Drug Response Research Articles

Improving Anticancer Drug Selection and Prioritization via Neural Learning to Rank.

Personalized cancer treatment requires a thorough understanding of complex interactions between drugs and cancer cell lines in varying genetic and molecular contexts. To address this, high-throughput screening has been used to generate large-scale drug response data, facilitating data-driven computational models. Such models can capture complex drug-cell line interactions across various contexts in a fully data-driven manner. However, accurately prioritizing the most effective drugs for each cell line still remains a significant challenge. To address this, we developed multiple neural ranking approaches that leverage large-scale drug response data across multiple cell lines from diverse cancer types. Unlike existing approaches that primarily utilize regression and classification techniques for drug response prediction, we formulated the objective of drug selection and prioritization as a drug ranking problem. In this work, we proposed multiple pairwise and listwise neural ranking methods that learn latent representations of drugs and cell lines and then use those representations to score drugs in each cell line via a learnable scoring function. Specifically, we developed neural pairwise and listwise ranking methods, Pair-PushC and List-One on top of the existing methods, pLETORg and ListNet, respectively. Additionally, we proposed a novel listwise ranking method, List-All, that focuses on all the effective drugs instead of the top effective drug, unlike List-One. We also provide an exhaustive empirical evaluation with state-of-the-art regression and ranking baselines on large-scale data sets across multiple experimental settings. Our results demonstrate that our proposed ranking methods mostly outperform the best baselines with significant improvements of as much as 25.6% in terms of selecting truly effective drugs within the top 20 predicted drugs (i.e., hit@20) across 50% test cell lines. Furthermore, our analyses suggest that the learned latent spaces from our proposed methods demonstrate informative clustering structures and capture relevant underlying biological features. Moreover, our comprehensive evaluation provides a thorough and objective comparison of the performance of different methods (including our proposed ones).

Mechanistic exploration and experimental validation of the Xiaochaihu decoction for the treatment of breast cancer by network pharmacology.

Xiaochaihu (XCH) decoction is a traditional Chinese prescription that has been recorded in the pharmacopeia of the People's Republic of China. In China, the XCH decoction is used clinically to treat a variety of tumors, including breast cancer. However, its potential mechanism of action is still undefined. The chemical compounds in the XCH decoction were identified via Q Exactive Orbitrap LC-MS/MS. Then, we screened the active ingredients and targets in the XCH decoction from the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP). Next, Cytoscape and Metascape were used to construct an active ingredient-target-disease network, which included a protein-protein interaction (PPI) network, GO enrichment analysis, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. Finally, we used molecular docking and in vitro experiments to verify the results of network pharmacology analysis. More than 70 major compounds were identified by Q Exactive Orbitrap LC-MS/MS analysis from the XCH decoction. A total of 162 active ingredients and 153 targets related to the XCH decoction and breast cancer were identified, and a compound-target-disease network was constructed. GO and KEGG analyses revealed that the XCH decoction regulated the drug response, apoptosis process, cancer pathway, and PI3K/Akt signaling pathway. Molecular docking and experimental validation indicated that the XCH decoction suppressed proliferation and induced apoptosis in breast cancer cells by regulating the expression of apoptosis-related proteins and inhibiting the PI3K/Akt pathway. This study suggested that the XCH decoction can be used to treat breast cancer by inhibiting cell proliferation, inducing apoptosis and downregulating the PI3K/Akt signaling pathway.

Precision oncology in neurofibromatosis type 1: quantification of differential sensitivity to selumetinib in plexiform neurofibromas using single-cell RNA sequencing.

Selumetinib is an FDA-approved targeted therapy for plexiform neurofibromas in neurofibromatosis type 1(NF1) with durable response rates seen in most, but not all patients. In this proof-of-concept study, we demonstrate single-cell RNA sequencing(scRNAseq) as a technique for quantifying drug response to selumetinib at the single cell level. scRNAseq data from neurofibroma biopsies was obtained from a public genomics repository. Schwann cell populations were identified through standard clustering techniques and single-cell selumetinib sensitivity was quantified on a scale of 0(resistant) to 1(sensitive) based on the expression pattern of a 500 gene selumetinib sensitivity signature from the BeyondCell sensitivity library. A total of seven plexiform neurofibromas were included in our final analysis. The median absolute number of Schwann cells across samples was 658 cells (IQR: 1,029 cells, Q1-Q3: 135 cells to 1,163 cells). There was a statistically significant difference in selumetinib sensitivity profiles across samples (p < 0.001). The tumor with the highest median selumetinib sensitivity score had a median selumetinib sensitivity score of 0.64(IQR: 0.14, Q1-Q3: 0.59-0.70, n = 112 cells) and the tumor with the lowest median selumetinib sensitivity score had a median score of 0.37 (IQR: 0.21, Q1-Q3: 0.27-0.48, n = 1,034 cells). scRNAseq of plexiform neurofibroma biopsies reveals differential susceptibilities to selumetinib on a single cell level. These findings may explain the partial responses seen in clinical trials of selumetinib for NF1 and demonstrate the value of collecting scRNAseq data for future NF1 trials.

Pathformer: a biological pathway informed transformer for disease diagnosis and prognosis using multi-omics data.

Multi-omics data provide a comprehensive view of gene regulation at multiple levels, which is helpful in achieving accurate diagnosis of complex diseases like cancer. However, conventional integration methods rarely utilize prior biological knowledge and lack interpretability. To integrate various multi-omics data of tissue and liquid biopsies for disease diagnosis and prognosis, we developed a biological pathway informed Transformer, Pathformer. It embeds multi-omics input with a compacted multi-modal vector and a pathway-based sparse neural network. Pathformer also leverages criss-cross attention mechanism to capture the crosstalk between different pathways and modalities. We first benchmarked Pathformer with 18 comparable methods on multiple cancer datasets, where Pathformer outperformed all the other methods, with an average improvement of 6.3%-14.7% in F1 score for cancer survival prediction, 5.1%-12% for cancer stage prediction, and 8.1%-13.6% for cancer drug response prediction. Subsequently, for cancer prognosis prediction based on tissue multi-omics data, we used a case study to demonstrate the biological interpretability of Pathformer by identifying key pathways and their biological crosstalk. Then, for cancer early diagnosis based on liquid biopsy data, we used plasma and platelet datasets to demonstrate Pathformer's potential of clinical applications in cancer screening. Moreover, we revealed deregulation of interesting pathways (e.g., scavenger receptor pathway) and their crosstalk in cancer patients' blood, providing potential candidate targets for cancer microenvironment study. Pathformer is implemented and freely available at https://github.com/lulab/Pathformer. Supplementary data are available at Bioinformatics online.

Biomimetic hydrogels with mesoscale collagen architecture for patient-derived tumor organoids culture

Patient-derived tumor organoids (PDTOs) shows great potential as a preclinical model. However, the current methods for establishing PDTOs primarily focus on modulating local properties, such as sub-micrometer topographies. Nevertheless, they neglect to capture the global millimeter or intermediate mesoscale architecture that have been demonstrated to influence tumor response to therapeutic treatment and tumor progression. In this study, we present a rapid technique for generating collagen bundles with an average length of 90 ± 27 μm and a mean diameter of 5 ± 1.5 μm from tumor tissue debris that underwent mechanical agitation following enzymatic digestion. The collagen bundles were subsequently utilized for the fabrication of biomimetic hydrogels, incorporating microbial transglutaminase (mTG) crosslinked gelatin. These biomimetic hydrogels, referred to as MC-gel, were specifically designed for patient-derived tumor organoids. The lung cancer organoids cultured in MC-gel exhibited larger diameters and higher cell viability compared to those cultured in gels lacking the mesoscale collagen bundle; moreover, their irregular morphology more closely resembled that observed in vivo. The MC-gel-based lung cancer organoids effectively replicated the histology and mutational landscapes observed in the original donor patient’s tumor tissue. Additionally, these lung cancer organoids showed a remarkable similarity in their gene expression and drug response across different matrices. This recently developed model holds great potential for investigating the occurrence, progression, metastasis, and management of tumors, thereby offering opportunities for personalized medicine and customized treatment options.

Sex affects transcriptional associations with schizophrenia across the dorsolateral prefrontal cortex, hippocampus, and caudate nucleus

Schizophrenia is a complex neuropsychiatric disorder with sexually dimorphic features, including differential symptomatology, drug responsiveness, and male incidence rate. Prior large-scale transcriptome analyses for sex differences in schizophrenia have focused on the prefrontal cortex. Analyzing BrainSeq Consortium data (caudate nucleus: n = 399, dorsolateral prefrontal cortex: n = 377, and hippocampus: n = 394), we identified 831 unique genes that exhibit sex differences across brain regions, enriched for immune-related pathways. We observed X-chromosome dosage reduction in the hippocampus of male individuals with schizophrenia. Our sex interaction model revealed 148 junctions dysregulated in a sex-specific manner in schizophrenia. Sex-specific schizophrenia analysis identified dozens of differentially expressed genes, notably enriched in immune-related pathways. Finally, our sex-interacting expression quantitative trait loci analysis revealed 704 unique genes, nine associated with schizophrenia risk. These findings emphasize the importance of sex-informed analysis of sexually dimorphic traits, inform personalized therapeutic strategies in schizophrenia, and highlight the need for increased female samples for schizophrenia analyses.

HiPSC-CM Electrophysiology: Impact of Temporal Changes and Study Parameters on Experimental Reproducibility.

Human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) are frequently used for preclinical cardiotoxicity testing and remain an important tool for confirming model-based predictions of drug effects in accordance with Comprehensive in Vitro Proarrhythmia Assay (CiPA). Despite the considerable benefits hiPSC-CMs provide, concerns surrounding experimental reproducibility have emerged. We investigated the effects of temporal changes and experimental parameters on hiPSC-CM electrophysiology. iCell cardiomyocytes2 were cultured and biosignals were acquired using a microelectrode array (MEA) system (1-14 days). Continuous recordings revealed a 22.6% increase in the beating rate and 7.7% decrease in the field potential duration (FPD) during a 20-minute equilibration period. Location specific differences across a multiwell plate were also observed, with iCell cardiomyocytes2 in the outer rows beating 8.8 beats per minute (BPM) faster than the inner rows. Cardiac endpoints were also impacted by cell culture duration; from 2-14 days the beating rate decreased (-12.7 BPM), FPD lengthened (+257 ms), and spike amplitude increased (+3.3 mV). Cell culture duration (4-10 days) also impacted cardiomyocyte drug responsiveness (E-4031, nifedipine, isoproterenol). qRT-PCR results suggest that daily variations in cardiac metrics may be linked to the continued maturation of hiPSC-CMs in culture (2-30 days). Daily experiments were also repeated using a second cell line (Cor.4U). Collectively, our study highlights multiple sources of variability to consider and address when performing hiPSC-CM MEA studies. To improve reproducibility and data interpretation, MEA-based studies should establish a standardized protocol and report key experimental conditions (e.g., cell line, culture time, equilibration time, electrical stimulation settings, raw data values).

Time-dependent proteomics and drug response in expanding cancer cells

Cancer cell line is commonly used for discovery and development of anti-cancer drugs. It is generally considered that drug response remains constant for a certain cell line due to the identity of genetics thus protein patterns. Here, we demonstrated that cancer cells continued dividing even after reaching confluence, in that the proteomics was changed continuously and dramatically with strong relevance to cell division, cell adhesion and cell metabolism, indicating time-dependent intrinsically reprogramming of cells during expansion. Of note, the inhibition effect of most anti-cancer drugs was strikingly attenuated in culture cells along with cell expansion, with the strongest change at the third day when cells were still expanding. Profiling of an FDA-approved drug library revealed that attenuation of response with cell expansion is common for most drugs, an exception was TAK165 that was a selective inhibitor of mitochondrial respiratory chain complex I. Finally, we screened a panel of natural products and identified four pentacyclic triterpenes as selective inhibitors of cancer cells under prolonged growth. Taken together, our findings underscore that caution should be taken in evaluation of anti-cancer drugs using culture cells, and provide agents selectively targeting overgrowth cancer cells.

Prediction of anticancer drug sensitivity using an interpretable model guided by deep learning

BackgroundThe prediction of drug sensitivity plays a crucial role in improving the therapeutic effect of drugs. However, testing the effectiveness of drugs is challenging due to the complex mechanism of drug reactions and the lack of interpretability in most machine learning and deep learning methods. Therefore, it is imperative to establish an interpretable model that receives various cell line and drug feature data to learn drug response mechanisms and achieve stable predictions between available datasets.ResultsThis study proposes a new and interpretable deep learning model, DrugGene, which integrates gene expression, gene mutation, gene copy number variation of cancer cells, and chemical characteristics of anticancer drugs to predict their sensitivity. This model comprises two different branches of neural networks, where the first involves a hierarchical structure of biological subsystems that uses the biological processes of human cells to form a visual neural network (VNN) and an interpretable deep neural network for human cancer cells. DrugGene receives genotype input from the cell line and detects changes in the subsystem states. We also employ a traditional artificial neural network (ANN) to capture the chemical structural features of drugs. DrugGene generates final drug response predictions by combining VNN and ANN and integrating their outputs into a fully connected layer. The experimental results using drug sensitivity data extracted from the Cancer Drug Sensitivity Genome Database and the Cancer Treatment Response Portal v2 reveal that the proposed model is better than existing prediction methods. Therefore, our model achieves higher accuracy, learns the reaction mechanisms between anticancer drugs and cell lines from various features, and interprets the model’s predicted results.ConclusionsOur method utilizes biological pathways to construct neural networks, which can use genotypes to monitor changes in the state of network subsystems, thereby interpreting the prediction results in the model and achieving satisfactory prediction accuracy. This will help explore new directions in cancer treatment. More available code resources can be downloaded for free from GitHub (https://github.com/pangweixiong/DrugGene).

Molecular mechanisms and clinical relevance of endothelial cell cross-talk in clear cell renal cell carcinoma

Background: Clear cell renal cell carcinoma (ccRCC) is the most common renal cancer in adults and stands out as one of the most vascularized and immune-infiltrated solid tumors. Overproduction of vascular endothelial growth factor A promotes uncontrolled growth of abnormal vessels and immunosuppression, and the tumor microenvironment (TME) has a prominent role in disease progression, drug targeting and drug response, and for patient outcome. Methods: Studies of experimental models, large-scale omics approaches, and patient prognosis and therapy prediction, using gene expression signatures and tissue biomarker analysis, have been reviewed for enhanced understanding of the endothelium in ccRCC and the interplay with the surrounding TME. Results: Preclinical and clinical studies have discovered molecular mechanisms of endothelial cross-talk of relevance for disease progression, patient prognosis, and therapy prediction. There is, however, a lack of representative ccRCC experimental models. Omics approaches have identified clinically relevant subsets of angiogenic and immune-infiltrated tumors with distinct molecular signatures and distinct endothelial cell and immune cell populations in patients. Conclusions: Recent genetically engineered ccRCC mouse models together with emerging evidence from single cell RNA sequencing data open up for future validation studies, including multiplex imaging of ccRCC patient cohorts. These studies are of importance for therapy benefit and personalized treatment of ccRCC patients.

Pharmacogenomic analysis in adrenocortical carcinoma reveals genetic features associated with mitotane sensitivity and potential therapeutics

BackgroundAdrenocortical carcinoma (ACC) is an aggressive endocrine malignancy with limited therapeutic options. Treating advanced ACC with mitotane, the cornerstone therapy, remains challenging, thus underscoring the significance to predict mitotane response prior to treatment and seek other effective therapeutic strategies.ObjectiveWe aimed to determine the efficacy of mitotane via an in vitro assay using patient-derived ACC cells (PDCs), identify molecular biomarkers associated with mitotane response and preliminarily explore potential agents for ACC.MethodsIn vitro mitotane sensitivity testing was performed in 17 PDCs and high-throughput screening against 40 compounds was conducted in 8 PDCs. Genetic features were evaluated in 9 samples using exomic and transcriptomic sequencing.ResultsPDCs exhibited variable sensitivity to mitotane treatment. The median cell viability inhibition rate was 48.4% (IQR: 39.3-59.3%) and -1.2% (IQR: -26.4-22.1%) in responders (n=8) and non-responders (n=9), respectively. Median IC50 and AUC were remarkably lower in responders (IC50: 53.4 µM vs 74.7 µM, P&amp;lt;0.0001; AUC: 158.0 vs 213.5, P&amp;lt;0.0001). Genomic analysis revealed CTNNB1 somatic alterations were only found in responders (3/5) while ZNRF3 alterations only in non-responders (3/4). Transcriptomic profiling found pathways associated with lipid metabolism were upregulated in responder tumors whilst CYP27A1 and ABCA1 expression were positively correlated to in vitro mitotane sensitivity. Furthermore, pharmacologic analysis identified that compounds including disulfiram, niclosamide and bortezomib exhibited efficacy against PDCs.ConclusionACC PDCs could be useful for testing drug response, drug repurposing and guiding personalized therapies. Our results suggested response to mitotane might be associated with the dependency on lipid metabolism. CYP27A1 and ABCA1 expression could be predictive markers for mitotane response, and disulfiram, niclosamide and bortezomib could be potential therapeutics, both warranting further investigation.

A new evaluation system for drug–microbiota interactions

AbstractThe drug response phenotype is determined by a combination of genetic and environmental factors. The high clinical conversion failure rate of gene‐targeted drugs might be attributed to the lack of emphasis on environmental factors and the inherent individual variability in drug response (IVDR). Current evidence suggests that environmental variables, rather than the disease itself, are the primary determinants of both gut microbiota composition and drug metabolism. Additionally, individual differences in gut microbiota create a unique metabolic environment that influences the in vivo processes underlying drug absorption, distribution, metabolism, and excretion (ADME). Here, we discuss how gut microbiota, shaped by both genetic and environmental factors, affects the host's ADME microenvironment within a new evaluation system for drug–microbiota interactions. Furthermore, we propose a new top‐down research approach to investigate the intricate nature of drug–microbiota interactions in vivo. This approach utilizes germ‐free animal models, providing foundation for the development of a new evaluation system for drug–microbiota interactions.

Decrypting the molecular basis of cellular drug phenotypes by dose-resolved expression proteomics.

Proteomics is making important contributions to drug discovery, from target deconvolution to mechanism of action (MoA) elucidation and the identification of biomarkers of drug response. Here we introduce decryptE, a proteome-wide approach that measures the full dose-response characteristics of drug-induced protein expression changes that informs cellular drug MoA. Assaying 144 clinical drugs and research compounds against 8,000 proteins resulted in more than 1 million dose-response curves that can be interactively explored online in ProteomicsDB and a custom-built Shiny App. Analysis of the collective data provided molecular explanations for known phenotypic drug effects and uncovered new aspects of the MoA of human medicines. We found that histone deacetylase inhibitors potently and strongly down-regulated the T cell receptor complex resulting in impaired human T cell activation in vitro and ex vivo. This offers a rational explanation for the efficacy of histone deacetylase inhibitors in certain lymphomas and autoimmune diseases and explains their poor performance in treating solid tumors.

Laminin-associated integrins mediate Diffuse Intrinsic Pontine Glioma infiltration and therapy response within a neural assembloid model

Diffuse Intrinsic Pontine Glioma (DIPG) is a highly aggressive and fatal pediatric brain cancer. One pre-requisite for tumor cells to infiltrate is adhesion to extracellular matrix (ECM) components. However, it remains largely unknown which ECM proteins are critical in enabling DIPG adhesion and migration and which integrin receptors mediate these processes. Here, we identify laminin as a key ECM protein that supports robust DIPG cell adhesion and migration. To study DIPG infiltration, we developed a DIPG-neural assembloid model, which is composed of a DIPG spheroid fused to a human induced pluripotent stem cell-derived neural organoid. Using this assembloid model, we demonstrate that knockdown of laminin-associated integrins significantly impedes DIPG infiltration. Moreover, laminin-associated integrin knockdown improves DIPG response to radiation and HDAC inhibitor treatment within the DIPG-neural assembloids. These findings reveal the critical role of laminin-associated integrins in mediating DIPG progression and drug response. The results also provide evidence that disrupting integrin receptors may offer a novel therapeutic strategy to enhance DIPG treatment outcomes. Finally, these results establish DIPG-neural assembloid models as a powerful tool to study DIPG disease progression and enable drug discovery.

Clinicians’ perspectives on race-specific guidelines for hypertensive treatment

Despite the general consensus that there is no biological basis to race, racial categorization is still used by clinicians to guide diagnosis and treatment plans for certain diseases. In medicine, race is commonly used as a rough proxy for unmeasured social, environmental, and genetic factors. The American College of Cardiology's Eighth Joint National Committee's (JNC 8) guidelines for the treatment of hypertension provide race-specific medication recommendations for Black versus non-Black patients, without strong evidence for race-specific physiological differences in drug response. Clinicians practicing family or geriatric medicine (n = 21) were shown a video of a mock hypertensive patient with genetic ancestry test results that could be viewed as discordant with their phenotype and self-identified race. After viewing the videos, we conducted in-depth interviews to examine how clinicians value and prioritize different cues about race -- namely genetic ancestry data, phenotypic appearance, and self-identified racial classifications – when making treatment decisions in the context of race-specific guidelines, particularly in situations when patients claim mixed-race or complex racial identities. Results indicate that clinicians inconsistently follow the race-specific guidelines for patients whose genetic ancestry test results do not match neatly with their self-identified race or phenotypic features. However, many clinicians also emphasized the importance of clinical experience, side effects, and other factors in their decision making. Clinicians' definitions of race, categorization of the patient's race, and prioritization of racial cues greatly varied. The existence of the race-specific guidelines clearly influences treatment decisions, even as clinicians' express uncertainty about how to incorporate consideration of a patient's genetic ancestry. In light of widespread debate about removal of race from medical diagnostics, researchers should revisit the clinical justification for maintaining these race-specific guidelines. Based on our findings and prior studies indicating a lack of convincing evidence for biological differences by race in medication response, we suggest removing race from the JNC 8 guidelines to avoid risk of perpetuating or exacerbating health disparities in hypertension.

Serum levels of tumour necrosis factor (TNF-α) and interleukin-17 (IL-17) in patients with nail psoriasis: A cross-sectional study

Background Psoriasis is a common chronic inflammatory disorder affecting all aspects of a patient’s life. Nail involvement is frequent, but little is known about its associated inflammatory biomarker profile, including similarities or differences from cutaneous disease. Aims We conducted this cross-sectional study to evaluate serum levels of inflammatory cytokines [tumour necrosis factor-alpha (TNF-α) and interleukin -17 (IL-17)] in patients with nail psoriasis and compared these to psoriasis patients without nail involvement, as well as in non-psoriatic healthy controls. Methods Adult psoriasis patients with (Group I, n = 30) and without nail involvement (Group-II, n = 30) were sequentially recruited. In addition, non-psoriatic healthy controls (Group-III, n = 20) were recruited. The nail disease severity by NAPSI score was determined for patients in Group I. Cutaneous disease severity (by PASI score) and presence of psoriatic arthritis (through CASPAR criteria) were evaluated for patients in Groups I and II. Serum levels of TNF-α, IL-17, erythrocyte sedimentation rate (ESR), rheumatoid factor (RA factor), and anti-cyclic citrullinated peptide antibody (Anti-CCP) were evaluated for all three groups. Results The median age was significantly higher for Group I as compared to Group II patients (41 ± 12.6 years vs 30 ± 12.4 years, p = 0.017). Group I patients also had higher median PASI score than Group II patients, although the difference was not statistically significant (10 ± 11.41 vs 6.50 ± 5.46, p = 0.275). The mean serum IL-17 levels were significantly higher for Group-I (113.39 ± 251.30 pg/mL) than Group II (27.91 ± 18.22 pg/mL, p = 0.002) and Group III (25.67 ± 12.08 pg/mL, p = 0.005). A weak positive correlation was found between NAPSI and serum IL-17 levels (Spearman’s Rho = 0.355) though not statistically significant (p = 0.054). Correlation between serum IL-17 and PASI was poor for Group-I patients (Spearman’s Rho = 0.13, p = 0.944) and strongly negative for Group-II patients (Spearman’s Rho = −0.368, statistically significant with p = 0.045). The mean serum levels of TNF-α were below the detection threshold of the assay kit, hence no meaningful comparison could be made. Limitations A small sample size and low sensitivity of TNF-α assay kit. Conclusion Our study showed that nail psoriasis could be independently associated with an elevation of IL-17. This can help choose appropriate drugs and estimate drug response in patients with nail psoriasis.

Phenomic landscape and pharmacogenomic implications for HLA region in a Taiwan Han Chinese population

BackgroundThe human leukocyte antigen (HLA) genes, exhibiting significant genetic diversity, are associated with susceptibility to various clinical diseases and diverse in drug responses. High costs of HLA sequencing and the population-specific architecture of this genetic region necessitate the establishment of a population-specific HLA imputation reference panel. Moreover, there is a lack of understanding about the genetic and phenotypic landscape of HLA variations within the Taiwanese population.MethodsWe created models for a Taiwanese-specific HLA imputation reference panel. These models were trained with the array genotype data and HLA sequencing data from 845 Taiwanese subjects. HLA imputation was applied for 59,448 Taiwanese subjects to characterize the HLA allele and haplotype frequencies. Additionally, a phenome-wide association study (PheWAS) was conducted to identify the phenotypes associated with HLA variations. The association of the biallelic HLA variants with the binary and quantitative traits were evaluated with additive logistic and linear regression models, respectively. Furthermore, an omnibus test with likelihood-ratio test was applied for each HLA amino acid position in the multiallelic HLA amino acid polymorphisms to compare the difference between a fitted model and a null model following a χ2 distribution of n-1 degree of freedom at a position with n residues. Finally, we estimated the prevalence of adverse drug reactions (ADR)-related HLA alleles in the Taiwanese population.ResultsIn this study, the reference panel models displayed remarkable accuracy, with averages of 99.3%, 98.9%, and 99.1% for 2-, 4-, 6-digit alleles of the eight classical HLA genes, respectively. For PheWAS, a total of 18,136 significant associations with HLA variants across 26 phenotypes are identified (p < 5×10-8), highlighting the pleiotropy feature of the HLA region. Among the independent signals, 15 are novel, including the association of HLA-B pos 138 variation with ankylosing spondylitis (AS), and rs9266290 and rs9266292 with allergy. Through an analysis spanning the entire HLA region, we identified clusters of phenotype correlations. Finally, the carriers of pharmacogenomic related HLA alleles, including HLA-C*01:02 (35.86%), HLA-B*58:01 (20.9%), and HLA-B*15:02 (8.38%), were characterized in the Taiwanese general population.ConclusionsWe successfully delivered the HLA imputation for 59,448 Taiwanese subjects and characterized the genetic and phenotypic landscapes of the HLA variations. In addition, we quantified the estimated prevalence of the ADR-related HLA alleles in the Taiwanese population. The developed HLA imputation reference panel could be used for estimation of population HLA allele frequencies, which can facilitate further studies in the role of HLA variants in a wider range of phenotypes in the population.

Abstract PO2-14-08: Development and characterization of breast cancer organoids representing epithelial heterogeneity and drug response

Abstract Organoids serve as an important preclinical model in multiple cancers including breast cancer, demonstrating preservation of biological features and feasibility of drug screening, with reasonable cost and high scalability. It has been reported that patient-derived organoids (PDOs) can be robustly derived from both primary and metastatic breast tumors, with high resemblance to source tissues in histological, genomic, and transcriptomic features, as well as consistency in drug response compared to xeno-transplantations or patients (Sachs et al, 2018). Since 2018, the Institute for Precision Medicine (IPM) has developed a panel of breast cancer organoids. The current inventory now consists of 128 PDOs from 94 primary tumors, 12 metastases, 12 rapid-autopsy tumors, and 10 normal breast tissues with a 60% success rate of culture; as well as 33 organoids from animal models including 11 patient-derived xenograft organoids (PDXOs), 15 rat (RDO) and 7 mouse (MDO) tumor organoids with 85% success rate. For PDOs, the protocol includes in-lab processing of fresh, deidentified tissue within 60 minutes of surgical operation, thanks to close collaboration with surgeons, consented patients, and an institutional biospecimen core. BC organoids have been comprehensively characterized in terms of morphology, histology, genomics, transcriptomics, and functional experiments. Key biological features are preserved. PDOs (n=56) exhibit a mutational spectrum consistent with human breast cancers. ER expression is detectable in a subset of cultures with robust estradiol response indicated by GREB1 expression; and PDOs from invasive lobular carcinoma (ILC) mostly showed more discohesive structures than PDOs from no special type (NST), with clear E-cadherin loss under immunofluorescence. We have also successfully constructed transcriptionally and genetic modified BC PDOs with shRNA and CRISPR, which showed replicable performance in proliferation assay, drug response, and signaling by western blot. Of note, longitudinal tracking of PDOs with single cell RNA-sequencing (scRNA-seq) revealed well-preserved heterogeneity compared to primary tumor, while transcriptomic clonality was mildly decreased in later passages from pilot data. Future work will include expansion of current collection from various sources, longitudinal characterization with scRNA-seq and DNA sequencing, as well as more functional studies. We aim to construct a robust resource of well-characterized and reliably-usable model bank for preclinical breast cancer research. Citation Format: Daniel Brown, Kai Ding, Fangyuan Chen, Jian Chen, Steffi Oesterreich, Peter Lucas, Priscilla McAuliffe, Jennifer Atkinson, ADRIAN LEE. Development and characterization of breast cancer organoids representing epithelial heterogeneity and drug response [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-14-08.

Abstract PO4-15-11: HER2 STATUS AND TUMOR HETEROGENEITY IN INVASIVE BREAST CARCINOMAS: CLINICAL PATHOLOGICAL IMPACT

Abstract Human epidermal growth factor receptor 2 (HER2) is a major player in breast carcinogenesis with a range of 15-20% of cases harboring gene amplification and overexpression turning the tumor eligible to over eight anti-HER2 drugs in the past 15 years. Recently, novel anti-HER2 antibody-drug conjugates (ADCs) treatment was recently approved to metastatic breast cancer bearing some HER2 expression albeit not due to gene amplification. Immunohistochemically classified, by ASCO-CAP Guidelines (2018), as 1+ cases were enrolled in clinical trials and benefited by the addition of ADCs to chemotherapy. Despite being classified as HER2 negative, most tumors exhibit detectable levels of HER2 protein, reinforcing the importance of HER2 expression evaluation. HER-2 intratumoral heterogeneity (ITH) is reported in up to 40% of breast cancers and is defined as the coexistence of subpopulations of tumor cells with different HER2 gene or protein expression. In this scenario, it becomes very important to evaluate the spectrum of HER-2 expression in correlation with pathological factors as ITH is associated with poor prognosis and anti-HER2 resistance. We designed a retrospective linear study to reanalyze all HER2 immunohistochemical core biopsy slides of patients with primary invasive breast carcinomas (IBC) diagnosed and treated at São Paulo Federal University Hospital between 2019 and 2023. Clinical variables (age, laterality, tumor size by ultrasound imaging, BI-RADS) were collected from clinical files. HER2 slides were reviewed by three observers and reported regarding status by ASCO-CAP 2018 guidelines, HER2 heterogeneity and patterns. Statistical analyses were performed with Chi-Square test of independence, Fisher's exact test and Mann–Whitney U test using IBM® SPSS® Statistics (26.0) software. 353 cases were included; the mean age of the patients at the time of diagnosis was 58 years old. 191 cases (54,1%) were left sided and 162 cases (45,9%) were right sided. Regarding BI-RADS report, the major category was 4 (140 cases), followed by 5 (119 cases). Out of the 164 patients who underwent clinical follow-up at our institute, 14 deaths were reported. Concerning pathological features, IBC of no special type was the prevalent histological subtype with 304 cases (86,1%), followed by invasive lobular carcinoma, with 26 cases (7,4%). 296 cases (83,9%) were HER2 negative and 42 (11,9%) were HER2 positive; 15 cases were HER2 2+ (4,2%). Amongst negative cases, 235 were HER2 0+ and 61 were HER2 1+. Considering ITH, 287 cases (81,3%) were HER2 homogenous with 0+ score as the most prevalent, presenting with 232 cases (80%). 66 cases (18,7%) had ITH, with 1+ as the most prevalent primary score, with 34 cases (51,5%), followed by 3+, with 17 cases (25,8%). Not only HER2 positive carcinomas, but also the presence of heterogeneity in HER2 expression was associated with death outcome (p=0,003 and p=0,001, respectively). Also, HER2 positive IBC had larger tumor size compared to HER2 negative carcinomas (mean size of 25,6 mm x 31,3 mm, respectively; p=0,012); HER2 heterogeneity showed no statistical significance when associated with tumor size (p=0,165). We conclude that ITH is prevalent in HER2 expression and should be addressed in pathology reports since it may play an additional role in tumor progression and drug response, especially in the ADCs scenario. Citation Format: Karla Calaça Kabbach Prigenzi, Lisandra Gonzalez Porta Nova, Adilson Monteiro dos Santos Filho, Ana Luiza da Cruz, Lucas de Figueiredo Barbosa, Angela Flavia Waitzberg. HER2 STATUS AND TUMOR HETEROGENEITY IN INVASIVE BREAST CARCINOMAS: CLINICAL PATHOLOGICAL IMPACT [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO4-15-11.

Abstract PO4-14-12: Predicting Response to HER2 Tyrosine Kinase Inhibitors and Antibody Drug Conjugates in HER2 Mutant Invasive Lobular Carcinoma Using CRISPR/Cas9 Knock-in Cell lines and Patient-derived Organoids

Abstract Background: Activating mutations in ERBB2 (HER2) are enriched &amp;gt;4-fold in invasive lobular carcinoma (ILC) with a rate of up to 19% in metastatic ILC. ILC is a histologic subtype of breast cancer characterized by loss of E-cadherin (CDH1), suggesting a potential interaction between loss of CDH1 and mutations in ERBB2. Recent trials have demonstrated promising single agent efficacy using the irreversible pan-HER tyrosine kinase inhibitor (TKI), neratinib, in patients with metastatic ERBB2 mutant ILC. However, further studies on combination therapies with other anticancer agents are needed to increase response rate and progression free survival for these patients. HER2-targeted antibody drug conjugates (ADC), particularly trastuzumab deruxtecan (T-DXd), have shown great promise in HER2-low metastatic breast cancer. Yet, their efficacy as a single agent or with HER2 TKIs in HER2-low and -mutant ILC is unknown and warrants investigation. Methods: Past studies analyzing ERBB2 mutations used overexpression of ERBB2 mutant cDNA, but this approach does not faithfully recapitulate the human disease. To model ERBB2 missense mutations as found in human breast cancers, we used CRISPR-based prime editing to generate a panel of isogenic ILC cell lines and patient-derived organoids (PDO) harboring ERBB2 wild-type (WT) or ERBB2 mutations (S310F or V777L). Both of these ERBB2 mutations have been previously characterized and are known to be activating mutations. We then used them to test neratinib and other TKIs with ADCs, including T-DXd and trastuzumab emtansine (T-DM1). Results: We successfully introduced single copy, heterozygous activating ERBB2 mutations (S310F or V777L) into two ERBB2-nonamplified metastatic ILC cell lines (MDA-MB-134 and SUM44PE) and one ERBB2-nonamplified metastatic ILC PDO (IPM-BO-053). Positive clones carrying the mutations were verified by Sanger sequencing and droplet digital PCR and subsequently pooled together. We further demonstrated that these mutations hyperactivated HER2 and downstream signaling pathways. ILC cell lines harboring these mutations showed enhanced sensitivity to HER2 TKIs but not ADCs. In contrary, ERBB2 mutations did not alter responses of IPM-BO-053 PDOs to HER2 TKIs but significantly increased responses to ADCs. Interestingly, we also observed accelerated HER2 protein degradation upon heregulin stimulation in ERBB2 mutant ILC PDOs, suggesting activating ERBB2 mutations may enhance ADC/HER2 complex internalization, degradation, and release of payloads. Lastly, we explored drug synergy between HER2 TKIs and ADCs in isogenic IPM-BO-053 PDOs and found that combination of T-DXd and neratinib or afatinib showed synergy (Combination Index &amp;lt; 1). Conclusions: Although the reason for the discrepancies in drug response between ILC cell lines and PDOs is not clear, we hypothesize that response in 3D PDOs might be more faithfully representing response seen in patients. We will generate additional ILC PDOs with knock-in ERBB2 mutations to validate our findings. Irreversible HER2 TKIs, such as neratinib and afatinib, showed synergy with T-DXd in ERBB2 mutant ILC PDOs. This holds important therapeutic implications in light of current treatment options for ILC. In future experiments, we will test if neratinib or afatinib increases endocytic uptake of T-DXd using a fluorescently labeled endocytosis tracker. Further, an in-depth molecular characterization of our isogenic cell lines and PDOs models is ongoing to gain mechanistic insights into how activating ERBB2 mutations increase HER2 internalization and degradation. Our in vitro studies provide a strong foundation for in vivo testing of neratinib or afatinib with T-DXd for ERBB2 mutant ILC. Citation Format: Jie Bin Liu, Danielle Tseng, Jagmohan Hooda, Daniel Brown, Adrian Lee, Steffi Oesterreich. Predicting Response to HER2 Tyrosine Kinase Inhibitors and Antibody Drug Conjugates in HER2 Mutant Invasive Lobular Carcinoma Using CRISPR/Cas9 Knock-in Cell lines and Patient-derived Organoids [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO4-14-12.