Screening of normal endoscopic large bowel biopsies with interpretable graph learning: a retrospective study.

OBJECTIVE: To develop an interpretable artificial intelligence algorithm to rule out normal large bowel endoscopic biopsies, saving pathologist resources and helping with early diagnosis. DESIGN: A graph neural network was developed incorporating pathologist domain knowledge to classify 6591 whole-slides images (WSIs) of endoscopic large bowel biopsies from 3291 patients (approximately 54% female, 46% male) as normal or abnormal (non-neoplastic and neoplastic) using clinically driven interpretable features. One UK National Health Service (NHS) site was used for model training and internal validation. External validation was conducted on data from two other NHS sites and one Portuguese site. RESULTS: Model training and internal validation were performed on 5054 WSIs of 2080 patients resulting in an area under the curve-receiver operating characteristic (AUC-ROC) of 0.98 (SD=0.004) and AUC-precision-recall (PR) of 0.98 (SD=0.003). The performance of the model, named Interpretable Gland-Graphs using a Neural Aggregator (IGUANA), was consistent in testing over 1537 WSIs of 1211 patients from three independent external datasets with mean AUC-ROC=0.97 (SD=0.007) and AUC-PR=0.97 (SD=0.005). At a high sensitivity threshold of 99%, the proposed model can reduce the number of normal slides to be reviewed by a pathologist by approximately 55%. IGUANA also provides an explainable output highlighting potential abnormalities in a WSI in the form of a heatmap as well as numerical values associating the model prediction with various histological features. CONCLUSION: The model achieved consistently high accuracy showing its potential in optimising increasingly scarce pathologist resources. Explainable predictions can guide pathologists in their diagnostic decision-making and help boost their confidence in the algorithm, paving the way for its future clinical adoption.

Periostin modulates myofibroblast differentiation during full-thickness cutaneous wound repair.

The matricellular protein periostin is expressed in the skin. Although periostin has been hypothesized to contribute to dermal homeostasis and repair, this has not been directly tested. To assess the contribution of periostin to dermal healing, 6 mm full-thickness excisional wounds were created in the skin of periostin-knockout and wild-type, sex-matched control mice. In wild-type mice, periostin was potently induced 5-7 days after wounding. In the absence of periostin, day 7 wounds showed a significant reduction in myofibroblasts, as visualized by expression of α-smooth muscle actin (α-SMA) within the granulation tissue. Delivery of recombinant human periostin by electrospun collagen scaffolds restored α-SMA expression. Isolated wild-type and knockout dermal fibroblasts did not differ in in vitro assays of adhesion or migration; however, in 3D culture, periostin-knockout fibroblasts showed a significantly reduced ability to contract a collagen matrix, and adopted a dendritic phenotype. Recombinant periostin restored the defects in cell morphology and matrix contraction displayed by periostin-deficient fibroblasts in a manner that was sensitive to a neutralizing anti-ß1-integrin and to the FAK and Src inhibitor PP2. We propose that periostin promotes wound contraction by facilitating myofibroblast differentiation and contraction.

Social network analysis of cell networks improves deep learning for prediction of molecular pathways and key mutations in colorectal cancer.

Colorectal cancer (CRC) is a primary global health concern, and identifying the molecular pathways, genetic subtypes, and mutations associated with CRC is crucial for precision medicine. However, traditional measurement techniques such as gene sequencing are costly and time-consuming, while most deep learning methods proposed for this task lack interpretability. This study offers a new approach to enhance the state-of-the-art deep learning methods for molecular pathways and key mutation prediction by incorporating cell network information. We build cell graphs with nuclei as nodes and nuclei connections as edges of the network and leverage Social Network Analysis (SNA) measures to extract abstract, perceivable, and interpretable features that explicitly describe the cell network characteristics in an image. Our approach does not rely on precise nuclei segmentation or feature extraction, is computationally efficient, and is easily scalable. In this study, we utilize the TCGA-CRC-DX dataset, comprising 499 patients and 502 diagnostic slides from primary colorectal tumours, sourced from 36 distinct medical centres in the United States. By incorporating the SNA features alongside deep features in two multiple instance learning frameworks, we demonstrate improved performance for chromosomal instability (CIN), hypermutated tumour (HM), TP53 gene, BRAF gene, and Microsatellite instability (MSI) status prediction tasks (2.4%-4% and 7-8.8% improvement in AUROC and AUPRC on average). Additionally, our method achieves outstanding performance on MSI prediction in an external PAIP dataset (99% AUROC and 98% AUPRC), demonstrating its generalizability. Our findings highlight the discrimination power of SNA features and how they can be beneficial to deep learning models' performance and provide insights into the correlation of cell network profiles with molecular pathways and key mutations.

An aggregation of aggregation methods in computational pathology.

Image analysis and machine learning algorithms operating on multi-gigapixel whole-slide images (WSIs) often process a large number of tiles (sub-images) and require aggregating predictions from the tiles in order to predict WSI-level labels. In this paper, we present a review of existing literature on various types of aggregation methods with a view to help guide future research in the area of computational pathology (CPath). We propose a general CPath workflow with three pathways that consider multiple levels and types of data and the nature of computation to analyse WSIs for predictive modelling. We categorize aggregation methods according to the context and representation of the data, features of computational modules and CPath use cases. We compare and contrast different methods based on the principle of multiple instance learning, perhaps the most commonly used aggregation method, covering a wide range of CPath literature. To provide a fair comparison, we consider a specific WSI-level prediction task and compare various aggregation methods for that task. Finally, we conclude with a list of objectives and desirable attributes of aggregation methods in general, pros and cons of the various approaches, some recommendations and possible future directions.

Decoding mTOR signalling heterogeneity in the tumour microenvironment using multiplexed imaging and graph convolutional networks.

Evaluating the contribution of the tumour microenvironment (TME) in tumour progression has proven a complex challenge due to the intricate interactions within the TME. Multiplexed imaging is an emerging technology that allows concurrent assessment of multiple of these components simultaneously. Here we utilise a highly multiplexed dataset of 61 markers across 746 colorectal tumours to investigate how complex mTOR signalling in different tissue compartments influences patient prognosis. We found that the signalling of mTOR pathway can have heterogeneous activation patterns in tumour and immune compartments which correlate with patient prognosis. Using graph neural networks, we determined the most predictive features of mTOR activity in immune cells and identified relevant cellular subpopulations. We validated our observations using spatial transcriptomics data analysis in an independent patient cohort. Our work provides a framework for studying complex cell signalling and reveals important insights for developing mTOR-based therapies.

Cross-linking breast tumor transcriptomic states and tissue histology.

Identification of the gene expression state of a cancer patient from routine pathology imaging and characterization of its phenotypic effects have significant clinical and therapeutic implications. However, prediction of expression of individual genes from whole slide images (WSIs) is challenging due to co-dependent or correlated expression of multiple genes. Here, we use a purely data-driven approach to first identify groups of genes with co-dependent expression and then predict their status from WSIs using a bespoke graph neural network. These gene groups allow us to capture the gene expression state of a patient with a small number of binary variables that are biologically meaningful and carry histopathological insights for clinical and therapeutic use cases. Prediction of gene expression state based on these gene groups allows associating histological phenotypes (cellular composition, mitotic counts, grading, etc.) with underlying gene expression patterns and opens avenues for gaining biological insights from routine pathology imaging directly.

Malignant Mesothelioma subtyping via sampling driven multiple instance prediction on tissue image and cell morphology data.

Malignant Mesothelioma is a difficult to diagnose and highly lethal cancer usually associated with asbestos exposure. It can be broadly classified into three subtypes: Epithelioid, Sarcomatoid, and a hybrid Biphasic subtype in which significant components of both of the previous subtypes are present. Early diagnosis and identification of the subtype informs treatment and can help improve patient outcome. However, the subtyping of malignant mesothelioma, and specifically the recognition of transitional features from routine histology slides has a high level of inter-observer variability. In this work, we propose an end-to-end multiple instance learning (MIL) approach for malignant mesothelioma subtyping. This uses an adaptive instance-based sampling scheme for training deep convolutional neural networks on bags of image patches that allows learning on a wider range of relevant instances compared to max or top-N based MIL approaches. We also investigate augmenting the instance representation to include aggregate cellular morphology features from cell segmentation. The proposed MIL approach enables identification of malignant mesothelial subtypes of specific tissue regions. From this a continuous characterisation of a sample according to predominance of sarcomatoid vs epithelioid regions is possible, thus avoiding the arbitrary and highly subjective categorisation by currently used subtypes. Instance scoring also enables studying tumor heterogeneity and identifying patterns associated with different subtypes. We have evaluated the proposed method on a dataset of 234 tissue micro-array cores with an AUROC of 0.89±0.05 for this task. The dataset and developed methodology is available for the community at: https://github.com/measty/PINS.

MesoGraph: Automatic profiling of mesothelioma subtypes from histological images.

Mesothelioma is classified into three histological subtypes, epithelioid, sarcomatoid, and biphasic, according to the relative proportions of epithelioid and sarcomatoid tumor cells present. Current guidelines recommend that the sarcomatoid component of each mesothelioma is quantified, as a higher percentage of sarcomatoid pattern in biphasic mesothelioma shows poorer prognosis. In this work, we develop a dual-task graph neural network (GNN) architecture with ranking loss to learn a model capable of scoring regions of tissue down to cellular resolution. This allows quantitative profiling of a tumor sample according to the aggregate sarcomatoid association score. Tissue is represented by a cell graph with both cell-level morphological and regional features. We use an external multicentric test set from Mesobank, on which we demonstrate the predictive performance of our model. We additionally validate our model predictions through an analysis of the typical morphological features of cells according to their predicted score.

Expression of integrin beta1 by fibroblasts is required for tissue repair in vivo.

In tissue repair, fibroblasts migrate into the wound to produce and remodel extracellular matrix (ECM). Integrins are believed to be crucial for tissue repair, but their tissue-specific role in this process is poorly understood. Here, we show that mice containing a fibroblast-specific deletion of integrin ß1 exhibit delayed cutaneous wound closure and less granulation tissue formation, including reduced production of new ECM and reduced expression of α-smooth muscle actin (α-SMA). Integrin-ß1-deficient fibroblasts showed reduced expression of type I collagen and connective tissue growth factor, and failed to differentiate into myofibroblasts as a result of reduced α-SMA stress fiber formation. Loss of integrin ß1 in adult fibroblasts reduced their ability to adhere to, to spread on and to contract ECM. Within stressed collagen matrices, integrin-ß1-deficient fibroblasts showed reduced activation of latent TGFß. Addition of active TGFß alleviated the phenotype of integrin-ß1-deficient mice. Thus integrin ß1 is essential for normal wound healing, where it acts, at least in part, through a TGFß-dependent mechanism in vivo.

Post-transcriptional regulation of alpha-smooth muscle actin determines the contractile phenotype of Dupuytren's nodular cells.

The objective was to study Dupuytren's myofibroblast cells in constrained collagen matrices in order to more closely emulate their in vivo environment and, to correlate their contractility with alpha-smooth muscle actin (alpha-SMA) expression and determine if dermal fibroblasts regulate Dupuytren's myofibroblast phenotype. Isotonic and isometric force contraction by cells isolated from Dupuytren's nodules, palmar and non-palmar skin fibroblasts was measured in collagen matrices. The effect of co-culturing nodule cells with dermal fibroblasts on isometric contraction was examined. Isometric contraction was correlated with levels of alpha-SMA mRNA by pcr and protein by Western blotting, and alpha-SMA distribution assessed by immunofluorescence. Dupuytren's nodule cells exhibited similar levels of isotonic contraction to both palmar and non-palmar dermal fibroblasts. However, nodule cells generated high levels of isometric force (mean: 3.5 dynes/h), which continued to increase over 24 h to a maximum of 173 dynes. In contrast, dermal fibroblasts initially exhibited low levels of contraction (mean: 0.5 dynes/h) and reached tensional homeostasis on average after 15 h (range: 4-20 h), with a maximum force of 52 dynes. Although all three cell types had similar alpha-SMA mRNA levels, increased levels of alpha-SMA protein were observed in nodule cells compared to dermal fibroblasts. alpha-SMA localised to stress fibres in 35% (range: 26-50%) of nodule cells compared to only 3% (range:0-6%) of dermal fibroblasts. Co-cultures of Dupuytren's cells and dermal fibroblasts showed no contractile differences. The contractile phenotype of Dupuytren's myofibroblasts is determined by increased alpha-SMA protein distributed in stress fibres, not by cellular mRNA levels. Dupuytren's cell contractility is not influenced by dermal fibroblasts.

Rosiglitazone alleviates the persistent fibrotic phenotype of lesional skin scleroderma fibroblasts.

OBJECTIVE: The transcription factor peroxisome proliferator-activated receptor (PPAR)-gamma plays an important role in controlling cell differentiation. The aim of the present study was to examine whether PPAR-gamma expression was reduced in skin scleroderma fibroblasts and whether PPAR-gamma agonists could suppress the persistent fibrotic phenotype of skin scleroderma fibroblasts. METHODS: Dermal fibroblasts were isolated from site-, age- and sex-matched healthy individuals and lesional areas of individuals with dcSSc. Western blot and collagen gel contraction analyses were used to detect protein expression in the presence or absence of the PPAR-gamma agonist rosiglitazone. RESULTS: PPAR-gamma expression was reduced in dcSSc fibroblasts. The PPAR-gamma agonist rosiglitazone alleviated the persistent fibrotic phenotype of dcSSc fibroblasts. CONCLUSION: Rosiglitazone may alleviate the extent of fibrosis in dcSSc.

FAK is required for TGFbeta-induced JNK phosphorylation in fibroblasts: implications for acquisition of a matrix-remodeling phenotype.

Transforming growth factor beta (TGFbeta) plays a critical role in connective tissue remodeling by fibroblasts during development, tissue repair, and fibrosis. We investigated the molecular pathways in the transmission of TGFbeta signals that lead to features of connective tissue remodeling, namely formation of an alpha-smooth muscle actin (alpha-SMA) cytoskeleton, matrix contraction, and expression of profibrotic genes. TGFbeta causes the activation of focal adhesion kinase (FAK), leading to JNK phosphorylation. TGFbeta induces JNK-dependent actin stress fiber formation, matrix contraction, and expression of profibrotic genes in fak+/+, but not fak-/-, fibroblasts. Overexpression of MEKK1, a kinase acting upstream of JNK, rescues TGFbeta responsiveness of JNK-dependent transcripts and actin stress fiber formation in FAK-deficient fibroblasts. Thus we propose a FAK-MEKK1-JNK pathway in the transmission of TGFbeta signals leading to the control of alpha-SMA cytoskeleton reorganization, matrix contraction, and profibrotic gene expression and hence to the physiological and pathological effects of TGFbeta on connective tissue remodeling by fibroblasts.

Force generation of different human cardiac valve interstitial cells: relevance to individual valve function and tissue engineering.

BACKGROUND AND AIM OF THE STUDY: Cardiac valves perform highly sophisticated functions that depend upon the specific characteristics of the component interstitial cells (ICs). The ability of valve ICs to contribute to these functions may be related to the generation of different types of tension within the valve structure. The study aim was to characterize cellular morphology and the forces generated by valve ICs and to compare this with morphology and forces generated by other cell types. METHODS: Cultured human valve ICs, pericardial fibroblasts and vascular smooth muscle cells were seeded in 3-D collagen gels and placed in a device that accurately measures the forces generated. Cell morphology was determined in seeded gels fixed in glutaraldehyde, stained with toluidine blue and visualized using a high-definition stereo light microscope. RESULTS: Valve ICs generated an average peak force of 30.9 +/- 10.4 dynes over a 24-h period which, unlike other cell types tested, increased as cell density decreased (R = 0.67, p <0.0001). The temporal pattern of force generation in mitral valve cells was significantly faster than in aortic or tricuspid cells (p <0.05). Microscopic examination revealed the formation of cellular processes establishing a cell/cell and cell/matrix network. When externally induced changes in matrix tension occurred, the valve ICs unlike the other cell types - did not respond to restore the previous level of tension. CONCLUSION: Human cardiac valve ICs produce a specific pattern of force generation that may be related to the individual function of each heart valve. The specialized function of these cells may serve as a guide for the choice of candidate cells for tissue engineering heart valves.

Endothelin-1 promotes myofibroblast induction through the ETA receptor via a rac/phosphoinositide 3-kinase/Akt-dependent pathway and is essential for the enhanced contractile phenotype of fibrotic fibroblasts.

The endothelins are a family of endothelium-derived peptides that possess a variety of functions, including vasoconstriction. Endothelin-1 (ET-1) is up-regulated during tissue repair and promotes myofibroblast contraction and migration, hence contributing to matrix remodeling during tissue repair. Here, we show that addition of ET-1 to normal lung fibroblasts induces expression of proteins that contribute to a contractile phenotype, including alpha-smooth muscle actin (alpha-SMA), ezrin, moesin, and paxillin. We confirm that ET-1 enhances the ability of lung fibroblasts to contract extracellular matrix, a function essential for tissue repair, through induction of de novo protein synthesis. Blockade of the Akt/phosphoinositide 3-kinase (PI3-kinase) pathway with LY294002 and wortmannin prevents the ability of ET-1 to induce alpha-SMA, ezrin, paxillin, and moesin and to promote matrix contraction. Dominant negative rac and Akt blocked the ability of ET-1 to promote formation of alpha-SMA stress fibers. Using specific ET-1 receptor inhibitors, we show that ET-1 induces collagen matrix contraction through the ETA, but not the ETB, receptor. Relative to normal pulmonary fibroblasts, fibroblasts cultured from scars of patients with the fibrotic disease systemic sclerosis (scleroderma) show enhanced ET-1 expression and binding. Systemic sclerosis lung fibroblasts show increased ability to contract a collagen matrix and elevated expression of the procontractile proteins alpha-SMA, ezrin, paxillin, and moesin, which are greatly reduced by antagonizing endogenous ET-1 signaling. Thus, blocking ET-1 or the PI3-kinase/Akt cascades might be beneficial in reducing scar formation in pulmonary fibrosis.

Erratum to: Thrombospondin 1 is a key mediator of transforming growth factor b-mediated cell contractility in systemic sclerosis via a mitogen-activated protein kinase kinase (MEK)/extracellular signal-regulated kinase (ERK)-dependent mechanism.

[This corrects the article DOI: 10.1186/1755-1536-4-9.].

New multi-cue bioreactor for tissue engineering of tubular cardiovascular samples under physiological conditions.

In the present study we have developed a multi-cue bioreactor (MCB) that is capable of delivering a range of stimuli to assist the development of a tissue-engineered construct. The MCB provides an accurate and utilizable computer-controlled pulsatile pump and strain induction mechanism and it has the capability of applying physiological conditions to samples. The device described here emulates the pressure and straining environment found at the aortic root. This function, along with an integral perfusion and sterile containment system, allows for long-term culture and whole-tissue testing capability. Aortic and pulmonary arteries were obtained from freshly isolated porcine hearts and subjected to various loading regimens (Deltapressure/flow/force). Through analyzing data acquired by the MCB transducer array it was possible to differentiate the dynamic mechanical properties of the tissue types tested. In addition, the MCB illustrates a novel concept in cardiovascular tissue engineering: being able to support long-term tissue culture of cell-seeded substrates while they are under the influence of mechanical cues. After 7 days of pulsation in the MCB cell alignment was observed. The MCB represents a versatile model that will enable the development of tissue engineering not only for cardiovascular tissue, but for all tubular tissues such as esophageal, tracheal, and bronchial systems.

Thrombospondin 1 is a key mediator of transforming growth factor ß-mediated cell contractility in systemic sclerosis via a mitogen-activated protein kinase kinase (MEK)/extracellular signal-regulated kinase (ERK)-dependent mechanism.

BACKGROUND: The mechanism underlying the ability of fibroblasts to contract a collagen gel matrix is largely unknown. Fibroblasts from scarred (lesional) areas of patients with the fibrotic disease scleroderma show enhanced ability to contract collagen relative to healthy fibroblasts. Thrombospondin 1 (TSP1), an activator of latent transforming growth factor (TGF)ß, is overexpressed by scleroderma fibroblasts. In this report we investigate whether activation of latent TGFß by TSP1 plays a key role in matrix contraction by normal and scleroderma fibroblasts. METHODS: We use the fibroblast populated collagen lattices (FPCL) model of matrix contraction to show that interfering with TSP1/TGFß binding and knockdown of TSP1 expression suppressed the contractile ability of normal and scleroderma fibroblasts basally and in response to TGFß. Previously, we have shown that ras/mitogen-activated protein kinase kinase (MEK)/extracellular signal-regulated kinase (ERK) mediates matrix contraction basally and in response to TGFß. RESULTS: During mechanical stimulation in the FPCL system, using a multistation tensioning-culture force monitor (mst-CFM), TSP1 expression and p-ERK activation in fibroblasts are enhanced. Inhibiting TSP1 activity reduced the elevated activation of MEK/ERK and expression of key fibrogenic proteins. TSP1 also blocked platelet-derived growth factor (PDGF)-induced contractile activity and MEK/ERK activation. CONCLUSIONS: TSP1 is a key mediator of matrix contraction of normal and systemic sclerosis fibroblasts, via MEK/ERK.

Inducible lineage-specific deletion of TbetaRII in fibroblasts defines a pivotal regulatory role during adult skin wound healing.

Previous attempts to delete type II TGFbeta receptor (TbetaRII) in fibroblasts have precluded examination of adult mice due to early mortality. We have selectively deleted TbetaRII postnatally in differentiated connective tissue fibroblasts using an inducible Cre-Lox strategy. Tamoxifen-dependent Cre recombinase linked to a fibroblast-specific regulatory sequence from the proalpha2(I)collagen gene permitted deletion of floxed TbetaRII alleles. After postnatal deletion of TbetaRII in fibroblasts, healing of excisional skin wounds in adults showed markedly attenuated dermal scar formation, defective wound contraction and enhanced epidermal proliferation. These findings support a pivotal role for transforming growth factor beta (TGFbeta) signalling in fibroblasts in regulating normal skin wound healing. Explanted dermal fibroblasts from TbetaRII-null-fib mice showed impaired migration and did not generate normal contractile biomechanical forces in fixed collagen gels nor develop alpha-smooth muscle antigen-rich stress fibers in response to TGFbeta1. Surprisingly, some TGFbeta-regulated proteins, including connective tissue growth factor (CTGF), were basally upregulated in TbetaRII-null fibroblasts and this was dependent on extracellular signal-regulated kinase 1/2 activity in these cells. This suggests that other intracellular pathways regulating CTGF expression may partially compensate for disruption of TGFbeta signalling in fibroblasts. Together, our data confirm that expression of TbetaRII in differentiated dermal fibroblasts is essential for normal wound healing and demonstrate a critical role in the development and function of myofibroblasts.

Rac inhibition reverses the phenotype of fibrotic fibroblasts.

BACKGROUND: Fibrosis, the excessive deposition of scar tissue by fibroblasts, is one of the largest groups of diseases for which there is no therapy. Fibroblasts from lesional areas of scleroderma patients possess elevated abilities to contract matrix and produce alpha-smooth muscle actin (alpha-SMA), type I collagen and CCN2 (connective tissue growth factor, CTGF). The basis for this phenomenon is poorly understood, and is a necessary prerequisite for developing novel, rational anti-fibrotic strategies. METHODS AND FINDINGS: Compared to healthy skin fibroblasts, dermal fibroblasts cultured from lesional areas of scleroderma (SSc) patients possess elevated Rac activity. NSC23766, a Rac inhibitor, suppressed the persistent fibrotic phenotype of lesional SSc fibroblasts. NSC23766 caused a decrease in migration on and contraction of matrix, and alpha-SMA, type I collagen and CCN2 mRNA and protein expression. SSc fibroblasts possessed elevated Akt phosphorylation, which was also blocked by NSC23766. Overexpression of rac1 in normal fibroblasts induced matrix contraction and alpha-SMA, type I collagen and CCN2 mRNA and protein expression. Rac1 activity was blocked by PI3kinase/Akt inhibition. Basal fibroblast activity was not affected by NSC23766. CONCLUSION: Rac inhibition may be considered as a novel treatment for the fibrosis observed in SSc.

Requirement of transforming growth factor beta-activated kinase 1 for transforming growth factor beta-induced alpha-smooth muscle actin expression and extracellular matrix contraction in fibroblasts.

OBJECTIVE: Fibrosis is believed to occur through normal tissue remodeling failing to terminate. Tissue repair intimately involves the ability of fibroblasts to contract extracellular matrix (ECM), and enhanced ECM contraction is a hallmark of fibrotic cells in various conditions, including scleroderma. Some fibrogenic transcriptional responses to transforming growth factor beta (TGFbeta), including alpha-smooth muscle actin (alpha-SMA) expression and ECM contraction, require focal adhesion kinase/Src (FAK/Src). The present study was undertaken to assess whether TGFbeta-activated kinase 1 (TAK1) acts downstream of FAK/Src to mediate fibrogenic responses in fibroblasts. METHODS: We used microarray, real-time polymerase chain reaction, Western blot, and collagen gel contraction assays to assess the ability of wild-type and TAK1-knockout fibroblasts to respond to TGFbeta1. RESULTS: The ability of TGF to induce TAK1 was blocked by the FAK/Src inhibitor PP2. JNK phosphorylation in response to TGFbeta1 was impaired in the absence of TAK1. TGFbeta could not induce matrix contraction or expression of a group of fibrotic genes, including alpha-SMA, in the absence of TAK1. CONCLUSION: These results suggest that TAK1 operates downstream of FAK/Src in mediating fibrogenic responses and that targeting of TAK1 may be a viable antifibrotic strategy in the treatment of certain disorders, including scleroderma.