bims-mascan Biomed News
on Mass spectrometry in cancer research
Issue of 2020–11–08
23 papers selected by
Giovanny Rodriguez Blanco, University of Edinburgh



  1. Anal Chim Acta. 2020 Nov 15. pii: S0003-2670(20)30908-9. [Epub ahead of print]1137 74-84
      Effects of blood collection tubes, the time period, the sample origin, and the method used on the lipidomic profile are investigated by ultrahigh-performance supercritical fluid chromatography - mass spectrometry (UHPSFC/MS) and hydrophilic interaction liquid chromatography ultrahigh-performance liquid chromatography - mass spectrometry (HILIC-UHPLC/MS). Heparin plasma samples have been obtained from 99 healthy volunteers at three time points separated by six-month intervals together with one collection for EDTA plasma and serum. Furthermore, lipid concentrations in heparin plasma collected at two different sites are compared. 171 lipid species from eight lipid classes are quantified with UHPSFC/MS, and 122 lipid species from four lipid classes with HILIC-UHPLC/MS. The accuracy of both methods is monitored by the quantitation error using two internal standards (IS) per individual lipid classes. No significant differences in lipid profiles are observed for different time points and types of collection tubes (heparin plasma, EDTA plasma, and serum). Most pronounced lipid concentration differences are observed for the comparison of NIST SRM 1950 human plasma and mean lipid concentrations of the investigated cohort. Furthermore, differences in lipid concentrations are observed between employed methods (UHPSFC/MS vs. HILIC-UHPLC/MS), which can be compensated by the normalization using NIST SRM 1950 human plasma used as the quality control sample.
    Keywords:  Blood collection; Human plasma; Lipidomics; Lipids; Liquid chromatography; Mass spectrometry; Serum; Supercritical fluid chromatography
    DOI:  https://doi.org/10.1016/j.aca.2020.08.061
  2. Cancer Discov. 2020 Nov 02.
      Cancer cells continuously rewire their metabolism to fulfill their need for rapid growth and survival while subject to changes in environmental cues. Thus, a vital component of a cancer cell lies in its metabolic adaptability. The constant demand for metabolic alterations requires flexibility, that is, the ability to utilize different metabolic substrates; as well as plasticity, that is, the ability to process metabolic substrates in different ways. In this review, we discuss how dynamic changes in cancer metabolism affect tumor progression and the consequential implications for cancer therapy. SIGNIFICANCE: Recognizing cancer dynamic metabolic adaptability as an entity can lead to targeted therapy that is expected to decrease drug resistance.
    DOI:  https://doi.org/10.1158/2159-8290.CD-20-0844
  3. Cancers (Basel). 2020 Nov 03. pii: E3244. [Epub ahead of print]12(11):
      Cancer is a complex disease that includes the reprogramming of metabolic pathways by malignant proliferating cells, including those affecting the tumor microenvironment (TME). The "TME concept" was introduced in recognition of the roles played by factors other than tumor cells in cancer progression. In response to the hypoxic or semi-hypoxic characteristic of the TME, cancer cells generate a large amount of lactate via the metabolism of glucose and glutamine. Export of this newly generated lactate by the tumor cells together with H+ prevents intracellular acidification but acidifies the TME. In recent years, the importance of lactate and acidosis in carcinogenesis has gained increasing attention, including the role of lactate as a tumor-promoting metabolite. Here we review the existing literature on lactate metabolism in tumor cells and the ability of extracellular lactate to direct the metabolic reprogramming of those cells. Studies demonstrating the roles of lactate in biological processes that drive or sustain carcinogenesis (tumor promotion, angiogenesis, metastasis and tumor resistance) and lactate's role as an immunosuppressor that contributes to tumor evasion are also considered. Finally, we consider recent therapeutic efforts using available drugs directed at and interfering with lactate production and transport in cancer treatment.
    Keywords:  LDH; MCTs; lactate; tumor microenvironment (TME), acidosis
    DOI:  https://doi.org/10.3390/cancers12113244
  4. Cell Metab. 2020 Oct 28. pii: S1550-4131(20)30550-7. [Epub ahead of print]
      Pancreatic ductal adenocarcinoma (PDAC) cells require substantial metabolic rewiring to overcome nutrient limitations and immune surveillance. However, the metabolic pathways necessary for pancreatic tumor growth in vivo are poorly understood. To address this, we performed metabolism-focused CRISPR screens in PDAC cells grown in culture or engrafted in immunocompetent mice. While most metabolic gene essentialities are unexpectedly similar under these conditions, a small fraction of metabolic genes are differentially required for tumor progression. Among these, loss of heme synthesis reduces tumor growth due to a limiting role of heme in vivo, an effect independent of tissue origin or immune system. Our screens also identify autophagy as a metabolic requirement for pancreatic tumor immune evasion. Mechanistically, autophagy protects cancer cells from CD8+ T cell killing through TNFα-induced cell death in vitro. Altogether, this resource provides metabolic dependencies arising from microenvironmental limitations and the immune system, nominating potential anti-cancer targets.
    Keywords:  cancer metabolism; in vivo CRISPR screen; pancreatic cancer; tumor immune evasion
    DOI:  https://doi.org/10.1016/j.cmet.2020.10.017
  5. Cell Metab. 2020 Oct 28. pii: S1550-4131(20)30551-9. [Epub ahead of print]
      Pancreatic ductal adenocarcinoma (PDA) is a deadly cancer characterized by complex metabolic adaptations that promote survival in a severely hypoxic and nutrient-limited tumor microenvironment (TME). Modeling microenvironmental influences in cell culture has been challenging, and technical limitations have hampered the comprehensive study of tumor-specific metabolism in vivo. To systematically interrogate metabolic vulnerabilities in PDA, we employed parallel CRISPR-Cas9 screens using in vivo and in vitro systems. This work revealed striking overlap of in vivo metabolic dependencies with those in vitro. Moreover, we identified that intercellular nutrient sharing can mask dependencies in pooled screens, highlighting a limitation of this approach to study tumor metabolism. Furthermore, metabolic dependencies were similar between 2D and 3D culture, although 3D culture may better model vulnerabilities that influence certain oncogenic signaling pathways. Lastly, our work demonstrates the power of genetic screening approaches to define in vivo metabolic dependencies and pathways that may have therapeutic utility.
    Keywords:  cancer cell signaling; metabolism; nutrient crosstalk; pancreatic cancer; tumor microenvironment
    DOI:  https://doi.org/10.1016/j.cmet.2020.10.018
  6. Cancer Lett. 2020 Oct 24. pii: S0304-3835(20)30490-0. [Epub ahead of print]
      The acyl-CoA thioesterase (ACOT) family catalyses the hydrolysis of acyl-CoA thioesters to their corresponding non-esterified fatty acid and coenzyme A (CoA). Increasing evidence suggests that cancer cells generally have altered lipid metabolism in different aspects. However, the roles of the ACOT family in cancer, especially in pancreatic ductal carcinoma (PDAC), are largely unknown. In the present study, we mined data to determine the clinical significance of all eleven ACOT genes among nine major solid tumour types from TCGA database and found that the expression of ACOT4 in PDAC was negatively correlated with patient survival, establishing ACOT4 as a potential biomarker of PDAC. Depletion of ACOT4 attenuated the proliferation and tumour formation of PDAC cells. Using mass spectrometry, HSPA1A was found to associate with ACOT4. Furthermore, we found that phosphorylation of ACOT4 at S392 by AKT decreased the binding of ACOT4 to HSPA1A, resulting in ACOT4 accumulation. The ACOT4 elevation promotes pancreatic tumourigenesis by producing excessive CoA to support tumour cell metabolism. Thus, our study expands the relationship between AKT signalling and lipid metabolism and establishes a functional role of ACOT4 in PDAC.
    Keywords:  ACOT4; AKT; Coenzyme A; HSPA1A; Pancreatic cancer
    DOI:  https://doi.org/10.1016/j.canlet.2020.09.022
  7. FASEB J. 2020 Nov 01.
      Fragile X-associated Tremor/Ataxia Syndrome (FXTAS) is a neurodegenerative disorder associated with the FMR1 premutation. It is currently unknown when, and if, individual premutation carriers will develop FXTAS. Thus, with the aim of identifying biomarkers for early diagnosis, development, and progression of FXTAS, we performed global metabolomic profiling of premutation carriers (PM) who, as part of an ongoing longitudinal study, emerged into two distinct categories: those who developed symptoms of FXTAS (converters, CON) at subsequent visits and those who did not (non-converters, NCON) and we compared to age-matched healthy controls (HC). We assessed CGG repeat allele size by Southern Blot and PCR analysis. Metabolomic profile was obtained by ultra-performance liquid chromatography, accurate mass spectrometer, and an Orbitrap mass analyzer. In this study we found 47 metabolites were significantly dysregulated between HC and the premutation groups (PM). Importantly, we identified 24 metabolites that showed significant changes in expression in the CON as compared to the NCON both at V1 and V2, and 70 metabolites in CON as compared to NCON but only at V2. These findings suggest the potential role of the identified metabolites as biomarkers for early diagnosis and for FXTAS disease progression, respectively. Interestingly, the majority of the identified metabolites were lipids, followed by amino acids. To our knowledge, this the first report of longitudinal metabolic profiling and identification of unique biomarkers of FXTAS. The lipid metabolism and specifically the sub pathways involved in mitochondrial bioenergetics, as observed in other neurodegenerative disorders, are significantly altered in FXTAS.
    Keywords:   FMR1 ; fatty acids; fragile X-associated tremor/ataxia syndrome; metabolomic; molecular biomarkers
    DOI:  https://doi.org/10.1096/fj.202001880R
  8. Anal Chim Acta. 2020 Nov 15. pii: S0003-2670(20)30993-4. [Epub ahead of print]1137 156-169
      Lipids are vital biological molecules and play multiple roles in cellular function of mammalian organisms such as cellular membrane anchoring, signal transduction, material trafficking and energy storage. Driven by the biological significance of lipids, lipidomics has become an emerging science in the field of omics. Lipidome in biological systems consists of hundreds of thousands of individual lipid molecules that possess complex structures, multiple categories, and diverse physicochemical properties assembled by different combinations of polar headgroups and hydrophobic fatty acyl chains. Such structural complexity poses a huge challenge for comprehensive lipidome analysis. Thanks to the great innovations in chromatographic separation techniques and the continuous advances in mass spectrometric detection tools, analytical strategies for lipidomics have been highly diversified so that the depth and breadth of lipidomics have been greatly enhanced. This review will present the current state of mass spectrometry-based analytical strategies including untargeted, targeted and pseudotargeted lipidomics. Recent typical applications of lipidomics in biomarker discovery, pathogenic mechanism and therapeutic strategy are summarized, and the challenges facing to the field of lipidomics are also discussed.
    Keywords:  Biomarker discovery; High-resolution mass spectrometry; Liquid chromatography; Pseudotargeted lipidomics; Targeted lipidomics; Untargeted lipidomics
    DOI:  https://doi.org/10.1016/j.aca.2020.09.060
  9. Expert Rev Proteomics. 2020 Nov 03.
      Introduction. - Saliva is an ideal biofluid that can be collected in a non-invasive manner, enabling safe and frequent screening of various diseases. Recent studies have revealed that salivary metabolomics analysis has the potential to detect both oral and systemic cancers. Area covered. - We reviewed the technical aspects, as well as applications, of salivary metabolomics for cancer detection. The topics include the effects of preconditioning and the method of sample collection, sample storage, processing, measurement, data analysis, and validation of the results. We also examined the rational relationship between salivary biomarkers and tumors distant from the oral cavity. A strategy to establish standard operating protocols for obtaining reproducible quantification data is also discussed Expert opinion. - Salivary metabolomics reflects oral and systematic health status, which potently enables cancer detection. The sensitivity and specificity of each marker and their combinations have been well evaluated, but a validation study is required. Further, the standard operating protocol for each procedure should be established to obtain reproducible data before clinical usage.
    Keywords:  Biomarker; cancer; mass spectrometry; metabolomics; saliva
    DOI:  https://doi.org/10.1080/14789450.2020.1846524
  10. Nat Commun. 2020 Nov 06. 11(1): 5625
      The human metabolome provides a window into the mechanisms and biomarkers of various diseases. However, because of limited availability, many sample types are still difficult to study by metabolomic analyses. Here, we present a mass spectrometry (MS)-based metabolomics strategy that only consumes sub-nanoliter sample volumes. The approach consists of combining a customized metabolomics workflow with a pulsed MS ion generation method, known as triboelectric nanogenerator inductive nanoelectrospray ionization (TENGi nanoESI) MS. Samples tested with this approach include exhaled breath condensate collected from cystic fibrosis patients as well as in vitro-cultured human mesenchymal stromal cells. Both test samples are only available in minimum amounts. Experiments show that picoliter-volume spray pulses suffice to generate high-quality spectral fingerprints, which increase the information density produced per unit sample volume. This TENGi nanoESI strategy has the potential to fill in the gap in metabolomics where liquid chromatography-MS-based analyses cannot be applied. Our method opens up avenues for future investigations into understanding metabolic changes caused by diseases or external stimuli.
    DOI:  https://doi.org/10.1038/s41467-020-19444-y
  11. Br J Cancer. 2020 Nov 03.
       BACKGROUND: To circumvent Warburg effect, several clinical trials for different cancers are utilising a combinatorial approach using metabolic reprogramming and chemotherapeutic agents including metformin. The majority of these metabolic interventions work via indirectly activating AMP-activated protein kinase (AMPK) to alter cellular metabolism in favour of oxidative phosphorylation over aerobic glycolysis. The effect of these drugs is dependent on glycaemic and insulin conditions.  Therefore, development of small molecules, which can activate AMPK, irrespective of the energy state, may be a better approach for triple-negative breast cancer (TNBC) treatment.
    METHODS: Therapeutic effect of SU212 on TNBC cells was examined using in vitro and in vivo models.
    RESULTS: We developed and characterised the efficacy of novel AMPK activator (SU212) that selectively induces oxidative phosphorylation and decreases glycolysis in TNBC cells, while not affecting these pathways in normal cells.   SU212 accomplished this metabolic reprogramming by activating AMPK independent of energy stress and irrespective of the glycaemic/insulin state. This leads to mitotic phase arrest and apoptosis in TNBC cells. In vivo, SU212 inhibits tumour growth, cancer progression and metastasis.
    CONCLUSIONS: SU212 directly activates AMPK in TNBC cells, but does not hamper glucose metabolism in normal cells. Our study provides compelling preclinical data for further development of SU212 for the treatment of TNBC.
    DOI:  https://doi.org/10.1038/s41416-020-01137-4
  12. Mol Cell. 2020 Nov 05. pii: S1097-2765(20)30689-4. [Epub ahead of print]80(3): 554
      
    DOI:  https://doi.org/10.1016/j.molcel.2020.10.006
  13. Arch Biochem Biophys. 2020 Oct 31. pii: S0003-9861(20)30668-8. [Epub ahead of print] 108659
      Metabolic reprogramming confers cancer cells plasticity and viability under harsh conditions. Such active alterations lead to cell metabolic dependency, which can be exploited as an attractive target in development of effective antitumor therapies. Similar to cancer cells, activated T cells also execute global metabolic reprogramming for their proliferation and effector functions when recruited to the tumor microenvironment (TME). However, the high metabolic activity of rapidly proliferating cancer cells can compete for nutrients with immune cells in the TME, and consequently, suppressing their anti-tumor functions. Thus, therapeutic strategies could aim to restore T cell metabolism and anti-tumor responses in TME by targeting the metabolic dependence of cancer cells. In this review, we highlight current research progress on metabolic reprogramming and the interplay between cancer cells and immune cells. We also discuss potential therapeutic intervention strategies for targeting metabolic pathways to improve tumor immunotherapy efficacy.
    Keywords:  Metabolic reprogramming; Nutrients competing; Tumor microenvironment; Tumor-infiltrating lymphocytes
    DOI:  https://doi.org/10.1016/j.abb.2020.108659
  14. Mol Cell. 2020 Oct 26. pii: S1097-2765(20)30693-6. [Epub ahead of print]
      Cancer-associated mutations that stabilize NRF2, an oxidant defense transcription factor, are predicted to promote tumor development. Here, utilizing 3D cancer spheroid models coupled with CRISPR-Cas9 screens, we investigate the molecular pathogenesis mediated by NRF2 hyperactivation. NRF2 hyperactivation was necessary for proliferation and survival in lung tumor spheroids. Antioxidant treatment rescued survival but not proliferation, suggesting the presence of distinct mechanisms. CRISPR screens revealed that spheroids are differentially dependent on the mammalian target of rapamycin (mTOR) for proliferation and the lipid peroxidase GPX4 for protection from ferroptosis of inner, matrix-deprived cells. Ferroptosis inhibitors blocked death from NRF2 downregulation, demonstrating a critical role of NRF2 in protecting matrix-deprived cells from ferroptosis. Interestingly, proteomics analyses show global enrichment of selenoproteins, including GPX4, by NRF2 downregulation, and targeting NRF2 and GPX4 killed spheroids overall. These results illustrate the value of spheroid culture in revealing environmental or spatial differential dependencies on NRF2 and reveal exploitable vulnerabilities of NRF2-hyperactivated tumors.
    Keywords:  oxidative stress, cancer, cell death, ferroptosis, 3D culture, NRF2, CRISPR screening, selenoprotein, GPX4
    DOI:  https://doi.org/10.1016/j.molcel.2020.10.010
  15. Anal Chim Acta. 2020 Nov 15. pii: S0003-2670(20)30912-0. [Epub ahead of print]1137 37-46
      Despite the growing popularity of liquid chromatography-mass spectrometry (LC-MS)-based metabolomics, no study has yet to systematically compare the performance of different data acquisition modes in the discovery of significantly altered metabolic features, which is an important task of untargeted metabolomics for identifying clinical biomarkers and elucidating disease mechanism in comparative samples. In this work, we performed a comprehensive comparison of three most commonly used data acquisition modes, including full-scan, data-dependent acquisition (DDA), and data-independent acquisition (DIA), using a metabolomics study of human plasma samples from leukemia patients before and after one-month chemotherapy. After optimization of data processing parameters, we extracted and compared statistically significant metabolic features from the results of each data acquisition mode. We found that most significant features can be consistently found in all three data acquisition modes with similar statistical performance as evaluated by Pearson correlation and receiver operating characteristic (ROC) analysis. Upon comparison, DDA mode consistently generated fewer uniquely found significant features than full-scan and DIA modes. We then manually inspected over 2000 uniquely discovered significant features in each data acquisition mode and showed that these features can be generally categorized into four major types. Many significant features were missed in DDA mode, primarily due to its low capability of detecting or extracting these features from raw LC-MS data. We thus proposed a bioinformatic solution to rescue these missing significant features from the raw DDA data with good reproducibility and accuracy. Overall, our work asserts that data acquisition modes can influence metabolomics results, suggesting room for improvement of data acquisition modes for untargeted metabolomics.
    Keywords:  Data acquisition mode; Data-dependent acquisition; Data-independent acquisition; Full-scan; Liquid chromatography-mass spectrometry; Untargeted metabolomics
    DOI:  https://doi.org/10.1016/j.aca.2020.08.065
  16. J Pharm Sci. 2020 Oct 24. pii: S0022-3549(20)30626-2. [Epub ahead of print]
      Determination of abundances of proteins involved in uptake, distribution, metabolism and excretion of xenobiotics is a prerequisite to understand and predict elimination mechanisms in tissue. Mass spectrometry promises simple and accurate measurements of individual proteins in complex mixtures using isotopically labeled peptide standards. However, comparisons of measurements performed in different laboratories have shown considerable discrepancies in the data generated. Even when very similar approaches are compared, the results differ significantly. An alternative method of measuring protein titers is global proteomics. Depending on sample type, this allows quantification of hundreds to thousands of proteins in a single analysis. It enables system-wide insights by providing protein copy numbers and cell sizes. Regardless of differences, the workflows of both the labeled standard-based and the proteomic approach share several steps. Each can be critical. Selection of optimal techniques is the prerequisite for accurate and reproducible protein quantification.
    Keywords:  Absorption, distribution, metabolism, and excretion (ADME); Enzyme(s); Omics; Proteomic; Transporter(s)
    DOI:  https://doi.org/10.1016/j.xphs.2020.10.034
  17. Anal Bioanal Chem. 2020 Nov 02.
      Mass spectrometry (MS) has become the de facto tool for routine quantitative analysis of biomolecules. MS is increasingly being used to reveal the spatial distribution of proteins, metabolites, and pharmaceuticals in tissue and interest in this area has led to a number of novel spatially resolved MS technologies. Most spatially resolved MS measurements are qualitative in nature due to a myriad of potential biases, such as sample heterogeneity, sampling artifacts, and ionization effects. As applications of spatially resolved MS in the pharmacological and clinical fields increase, demand has become high for quantitative MS imaging and profiling data. As a result, several varied technologies now exist that provide differing levels of spatial and quantitative information. This review provides an overview of MS profiling and imaging technologies that have demonstrated quantitative analysis from tissue. Focus is given on the fundamental processes affecting quantitative analysis in an array of MS imaging and profiling technologies and methods to address these biases.Graphical abstract.
    Keywords:  Absolute; Imaging; Mass spectrometry; Profiling; Quantitation; Tissue
    DOI:  https://doi.org/10.1007/s00216-020-02964-3
  18. Biochim Biophys Acta Mol Cell Biol Lipids. 2020 Nov 03. pii: S1388-1981(20)30246-8. [Epub ahead of print] 158854
      Inflammatory bowel diseases (IBDs), such as Crohn's disease and ulcerative colitis, are lifelong diseases that remain challenging to treat. IBDs are characterized by alterations in intestinal barrier function and dysregulation of the innate and adaptive immunity. An increasing number of lipids are found to be important regulators of inflammation and immunity as well as gut physiology. Therefore, the study of lipid mediators in IBDs is expected to improve our understanding of disease pathogenesis and lead to novel therapeutic opportunities. Here, through selected examples - such as fatty acids, specialized proresolving mediators, lysophospholipids, endocannabinoids, and oxysterols - we discuss how lipid signaling is involved in IBD physiopathology and how modulating lipid signaling pathways could affect IBD.
    Keywords:  SPM; autotaxin; docosahexaenoic acid; eicosanoids; palmitoylethanolamide; short chain fatty acids
    DOI:  https://doi.org/10.1016/j.bbalip.2020.158854
  19. Biochim Biophys Acta Mol Cell Biol Lipids. 2020 Nov 03. pii: S1388-1981(20)30241-9. [Epub ahead of print] 158849
      This review provides an overview of lipids and lipid metabolism in relation to COVID-19, with special attention on cholesterol. Cholesterol enriched lipid rafts represent a platform for viruses to enter the host cell by endocytosis. Generally, higher membrane cholesterol coincides with higher efficiency of COVID-19 entry. Inversely, patients with COVID-19 show lowered levels of blood cholesterol, high-density and low-density lipoproteins. The modulated efficiency of viral entry can be explained by availability of SR-B1 and LDL-receptors. Especially HDL seems to have a variety of roles, from being itself a scavenger for viruses, an immune modulator and mediator of viral entry. Due to inverse roles of membrane cholesterol and lipoprotein cholesterol in COVID-19 infected patients, treatment of these patients with cholesterol lowering statins remains controversial. In conclusion, cholesterol and lipoproteins are potential markers for monitoring the viral infection status where mechanistic inconsistencies warrant immediate further research.
    Keywords:  COVID-19; SARS-CoV-2; cholesterol; high density lipoprotein; low density lipoprotein; statins
    DOI:  https://doi.org/10.1016/j.bbalip.2020.158849
  20. J Clin Med. 2020 Oct 29. pii: E3490. [Epub ahead of print]9(11):
      In these recent years, a growing interest with regard to polyunsaturated fatty acids (PUFAs) and microbiota relationship has been noted [...].
    Keywords:  DHA; EPA; Omega-3 fatty acids; cancer; docosahexaenoic acid; eicosapentaenoic acid; microbiome; microbiota; omega-6 fatty acids; tumor
    DOI:  https://doi.org/10.3390/jcm9113490
  21. Nat Commun. 2020 11 05. 11(1): 5587
      Human skin provides both physical integrity and immunological protection from the external environment using functionally distinct layers, cell types and extracellular matrix. Despite its central role in human health and disease, the constituent proteins of skin have not been systematically characterized. Here, we combine advanced tissue dissection methods, flow cytometry and state-of-the-art proteomics to describe a spatially-resolved quantitative proteomic atlas of human skin. We quantify 10,701 proteins as a function of their spatial location and cellular origin. The resulting protein atlas and our initial data analyses demonstrate the value of proteomics for understanding cell-type diversity within the skin. We describe the quantitative distribution of structural proteins, known and previously undescribed proteins specific to cellular subsets and those with specialized immunological functions such as cytokines and chemokines. We anticipate that this proteomic atlas of human skin will become an essential community resource for basic and translational research ( https://skin.science/ ).
    DOI:  https://doi.org/10.1038/s41467-020-19383-8