SEARCH · PubMed Health

Results for “Drug response prediction”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 recordsLinked to original sources

PubMed2025

ASGCL: Adaptive Sparse Mapping-based graph contrastive learning network for cancer drug response prediction.

Personalized cancer drug treatment is emerging as a frontier issue in modern medical research. Considering the genomic differences among cancer patients, determining the most effective drug treatment plan is a complex and crucial task. In response to these challenges, this study introduces the Adaptive Sparse Graph Contrastive Learning Network (ASGCL), an innovative approach to unraveling latent interactions in the complex context of cancer cell lines and drugs. The core of ASGCL is the GraphMorpher module, an innovative component that enhances the input graph structure via strategic node attribute masking and topological pruning. By contrasting the augmented graph with the original input, the model delineates distinct positive and negative sample sets at both node and graph levels. This dual-level contrastive approach significantly amplifies the model's discriminatory prowess in identifying nuanced drug responses. Leveraging a synergistic combination of supervised and contrastive loss, ASGCL accomplishes end-to-end learning of feature representations, substantially outperforming existing methodologies. Comprehensive ablation studies underscore the efficacy of each component, corroborating the model's robustness. Experimental evaluations further illuminate ASGCL's proficiency in predicting drug responses, offering a potent tool for guiding clinical decision-making in cancer therapy.

Humans↗

PubMed2025

Anticancer drug response prediction integrating multi-omics pathway-based difference features and multiple deep learning techniques.

Individualized prediction of cancer drug sensitivity is of vital importance in precision medicine. While numerous predictive methodologies for cancer drug response have been proposed, the precise prediction of an individual patient's response to drug and a thorough understanding of differences in drug responses among individuals continue to pose significant challenges. This study introduced a deep learning model PASO, which integrated transformer encoder, multi-scale convolutional networks and attention mechanisms to predict the sensitivity of cell lines to anticancer drugs, based on the omics data of cell lines and the SMILES representations of drug molecules. First, we use statistical methods to compute the differences in gene expression, gene mutation, and gene copy number variations between within and outside biological pathways, and utilized these pathway difference values as cell line features, combined with the drugs' SMILES chemical structure information as inputs to the model. Then the model integrates various deep learning technologies multi-scale convolutional networks and transformer encoder to extract the properties of drug molecules from different perspectives, while an attention network is devoted to learning complex interactions between the omics features of cell lines and the aforementioned properties of drug molecules. Finally, a multilayer perceptron (MLP) outputs the final predictions of drug response. Our model exhibits higher accuracy in predicting the sensitivity to anticancer drugs comparing with other methods proposed recently. It is found that PARP inhibitors, and Topoisomerase I inhibitors were particularly sensitive to SCLC when analyzing the drug response predictions for lung cancer cell lines. Additionally, the model is capable of highlighting biological pathways related to cancer and accurately capturing critical parts of the drug's chemical structure. We also validated the model's clinical utility using clinical data from The Cancer Genome Atlas. In summary, the PASO model suggests potential as a robust support in individualized cancer treatment. Our methods are implemented in Python and are freely available from GitHub (https://github.com/queryang/PASO).

Deep Learning↗

PubMed2026

A Knowledge-Enhanced Multimodal Framework with Genomic Reconstruction for DLBCL Drug Response Prediction.

Diffuse large B-cell lymphoma (DLBCL) exhibits substantial biological heterogeneity, leading to pronounced variability in patient response to therapy. Accurate drug response prediction is therefore critical for precision treatment but remains challenging in clinical settings where genomic sequencing, a highly informative modality, is frequently incomplete. Existing methods, often developed from cell-line pharmacogenomic datasets or single-modality data, typically assume fully observed molecular profiles and thus show limited robustness under missing genomic data. To address this limitation, a knowledge-enhanced multimodal framework with genomic reconstruction (KeM-DRP) is proposed for individualized drug response prediction in DLBCL. The framework models the central role of genomics by integrating biological prior knowledge through a gene-pathway-biological process hierarchy, enabling robust representation learning from sparse observations. To compensate for missing genomic measurements, a cross-modal genomic compensation module reconstructs genomically informed latent features from routinely available clinical modalities. Furthermore, a genomics-guided adaptive fusion strategy dynamically integrates heterogeneous modalities conditioned on observed or reconstructed genomic representation. Experiments on a real-world DLBCL cohort demonstrate that KeM-DRP consistently outperforms competitive baselines. The reconstructed genomic representation represents most predictive utility, highlighting the robustness and practical value of the framework under incomplete genomic data.

Journal Article↗

PubMed2025

Transfer learning with multiomics integration and deep neural networks reveals drug resistance mechanisms in cancer.

Drug resistance remains one of the primary challenges in effective cancer therapy. In this study, we employed a deep neural network (DNN)-based transfer learning (TL) approach to predict drug response and uncover drug resistance mechanisms. We integrated gene expression, somatic mutation, and copy number aberration (CNA) data with drug response profiles using multi-omics integration (MI). We used the Genomics of Drug Sensitivity in Cancer (GDSC) data for training and incorporated drugs with same pathways into the training models. We then evaluated drug response predictions on independent in-vivo PDX Encyclopedia (PDX) and ex-vivo the Cancer Genome Atlas (TCGA) datasets. In addition, we conducted pathway enrichment analyses to elucidate the mechanisms underlying drug resistance for paclitaxel, 5-fluorouracil (5-FU), gemcitabine, and cetuximab. We also applied Fisher's exact test (FET) to assess potential associations between drug resistance and the presence of mutations or CNAs. Our pan-drug models outperformed other methods based on the area under the precision-recall curve (AUCPR). Our pathway enrichment analyses revealed LDHB-mediated pyruvate metabolism and FYN-mediated focal adhesion might have pivotal roles in paclitaxel resistance, while PINK1-mediated mitophagy might be critical in 5-FU resistance. In addition to transcriptional activation, FET suggested that CNAs in LDHB and PINK1 may also be associated with resistance to paclitaxel and 5-FU, respectively. Furthermore, enrichment results for paclitaxel and cetuximab indicated shared resistance mechanisms between the two drugs. Importantly, our findings are consistent with prior experimental studies, providing literature-based validation of our results. Overall, our DNN-based TL approach achieved strong predictive performance across PDX & TCGA datasets and enrichment analyses provided valuable biological insights into drug resistance mechanisms.

Humans↗

PubMed2026

ZUP1 as a Novel Potential Oncogenic Driver and Prognostic Biomarker in Breast Cancer.

INTRODUCTION: Breast cancer is one of the main causes of cancer death in women globally. Identifying new predictive markers and therapeutic targets is important for improving patient outcomes. Zinc finger-containing U-rich RNA-binding protein 1 (ZUP1) is an RNA-binding protein containing a zinc finger structure that has not been systematically analyzed in breast cancer research. MATERIALS AND METHODS: The study used data from 1,231 samples from the Cancer Genome Atlas (TCGA) database. The ZUP1 expression in tumor tissues and normal tissues was compared. Its predictive value was assessed using survival analysis and regression models. Its biological role was explored through gene functional analysis. The immune cell analysis method was used to study the tumor immune environment, and the drug susceptibility database was used to predict drug responses. Predictive models were also built and validated. RESULTS: ZUP1 expression was significantly higher in breast cancer tissues than in normal tissues. High expression of ZUP1 is related to advanced tumor stage and is an independent indicator of poor survival prognosis in univariate and multivariate analyses. Functional enrichment revealed that ZUP1 is closely linked to cell cycle progression, DNA replication, and the Fanconi anemia (FA) pathway. Immune infiltration analysis demonstrated a significant negative link between ZUP1 levels and the abundance of resting mast cells and activated NK cells. Furthermore, high ZUP1 expression was associated with increased sensitivity to several targeted therapies, including Nutlin-3a and PD-0325901. A clinically applicable nomogram combining ZUP1 expression with key clinical factors (age, stage, T, N, M) was developed to predict 3- and 5-year OS with good calibration and discrimination. DISCUSSION: Our study identifies ZUP1 as a potential oncogenic factor and a robust independent prognostic biomarker in breast cancer. Its involvement in critical cellular processes and modulation of the tumor immune microenvironment highlights its potential as a novel therapeutic target. Functional experiments, including immunohistochemical staining and CCK8 proliferation assays, further supported the oncogenic role of ZUP1. The established nomogram provides a valuable tool for personalized risk assessment and clinical decision-making. CONCLUSION: Our findings suggest that ZUP1 is a novel multifaceted biomarker with significant implications for personalized treatment strategies in breast cancer.

ZUP1↗

PubMed2025

Understanding the sources of performance in deep drug response models reveals insights and improvements.

MOTIVATION: Anti-cancer drug response prediction (DRP) using cancer cell lines (CLs) is crucial in stratified medicine and drug discovery. Recently, new deep learning models for DRP have improved performance over their predecessors. However, different models use different input data types and architectures making it hard to find the source of these improvements. Here we consider published DRP models that report state-of-the-art performance predicting continuous response values. These models take chemical structures of drugs and omics profiles of CLs as input. RESULTS: By experimenting with these models and comparing with our simple baselines, we show that no performance comes from drug features, instead, performance is due to the transcriptomics CL profiles. Furthermore, we show that, depending on the testing type, much of the current reported performance is a property of the training target values. We address these limitations by creating BinaryET and BinaryCB that predict binary drug response values, guided by the hypothesis that this reduces the noise in the drug efficacy data. Thus, better aligning them with biochemistry that can be learnt from the input data. BinaryCB leverages a chemical foundation model, while BinaryET is trained from scratch using a transformer-type architecture. We show that these models learn useful chemical drug features, which is the first time this has been demonstrated for multiple testing types to our knowledge. We further show binarizing the drug response values causes the models to learn useful chemical drug features. We also show that BinaryET improves performance over BinaryCB, and the published models that report state-of-the-art performance. AVAILABILITY AND IMPLEMENTATION: Code is available from https://github.com/Nik-BB/Understanding_DRP_models.

Humans↗

PubMed2025

Personalizing chemotherapy drug selection using a novel transcriptomic chemogram.

Gene signatures predictive of chemotherapeutic response have the potential to extend the reach of precision medicine by allowing oncologists to optimize treatment for individuals. Most published predictive signatures are only capable of predicting response for individual drugs, but most chemotherapy regimens utilize combinations of different agents. We propose a unified framework, called the chemogram, that uses predictive signatures to rank the relative predicted sensitivity of different drugs for individual tumors. Using this approach, providers could efficiently screen against many therapeutics to optimize chemotherapy at any time, whether it be for a treatment-naive tumor or a chemo-resistant tumor requiring a new treatment strategy. To demonstrate the utility of the chemogram, we used predictive signatures (extracted from a previously established method) in our framework to rank predicted sensitivity among drugs within cell lines. We then compared the rank order of predicted and observed response against each drug. Across most cancer types, chemogram-generated predictions were more accurate than predictions made by randomly generated gene signatures, signatures extracted from differential expression alone, and was comparable to another established method of drug response prediction. Our framework demonstrates the ability of transcriptomic signatures to not only predict chemotherapeutic response, but also correctly assign rankings of drug sensitivity on an individual basis. Additionally, scaling the chemogram to include more drugs does not compromise accuracy.

Humans↗

PubMed2026

PLK1/FOXM1-associated tumor-cell state and macrophage-related immune features in endometrial cancer.

BACKGROUND: Polo-like kinase 1 (PLK1) and forkhead box M1 (FOXM1) have been widely studied in various cancers; however, their expression characteristics in endometrial cancer (EC) and their potential association with tumor microenvironment remodeling remain insufficiently characterized. METHODS: This study integrated The Cancer Genome Atlas uterine corpus endometrial carcinoma cohort, Gene Expression Omnibus, pan-cancer transcriptomic data, Human Protein Atlas/Clinical Proteomic Tumor Analysis Consortium, and local immunohistochemistry data to evaluate PLK1 expression and clinicopathological relevance across transcriptomic, proteomic, and histopathological data. Differential expression, survival, gene-set enrichment, transcription-factor enrichment, and immune-infiltration analyses characterized PLK1-associated features. In vitro experiments combined EC cell lines AN3CA and HEC-1A with co-immunoprecipitation, Western blotting, Transwell assays, and a THP-1 conditioned-medium model. Drug-response prediction and structure-based analysis prioritized candidate therapeutic hypotheses. RESULTS: PLK1 was consistently upregulated at both mRNA and protein levels in EC and was associated with higher tumor grade and International Federation of Gynecology and Obstetrics (FIGO) stage. In survival analysis, higher PLK1 expression was associated with poorer overall survival in univariable models but not after adjustment for age, tumor grade, and FIGO stage. Functional enrichment analysis showed that PLK1-associated genes were mainly involved in cell-cycle and mitotic processes. FOXM1 was identified as a potential candidate component of the PLK1-associated transcriptional program and was positively correlated with PLK1 expression and cell-cycle-related features. In vitro experiments supported an interaction between PLK1 and FOXM1 and suggested that FOXM1 Thr600 phosphorylation-related alterations were associated with migration and invasion phenotypes. Furthermore, the PLK1/FOXM1-associated tumor-cell state was linked to macrophage-related immune features and changes in the M2-like marker profile of THP-1-derived macrophage-like cells. Drug response analyses suggested differential predicted sensitivity patterns in PLK1-high tumors, providing candidate therapeutic hypotheses for further validation. CONCLUSION: The PLK1/FOXM1-associated tumor-cell state may represent a distinct molecular feature associated with proliferative activity, invasive phenotypes, and macrophage-related immune features in EC. This study provides preliminary evidence supporting the biological relevance of this molecular feature and highlights potential therapeutic directions for future investigation.

FoxM1↗

PubMed1978

Hyperkinetic/aggressive boys in treatment: predictors of clinical response to methylphenidate.

Data on 84 nonretarded boys aged 6--12 with the hyperkinetic/MBD syndrome were drawn from a comprehensive, longitudinal investigation in the interest of identifying factors that contribute significantly to variation in clinically rated improvement during treatment with methylphenidate. The size of the multiple correlation (R = .50) indicates that 25% of the variation in the children's response to methylphenidate is jointly predictable from age at referral, degree of perinatal complications, and score on the hyperactivity factor. The authors discuss stepwise multiple regression analysis as the method of choice in drug response prediction studies and the possible effects on the results of such studies of differing definitions of improvements.

Age Factors↗

PubMed2026

GBFN: A gated bimodal fusion network leveraging foundation model embeddings for cancer drug sensitivity prediction.

Despite recent progress in deep learning for cancer drug sensitivity prediction, many existing models still rely on task-specific representation learning or relatively simple multimodal fusion, which may limit their ability to capture complex drug-cell interactions. To address this issue, we developed GBFN, a gated bimodal fusion network for continuous IC50 prediction that integrates pretrained drug and cell-line representations. Specifically, drug embeddings were obtained from SMI-TED, whereas cell-line embeddings were derived from transcriptomic profiles using BulkFormer. These two modalities were then combined through a dimension-wise gated fusion module and used to predict IC50 values in matched drug-cell line pairs. On the CCLE-based benchmark, GBFN outperformed representative neural baselines, including GraphDRP, TGSA, and TransEDRP, and achieved the best overall performance, with an R² of 0.8714 and an RMSE of 0.8938. Moreover, ablation analysis showed that the model using drug features and cell-line expression data with gated fusion performed better than the corresponding model using direct concatenation, indicating that the improvement was associated with the fusion strategy rather than with the input modalities alone. In addition, cell-line expression data were more informative than mutation data in the present setting, and adding mutation data to the model using drug features and expression data did not further improve performance. Across major cancer types, GBFN maintained generally high cell-line-level predictive performance, and perturbation-based attribution identified biologically relevant transcriptomic programs in selected drug-cell line settings. Together, these findings support GBFN as a compact and effective framework for continuous drug response prediction.

Humans↗

PubMed2026

Multi-omics characterization of a GPRC5A+ epithelial subpopulation associated with malignant features in colorectal cancer.

BACKGROUND: Colorectal cancer (CRC) exhibits marked cellular heterogeneity, and the cellular context of malignancy-associated epithelial programs remains incompletely defined. METHODS: We integrated 2,993 CRC samples spanning bulk RNA-seq (n = 2,568; two OS/RFS cohorts), scRNA-seq (281,961 cells/152 specimens), spatial transcriptomics (n = 6), and proteomics (n = 267). Analyses included single-cell integration/annotation, GSVA/HALLMARK, interactome, pseudotime, and ligand-receptor mapping; functional CRISPR assays, EMT immunoblotting, and xenografts; TF profiling (SCENIC/JASPAR/ChIP-qPCR); and exploratory drug-response prediction (OncoPredict), cell-sensitivity assays, and docking/MD modeling. RESULTS: We constructed a stage-stratified single-cell atlas and resolved eleven malignant epithelial subsets, characterizing Epi_4 as late-stage-enriched with EMT, hypoxia, and inflammatory programs and adverse OS/RFS. GPRC5A marked this subset, which we define as GPRC5A+Epi; its expression rose from stage I→IV and was associated with poor outcomes across cohorts, with concordant spatial/proteomic observations. GPRC5A perturbation affected CRC proliferation, migration/invasion, EMT, and xenograft tumorigenicity, supporting a functionally important role in the tested models. SCENIC and ChIP-qPCR supported FOSL1 as an upstream regulator that occupies the GPRC5A promoter. Spatial and ligand-receptor analyses predicted close association and potentially reciprocal signaling between GPRC5A+Epi and POSTN+fibroblasts (COL1A1-SDC4, COL1A1/1A2-ITGA2/ITGB1, PPIA-BSG); concurrent high GPRC5A+Epi/POSTN+Fib signatures were associated with inferior OS/RFS. Drug-response analyses identified an association between GPRC5A status and trametinib sensitivity. Docking/MD produced a computational model of a possible trametinib-GPRC5A interaction, which remains experimentally unvalidated. CONCLUSIONS: GPRC5A⁺Epi is a malignancy-associated epithelial state in CRC, and GPRC5A is functionally important for malignant phenotypes in the tested models. Its inferred relationships with POSTN⁺ fibroblasts and the trametinib findings should be regarded as hypothesis-generating pending functional crosstalk, direct-binding, and therapeutic validation.

Humans↗

PubMed2026

scRNA-seq and bulk RNA-seq reveal the characteristics of macrophage copper metabolism and establish a risk signature in hepatocellular carcinoma.

BACKGROUND: Hepatocellular carcinoma (HCC) is a prevalent malignancy with an urgent need for improved prognostic stratification and treatment-response prediction. This study aimed to explore a macrophage copper metabolism-associated prognostic model and to investigate the relationship between this risk model and the tumor immune microenvironment. METHODS: The FindClusters function was used to analyze cell clusters, and CellChat and CellPhoneDB/LIANA were employed for cell-cell communication analysis. Copper metabolism-related genes were sourced from the MSigDB database. A prognostic risk model was established using least absolute shrinkage and selection operator (LASSO) analysis and multivariate Cox regression analysis, and a nomogram was constructed by integrating the prognostic model with clinicopathological factors. Additional analyses were performed to map the seven model genes in single-cell data, assess model uncertainty and robustness, evaluate macrophage/copper/cuproptosis-related transcriptional programs, and examine the correlations between risk score, immune infiltration and predicted drug sensitivity. RESULTS: Using single-cell RNA sequencing (scRNA-seq) data, we identified four macrophage subpopulations. Macrophages with high SPP1 expression showed close interaction with T cell populations and were associated with copper ion metabolism. By incorporating 141 copper metabolism-related genes and using The Cancer Genome Atlas Liver Hepatocellular Carcinoma (TCGA-LIHC) cohort, we constructed a seven-gene risk prediction model. Additional single-cell mapping showed that the model genes were detectable in the HCC single-cell dataset and showed a macrophage-associated expression pattern. The model showed moderate prognostic discrimination in TCGA-LIHC, whereas its external performance was heterogeneous and remained evaluable across external cohorts, with performance varying among datasets. Immune and mechanism-related analyses suggested that the risk signature was associated with macrophage-related infiltration, copper metabolism and cuproptosis-related transcriptional programs. Drug sensitivity analysis nominated Daporinad as a computationally predicted candidate compound, supporting Daporinad as a pharmacogenomic candidate for follow-up investigation. CONCLUSIONS: By integrating scRNA-seq and bulk RNA sequencing (RNA-seq) data, we constructed a macrophage copper metabolism-associated prognostic signature for HCC. The risk score was associated with survival, immune microenvironment features and predicted drug response, providing a transcriptomic framework for risk stratification and therapeutic hypothesis generation.

Hepatocellular carcinoma (HCC)↗

PubMed2025

BOGO: A Proteome-Wide Gene Overexpression Platform for Discovering Rational Cancer Combination Therapies.

Cancer drug resistance remains a major barrier to durable treatment success, often leading to relapse despite advances in precision oncology. While combination therapies are being increasingly investigated, such as chemotherapy with small molecule inhibitors, predicting drug response and identifying rational drug combinations based on resistance mechanisms remain major challenges. Therefore, a proteome-wide, single-gene overexpression screening platform is essential for guiding rational therapy selection. Here, we present BOGO (Bxb1-landing pad human ORFeome-integrated system for a proteome-wide Gene Overexpression), a robust, scalable, and reproducible screening platform that enables single-copy, site-specific integration and overexpression of ~19,000 human open across cancer cell models. Using BOGO, we identified drug-specific response drivers for 16 chemotherapeutic agents and integrated clinical datasets to uncover proliferation and resistance-associated genes with prognostic potential. Drug response similarity networks revealed both shared and unique mechanisms, highlighting key pathways such as autophagy, apoptosis, and Wnt signaling, and notable resistance-associated genes including BCL2, POLD2, and TRADD. In particular, we proposed a synergistic combination of the BCL2 family inhibitor ABT-263 (Navitoclax®) and the DNA analog TAS-102 (Lonsurf®), which revealed that lysosomal modulation is a key mechanism driving DNA analog resistance. This combination therapy selectively enhanced cytotoxicity in colorectal and pancreatic cancer cells in vitro, and demonstrated therapeutic benefit in vivo in both cell line-derived xenograft (CDX) and patient-derived xenograft (PDX) models. Together, these findings establish BOGO as a powerful gene overexpression perturbation platform for systematically identifying chemoresistance and chemosensitization drivers, and for discovering rational combination therapies. Its scalability and reproducibility position BOGO as a broadly applicable tool for functional genomics and therapeutic discovery beyond cancer resistance.

Journal Article↗

PubMed2026

Biological Foundation Models for Complex Disease Research and Clinical Translation.

Complex diseases, including cancer, rare genetic disorders, neurodevelopmental and psychiatric conditions, and neurodegenerative diseases, arise from interactions among genetic variation, gene regulation, and cellular states that are difficult to capture using a single data type or biological scale. Biological foundation models address this challenge by treating nucleotides and genes as tokens and learning representations that can be transferred to downstream biomedical and clinical tasks. In this review, we examine two major model classes, genomic sequence foundation models and cell foundation models, and compare their tokenization strategies, model architectures, pretraining objectives, and adaptation methods. We summarize their emerging applications in regulatory variant interpretation, disease-associated cell-state analysis, drug-response prediction, and therapeutic target discovery across complex diseases. We distinguish applications supported by experimental or retrospective validation from those that remain primarily computational or conceptual. We further discuss key challenges to clinical translation, including multimodal data integration, model interpretability, benchmarking, patient-specific prediction, and privacy protection. We highlight future opportunities to integrate biological foundation models with emerging frameworks of medical digital twins, agentic AI, and federated learning. By linking model design to translational goals, this review provides a practical framework for evaluating biological foundation models and their readiness for complex disease research and clinical use.

biological foundation model↗

PubMed2026

Methylation-Associated Differentiation Features Define Biological and Prognostic Heterogeneity in CMS4 Colorectal Cancer.

Consensus molecular subtype 4 (CMS4) colorectal cancer (CRC) is associated with an aggressive clinical course and poor survival, yet the biological basis of heterogeneity within this subtype remains incompletely understood. DNA methylation is an epigenetic mechanism involved in transcriptional regulation, cellular differentiation, and colorectal tumorigenesis. Here, we integrated single-cell RNA sequencing (scRNA-seq), bulk data, and promoter DNA methylation data to characterize CMS4-associated cancer cell states and methylation-related features. Using the scAB algorithm, we integrated scRNA-seq with bulk CMS4 data and identified CMS4-related cells distributed across multiple patients. Single-cell analyses of cell-cell communication and transcriptional regulation revealed a CMS4-related cancer cell population characterized by macrophage migration inhibitory factor (MIF)-centered intercellular communication, enhanced caudal type homeobox 1 (CDX1) and Kruppel-like factor 5 (KLF5) regulon activity, and gene modules enriched in differentiation-related pathways. CytoTRACE analysis further stratified CMS4 cancer cells into poorly and well-differentiated states, yielding 802 differentially expressed genes (DEGs). Linking these differentiation-associated DEGs with bulk expression and promoter methylation data identified 218 methylation-associated DEGs showing significant inverse methylation expression correlations, suggesting a link between differentiation-related heterogeneity and promoter methylation. Univariable Cox regression followed by LASSO regression further prioritized eight genes for construction of the methylation and differentiation-related prognostic model (MeDiff-PM). MeDiff-PM consistently stratified overall survival in the TCGA CMS4 cohort and two independent validation cohorts, with cutoff-independent continuous Cox analyses further supporting its prognostic association across cohorts. And MeDiff-PM remained prognostically significant after adjustment for available clinical variables. High MeDiff-PM risk scores were associated with activation of P53, WNT, and ubiquitin-mediated proteolysis pathways and with consistent predicted drug response differences for compounds across three CMS4 cohorts. While individual in silico knockout analysis suggested links between MeDiff-PM genes and metallothionein-related and immune-associated transcriptional responses. Collectively, these findings indicate that methylation-associated differentiation features represent a molecular dimension of intra-CMS4 heterogeneity and provide a biologically informed framework for prognostic stratification within CMS4 CRC.

Humans↗

PubMed1978

Heterogeneity in drug sensitivity among tumor cell subpopulations of a single mammary tumor.

Three distinct subpopulations of tumor cells derived from a single parent strain BALB/cfC3H mammary adenocarcinoma were tested in vivo for sensitivity to cyclophosphamide, methotrexate, and 5-fluorouracil. Treatment was begun either 2 days after s.c. tumor cell injection or at the time when the tumors became palpable. It was given on a weekly basis for 4 weeks. The mice were observed for growth of the primary implant and for development of spontaneous metastases. The three subpopulations differed markedly in their sensitivity to the drugs. The effects of the drugs ranged from induction of regression of the "primary" to enhancement of metastases. The effect on primary growth was independent of that on metastasis. The effect of the time of administration of the drugs also varied among the subpopulations. The sublines were also tested in vitro with methotrexate and 5-fluorouracil. Again there were marked differences in sensitivity to inhibition of cell division by the drugs. The relative sensitivities in vitro did not correlate with observations in vivo. The existence of subpopulations of tumor cells, differing in sensitivity to therapeutic agents, within a single neoplasm, presents a challenge to development of assays capable of predicting drug response and to the selection of combination therapies.

Adenocarcinoma↗

PubMed1976

Predicting individual responses to drug treatment in schizophrenia: a test dose model.

The literature and the findings from the Camarillo Schizophrenia Research Project reported in this paper indicate that a satisfactory method for predicting the response of an individual schizophrenic patient to antipsychotic drugs has yet to be devised. A test dose procedure is described which offers promise of a practical approach to selecting the most appropriate drug and dosage for a particular patient and tailoring blood concentrations to the needs of the individual case. Preliminary findings indicate that the test dose procedure is feasible; that detectable changes occur after a single test dose; and that measurements made during the test dose period may be predictive of eventual outcome. These findings are, of course, only a report of a preliminary pilot experiment, subject to important caveats about small number of cases, interpretation of large numbers of correlation coefficients, and need for cross-validation. Nevertheless, they are encouraging and suggest that the test dose approach has considerable potential for further research.

Antipsychotic Agents↗

PubMed1979

Pharmacokinetics of digoxin: relationship between response intensity and predicted compartmental drug levels in man.

A study designed to investigate the relationship between the pharmacokinetics of digoxin and a measure of its pharmacological effect has been conducted. Serum digoxin concentrations and systolic time intervals were measured concurrently in 12 normal male volunteers following a 1.0 mg i.v. bolus injection. The averaged serum digoxin concentration--time and response--time data were analyzed pharmacokinetically using a three-compartment open model and nonlinear least-squares fitting. When only the serum level--time data were analyzed, a close relationship was found between calculated digoxin levels in the slowly distributing (deep) peripheral compartment and response of the heart to digoxin, as measured by changes in the QS2 index (delta QS2I). Although it was not possible to distinguish clearly a linear from a nonlinear relationship between digoxin levels in the deep compartment and delta QS2I, the nonlinear relationship gave the best overall fit when both serum digoxin and delta QS2I data were fitted simultaneously. The simultaneous fit yielded a total body clearance of digoxin of 3.6 ml/min/kg and a terminal t1/2 of 42 hr.

Digoxin↗