Plasma infrared fingerprinting with machine learning enables single-measurement multi-phenotype health screening.

Infrared spectroscopy is a powerful technique for probing the molecular profiles of complex biofluids, offering a promising avenue for high-throughput in vitro diagnostics. While several studies showcased its potential in detecting health conditions, a large-scale analysis of a naturally heterogeneous potential patient population has not been attempted. Using a population-based cohort, here we analyze 5,184 blood plasma samples from 3,169 individuals using Fourier transform infrared (FTIR) spectroscopy. Applying a multi-task classification to distinguish between dyslipidemia, hypertension, prediabetes, type 2 diabetes, and healthy states, we find that the approach can accurately single out healthy individuals and characterize chronic multimorbid states. We further identify the capacity to forecast the development of metabolic syndrome years in advance of onset. Dataset-independent testing confirms the robustness of infrared signatures against variations in sample handling, storage time, and measurement regimes. This study provides the framework that establishes infrared molecular fingerprinting as an efficient modality for populational health diagnostics.

Probing Blood Plasma Protein Glycosylation with Infrared Spectroscopy.

The health state of an individual is closely linked to the glycosylation patterns of his or her blood plasma proteins. However, obtaining this information requires cost- and time-efficient analytical methods. We put forward infrared spectroscopy, which allows label-free analysis of protein glycosylation but so far has only been applied to analysis of individual proteins. Although spectral information does not directly provide the molecular structure of the glycans, it is sensitive to changes therein and covers all types of glycosidic linkages. Combining single-step ion exchange chromatography with infrared spectroscopy, we developed a workflow that enables the separation and analysis of major protein classes in blood plasma. Our results demonstrate that infrared spectroscopy can identify different patterns and global levels of glycosylation of intact plasma proteins. To showcase the strengths and limitations of the proposed approach, we compare the glycoforms of human and bovine alpha-1-acid glycoproteins, which exhibit highly variable global levels of glycosylation. To independently evaluate our conclusions, the glycan moieties of human alpha-1-acid glycoprotein were further analyzed using an established glycomics workflow. Importantly, the chromatographic separation of blood plasma improves the detection of aberrant glycoforms of a given protein as compared to infrared spectroscopy of bulk plasma. The presented approach allows a time-efficient comparison of glycosylation patterns of multiple plasma proteins, opening new avenues for biomedical probing.

Limits and Prospects of Molecular Fingerprinting for Phenotyping Biological Systems Revealed through In Silico Modeling.

Molecular fingerprinting via vibrational spectroscopy characterizes the chemical composition of molecularly complex media which enables the classification of phenotypes associated with biological systems. However, the interplay between factors such as biological variability, measurement noise, chemical complexity, and cohort size makes it challenging to investigate their impact on how the classification performs. Considering these factors, we developed an in silico model which generates realistic, but configurable, molecular fingerprints. Using experimental blood-based infrared spectra from two cancer-detection applications, we validated the model and subsequently adjusted model parameters to simulate diverse experimental settings, thereby yielding insights into the framework of molecular fingerprinting. Intriguingly, the model revealed substantial improvements in classifying clinically relevant phenotypes when the biological variability was reduced from a between-person to a within-person level and when the chemical complexity of the spectra was reduced. These findings quantitively demonstrate the potential benefits of personalized molecular fingerprinting and biochemical fractionation for applications in health diagnostics.

Breast-cancer detection using blood-based infrared molecular fingerprints.

BACKGROUND: Breast cancer screening is currently predominantly based on mammography, tainted with the occurrence of both false positivity and false negativity, urging for innovative strategies, as effective detection of early-stage breast cancer bears the potential to reduce mortality. Here we report the results of a prospective pilot study on breast cancer detection using blood plasma analyzed by Fourier-transform infrared (FTIR) spectroscopy - a rapid, cost-effective technique with minimal sample volume requirements and potential to aid biomedical diagnostics. FTIR has the capacity to probe health phenotypes via the investigation of the full repertoire of molecular species within a sample at once, within a single measurement in a high-throughput manner. In this study, we take advantage of cross-molecular fingerprinting to probe for breast cancer detection. METHODS: We compare two groups: 26 patients diagnosed with breast cancer to a same-sized group of age-matched healthy, asymptomatic female participants. Training with support-vector machines (SVM), we derive classification models that we test in a repeated 10-fold cross-validation over 10 times. In addition, we investigate spectral information responsible for BC identification using statistical significance testing. RESULTS: Our models to detect breast cancer achieve an average overall performance of 0.79 in terms of area under the curve (AUC) of the receiver operating characteristic (ROC). In addition, we uncover a relationship between the effect size of the measured infrared fingerprints and the tumor progression. CONCLUSION: This pilot study provides the foundation for further extending and evaluating blood-based infrared probing approach as a possible cross-molecular fingerprinting modality to tackle breast cancer detection and thus possibly contribute to the future of cancer screening.

Infrared molecular fingerprinting of blood-based liquid biopsies for the detection of cancer.

Recent omics analyses of human biofluids provide opportunities to probe selected species of biomolecules for disease diagnostics. Fourier-transform infrared (FTIR) spectroscopy investigates the full repertoire of molecular species within a sample at once. Here, we present a multi-institutional study in which we analysed infrared fingerprints of plasma and serum samples from 1639 individuals with different solid tumours and carefully matched symptomatic and non-symptomatic reference individuals. Focusing on breast, bladder, prostate, and lung cancer, we find that infrared molecular fingerprinting is capable of detecting cancer: training a support vector machine algorithm allowed us to obtain binary classification performance in the range of 0.78-0.89 (area under the receiver operating characteristic curve [AUC]), with a clear correlation between AUC and tumour load. Intriguingly, we find that the spectral signatures differ between different cancer types. This study lays the foundation for high-throughput onco-IR-phenotyping of four common cancers, providing a cost-effective, complementary analytical tool for disease recognition.

Molecular Origin of Blood-Based Infrared Spectroscopic Fingerprints*.

Infrared spectroscopy of liquid biopsies is a time- and cost-effective approach that may advance biomedical diagnostics. However, the molecular nature of disease-related changes of infrared molecular fingerprints (IMFs) remains poorly understood, impeding the method's applicability. Here we probe 148 human blood sera and reveal the origin of the variations in their IMFs. To that end, we supplemented infrared spectroscopy with biochemical fractionation and proteomic profiling, providing molecular information about serum composition. Using lung cancer as an example of a medical condition, we demonstrate that the disease-related differences in IMFs are dominated by contributions from twelve highly abundant proteins-that, if used as a pattern, may be instrumental for detecting malignancy. Tying proteomic to spectral information and machine learning advances our understanding of the infrared spectra of liquid biopsies, a framework that could be applied to probing of any disease.

Stability of person-specific blood-based infrared molecular fingerprints opens up prospects for health monitoring.

Health state transitions are reflected in characteristic changes in the molecular composition of biofluids. Detecting these changes in parallel, across a broad spectrum of molecular species, could contribute to the detection of abnormal physiologies. Fingerprinting of biofluids by infrared vibrational spectroscopy offers that capacity. Whether its potential for health monitoring can indeed be exploited critically depends on how stable infrared molecular fingerprints (IMFs) of individuals prove to be over time. Here we report a proof-of-concept study that addresses this question. Using Fourier-transform infrared spectroscopy, we have fingerprinted blood serum and plasma samples from 31 healthy, non-symptomatic individuals, who were sampled up to 13 times over a period of 7 weeks and again after 6 months. The measurements were performed directly on liquid serum and plasma samples, yielding a time- and cost-effective workflow and a high degree of reproducibility. The resulting IMFs were found to be highly stable over clinically relevant time scales. Single measurements yielded a multiplicity of person-specific spectral markers, allowing individual molecular phenotypes to be detected and followed over time. This previously unknown temporal stability of individual biochemical fingerprints forms the basis for future applications of blood-based infrared spectral fingerprinting as a multiomics-based mode of health monitoring.

Field-resolved infrared spectroscopy of biological systems.

The proper functioning of living systems and physiological phenotypes depends on molecular composition. Yet simultaneous quantitative detection of a wide variety of molecules remains a challenge1-8. Here we show how broadband optical coherence opens up opportunities for fingerprinting complex molecular ensembles in their natural environment. Vibrationally excited molecules emit a coherent electric field following few-cycle infrared laser excitation9-12, and this field is specific to the sample's molecular composition. Employing electro-optic sampling10,12-15, we directly measure this global molecular fingerprint down to field strengths 107 times weaker than that of the excitation. This enables transillumination of intact living systems with thicknesses of the order of 0.1 millimetres, permitting broadband infrared spectroscopic probing of human cells and plant leaves. In a proof-of-concept analysis of human blood serum, temporal isolation of the infrared electric-field fingerprint from its excitation along with its sampling with attosecond timing precision results in detection sensitivity of submicrograms per millilitre of blood serum and a detectable dynamic range of molecular concentration exceeding 105. This technique promises improved molecular sensitivity and molecular coverage for probing complex, real-world biological and medical settings.

Tcf7l2 is required for left-right asymmetric differentiation of habenular neurons.

BACKGROUND: Although left-right asymmetries are common features of nervous systems, their developmental bases are largely unknown. In the zebrafish epithalamus, dorsal habenular neurons adopt medial (dHbm) and lateral (dHbl) subnuclear character at very different frequencies on the left and right sides. The left-sided parapineal promotes the elaboration of dHbl character in the left habenula, albeit by an unknown mechanism. Likewise, the genetic pathways acting within habenular neurons to control their asymmetric differentiated character are unknown. RESULTS: In a forward genetic screen for mutations that result in loss of habenular asymmetry, we identified two mutant alleles of tcf7l2, a gene that encodes a transcriptional regulator of Wnt signaling. In tcf7l2 mutants, most neurons on both sides differentiate with dHbl identity. Consequently, the habenulae develop symmetrically, with both sides adopting a pronounced leftward character. Tcf7l2 acts cell automously in nascent equipotential neurons, and on the right side, it promotes dHbm and suppresses dHbl differentiation. On the left, the parapineal prevents this Tcf7l2-dependent process, thereby promoting dHbl differentiation. CONCLUSIONS: Tcf7l2 is essential for lateralized fate selection by habenular neurons that can differentiate along two alternative pathways, thereby leading to major neural circuit asymmetries.

Molecular dissection of Wnt3a-Frizzled8 interaction reveals essential and modulatory determinants of Wnt signaling activity.

BACKGROUND: Wnt proteins are a family of secreted signaling molecules that regulate key developmental processes in metazoans. The molecular basis of Wnt binding to Frizzled and LRP5/6 co-receptors has long been unknown due to the lack of structural data on Wnt ligands. Only recently, the crystal structure of the Wnt8-Frizzled8-cysteine-rich-domain (CRD) complex was solved, but the significance of interaction sites that influence Wnt signaling has not been assessed. RESULTS: Here, we present an extensive structure-function analysis of mouse Wnt3a in vitro and in vivo. We provide evidence for the essential role of serine 209, glycine 210 (site 1) and tryptophan 333 (site 2) in Fz binding. Importantly, we discovered that valine 337 in the site 2 binding loop is critical for signaling without contributing to binding. Mutations in the presumptive second CRD binding site (site 3) partly abolished Wnt binding. Intriguingly, most site 3 mutations increased Wnt signaling, probably by inhibiting Wnt-CRD oligomerization. In accordance, increasing amounts of soluble Frizzled8-CRD protein modulated Wnt3a signaling in a biphasic manner. CONCLUSIONS: We propose a concentration-dependent switch in Wnt-CRD complex formation from an inactive aggregation state to an activated high mobility state as a possible modulatory mechanism in Wnt signaling gradients.

Hoxb1b controls oriented cell division, cell shape and microtubule dynamics in neural tube morphogenesis.

Hox genes are classically ascribed to function in patterning the anterior-posterior axis of bilaterian animals; however, their role in directing molecular mechanisms underlying morphogenesis at the cellular level remains largely unstudied. We unveil a non-classical role for the zebrafish hoxb1b gene, which shares ancestral functions with mammalian Hoxa1, in controlling progenitor cell shape and oriented cell division during zebrafish anterior hindbrain neural tube morphogenesis. This is likely distinct from its role in cell fate acquisition and segment boundary formation. We show that, without affecting major components of apico-basal or planar cell polarity, Hoxb1b regulates mitotic spindle rotation during the oriented neural keel symmetric mitoses that are required for normal neural tube lumen formation in the zebrafish. This function correlates with a non-cell-autonomous requirement for Hoxb1b in regulating microtubule plus-end dynamics in progenitor cells in interphase. We propose that Hox genes can influence global tissue morphogenesis by control of microtubule dynamics in individual cells in vivo.

Brain on the stage - spotlight on nervous system development in zebrafish: EMBO practical course, KIT, Sept. 2013.

During the EMBO course 'Imaging of Neural Development in Zebrafish', held on September 9-15th 2013, researchers from different backgrounds shared their latest results, ideas and practical expertise on zebrafish as a model to address open questions regarding nervous system development.

Zebrafish neural tube morphogenesis requires Scribble-dependent oriented cell divisions.

How control of subcellular events in single cells determines morphogenesis on the scale of the tissue is largely unresolved. The stereotyped cross-midline mitoses of progenitors in the zebrafish neural keel provide a unique experimental paradigm for defining the role and control of single-cell orientation for tissue-level morphogenesis in vivo. We show here that the coordinated orientation of individual progenitor cell division in the neural keel is the cellular determinant required for morphogenesis into a neural tube epithelium with a single straight lumen. We find that Scribble is required for oriented cell division and that its function in this process is independent of canonical apicobasal and planar polarity pathways. We identify a role for Scribble in controlling clustering of α-catenin foci in dividing progenitors. Loss of either Scrib or N-cadherin results in abnormally oriented mitoses, reduced cross-midline cell divisions, and similar neural tube defects. We propose that Scribble-dependent nascent cell-cell adhesion clusters between neuroepithelial progenitors contribute to define orientation of their cell division. Finally, our data demonstrate that while oriented mitoses of individual cells determine neural tube architecture, the tissue can in turn feed back on its constituent cells to define their polarization and cell division orientation to ensure robust tissue morphogenesis.

Mitotic activation of the kinase Aurora-A requires its binding partner Bora.

The protein kinase Aurora-A is required for centrosome maturation, spindle assembly, and asymmetric protein localization during mitosis. Here, we describe the identification of Bora, a conserved protein that is required for the activation of Aurora-A at the onset of mitosis. In the Drosophila peripheral nervous system, bora mutants have defects during asymmetric cell division identical to those observed in aurora-A. Furthermore, overexpression of bora can rescue defects caused by mutations in aurora-A. Bora is conserved in vertebrates, and both Drosophila and human Bora can bind to Aurora-A and activate the kinase in vitro. In interphase cells, Bora is a nuclear protein, but upon entry into mitosis, Bora is excluded from the nucleus and translocates into the cytoplasm in a Cdc2-dependent manner. We propose a model in which activation of Cdc2 initiates the release of Bora into the cytoplasm where it can bind and activate Aurora-A.

Mammalian inscuteable regulates spindle orientation and cell fate in the developing retina.

During mammalian neurogenesis, progenitor cells can divide with the mitotic spindle oriented parallel or perpendicular to the surface of the neuroepithelium. Perpendicular divisions are more likely to be asymmetric and generate one progenitor and one neuronal precursor. Whether the orientation of the mitotic spindle actually determines their asymmetric outcome is unclear. Here, we characterize a mammalian homolog of Inscuteable (mInsc), a key regulator of spindle orientation in Drosophila. mInsc is expressed temporally and spatially in a manner that suggests a role in orienting the mitotic spindle in the developing nervous system. Using retroviral RNAi in rat retinal explants, we show that downregulation of mInsc inhibits vertical divisions. This results in enhanced proliferation, consistent with a higher frequency of symmetric divisions generating two proliferating cells. Our results suggest that the orientation of neural progenitor divisions is important for cell fate specification in the retina and determines their symmetric or asymmetric outcome.

X-linked adrenoleukodystrophy mice demonstrate abnormalities in cholesterol metabolism.

The neurodegenerative disorder X-linked adrenoleukodystrophy (X-ALD) is caused by ABCD1 mutations and characterized by very long-chain fatty acid (VLCFA) accumulation. Cholesterol-lowering normalized VLCFA in fibroblasts and plasma of X-ALD patients. We show that in cultured cells, cholesterol-loading induces ABCD1. In X-ALD mice, plasma cholesterol is elevated and not further increasable by cholesterol-feeding, whereas hepatic HMG-CoA reductase and Abcd2 are downregulated. Upon cholesterol modulation, brain VLCFA increased in X-ALD mice, but decreased in controls. In murine X-ALD fibroblasts, cholesterol-lowering did not normalize VLCFA. Thus, ALDP-deficiency and VLCFA are linked to cholesterol but species differences complicate evaluating cholesterol-lowering drugs in X-ALD mice.

Murine bubblegum orthologue is a microsomal very long-chain acyl-CoA synthetase.

It has been suggested that a gene termed bubblegum (Bgm), encoding an acyl-CoA synthetase, could be involved in the pathogenesis of the inherited neurodegenerative disorder X-ALD (X-linked adrenoleukodystrophy). The precise function of the ALDP (ALD protein) still remains unclear. Aldp deficiency in mammals and Bgm deficiency in Drosophila led to accumulation of VLCFAs (very long-chain fatty acids). As a first step towards studying this interaction in wild-type versus Aldp-deficient mice, we analysed the expression pattern of the murine orthologue of the Bgm gene. In contrast with the ubiquitously expressed Ald gene, Bgm expression is restricted to the tissues that are affected by X-ALD such as brain, testis and adrenals. During mouse brain development, Bgm mRNA was first detected by Northern-blot analysis on embryonic day 18 and increased steadily towards adulthood, whereas the highest level of Ald mRNA was found on embryonic day 12 and decreased gradually during differentiation. Protein fractionation and confocal laser imaging of Bgm-green fluorescent protein fusion proteins revealed a microsomal localization that was different from peroxisomes (where Aldp is detected), endoplasmic reticulum and Golgi. Mouse Bgm showed acyl-CoA synthetase activity towards a VLCFA substrate in addition to LCFAs, and this activity was enriched in the microsomal compartment. Speculating that Bgm expression could be regulated by Ald deficiency, we compared the abundance of Bgm mRNA in wild-type and Ald knockout mice but observed no difference. Although mouse Bgm is capable of activating VLCFA, we conclude that a direct interaction between the mouse Bgm and the Aldp seems unlikely.

Cholesterol regulates ABCD2 expression: implications for the therapy of X-linked adrenoleukodystrophy.

X-linked adrenoleukodystrophy (X-ALD) is a severe neurodegenerative disorder with impaired very long-chain fatty acid (VLCFA) metabolism. The disease-associated ABCD1 (ALD) gene encodes a peroxisomal membrane protein, which belongs to the superfamily of ATP-binding cassette transporters. Several treatment regimes have been tried without satisfactory clinical benefit. Recently, the cholesterol-lowering drug lovastatin was reported to normalize VLCFA levels in two out of three clinical studies. This investigation aimed to disclose the molecular mechanism of successful reduction of VLCFA accumulation in order to fill in the gap in the understanding how dietary cholesterol lowering affects the levels of VLCFA in patients with X-ALD and to allow more efficacious treatment. Overexpression of ABCD2 (ALDR), the closest relative of ABCD1, restores VLCFA accumulation in cultured ABCD1-deficient cells. Here we show by real-time PCR that the ABCD2 gene is induced in cultured human fibroblasts and monocytes upon sterol depletion via a mechanism requiring the activation of sterol regulatory element-binding proteins (SREBPs), a family of transcription factors that control the metabolism of cholesterol and fatty acids. This is unexpected and the first report that extends the mechanism of transcriptional regulation by SREBPs to a peroxisomal protein, thus providing a closer link between peroxisomes, cholesterol and fatty acid biosynthesis. Using reporter gene studies, site-directed mutagenesis and gel shift assays, we identified a functional sterol regulatory element in the proximal promoter region of ABCD2. Finally, we demonstrated that ABCD2 induction by sterol depletion significantly reduced the accumulation of VLCFA in X-ALD fibroblasts. Thus, lowering cholesterol leads to SREBP maturation, increased ABCD2 expression and reduced VLCFA accumulation.