue compliance conditions in vitro, we and others use ECM-coated hydrogels. In our laboratory, the hydrogels are based on polyacrylamide which can mimic the range of tissue compliances seen biologically. "Reactive" cover slips are generated by incubation with NaOH followed by addition of 3-APTMS. Glutaraldehyde is used to cross-link the 3-APTMS and the polyacrylamide gel. A solution of acrylamide (AC), bis-acrylamide (Bis-AC) and ammonium persulfate is used for the polymerization of the hydrogel. N-hydroxysuccinimide (NHS) is incorporated into the AC solution to crosslink ECM protein to the hydrogel. Following polymerization of the hydrogel, the gel surface is coated with an ECM protein of choice such as fibronectin, vitronectin, collagen, etc. The stiffness of a hydrogel can be determined by rheology or atomic force microscopy (AFM) and adjusted by varying the percentage of AC and/or bis-AC in the solution. In this manner, substratum stiffness can be matched to the stiffness of biological tissues which can also be quantified using rheology or AFM. Cells can then be seeded on these hydrogels and cultured based upon the experimental conditions required. Imaging of the cells and their recovery for molecular analysis is straightforward. For this article, we define soft substrata as those having elastic moduli (E) < 3000 Pascal and stiff substrata/tissues as those with E > 20,000 Pascal.

Stress sensor Gadd45 genes as therapeutic targets in cancer.

Gadd45 genes have been implicated in stress signaling responses to various physiological or environmental stressors, resulting in cell cycle arrest, DNA repair, cell survival and senescence, or apoptosis. Evidence accumulated up to date suggests that Gadd45 proteins function as stress sensors, mediating their activity through a complex interplay of physical interactions with other cellular proteins that are implicated in cell cycle regulation and the response of cells to stress. These include PCNA, p21, cdc2/cyclinB1, and the p38 and JNK stress response kinases. Disregulated expression of Gadd45 has been observed in multiple types of solid tumors as well as in hematopoietic malignancies. Also, evidence has accumulated that Gadd45 proteins are intrinsically associated with the response of tumor cells to a variety of cancer therapeutic agents. Thus, Gadd45 proteins may represent a novel class of targets for therapeutic intervention in cancer. Additional research is needed to better understand which of the Gadd45 stress response functions may be targeted for chemotherapeutic drug design in cancer therapy.

Impact of the non-cellular tumor microenvironment on metastasis: potential therapeutic and imaging opportunities.

Evidence is accumulating that the malignant phenotype of a given tumor is dependent not only on the intrinsic characteristics of tumor cells, but also on the cooperative interactions of non-neoplastic cells, soluble secreted factors and the non-cellular solid-state ECM network that comprise the tumor microenvironment. Given the ability of the tumor microenvironment to regulate the cellular phenotype, recent efforts have focused on understanding the molecular mechanisms by which cells sense, assimilate, interpret, and ultimately respond to their immediate surroundings. Exciting new studies are beginning to unravel the complex interactions between the numerous cell types and regulatory factors within the tumor microenvironment that function cooperatively to control tumor cell invasion and metastasis. Here, we will focus on studies concerning a common theme, which is the central importance of the non-cellular solid-state compartment as a master regulator of the malignant phenotype. We will highlight the non-cellular solid-state compartment as a relatively untapped source of therapeutic and imaging targets and how cellular interactions with these targets may regulate tumor metastasis.

Disruption of endothelial cell interactions with the novel HU177 cryptic collagen epitope inhibits angiogenesis.

PURPOSE: The importance of cellular communication with the extracellular matrix in regulating cellular invasion is well established. Selective disruption of communication links between cells and the local microenvironment by specifically targeting non-cellular matrix-immobilized cryptic extracellular matrix epitopes may represent an effective new clinical approach to limit tumor-associated angiogenesis. Therefore, we sought to determine whether the HU177 cryptic collagen epitope plays a functional role in regulating angiogenesis in vivo. EXPERIMENTAL DESIGN: We examined the expression and characterized the HU177 cryptic collagen epitope in vitro and in vivo using immunohistochemistry and ELISA. We examined potential mechanisms by which this cryptic collagen epitope may regulate angiogenesis using in vitro cell adhesion, migration, proliferation, and biochemical assays. Finally, we examined the whether blocking cellular interactions with the HU177 cryptic epitope plays a role in angiogenesis and tumor growth in vivo using the chick embryo model. RESULTS: The HU177 cryptic epitope was selectively exposed within tumor blood vessel extracellular matrix, whereas little was associated with quiescent vessels. An antibody directed to this cryptic site selectively inhibited endothelial cell adhesion, migration, and proliferation on denatured collagen type IV and induced increased levels of cyclin-dependent kinase inhibitor p27(KIP1). Systemic administration of mAb HU177 inhibited cytokine- and tumor-induced angiogenesis in vivo. CONCLUSIONS: We provide evidence for a new functional cryptic regulatory element within collagen IV that regulates tumor angiogenesis. These findings suggest a novel and highly selective approach for regulating angiogenesis by targeting a non-cellular cryptic collagen epitope.

Inhibition of experimental metastasis by targeting the HUIV26 cryptic epitope in collagen.

Metastasis from the primary tumor to distant sites involves an array of molecules that function in an integrated manner. Proteolytic remodeling and subsequent tumor cell interactions with the extracellular matrix regulate tumor invasion. In previous studies, we have identified a cryptic epitope (HUIV26) that is specifically exposed after alterations in the triple helical structure of type IV collagen. Exposure of this cryptic epitope plays a fundamental role in the regulation of angiogenesis in vivo. However, little is known concerning the ability of tumor cells to interact with this cryptic site or whether this site regulates tumor cell metastasis in vivo. In this regard, many of the same cellular processes that regulate angiogenesis also contribute to tumor metastasis. Here we provide evidence that tumor cells such as B16F10 melanoma interact with denatured collagen type IV in part by recognizing the HUIV26 cryptic site. Systemic administration of a HUIV26 monoclonal antibody inhibited experimental metastasis of B16F10 melanoma in vivo. Taken together, our findings suggest that tumor cell interactions with the HUIV26 cryptic epitope play an important role in regulating experimental metastasis and that this cryptic element may represent a therapeutic target for controlling the spread of tumor cells to distant sites.

Targeting integrins for the control of tumour angiogenesis.

The crucial role of cell extracellular matrix communication in angiogenesis is well established; thus, it is not surprising that integrins have gained considerable attention as targets for the treatment of neovascular disease. Given the diversity of ligands and complexity of integrin signalling, a new appreciation for the divergent roles of integrins in angiogenesis is emerging. It is becoming clear that integrins regulate angiogenesis in both a positive and negative manner. New studies have provided a better understanding of integrin structure as it relates to ligand binding and signalling. This new insight has opened exciting possibilities for the design of novel inhibitors for clinical applications. In this review, studies concerning the cooperative interactions between integrins and regulatory molecules and possible new strategies for controlling angiogenesis will be discussed.

Human and rat glioma growth, invasion, and vascularization in a novel chick embryo brain tumor model.

The mechanisms that control the insidiously invasive nature of malignant gliomas are poorly understood, and their study would be facilitated by an in vivo model that is easy to manipulate and inexpensive. The developing chick embryo brain was assessed as a new xenograft model for the production, growth, and study of human and rat glioma cell lines. Three established glioma lines (U-87 MG, C6, and 9L) were injected into chick embryo brain ventricles on embryonic day (E) 5 and brains were examined after several days to two weeks after injection. All glioma lines survived, produced vascularized intraventricular tumors, and invaded the brain in a manner similar to that in rodents. Rat C6 glioma cells spread along vasculature and also invaded the neural tissue. Human U-87 glioma cells migrated along vasculature and exhibited slight invasion of neural tissue. Rat 9L gliosarcoma cells were highly motile, but migrated only along the vasculature. A derivative of 9L cells that stably expressed the cell surface adhesion molecule NgCAM/L1 was produced and also injected into chick embryo brain ventricles to see if this protein could facilitate tumor cell migration away from the vasculature into areas such as axonal tracts. 9L/NgCAM cells, however, did not migrate away from the vasculature and, thus, this protein alone cannot be responsible for diffuse invasiveness of some gliomas. 9L/NgCAM cell motility was assessed in vitro using sophisticated time-lapse microscopy and quantitative analysis, and was significantly altered compared to parental 9L cells. These studies demonstrate that the chick embryo brain is a successful and novel xenograft model for mammalian gliomas and demonstrate the potential usefulness of this new model for studying glioma tumor cell growth, vascularization, and invasiveness.