High-content screening of drug combinations of an mPGES-1 inhibitor in multicellular tumor spheroids leads to mechanistic insights into neuroblastoma chemoresistance

High-throughput drug screening enables the discovery of new anticancer drugs. Although monolayer cell cultures are commonly used for screening, their limited complexity and translational efficiency require alternative models. Three-dimensional cell cultures, such as multicellular tumor spheroids (MCTS), mimic tumor architecture and offer promising opportunities for drug discovery. In this study, we developed a neuroblastoma MCTS model for high-content drug screening. We also aimed to decipher the mechanisms underlying synergistic drug combinations in this disease model. Several agents from different therapeutic categories and with different mechanisms of action were tested alone or in combination with selective inhibition of prostaglandin E2 by pharmacological inhibition of microsomal prostaglandin E synthase-1 (mPGES-1). After a systematic investigation of the sensitivity of individual agents and the effects of pairwise combinations, GFP-transfected MCTS were used in a confirmatory screen to validate the hits. Finally, inhibitory effects on multidrug resistance proteins were examined. In summary, we demonstrate how MCTS-based high-throughput drug screening has the potential to uncover effective drug combinations and provide insights into the mechanism of synergy between an mPGES-1 inhibitor and chemotherapeutic agents.

Data-driven approaches for identifying hyperparameters in multi-step retrosynthesis

The multi-step retrosynthesis problem can be solved by a search algorithm, such as Monte Carlo tree search (MCTS). The performance of multistep retrosynthesis, as measured by a trade-off in search time and route solvability, therefore depends on the hyperparameters of the search algorithm. In this paper, we demonstrated the effect of three MCTS hyperparameters (number of iterations, tree depth, and tree width) on metrics such as Linear integrated speed-accuracy score (LISAS) and Inverse efficiency score which consider both route solvability and search time. This exploration was conducted by employing three data-driven approaches, namely a systematic grid search, Bayesian optimization over an ensemble of molecules to obtain static MCTS hyperparameters, and a machine learning approach to dynamically predict optimal MCTS hyperparameters given an input target molecule. With the obtained results on the internal dataset, we demonstrated that it is possible to identify a hyperparameter set which outperforms the current AiZynthFinder default setting. It appeared optimal across a variety of target input molecules, both on proprietary and public datasets. The settings identified with the in-house dataset reached a solvability of 93 % and median search time of 151 s for the in-house dataset, and a 74 % solvability and 114 s for the ChEMBL dataset. These numbers can be compared to the current default settings which solved 85 % and 73 % during a median time of 110s and 84 s, for in-house and ChEMBL, respectively.

AlphaDrug: protein target specific de novo molecular generation

Traditional drug discovery is very laborious, expensive, and time-consuming, due to the huge combinatorial complexity of the discrete molecular search space. Researchers have turned to machine learning methods for help to tackle this difficult problem. However, most existing methods are either virtual screening on the available database of compounds by protein-ligand affinity prediction, or unconditional molecular generation, which does not take into account the information of the protein target. In this paper, we propose a protein target-oriented de novo drug design method, called AlphaDrug. Our method is able to automatically generate molecular drug candidates in an autoregressive way, and the drug candidates can dock into the given target protein well. To fulfill this goal, we devise a modified transformer network for the joint embedding of protein target and the molecule, and a Monte Carlo tree search (MCTS) algorithm for the conditional molecular generation. In the transformer variant, we impose a hierarchy of skip connections from protein encoder to molecule decoder for efficient feature transfer. The transformer variant computes the probabilities of next atoms based on the protein target and the molecule intermediate. We use the probabilities to guide the look-ahead search by MCTS to enhance or correct the next-atom selection. Moreover, MCTS is also guided by a value function implemented by a docking program, such that the paths with many low docking values are seldom chosen. Experiments on diverse protein targets demonstrate the effectiveness of our methods, indicating that AlphaDrug is a potentially promising solution to target-specific de novo drug design.

Natural and Synthetic Biomaterials for Engineering Multicellular Tumor Spheroids

The lack of in vitro models that represent the native tumor microenvironment is a significant challenge for cancer research. Two-dimensional (2D) monolayer culture has long been the standard for in vitro cell-based studies. However, differences between 2D culture and the in vivo environment have led to poor translation of cancer research from in vitro to in vivo models, slowing the progress of the field. Recent advances in three-dimensional (3D) culture have improved the ability of in vitro culture to replicate in vivo conditions. Although 3D cultures still cannot achieve the complexity of the in vivo environment, they can still better replicate the cell-cell and cell-matrix interactions of solid tumors. Multicellular tumor spheroids (MCTS) are three-dimensional (3D) clusters of cells with tumor-like features such as oxygen gradients and drug resistance, and represent an important translational tool for cancer research. Accordingly, natural and synthetic polymers, including collagen, hyaluronic acid, Matrigel^®, polyethylene glycol (PEG), alginate and chitosan, have been used to form and study MCTS for improved clinical translatability. This review evaluates the current state of biomaterial-based MCTS formation, including advantages and disadvantages of the different biomaterials and their recent applications to the field of cancer research, with a focus on the past five years.

Numerical Analysis of Fatigue Crack Growth Path and Life Predictions for Linear Elastic Material

The main objective of this work was to present a numerical modelling of crack growth path in linear elastic materials under mixed-mode loadings, as well as to study the effect of presence of a hole on fatigue crack propagation and fatigue life in a modified compact tension specimen under constant amplitude loading condition. The ANSYS Mechanical APDL 19.2 is implemented for accurate prediction of the crack propagation paths and the associated fatigue life under constant amplitude loading conditions using a new feature in ANSYS which is the smart crack growth technique. The Paris law model has been employed for the evaluation of the mixed-mode fatigue life for the modified compact tension specimen (MCTS) with different configuration of MCTS under the linear elastic fracture mechanics (LEFM) assumption. The approach involves accurate evaluation of stress intensity factors (SIFs), path of crack growth and a fatigue life evaluation through an incremental crack extension analysis. Fatigue crack growth results indicate that the fatigue crack has always been attracted to the hole, so either it can only curve its path and propagate towards the hole, or it can only float from the hole and grow further once the hole has been lost. In terms of trajectories of crack propagation under mixed-mode load conditions, the results of this study are validated with several crack propagation experiments published in literature showing the similar observations. Accurate results of the predicted fatigue life were achieved compared to the two-dimensional data performed by other researchers.

A Triple Combination of Metformin, Acetylsalicylic Acid, and Oseltamivir Phosphate Impacts Tumour Spheroid Viability and Upends Chemoresistance in Triple-Negative Breast Cancer

Introduction

Targeted multimodal approaches need to be strategically developed to control tumour growth and prevent metastatic burden successfully. Breast cancer presents a unique clinical problem because of the variety of cellular subtypes that arise. The tumour stage and cellular subtypes often dictate the appropriate clinical treatment regimen. Also, the development of chemoresistance is a common clinical challenge with breast cancer. Higher doses and additional drug agents can produce additional adverse effects leading to a more aggressive malignancy. Acetylsalicylic acid (ASA), metformin (Met), and oseltamivir phosphate (OP) were investigated for their efficacy to sensitize MDA-MB-231 triple-negative breast cancer and its tamoxifen (Tmx) resistant variant (MDA-MB-231-TmxR) together in combination with Tmx treatment.

Methods

Microscopic imaging, the formation of 3D multicellular tumour spheroids, immunocytochemistry, flow cytometry, Annexin V Assay, Caspase 3/7 Apoptosis Assay, tube formation assay and analysis, and WST-1 cell viability assay evaluated the formation of MCTS, morphologic changes, cell viability, apoptosis activity and the expression levels of ALDH1A1, CD44 and CD24 on the cell surface, MDA-MB231 triple-negative breast cancer, tamoxifen (Tmx) resistant variant (MDA-MB-231-TmxR).

Results

The results using a triple combination of ASA, Met and OP on MDA-MB-231 and MDA-MB-231-TmxR cells and their matrix-free 3D multicellular tumour spheroids (MCTS) formed by using the cyclic Arg-Gly-Asp-D-Phe-Lys peptide modified with 4-carboxybutyl-triphenylphosphonium bromide (cyclo-RGDfK(TPP)) peptide method demonstrate a consistent and significant decrease in cell and tumour spheroid viability and volume with increased apoptotic activity, and increased sensitivity to Tmx therapy. Tmx treatment of MDA-MB-231 cells in combination with ASA, Met and OP markedly reduced the CD44/CD24 ratio by 6.5-fold compared to the untreated control group. Tmx treatment of MDA-MB-231-TmxR cells in combination with ASA, Met and OP markedly reduced the ALDH1A1 by 134-fold compared to the same treatment for the parental cell line. Also, the triple combination treatment of ASA, Met, and OP inhibited vasculogenic endothelial cell tube formation and induced endothelial cell apoptosis.

Conclusion

For the first time, the findings demonstrate that repurposing ASA, Met, and OP provides a novel and promising targeted multimodal approach in the treatment of triple-negative breast cancer and its chemoresistant variant.

Impact of Fucosylation on Self-Assembly of Prostate and Breast Tumor Spheroids by Using Cyclo-RGDfK(TPP) Peptide and Image Object Detection

Introduction

Core fucosylation of N-glycans on the integrin β1 subunit is essential for the functional activity of the integrin. The binding of α5β1 integrin with the tripeptide Arg-Gly-Asp (RGD) motif within the extracellular matrix protein fibronectin may be influenced by the α-1,6-fucose core or α-1,2-fucose and α-1,3/4-fucose peripheral N-glycan profiles. Here, we investigated whether fucosylation impacts the formation of matrix-free 3D multicellular tumor spheroids (MCTS) from human triple negative breast MDA-MB231 cell line, prostate PC3 and DU145 cell lines and DU145 gemcitabine resistant (GemR) variant by using the cyclic Arg-Gly-Asp-D-Phe-Lys peptide modified with 4-carboxybutyl-triphenylphosphonium bromide (cyclo-RGDfK(TPP)) peptide method.

Methods

Microscopic imaging, lectin histochemistry, flow cytometry, WST-1 cell viability assay and You Only Look Once version 2 (YOLOv2) training object detection using cyclic learning rates were used to evaluate the formation of MCTS, morphologic changes, and the expression levels of α-1,6-fucose and α-1,2-fucose linkages on the cell surface.

Results

DU145 prostate cancer cells expressed higher α-1,6-fucose than α-1,2-fucose linkages on their cell surface, as determined by lectin cytochemistry and flow cytometry. Blockage of the α-1,6- and α-1,2-fucose linkages with Aspergillus oryzae lectin (AOL) and Ulex Europaeus agglutinin I (UEA I) one hour before the addition of cyclic-RGDfK(TPP) peptide to the monolayer of the cancer cells resulted in a statistically significant dose-dependent reduction in spheroid volumes using threshold diameters of 40 and 60 µm. Application of a 40 µm threshold diameter measurements of spheroids resulted in fewer false-positive ones compared to the 60 µm diameter threshold previously used in our studies. A state-of-the-art, image object detection system YOLOv2 was used to automate the analysis of spheroid measurements and volumes. The results showed that YOLOv2 corroborated manual spheroid detection and volume measurements with high precision and accuracy.

Conclusion

For the first time, the findings demonstrate that α-1,6- and α-1,2-fucose linkages of N-glycans on the cell surface receptors facilitate cyclo-RGDfK(TPP)-mediated self-assembly of cancer cells to form 3D multicellular tumor spheroids.

Microenvironment-regulated lncRNA-HAL is able to promote stemness in breast cancer cells

Multicellular Tumor Spheroids culture (MCTS) is an in vitro model mimicking the characteristics of the tumor microenvironment, such as hypoxia and acidosis, resulting in the presence of both proliferating and quiescent cell populations. lncRNA's is a novel group of regulatory molecules that participates in the acquisition of tumorigenic phenotypes. In the present work we evaluated the oncogenic association of an uncharacterized lncRNA (lncRNA-HAL) in the tumorigenic phenotype induced by the MCTS microenvironment. We measured lncRNA-HAL expression level in MCF-7-MCTS populations and under different hypoxic conditions by RT-qPCR. Afterwards, we silenced lncRNA-HAL expression by shRNAs and evaluated its effect in MCF-7 transcriptome (by RNAseq) and validated the modified cellular processes by proliferation, migration, and stem cells assays. Finally, we analyzed which proteins interacts with lncRNA-HAL by ChIRP assay, to propose a possible molecular mechanism for this lncRNA. We found that lncRNA-HAL is overexpressed in the internal quiescent populations (p27 positive populations) of MCF-7-MCTS, mainly in the quiescent stem cell population, being hypoxia one of the microenvironmental cues responsible of its overexpression. Transcriptome analysis of lncRNA-HAL knockdown MCF7 cells revealed that lncRNA-HAL effect is associated with proliferation, migration and cell survival mechanisms; moreover, lncRNA-HAL silencing increased cell proliferation and impaired cancer stem cell proportion and function, resulting in decreased tumor grafting in vivo. In addition, we found that this lncRNA was overexpressed in triple-negative breast cancer patients. Analysis by ChIRP assay showed that this nuclear lncRNA binds to histones and hnRNPs suggesting a participation at the chromatin level and transcriptional regulation. The results obtained in the present work suggest that the function of lncRNA-HAL is associated with quiescent stem cell populations, which in turn is relevant due to its implications in cancer cell survival and resistance against treatment in vivo. Altogether, our data highlights a new lncRNA whose expression is regulated by the tumor microenvironment and associated to stemness in breast cancer.

Cationic Oligospermine-Oligonucleotide Conjugates Provide Carrier-free Splice Switching in Monolayer Cells and Spheroids

We report the evaluation of 18-mer 2′-O-methyl-modified ribose oligonucleotides with a full-length phosphorothioate backbone chemically conjugated at the 5′ end to the oligospermine units (S_n-: n = 5, 15, 20, 25, and 30 [number of spermine units]) as splice switching oligonucleotides (SSOs). These conjugates contain, in their structure, covalently linked oligocation moieties, making them capable of penetrating cells without transfection vector. In cell culture, we observed efficient cytoplasmic and nuclear delivery of fluorescein-labeled S₂₀-SSO by fluorescent microscopy. The SSO conjugates containing more than 15 spermine units induced significant carrier-free exon skipping at nanomolar concentration in the absence and in the presence of serum. With an increasing number of spermine units, the conjugates became slightly toxic but more active. Advantages of these molecules were particularly demonstrated in three-dimensional (3D) cell culture (multicellular tumor spheroids [MCTSs]) that mimics living tissues. Whereas vector-complexed SSOs displayed a drastically reduced splice switching in MCTS compared with the assay in monolayer culture, an efficient exon skipping without significant toxicity was observed with oligospermine-grafted SSOs (S₁₅- and S₂₀-SSOs) transfected without vector. It was shown, by flow cytometry and confocal microscopy, that the fluorescein-labeled S₂₀-SSO was freely diffusing and penetrating the innermost cells of MCTS, whereas the vector-complexed SSO penetrated only the cells of the spheroid’s outer layer.

3D tissue-engineered model of Ewing's sarcoma

Despite longstanding reliance upon monolayer culture for studying cancer cells, and numerous advantages from both a practical and experimental standpoint, a growing body of evidence suggests that more complex three-dimensional (3D) models are necessary to properly mimic many of the critical hallmarks associated with the oncogenesis, maintenance and spread of Ewing's sarcoma (ES), the second most common pediatric bone tumor. And as clinicians increasingly turn to biologically-targeted therapies that exert their effects not only on the tumor cells themselves, but also on the surrounding extracellular matrix, it is especially important that preclinical models evolve in parallel to reliably measure antineoplastic effects and possible mechanisms of de novo and acquired drug resistance. Herein, we highlight a number of innovative methods used to fabricate biomimetic ES tumors, encompassing both the surrounding cellular milieu and the extracellular matrix (ECM), and suggest potential applications to advance our understanding of ES biology, preclinical drug testing, and personalized medicine.

Necrosis in DU145 prostate cancer spheroids induces COX-2/mPGES-1-derived PGE2 to promote tumor growth and to inhibit T cell activation

Cyclooxygenase (COX)-2-derived prostaglandin E2 (PGE2 ) supports the growth of a spectrum of cancers. The potential benefit of COX-2-inhibiting non-steroidal anti-inflammatory drugs (NSAIDs) for cancer treatment is however limited by their well-known cardiovascular side-effects. Therefore, targeting microsomal PGE synthase 1 (mPGES-1), the downstream enzyme in the COX-2-dependent pathway of PGE2 production might be attractive, although conflicting data regarding a potential tumor-supporting function of mPGES-1 were reported. We determined the impact of mPGES-1 in human DU145 prostate cancer cell growth. Surprisingly, knockdown of mPGES-1 did not alter growth of DU145 monolayer cells, but efficiently inhibited the growth of DU145 multicellular tumor spheroids (MCTS). Opposed to MCTS, monolayer cells did not secrete PGE2 due to a lack of COX-2 expression, which was induced during spheroid formation. Pharmacological inhibition of COX-2 and mPGES-1 supported the crucial role of PGE2 for growth of MCTS. The functionality of spheroid-derived PGE2 was demonstrated by its ability to inhibit cytotoxic T cell activation. When investigating mechanisms of spheroid-induced COX-2 induction, we observed that among microenvironmental factors neither glucose deprivation, hypoxia nor tumor cell apoptosis enhanced COX-2 expression. Interestingly, interfering with apoptosis in spheroids triggered a shift towards necrosis, thus augmenting COX-2 expression. We went on to demonstrate that necrotic cells induced COX-2 mRNA expression and PGE2 secretion from live tumor cells. In conclusion, necrosis-dependent COX-2 upregulation in MCTS promoted PGE2 -dependent tumor growth and inhibited activated cytotoxic T cells. Hence, blocking mPGES-1 as a therapeutic option may be considered for COX-2/mPGES-1-positive solid cancers.