Stellate cells are mesenchymal stem cells

Continually adapting pre-trained language model to universal annotation of single-cell RNA-seq data

MOTIVATION

Cell-type annotation of single-cell RNA-sequencing (scRNA-seq) data is a hallmark of biomedical research and clinical application. Current annotation tools usually assume the simultaneous acquisition of well-annotated data, but without the ability to expand knowledge from new data. Yet, such tools are inconsistent with the continuous emergence of scRNA-seq data, calling for a continuous cell-type annotation model. In addition, by their powerful ability of information integration and model interpretability, transformer-based pre-trained language models have led to breakthroughs in single-cell biology research. Therefore, the systematic combining of continual learning and pre-trained language models for cell-type annotation tasks is inevitable.

RESULTS

We herein propose a universal cell-type annotation tool, called CANAL, that continuously fine-tunes a pre-trained language model trained on a large amount of unlabeled scRNA-seq data, as new well-labeled data emerges. CANAL essentially alleviates the dilemma of catastrophic forgetting, both in terms of model inputs and outputs. For model inputs, we introduce an experience replay schema that repeatedly reviews previous vital examples in current training stages. This is achieved through a dynamic example bank with a fixed buffer size. The example bank is class-balanced and proficient in retaining cell-type-specific information, particularly facilitating the consolidation of patterns associated with rare cell types. For model outputs, we utilize representation knowledge distillation to regularize the divergence between previous and current models, resulting in the preservation of knowledge learned from past training stages. Moreover, our universal annotation framework considers the inclusion of new cell types throughout the fine-tuning and testing stages. We can continuously expand the cell-type annotation library by absorbing new cell types from newly arrived, well-annotated training datasets, as well as automatically identify novel cells in unlabeled datasets. Comprehensive experiments with data streams under various biological scenarios demonstrate the versatility and high model interpretability of CANAL.

AVAILABILITY

An implementation of CANAL is available from https://github.com/aster-ww/CANAL-torch.

CONTACT

dengmh@pku.edu.cn.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Journal Name online.

Emerging role of pancreatic stellate cell-derived extracellular vesicles in pancreatic cancer

Pancreatic ductal adenocarcinoma (PDAC) is a highly aggressive cancer that is characterised by a prominent collagenous stromal reaction/desmoplasia surrounding tumour cells. Pancreatic stellate cells (PSCs) are responsible for the production of this stroma and have been shown to facilitate PDAC progression. Recently, extracellular vesicles (EVs), in particular, small extracellular vesicles (exosomes) have been a topic of interest in the field of cancer research for their emerging roles in cancer progression and diagnosis. EVs act as a form of intercellular communication by carrying their molecular cargo from one cell to another, regulating functions of the recipient cells. Although the knowledge of the bi-directional interactions between the PSCs and cancer cells that promote disease progression has advanced significantly over the past decade, studies on PSC-derived EVs in PDAC are currently rather limited. This review provides an overview of PDAC, pancreatic stellate cells and their interactions with cancer cells, as well as the currently known role of extracellular vesicles derived from PSCs in PDAC progression.

Perivascular Mesenchymal Stem/Stromal Cells, an Immune Privileged Niche for Viruses?

Mesenchymal stem cells (MSCs) play a critical role in response to stress such as infection. They initiate the removal of cell debris, exert major immunoregulatory activities, control pathogens, and lead to a remodeling/scarring phase. Thus, host-derived ‘danger’ factors released from damaged/infected cells (called alarmins, e.g., HMGB1, ATP, DNA) as well as pathogen-associated molecular patterns (LPS, single strand RNA) can activate MSCs located in the parenchyma and around vessels to upregulate the expression of growth factors and chemoattractant molecules that influence immune cell recruitment and stem cell mobilization. MSC, in an ultimate contribution to tissue repair, may also directly trans- or de-differentiate into specific cellular phenotypes such as osteoblasts, chondrocytes, lipofibroblasts, myofibroblasts, Schwann cells, and they may somehow recapitulate their neural crest embryonic origin. Failure to terminate such repair processes induces pathological scarring, termed fibrosis, or vascular calcification. Interestingly, many viruses and particularly those associated to chronic infection and inflammation may hijack and polarize MSC’s immune regulatory activities. Several reports argue that MSC may constitute immune privileged sanctuaries for viruses and contributing to long-lasting effects posing infectious challenges, such as viruses rebounding in immunocompromised patients or following regenerative medicine therapies using MSC. We will herein review the capacity of several viruses not only to infect but also to polarize directly or indirectly the functions of MSC (immunoregulation, differentiation potential, and tissue repair) in clinical settings.

Autophagy: A Key Player in Pancreatic Cancer Progression and a Potential Drug Target

Pancreatic cancer is known to have the lowest survival outcomes among all major cancers, and unfortunately, this has only been marginally improved over last four decades. The innate characteristics of pancreatic cancer include an aggressive and fast-growing nature from powerful driver mutations, a highly defensive tumor microenvironment and the upregulation of advantageous survival pathways such as autophagy. Autophagy involves targeted degradation of proteins and organelles to provide a secondary source of cellular supplies to maintain cell growth. Elevated autophagic activity in pancreatic cancer is recognized as a major survival pathway as it provides a plethora of support for tumors by supplying vital resources, maintaining tumour survival under the stressful microenvironment and promoting other pathways involved in tumour progression and metastasis. The combination of these features is unique to pancreatic cancer and present significant resistance to chemotherapeutic strategies, thus, indicating a need for further investigation into therapies targeting this crucial pathway. This review will outline the autophagy pathway and its regulation, in addition to the genetic landscape and tumor microenvironment that contribute to pancreatic cancer severity. Moreover, this review will also discuss the mechanisms of novel therapeutic strategies that inhibit autophagy and how they could be used to suppress tumor progression.

Human Induced Pluripotent Stem Cell-Derived Mesenchymal Stem Cells Acquire Rejuvenation and Reduced Heterogeneity

Despite the uniform selection criteria for the isolation of human mesenchymal stem cells (MSCs), considerable heterogeneity exists which reflects the distinct tissue origins and differences between individuals with respect to their genetic background and age. This heterogeneity is manifested by the variabilities seen in the transcriptomes, proteomes, secretomes, and epigenomes of tissue-specific MSCs. Here, we review literature on different aspects of MSC heterogeneity including the role of epigenetics and the impact of MSC heterogeneity on therapies. We then combine this with a meta-analysis of transcriptome data from distinct MSC subpopulations derived from bone marrow, adipose tissue, cruciate, tonsil, kidney, umbilical cord, fetus, and induced pluripotent stem cells derived MSCs (iMSCs). Beyond that, we investigate transcriptome differences between tissue-specific MSCs and pluripotent stem cells. Our meta-analysis of numerous MSC-related data sets revealed markers and associated biological processes characterizing the heterogeneity and the common features of MSCs from various tissues. We found that this heterogeneity is mainly related to the origin of the MSCs and infer that microenvironment and epigenetics are key drivers. The epigenomes of MSCs alter with age and this has a profound impact on their differentiation capabilities. Epigenetic modifications of MSCs are propagated during cell divisions and manifest in differentiated cells, thus contributing to diseased or healthy phenotypes of the respective tissue. An approach used to reduce heterogeneity caused by age- and tissue-related epigenetic and microenvironmental patterns is the iMSC concept: iMSCs are MSCs generated from induced pluripotent stem cells (iPSCs). During iMSC generation epigenetic and chromatin remodeling result in a gene expression pattern associated with rejuvenation thus allowing to overcome age-related shortcomings (e.g., limited differentiation and proliferation capacity). The importance of the iMSC concept is underlined by multiple clinical trials. In conclusion, we propose the use of rejuvenated iMSCs to bypass tissue- and age-related heterogeneity which are associated with native MSCs.

The double edge sword of fibrosis in cancer

Cancer-associated fibrosis is a critical component of the tumor microenvironment (TME) which significantly impacts cancer behavior. However, there is significant controversy regarding fibrosis as a predominantly tumor promoting or tumor suppressing factor. Cells essential to the generation of tissue fibrosis such as fibroblasts and mesenchymal stem cells (MSCs) have dual phenotypes dependent upon their independence or association with cancer cells. Cancer-associated fibroblasts and cancer-associated MSCs have unique molecular profiles which facilitate cancer cell cross talk, influence extracellular matrix deposition, and direct the immune system to generate a protumorigenic environment. In contrast, normal tissue fibroblasts and MSCs are important in restraining cancer initiation, influencing epithelial cell differentiation, and limiting cancer cell invasion. We propose this apparent dichotomy of function is due to (1) cancer mediated stromal reprogramming; (2) tissue stromal source; (3) unique subtypes of fibrosis; and (4) the impact of fibrosis on other TME elements. First, as cancer progresses, tumor cells influence their surrounding stroma to move from a cancer restraining phenotype into a cancer supportive role. Second, cancer has specific organ tropism, thus stroma derived from preferred metastatic organs support growth while less preferred metastatic tissues do not. Third, there are subtypes of fibrosis which have unique function to support or inhibit cancer growth. Fourth, depleting fibrosis influences other TME components which drive the cancer response. Collectively, this review highlights the complexity of cancer-associated fibrosis and supports a dual function of fibrosis which evolves during the continuum of cancer growth.

Stem Cell Differentiation is Regulated by Extracellular Matrix Mechanics

Stem cells mechanosense the stiffness of their microenvironment, which impacts differentiation. Although tissue hydration anti-correlates with stiffness, extracellular matrix (ECM) stiffness is clearly transduced into gene expression via adhesion and cytoskeleton proteins that tune fates. Cytoskeletal reorganization of ECM can create heterogeneity and influence fates, with fibrosis being one extreme.

A Rising Star in Pancreatic Diseases: Pancreatic Stellate Cells

Pancreatic stellate cell (PSC) is a type of pluripotent cell located between pancreatic lobules and the surrounding area of acinars. When activated, PSC can be transformed into myofibroblast-like cell. A number of evidences suggest that activated PSC is the main source of the accumulation of extracellular matrix (ECM) protein under the pathological conditions, which lead to pancreatic fibrosis in chronic pancreatitis and pancreatic cancer. Recent studies have found that PSC also plays an important role in the endocrine cell function, islet fibrosis and diabetes. In order to provide new strategies for the treatment of pancreatic diseases, this paper systematically summarizes the recent researches about the biological behaviors of PSC, including its stem/progenitor cell characteristics, secreted exosomes, cellular senescence, epithelial mesenchymal transformation (EMT), energy metabolism and direct mechanical reprogramming.

Expression and Clinical Significance of Cancer Stem Cell Markers CD24, CD44, and CD133 in Pancreatic Ductal Adenocarcinoma and Chronic Pancreatitis

Cancer stem cells (CSC) play an important role in pancreatic carcinogenesis and prognosis. The study aimed at examining the expression of CD24, CD44, and CD133 in human PDAC and CP in order to evaluate its clinicopathological correlations and the clinical significance. Surgical specimens from 23 patients with PDAC and 15 patients with chronic pancreatitis after pancreatic resection were stained with CD24, CD44, and CD133 antibodies. The intensity of staining was scored from 0 (negative) to 3 (strongly positive). Results. Mean CD24 staining score in PDAC was 1.38 ± 0.76 and was significantly higher than that in CP: 0.70 ± 0.53 (p < 0.01); CD44 score in PDAC was 2.23 ± 0.42 and was significantly higher than that in CP: 1.87 ± 0.55 (p < 0.05); CD133 score 0.93 ± 0.58 was not different from CP: 0.71 ± 0.43 (p > 0.05). CD44 immunoreactivity was significantly higher (p < 0.05) in pT1 and pT2 patients together as regards pT3: 2.45 ± 0.37 versus 2.06 ± 0.38 as well as in N0 patients compared to N1 patients: 2.5 ± 0.38 versus 2.04 ± 0.34. Conclusions. CD24 and CD44 are upregulated in human pancreatic cancer compared to chronic pancreatitis. CD44 immunoreactivity decreases with the tumor advancement and may represent the negative PDAC prognostic factor. Each CSC marker was differently related to PDAC advancement. CD133 may lack clinical significance in PDAC.