1
|
Withnell E, Zhang X, Sun K, Guo Y. XOmiVAE: an interpretable deep learning model for cancer classification using high-dimensional omics data. Brief Bioinform 2021; 22:bbab315. [PMID: 34402865 PMCID: PMC8575033 DOI: 10.1093/bib/bbab315] [Citation(s) in RCA: 19] [Impact Index Per Article: 6.3] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/26/2021] [Revised: 07/04/2021] [Accepted: 07/20/2021] [Indexed: 12/26/2022] Open
Abstract
The lack of explainability is one of the most prominent disadvantages of deep learning applications in omics. This 'black box' problem can undermine the credibility and limit the practical implementation of biomedical deep learning models. Here we present XOmiVAE, a variational autoencoder (VAE)-based interpretable deep learning model for cancer classification using high-dimensional omics data. XOmiVAE is capable of revealing the contribution of each gene and latent dimension for each classification prediction and the correlation between each gene and each latent dimension. It is also demonstrated that XOmiVAE can explain not only the supervised classification but also the unsupervised clustering results from the deep learning network. To the best of our knowledge, XOmiVAE is one of the first activation level-based interpretable deep learning models explaining novel clusters generated by VAE. The explainable results generated by XOmiVAE were validated by both the performance of downstream tasks and the biomedical knowledge. In our experiments, XOmiVAE explanations of deep learning-based cancer classification and clustering aligned with current domain knowledge including biological annotation and academic literature, which shows great potential for novel biomedical knowledge discovery from deep learning models.
Collapse
Affiliation(s)
- Eloise Withnell
- Data Science Institute Imperial College London, SW7 2AZ London, UK
- Department of Health Informatics University College London, WC1E 6BT London, UK
| | - Xiaoyu Zhang
- Data Science Institute Imperial College London, SW7 2AZ London, UK
| | - Kai Sun
- Data Science Institute Imperial College London, SW7 2AZ London, UK
| | - Yike Guo
- Data Science Institute Imperial College London, SW7 2AZ London, UK
- Department of Computer Science Hong Kong Baptist University, Hong Kong China
| |
Collapse
|
2
|
A comprehensive characterization of the cell-free transcriptome reveals tissue- and subtype-specific biomarkers for cancer detection. Nat Commun 2021; 12:2357. [PMID: 33883548 PMCID: PMC8060291 DOI: 10.1038/s41467-021-22444-1] [Citation(s) in RCA: 95] [Impact Index Per Article: 31.7] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/03/2020] [Accepted: 03/10/2021] [Indexed: 12/12/2022] Open
Abstract
Cell-free RNA (cfRNA) is a promising analyte for cancer detection. However, a comprehensive assessment of cfRNA in individuals with and without cancer has not been conducted. We perform the first transcriptome-wide characterization of cfRNA in cancer (stage III breast [n = 46], lung [n = 30]) and non-cancer (n = 89) participants from the Circulating Cell-free Genome Atlas (NCT02889978). Of 57,820 annotated genes, 39,564 (68%) are not detected in cfRNA from non-cancer individuals. Within these low-noise regions, we identify tissue- and cancer-specific genes, defined as “dark channel biomarker” (DCB) genes, that are recurrently detected in individuals with cancer. DCB levels in plasma correlate with tumor shedding rate and RNA expression in matched tissue, suggesting that DCBs with high expression in tumor tissue could enhance cancer detection in patients with low levels of circulating tumor DNA. Overall, cfRNA provides a unique opportunity to detect cancer, predict the tumor tissue of origin, and determine the cancer subtype. Cell-free RNA (cfRNA) is a promising analyte for cancer diagnosis. Here, the authors determine the baseline cell-free transcriptome in the absence of cancer and identify tissue- and subtype-specific cfRNA biomarkers in breast and lung cancer patients.
Collapse
|
3
|
Dexamethasone Treatment Increases the Intracellular Calcium Level Through TRPV6 in A549 Cells. Int J Mol Sci 2020; 21:ijms21031050. [PMID: 32033337 PMCID: PMC7037905 DOI: 10.3390/ijms21031050] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/13/2019] [Revised: 02/03/2020] [Accepted: 02/04/2020] [Indexed: 12/04/2022] Open
Abstract
This study investigated the effect of dexamethasone (DEX) on intracellular calcium levels and the expressions of transient receptor potential cation channel subcomponent V member 6 (TRPV6), sodium-calcium exchanger 1 (NCX1), and plasma membrane calcium ATPase 1 (PMCA1) in A549 cells. The intracellular calcium level, by using the calcium indicator pGP-CMV-GCaMP6f, increased following DEX treatment for 6, 12, and 24 h in A549 cells. In addition, Rhod-4 assay after DEX treatment for 24 h showed that DEX increased the level of intracellular calcium. The expression of the calcium influx TRPV6 gene significantly increased, whereas the expressions of the calcium outflow NCX1 and PMCA1 genes significantly decreased with DEX treatment. The mRNA levels of surfactant protein genes SFTPA1, SFTPB, SFTPC, and SFTPD and the secreted airway mucin genes MUC1 and MUC5AC were investigated by treating cells with DEX. The DEX treatment decreased the mRNA levels of SFTPA1 and SFTPB but increased the mRNA levels of SFTPC and SFTPD. The MUC1 mRNA level was increased by DEX treatment, whereas MUC5AC mRNA was significantly decreased. These results indicate that DEX influences the intracellular calcium level through TRPV6, and affects pulmonary surfactant genes and secreted airway mucin genes in A549 cells.
Collapse
|
4
|
He W, Li W, Jiang B, Chang L, Jin C, Tu C, Li Y. Correlation between epidermal growth factor receptor tyrosine kinase inhibitor efficacy and circulating tumor cell levels in patients with advanced non-small cell lung cancer. Onco Targets Ther 2016; 9:7515-7520. [PMID: 28003764 PMCID: PMC5161394 DOI: 10.2147/ott.s115221] [Citation(s) in RCA: 16] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/10/2023] Open
Abstract
Objective The aim of this study was to investigate the correlation between the efficacy of epidermal growth factor receptor tyrosine kinase inhibitor (EGFR-TKI) and circulating tumor cell (CTC) levels in patients with advanced non-small cell lung cancer (NSCLC). The efficacy of EGFR-TKIs in reducing CTC counts in patients with advanced NSCLC was studied. Patients and methods A total of 66 patients with advanced NSCLC were enrolled and divided into two groups (those with high CTC counts and those with low CTC counts) based on the patients’ median CTC counts. All the patients were treated with an EGFR-TKI, and the treatment efficacy and prognoses were compared. Results The treatment efficacies were 53.3% (16/30) and 27.8% (10/36) for the low CTC group and high CTC group, respectively, and this difference was statistically significant (P<0.05). The median overall survival was 22.8 months (95% confidence interval [CI]: 18.9–26.8 months) for the low CTC group and 18.3 months (95% CI: 2.9–8.2 months) for the high CTC group. The median progression-free survival was 11.5 months (95% CI: 8.1–15 months) and 5.6 months (95% CI: 2.9–8.2 months) for the low and high CTC groups, respectively, and the difference was statistically significant (P<0.05). Conclusion The CTC count can be used as an index for predicting the EGFR-TKI effect on patients with advanced NSCLC. Efficacy and prognosis of EGFR-TKI treatment and CTC count were considered important, and the CTC count could be used to predict the efficacy of EGFR-TKI treatment and prognosis of advanced NSCLC. The change in CTC expression levels can be used as an index for evaluating the prognosis of patients with advanced NSCLC.
Collapse
Affiliation(s)
- Wenjie He
- Department of Cadre's Medical Oncology, Yunnan Tumor Hospital, The Third Affiliated Hospital of Kunming Medical University
| | - Wenhui Li
- Department of Cadre's Medical Oncology, Yunnan Tumor Hospital, The Third Affiliated Hospital of Kunming Medical University
| | - Bo Jiang
- Department of Cadre's Medical Oncology, Yunnan Tumor Hospital, The Third Affiliated Hospital of Kunming Medical University
| | - Li Chang
- Department of Cadre's Medical Oncology, Yunnan Tumor Hospital, The Third Affiliated Hospital of Kunming Medical University
| | - Congguo Jin
- Department of Oncology Research Institution, Kunming, People's Republic of China
| | - Changlin Tu
- Department of Cadre's Medical Oncology, Yunnan Tumor Hospital, The Third Affiliated Hospital of Kunming Medical University
| | - Yunfen Li
- Department of Cadre's Medical Oncology, Yunnan Tumor Hospital, The Third Affiliated Hospital of Kunming Medical University
| |
Collapse
|
5
|
Li J, Shi SB, Shi WL, Wang Y, Yu LC, Zhu LR, Ge LP. LUNX mRNA-positive cells at different time points predict prognosis in patients with surgically resected nonsmall cell lung cancer. Transl Res 2014; 163:27-35. [PMID: 24120632 DOI: 10.1016/j.trsl.2013.09.010] [Citation(s) in RCA: 15] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 07/16/2013] [Revised: 09/15/2013] [Accepted: 09/19/2013] [Indexed: 01/09/2023]
Abstract
LUNX is a lung-specific gene whose messenger ribonucleic acid (mRNA) expression is strictly limited to normal lung tissue and nonsmall cell lung cancer (NSCLC) tissue. The aim of this study was to investigate whether the detection of LUNX mRNA-positive circulating tumor cells (CTC)s in peripheral blood at different time points is useful for predicting disease recurrence, disease-free survival (DFS), and overall survival (OS) in NSCLC patients undergoing surgery. Serial blood samples from 68 patients with stage I-IIIA NSCLC were examined by real-time quantitative polymerase chain reaction assay targeting LUNX mRNA before (T0) and after surgery (T1) and after the completion of adjuvant chemotherapy (T2). Results showed that LUNX mRNA-positive CTCs were detected in 40 of 68 NSCLC patients (58.8%) before surgery; the detection rates of LUNX mRNA-positive CTCs at T1 and T2 time points were 32.4% (22/68) and 33.3% (20/60), respectively. The detection of LUNX mRNA-positive CTC at 3 time points was associated with lymph node status and pathologic stage. During the follow-up period, patients with LUXN mRNA-positive CTC at 3 time points had a higher relapse rate and a shorter DFS and OS than those without. Multivariate analysis revealed that presence of LUNX mRNA-positive CTC at T1 and T2 time points was an independent unfavorable factor for DFS and OS. In conclusion, detection of LUNX mRNA-positive CTC after surgery and the completion of adjuvant chemotherapy in patients with stage I-IIIA NSCLC are highly predictive for DFS and OS. This technique could aid in the prediction of prognosis and design of tailored treatment.
Collapse
Affiliation(s)
- Jian Li
- Department of Pulmonary Medicine, Department of Thoracic Surgery, and Center of Experimental Medicine, Affiliated Hospital of Jiangsu University, Zhenjiang, China.
| | - Shun-Bing Shi
- Department of Pulmonary Medicine, Department of Thoracic Surgery, and Center of Experimental Medicine, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| | - Wei-Lin Shi
- Department of Pulmonary Medicine, Department of Thoracic Surgery, and Center of Experimental Medicine, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| | - Yi Wang
- Department of Pulmonary Medicine, Department of Thoracic Surgery, and Center of Experimental Medicine, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| | - Li-Chao Yu
- Department of Pulmonary Medicine, Department of Thoracic Surgery, and Center of Experimental Medicine, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| | - Li-Rong Zhu
- Department of Pulmonary Medicine, Department of Thoracic Surgery, and Center of Experimental Medicine, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| | - Li-Ping Ge
- Department of Pulmonary Medicine, Department of Thoracic Surgery, and Center of Experimental Medicine, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| |
Collapse
|
6
|
Kurotani R, Kumaki N, Naizhen X, Ward JM, Linnoila RI, Kimura S. Secretoglobin 3A2/uteroglobin-related protein 1 is a novel marker for pulmonary carcinoma in mice and humans. Lung Cancer 2010; 71:42-8. [PMID: 20466451 DOI: 10.1016/j.lungcan.2010.04.001] [Citation(s) in RCA: 17] [Impact Index Per Article: 1.2] [Reference Citation Analysis] [Abstract] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/02/2009] [Revised: 02/06/2010] [Accepted: 04/01/2010] [Indexed: 10/19/2022]
Abstract
Secretoglobin (SCGB) 3A2, also called uteroglobin-related protein (UGRP) 1, is a downstream target for a homeodomain transcription factor NKX2-1, which is critical for the development of lung, thyroid and ventral forebrain. Both SCGB3A2 and NKX2-1 are expressed in airway epithelial cells and the latter also in alveolar Type II cells. NKX2-1 has been used clinically for diagnosis of human pulmonary tumors. Recently, the expression of SCGB3A2 was reported in human carcinomas, suggesting the use of this protein as a tumor marker. In this study, 28 lung tumors from aging B6;129 mice and nine lung adenocarcinomas from CC10TAg transgenic mice that express SV40 large T antigen under the mouse Scgb1a1 (CC10) gene promoter, were subjected to histopathological and immunohistochemical analyses for the expression of NKX2-1 and SCGB3A2. NKX2-1 was expressed in all types of tumors albeit more focally in carcinomas. In contrast, SCGB3A2 normally expressed in Clara cells, was negative in Type II cell hyperplasias and adenomas. However, it was expressed in alveolar Type II cell carcinomas and Clara cell adenocarcinomas. In these carcinomas, SCGB3A2 expression was observed in the portion of the tumor where NKX2-1 expression was reduced or almost abolished. As a comparison, the expression of SCGB3A2 and NKX2-1 from 23 human non-small cell lung carcinoma specimens was also examined. The results demonstrate that SCGB3A2 is a useful marker for diagnosis of pulmonary tumors both in mice and humans.
Collapse
Affiliation(s)
- Reiko Kurotani
- Laboratory of Metabolism, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892 USA
| | | | | | | | | | | |
Collapse
|
7
|
|
8
|
Clinical significance of detecting survivin-expressing circulating cancer cells in patients with non-small cell lung cancer. Lung Cancer 2008; 63:284-90. [PMID: 18606477 DOI: 10.1016/j.lungcan.2008.05.024] [Citation(s) in RCA: 46] [Impact Index Per Article: 2.9] [Reference Citation Analysis] [Abstract] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/20/2008] [Revised: 05/28/2008] [Accepted: 05/28/2008] [Indexed: 11/24/2022]
Abstract
We previously demonstrated that the detection of circulating cancer cells (CCC) expressing survivin mRNA could provide valuable information for predicting metastasis and recurrence in breast cancer. The objective of this study was to investigate the significance of detecting survivin-expressing CCC on the clinical outcomes of patients with non-small cell lung cancer (NSCLC). Peripheral blood samples collected from 143 NSCLC patients and 177 healthy volunteers were quantitatively evaluated using a technique developed in our laboratory that detected reverse transcription-polymerase chain reaction (RT-PCR) products based on a hybridisation-enzyme linked immunosorbant essay (ELISA), which we called RT-PCR ELISA. The presence of survivin-expressing CCC was detected in 63 cancer patients (44.1%) and was significantly associated with pathological T classification, nodal status, and disease stages (all P<0.001). During a follow-up period of 36 months, patients who had positive survivin expressions at the time of the initial assay test had a higher relapse rate and shorter survival time when compared to those who had negative survivin expressions (all P<0.001). Through multivariate analysis, the detection of survivin-expressing CCC was found to be an independent predictor for cancer recurrence (HR=43.5; 95% CI=2.67-70.9; P=0.008) and survival (HR=1.35; 95% CI=1.02-4.31; P=0.049). Thus, detection of survivin-expressing CCC could be used in the prediction of disease recurrence as well as in the prognosis of NSCLC.
Collapse
|
9
|
Berger MJ, Adams SD, Tigges BM, Sprague SL, Wang XJ, Collins DP, McKenna DH. Differentiation of umbilical cord blood-derived multilineage progenitor cells into respiratory epithelial cells. Cytotherapy 2006; 8:480-7. [PMID: 17050253 DOI: 10.1080/14653240600941549] [Citation(s) in RCA: 79] [Impact Index Per Article: 4.4] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/24/2022]
Abstract
BACKGROUND Umbilical cord blood (UCB) has been examined for the presence of stem cells capable of differentiating into cell types of all three embryonic layers (i.e. endo-, ecto- and mesoderm). The few groups reporting success have typically confirmed endodermal potential using hepatic differentiation. We report differentiation of human UCB-derived multipotent stem cells, termed multilineage progenitor cells (MLPC), into respiratory epithelial cells (i.e. type II alveolar cells). METHODS Using a cell separation medium (PrepaCyte-MLPC; BioE Inc.) and plastic adherence, MLPC were isolated from four of 16 UCB units (American Red Cross) and expanded. Cultures were grown to 80% confluence in mesenchymal stromal cell growth medium (MSCGM; Cambrex BioScience) prior to addition of small airway growth medium (SAGM; Cambrex BioScience), an airway maintenance medium. Following a 3-8-day culture, cells were characterized by light microscopy, transmission electron microscopy, immunofluorescence and reverse transcriptase (RT)-PCR. RESULTS MLPC were successfully differentiated into type II alveolar cells (four of four mixed lines; two of two clonal lines). Differentiated cells were characterized by epithelioid morphology with lamellar bodies. Both immunofluorescence and RT-PCR confirmed the presence of surfactant protein C, a protein highly specific for type II cells. DISCUSSION MLPC were isolated, expanded and then differentiated into respiratory epithelial cells using an off-the-shelf medium designed for maintenance of fully differentiated respiratory epithelial cells. To the best of our knowledge, this is the first time human non-embryonic multipotent stem cells have been differentiated into type II alveolar cells. Further studies to evaluate the possibilities for both research and therapeutic applications are necessary.
Collapse
Affiliation(s)
- M J Berger
- Department of Laboratory Medicine and Pathology, University of Minnesota Medical School, Minneapolis, Minnesota, USA
| | | | | | | | | | | | | |
Collapse
|