Organogenetic transcriptomes of the Drosophila embryo at single cell resolution.

To gain insight into the transcription programs activated during the formation of Drosophila larval structures, we carried out single cell RNA sequencing during two periods of Drosophila embryogenesis: stages 10-12, when most organs are first specified and initiate morphological and physiological specialization; and stages 13-16, when organs achieve their final mature architectures and begin to function. Our data confirm previous findings with regards to functional specialization of some organs - the salivary gland and trachea - and clarify the embryonic functions of another - the plasmatocytes. We also identify two early developmental trajectories in germ cells and uncover a potential role for proteolysis during germline stem cell specialization. We identify the likely cell type of origin for key components of the Drosophila matrisome and several commonly used Drosophila embryonic cell culture lines. Finally, we compare our findings with other recent related studies and with other modalities for identifying tissue-specific gene expression patterns. These data provide a useful community resource for identifying many new players in tissue-specific morphogenesis and functional specialization of developing organs.

Analysis of the human kidney transcriptome and plasma proteome identifies markers of proximal tubule maladaptation to injury.

Acute kidney injury (AKI) is a major risk factor for long-term adverse outcomes, including chronic kidney disease. In mouse models of AKI, maladaptive repair of the injured proximal tubule (PT) prevents complete tissue recovery. However, evidence for PT maladaptation and its etiological relationship with complications of AKI is lacking in humans. We performed single-nucleus RNA sequencing of 120,985 nuclei in kidneys from 17 participants with AKI and seven healthy controls from the Kidney Precision Medicine Project. Maladaptive PT cells, which exhibited transcriptomic features of dedifferentiation and enrichment in pro-inflammatory and profibrotic pathways, were present in participants with AKI of diverse etiologies. To develop plasma markers of PT maladaptation, we analyzed the plasma proteome in two independent cohorts of patients undergoing cardiac surgery and a cohort of marathon runners, linked it to the transcriptomic signatures associated with maladaptive PT, and identified nine proteins whose genes were specifically up- or down-regulated by maladaptive PT. After cardiac surgery, both cohorts of patients had increased transforming growth factor-ß2 (TGFB2), collagen type XXIII-α1 (COL23A1), and X-linked neuroligin 4 (NLGN4X) and had decreased plasminogen (PLG), ectonucleotide pyrophosphatase/phosphodiesterase 6 (ENPP6), and protein C (PROC). Similar changes were observed in marathon runners with exercise-associated kidney injury. Postoperative changes in these markers were associated with AKI progression in adults after cardiac surgery and post-AKI kidney atrophy in mouse models of ischemia-reperfusion injury and toxic injury. Our results demonstrate the feasibility of a multiomics approach to discovering noninvasive markers and associating PT maladaptation with adverse clinical outcomes.

Single cell and spatial transcriptomics analysis of kidney double negative T lymphocytes in normal and ischemic mouse kidneys.

T cells are important in the pathogenesis of acute kidney injury (AKI), and TCR+CD4-CD8- (double negative-DN) are T cells that have regulatory properties. However, there is limited information on DN T cells compared to traditional CD4+ and CD8+ cells. To elucidate the molecular signature and spatial dynamics of DN T cells during AKI, we performed single-cell RNA sequencing (scRNA-seq) on sorted murine DN, CD4+, and CD8+ cells combined with spatial transcriptomic profiling of normal and post AKI mouse kidneys. scRNA-seq revealed distinct transcriptional profiles for DN, CD4+, and CD8+ T cells of mouse kidneys with enrichment of Kcnq5, Klrb1c, Fcer1g, and Klre1 expression in DN T cells compared to CD4+ and CD8+ T cells in normal kidney tissue. We validated the expression of these four genes in mouse kidney DN, CD4+ and CD8+ T cells using RT-PCR and Kcnq5, Klrb1, and Fcer1g genes with the NIH human kidney precision medicine project (KPMP). Spatial transcriptomics in normal and ischemic mouse kidney tissue showed a localized cluster of T cells in the outer medulla expressing DN T cell genes including Fcer1g. These results provide a template for future studies in DN T as well as CD4+ and CD8+ cells in normal and diseased kidneys.

A systematic comparison of computational methods for expression forecasting.

Due to the abundance of single cell RNA-seq data, a number of methods for predicting expression after perturbation have recently been published. Expression prediction methods are enticing because they promise to answer pressing questions in fields ranging from developmental genetics to cell fate engineering and because they are faster, cheaper, and higher-throughput than their experimental counterparts. However, the absolute and relative accuracy of these methods is poorly characterized, limiting their informed use, their improvement, and the interpretation of their predictions. To address these issues, we created a benchmarking platform that combines a panel of large-scale perturbation datasets with an expression forecasting software engine that encompasses or interfaces to current methods. We used our platform to systematically assess methods, parameters, and sources of auxiliary data. We found that uninformed baseline predictions, which were not always included in prior evaluations, yielded the same or better mean absolute error than benchmarked methods in all test cases. These results cast doubt on the ability of current expression forecasting methods to provide mechanistic insights or to rank hypotheses for experimental follow-up. However, given the rapid pace of innovation in the field, new approaches may yield more accurate expression predictions. Our platform will serve as a neutral benchmark to improve methods and to identify contexts in which expression prediction can succeed.

Platform-agnostic CellNet enables cross-study analysis of cell fate engineering protocols.

Optimization of cell engineering protocols requires standard, comprehensive quality metrics. We previously developed CellNet, a computational tool to quantitatively assess the transcriptional fidelity of engineered cells compared with their natural counterparts, based on bulk-derived expression profiles. However, this platform and others were limited in their ability to compare data from different sources, and no current tool makes it easy to compare new protocols with existing state-of-the-art protocols in a standardized manner. Here, we utilized our prior application of the top-scoring pair transformation to build a computational platform, platform-agnostic CellNet (PACNet), to address both shortcomings. To demonstrate the utility of PACNet, we applied it to thousands of samples from over 100 studies that describe dozens of protocols designed to produce seven distinct cell types. We performed an in-depth examination of hepatocyte and cardiomyocyte protocols to identify the best-performing methods, characterize the extent of intra-protocol and inter-lab variation, and identify common off-target signatures, including a surprising neural/neuroendocrine signature in primary liver-derived organoids. We have made PACNet available as an easy-to-use web application, allowing users to assess their protocols relative to our database of reference engineered samples, and as open-source, extensible code.

Sympathetic Innervation Regulates Osteocyte-Mediated Cortical Bone Resorption during Lactation.

Bone undergoes constant remodeling by osteoclast bone resorption coupled with osteoblast bone formation at the bone surface. A third major cell type in the bone is osteocytes, which are embedded in the matrix, are well-connected to the lacunar network, and are believed to act as mechanical sensors. Here, it is reported that sympathetic innervation directly regulates lacunar osteocyte-mediated bone resorption inside cortical bone. It is found that sympathetic activity is elevated in different mouse models of bone loss, including lactation, ovariectomy, and glucocorticoid treatment. Further, during lactation elevated sympathetic outflow induces netrin-1 expression by osteocytes to further promote sympathetic nerve sprouting in the cortical bone endosteum in a feed-forward loop. Depletion of tyrosine hydroxylase-positive (TH+ ) sympathetic nerves ameliorated osteocyte-mediated perilacunar bone resorption in lactating mice. Moreover, norepinephrine activates ß-adrenergic receptor 2 (Adrb2) signaling to promote secretion of extracellular vesicles (EVs) containing bone-degrading enzymes for perilacunar bone resorption and inhibit osteoblast differentiation. Importantly, osteocyte-specific deletion of Adrb2 or treatment with a ß-blocker results in lower bone resorption in lactating mice. Together, these findings show that the sympathetic nervous system promotes osteocyte-driven bone loss during lactation, likely as an adaptive response to the increased energy and mineral demands of the nursing mother.

RANK+TLR2+ myeloid subpopulation converts autoimmune to joint destruction in rheumatoid arthritis.

Joint destruction is the major clinic burden in patients with rheumatoid arthritis (RA). It is unclear, though, how this autoimmune disease progresses to the point of deterioration of the joint. Here, we report that in a mouse model of RA the upregulation of TLR2 expression and its α(2,3) sialylation in RANK+ myeloid monocytes mediate the transition from autoimmunity to osteoclast fusion and bone resorption, resulting in joint destruction. The expression of α(2,3) sialyltransferases was significantly increased in RANK+TLR2+ myeloid monocytes, and their inhibition or treatment with a TLR2 inhibitor blocked osteoclast fusion. Notably, analysis of our single-cell RNA-sequencing (scRNA-seq) libraries generated from RA mice revealed a novel RANK+TLR2- a subset that negatively regulated osteoclast fusion. Importantly, the RANK+TLR2+ subset was significantly diminished with the treatments, whereas the RANK+TLR2- subset was expanded. Moreover, the RANK+TLR2- subset could differentiate into a TRAP+ osteoclast lineage, but the resulting cells did not fuse to form osteoclasts. Our scRNA-seq data showed that Maf is highly expressed in the RANK+TLR2- subset, and the α(2,3) sialyltransferase inhibitor-induced Maf expression in the RANK+TLR2+ subset. The identification of a RANK+TLR2- subset provides a potential explanation for TRAP+ mononuclear cells in bone and their anabolic activity. Further, TLR2 expression and its α(2,3) sialylation in the RANK+ myeloid monocytes could be effective targets to prevent autoimmune-mediated joint destruction.

Gene regulation network inference using k-nearest neighbor-based mutual information estimation: revisiting an old DREAM.

BACKGROUND: A cell exhibits a variety of responses to internal and external cues. These responses are possible, in part, due to the presence of an elaborate gene regulatory network (GRN) in every single cell. In the past 20 years, many groups worked on reconstructing the topological structure of GRNs from large-scale gene expression data using a variety of inference algorithms. Insights gained about participating players in GRNs may ultimately lead to therapeutic benefits. Mutual information (MI) is a widely used metric within this inference/reconstruction pipeline as it can detect any correlation (linear and non-linear) between any number of variables (n-dimensions). However, the use of MI with continuous data (for example, normalized fluorescence intensity measurement of gene expression levels) is sensitive to data size, correlation strength and underlying distributions, and often requires laborious and, at times, ad hoc optimization. RESULTS: In this work, we first show that estimating MI of a bi- and tri-variate Gaussian distribution using k-nearest neighbor (kNN) MI estimation results in significant error reduction as compared to commonly used methods based on fixed binning. Second, we demonstrate that implementing the MI-based kNN Kraskov-Stoögbauer-Grassberger (KSG) algorithm leads to a significant improvement in GRN reconstruction for popular inference algorithms, such as Context Likelihood of Relatedness (CLR). Finally, through extensive in-silico benchmarking we show that a new inference algorithm CMIA (Conditional Mutual Information Augmentation), inspired by CLR, in combination with the KSG-MI estimator, outperforms commonly used methods. CONCLUSIONS: Using three canonical datasets containing 15 synthetic networks, the newly developed method for GRN reconstruction-which combines CMIA, and the KSG-MI estimator-achieves an improvement of 20-35% in precision-recall measures over the current gold standard in the field. This new method will enable researchers to discover new gene interactions or better choose gene candidates for experimental validations.

Immune Checkpoint Molecule TIGIT Regulates Kidney T Cell Functions and Contributes to AKI.

SIGNIFICANCE STATEMENT: T cells mediate pathogenic and reparative processes during AKI, but the exact mechanisms regulating kidney T cell functions are unclear. This study identified upregulation of the novel immune checkpoint molecule, TIGIT, on mouse and human kidney T cells after AKI. TIGIT-expressing kidney T cells produced proinflammatory cytokines and had effector (EM) and central memory (CM) phenotypes. TIGIT-deficient mice had protection from both ischemic and nephrotoxic AKI. Single-cell RNA sequencing led to the discovery of possible downstream targets of TIGIT. TIGIT mediates AKI pathophysiology, is a promising novel target for AKI therapy, and is being increasingly studied in human cancer therapy trials. BACKGROUND: T cells play pathogenic and reparative roles during AKI. However, mechanisms regulating T cell responses are relatively unknown. We investigated the roles of the novel immune checkpoint molecule T cell immunoreceptor with Ig and immunoreceptor tyrosine-based inhibitory motif domains (TIGIT) in kidney T cells and AKI outcomes. METHODS: TIGIT expression and functional effects were evaluated in mouse kidney T cells using RNA sequencing (RNA-Seq) and flow cytometry. TIGIT effect on AKI outcomes was studied with TIGIT knockout (TIGIT-KO) mice in ischemia reperfusion (IR) and cisplatin AKI models. Human kidney T cells from nephrectomy samples and single cell RNA sequencing (scRNA-Seq) data from the Kidney Precision Medicine Project were used to assess TIGIT's role in humans. RESULTS: RNA-Seq and flow cytometry analysis of mouse kidney CD4+ T cells revealed increased expression of TIGIT after IR injury. Ischemic injury also increased TIGIT expression in human kidney T cells, and TIGIT expression was restricted to T/natural killer cell subsets in patients with AKI. TIGIT-expressing kidney T cells in wild type (WT) mice had an effector/central memory phenotype and proinflammatory profile at baseline and post-IR. Kidney regulatory T cells were predominantly TIGIT+ and significantly reduced post-IR. TIGIT-KO mice had significantly reduced kidney injury after IR and nephrotoxic injury compared with WT mice. scRNA-Seq analysis showed enrichment of genes related to oxidative phosphorylation and mTORC1 signaling in Th17 cells from TIGIT-KO mice. CONCLUSIONS: TIGIT expression increases in mouse and human kidney T cells during AKI, worsens AKI outcomes, and is a novel therapeutic target for AKI.

Engineering Skeletal Muscle Grafts with PAX7::GFP-Sorted Human Pluripotent Stem Cell-Derived Myogenic Progenitors on Fibrin Microfiber Bundles for Tissue Regeneration.

Tissue engineering strategies that combine human pluripotent stem cell-derived myogenic progenitors (hPDMs) with advanced biomaterials provide promising tools for engineering 3D skeletal muscle grafts to model tissue development in vitro and promote muscle regeneration in vivo. We recently demonstrated (i) the potential for obtaining large numbers of hPDMs using a combination of two small molecules without the overexpression of transgenes and (ii) the application of electrospun fibrin microfiber bundles for functional skeletal muscle restoration following volumetric muscle loss. In this study, we aimed to demonstrate that the biophysical cues provided by the fibrin microfiber bundles induce hPDMs to form engineered human skeletal muscle grafts containing multinucleated myotubes that express desmin and myosin heavy chains and that these grafts could promote regeneration following skeletal muscle injuries. We tested a genetic PAX7 reporter line (PAX7::GFP) to sort for more homogenous populations of hPDMs. RNA sequencing and gene set enrichment analyses confirmed that PAX7::GFP-sorted hPDMs exhibited high expression of myogenic genes. We tested engineered human skeletal muscle grafts derived from PAX7::GFP-sorted hPDMs within in vivo skeletal muscle defects by assessing myogenesis, engraftment and immunogenicity using immunohistochemical staining. The PAX7::GFP-sorted groups had moderately high vascular infiltration and more implanted cell association with embryonic myosin heavy chain (eMHC) regions, suggesting they induced pro-regenerative microenvironments. These findings demonstrated the promise for the use of PAX7::GFP-sorted hPDMs on fibrin microfiber bundles and provided some insights for improving the cell-biomaterial system to stimulate more robust in vivo skeletal muscle regeneration.

Reconstruction of dynamic regulatory networks reveals signaling-induced topology changes associated with germ layer specification.

Elucidating regulatory relationships between transcription factors (TFs) and target genes is fundamental to understanding how cells control their identity and behavior. Unfortunately, existing computational gene regulatory network (GRN) reconstruction methods are imprecise, computationally burdensome, and fail to reveal dynamic regulatory topologies. Here, we present Epoch, a reconstruction tool that uses single-cell transcriptomics to accurately infer dynamic networks. We apply Epoch to identify the dynamic networks underpinning directed differentiation of mouse embryonic stem cells (ESCs) guided by multiple signaling pathways, and we demonstrate that modulating these pathways drives topological changes that bias cell fate potential. We also find that Peg3 rewires the pluripotency network to favor mesoderm specification. By integrating signaling pathways with GRNs, we trace how Wnt activation and PI3K suppression govern mesoderm and endoderm specification, respectively. Finally, we identify regulatory circuits of patterning and axis formation that distinguish in vitro and in vivo mesoderm specification.

Single-Cell Transcriptomic Census of Endothelial Changes Induced by Matrix Stiffness and the Association with Atherosclerosis.

Vascular endothelial cell (EC) plasticity plays a critical role in the progression of atherosclerosis by giving rise to mesenchymal phenotypes in the plaque lesion. Despite the evidence for arterial stiffening as a major contributor to atherosclerosis, the complex interplay among atherogenic stimuli in vivo has hindered attempts to determine the effects of extracellular matrix (ECM) stiffness on endothelial-mesenchymal transition (EndMT). To study the regulatory effects of ECM stiffness on EndMT, an in vitro model is developed in which human coronary artery ECs are cultured on physiological or pathological stiffness substrates. Leveraging single-cell RNA sequencing, cell clusters with mesenchymal transcriptional features are identified to be more prevalent on pathological substrates than physiological substrates. Trajectory inference analyses reveal a novel mesenchymal-to-endothelial reverse transition, which is blocked by pathological stiffness substrates, in addition to the expected EndMT trajectory. ECs pushed to a mesenchymal character by pathological stiffness substrates are enriched in transcriptional signatures of atherosclerotic ECs from human and murine plaques. This study characterizes at single-cell resolution the transcriptional programs that underpin EC plasticity in both physiological or pathological milieus, and thus serves as a valuable resource for more precisely defining EndMT and the transcriptional programs contributing to atherosclerosis.

Spatial transcriptomics reveals a role for sensory nerves in preserving cranial suture patency through modulation of BMP/TGF-ß signaling.

The patterning and ossification of the mammalian skeleton requires the coordinated actions of both intrinsic bone morphogens and extrinsic neurovascular signals, which function in a temporal and spatial fashion to control mesenchymal progenitor cell (MPC) fate. Here, we show the genetic inhibition of tropomyosin receptor kinase A (TrkA) sensory nerve innervation of the developing cranium results in premature calvarial suture closure, associated with a decrease in suture MPC proliferation and increased mineralization. In vitro, axons from peripheral afferent neurons derived from dorsal root ganglions (DRGs) of wild-type mice induce MPC proliferation in a spatially restricted manner via a soluble factor when cocultured in microfluidic chambers. Comparative spatial transcriptomic analysis of the cranial sutures in vivo confirmed a positive association between sensory axons and proliferative MPCs. SpatialTime analysis across the developing suture revealed regional-specific alterations in bone morphogenetic protein (BMP) and TGF-ß signaling pathway transcripts in response to TrkA inhibition. RNA sequencing of DRG cell bodies, following direct, axonal coculture with MPCs, confirmed the alterations in BMP/TGF-ß signaling pathway transcripts. Among these, the BMP inhibitor follistatin-like 1 (FSTL1) replicated key features of the neural-to-bone influence, including mitogenic and anti-osteogenic effects via the inhibition of BMP/TGF-ß signaling. Taken together, our results demonstrate that sensory nerve-derived signals, including FSTL1, function to coordinate cranial bone patterning by regulating MPC proliferation and differentiation in the suture mesenchyme.

Computational reconstruction of the signalling networks surrounding implanted biomaterials from single-cell transcriptomics.

The understanding of the foreign-body responses to implanted biomaterials would benefit from the reconstruction of intracellular and intercellular signalling networks in the microenvironment surrounding the implant. Here, by leveraging single-cell RNA-sequencing data from 42,156 cells collected from the site of implantation of either polycaprolactone or an extracellular-matrix-derived scaffold in a mouse model of volumetric muscle loss, we report a computational analysis of intercellular signalling networks reconstructed from predictions of transcription-factor activation. We found that intercellular signalling networks can be clustered into modules associated with specific cell subsets, and that biomaterial-specific responses can be characterized by interactions between signalling modules for immune, fibroblast and tissue-specific cells. In a Il17ra-/- mouse model, we validated that predicted interleukin-17-linked transcriptional targets led to concomitant changes in gene expression. Moreover, we identified cell subsets that had not been implicated in the responses to implanted biomaterials. Single-cell atlases of the cellular responses to implanted biomaterials will facilitate the design of implantable biomaterials and the understanding of the ensuing cellular responses.

Evaluating the transcriptional fidelity of cancer models.

BACKGROUND: Cancer researchers use cell lines, patient-derived xenografts, engineered mice, and tumoroids as models to investigate tumor biology and to identify therapies. The generalizability and power of a model derive from the fidelity with which it represents the tumor type under investigation; however, the extent to which this is true is often unclear. The preponderance of models and the ability to readily generate new ones has created a demand for tools that can measure the extent and ways in which cancer models resemble or diverge from native tumors. METHODS: We developed a machine learning-based computational tool, CancerCellNet, that measures the similarity of cancer models to 22 naturally occurring tumor types and 36 subtypes, in a platform and species agnostic manner. We applied this tool to 657 cancer cell lines, 415 patient-derived xenografts, 26 distinct genetically engineered mouse models, and 131 tumoroids. We validated CancerCellNet by application to independent data, and we tested several predictions with immunofluorescence. RESULTS: We have documented the cancer models with the greatest transcriptional fidelity to natural tumors, we have identified cancers underserved by adequate models, and we have found models with annotations that do not match their classification. By comparing models across modalities, we report that, on average, genetically engineered mice and tumoroids have higher transcriptional fidelity than patient-derived xenografts and cell lines in four out of five tumor types. However, several patient-derived xenografts and tumoroids have classification scores that are on par with native tumors, highlighting both their potential as faithful model classes and their heterogeneity. CONCLUSIONS: CancerCellNet enables the rapid assessment of transcriptional fidelity of tumor models. We have made CancerCellNet available as a freely downloadable R package ( https://github.com/pcahan1/cancerCellNet ) and as a web application ( http://www.cahanlab.org/resources/cancerCellNet_web ) that can be applied to new cancer models that allows for direct comparison to the cancer models evaluated here.

Challenges for Computational Stem Cell Biology: A Discussion for the Field.

The first meetup for Computational Stem Cell Biologists was held at the 2020 annual meeting of the International Society for Stem Cell Research. The discussions highlighted opportunities and barriers to computational stem cell research that require coordinated action across the stem cell sector.

Computational Stem Cell Biology: Open Questions and Guiding Principles.

Computational biology is enabling an explosive growth in our understanding of stem cells and our ability to use them for disease modeling, regenerative medicine, and drug discovery. We discuss four topics that exemplify applications of computation to stem cell biology: cell typing, lineage tracing, trajectory inference, and regulatory networks. We use these examples to articulate principles that have guided computational biology broadly and call for renewed attention to these principles as computation becomes increasingly important in stem cell biology. We also discuss important challenges for this field with the hope that it will inspire more to join this exciting area.

Gene Regulatory Network Analysis and Engineering Directs Development and Vascularization of Multilineage Human Liver Organoids.

Pluripotent stem cell (PSC)-derived organoids have emerged as novel multicellular models of human tissue development but display immature phenotypes, aberrant tissue fates, and a limited subset of cells. Here, we demonstrate that integrated analysis and engineering of gene regulatory networks (GRNs) in PSC-derived multilineage human liver organoids direct maturation and vascular morphogenesis in vitro. Overexpression of PROX1 and ATF5, combined with targeted CRISPR-based transcriptional activation of endogenous CYP3A4, reprograms tissue GRNs and improves native liver functions, such as FXR signaling, CYP3A4 enzymatic activity, and stromal cell reactivity. The engineered tissues possess superior liver identity when compared with other PSC-derived liver organoids and show the presence of hepatocyte, biliary, endothelial, and stellate-like cell populations in single-cell RNA-seq analysis. Finally, they show hepatic functions when studied in vivo. Collectively, our approach provides an experimental framework to direct organogenesis in vitro by systematically probing molecular pathways and transcriptional networks that promote tissue development.

Hepatocyte-like cells derived from human induced pluripotent stem cells using small molecules: implications of a transcriptomic study.

BACKGROUND: Hepatocyte-like cells (HLCs) derived from human induced pluripotent stem cells (iPSCs) hold great promise in toxicological applications as well as in regenerative medicine. Previous efforts on hepatocyte differentiation have mostly relied on the use of growth factors (GFs) to recapitulate developmental signals under in vitro conditions. Recently, the use of small molecules (SMs) has emerged as an attractive tool to induce cell fate transition due to its superiority in terms of both quality and cost. However, HLCs derived using SMs have not been well characterized, especially on the transcriptome level. METHODS: HLCs were differentiated from human iPSCs using a protocol that only involves SMs and characterized by transcriptomic analysis using whole genome microarrays. RESULTS: HLCs derived using the SM protocol (HLC_SM) displayed specific hepatic marker expression and demonstrated key hepatic functions. Transcriptomic analysis of the SM-driven differentiation defined a hepatocyte differentiation track and characterized the expression of some key marker genes in major stages of hepatocyte differentiation. In addition, HLC_SM were scored with CellNet, a bioinformatics tool quantifying how closely engineered cell populations resemble their target cell type, and compared to primary human hepatocytes (PHHs), adult liver tissue, fetal liver tissue, HLCs differentiated using GFs (HLC_GF), and commercially available HLCs. Similar to HLC_GF, HLC_SM displayed a mixed phenotype of fetal and adult hepatocytes and had relatively low expression of metabolic enzymes, transporters, and nuclear receptors compared to PHHs. Finally, the differentially expressed genes in HLC_SM compared to HLC_GF and to PHHs were analyzed to identify pathways and upstream transcription regulators which could potentially be manipulated to improve the differentiation of HLCs. CONCLUSIONS: Overall, the present study demonstrated the usefulness of the SM-based hepatocyte differentiation method, offered new insights into the molecular basis of hepatogenesis and associated gene regulation, and suggested ways for further improvements in hepatocyte differentiation in order to obtain more mature HLCs that could be used in toxicological studies.