scholarly journals Drug Response Prediction as a Link Prediction Problem

2017 ◽  
Vol 7 (1) ◽  
Author(s):  
Zachary Stanfield ◽  
Mustafa Coşkun ◽  
Mehmet Koyutürk

Abstract Drug response prediction is a well-studied problem in which the molecular profile of a given sample is used to predict the effect of a given drug on that sample. Effective solutions to this problem hold the key for precision medicine. In cancer research, genomic data from cell lines are often utilized as features to develop machine learning models predictive of drug response. Molecular networks provide a functional context for the integration of genomic features, thereby resulting in robust and reproducible predictive models. However, inclusion of network data increases dimensionality and poses additional challenges for common machine learning tasks. To overcome these challenges, we here formulate drug response prediction as a link prediction problem. For this purpose, we represent drug response data for a large cohort of cell lines as a heterogeneous network. Using this network, we compute “network profiles” for cell lines and drugs. We then use the associations between these profiles to predict links between drugs and cell lines. Through leave-one-out cross validation and cross-classification on independent datasets, we show that this approach leads to accurate and reproducible classification of sensitive and resistant cell line-drug pairs, with 85% accuracy. We also examine the biological relevance of the network profiles.

2021 ◽  
Author(s):  
David Earl Hostallero ◽  
Yihui Li ◽  
Amin Emad

Motivation: The increasing number of publicly available databases containing drugs' chemical structures, their response in cell lines, and molecular profiles of the cell lines has garnered attention to the problem of drug response prediction. However, many existing methods do not fully leverage the information that is shared among cell lines and drugs with similar structure. As such, drug similarities in terms of cell line responses and chemical structures could prove to be useful in forming drug representations to improve drug response prediction accuracy. Results: We present two deep learning approaches, BiG-DRP and BiG-DRP+, for drug response prediction. Our models take advantage of the drugs' chemical structure and the underlying relationships of drugs and cell lines through a bipartite graph and a heterogenous graph convolutional network that incorporate sensitive and resistant cell line information in forming drug representations. Evaluation of our methods and other state-of-the-art models in different scenarios show that incorporating this bipartite graph significantly improve the prediction performance. Additionally, genes that contribute significantly to the performance of our models also point to important biological processes and signaling pathways.


2017 ◽  
Vol 7 (1) ◽  
Author(s):  
Zachary Stanfield ◽  
Mustafa Coşkun ◽  
Mehmet Koyutürk

1974 ◽  
Vol 16 (3) ◽  
pp. 603-621
Author(s):  
C. ALTANER ◽  
J. MATOSKA

Hamster cells transformed with the Schmidt-Ruppin strain of avian sarcoma virus were selected for resistance to ethidium bromide (EB). The resistant cell lines proliferated in the presence of up to 30 µg/ml EB. From avian sarcoma virus-transformed hamster cells already resistant to bromodeoxy-uridine (BrdU), ethidium bromide-resistant cells which were able to grow in 10 µg/ml EB were also prepared. These cells remain deficient in thymidine kinase activity and are suitable for selective preparation of hybrid cells. The EB resistance was genetically stable. The EB-resistant cell lines, and doubly resistant cells (BrdU, EB) showed no differences in mitochondrial ultrastructure compared with the original cell lines. Thymidine incorporation into mitochondrial DNA was not influenced by EB resistance. All resistant cell lines, including the doubly resistant cell line, contained the avian sarcoma virus genome. The number of cells needed for positive rescue experiments for avian sarcoma virus genome by cell fusion with permissive chicken embryo cells was the same as with the original cell lines. The single EB-resistant cell lines contained R-type virus-like particles, while in BrdU-resistant and doubly resistant cells the R-type particles were absent. The possible nature of EB resistance is discussed.


Blood ◽  
2005 ◽  
Vol 106 (11) ◽  
pp. 746-746
Author(s):  
Kavita B. Kalra ◽  
Xiangfei Cheng ◽  
Marion Womak ◽  
Christopher Gocke ◽  
Jyoti B. Patel ◽  
...  

Abstract All trans retinoic acid (ATRA) has been used in differentiation therapy for APL and other types of cancers. However, the rapid emergence of ATRA resistance due in part to ATRA-induced acceleration of ATRA metabolism limits its use. A novel strategy to overcome the limitation associated with exogenous ATRA therapy has been developed by inhibiting the cytochrome P450-dependent ATRA-4-hydroxylase enzyme responsible for ATRA metabolism. These inhibitors are referred to as RAMBAs. Novel RAMBAs were developed which demonstrated a superior apoptosis, cell growth inhibition, in vivo anti-tumor effect in addition to the differentiation effect in breast cancer cell lines (Patel JB et al. J. Med. Chem2004,47:6716). We tested 3 RAMBAs, VN/14-1, 50-1, and 66-1 to investigate their activities against APL cell lines. RAMBAs did not confer cytotoxicity or apoptosis induction in vitro at the concentration between 0.5 to 5 μM as opposed to breast or prostate cancer cell lines. However, the differentiation effect was demonstrated by morphological and phenotypic changes using Wright-Giemsa stain and CD11b staining measured by flow cytometric analysis. VN/14-1 and VN/66-1 induced differentiation and apoptosis morphologically and phenotypically in HL60 cells. VN/14-1 and VN/50-1 showed superior differentiation in NB4 cell line compared to ATRA (70%, 69%, and 45%, respectively). Interestingly, HL60 ATRA resistant cell line was induced to undergo differentiation by VN/14-1 (0.5μM) at 55% whereas ATRA (0.5, 1, 5μM) showed less than 5% by flow cytometry analysis. VN/14-1 inhibited cell cycle at S phase whereas ATRA did not attenuate the cell cycle at the same concentration. We also tested the effect of RAMBAs on human CD34+ enriched cell colony formation. RAMBAs were added to the methylcellulose culture plates with CD34+ cells and colonies were determined after 14 days. There was no difference in the CFU-GM or BFU-E colony count between the control and the RAMBAs group. In summary, RAMBAs are promising differentiation agents in the treatment of APL, possibly through an inhibition of Cyp26A leading to increased endogenous ATRA levels. In addition, cell cycle inhibition may be a mechanism of differentiation induction in ATRA resistant cell lines. RAMBAs did not affect normal hematopoietic stem cells. We are currently testing whether RAMBAs can induce acetylation of histones in APL cell lines.


Blood ◽  
2021 ◽  
Vol 138 (Supplement 1) ◽  
pp. 2668-2668
Author(s):  
Yuan Xiao Zhu ◽  
Laura Ann Bruins ◽  
Joseph Ahmann ◽  
Cecilia Bonolo De Campos ◽  
Esteban Braggio ◽  
...  

Abstract Venetoclax (VTX) is a selective small-molecule inhibitor of BCL-2 that exhibits antitumoral activity against MM cells presenting lymphoid features and those with translocation t(11;14). Despite its impressive clinical activity, VTX therapy for a prolonged duration can lead to drug resistance. Therefore, it is important to understand the underlying mechanisms of resistance in order to develop strategies to prevent or overcome resistance. In the present study, we established four VTX resistant human myeloma cell lines (HMCLs) from four sensitive HMCLs, including three with t(11;14), in culture with a stepwise increase in treatment dose with VTX. To identify the molecular basis of acquired VTX resistance, whole exon sequencing (WES), mRNA-sequencing (mRNAseq), and protein expression assays were performed in the four isogenic VTX-sensitive/resistant HMCLs and three MM patients with samples collected before VTX administration and after clinical resistance to the drug. Compared with sensitive cell lines and patient samples collected before VTX administration, mRNAseq analysis identified downregulation of BIM and upregulation of BCLXL in both resistant cell lines and MM cells from relapse patients. Other transcriptional changes detected included upregulation of AURKA, BIRC3, BIRC5, and IL32. Enrichment analysis of differentially expressed genes suggested involvement of PI3K and MAPK signaling, likely associated with cytokines, growth factors (EGF, FGF and IGF family members), and receptor tyrosine kinase (EGF and FGF). Western blot analysis was performed to compare BCL2 family expression in resistant cell lines versus sensitive cell lines and it showed upregulation of BCL2 survival members (such as MCL-1 and BCLXL), and downregulation of pro-apoptotic BH3 members (such as BIM and PUMA). BIM expression was completely lost in one resistant cell line, and introduction of exogenous BIM into this cell line enhanced VTX sensitivity. Interestingly, BCL2 was upregulated in some resistant cell lines generated after a long-term treatment with VTX, suggesting BCL2 expression level may not be suitable as a marker of VTX sensitivity for acquired resistance. Unlike in CLL, BCL2 mutations were not identified through WES in any resistant cell lines or primary patient sample harvested after relapse. While 8 genes were mutated in two resistant samples , no clear mutational pattern emerged . Based on the above, we further tested some specific inhibitors in in vitro or ex vivo cell models to help understanding resistant mechanism and identify strategies to overcome VTX resistance. We found that inhibition of MCL-1, with the compound S68345, substantially enhanced VTX sensitivity in three resistant HMCLs and in primary cells from one relapsed MM patient. A BCLXL inhibitor (A155463) only significantly enhanced VTX sensitivity in one resistant cell line after co-treatment with VTX. Co-treatment of the other three resistant cell lines with VTX, S68345 and A155463 resulted in the most synergistic anti-myeloma activity, suggesting those cell lines are co-dependent on MCL-1, BCLXL, and BCL2 for survival, although they are more dependent on MCL-1. We also found that inhibition of PI3K signaling, IGF1, RTK (EGF and FGF) and AURKA significantly increased VTX sensitivity, partially through downregulation of MCL-1, and BCLXL, and upregulation of BIM. Conventional anti-MM drugs such as dexamethasone, bortezomib and lenalidomide, were shown to have little activity on augmenting VTX sensitivity in most resistant cell lines. In summary, we find that acquired resistance to VTX in MM is largely associated with BCL2 family regulation, including upregulation of survival members such as MCL-1, BCLXL, BCL2, and downregulation of pro-apoptotic members, especially BIM. Our study also indicates that upstream signaling involved in BCL2 family regulation during acquired resistance is likely related to cytokine, growth factor, and/or RTK-induced cell signaling such as PI3K. Co-inhibition of MCL-1, or BCLXL, as well as the upstream PI3K, RTK (FGF and EGF), IGF-1 mediated signaling were effective in overcoming VTX resistance. Disclosures Fonseca: Mayo Clinic in Arizona: Current Employment; Amgen: Consultancy; BMS: Consultancy; Celgene: Consultancy; Takeda: Consultancy; Bayer: Consultancy; Janssen: Consultancy; Novartis: Consultancy; Pharmacyclics: Consultancy; Sanofi: Consultancy; Merck: Consultancy; Juno: Consultancy; Kite: Consultancy; Aduro: Consultancy; OncoTracker: Consultancy, Membership on an entity's Board of Directors or advisory committees; GSK: Consultancy; AbbVie: Consultancy; Patent: Prognosticaton of myeloma via FISH: Patents & Royalties; Scientific Advisory Board: Adaptive Biotechnologies: Membership on an entity's Board of Directors or advisory committees; Caris Life Sciences: Membership on an entity's Board of Directors or advisory committees.


Blood ◽  
1994 ◽  
Vol 83 (4) ◽  
pp. 1051-1059 ◽  
Author(s):  
N Bonnefoy-Berard ◽  
L Genestier ◽  
M Flacher ◽  
JP Rouault ◽  
G Lizard ◽  
...  

Antilymphocyte and antithymocyte globulins (ALG) are currently used as immunosuppressive agents in clinical transplantation and for the treatment of severe aplastic anemia. ALG contain a mixture of antibodies that recognize T- and B-cell-specific antigens but mostly nonlineage-specific molecules. We reported previously that ALG could inhibit the proliferation of activated B cells and B cell lines (Bonnefoy-Berard et al, Blood 79:2164, 1992). We show here that ALG induce apoptosis of several human hematopoietic cell lines, as shown by nuclear condensation and fragmentation in fluorescence and electronic microscopy and by double-strand DNA breaks shown by DNA electrophoresis. Apoptosis was achieved without elevation of intracellular Ca2+ and requirement for mRNA and protein synthesis. Most of the B-cell lines tested (Epstein-Barr virus [EBV]-transformed lymphoblastoid cell lines, EBV-negative and groups I/III EBV-positive Burkitt's lymphoma cell lines, as well as other B-lymphoma cell lines) were susceptible to ALG-induced cytotoxicity. Myelomonocytic and T-cell lines were much less susceptible than B-cell lines. Susceptibility to ALG-induced cytotoxicity was not correlated with intracellular Bcl-2 level. Most cell lines that express high levels of Fas/Apo-1 antigen were susceptible to ALG. However, several lines of evidence support the conclusion that, in addition to Fas/Apo-1, other cell surface molecules can mediate ALG-induced apoptosis. The cytotoxic activity could be fully removed by adsorption on susceptible cell lines but not on a resistant cell line, indicating that it was mediated by antibodies specific for surface antigens expressed only on susceptible cell lines. Apoptosis was triggered by ALG F(ab')2 fragments as well as by intact ALG. This cytotoxic property of ALG may account for their antiproliferative effect and might contribute to some extent to the relatively lower risk of posttransplant lymphoproliferative disorders previously reported in ALG-treated patients.


2020 ◽  
Author(s):  
Nithya Balasundaram ◽  
Saravanan Ganesan ◽  
Ezhilarasi Chendamarai ◽  
Hamenth Kumar Palani ◽  
Arvind Venkatraman ◽  
...  

AbstractAcquired genetic mutations can confer resistance to arsenic trioxide (ATO) in the treatment of acute promyelocytic leukemia (APL). However, such resistance-conferring mutations are rare and do not explain the majority of disease recurrence seen in the clinic. We have generated a stable ATO resistant promyelocytic cell from a ATO sensitive NB4 cell line. We also noted that another ATRA resistant cell line (UF1) was cross resistant to ATO. We have characterized these resistant cell lines and observed that they significantly differed in their immunophenotype, drug transporter expression, drug resistance mutation profile and were also cross-resistant to other conventional chemotherapeutic agents. The NB4 derived resistant cell line had the classical A216V PML-B2 domain mutation while the UF1 cell line did not. Gene expression profiling revealed prominent dysregulation of the cellular metabolic pathways in the resistant cell lines. Glycolytic inhibition by 2-DG was efficient and comparable to the standard of care (ATO) in targeting the sensitive APL cell lines and was also effective in the in vivo transplantable APL mouse model; however, it did not affect the ATO resistant cell lines. The survival of the resistant cell lines was significantly affected by compounds targeting the mitochondrial respiration irrespective of the existence of ATO resistance-conferring genetic mutations. Our data demonstrate the addition of mitocans can overcome ATO resistance. We further demonstrated that the combination of ATO and mitocans has the potential in the treatment of non-M3 AML and the translation of this approach in the clinic needs to be explored further.Key pointsMetabolic rewiring promotes ATO resistance, which can be overcome by targeting mitochondrial oxidative phosphorylation.Combination of ATO and mitocans can be exploited as a potential therapeutic option for relapsed APL and in non-M3 AML patients.


Sign in / Sign up

Export Citation Format

Share Document