Sustained AWT1 expression by Dupuytren’s disease myofibroblasts promotes a proinflammatory milieu

Abstract

Palmar fibromatosis, also known as Dupuytren’s disease (DD), is a common and heritable fibrosis of the hand. It is characterized by the formation of myofibroblastic nodules that can progress to palmar-digital contractures and permanent loss of dexterity. The presence of inflammatory cell infiltrate within these nodules has been interpreted to suggest a pathogenesis mediated by a proinflammatory microenvironment. However, the molecular mechanisms driving the formation of pro-fibrotic microenvironments in this and other fibroses remain unclear. To gain insights into this process, we have assessed the contributions of an alternatively spliced, multi-functional transcription factor, Wilms Tumor 1 (WT1), previously shown to be upregulated in primary myofibroblasts derived from DD tissues. Proinflammatory cytokine stimuli of DD myofibroblasts enhanced the expression of several distinct WT1 variants, the most sustained being a 5′ truncated version of WT1, alternative WT1 (AWT1). Constitutive adenoviral expression of AWT1 in myofibroblasts derived from phenotypically non-fibrotic palmar fascia significantly induced the expression and secretion of proinflammatory cytokines, including some with potential as novel therapeutic targets. In summary, these data implicate roles for sustained AWT1 expression in DD as a transcriptional driver of a proinflammatory fascial milieu.

Introduction

The innate immune system is the widely accepted “front-line” response to tissue damage, and abnormally excessive and/or prolonged activation of this system can result in a disease-promoting microenvironment (Mata et al. 2021). The parallel immunoregulatory roles played by stromal cells in inflammation-associated diseases have been historically overlooked and underappreciated despite ample evidence of complex interactions between immune system cells and a “reactive” stroma (Crowley et al. 2018). One aspect of these interactions that remains poorly understood is how a proinflammatory milieu modifies gene transcription in stromal cells to induce the maintenance of this disease-promoting microenvironment.

Dupuytren’s disease (DD) is a common, heritable fibroproliferative and myofibroblastic fibrosis that affects the palmar fascia and superficial palmar tissues of the hand and digits (Prosser and Conolly 1996; Rajesh et al. 2000; Anwar et al. 2007; Mella et al. 2018). Like many fibroses (Borthwick et al. 2012), DD development has been causally linked to the presence of a local proinflammatory milieu (Verjee et al. 2013; Bianchi et al. 2015). This milieu is hypothesized to be a primary initiator of nodular fibrosis and, in a sub-set of patients, progression to permanent and debilitating palmar-digital contractures.

We have demonstrated that palmar fascia myofibroblasts (PFMs) derived from patients with DD express and secrete a unique and complex proinflammatory cytokine profile when “activated” by co-culture with immortalized immune cells in a biomimetic of a proinflammatory fascial microenvironment (Gonga-Cavé et al. 2021). While DD PFMs are clearly hyperreactive to the presence of immune cells and the cytokines they secrete, little is known about the transcription factors that mediate their responses.

In this study, we have expanded on our previous findings of increased WT1 gene expression and Wilms Tumor 1 (WT1) immunoreactivity in DD myofibroblasts and surgically resected tissues (Crawford et al. 2015). The WT1 gene encodes multiple gene transcripts derived from different transcriptional start sites and alternative mRNA splicing, resulting in the translation of an array of proteins that function, amongst other roles, as transcription factors in cancers (Langerak et al. 1995; Kennedy et al. 1996; Dallosso et al. 2007; Amin et al. 2011). Beside DD, increased WT1 expression levels have been reported in other non-malignant, inflammation-associated fibroses, such as idiopathic pulmonary fibrosis, cardiac fibrosis and hepatic cirrhosis (Berasain et al. 2003; Braitsch et al. 2013; Karki et al. 2015; Sontake et al. 2015). Little is known about the identities or roles of the specific WT1 variants expressed in these diseases.

Our findings indicate that the most consistently expressed variant of WT1 in DD PFMs is a 5′ truncated version of “canonical” WT1, known as alternative WT1 (AWT1). We demonstrate that proinflammatory cytokine stimuli transiently induced the expression of AWT1 and other WT1 mRNAs in PFMs derived from non-fibrotic palmar fascia, and that transduction of these cells with an adenoviral vector expressing AWT1 resulted in a significant increase in the expression and secretion of proinflammatory cytokines. Our findings are consistent with the hypothesis that sustained AWT1 expression in DD PFMs is a transcriptional driver of a proinflammatory fascial milieu.

Materials and methods

Derivation and maintenance of primary palmar fascia myofibroblasts (PFMs)

Palmar fascia tissue samples were surgically resected from patients with Dupuytren’s Disease (DD) and from patients undergoing hand surgery for unrelated conditions. Primary palmar fascia myofibroblasts (PFMs) were derived from visibly fibrotic palmar fascia tissue (DD PFMs) and from visibly non-fibrotic palmar fascia tissue in an adjacent digit of the same patients exposed during hand surgery (PF PFMs) as syngeneic controls. Normal palmar fascia fibroblasts (CT PFMs) were derived from patients with no history of DD undergoing hand surgery for unrelated conditions, typically carpal tunnel release surgeries, and were utilized as allogeneic normal controls. Surgical resections to obtain palmar fascia tissues for cell derivation were performed by surgeons of the Roth McFarlane Hand and Upper Limb Clinic at St. Joseph’s Hospital in London, Ontario, in accordance with institutional ethics guidelines and approval (HSREB 104888). Cells were routinely cultured in Dulbecco Modified Eagle Medium (DMEM) (Gibco) supplemented with 8% fetal bovine serum (FBS) (Gibco), L-glutamine and antibiotic–antimycotic at 37 °C and under 5% CO₂. All experiments were performed on cells that had been consecutively passaged in vitro less than 7 times to minimize any in vitro culture-induced selection for cells with non-representative gene expression and/or phenotype.

RNA extraction and cDNA synthesis

Total RNA was isolated from PFMs in TRIzol (Thermofisher Scientific) using Direct-zol RNA extraction kit (Zymo Research, Cat# R2072) according to manufacturer’s protocol. RNA isolates were DNase treated for 25 min to remove genomic or viral DNA contaminants and were assessed for quantity and quality (A_260/280 and A_260/230) using a DeNovix DS-11 Spectrophotometer. High quality RNA (2 μg) was reversed transcribed into first strand cDNA using the High-Capacity cDNA Archive Kit (Applied Biosystems) according to manufacturer’s instructions using a single thermocycle of 25 °C for 10 min, 37 °C for 2 h, 85 °C for 5 min, 4 °C until sample retrieval. All cDNA samples were stored at − 20 °C.

Qualitative PCR

First strand cDNA (100 ng) was added into PCR master mix containing 10 × PCR buffer (5 μl), 50 mM MgCl₂ (1.5 μl), 25 mM dNTP (0.5 μl), forward primer (1 μl), reverse primer (1 μl), Platinum Taq Polymerase (0.4 μl) and ddH₂O to a final volume of 50 μl. Forward and reverse primers were designed in house (Appendix 1), synthesized by Sigma Aldrich, and reconstituted to a concentration of 10 μM in ddH₂O. Forward and reverse primer pairs were chosen to cross exon-exon boundaries to avoid amplification of genomic DNA. Touchdown PCR was carried out with initial denaturation at 95 °C for 5 min followed by 35 cycles of 95 °C for 30 s, annealing of 68 °C for 25 s and extension of 72 °C for ≤ 1 min. Final extension was carried out at 72 °C for 10 min and held at 4 °C or frozen at − 20 °C until needed. Amplified products were separated on 10% polyacrylamide gels at 90 V for ~ 1.5 h, stained with ethidium bromide and imaged using a gel imaging dock (Bio-Rad). Alternatively, amplified products were separated in 1% agarose at 90 V for ~ 1.5 h, stained with ethidium bromide and bands were excised for cDNA extraction using QIAquick Gel Extraction Kit (Qiagen, Cat# 28704). Amplified DNAs samples were sent to the London Regional Genomics Center at Robarts Research Institute for DNA sequencing.

5′ Rapid amplification of cDNA ends (5′RACE)

DD PFMs were cultured as FPCLs and stimulated with TNF, IFN-γ and IL-1β (0.5 ng/ml) for 24 h. RNA was extracted from DD PFMs and 5 μg of total RNA was used for 5′RACE in accordance with manufacturer’s protocol (ThermoFisher Scientific, Cat# 18374058). In brief, mRNA was copied into cDNA using GSP1 (Appendix 2) and superscript™ II reverse transcriptase and subsequently RNase treated to remove mRNA. TdT (terminal deoxynucleotidyl transferase) tailing of cDNA was then performed to create an abridged anchor primer binding site on the 3′-end of the cDNA for downstream amplification using abridged anchor primer (provided in kit) and GSP2 (Appendix 2). Amplified products from 5′RACE experiments were reamplified using nested primers to identify amplified WT1 transcripts by sequencing (Appendix 1). Touchdown PCR was carried out with an initial denaturation at 95 °C for 5 min followed by 35 cycles of 95 °C for 30 s, annealing of 68 °C for 25 s and extension of 72 °C for 1 min.

Quantitative PCR

One microlitre aliquots of first strand cDNA were assessed in triplicate for qPCR analyses using Taqman Primers (Thermofisher Scientific, Appendix 2) and Fast Advanced Taqman Master Mix (ThermoFisher Scientific) at a total volume of 10 μl. qPCR reaction was set for initial denaturation at 95 °C for 20 s followed by 40 cycles of denaturation at 95 °C for 1 s and annealing/extension at 60 °C for 20 s using Quantstudio 5 Thermocycler (ThermoFisher Scientific). The expression levels of target genes were calculated using the ∆∆Ct method and normalized to the housekeeping gene RPLP0. All qPCR analyses were performed three times in triplicate (N = 3, n = 3) at minimum.

Enzyme-linked ImmunoSorbent Assay (ELISA) analyses

Protein Isolation and Quantification by Enzyme-Linked Immunosorbent Assay (ELISA) Fibroblasts were cultured in T75 flasks until 70% confluence. The cells were detached with trypsin (Gibco) for 5 min at 37 °C, and the detached cells were centrifuged at 700 × g. The cell pellets were isolated and resuspended in supplemented (0.1 M NaF, 10 mM PMSF and 10 mM Na₃VO₄) RIPA buffer (Teknova), aspirated through a 27.5G needle to rupture the cells, and then centrifuged (12,000×g) for 2 min. Supernatants containing protein lysate were quantified using the Bicinchoninic Acid (BCA) protein assay kit (Pierce BCA Protein Assay Kit, ThermoFisher Scientific, Cat# 23227) according to manufacturer’s protocol. 1 μg of total protein lysate was used for WT1 quantification by human Wilms tumor protein ELISA kit (Mybiosource, Cat# MBS761090) according to manufacturer’s protocol.

Adenoviral amplification and transduction

HEK293 cells were cultured in DMEM supplemented with 8% FBS, L-glutamine and antibiotic–antimycotic until 60–70% confluency. 1 mL viral stock solutions of containing either cDNA encoding AWT1 (Ex5+, KTS−) (Vigene Biosciences VH801783) or cDNA encoding green fluorescent protein (GFP) (Vigene Biosciences CV10001) was added to infect cells. Infected HEK293 cells were cultured until 95% of the cells detached from the dishes, collected and subjected to 3 freeze/thaw cycles in − 80 °C freezer and 37 °C water bath. Cellular debris was removed by centrifugation for 10 min at 700×g. Viral supernatant was collected, aliquoted and stored at − 80 °C. CT and PF fibroblasts were transduced with adenovirus expressing either AWT1 (AdAWT1, Ex5+/KTS−) or GFP (AdGFP as viral control) at 1:3000 in DMEM in 30 mm culture dishes for 24 h. Exogenous expression of AWT1 was maintained for 72 h prior to RNA extraction for qPCR or RNA sequencing analyses.

Western immunoblotting

20 µg of total protein lysates in RIPA buffer were derived from the OVCAR3 ovarian cancer cell line (positive control for WT1 protein) and from CT fibroblasts transduced with either AdAWT1 or AdGFP and loaded onto Mini-PROTEAN® TGX™ Precast Gels (Bio-Rad, 4–15% SDS) for electrophoresis at 100 V for ~ 1.5 h. The polyacrylamide gels were transferred to an iBlot™ PVDF Transfer Stack (ThermoFisher Scientific, Cat# IB401002) and transferred to PVDF using an iBlot™ Dry Blotting System (ThermoFisher Scientific). After protein transfer, the PVDF membrane was incubated with blocking solution consisting of 1 × Tris-buffered saline supplemented with 0.1% TWEEN-20 (TBS-T) and 5% w/v bovine serum albumin (BSA) for 45 min at room temperature. Primary WT1 antibody (D6M6S Rabbit mAb, Cell Signalling, Cat# 13580) was diluted at 1:1000 in TBS-T (5% w/v BSA) and incubated with PVDF membranes at 4 °C overnight. PVDF membranes were washed in TBS-T 3 times for 5 min each and incubated with anti-rabbit-HRP conjugated secondary anti-body for 1.5 h at room temperature, washed in TBS-T 3 times for 5 min each and incubated briefly in peroxidase: luminol solution at a 1:1 ratio of. Chemiluminescent images were captured using a Gel Doc XR + System (Bio-Rad).

RNA sequencing

Total RNA samples were further assessed for quantity and quality using Agilent 2100 Bioanalyzer. 1 μg of high-quality RNA from each sample was submitted to London Genomics Sequencing Center at Robarts for RNA sequencing (RNA-seq) and library preparation using Illumina NextSeq Mid Output Kit (Vazyme VAHTS Total RNA-seq (H/M/R) Library Prep Kit for Illumina®). The raw data generated by library preparation were then analyzed using Partek Flow® data analysis software. Raw data were aligned to reference human genome (hg38) using Spliced Transcripts Alignment to a Reference (STAR) tool, normalized using Trimmed Mean of M values (TMM) and differential analysis was performed using Gene Specific Analysis (GSA). Inclusion criteria for downstream analysis were p < 0.05 and a fold change in gene expression of ≥ 1.5 or ≤ − 1.5. Heat-map, gene set enrichment, and pathway enrichment analyses were performed and generated using Partek Flow® software.

Multiplex protein analyses

Secreted cytokines were measured using multiplexed immunoassay kits (ProcartaPlex Multiplex Immunoassay custom 19plex, Thermo Fisher Scientific, Waltham, MA, Cat# PPX-19-MXAAA6K) according to the manufacturers’ instructions (R&D Systems, Minneapolis, MN, Cat#). A Bio-Plex 200 readout system was used (Bio-Rad, Mississauga, ON, Cat# 171000201), which utilizes Luminex® xMAP fluorescent bead-based technology (Luminex Corporation, Austin, TX). Cytokine levels (pg/μg of total protein) were calculated from standard curves using Bio-Plex Manager software (v. 6.1, Bio-Rad, Mississauga, ON, Cat# 10022815). Cell cultures were incubated for 72 h to provide sufficient time for translational impacts on cytokine secretion. After 72 h, the conditioned media was removed, treated with anti-proteolytic solution and stored at − 80 °C.

Illustrations and graphical information

All illustrations were created in BioRender and all graphical information was created in GraphPad Prism.

Results

Our original report of increased WT1 mRNA levels and WT1 immunoreactivity in DD PFMs and tissues (Crawford et al. 2015) did not identify the specific WT1 transcript variants involved. As shown in Fig. 1A (A and B), the multiple WT1 mRNA transcript variants evident in various cancers were reported to be derived from at least three distinct transcriptional start sites and two sites of alternative splicing. Relative to the transcriptional start site of “canonical” WT1 transcripts, start site variants include 5′ “extended” (ExtWT1) transcripts that initiate at a CUG upstream of the canonical AUG start site, “alternative” (AWT1) transcripts that initiate at an AUG encoded within intron 1 of the canonical sequence, and truncated transcript variants (TrWT1) that initiate from an AUG encoded within intron 5 of the canonical sequence. As our original WT1 transcript analyses utilized primers that amplified WT1 exon sequences within canonical WT1, ExtWT1 and AWT1 transcripts, they did not distinguish between them, nor did they detect any transcripts with start sites downstream of exon 4, such as TrWT1 transcripts. Therefore, we revisited our original analyses using transcript-specific primers derived from previous reports (Bruening and Pelletier 1996; Dallosso et al. 2004; Hossain et al. 2006; Hastie 2017; Lee et al. 2017). PF and CT PFMs were shown to express TrWT1 mRNA transcripts, but no other WT1 mRNA variants were detected under routine cell culture conditions. In contrast, DD PFMs consistently expressed AWT1 transcripts under routine cell culture conditions (Fig. 1B). To determine if the expression of different WT1 variants in DD, PF and CT PFMs impacted the total amount of WT1 proteins they expressed, we assessed total WT1 levels in cells lysates by ELISA. As shown in Fig. 1C, WT1 proteins could be detected in all PFM lysates assessed and a modest but statistically significant increase in WT1 levels was evident in DD PFM lysates relative to CT PFM (normal control) lysates.

Fig. 1.

WT1 variants and their expression in palmar fascia myofibroblasts (PFMs). A The WT1 gene is located on the p arm of chromosome 11 and contains 10 major exons. (A) Sites of alternative mRNA splicing include exon 5 (Ex5) and the end of exon 9, encoding a KTS motif. (B) Alternative transcriptional start sites relative to the start site for canonical WT1 mRNA transcripts include intron 1 (alternative WT1), intron 5 (truncated WT1) and the 5′ untranslated region (extended WT1). Adapted from (Kramarzova et al. 2012). B Transcript-specific PCR analyses illustrating (A) the absence of canonical WT1 mRNA expression DD PFM lysates (DD samples), (B) the expression of AWT1 mRNAs in DD samples but not syngeneic (PF) or allogeneic (CT) sample controls, and (C) the expression of TrWT1 mRNA in PF and CT samples. OVCAR3 cell lysates were used as positive controls for WT1 expression. The transcript specific primer sequences are shown in Appendix 1. C ELISA analyses of total WT1 protein in lysates derived from DD, PF and CT PFMs. WT1 protein levels were normalized to CT PFM (allogeneic normal) controls. N = 3/group, **p < 0.01 by ANOVA. Standard errors (SE) for DD samples were too small to be illustrated. D Transcript-specific PCR analyses for the alternatively spliced WT1 mRNAs illustrated in A. From left: PCR analyses to detect amplicons from canonical WT1 exons 6–7, A (Ex5−/KTS−), B (Ex5+/KTS−), C (Ex5−/KTS+), D: Ex5+/KTS− in lysates from DD PFMs, syngeneic (PF PFMs) controls and OVCAR3 cells (positive control). Primer sequences were adapted from (Kramarzova et al. 2012)

In addition to different transcriptional start sites, WT1 mRNAs can be modified by alternative splicing. Known splice variants of WT1 transcripts include the excision of exon 5 and/or of a region at the end of exon 9, encoding a Lysine-Threonine-Serine (KTS) motif. Using PCR primers designed to distinguish these variants in leukemias (Kramarzova et al. 2012), we demonstrated that DD PFMs, like the ovarian cancer OVCAR3 cell line and positive control, expressed all four splice variants (Ex5−/KTS−, Ex5+/KTS−, Ex5−/KTS+, and Ex5+/KTS−) under routine cell culture conditions (Fig. 1D).

Since DD PFMs are derived from a proinflammatory microenvironment, we treated DD, PF and CT PFMs with a “proinflammatory cytomix” (TNF, IL-1β, IFN-γ, each 0.5 ng/ml) for 24 h to determine any impacts of cytokine stimuli on WT1 transcript expression. As shown in Fig. 2A, proinflammatory cytokine stimuli significantly induced WT1 mRNA transcript levels in all three cell groups by ~ tenfold. Unlike unstimulated cells, canonical WT1 and AWT1 mRNA transcripts were detectible in cytokine-stimulated PF and CT PFMs, as were all four of the splice variants previously only evident in DD PFMs (Ex5−/KTS−, Ex5+/KTS−, Ex5−/KTS+, and Ex5+/KTS−) (Fig. 2B). To further enrich for rare WT1 transcripts in cytokine-stimulated DD PFMs, 5′ Rapid Amplification of cDNA ends (5′RACE) and sequencing analyses were performed using nested WT1-specific primers (Appendix 1). As shown in Fig. 2C, PCR amplicons consistent with canonical WT1, AWT1 and ExtWT1 mRNAs could be detected in cytokine-stimulated DD PFMs.

Fig. 2.

WT1 variant expression is modified by proinflammatory cytokine stimuli of PFMs. A Taqman qPCR analyses of WT1 mRNA expression in lysates derived from DD PFMs (DD cells), PF PFMs (PF cells) and CT PFMs (CT cells), with (-Cyt) or without “proinflammatory cytomix” (TNF, IL-1β, IFN-γ, each 0.5 ng/ml) stimulation. The Taqman primers amplified a region encoding exons 3 and 4 (primers shown as blue arrows, probe as a yellow star) relative to canonical WT1 gene sequence. N = 3, n = 3, *p < 0.05, **p < 0.01 by ANOVA. B Transcript-specific PCR analyses illustrating (A) the expression of canonical WT1 mRNA transcripts in PF PFM and CT PFM lysates (PF samples and CT samples respectively), (B) the expression of AWT1 mRNAs in PF samples and CT samples respectively, and (C) the expression of canonical WT1 mRNA transcripts in DD PFM lysates (DD samples), in all cases after proinflammatory cytomix stimulation. The transcript specific primer sequences are shown in Appendix 1. C PCR reamplification of 5′RACE enriched WT1 transcripts with transcript-specific primers. From left, molecular weight markers (MW), A: Nested forward and reverse primers targeting exons 1 and 3 of canonical WT1 sequence (234 bp amplicon), B: Nested forward and reverse primers targeting intron 1 and exon 3 to detect AWT1 transcripts (219 bp amplicon), C: Nested forward and reverse primers targeting a region encoding the 5′ untranslated region of canonical WT1 and exon 1 to detect ExtWT1 transcript amplification (152 bp amplicon), D: Sample C omitting reverse transcriptase as a control for DNA contamination. The locations of transcript-specific primers are illustrated as red arrows and the Taqman primer set used to detect WT1 mRNA transcripts in Crawford et al. (2015) is shown as blue arrows with a yellow star indicating Taqman probe location

Having established that, unlike unstimulated PF and CT PFM controls, DD PFMs exhibit sustained expression of AWT1, we chose to assess the consequences of constitutive AWT1 expression in PF and CT PFM controls to mimic DD PFMs. An adenoviral vector encoding the Ex5+/KTS− variant of AWT1 was chosen to transduce three genetically unrelated PF and three genetically unrelated CT PFM cultures. Gene expression and western immunoblotting analyses confirmed the sustained expression and translation of AWT1 in all six AdAWT1-transduced cultures, unlike controls transduced with the adenoviral vector expressing GFP (example shown in Fig. 3A). To assess the consequences of AWT1 expression, we assessed RNA expression in all 12 groups (3 × PF AdAWT1, 3 × PF AdGFP, 3 × CT AdAWT1 and 3 × CT AdGFP) by RNA sequencing (RNA-seq). RNA-seq analysis identified 3017 gene transcripts that were differentially regulated in PF and CT fibroblasts expressing AWT1 relative to vector expressing fibroblasts (pooled, p < 0.05). Of the 3017 significantly up-or down-regulated gene transcripts, 961 differentially regulated gene transcripts (~ 32%) were identified as unique to AWT1 expressing PF fibroblasts relative to AWT1-expressing CT fibroblasts. Heatmap and clustering analyses of the RNA-seq data confirmed the presence of gene transcripts that were differentially expressed in PF (WT1 versus GFP) and in CT (WT1 versus GFP) groups and that each of the four groups had similar transcriptional profiles that were consistent within groups and distinct from the other three groups (Fig. 3B).

Fig. 3.

Adenoviral expression of AWT1 in PF and CT PFMs. A (A) qPCR analyses of WT1 gene expression in CT (N = 3) and PF (N = 3) PFMs transduced with (from left) tissue culture media as a vehicle control, (Veh), adenoviral vector encoding green fluorescent protein (AdGFP) or adenoviral vector encoding AWT1 (AdWT1). The Taqman primer set used to detect WT1 mRNA transcript expression are shown as blue arrows with a yellow star indicating the Taqman probe, **p < 0.01 by ANOVA. (B) Western immunoblotting with a WT1 antibody (D6M6S, Cell Signaling Technology) against total protein lysates of CT PFMs transduced with adenovirus encoding GFP (CT-AdGFP), OVCAR3 total protein lysate (positive control) and CT PFMs transduced with adenovirus encoding AWT1 (CT-AdWT1). Based on AWT1 cDNA sequence data, the predicted MW of AWT1 is 34,447 Daltons. B Heatmap and Clustering analysis of PF and CT PFMs transduced with adenoviral vector encoding GFP (PF-GFP and CT-GFP, each N = 3) or AWT1 (PF-AWT1 and CT-AWT1, each N = 3) respectively. Significant differential gene expression (p < 0.05) indicated by green (negative fold-change ≤ 1.6) or red (positive fold-change ≥ 1.6) in AWT1 expressing fibroblasts relative to GFP-expressing controls

Gene set enrichment and pathway enrichment analyses were performed to identify overrepresented biological processes or pathways. The highest enrichment score from the pathway enrichment analysis was for cytokine–cytokine receptor interaction (Table 1). As shown in Table 2, detailed analyses of the cytokine–cytokine receptor interaction subgroup revealed that ~ 69% of the modified gene transcripts were categorized as contributing to proinflammatory signaling pathways.

Table 1.

A Pathway enrichment analysis identifying overrepresented biological pathways in AWT1 expressing PF and CT PFMs (pooled) relative to PF/CT PFM controls. Inclusion criteria were p-values < 0.05 and fold change in gene expression of > 1.5 or < −1.5. Enrichment scores indicate the negative natural logarithm of the p-value. As shown, the highest enrichment score was associated with cytokine-cytokine receptor interactions

Path term	Enrichment score
Cytokine–cytokine receptor interaction	21.47
MAPK signaling pathway	10.3
ABC transporters	7.35
Bile secretion	7.08
Gastric cancer	6.84
Neuroactive ligand-receptor interaction	6.65
Pathways in cancer	6.62
Transcriptional misregulation in cancer	6.43
Viral protein interaction with cytokine and cytokine receptor	6.34
MicroRNAs in cancer	6.29
Inflammatory bowel disease (IBD)	6.07
African trypanosomiasis	5.78
Systemic lupus erythematosus	5.74
Proximal tubule bicarbonate reclamation	5.69
Type I diabetes mellitus	5.68
Cell adhesion molecules (CAMs)	5.66
Toll-like receptor signaling pathway	5.59
Influenza A	5.34
Measles	5.31
NF -kappa B signaling pathway	5.21
RIG-I-like receptor signaling pathway	5.04
B cell receptor signaling pathway	4.85
Leishmaniasis	4.71
Hematopoietic cell lineage	4.58
Th17 cell differentiation	4.5
Rheumatoid arthritis	4.47
Aldosterone-regulated sodium reabsorption	4.42
Graft-versus-host disease	4.4
Carbohydrate digestion and absorption	4.23
Ras signaling pathway	4.22
Epstein-Barr virus infection	4.21
Fat digestion and absorption	4.19
Bladder cancer	4.13
JAK-STAT signaling pathway	3.95
Basal cell carcinoma	3.81
Protein digestion and absorption	3.78
cAMP signaling pathway	3.72
PI3K-Akt signaling pathway	3.71
Staphylococcus aureus infection	3.68
Allograft rejection	3.56
Arachidonic acid metabolism	3.54
IL-17 signaling pathway	3.53
ErbB signaling pathway	3.49
Apoptosis—multiple species	3.46
Intestinal immune network for IgA production	3.46
Mineral absorption	3.41
Hepatitis B	3.35
Steroid biosynthesis	3.32
p53 signaling pathway	3.28
EGFR tyrosine kinase inhibitor resistance	3.24

Table 2.

The category with the highest enrichment score in Table 1, cytokine cytokine receptor interaction, was reviewed at the level of individual gene expression and, where possible, allocated to one of two general categories, “pro-inflammatory” or “anti-inflammatory”, based on literature review. As shown, ~ 69% of the total number of gene transcripts could be readily categorized as classically pro- or antiinflammatory, with ~ 82% of these (29/35) categorized as pro-inflammatory

Gene symbol	p value	Fold change
Pro-inflammatory genes
CCL1	4.34E−02	3.59E+00
CCL4	4.93E−02	2.56E+00
CCL11	3.63E−02	5.62E+00
CCL26	1.04E−02	− 1.95E+00
CCL20	3.24E−03	1.75E+00
CXCL6	1.16E−03	5.38E+00
CXCL8	3.89E−02	2.58E+00
CXCL9	1.96E−03	8.75E+00
CXCL10	1.77E−08	1.27E+02
CXCL11	8.01E−04	4.27E+00
CXCL14	5.13E−05	5.93E+00
CXCR4	6.66E−04	6.17E+00
TSLP	1.02E−02	1.90E+00
IL12A	4.13E−03	− 1.89E+00
IL12RB2	5.25E−05	5.89E+00
IL23A	2.57E+00	2.41E+00
IL19	1.13E−02	2.36E+00
IL20	1.81E−02	6.31E+00
IL1A	4.20E−03	2.37E+00
IL1B	7.66E−03	3.09E+00
FASLG	3.96E−02	2.33E+00
TNFSF15	1.20E−02	2.08E+00
TNFRSF21	3.60E−02	1.60E+00
TNFSFR8	5.45E−04	3.85E+00
TNFSF4	1.11E−04	2.88E+00
RELT	2.08E−02	1.53E+00
GDF15	1.66E−03	1.75E+00
ACVR1	2.16E−04	− 1.81E+00
IL31RA	4.90E−02	1.82E+00
Anti-inflammatory genes
CLCF1	8.86E−04	− 2.02E+00
CNTFR	2.76E−02	2.73E+00
TGFB1	2.77E−04	− 1.66E+00
TGFBR1	7.32E−03	− 1.52E+00
BMP4	7.17E−03	− 1.57E+00
GDF6	1.81E−03	3.86E+00

To independently confirm the changes in gene expression identified by RNA-seq, Taqman qPCR analyses were performed for six genes with upregulated expression in AWT1 expressing PF and CT PFMs (ICAM1, CXCL14, CXCR4, CXCL10, TNFSF4, IL1B, primer details in Appendix 2) and which had been previously shown to be associated with DD or other fibroses (Satish et al. 2012; Xia et al. 2014; Qiu et al. 2021; Miądlikowska et al. 2021; Ding et al. 2022). The expression levels of all these genes were consistently and significantly upregulated in AWT1 expressing PF fibroblasts, and many were either significantly upregulated or trended toward an increase in expression levels in AdAWT1 expressing CT PFMs (Fig. 4A).

Fig. 4.

Cytokine gene expression levels in PF and CT PFMs expressing AWT1. A qPCR analysis of ICAM1, CXCL14, CXCR4, CXCL10, TNFSF4 and IL-1B expression levels in PF and CT PFMs transduced with AdAWT1 or AdGFP (N = 3/group, *p < 0.05, **p < 0.01 and ***p < 0.001 by t-test relative to corresponding AdGFP controls). B TNFSF4 mRNA expression assessed using Taqman primers in lysates from DD, PF and CT PFMs (each N = 3) under routine culture conditions, **p < 0.001 by ANOVA

Multiplex protein analyses and ELISAs were performed to confirm the translation and secretion of a representative subset of the cytokines with increased expression in AdAWT1 PFMs. As shown in Fig. 5, constitutive AdAWT1 expression in PF and CT PFMs was associated with either significant increases, or trends towards increases, in IL-1α, IL-1ß, IL-6, IL-8, TNF and IP-10 secretion into cell culture media relative to the levels secreted by AdGFP-expressing controls.

Fig. 5.

Cytokine secretion by PF and CT PFMs expressing AWT1. The levels of Interleukin-6 (IL-6), IL-8, IL-1α, IL-1ß, Interferon-inducible Protein-10 (IP-10) and Tumor Necrosis Factor (TNF) secreted by PF and CT PFMs transduced with AdAWT1 or AdGFP (N = 3/group) into serum-free cell culture media over 72 h. were measured using multiplexed immunoassay kits (ProcartaPlex Multiplex Immunoassay custom 19plex, Thermo Fisher Scientific, Waltham, MA). Significant changes in cytokine levels were detected either by One Way ANOVA – Tukey multiple comparison test or by t-test relative to corresponding AdGFP controls as indicated, *p < 0.05, **p < 0.01 and ***p < 0.001

Discussion

To our knowledge, this is the first report to link the enhanced expression of AWT1 in myofibroblasts to maintenance of a proinflammatory microenvironment. The circular nature of our findings, i.e., that inflammation induces AWT1 expression in myofibroblasts, which in turn induces secretion of inflammatory cytokines, is consistent with the hypotheses that the development of DD is causally linked to a sustained local proinflammatory milieu (Verjee et al. 2013; Bianchi et al. 2015) and that this milieu is derived from both immune cell infiltrate and from a reactive palmar fascial stroma (Fig. 6).

Fig. 6.

AWT1 expression contributes to the DD fascial milieu. Immune cell infiltrate into and/or adjacent to the palmar fascia is hypothesized to induce a reactive stroma response from PFMs, which includes enhanced expression of AWT1. AWT1 induces the transcription of cytokine-encoding genes, feeding back on the immune cell infiltrate, generating a positive feedback loop and a chronic proinflammatory cytokine-enriched microenvironment. Based on the data reported here and in Gonga-Cavé et al. (2021)), some cytokines, including latent TGFß-1, TNF, IL-6, IL-8 and IP-10, are predicted to be derived from both immune cells and PFMs, some (e.g., MCP-1, MCP-4, IL-10) are predicted to be primarily derived from local macrophages, while others (SDF-1 and OX-40L) may be primarily derived from DD PFMs. Image created with BioRender

That proinflammatory stimuli can transiently induce AWT1 expression in PFMs derived from normal palmar fascia (CT PFMs) may imply that AWT1 is normally expressed in PFMs during the proinflammatory stage of tissue repair, and that its expression is normally attenuated again during transition to the resolving stage. If this hypothesis is verified in future studies, it would identify sustained AWT1 expression as a biomarker of sustained inflammation and increased risk of fibrosis development.

AWT1 expression in myeloid leukemias and hepatocellular carcinomas is regulated by intron 1 start site methylation and the expression levels of an antisense lncRNA transcript, WT1-AS (Lv et al. 2015; McCarty and Loeb 2015). Future studies could determine if WT1-AS transcription and/or AWT1 start site methylation are also (dys)regulated by proinflammatory cytokine signaling in DD PFMs. It could be hypothesized that long term exposure to a proinflammatory microenvironment permanently alters AWT1 start site methylation and or WT1-AS expression, and that these epigenetic modifications explain the sustained expression of AWT1 in DD PFMs in routine culture in the absence of proinflammatory stimuli. AWT1 has also been reported to be monoallelically expressed and subject to genomic imprinting (Dallosso et al. 2004; Hancock et al. 2007), and inflammation-induced loss of imprinting, resulting in biallelic AWT1 expression, in DD PFMs is another intriguing possibility for investigation.

PF and CT PFMs were shown to express TrWT1 transcripts and to exhibit measurable levels of WT1 proteins in the absence of proinflammatory stimuli. These findings may indicate normal roles for “truncated” WT1 proteins in PFM homeostasis. To our knowledge, the functional properties of proteins derived from TrWT1 transcripts are currently unknown. Since the sequence encoded by TrWT1 mRNA transcripts is downstream of, and included in, the sequence encoded by AWT1 transcripts (Fig. 1A), we were unable to determine if DD PFMs express both AWT1 and TrWT1 transcripts simultaneously using qualitative PCR. Carefully designed qPCR analyses may be able to address this question in future studies.

Despite being derived from six randomly selected and genetically unrelated patients, unbiased heatmap and clustering analyses of the RNA-seq data (Fig. 3B) revealed that the three AWT1-expressing PF PFM and three AWT1-expressing CT PFM data sets clustered together into distinct groups. These findings are consistent with our previous reports indicating that, despite being derived from visibly non-fibrotic palmar fascia, PF PFMs (syngeneic controls from visibly non-fibrotic palmar fascia in patients with DD in an adjacent area and who are potentially genetically predisposed to develop DD) and CT PFMs (allogeneic controls derived from unaffected individuals) are transcriptionally and functionally distinct (Satish et al. 2008, 2012). We interpret these data to indicate that PF PFMs are in a state of “prodromal fibrosis”, a poorly characterized, third state that is distinct from normal PFMs in homeostasis and from myofibroblasts actively contributing to ongoing fibrosis. Characterization of PFMs in a state of prodromal fibrosis could be interesting from a clinical perspective, as they may secrete unique biomarkers that could identify patients primed to develop DD before disease onset.

While our findings implicate roles for AWT1 in maintenance of a proinflammatory milieu, it is unlikely to be the only transcription factor contributing to this disease process. NFκB is arguably the most extensively studied central mediator of inflammatory responses (Taniguchi and Karin 2018). The WT1 promoter includes a consensus NFκB binding site, and complex regulation of NFκB-induced WT1 transcription has been reported in podocytes (Zhang et al. 2021). SP1, c-MYB, ETS-1, AP1, and the histone acetyltransferase p300/CBP have well established roles in the transcriptional regulation of genes encoding cytokines and other inflammation-associated molecules (Rius-Pérez et al. 2020), and each have been demonstrated to cooperatively enhance WT1 expression levels in various human cell cultures (Shao et al. 2007; Anuchapreeda et al. 2019). Consistent with its complex roles in adaptive and innate immunity responses (White et al. 2019), Hippo-Yap regulation of WT1 expression appears to be complex and potentially cell-type specific, with reports of Yap-enhanced WT1 expression in mouse epicardium (Singh et al. 2016) contrasting with YAP-downregulated WT1 expression in human podocytes (Xie et al. 2019). Future studies could focus on these and many other transcriptional regulators of cytokine-encoding genes, and more specifically, of AWT1-mediated inflammatory cytokine expression in DD.

One limitation of this study, and of most studies utilizing viral vector transduction technologies, was our inability to “dial down” AWT1 expression in PF and CT PFMs to physiologically relevant levels i.e., to the expression levels evident in DD PFMs. Our efforts to achieve this by serial dilution of adenoviral vector concentration prior to transfection were unsuccessful, and we estimate that AWT1 expression levels in our transduced PF and CT PFMs was ~ 100 fold higher than the AWT1 expression levels in DD PFMs. It is probable, therefore, that some of the cytokines identified as transcriptionally regulated by AWT1 in this study may not be noticeably impacted by the lower concentrations of AWT1 in DD PFMs. It is also currently unclear how many of the cytokines identified as transcriptionally regulated by AWT1 in transduced PF or CT PFMs are directly versus indirectly regulated, where AWT1 induces the expression of an intermediate factor that then induces the expression of a cytokine-encoding gene.

One of many ways to distinguish between direct and indirect AWT1 targets would be to screen these genes for promoters containing the WT1 consensus binding site, GCGGGGGCG (Hamilton et al. 1995; Hewitt et al. 1996; Ye et al. 1996). In this context, we were intrigued by one of the novel AWT1 targets identified in this study, TNFSF4. TNFSF4 is a TNF superfamily member that contains the WT1 consensus binding site in its promoter and encodes OX40 ligand (OX40L). OX40L is a recognized therapeutic target for several inflammatory diseases including systemic sclerosis (SSC) (Elhai et al. 2016) and humanized monoclonal antibodies are currently in phase II clinical trials for allergen-induced asthma. As shown in Fig. 4A and B, AWT1 induced a significant increase in TNFSF4 expression in transduced PF and CT PFMs, and DD PFMs exhibit a significant increase in TNFSF4 expression relative to PF and CT PFMs under routine culture conditions. Additional studies could determine if serum levels of OX40L are predictive biomarkers for DD, as they are for SSC (Elhai et al. 2016) and other inflammatory diseases (Mao et al. 2020; Chen et al. 2021; Li et al. 2021; Sani et al. 2021).

In addition to enhanced TNFSF4 expression, we noted significant increases in the expression levels of CXCR4 and CXCL10, encoding Receptor for the CXC chemokine ligand 12 (CXCL12)/Stromal Cell-derived Factor I receptor (SDF-1) and Interferon-inducible protein 10 (IP-10) respectively in AWT1-expressing PFMs (Figs. 4 and 5). Both CXCR4/CXCL12/SDF-1 (Scotton and Chambers 2007) and CXCL10/IP-10 (Miądlikowska et al. 2021) have been identified as potential biomarkers of pulmonary fibrosis, however neither have been previously linked to the enhanced WT1 expression levels in this disease (Karki et al. 2015; Sontake et al. 2015). Based on our previous studies (Gonga-Cavé et al. 2021), it is likely that IP-10 is jointly derived from both DD PFMs and from immune cells in DD.

In summary, our findings implicate sustained AWT1 expression in DD PFMs as a transcriptional driver of a proinflammatory fascial milieu and we identify some novel proinflammatory cytokines regulated by AWT1 that may serve as potential screening tools and/or therapeutic targets for DD.

Appendix 1

See Table 3.

Table 3.

Custom designed primers used for detection of WT1 mRNA transcripts

Target exon	Sequence 5′–3′
Exon 1—Forward	AGCCCGCTATTCGCAATCAG
Exon 2—Forward	TTACAGCACGGTCACCTTCG
Exon 3—Reverse	TCCTCAGCAGCAAAGCCTGG
Intron 1—Forward	GAGAAGGGTTACAGCACGGTC
Intron 5—Forward	GACAGAAGGGCAGAGCAA
CUG—Forward	TACAGCAGCCAGAGCAGCAG
CUG—Reverse	GTCCCGCACGTCGGAGCCCAT
GSP1—Reverse	GTGTGTATTCTGTATT
GSP2—Reverse	CAACGCCCATCCTCTGCGGA

Table listing WT1 Primers designed in-house and verified by DNA sequencing. CUG primers were designed to recognize ExtWT1, upstream of exon 1 (CUG start site), whereas intron 1 forward primers recognize ATG start site found within intron 1 of WT1. Primers designed to recognize specific WT1 variants were also used as nested primers for 5′RACE. Gene specific primers 1/2 (GSP1/2) were designed in-house to recognize exon 7 and 6 respectively to generate 5′RACE products

Appendix 2

See Table 4.

Table 4.

Taqman primers used for qPCR analyses

Gene ID	Cat #	Amplicon length (bp)
WT1	Hs01103751_ml	72
RPLPO	Hs99999902_ml	105
ICAM1	Hs00l64932_ml	87
CXCL14	Hs01557413_ml	66
CXCR4	Hs00237052_ml	78
CXCL1O	Hs00171042_ml	98
TNFSF4	Hs00182411_ml	72
IL1B	Hs01555410_ml	91

Table listing qPCR Taqman Primers (Thermofisher Scientific), gene ID, their corresponding catalog number and amplicon length (bp). All Taqman primers used amplified gene products that crossed exon-exon boundaries