High throughput screening of mesenchymal stem cell lines using deep learning

Abstract

Mesenchymal stem cells (MSCs) are increasingly used as regenerative therapies for patients in the preclinical and clinical phases of various diseases. However, the main limitations of such therapies include functional heterogeneity and the lack of appropriate quality control (QC) methods for functional screening of MSC lines; thus, clinical outcomes are inconsistent. Recently, machine learning (ML)-based methods, in conjunction with single-cell morphological profiling, have been proposed as alternatives to conventional in vitro/vivo assays that evaluate MSC functions. Such methods perform in silico analyses of MSC functions by training ML algorithms to find highly nonlinear connections between MSC functions and morphology. Although such approaches are promising, they are limited in that extensive, high-content single-cell imaging is required; moreover, manually identified morphological features cannot be generalized to other experimental settings. To address these limitations, we propose an end-to-end deep learning (DL) framework for functional screening of MSC lines using live-cell microscopic images of MSC populations. We quantitatively evaluate various convolutional neural network (CNN) models and demonstrate that our method accurately classifies in vitro MSC lines to high/low multilineage differentiating stress-enduring (MUSE) cells markers from multiple donors. A total of 6,120 cell images were obtained from 8 MSC lines, and they were classified into two groups according to MUSE cell markers analyzed by immunofluorescence staining and FACS. The optimized DenseNet121 model showed area under the curve (AUC) 0.975, accuracy 0.922, F1 0.922, sensitivity 0.905, specificity 0.942, positive predictive value 0.940, and negative predictive value 0.908. Therefore, our DL-based framework is a convenient high-throughput method that could serve as an effective QC strategy in future clinical biomanufacturing processes.

Advancing in vitro virtual nuclei staining of stem cells through a cross-structure, artifact-free U-Net approach

“…In the context of non-invasive imaging methods for MSCs, deep learning exploration was conducted by Zhang et al [11], which centered around label-free nuclei detection from implicit phase information of MSCs. High throughput screening of MSCs has been addressed by Kim et al (2022) [12] and Imboden et al (2021) [13] through leveraging deep learning algorithms.…”

“…Despite the strides made in high-throughput microscopy screening contributing to the development of promising automated image analysis tools, background noise in biological samples presents a significant barrier to extract in-depth and accurate details from raw data that are close to intrinsic noise. Current techniques typically involve either label-based methods which can perturb cellular properties or complex label-free methods, which require extensive post-processing and manual intervention of cellular analytics [11][12][13]. Effective analysis of the heterogeneous characteristics of MSCs from bright-field (BF) images still has limited practical applications [14][15][16][17][18].…”

Unleashing the power of high-throughput bright-field imaging for enhanced mesenchymal cell separation: a novel supervised clustering approach in vitro augmentation of healthy and stressful conditions

“…Yet, manufacturing these therapeutic agents introduces unique challenges, including issues related to donor variability, tissue source, and differences in the media environment [5,6]. To address these constraints, the use of high-throughput imaging and artificial intelligence (AI) technologies has been recently advanced, offering speedy and in-depth insights to enhance bioprocess analytics in CT manufacturing [7][8][9].…”

“…However, this approach is not only expensive and time-consuming but can also be potentially damaging to viable cells. Despite recent progress in understanding cellular features in CT bioprocesses, systematic noise in high-throughput microscopy screening still leads to false-negative outcomes in AI models employed for virtual cell staining [8,10]. AI and imaging technologies have been increasingly adopted to enhance CT manufacturing, with several researchers exploring diverse strategies to address CT manufacturing challenges.…”

“…AI and imaging technologies have been increasingly adopted to enhance CT manufacturing, with several researchers exploring diverse strategies to address CT manufacturing challenges. Examples include the work of Zhang et al, who employed deep learning for label-free nuclei detection from MSC implicit phase information [7], and Kim et al conducted a high-throughput screening of MSC lines using deep learning techniques [8]. Further, Imboden et al, who used AI-driven label-free imaging to examine MSC heterogeneities [9].…”