Advancements in Breast Cancer Research: A Comprehensive Overview of Patient-Derived Models and Precision Oncology
The development of effective cancer therapies is often hindered by the immense heterogeneity of human cancers and the variability in treatment responses. To address these challenges, there is a pressing need for model systems that accurately replicate the complexity of human tumors and the diversity of treatment outcomes. Patient-derived xenografts (PDX) and patient-derived organoids (PDO) are emerging as valuable tools in cancer research, offering the potential to advance drug development and personalized treatment strategies.
Patient-Derived Xenografts (PDX): Modeling Human Tumors In Vivo
PDX models involve the implantation of human tumor fragments directly into immune-deficient mice, allowing the tumors to grow and be serially transplanted. These models have demonstrated a remarkable ability to recapitulate human tumors with high fidelity, exhibiting treatment responses that align with those observed in the original patients. While not perfect, PDX models currently represent the most reliable method for modeling diverse human tumors in a laboratory setting.
PDX models serve as invaluable tools for preclinical, co-clinical, and clinical research studies. They have been instrumental in pioneering precision medicine programs, such as The Mouse Hospital and Co-Clinical Trial Project, which aim to tailor patient-specific therapies for various cancer types. Furthermore, PDX models facilitate investigations into drug responses and resistance mechanisms, the study of tumor heterogeneity and evolution, and the modeling of metastatic disease.
However, PDX-based studies face limitations due to their high cost and low throughput, highlighting the need for complementary approaches.
Patient-Derived Organoids (PDO): Three-Dimensional Models for Enhanced Representation
The development of three-dimensional (3D) organoid models from patient tumors and PDX models is becoming increasingly feasible for several solid tumor types. PDOs offer a more representative model of human cancer compared to traditional two-dimensional (2D) cultures. Accumulating evidence suggests a strong biological concordance between PDOs and the tumors from which they originate.
Read also: University of Georgia Sorority Guide
PDOs have been extensively used to model pancreatic and colorectal tumors, demonstrating their utility in predicting therapeutic responses and facilitating precision medicine for patients. Additionally, PDOs or PDX-derived organoids (PDxO) have been successfully developed for various other cancer types, including hepatocellular carcinoma, hepatoblastoma, glioblastoma, prostate, bladder, ovarian, breast, gastric, lung, esophageal, kidney, and head and neck cancers.
The Role of Genomic Testing in Precision Oncology
Genomic testing is gaining prominence in precision oncology, enabling the personalization of cancer therapy for improved outcomes. Molecular analysis of patients with diverse malignancies has revealed that a significant proportion of patients harbor mutations that match to at least one drug, including combination therapies. Patients receiving treatments aligned with their molecular profiles have demonstrated longer progression-free survival times compared to those receiving physician's choice of drug.
However, functional drug testing using patient-derived models may offer distinct advantages over relying solely on genomics for personalized therapy. Studies have shown that genomics alone may identify viable therapeutic options for a limited percentage of patients with advanced disease, whereas organoids or PDX can be grown from a larger proportion of cases. In proof-of-concept studies, combined genomic testing and functional drug testing have identified effective targeted agents and combinations, highlighting the potential of functional screening to reveal different drug responses despite similar driver mutations.
Addressing the Challenges in Breast Cancer Therapy
Breast cancer poses particular challenges in identifying successful therapies based on genomic alterations. Despite the discovery of numerous genetic and epigenetic drivers, the molecular heterogeneity of metastatic breast cancer hinders the identification and development of effective targeted therapies. While clinically-actionable mutations can be identified in a significant percentage of cases, matching therapies to these variants has not consistently translated into clinical benefits.
These findings underscore the need for parallel functional modeling of candidate driver gene dependence and drug response to improve outcomes for breast cancer patients. The development of a larger biobank of advanced breast cancer models, along with in vitro methods to propagate these tumors for more feasible experimental manipulation, is crucial for accelerating progress in understanding sensitivity and resistance to various therapies across diverse breast cancer subtypes.
Read also: History of the Block 'M'
Generation of New Breast Cancer PDX Models
Researchers continue to generate breast cancer PDX models, focusing on those representing the greatest unmet medical and research needs, including:
- ER+ endocrine-resistant tumors
- ER+ and HER2+ co-expressing tumors
- Unusually aggressive tumor types (e.g., metaplastic breast cancer)
- Drug-resistant tumors
- PDXs representing primary-metastatic pairs or longitudinal collections from the same patient
These models provide valuable resources for studying the complexities of breast cancer and developing more effective treatment strategies.
Credentialing and Characterization of PDX Models
Each PDX line undergoes a rigorous credentialing process to ensure its quality and reliability. This process includes testing for human and mouse pathogens, validating the expression of breast epithelial markers and human mitochondria, and confirming the absence of lymphoma markers. ER, PR, and HER2 staining is conducted and compared to the original tumor material to ensure concordant results.
PDX models are also characterized through DNA mutation and copy number variant (CNV) analysis, as well as RNA sequencing. Genomic analysis reveals that PDX models reflect the heterogeneity of human breast cancer with respect to both driver mutations and intrinsic subtypes.
ScreenDL: A Deep Learning Model for Cancer Drug Response Prediction
To address the limitations of current biomarker-based treatment selection strategies, researchers have developed ScreenDL, a novel deep learning (DL)-based cancer drug response prediction model. ScreenDL leverages the combination of tumor omic and functional drug screening data to predict the most efficacious treatments.
Read also: Legacy of Fordham University
The model learns a function that maps a drug's chemical structure and a tumor's transcriptomic profile to a predicted response. Drugs are encoded as Morgan fingerprints, while tumors are represented as vectors encoding the z-score normalized expression of hallmark genes from the Molecular Signatures Database (MSigDB) hallmark gene set collection.
ScreenDL undergoes a three-phase training process:
- Pretraining: Extracts generalizable associations between tumor omics and drug response from large-scale cell line pharmaco-omic databases.
- Domain-Specific Fine-Tuning: Integrates context-specific pharmaco-omic data derived from more clinically relevant PDXOs, adapting the model to a more clinically relevant response prediction context.
- Patient-Specific Fine-Tuning: Utilizes the ScreenAhead module to integrate patient-level transcriptomic features with limited functional drug screening in PDMCs derived from the patient, generating a personalized response prediction model.
ScreenDL has demonstrated superior performance compared to existing DL models across a panel of clinically relevant benchmarks in cell lines and PDMCs. Personalization with ScreenAhead further improves response prediction, highlighting the power of this combined computational/experimental strategy.
Gene.iobio: A Real-Time Platform for Variant Interrogation and Prioritization
Gene.iobio is a real-time, intuitive, and interactive web application designed for clinically-driven variant interrogation and prioritization. This platform empowers clinical care providers to directly interact with patient genomic data, enabling sophisticated genomic analyses that were previously accessible only through complex command-line tools.
Gene.iobio accepts file-format compliant indexed BAM/CRAM and indexed (unannotated or annotated) VCF files, allowing users to analyze exome and genome sequencing data in real-time. The platform provides visual summaries of pertinent variant annotations, such as biological impact, population frequency, ClinVar assertion, REVEL score, and evolutionary conservation, with visual cues for how each annotation might contribute to pathogenicity.
Gene.iobio integrates numerous public datasets to present up-to-date gene and variant annotations, including ClinVar, gnomAD, NCBI E-utilities, UCSC, GENCODE, and RefSeq. The platform also provides links to external resources such as MARRVEL, VarSome, OMIM, DECIPHER, GeneCards, and GTEx.
tags: #Yi #Qiao #University #of #Utah #history
