Recognition of breast cancer subtypes using FTIR hyperspectral data

Fourier-transform infrared spectroscopy (FTIR) is a powerful, non-destructive, highly sensitive and a promising analytical technique to provide spectrochemical signatures of biological samples, where markers like carbohydrates, proteins, and phosphate groups of DNA can be recognized in biological micro-environment. However, method of measurements of large cells need an excessive time to achieve high quality images, making its clinical use difficult due to speed of data-acquisition and lack of optimized computational procedures. To address such challenges, Machine Learning (ML) based technologies can assist to assess an accurate prognostication of breast cancer (BC) subtypes with high performance. Here, we applied FTIR spectroscopy to identify breast cancer subtypes in order to differentiate between luminal (BT474) and non-luminal (SKBR3) molecular subtypes. For this reason, we tested multivariate classification technique to extract feature information employing three-dimension (3D)-discriminant analysis approach based on 3D-principle component analysis-linear discriminant analysis (3D-PCA-LDA) and 3D-principal component analysis-quadratic discriminant analysis (3D-PCA-QDA), showing an improvement in sensitivity (98%), specificity (94%) and accuracy (98%) parameters compared to conventional unfolded methods. Our results evidence that 3D-PCA-LDA and 3D-PCA-QDA are potential tools for discriminant analysis of hyperspectral dataset to obtain superior classification assessment.

Stratification of tumour cell radiation response and metabolic signatures visualization with Raman spectroscopy and explainable convolutional neural network

Reprogramming of cellular metabolism is a driving factor of tumour progression and radiation therapy resistance. Identifying biochemical signatures associated with tumour radioresistance may assist with the development of targeted treatment strategies to improve clinical outcomes. Raman spectroscopy (RS) can monitor post-irradiation biomolecular changes and signatures of radiation response in tumour cells in a label-free manner. Convolutional Neural Networks (CNN) perform feature extraction directly from data in an end-to-end learning manner, with high classification performance. Furthermore, recently developed CNN explainability techniques help visualize the critical discriminative features captured by the model. In this work, a CNN is developed to characterize tumour response to radiotherapy based on its degree of radioresistance. The model was trained to classify Raman spectra of three human tumour cell lines as radiosensitive (LNCaP) or radioresistant (MCF7, H460) over a range of treatment doses and data collection time points. Additionally, a method based on Gradient-Weighted Class Activation Mapping (Grad-CAM) was used to determine response-specific salient Raman peaks influencing the CNN predictions. The CNN effectively classified the cell spectra, with accuracy, sensitivity, specificity, and F1 score exceeding 99.8%. Grad-CAM heatmaps of H460 and MCF7 cell spectra (radioresistant) exhibited high contributions from Raman bands tentatively assigned to glycogen, amino acids, and nucleic acids. Conversely, heatmaps of LNCaP cells (radiosensitive) revealed activations at lipid and phospholipid bands. Finally, Grad-CAM variable importance scores were derived for glycogen, asparagine, and phosphatidylcholine, and we show that their trends over cell line, dose, and acquisition time agreed with previously established models. Thus, the CNN can accurately detect biomolecular differences in the Raman spectra of tumour cells of varying radiosensitivity without requiring manual feature extraction. Finally, Grad-CAM may help identify metabolic signatures associated with the observed categories, offering the potential for automated clinical tumour radiation response characterization.

Vibrational spectroscopies for biochemical investigation of X-ray exposure effects on SH-SY5Y human neuroblastoma cells

Neuroblastoma is the most recurring cancer in childhood and adolescence. The SH-SY5Y neuroblastoma cell line is generally adopted for elaborating new therapeutical approaches and/or elaborating strategies for the prevention of central nervous system disturbances. In fact, it represents a valid model system for investigating in vitro the effects on the brain of X-ray exposure using vibrational spectroscopies that can detect early radiation-induced molecular alterations of potential clinical usefulness. In recent years, we dedicated significant efforts in the use of Fourier-transform and Raman microspectroscopy techniques for characterizing such radiation-induced effects on SH-SY5Y cells by examining the contributions from different cell components (DNA, proteins, lipids, and carbohydrates) to the vibrational spectra. In this review, we aim at revising and comparing the main results of our studies to provide a wide outlook of the latest outcomes and a framework for future radiobiology research using vibrational spectroscopies. A short description of our experimental approaches and data analysis procedures is also reported.

Glycogen synthase 1 targeting reveals a metabolic vulnerability in triple-negative breast cancer

BACKGROUND

Hypoxia-induced glycogen turnover is implicated in cancer proliferation and therapy resistance. Triple-negative breast cancers (TNBCs), characterized by a hypoxic tumor microenvironment, respond poorly to therapy. We studied the expression of glycogen synthase 1 (GYS1), the key regulator of glycogenesis, and other glycogen-related enzymes in primary tumors of patients with breast cancer and evaluated the impact of GYS1 downregulation in preclinical models.

METHODS

mRNA expression of GYS1 and other glycogen-related enzymes in primary breast tumors and the correlation with patient survival were studied in the METABRIC dataset (n = 1904). Immunohistochemical staining of GYS1 and glycogen was performed on a tissue microarray of primary breast cancers (n = 337). In four breast cancer cell lines and a mouse xenograft model of triple-negative breast cancer, GYS1 was downregulated using small-interfering or stably expressed short-hairpin RNAs to study the effect of downregulation on breast cancer cell proliferation, glycogen content and sensitivity to various metabolically targeted drugs.

RESULTS

High GYS1 mRNA expression was associated with poor patient overall survival (HR 1.20, P = 0.009), especially in the TNBC subgroup (HR 1.52, P = 0.014). Immunohistochemical GYS1 expression in primary breast tumors was highest in TNBCs (median H-score 80, IQR 53-121) and other Ki67-high tumors (median H-score 85, IQR 57-124) (P < 0.0001). Knockdown of GYS1 impaired proliferation of breast cancer cells, depleted glycogen stores and delayed growth of MDA-MB-231 xenografts. Knockdown of GYS1 made breast cancer cells more vulnerable to inhibition of mitochondrial proteostasis.

CONCLUSIONS

Our findings highlight GYS1 as potential therapeutic target in breast cancer, especially in TNBC and other highly proliferative subsets.

Enhanced characterization of breast cancer phenotypes using Raman micro-spectroscopy on stainless steel substrate

Biochemical insights into varying breast cancer (BC) phenotypes can provide a fundamental understanding of BC pathogenesis, while identifying novel therapeutic targets. Raman spectroscopy (RS) can gauge these biochemical differences with high specificity. For routine RS, cells are traditionally seeded onto calcium fluoride (CaF₂) substrates that are costly and fragile, limiting its widespread adoption. Stainless steel has been interrogated previously as a less expensive alternative to CaF₂ substrates, while reporting increased Raman signal intensity than the latter. We sought to further investigate and compare the Raman signal quality measured from stainless steel versus CaF₂ substrates by characterizing different BC phenotypes with altered human epidermal growth factor receptor 2 (HER2) expression. Raman spectra were obtained on stainless steel and CaF₂ substrates for HER2 negative cells - MDA-MB-231, MDA-MB-468 and HER2 overexpressing cells - AU565, SKBr3. Upon analyzing signal-to-noise ratios (SNR), stainless steel provided a stronger Raman signal, improving SNR by 119% at 1450 cm^-1 and 122% at 2925 cm^-1 on average compared to the CaF₂ substrate. Utilizing only 22% of laser power on sample relative to the CaF₂ substrate, stainless steel still yielded improved spectral characterization over CaF₂, achieving 96.0% versus 89.8% accuracy in BC phenotype discrimination and equivalent 100.0% accuracy in HER2 status classification. Spectral analysis further highlighted increased lipogenesis and altered metabolism in HER2 overexpressing cells, which was subsequently visualized with coherent anti-Stokes Raman scattering microscopy. Our findings demonstrate that stainless steel substrates deliver improved Raman signal and enhanced spectral characterization, underscoring its potential as a cost-effective alternative to CaF₂ for non-invasively monitoring cellular biochemical dynamics in translational cancer research.

Raman microspectroscopy and machine learning for use in identifying radiation-induced lung toxicity

OBJECTIVE

In this work, we explore and develop a method that uses Raman spectroscopy to measure and differentiate radiation induced toxicity in murine lungs with the goal of setting the foundation for a predictive disease model.

METHODS

Analysis of Raman tissue data is achieved through a combination of techniques. We first distinguish between tissue measurements and air pockets in the lung by using group and basis restricted non-negative matrix factorization. We then analyze the tissue spectra using sparse multinomial logistic regression to discriminate between fibrotic gradings. Model validation is achieved by splitting the data into a training set containing 70% of the data and a test set with the remaining 30%; classification accuracy is used as the performance metric. We also explore several other potential classification tasks wherein the response considered is the grade of pneumonitis and fibrosis sickness.

RESULTS

A classification accuracy of 91.6% is achieved on the test set of fibrotic gradings, illustrating the ability of Raman measurements to detect differing levels of fibrotic disease among the murine lungs. It is also shown via further modeling that coarser consideration of fibrotic grading via binning (ie. 'Low', 'Medium', 'High') does not degrade performance. Finally, we consider preliminary models for pneumonitis discrimination using the same methodologies.

Group and Basis Restricted Non-Negative Matrix Factorization and Random Forest for Molecular Histotype Classification and Raman Biomarker Monitoring in Breast Cancer

Raman spectroscopy is a non-invasive optical technique that can be used to investigate biochemical information embedded in cells and tissues exposed to ionizing radiation used in cancer therapy. Raman spectroscopy could potentially be incorporated in personalized radiation treatment design as a tool to monitor radiation response in at the metabolic level. However, tracking biochemical dynamics remains challenging for Raman spectroscopy. Here we developed a novel analytical framework by combining group and basis restricted non-negative matrix factorization and random forest (GBR-NMF-RF). This framework can monitor radiation response profiles in different molecular histotypes and biochemical dynamics in irradiated breast cancer cells. Five subtypes of; human breast cancer (MCF-7, BT-474, MDA-MB-230, and SK-BR-3) and normal cells derived from human breast tissue (MCF10A) which had been exposed to ionizing radiation were tested in this framework. Reference Raman spectra of 20 biochemicals were collected and used as the constrained Raman biomarkers in the GBR-NMF-RF framework. We obtained scores for individual biochemicals corresponding to the contribution of each Raman reference spectrum to each spectrum obtained from the five cell types. A random forest classifier was then fitted to the chemical scores for performing molecular histotype classifications (HER2, PR, ER, Ki67, and cancer versus non-cancer) and assessing the importance of the Raman biochemical basis spectra for each classification test. Overall, the GBR-NMF-RF framework yields classification results with high accuracy (>97%), high sensitivity (>97%), and high specificity (>97%). Variable importance calculated in the random forest model indicated high contributions from glycogen and lipids (cholesterol, phosphatidylserine, and stearic acid) in molecular histotype classifications.

The efficiency of photothermal action of gold shell-isolated nanoparticles against tumor cells depends on membrane interactions

Photoinduced hyperthermia with nanomaterials has been proven effective in photothermal therapy (PTT) of tumor tissues, but a precise control in PTT requires determination of the molecular-level mechanisms. In this paper, we determined the mechanisms responsible for the action of photoexcited gold shell-isolated nanoparticles (AuSHINs) in reducing the viability of MCF7 (glandular breast cancer) and especially A549 (lung adenocarcinoma) cells in vitro experiments, while the photoinduced damage to healthy cells was much smaller. The photoinduced effects were more significant than using other nanomaterials, and could be explained by the different effects from incorporating AuSHINs on Langmuir monolayers from lipid extracts of tumoral (MCF7 and A549) and healthy cells. The incorporation of AuSHINs caused similar expansion of the Langmuir monolayers, but Fourier-transform infrared spectroscopy (FTIR) data of Langmuir-Schaefer films (LS) indicated distinct levels of penetration into the monolayers. AuSHINs penetrated deeper into the A549 extract monolayers, affecting the vibrational modes of polar groups and carbon chains, while in MCF7 monolayers penetration was limited to the surroundings of the polar groups. Even smaller insertion was observed for monolayers of the healthy cell extract. The photochemical reactions were modulated by AuSHINs penetration, since upon irradiation the surface area of A549 monolayer decreased owing to lipid chain cleavage by oxidative reactions. For MCF7 monolayers, hydroperoxidation under illumination led to a ca. 5% increase in surface area. The monolayers of healthy cell lipid extract were barely affected by irradiation, consistent with the lowest degree of AuSHINs insertion. In summary, efficient photothermal therapy may be devised by producing AuSHINs capable of penetrating the chain region of tumor cell membranes.

Single Particle Automated Raman Trapping Analysis of Breast Cancer Cell-Derived Extracellular Vesicles as Cancer Biomarkers

Extracellular vesicles (EVs) secreted by cancer cells provide an important insight into cancer biology and could be leveraged to enhance diagnostics and disease monitoring. This paper details a high-throughput label-free extracellular vesicle analysis approach to study fundamental EV biology, toward diagnosis and monitoring of cancer in a minimally invasive manner and with the elimination of interpreter bias. We present the next generation of our single particle automated Raman trapping analysis─SPARTA─system through the development of a dedicated standalone device optimized for single particle analysis of EVs. Our visualization approach, dubbed dimensional reduction analysis (DRA), presents a convenient and comprehensive method of comparing multiple EV spectra. We demonstrate that the dedicated SPARTA system can differentiate between cancer and noncancer EVs with a high degree of sensitivity and specificity (&gt;95% for both). We further show that the predictive ability of our approach is consistent across multiple EV isolations from the same cell types. Detailed modeling reveals accurate classification between EVs derived from various closely related breast cancer subtypes, further supporting the utility of our SPARTA-based approach for detailed EV profiling.

Detection of acquired radioresistance in breast cancer cell lines using Raman spectroscopy and machine learning

Radioresistance-a living cell's response to, and development of resistance to ionising radiation-can lead to radiotherapy failure and/or tumour recurrence. We used Raman spectroscopy and machine learning to characterise biochemical changes that occur in acquired radioresistance for breast cancer cells. We were able to distinguish between wild-type and acquired radioresistant cells by changes in chemical composition using Raman spectroscopy and machine learning with 100% accuracy. In studying both hormone receptor positive and negative cells, we found similar changes in chemical composition that occur with the development of acquired radioresistance; these radioresistant cells contained less lipids and proteins compared to their parental counterparts. As well as characterising acquired radioresistance in vitro, this approach has the potential to be translated into a clinical setting, to look for Raman signals of radioresistance in tumours or biopsies; that would lead to tailored clinical treatments.

Molecular-level effects on cell membrane models to explain the phototoxicity of gold shell-isolated nanoparticles to cancer cells

Metallic nanoparticles are promising agents for photothermal cancer therapy (PTT) owing to their photostability and efficient light-to-heat conversion, but their possible aggregation remains an issue. In this paper, we report on the photoinduced heating of gold shell-isolated nanoparticles (AuSHINs) in in vitro experiments to kill human oropharyngeal (HEp-2) and breast (BT-474 and MCF-7) carcinoma cells, with cell viability reducing below 50 % with 2.2 × 10¹² AuSHINs/mL and 6 h of incubation. This toxicity to cancer cells is significantly higher than in previous works with gold nanoparticles. Considering the AuSHINs dimensions we hypothesize that cell uptake is not straightforward, and the mechanism of action involves accumulation on phospholipid membranes as the PTT target for photoinduced heating and subsequent generation of reactive oxygen species (ROS). Using Langmuir monolayers as simplified membrane models, we confirmed that AuSHINs have a larger effect on 1,2-dioleoyl-sn-glycero-3-phospho-l-serine (DOPS), believed to represent cancer cell membranes, than on 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) taken as representative of healthy eukaryotic cells. In particular, data from polarization-modulated infrared reflection absorption spectroscopy (PM-IRRAS) revealed an increased conformational order of DOPS tails due to the stronger adsorption of AuSHINs. Furthermore, light irradiation reduced the stability of AuSHINs containing DOPC and DOPS monolayers owing to oxidative reactions triggered by ROS upon photoinduced heating. Compared to DOPC, DOPS lost nearly twice as much material to the subphase, which is consistent with a higher rate of ROS formation in the vicinity of the DOPS monolayer.

Monitor Ionizing Radiation-Induced Cellular Responses with Raman Spectroscopy, Non-Negative Matrix Factorization, and Non-Negative Least Squares

Radiation therapy (RT) is one of the most commonly prescribed cancer treatments. New tools that can accurately monitor and evaluate individual patient responses would be a major advantage and lend to the implementation of personalized treatment plans. In this study, Raman spectroscopy (RS) was applied to examine radiation-induced cellular responses in H460, MCF7, and LNCaP cancer cell lines across different dose levels and times post-irradiation. Previous Raman data analysis was conducted using principal component analysis (PCA), which showed the ability to extract biological information of glycogen. In the current studies, the use of non-negative matrix factorization (NMF) allowed for the discovery of multiplexed biological information, specifically uncovering glycogen-like and lipid-like component bases. The corresponding scores of glycogen and previously unidentified lipids revealed the content variations of these two chemicals in the cellular data. The NMF decomposed glycogen and lipid-like bases were able to separate the cancer cell lines into radiosensitive and radioresistant groups. A further lipid phenotype investigation was also attempted by applying non-negative least squares (NNLS) to the lipid-like bases decomposed individually from three cell lines. Qualitative differences found in lipid weights for each lipid-like basis suggest the lipid phenotype differences in the three tested cancer cell lines. Collectively, this study demonstrates that the application of NMF and NNLS on RS data analysis to monitor ionizing radiation-induced cellular responses can yield multiplexed biological information on bio-response to RT not revealed by conventional chemometric approaches.

Multivariate Analysis of Difference Raman Spectra of the Irradiated Nucleus and Cytoplasm Region of SH-SY5Y Human Neuroblastoma Cells

Previous works showed that spatially resolved Raman spectra of cytoplasm and nucleus region of single cells exposed to X-rays evidence different features. The present work aims to introduce a new approach to profit from these differences to deeper investigate X-ray irradiation effects on single SH-SY5Y human neuroblastoma cells. For this aim, Raman micro-spectroscopy was performed in vitro on single cells after irradiation by graded X-ray doses (2, 4, 6, 8 Gy). Spectra from nucleus and cytoplasm regions were selectively acquired. The examination by interval Principal Component Analysis (i-PCA) of the difference spectra obtained by subtracting each cytoplasm-related spectrum from the corresponding one detected at the nucleus enabled us to reveal the subtle modifications of Raman features specific of different spatial cell regions. They were discussed in terms of effects induced by X-ray irradiation on DNA/RNA, lipids, and proteins. The proposed approach enabled us to evidence some features not outlined in previous investigations.