Non-targeted transcriptomic effects upon thyroid irradiation: similarity between in-field and out-of-field responses varies with tissue type

Development of a machine learning framework for radiation biomarker discovery and absorbed dose prediction

Background

Molecular radiation biomarkers are an emerging tool in radiation research with applications for cancer radiotherapy, radiation risk assessment, and even human space travel. However, biomarker screening in genome-wide expression datasets using conventional tools is time-consuming and underlies analyst (human) bias. Machine Learning (ML) methods can improve the sensitivity and specificity of biomarker identification, increase analytical speed, and avoid multicollinearity and human bias.

Aim

To develop a resource-efficient ML framework for radiation biomarker discovery using gene expression data from irradiated normal tissues. Further, to identify biomarker panels predicting radiation dose with tissue specificity.

Methods

A strategic search in the Gene Expression Omnibus database identified a transcriptomic dataset (GSE44762) for normal tissues radiation responses (murine kidney cortex and medulla) suited for biomarker discovery using an ML approach. The dataset was pre-processed in R and separated into train and test data subsets. High computational cost of Genetic Algorithm/k-Nearest Neighbor (GA/KNN) mandated optimization and 13 ML models were tested using the caret package in R. Biomarker performance was evaluated and visualized via Principal Component Analysis (PCA) and dose regression. The novelty of ML-identified biomarker panels was evaluated by literature search.

Results

Caret-based feature selection and ML methods vastly improved processing time over the GA approach. The KNN method yielded overall best performance values on train and test data and was implemented into the framework. The top-ranking genes were Cdkn1a, Gria3, Mdm2 and Plk2 in cortex, and Brf2, Ccng1, Cdkn1a, Ddit4l, and Gria3 in medulla. These candidates successfully categorized dose groups and tissues in PCA. Regression analysis showed that correlation between predicted and true dose was high with R² of 0.97 and 0.99 for cortex and medulla, respectively.

Conclusion

The caret framework is a powerful tool for radiation biomarker discovery optimizing performance with resource-efficiency for broad implementation in the field. The KNN-based approach identified Brf2, Ddit4l, and Gria3 mRNA as novel candidates that have been uncharacterized as radiation biomarkers to date. The biomarker panel showed good performance in dose and tissue separation and dose regression. Further training with larger cohorts is warranted to improve accuracy, especially for lower doses.

Clinical Efficacy of Levothyroxine Sodium plus I 131 in the Treatment of Patients with Thyroidectomy and Its Effect on the Levels of Thyroglobulin and Thyrotropin

Objective

This study was to evaluate the clinical efficacy of levothyroxine sodium tablets combined with I 131 in the treatment of patients after thyroidectomy and the effect on thyroglobulin (Tg) and thyroid stimulating hormone (TSH) levels.

Methods

80 patients with differentiated thyroid cancer who required thyroidectomy after surgery between July 2019 and January 2021 were recruited for prospective study, 40 patients in the control group received levothyroxine sodium tablets, and 40 patients in the experimental group received levothyroxine sodium treatment plus I 131 treatment. Treatment effect, serum Tg and TSH levels, and relapse were measured.

Results

The removal rate of residual thyroid tissue in the experimental group (87.50%) was significantly higher than that in the control group (57.50%) (P < 0.05). Levothyroxine sodium tablets plus I 131 was associated with a significantly higher efficacy versus levothyroxine sodium tablets (P < 0.05). There were no significant differences in the serum Tg levels between the two groups before treatment (P > 0.05). After treatment, the serum Tg levels in both groups were significantly decreased, and levothyroxine sodium tablets plus I 131 resulted in a significantly lower Tg level versus levothyroxine sodium tablets (P < 0.05). Before treatment, the two groups showed similar TSH levels (P > 0.05). After treatment, patients receiving levothyroxine sodium tablets plus I 131 had a significantly greater increase in the TSH levels versus levothyroxine sodium tablets (P < 0.05). The recurrence rate of the experimental group was lower than that of the control group (P < 0.05).

Conclusion

Levothyroxine sodium tablets plus I 131 for post-operative patients with differentiated thyroid cancer enhance the removal rate of residual thyroid tissue, effectively reduce serum Tg level, and increase TSH level, with significant therapeutic effects, low recurrence rates, and a high safety profile.

Age and sex effects across the blood proteome after ionizing radiation exposure can bias biomarker screening and risk assessment

Molecular biomarkers of ionizing radiation (IR) exposure are a promising new tool in various disciplines: they can give necessary information for adaptive treatment planning in cancer radiotherapy, enable risk projection for radiation-induced survivorship diseases, or facilitate triage and intervention in radiation hazard events. However, radiation biomarker discovery has not yet resolved the most basic features of personalized medicine: age and sex. To overcome this critical bias in biomarker identification, we quantitated age and sex effects and assessed their relevance in the radiation response across the blood proteome. We used high-throughput mass spectrometry on blood plasma collected 24 h after 0.5 Gy total body irradiation (15 MV nominal photon energy) from male and female C57BL/6 N mice at juvenile (7-weeks-old) or adult (18-weeks-old) age. We also assessed sex and strain effects using juvenile male and female BALB/c nude mice. We showed that age and sex created significant effects in the proteomic response regarding both extent and functional quality of IR-induced responses. Furthermore, we found that age and sex effects appeared non-linear and were often end-point specific. Overall, age contributed more to differences in the proteomic response than sex, most notably in immune responses, oxidative stress, and apoptotic cell death. Interestingly, sex effects were pronounced for DNA damage and repair pathways and associated cellular outcome (pro-survival vs. pro-apoptotic). Only one protein (AHSP) was identified as a potential general biomarker candidate across age and sex, while GMNN, REG3B, and SNCA indicated some response similarity across age. This low yield advocated that unisex or uniage biomarker screening approaches are not feasible. In conclusion, age- and sex-specific screening approaches should be implemented as standard protocol to ensure robustness and diagnostic power of biomarker candidates. Bias-free molecular biomarkers are a necessary progression towards personalized medicine and integral for advanced adaptive cancer radiotherapy and risk assessment.

Systems Biology Approaches and Precision Oral Health: A Circadian Clock Perspective

A vast majority of the pathophysiological and metabolic processes in humans are temporally controlled by a master circadian clock located centrally in the hypothalamic suprachiasmatic nucleus of the brain, as well as by specialized peripheral oscillators located in other body tissues. This circadian clock system generates a rhythmical diurnal transcriptional-translational cycle in clock genes and protein expression and activities regulating numerous downstream target genes. Clock genes as key regulators of physiological function and dysfunction of the circadian clock have been linked to various diseases and multiple morbidities. Emerging omics technologies permits largescale multi-dimensional investigations of the molecular landscape of a given disease and the comprehensive characterization of its underlying cellular components (e.g., proteins, genes, lipids, metabolites), their mechanism of actions, functional networks and regulatory systems. Ultimately, they can be used to better understand disease and interpatient heterogeneity, individual profile, identify personalized targetable key molecules and pathways, discover novel biomarkers and genetic alterations, which collectively can allow for a better patient stratification into clinically relevant subgroups to improve disease prediction and prevention, early diagnostic, clinical outcomes, therapeutic benefits, patient's quality of life and survival. The use of “omics” technologies has allowed for recent breakthroughs in several scientific domains, including in the field of circadian clock biology. Although studies have explored the role of clock genes using circadiOmics (which integrates circadian omics, such as genomics, transcriptomics, proteomics and metabolomics) in human disease, no such studies have investigated the implications of circadian disruption in oral, head and neck pathologies using multi-omics approaches and linking the omics data to patient-specific circadian profiles. There is a burgeoning body of evidence that circadian clock controls the development and homeostasis of oral and maxillofacial structures, such as salivary glands, teeth and oral epithelium. Hence, in the current era of precision medicine and dentistry and patient-centered health care, it is becoming evident that a multi-omics approach is needed to improve our understanding of the role of circadian clock-controlled key players in the regulation of head and neck pathologies. This review discusses current knowledge on the role of the circadian clock and the contribution of omics-based approaches toward a novel precision health era for diagnosing and treating head and neck pathologies, with an emphasis on oral, head and neck cancer and Sjögren's syndrome.

Deconvolution of expression microarray data reveals 131I-induced responses otherwise undetected in thyroid tissue

High-throughput gene expression analysis is increasingly used in radiation research for discovery of damage-related or absorbed dose-dependent biomarkers. In tissue samples, cell type-specific responses can be masked in expression data due to mixed cell populations which can preclude biomarker discovery. In this study, we deconvolved microarray data from thyroid tissue in order to assess possible bias from mixed cell type data. Transcript expression data [GSE66303] from mouse thyroid that received 5.9 Gy from ¹³¹I over 24 h (or 0 Gy from mock treatment) were deconvolved by cell frequency of follicular cells and C-cells using csSAM and R and processed with Nexus Expression. Literature-based signature genes were used to assess the relative impact from ionizing radiation (IR) or thyroid hormones (TH). Regulation of cellular functions was inferred by enriched biological processes according to Gene Ontology terms. We found that deconvolution increased the detection rate of significantly regulated transcripts including the biomarker candidate family of kallikrein transcripts. Detection of IR-associated and TH-responding signature genes was also increased in deconvolved data, while the dominating trend of TH-responding genes was reproduced. Importantly, responses in biological processes for DNA integrity, gene expression integrity, and cellular stress were not detected in convoluted data–which was in disagreement with expected dose-response relationships–but upon deconvolution in follicular cells and C-cells. In conclusion, previously reported trends of ¹³¹I-induced transcriptional responses in thyroid were reproduced with deconvolved data and usually with a higher detection rate. Deconvolution also resolved an issue with detecting damage and stress responses in enriched data, and may reduce false negatives in other contexts as well. These findings indicate that deconvolution can optimize microarray data analysis of heterogeneous sample material for biomarker screening or other clinical applications.

Predicting ionizing radiation exposure using biochemically-inspired genomic machine learning

Background: Gene signatures derived from transcriptomic data using machine learning methods have shown promise for biodosimetry testing. These signatures may not be sufficiently robust for large scale testing, as their performance has not been adequately validated on external, independent datasets. The present study develops human and murine signatures with biochemically-inspired machine learning that are strictly validated using k-fold and traditional approaches. Methods: Gene Expression Omnibus (GEO) datasets of exposed human and murine lymphocytes were preprocessed via nearest neighbor imputation and expression of genes implicated in the literature to be responsive to radiation exposure (n=998) were then ranked by Minimum Redundancy Maximum Relevance (mRMR). Optimal signatures were derived by backward, complete, and forward sequential feature selection using Support Vector Machines (SVM), and validated using k-fold or traditional validation on independent datasets. Results: The best human signatures we derived exhibit k-fold validation accuracies of up to 98% ( DDB2,  PRKDC, TPP2, PTPRE, and GADD45A) when validated over 209 samples and traditional validation accuracies of up to 92% ( DDB2,  CD8A,  TALDO1,  PCNA,  EIF4G2,  LCN2,  CDKN1A,  PRKCH,  ENO1,  and PPM1D) when validated over 85 samples. Some human signatures are specific enough to differentiate between chemotherapy and radiotherapy. Certain multi-class murine signatures have sufficient granularity in dose estimation to inform eligibility for cytokine therapy (assuming these signatures could be translated to humans). We compiled a list of the most frequently appearing genes in the top 20 human and mouse signatures. More frequently appearing genes among an ensemble of signatures may indicate greater impact of these genes on the performance of individual signatures. Several genes in the signatures we derived are present in previously proposed signatures. Conclusions: Gene signatures for ionizing radiation exposure derived by machine learning have low error rates in externally validated, independent datasets, and exhibit high specificity and granularity for dose estimation.