Pseudohypoxic HIF pathway activation dysregulates collagen structure-function in human lung fibrosis

  1. Christopher J Brereton
  2. Liudi Yao
  3. Elizabeth R Davies
  4. Yilu Zhou
  5. Milica Vukmirovic
  6. Joseph A Bell
  7. Siyuan Wang
  8. Robert A Ridley
  9. Lareb SN Dean
  10. Orestis G Andriotis
  11. Franco Conforti
  12. Lennart Brewitz
  13. Soran Mohammed
  14. Timothy Wallis
  15. Ali Tavassoli
  16. Rob M Ewing
  17. Aiman Alzetani
  18. Benjamin G Marshall
  19. Sophie V Fletcher
  20. Philipp J Thurner
  21. Aurelie Fabre
  22. Naftali Kaminski
  23. Luca Richeldi
  24. Atul Bhaskar
  25. Christopher J Schofield
  26. Matthew Loxham
  27. Donna E Davies
  28. Yihua Wang  Is a corresponding author
  29. Mark G Jones  Is a corresponding author
  1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, United Kingdom
  2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, United Kingdom
  3. Biological Sciences, Faculty of Environmental and Life Sciences, University of Southampton, United Kingdom
  4. Institute for Life Sciences, University of Southampton, United Kingdom
  5. Section of Pulmonary, Critical Care and Sleep Medicine, Department of Medicine, Yale University School of Medicine, United States
  6. Leslie Dan Faculty of Pharmacy, University of Toronto, Canada
  7. Institute of Lightweight Design and Structural Biomechanics, TU Wien, Austria
  8. Department of Chemistry and the Ineos Oxford Institute for Antimicrobial Research, Chemistry Research Laboratory, United Kingdom
  9. School of Chemistry, University of Southampton, United Kingdom
  10. University Hospital Southampton, United Kingdom
  11. Department of Histopathology, St. Vincent's University Hospital & UCD School of Medicine, University College Dublin, Ireland
  12. Unità Operativa Complessa di Pneumologia, Università Cattolica del Sacro Cuore, Fondazione Policlinico A. Gemelli IRCCS, Italy
  13. Faculty of Engineering and Physical Sciences, University of Southampton, United Kingdom

Abstract

Extracellular matrix (ECM) stiffening with downstream activation of mechanosensitive pathways is strongly implicated in fibrosis. We previously reported that altered collagen nanoarchitecture is a key determinant of pathogenetic ECM structure-function in human fibrosis (Jones et al., 2018). Here, through human tissue, bioinformatic and ex vivo studies we provide evidence that hypoxia-inducible factor (HIF) pathway activation is a critical pathway for this process regardless of the oxygen status (pseudohypoxia). Whilst TGFβ increased the rate of fibrillar collagen synthesis, HIF pathway activation was required to dysregulate post-translational modification of fibrillar collagen, promoting pyridinoline cross-linking, altering collagen nanostructure, and increasing tissue stiffness. In vitro, knockdown of Factor Inhibiting HIF (FIH), which modulates HIF activity, or oxidative stress caused pseudohypoxic HIF activation in the normal fibroblasts. By contrast, endogenous FIH activity was reduced in fibroblasts from patients with lung fibrosis in association with significantly increased normoxic HIF pathway activation. In human lung fibrosis tissue, HIF-mediated signalling was increased at sites of active fibrogenesis whilst subpopulations of human lung fibrosis mesenchymal cells had increases in both HIF and oxidative stress scores. Our data demonstrate that oxidative stress can drive pseudohypoxic HIF pathway activation which is a critical regulator of pathogenetic collagen structure-function in fibrosis.

Editor's evaluation

The reviewers found your manuscript of broad interest to researchers interested in lung biology, as the study builds upon the previous original work of the authors, by identifying a pathway that regulates collagen nanostructure and stiffness in lung fibrosis and demonstrating that this pathway it is independent of pathways regulating collagen synthesis. They also valued the elegant analysis you performed to validate the specificity of experimental finding, and demonstrate that HIF activation is required for the increased tissue stiffness associated with fibrosis.

https://doi.org/10.7554/eLife.69348.sa0

Introduction

We previously identified that in the lung tissue of patients with idiopathic pulmonary fibrosis (IPF) there is increased pyridinoline collagen cross-linking and altered collagen fibril nano-architecture, with individual collagen fibrils being structurally and functionally abnormal (Jones et al., 2018). This was associated with increased tissue expression of lysyl hydroxylase 2 (LH2/PLOD2, which catalyses telopeptide lysine hydroxylation to determine pyridinoline cross-linking) and the lysyl oxidase-like (LOXL) enzymes LOXL2 and LOXL3, which initiate covalent collagen cross-linking (Jones et al., 2018). This pyridinoline cross-linking, rather than any change in collagen deposition per se, determined increased IPF tissue stiffness. Inhibiting pyridinoline cross-linking normalised mechano-homeostasis and limited the self-sustaining effects of ECM on fibrosis progression. Whilst identifying the importance of altered collagen nanoarchitecture to human lung fibrosis pathogenesis, the upstream mechanisms that dysregulate collagen structure-function to promote progressive fibrosis rather than tissue repair were not determined. Here, we investigated possible mechanisms and established their relevance to human lung fibrosis.

Results

The pyridinoline collagen fibrillogenesis genes PLOD2 and LOXL2 are co-expressed at sites of active fibrogenesis

In our previous work comparing human IPF lung tissue with age-matched control lung tissue, we identified that in bulk IPF lung tissue lysates there are significant increases in the relative expression levels of the collagen modifying enzymes LOXL2, LOXL3, and LOXL4, as well as PLOD2 (also known as lysyl hydroxylase or LH2) (Jones et al., 2018). To further investigate this observation, we first studied the transcriptomic profiles of fibroblast foci, the sites of active fibrogenesis in IPF. We analysed a data set we recently generated by integrating laser-capture-microdissection and RNA-Seq (LCMD/RNA-seq) which enabled profiling of the in situ transcriptome of fibroblast foci as well as alveolar septae from control tissue and IPF tissue (Gene Expression Omnibus (GEO) GSE169500). The LOXL enzyme with the greatest expression in fibroblast foci was LOXL2 (Figure 1a–e). Whilst LOXL3 and LOXL4 expression was increased within fibroblast foci, only limited expression was identified (Figure 1d and e). Furthermore, PLOD2 expression was significantly increased within fibroblast foci (Figure 1f), and PLOD2 expression levels correlated (r = 0.63, P = 0.04) with those of LOXL2 (Figure 1g) but not with those of other LOXL enzymes (Figure 1—figure supplement 1a-d), suggesting possible co-ordinated regulation of PLOD2 and LOXL2 gene expression. By contrast, expression of the major collagen fibrillogenesis gene COL1A1 did not significantly correlate with their expression (Figure 1—figure supplement 1e), suggesting that, in lung fibrosis, distinct pathways might promote pyridinoline cross-linking to dysregulate collagen fibril nano-structure independently of pathways regulating major fibrillar collagen synthesis. We then performed RNA in situ hybridisation upon IPF lung tissue, with semi-quantitative analysis (Supplementary file 1a); the results showed that the greatest expression of LOXL2 and PLOD2 in IPF tissue was by mesenchymal cells within fibroblast foci, and that LOXL2 and PLOD2 are co-expressed within the same cells (Figure 1h) within areas of fibrillar collagen deposition (Figure 1—figure supplement 1f-h).

Figure 1 with 1 supplement see all
The collagen cross-linking enzymes PLOD2 and LOXL2 are co-expressed at sites of active fibrogenesis in IPF.

(A–F) Expression of LOX, LOXL1, LOXL2, LOXL3, LOXL4, and PLOD2 in healthy alveolar septae, IPF alveolar septae and IPF fibroblast foci (n = 10 individual healthy and IPF donors). Relative expression levels are calculated as Fragments Per Kilobase of transcript per Million mapped reads (FPKM). Bars represent standard geometric means. **p < 0.01; ****p < 0.0001 by Tukey’s multiple comparisons test. (G) Scatterplot of paired fibroblast foci data from (C) and (F) were plotted to compare expression of PLOD2 and LOXL2 (Spearman rank correlation coefficient r = 0.63, p = 0.04). (H) Representative image of mRNA expression of PLOD2 (red chromagen) and LOXL2 (green chromagen) in IPF lung tissue (n = 7 donors) using RNAscope RNA in-situ hybridisation. A fibroblastic focus is identified by * and arrows identify co-expression pattern. Left scale bar 100 μm, right scale bar 20 μm.

HIF pathway activation is a key inducer of PLOD2 and LOXL2 expression in lung fibroblasts

To investigate common regulators of PLOD2 and LOXL2 in lung fibrosis, we studied their expression in primary human lung fibroblasts over a 72-hr time course following activation (Figure 2—figure supplement 1a) of transforming growth factor beta (TGFβ), epidermal growth factor (EGF), hypoxia inducible factors (HIF) or Wnt signalling pathways, each of which have been implicated in fibrogenesis (Richeldi et al., 2017; Yao et al., 2019; Martin-Medina et al., 2018; Königshoff et al., 2008; Yue et al., 2010; Bodempudi et al., 2014; Hill et al., 2019a; Yao et al., 2021; Zhou et al., 2021). A prodrug form of the hypoxia mimetic and broad spectrum 2-oxoglutarate oxygenase inhibitor N-oxalylglycine (dimethyloxalylglycine, DMOG) (Chowdhury et al., 2013), which inhibits the HIF prolyl hydroxylases with consequent stabilisation of HIF1α and HIF2α, most strongly upregulated both PLOD2 and LOXL2 mRNA and protein levels (Figure 2a–c and Figure 2—figure supplement 1b) but did not induce expression of interstitial collagen genes (COL1A1, COL3A1) (Figure 2d and Figure 2—figure supplement 1c). In contrast, TGFβ1 strongly induced COL1A1 and COL3A1 and this was associated with smaller up-regulation of PLOD2 at 24 hr and of LOXL2 at 72 hr (Figure 2a–d; Figure 2—figure supplement 1c). No induction of PLOD2 or LOXL2 was identified with canonical Wnt (Wnt3a), non-canonical Wnt (Wnt5a) or EGF pathway activation (Figure 2a–c). We further extended these observations by showing that treatment with the selective HIF prolyl 2 hydroxylase inhibitor, N-[[1,2-dihydro-4-hydroxy-2-oxo-1-(phenylmethyl)–3-quinolinyl]carbonyl]-glycine (IOX2) (Chowdhury et al., 2013) or culture for 24 hr under hypoxic conditions induced expression of PLOD2 and LOXL2 (Figure 2—figure supplement 1d, e), with immunofluorescent staining confirming an increase in intracellular LOXL2 and PLOD2 expression following DMOG or IOX2 treatment in comparison to treatment with TGFβ1 (Figure 2e). Transcriptional activation of HIF pathways requires assembly of a heterodimer between HIF1α or HIF2α and their obligate binding partner HIF1β (Schödel and Ratcliffe, 2019; Schofield and Ratcliffe, 2004). To confirm the dependence of the induction of PLOD2 and LOXL2 expression upon HIF levels, siRNA knockdown against HIF1α(HIF1A), HIF2α(EPAS1), and HIF1β (ARNT) was performed (Figure 3a). The knockdown of HIF1α, but not HIF2α prevented DMOG induction of PLOD2 mRNA and protein expression, whilst LOXL2 required silencing of both HIF1α and HIF2α or HIF1β (Figure 3b–d). Together, these findings identify that HIF stabilisation is required to orchestrate induction of PLOD2 and LOXL2 expression in human lung fibroblasts.

Figure 2 with 1 supplement see all
Hypoxia mimetics strongly promote PLOD2 and LOXL2 expression in lung fibroblasts.

(A–B, D) Relative gene expression using the ΔΔCt method of PLOD2, LOXL2, and COL1A1 in healthy lung fibroblasts over a 72-hr time course in the presence of EGF, TGFβ1, the hypoxia mimetic DMOG, Wnt3a, Wnt5a, or vehicle control. n = 3 independent experiments. Bars indicate geometric means. *p < 0.05; ***p < 0.001; ****p < 0.0001 by Dunnett’s multiple comparisons test. (C) PLOD2 and LOXL2 protein levels at 72 hr. β-actin loading control. The full blots are shown in Figure 2—source data 1. (E) Representative immunofluorescence images of healthy lung fibroblasts with indicated treatment stained for LOXL2 (red), PLOD2 (green), and DAPI (blue). Scale bar 50 μm.

HIF pathway activation regulates PLOD2 and LOXL2 expression in lung fibroblasts from patients with IPF.

(A) Fold changes in mRNA levels of HIF1α (HIF1A), HIF2α (EPAS1), and HIF1β (ARNT) in primary human lung fibroblasts from patients with IPF transfected with indicated siRNA followed by treatment with DMOG. β-actin-normalised mRNA levels in control cells were used to set the baseline value at unity. Data are mean  ±  s.d. n  =  3 samples per group. (B, C) Fold change in mRNA levels of LOXL2 (B) and PLOD2 (C) in IPF fibroblasts transfected with indicated siRNA followed by treatment with DMOG or vehicle control. β-actin-normalised mRNA levels in control cells were used to set the baseline value at unity. Data are mean  ±  s.d. n  =  3 samples per group. ns (not significant, p > 0.05); *p < 0.05; ****p < 0.0001 by Dunnett’s multiple comparisons test. (D) PLOD2, LOXL2 and HIF1α and β-tubulin protein levels in IPF fibroblasts transfected with indicated siRNA followed by treatment of DMSO or DMOG. β-tubulin was used as a loading control. The full blots are shown in Figure 3—source data 1.

HIF pathway activation and TGFβ1 synergistically increase PLOD2 expression

Given that TGFβ1 strongly induced major collagen fibrillogenesis genes whilst HIF pathways most strongly increased PLOD2 and LOXL2 expression levels, we investigated the effects of activating these pathways individually or in combination using lung fibroblasts from patients with IPF. The effect of DMOG in the absence or presence of TGFβ1 upon PLOD2 and LOXL2 induction (Figure 4a–c) was comparable to that identified using normal control lung fibroblasts. When combined, a synergistic effect upon the induction of PLOD2 expression was apparent which was greater than either pathway alone (Figure 4a and c). Whilst expression of LOXL2 was also increased with the combination of HIF stabilisation and TGFβ1, a corresponding increase in LOXL2 protein levels within cell lysates was not apparent. As LOXL2 is processed intracellularly before being extracellularly secreted, we therefore investigated whether increased secretion of LOXL2 was occurring; this identified that under conditions with HIF stabilisation LOXL2 secretion was increased in both IPF fibroblasts (Figure 4d) and control fibroblasts (Figure 4—figure supplement 1).

Figure 4 with 2 supplements see all
HIF pathway activation promotes PLOD2 and LOXL2 gene expression relative to fibrillar collagen expression.

Lung fibroblasts from IPF donors (n = 3 across two independent experiments) were cultured in the presence or absence TGFβ1, DMOG, combined TGFβ1 and DMOG, or vehicle control for 48 hr. (A, B) Relative gene expression of PLOD2 (A) and LOXL2 (B) using the ΔΔCt method. Bars indicate geometric means. Data are mean  ±  s.d. **p < 0.01; ***p < 0.001; ****p < 0.0001 by Dunnett’s multiple comparisons test. (C) PLOD2 and LOXL2 protein levels. β-actin was used as a loading control. (D) Protein expression of LOXL2 in conditioned media. Ponceau S staining showing total protein levels. The full blots are shown in Figure 4—source data 1. Bars in graph indicate geometric means. Data are mean  ±  s.d. **p < 0.01; ***p < 0.001; ****p < 0.0001 by Dunnett’s multiple comparisons test. (E, F) Expression of PLOD2 and LOXL2 from (A and B) was divided by COL1A1 expression (shown in Figure 4—figure supplement 2) to calculate proportionate expression changes of cross-linking enzymes relative to collagen fibrillogenesis gene expression. Bars indicate geometric mean. Grouped analysis was performed using Dunnett’s multiple comparison test. * p < 0.05, ** p < 0.01, *** p < 0.001, ****p < 0.0001.

Although TGFβ1 alone was sufficient to induce interstitial collagen gene (COL1A1) expression (Figure 4—figure supplement 2), HIF stabilisation significantly increased the ratio of PLOD2 and LOXL2 gene expression relative to fibrillar collagen (COL1A1) gene expression while TGFβ1 did not (Figure 4e and f), suggesting that TGFβ activity alone may be insufficient to promote the altered collagen cross-linking that is present in IPF lung tissue. Together these findings demonstrate that whilst TGFβ1 has a dominant role in increasing the rate of synthesis of major fibrillar collagens, HIF pathways may have a key role in regulating pathological post-translational modifications and collagen structure in lung fibrosis.

HIF pathway activation alters collagen structure-function and increases tissue stiffness

To investigate whether HIF pathway activation acts as a mechanism that drives pathologic collagen crosslinking by disproportionate induction of collagen-modifying enzymes relative to TGFβ-induced collagen fibril synthesis, we employed our long-term (6 weeks) 3D in vitro model of lung fibrosis using primary human lung fibroblasts from patients with IPF, which we have previously described (Jones et al., 2018) and which allows direct evaluation of pyridinoline cross-linking, collagen nanostructure, and tissue biomechanics. We employed the selective HIF-prolyl hydroxylase inhibitor IOX2 to test within the in vitro fibrosis model, confirming HIF stabilisation by IOX2 following 2-week culture, and that in combination with TGFβ1 this promoted PLOD2 and LOXL2 expression (Figure 5—figure supplement 1a and b). Following 6 weeks of culture with TGFβ1 in the absence (control) or presence of IOX2 to drive HIF pathway activation, mature pyridinoline cross links (DPD/PYD) were significantly increased by the addition of IOX2 (Figure 5a) and these achieved a level comparable to our previous findings in IPF tissue (Jones et al., 2018). The biomechanical consequence of HIF stabilisation by IOX2 treatment was then investigated with parallel plate compression testing, identifying a greater than threefold increase in tissue stiffness by the addition of IOX2 (Figure 5b), with the mean (± SEM) compressive modulus measurement following IOX2 treatment of (107.1 ± 10.7) kPa comparable to the maximal stiffness of between 50 and 150 kPa we and others have previously identified in highly fibrotic areas in IPF tissue (Booth et al., 2012).

Figure 5 with 1 supplement see all
HIF pathway activation promotes pyridinoline cross-linking, alters collagen nano-architecture, and increases tissue stiffness.

Lung fibroblasts from IPF patients (n = 3 donors, two experiments per donor) were used in the 3D model of fibrosis in the presence of IOX2 or vehicle control. Bars indicate geometric mean + s.e.m. Analysis was performed using a Mann-Whitney t-test (two-tailed) **p < 0.01; ***p < 0.001; ****p < 0.0001. (A) Total mature trivalent (PYD+ DPD) collagen cross-links determined by ELISA. n = 6 samples from three IPF donors. (B) Tissue stiffness measured from parallel-plate compression testing (n = 12 samples from three IPF donors) determined by Young’s modulus and represented as proportion of control. (C) Representative images of histological sections of samples stained with picrosirius red and imaged under polarised light. Scale bar 20 μm. (D) Representative electron microscopy images of collagen fibrils within the 3D model of fibrosis. Scale bar 50 nm. (E) Collagen fibril diameter within the 3D model of fibrosis measured in transverse section (300 fibrils for each condition from two IPF donors, measured by a blinded investigator). (F) Atomic force microscopy indentation modulus of collagen fibrils (3–7 fibrils per donor) from control (n = 42 fibrils from eight donors) or IPF lung tissue (n = 57 fibrils from 10 donors) under non-hydrated conditions; each data point represents the mean of 30–50 force-displacement curves per fibril.

We next assessed collagen morphology. When visualised by polarised light Picrosirius red microscopy (Figure 5c), highly organised collagen fibrils were evident in vehicle-treated fibrotic control cultures as well as in those treated with IOX2 with no apparent morphological differences (Figure 5—figure supplement 1c). By contrast, ultrastructural analysis of the collagen fibrils using electron microscopy identified a change in collagen nanostructure with a significant decrease in fibril diameter (Figure 5d and e) when pyridinoline cross-linking was increased by IOX2, consistent with our previous observation that fibril diameter is increased by inhibition of pyridinoline cross-linking (Jones et al., 2018). In support of the disease relevance of our in vitro findings, non-hydrated collagen fibrils from patients with IPF have reduced diameters when measured by atomic force microscopy (Figure 5f), consistent with our previous findings that hydrated collagen fibrils extracted from IPF lung tissue have a reduced diameter compared to control samples (Jones et al., 2018). Together, these data identify HIF pathway activation to be a key regulator of pyridinoline cross-link density, collagen fibril nano-architecture, and tissue stiffness.

Pseudohypoxia and loss of FIH activity promotes HIF pathway activation in lung fibroblasts

Whilst canonical HIF pathway activation was observed in lung fibroblasts under hypoxic conditions, elevated levels of HIF1α and HIF2α in IPF fibroblasts under normoxic conditions have recently been reported (Aquino-Gálvez et al., 2019), suggesting a pseudohypoxic state that is a state in which cells express, at least some, hypoxia-associated genes and proteins, regardless of the oxygen status (Russell et al., 2017). To further investigate this possibility, we employed gene set variation analysis (GSVA) using a validated 15-gene HIF/hypoxia gene expression signature (Buffa et al., 2010) to published datasets, identifying that fibroblasts cultured under normoxic conditions from patients with a usual interstitial pneumonia pattern of fibrosis or systemic sclerosis associated lung fibrosis have a significantly increased HIF score (i.e. manifest evidence for HIF upregulation) compared to cultured control fibroblasts (Figure 6a), consistent with an oxygen independent increase in HIF activity. Furthermore, there was a significant increase in the HIF score in lung mesenchymal stromal cells of patients with progressive lung fibrosis compared to those with stable fibrosis (Figure 6b), suggesting that HIF pathway activation may be required for fibrosis progression.

Figure 6 with 1 supplement see all
Pseudohypoxia and loss of FIH activity promotes HIF pathway signalling in IPF fibroblasts and increases tissue stiffness.

(A) HIF GSVA scores calculated in human lung fibroblasts derived from control or patients with interstitial lung disease (scleroderma lung or a usual interstitial pneumonia / IPF pattern) (GSE40839). Data are mean  ±  s.d. ***p < 0.001; ****p < 0.0001 by Dunnett’s multiple comparisons test. (B) HIF GSVA scores calculated in human bronchoalveolar lavage derived mesenchymal stromal cells from patients with stable and progressive IPF (GSE73854). Data are mean  ±  s.d. *p < 0.05 by the unpaired t test. (C) PLOD2, LOXL2, HIF1β, FIH, and β-tubulin protein levels in lung fibroblasts from patients with IPF transfected with indicated siRNA. β-tubulin was used as a loading control. The full blots are shown in Figure 6—source data 1. (D) Diagram explaining the HIF1α CAD reporter assay in E and F. In brief, the FIH asparaginyl hydroxylase hydroxylates HIF1α CAD, inhibiting its binding with CBP/p300 and decreasing luciferase activity. When FIH is inhibited, the non-hydroxylated HIF1α CAD can bind with CBP/p300 increasing luciferase activity. (E) HIF1α CAD reporter assays in normal human lung fibroblasts (control fibroblasts) or IPF lung fibroblasts (IPF fibroblasts). Values represent the relative fold increase of firefly luciferase in relation to Renilla luciferase, normalised against control (1.0). Data are mean  ±  s.d. n  =  3 samples per group. **p <  0.01 by unpaired t test. (F) HIF1α CAD reporter assays in control fibroblasts with indicated treatment (hydrogen peroxide (T-hydro), DMOG, or vehicle control). Values represent relative fold of firefly luciferase in relation to Renilla luciferase, normalised against control (1.0). Data are mean  ±  s.d. n  =  3 samples per group. (G and H) Control lung fibroblasts (n = 3 donors, two experiments per donor) were used in the 3D model of fibrosis in the presence of IOX2 and/or DM-NOFD or vehicle control as indicated. (G) Total mature trivalent (PYD+ DPD) collagen cross-links determined by ELISA. n = 6 samples from three donors. (H) Tissue stiffness measured from parallel-plate compression testing (n = 6 samples from three donors) determined by Young’s modulus and represented as proportion of control. * p < 0.05, ** p < 0.01, *** p < 0.001, ****p < 0.0001 by Dunnett’s multiple comparisons test.

To further investigate the mechanism underlying pseudohypoxic HIF activity in lung fibrosis, we investigated the role of Factor Inhibiting HIF (FIH), a Fe (II)- and 2-oxoglutarate (2-OG)-dependent dioxygenase, which regulates HIF activity and likely the set of HIF target genes upregulated via hydroxylating a conserved asparagine (Asn) residue within the HIFα C-terminal activation domain (CAD), a post-translational modification that blocks interactions between the HIFα-CAD and the histone acetyl transferases CBP/p300 (Elkins et al., 2003; Hewitson et al., 2002; Lando et al., 2002; Mahon et al., 2001; McNeill et al., 2002; Chan et al., 2016). Whilst oxygen tension is the classical regulator of FIH activity, oxidative stress can also inactivate FIH so promoting HIF activity under normoxic conditions (Masson et al., 2012).

Initially, to investigate the potential role of reduced FIH activity in regulating collagen post-translational modifications, we silenced FIH under normoxic conditions; the results show that loss of FIH was sufficient to induce both PLOD2 and LOXL2 expression, and that this effect required HIF promoted transcription, since HIF1β knockdown prevented their induction (Figure 6c). Whilst FIH is stably constitutively expressed across tissues (Bracken et al., 2006; Stolze et al., 2004), the activity levels of FIH can vary (Wang et al., 2018; Tan et al., 2007; Kroeze et al., 2010); thus, we compared FIH activity in control or IPF fibroblasts using a UAS-luc/GAL4DBD-HIF1αCAD binary reporter system (HIF1α CAD reporter) (Coleman et al., 2007). In this assay, the activity of FIH is monitored by a Gal4-driven luciferase reporter that registers the activity of the heterologous Gal4‐HIF‐CAD fusion protein. Inhibition of FIH leads to a reduction in hydroxylation at Asn‐803 of the HIF‐CAD (C-terminal transactivation domain) fusion, which permits increased recruitment of the transcriptional co‐activators p300/CBP and enhanced reporter gene activity (Figure 6d). Consistent with a loss of function of FIH in lung fibrosis, we found FIH activity was significantly reduced in fibroblasts from patients with IPF compared to control fibroblasts (Figure 6e). We further confirmed that a reduction in FIH activity in normal lung fibroblasts could be caused under normoxia by oxidative stress, achieving a level of HIF CAD activity comparable to treatment with the hypoxia mimetic DMOG (Figure 6f). Thus, in lung fibroblasts a reduction in FIH activity may promote HIF pathway activation to dysregulate collagen structure-function.

We next employed the FIH-selective inhibitor DM-NOFD (McDonough et al., 2005) within our 3D model of fibrosis. We confirmed that FIH inhibition by DM-NOFD was sufficient to induce the HIF pathway activation marker gene carbonic anhydrase IX (CA9), PLOD2, and LOXL2 gene expression following 2-week culture (Figure 6—figure supplement 1), and in combination with HIF stabilisation (IOX2) this expression was further increased. Following 6 weeks of culture, DM-NOFD increased mature pyridinoline cross-links (Figure 6g) as well as tissue stiffness (Figure 6h), whilst the combination of DM-NOFD and IOX2 was additive. Thus, FIH inhibition can promote collagen post-translational modification and increase tissue stiffness.

HIF pathway activation localises in areas of active fibrogenesis to cells co-expressing LOXL2 and PLOD2

To support our in vitro studies, we investigated for evidence that HIF regulates PLOD2 and LOXL2 expression within the fibroblast foci of human IPF lung tissue. To assess for HIF activity, we applied GSVA using the 15-gene HIF/hypoxia gene expression signature (Buffa et al., 2010) to the transcriptome of each fibroblast focus, identifying that the HIF signature score, but not TGFβ score, significantly correlated with LOXL2/PLOD2 expression (Figure 7a and b). Furthermore, analysis of serial tissue sections using immunohistochemistry identified that HIF1α and the HIF pathway activation marker gene carbonic anhydrase IX (CA-IX) were expressed within fibroblast foci (Bodempudi et al., 2014; Loncaster et al., 2001), and that this expression localised to cells co-expressing LOXL2 and PLOD2 mRNA (Figure 7c; Figure 7—figure supplement 1). Finally, as FIH is more sensitive to inhibition by oxidative stress (Masson et al., 2012) compared to the PHDs, which are more sensitive to hypoxia than FIH (Masson et al., 2012), we investigated whether HIF activation occurs in lung mesenchymal cells in the context of oxidative stress. We applied GSVA to a published single cell RNAseq dataset (114,396 cells) from 10 control and 20 fibrotic lungs which identified 31 cell types including four fibroblast types (fibroblasts, myofibroblasts, Hyaluronan Synthase 1 (HAS1) high and Perilipin 2 (PLIN2)+ fibroblasts) (Figure 8—figure supplement 1; Habermann et al., 2020). Applying HIF signature or upregulated oxidative stress gene expression signatures to this dataset, we identified that, compared to fibroblasts and myofibroblasts, the HAS1 high and PLIN2+ cells, whose presence was almost exclusively derived from the IPF lung tissue, had significantly increased HIF and upregulated oxidative stress scores (Figure 8a–d) and that these two scores were significantly correlated (Figure 8e), consistent with an increase in pseudohypoxic HIF activity in these disease-specific mesenchymal cell types.

Figure 7 with 1 supplement see all
HIF pathway activation localises in areas of active fibrogenesis to cells co-expressing LOXL2 and PLOD2.

(A–B) Scatterplots showing correlations between LOXL2/PLOD2 expression and HIF scores (A) or TGFβ scores (B) in IPF fibroblast foci (n = 10) using the Spearman rank correlation coefficient. (C) Representative images of serial sections of lung tissue from patients with IPF (n = 3). mRNA expression of PLOD2 (red chromagen) and LOXL2 (green chromagen) using RNAscope RNA in-situ hybridisation with immunohistochemical staining for Carbonic anhydrase IX (CA-IX) and HIF1α using DAB (brown). A fibroblastic focus is identified by *. Scale bar 20 μm.

Figure 8 with 1 supplement see all
Gene set variance analysis of single-cell RNAseq fibroblast populations identifies co-enrichment of HIF score and oxidative stress genes.

(A) HIF score GSVA in control and IPF fibroblasts sequenced by single-cell RNAseq (GSE135893). Colours correspond to calculated GSVA score for each cell. (B) Plot of mean HIF GSVA scores for each fibroblast type in control and IPF fibroblast cell populations and compared using Dunnett’s multiple comparison test, ****p < 0.0001. (C) GSVA scores for genes upregulated in IPF in this dataset associated with the Gene Set: HALLMARK_REACTIVE_OXYGEN_SPECIES_PATHWAY (M5938). (D) Plot of upregulated oxidative stress GSVA scores for each fibroblast type in control and IPF cells. (E) Correlation plot of HIF score vs upregulated oxidative stress GSVA score for single cell RNAseq data. Correlation coefficient is Pearson’s product-moment coefficient.

Discussion

We previously reported that altered collagen fibril nanoarchitecture is a core determinant of dysregulated ECM structure-function in human lung fibrosis (Jones et al., 2018). Here, through ex vivo models, bioinformatics and human lung fibrosis tissue studies, we extend these observations leading to the discovery that HIF pathway activation promotes pathologic pyridinlone collagen crosslinking and tissue stiffness by disproportionate induction of collagen-modifying enzymes relative to TGFβ-induced collagen fibril synthesis. Furthermore, this may occur via pseudohypoxic oxygen-independent mechanisms, including the involvement of a decrease in FIH activity that can occur due to oxidative stress, which is thought to play a significant role in IPF pathogenesis (Cheresh et al., 2013). Consistent with this, oxidative stress is increased in subpopulations of IPF fibroblasts whilst FIH activity is significantly reduced in fibroblasts from patients with lung fibrosis resulting in HIF activation under normoxic conditions. Thus, we provide evidence that dysregulated HIF activity is a core regulator of ECM structure-function in human lung fibrosis, and that this may be a key determinant of pathologic tissue stiffness and progressive human lung fibrosis.

TGFβ is a multifunctional growth factor with key roles in normal development and wound healing. It is also considered the prototypic profibrogenic cytokine that promotes increased ECM deposition and has been associated with fibrosis across multiple organs (Yue et al., 2010). We identified that in lung fibroblasts, TGFβ1 increased fibrillar collagen mRNA transcription but its relative effects on PLOD2 or LOXL2 were more limited, suggesting that TGFβ pathway activation alone may be insufficient to cause a substantial increase in pathologic pyridinoline collagen crosslinking. HIF-mediated transcription appears to be relatively more important in inducing PLOD2/LOXL2 relative to interstitial collagen fibril synthesis, so promoting pyridinoline collagen cross-linking, altering collagen fibril nanostructure, and increasing tissue stiffness. While TGFβ has been reported to cause HIF stabilisation (Basu et al., 2011), our findings suggest that this effect is modest and that further HIF-mediated activation is likely required to drive matrix stiffening. This proposal is consistent with a recent result implying a hierarchical relationship in which HIF proteins play a relatively important role in the induction of PLOD2 expression, that is in this regard the effect of the HIF transcription factors appears to be more important relative to that of TGFβ stimulated SMAD proteins (Rosell-García et al., 2019). Thus, we propose that HIF pathway activation acts as a key pathologic ‘second hit’ which disrupts the normal wound healing role of TGFβ by altering collagen fibril nanoarchitecture so dysregulating ECM structure-function and promoting progressive lung fibrosis. In keeping with this concept, GSVA using a validated HIF score (Buffa et al., 2010) applied to microarray data for lung mesenchymal stromal cells showed that HIF activity was increased in cells from patients with progressive lung fibrosis compared with those with stable disease.

We investigated the functional consequences of our findings by employing our long-term 3D in vitro model of lung fibrosis. The results show that HIF pathway activation using a HIF stabilising PHD inhibitor and/or an FIH inhibitor increased pyridinoline cross-links to a level comparable to that identified in IPF tissue, and that the increase in cross-links is associated with an increase in tissue stiffness comparable to the extremes of stiffness identified in IPF tissue together with a reduction in fibril diameter similar to those present in IPF lung tissue. Together these observations support the human disease relevance of HIF pathway activation to IPF and define conditions for future mechanistic studies whereby the 3D in vitro model recapitulates key features of dysregulated collagen structure-function in IPF.

The LOX and LOXL enzymes play key roles in the process of fibrillar collagen production and are tightly regulated in normal development and under physiological conditions (Trackman, 2016). In our LCMD RNA-Seq analyses, LOXL2 was the most highly expressed LOX/LOXL family member as well as the only LOX/LOXL member which correlated with PLOD2 expression, whilst in our previous work investigating collagen structure-function dysregulation in human lung fibrosis, we identified that gene expression of LOXL2 was significantly increased in IPF tissue when compared to age-matched control lung tissue (Jones et al., 2018). Furthermore, using a small molecule LOXL inhibitor in our 3D model of fibrosis, we identified a greater than 50% reduction in mature pyridinoline cross-links using a concentration which completely inhibits LOXL2 but has minimal effects on LOX and LOXL1 (Jones et al., 2018). This is consistent with previous reports that LOXL2 has key pathologic roles in cancer and fibrosis (Barker et al., 2012; Barry-Hamilton et al., 2010). As our studies do not unequivocally exclude a potential role for other LOX/LOXL family members in human lung fibrosis an area of future study could be the systematic silencing of each LOX/LOXL family member using CRISPR gene editing.

The HIF signalling pathway has been reported to be active in lungs and fibroblasts from IPF patients, as determined by the abundance of HIF1α and HIF2α (Aquino-Gálvez et al., 2019; Bodempudi et al., 2014). These findings are consistent with our own observations of increased expression of the HIF-responsive gene, CA-IX. Hypoxia has been proposed to have a pathogenetic role in lung fibrosis through mechanisms including fibroblast proliferation, augmented ER stress, epithelial-mesenchymal transition, and glycolytic reprogramming (Bodempudi et al., 2014; Higgins et al., 2007; Senavirathna et al., 2018; Goodwin et al., 2018). Furthermore, a number of reports have proposed that cross-talk between TGFβ and hypoxia may promote fibrosis, with hypoxia and TGFβ1 synergistically increasing myofibroblast marker expression (Senavirathna et al., 2020), promoting experimental nickel oxide nanoparticle-induced lung fibrosis (Qian et al., 2015), and HIF1α mediating TGF-β-induced PAI-1 production in alveolar macrophages in the bleomycin model of lung fibrosis (Ueno et al., 2011). Here, we extended these previous observations by showing that in lung fibrosis, loss of FIH activity either by siRNA-mediated knockdown or exposure to oxidative stress induces HIF pathway activation independently of oxygen tension, so dysregulating collagen fibrillogenesis under normoxic conditions. FIH negatively regulates HIF activity by hydroxylation of N803, preventing the interaction of the HIFα CAD with CBP/p300 (Elkins et al., 2003; Hewitson et al., 2002; Lando et al., 2002; Mahon et al., 2001; McNeill et al., 2002). Whilst oxygen tension is the classical regulator of FIH activity, oxidative stress can inactivate FIH so promoting HIF activity, with FIH more sensitive to oxidative stress than the HIF prolyl hydroxylases (Masson et al., 2012). Oxidative stress has been implicated as an important profibrotic mechanism in the lungs and other organs (Cheresh et al., 2013; Purnomo et al., 2013; Sánchez-Valle et al., 2012); it can arise from exposure to environmental toxins (e.g. air pollution, tobacco, asbestos, silica, radiation, and drugs such as bleomycin) or from endogenous sources including mitochondria, NADPH oxidase (NOX) activity, and/or inadequate or deficient antioxidant defenses (Cheresh et al., 2013). In our bioinformatic studies, we observed subsets of disease-specific fibroblasts with elevated scores for oxidative stress and these same populations had evidence of HIF pathway activation. Further investigation is merited to understand the consequences of this upon the fibrotic microenvironment including possible dysregulation of epithelial-mesenchymal cross-talk.

To our knowledge whether perturbations in FIH activity could contribute to fibrosis has not been investigated previously. Whilst our studies have focused upon HIF pathways and collagen, functionally FIH, via both HIF-dependent and HIF-independent pathways, has been reported to regulate metabolism (Zhang et al., 2010; Scholz et al., 2016; Peng et al., 2012a; Sim et al., 2018), keratinocyte differentiation (Peng et al., 2012b), vascular endothelial cell survival (Kiriakidis et al., 2015), tumour growth (Pelletier et al., 2012; Kuzmanov et al., 2012) and metastasis (Kang et al., 2018) as well as Wnt signalling (Rodriguez et al., 2016), suggesting that the loss of FIH activity that we have identified could have pleiotropic effects in lung fibrosis, meriting further investigation.

In summary, this study identifies that HIF pathway activation via oxygen dependent and oxygen independent mechanisms promotes pyridinoline collagen cross-linking which is a defining feature of human lung fibrosis that dysregulates ECM structure-function to promote progressive lung fibrosis. Our findings suggest that therapeutically targeting HIF pathway activation might restore ECM homeostasis and so prevent fibrosis progression.

Materials and methods

Lung tissue sampling

Request a detailed protocol

Human lung experiments were approved by the Southampton and South West Hampshire and the Mid and South Buckinghamshire Local Research Ethics Committees (ref 07 /H0607/73), and all subjects gave written informed consent. Clinically indicated IPF lung biopsy tissue samples deemed surplus to clinical diagnostic requirements were formalin fixed and paraffin embedded. All IPF samples were from patients subsequently receiving a multidisciplinary diagnosis of IPF according to international consensus guidelines.

Transcriptomic analysis of in situ IPF fibroblast foci

Request a detailed protocol

We analysed a transcriptomic data set that we have recently established (GSE169500). Briefly, laser capture microdissection was performed upon Formalin-Fixed Paraffin-Embedded (FFPE) control non-fibrotic lung tissue (alveolar septae, [n = 10]) and usual interstitial pneumonia/idiopathic pulmonary fibrosis FFPE lung tissue (fibroblast foci, [n = 10] and adjacent non-affected alveolar septae, [n = 10]). Total RNA was isolated, cDNA libraries were prepared using Ion Ampli‐Seq‐transcriptome human gene expression kit (Life Technologies, Paisley, UK) and sequenced using Ion Torrent Proton Sequencer. A two-stage mapping strategy was used to map the reads to UCSC hg19 human genome. Cufflinks was used to calculate Fragments per Kilobase of exon per Million (FPKM) values.

RNA in-situ hybridisation

Request a detailed protocol

Simultaneous in situ detection of the LOXL2 and PLOD2 mRNA on human IPF formalin-fixed paraffin-embedded tissue sections from IPF lung tissue biopsy samples from seven patients were performed using duplex RNAscope technology (Advanced Cell Diagnostics, Biotechne, Abingdon, UK). LOXL2 was detected by C1-probe (Probe-Hs-LOXL2-C1, 311341) and PLOD2 was detected by C2-probe (Probe-Hs-PLOD2-C2, 547761-C2). Briefly, 5 μm human IPF lung tissue sections were baked at 60 °C, deparaffinised in xylene, followed by dehydration in an ethanol series. Target retrieval, hybridisation with target probes, amplification, and chromogenic detection were performed according to the manufacturer’s recommendations (RNAscope 2.5 Duplex Detection protocol for FFPE tissues). Sections were counterstained with Gill’s Hematoxylin, and mounted with Vectamount permanent mounting medium prior to imaging. Assays were performed with duplex positive (PPIB and POLR2A) and negative controls. For co-localisation studies, adjacent serial sections were stained using modified Movat’s pentachrome or hematoxylin and eosin stain as previously reported (Jones et al., 2016). Images were acquired using an Olympus Dotslide Scanner VS110 (Olympus UK, Southend-on-Sea, UK). Semi-quantitative analysis (range absent to +++) of LOXL2 and PLOD2 expression in cell types in IPF tissue was performed by an expert lung pathologist (AF).

2D cell culture, reagents, and transfections

Request a detailed protocol

Primary fibroblast cultures were established from lung parenchyma tissue of patients with IPF obtained by video-assisted thoracoscopic lung biopsy at University Hospital Southampton or non-fibrotic control lung parenchyma tissue (macroscopically normal lung sampled remote from a cancer site in patients undergoing surgery for early stage lung cancer) (Jones et al., 2018; Yao et al., 2019; Conforti et al., 2020; Hill et al., 2019b). MRC5 lung fibroblasts (RRID:CVCL_0440) were obtained from the European Collection of Authenticated Cell Cultures (ECACC). All cultures were tested and free of mycoplasma contamination. Demographic details for the primary lung fibroblast lines are provided in Supplementary file 1b.

Fibroblasts were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% foetal bovine serum (FBS), 50 units/ml penicillin, 50 μg/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate, and 1 x non-essential amino acids (DMEM/FBS) (Life Technologies, Paisley, UK). All cells were kept at 37  °C and 5% CO2. Hypoxic incubation of cells was carried out in a H35 Hypoxystation (Don Whitley Scientific) in which cells were cultured in humidified atmosphere of 1% O2, 5% CO2, and 94% N2 at 37 °C. Following hypoxic incubation, cells were kept in hypoxic condition until samples were collected.

For pro-fibrogenic mediator studies, control lung fibroblasts were treated in the presence of EGF (R&D systems, 236-GMP-200, 10 ng/mL), TGFβ1 (R&D systems, 240-GMP-010, 10 ng/mL), Dimethyloxaloylglycine (DMOG) (Merck, CAS89464-63-1, 1 mM), Wnt3a (R&D systems, 5036-WN-010, 100 ng/mL), Wnt5a (R&D systems, 645-WN-010, 100 ng/mL), or vehicle control (DMSO). For subsequent HIF studies fibroblasts were treated in the presence of DMOG (1 mM), IOX2 (50 μM or 250 μM), or vehicle control (DMSO).

Short interfering RNA (siRNA) oligos against HIF1A (HIF1α) (MU-00401805-01-0002), EPAS1 (HIF2α) (MU-004814-01-0002), ARNT (HIF1β) (MU-007207-01-0002) and HIF1AN (FIH) (MU-004073-02-0002), LOXL2(L-008020-01-0005) were from Dharmacon, Cambridge, UK. Sequences are available from Dharmacon, or Supplementary file 2. As a negative control, we used siGENOME RISC-Free siRNA (Dharmacon, D-001220–01). Human lung fibroblasts were transfected with the indicated siRNA at a final concentration of 35 nM using Lipofectamine RNAiMAX reagent (Invitrogen).

Reporter assay

Request a detailed protocol

FIH activity was evaluated using a UAS-luc/GAL4DBD-HIF1αCAD binary reporter system (HIF1α CAD reporter) (Coleman et al., 2007). For the luciferase reporter assays, human lung fibroblasts (control or IPF fibroblasts) were reverse transfected using Lipofectamine 3000 (Invitrogen) with 50  ng of phRL-CMV (Promega UK, Southampton, UK), which constitutively expresses the Renilla luciferase reporter, plus 225  ng of plasmid-GAL4DBD-HIF1αCAD and 225  ng of plasmid-UAS-luc per well. After 24 hour of transfection, a final concentration of 1 mM of DMOG, 1 mM DMSO or 20 μM freshly prepared T‐hydro (tert‐butyl hydroperoxide) (Sigma-Aldrich, Poole, UK) was dosed for 16 hours. T‐hydro was added to the cells every 2 hours. Finally, the transcriptional assay was carried out using the Dual-Luciferase reporter assay system (Promega) following the manufacturer’s protocol.

HIF score, TGFβ score, and oxidative stress GSVA analyses

Request a detailed protocol

Raw CEL files for GSE73854 and GSE40839 were downloaded from GEO and imported into RStudio (version 3.6). Raw data were normalised by Robust Multi-array Average (RMA) function in the affy package (version 1.64.0). Multiple probes relating to the same gene were deleted and summarised as the median value for further analysis.

A 15-gene expression signature (ACOT7, ADM, ALDOA, CDKN3, ENO1, LDHA, MIF, MRPS17, NDRG1, P4HA1, PGAM1, SLC2A1, TPI1, TUBB6, and VEGFA) was selected to classify HIF activity (Buffa et al., 2010). All parameters and variables can be found in the accompanying file (Source code 1). This gene signature was defined based on knowledge of gene function and analysis of in vivo co-expression patterns and was highly enriched for HIF-regulated pathways. The HIF score for each sample was calculated by using gene set variation analysis (GSVA) (Hänzelmann et al., 2013) based on this 15-gene expression signature. The TGFβ score for each sample was calculated by using GSVA based on a list of gene from Gene Set: HALLMARK_TGF_BETA_SIGNALING (M5896). All parameters and variables can be found in the accompanying file (Source code 2). The Student t-test was used to evaluate the statistical difference in HIF scores between different conditions.

For single-cell transcriptomic analyses raw CEL files for GSE135893 were downloaded from GEO. Data was processed using the Seurat R package (v3.2.1) in R version 4.0.2. Cell types were assigned based on the published metadata (Habermann et al., 2020). Fibroblast counts data were log-normalised, variable genes quantified and principal component analysis performed on these variable genes. T-stochastic nearest neighbour embedding (t-SNE) dimensional reduction was performed on the top 15 principal components to obtain embeddings for individual cells. GSVA was performed using the 15 genes used for HIF score calculation as above. An oxidative stress score for each cell was calculated using GSVA based on a list of genes upregulated in IPF cell populations (ABCC1, CDKN2D, FES, GCLC, GCLM, GLRX2, HHEX, IPCEF1, JUNB, LAMTOR5, LSP1, MBP, MGST1, MPO, NDUFA6, PFKP, PRDX1, PRDX2, PRDX4, PRNP, SBNO2, SCAF4, SOD1, SOD2, RXN1, TXN, TXNRD1) from Gene Set: HALLMARK_REACTIVE_OXYGEN_SPECIES_PATHWAY (M5938). All parameters and variables can be found in the accompanying file (Source code 3). Upregulated oxidative stress genes were those whose expression was higher in IPF populations than control. Calculated GSVA scores were mapped onto t-SNE plots. Student’s t-test was used to calculate statistical differences between GSVA scores of the different cellular populations.

3D in vitro model of fibrosis

Request a detailed protocol

Culture was performed as previously described (Jones et al., 2018). Briefly, peripheral lung fibroblasts were obtained as outgrowths from surgical lung biopsy tissue of patients (n = 3 donors) who were subsequently confirmed with a diagnosis of IPF. All primary cultures were tested and free of mycoplasma contamination. The fibroblasts were seeded in Transwell inserts in DMEM containing 10% FBS. After 24 hr, the media was replaced with DMEM/F12 containing 5% FBS, 10 μg/ml L-ascorbic acid-2-phosphate, 10 ng/ml EGF, and 0.5 μg/ml hydrocortisone with or without 50 μM or 250 μM IOX2 and/or 1 mM DM-NOFD (McDonough et al., 2005), as indicated; each experiment included a vehicle control (0.2% DMSO). TGF-β1 (3 ng/mL) was added to the cultures, and the medium replenished three times per week. After 2 weeks spheroids were lysed for western blotting. After 6 weeks, the spheroids were either snap frozen for parallel-plate compression testing, analysis of cross-linking, and histochemical staining, or fixed using 4% paraformaldehyde for histochemistry or 3% glutaraldehyde in 0.1 M cacodylate buffer at pH 7.4 for electron microscopy.

Reverse transcription quantitative polymerase chain reaction (RTqPCR)

Request a detailed protocol

RTqPCR was performed as previously described (Yao et al., 2019; Conforti et al., 2020; Hill et al., 2019b). Primers and TaqMan probe sets were obtained from Primer Design, Southampton, UK (LOXL2, COL1A1, Col3A1, PLOD2), ThermoFisher Scientific, Reading, UK (HIF1A [HIF1α], EPAS1 [HIF2α], ARNT [HIF1β]), and Qiagen, Manchester, UK (QuantiTect Primer Assays, HIF1A, EPAS1, ARNT, LOXL2, PLOD2, CA9, ACTB).

Western blotting

Request a detailed protocol

Fibroblasts were lysed using 2 x Laemmli SDS sample buffer or urea buffer (8 M Urea, 1 M Thiourea, 0.5% CHAPS, 50 mM DTT, and 24 mM Spermine). Western blotting of cellular lysates was performed for β-actin (1:100.000, Sigma-Aldrich, Poole, UK), LOXL2 (1:1000, R&D Systems, Abingdon, UK), HIF1α (1:1000, BD Biosciences, Wokingham, UK), FIH (1:200, mouse monoclonal 162 C) (Wang et al., 2018), β-tubulin (1:1000, Cell Signaling Technology, London, UK), HIF1 β (1:1000, Cell Signaling Technology), p-Smad2/3 (1:1000, Cell Signaling Technology), p-ERK (1:1000, Cell Signaling Technology), active β-catenin (1:1000, Cell Signaling Technology). Immunodetected proteins were identified using the enhanced chemiluminescence system (Clarity Western Blotting ECL Substrate, Bio-Rad Laboratories Ltd, Watford, UK) or Odyssey imaging system (LI-COR), and evaluated by ImageJ 1.42q software (National Institutes of Health).

Immunofluorescence staining

Request a detailed protocol

Cells were fixed with 4% paraformaldehyde followed by permeabilisation and staining with primary antibodies for LOXL2 (1:100, R&D Systems), PLOD2 (1:100, Proteintech) and tetramethylrhodamine (TRITC)-conjugated Phalloidin (1:1000, Millipore UK Limited, Watford, UK). The secondary antibodies used were Alexafluor 488 and 647 (1:1000, BioLegend UK Ltd, London, UK). Cell nuclei were counterstained with 4',6-Diamidino-2-Phenylindole, Dihydrochloride (DAPI) (1:1000, Millipore UK Limited, Watford, UK). Cells were imaged using an inverted confocal microscope (Leica TCS-SP5 Confocal Microscope, Leica Microsystems).

Immunohistochemistry

Request a detailed protocol

Control or IPF lung tissues (n = 3 donors) were fixed and embedded in paraffin wax; tissue sections (4 μm) were processed and stained as previously described (Yao et al., 2019; Hill et al., 2019b). Briefly, the tissue sections were de-waxed, rehydrated and incubated with 3% hydrogen peroxide in methanol for 10 min to block endogenous peroxidase activity. Sections were then blocked with normal goat serum and incubated at room temperature with a primary antibody against CA-IX (1:500, Novus Biologicals, Cambridge, UK) or HIF1α (1:500, Cayman Chemical, Michigan, USA), followed by a biotinylated secondary antibody (1:500, Vector Laboratories Ltd., UK); antibody binding was detected using streptavidin-conjugated horse-radish peroxidase and visualised using DAB before counter-staining with Gill’s Haematoxylin. Images were acquired using an Olympus Dotslide Scanner VS110.

Picrosirius red collagen area quantitation

Request a detailed protocol

Sample sections, stained with Picrosirius Red as previously described (Jones et al., 2018), were imaged under polarised light and 10 areas were selected at random for each condition (5 each from two donors). Images of dimension 1498 × 1221 pixels with a pixel size of 0.14 μm x 0.14 μm were taken using Olympus Olyvia software and converted through ImageJ to binary RGB images using pre-determined threshold levels (low 25, high 255) to demonstrate areas of collagen fibres only, as previously described (Hadi et al., 2011). The proportion of area composed of collagen fibres within total sample area was then calculated.

Protein, hydroxyproline and collagen cross-link assays performed as previously described (Jones et al., 2018).

Parallel plate compression testing: performed as previously described (Jones et al., 2018).

Transmission electron microscopy: performed as previously described (Jones et al., 2018).

Atomic force microscopy nanoindentation imaging of individual non-hydrated collagen fibrils: performed as previously described (Jones et al., 2018).

Statistics

Statistical analyses were performed in GraphPad Prism v7.02 (GraphPad Software Inc, San Diego, CA) unless otherwise indicated. No data were excluded from the studies and for all experiments, all attempts at replication were successful. For each experiment, sample size reflects the number of independent biological replicates and is provided in the figure legend. Normality of distribution was assessed using the D’Agostino-Pearson normality test. Statistical analyses of single comparisons of two groups utilised Student’s t-test or Mann-Whitney U-test for parametric and non-parametric data respectively. Where appropriate, individual t-test results were corrected for multiple comparisons using the Holm-Sidak method. For multiple comparisons, one-way analysis of variance (ANOVA) with Dunnett’s multiple comparison test or Kruskal-Wallis analysis with Dunn’s multiple comparison test were used for parametric and non-parametric data, respectively. Results were considered significant if p < 0.05, where *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Appendix 1

Appendix 1—key resources table
Reagent type (species) or resourceDesignationSource or
reference
IdentifiersAdditional information
transfected construct (human)GAL4DBD-HIF1αCAD (residues 652–826)
for
HIF1α CAD reporter assay
Ratcliffe lab (University of
Oxford) Coleman et al., 2007
transfected construct (Homo-sapiens)UAS-luc reporter
for
HIF1α CAD reporter assay
Ratcliffe lab (University of
Oxford) Coleman et al., 2007
transfected construct (Homo-sapiens)Plasmid for Dual-Luciferase Reporter AssayPromegaphRL-CMV
transfected construct (Homo-sapiens)siRNA to human HIF1AN (FIH)Dharmacon/
Thermo Fisher Scientific
MU-004073-02-0002
transfected construct (Homo-sapiens)siRNA to human
HIF1A (HIF1α)
Dharmacon/
Thermo Fisher Scientific
MU-00401805-05-0002
transfected construct (Homo-sapiens)siRNA to human EPAS1(HIF2α)Dharmacon/
Thermo Fisher Scientific
MU-004814-01-0002
transfected construct (Homo-sapiens)siRNA to
human ARNT
(HIF1β)
Dharmacon/
Thermo Fisher Scientific
MU-007207-01-0002
transfected construct (Homo-sapiens)siGENOME
RISC-Free
Dharmacon/
Thermo Fisher Scientific
D-001220-01-05
AntibodyAnti-CAIX
(Rabbit
polyclonal)
Novus BiologicalsCat. #: NB100-417IHC 1:500
AntibodyAnti-HIF1A
(Rabbit polyclonal)
Cayman ChemicalCat. #: 10006421IHC 1:500
antibodyanti-human
HIF1α (Mouse
polyclonal IgG1k)
BD BiosciencesCat #:610,958WB (1:1000)
antibodyAnti-HIF1β
(Rabbit polyclonal)
Cell Signaling TechnologyCat #:5,537WB (1:1000)
antibodyAnti-phospho-
Smad2 (Rabbit polyclonal)
Cell Signaling TechnologyCat #: 3,104WB (1:1000)
antibodyanti-β-tubulin
(Mouse polyclonal)
Cell Signaling TechnologyCat #: 86,298WB (1:1000)
antibodyanti-PLOD2
(Mouse
monoclonal
IgG2B)
R&D SystemsCat #: MAB4445WB (1:500)
antibodyAnti- human
LOXL2 (Goat polyclonal)
R&D SystemsCat #: AF2639WB (1:1000)IF (1:100)
antibodyanti-human FIH
(Mouse monoclonal 162 C)
Ratcliffe lab (University of
Oxford) Stolze et al., 2004
WB (1:200)
AntibodyAnti-P-ERK
(polyclonal rabbit Thr202/Tyr204)
Cell Signalling TechnologyCat #: 9,101WB (1:1000)
AntibodyAnti-P-SMAD2/3
(Rabbit polyclonal Ser465/467)
Cell Signalling TechnologyCat #: 8,828WB (1:1000)
AntibodyIRDye 800CW
Donkey anti-Goat
IgG Secondary
Antibody
LI-COR
Biosciences
Cat #: 926–32214WB (1:5000)
AntibodyIRDye 800CW Goat anti-Rabbit IgG Secondary
Antibody
LI-COR
Biosciences
Cat #: 926–32211WB (1:5000)
AntibodyIRDye 680LT Goat
anti-Mouse IgG Secondary
Antibody
LI-COR
Biosciences
Cat #: 926–68020WB (1:5000)
AntibodyAnti-non-phospho (active) β-catenin (Rabbit
monoclonal IgG)
Cell Signalling TechnologyCat. #: 8,814 SWB (1:1000)
AntibodyAnti-mouse IgG
HRP-linked
whole antibody
Life SciencesCat. #: NXA931WB (1:1000)
AntibodyAnti-goat Immunoglobulins/
HRP (affinity
isolated)
DakoCat. #: P0449WB (1:1000)
AntibodyAnti-goat IgG H&L (Alexa Fluor 647)AbcamCat. #: Ab150131ICC 1:100
sequence-based reagentHuman HIF1A (HIF1α)QiagenQuantiTect PCR
primersCat #: QT00083664
sequence-based reagentHuman EPAS1 (HIF2α)QiagenQuantiTect PCR
primersCat #: QT00069587
sequence-based reagentHuman ARNT (HIF1β)QiagenQuantiTect PCR
primersCat #: QT00023177
sequence-based reagentHuman ACTB(β-actin)QiagenQuantiTect PCR
primersCat #: QT01680476
Sequence-based reagentLOXL2Primer Design
Sequence-based reagentPLOD2Primer Design
Sequence-based reagentCOL1A1Primer Design
Sequence-based reagentCOL3A1Primer Design
Peptide,
recombinant
protein
Recombinant
Human TGF-
beta 1 Protein
R&D SystemsCat. #: 240-B-010
Peptide,
recombinant
protein
Recombinant
Human EGF
GMP Protein
R&D SystemsCat. #: 236-GMP-200
Commercial
assay or kit
RNAscope 2.5
HD Duplex Assay
Advanced Cell DiagnosticsCat. #: 322,430
Commercial
assay or kit
RNAscope probe-
Hs-LOXL2-C1
Advanced Cell DiagnosticsCat. #: 311,341
Commercial
assay or kit
RNAscope probe-
Hs-PLOD2-C2
Advanced Cell DiagnosticsCat. #: 547761-C2
Commercial
assay or kit
MicroVue Bone
PYD EIA
QuidelCat. #: 8,010
Commercial
assay or kit
Hydroxyproline
Assay Kit
MerckCat. #: MAK008
Commercial
assay or kit
Total Protein
Assay
QuickZyme BiosciencesCat. #: QZBtotprot
Commercial assay or kitPicro Sirius Red
Stain Kit
(Connective
Tissue Stain)
AbcamCat. #: Ab150681
commercial assay or kitLipofectamine 3,000Thermo Fisher ScientificCat. #: L3000008
commercial assay or kitLipofectamine RNAiMAXThermo Fisher ScientificCat. #: 13778–075
commercial assay or kitDual-Luciferase Reporter Assay
System
PromegaCat. #: E1910
commercial assay or kitQuantiNova SYBR Green RT-PCR kitsQiagenCat. #: 208052
chemical compound, drugDimethyloxaloylglycine (DMOG)Sigma AldrichCat #: D3695CAS: 89464-63-1
chemical compound, drugN-[[1,2-Dihydro-4-hydroxy-2-oxo-1-(phenylmethyl)–3-quinolinyl]carbonyl]-glycine (IOX2)Selleck ChemicalsCat #: S2919CAS: 931398-72-0
chemical compound, drugDimethyl N-oxalyl-D-phenylalanine (DM-NOFD)Schofield lab (University of Oxford) McDonough et al., 2005
chemical compound, drugDMSOSigma AldrichCat #: 276,855CAS: 67-68-5
chemical compound, drugT-hydro (tert-butyl hydroperoxide)Sigma AldrichCat #: 19,999CAS: 75-91-220 µM (fresh prepared)

Data availability

All data generated or analysed during this study are included in the manuscript and supporting files.

The following previously published data sets were used
    1. Thannickal VJ
    2. Chanda D
    (2017) NCBI Gene Expression Omnibus
    ID GSE73854. Developmental programming in Idiopathic pulmonary fibrosis (IPF).
    1. Vukmirovic M
    2. Jones MG
    3. Kaminski N
    (2021) NCBI Gene Expression Omnibus
    ID GSE169500. Spatial transcriptome profiling identifies CREB1 as a regulator of core transcriptional programs in idiopathic pulmonary fibrosis.

References

    1. Loncaster JA
    2. Harris AL
    3. Davidson SE
    4. Logue JP
    5. Hunter RD
    6. Wycoff CC
    7. Pastorek J
    8. Ratcliffe PJ
    9. Stratford IJ
    10. West CM
    (2001)
    Carbonic anhydrase (CA IX) expression, a potential new intrinsic marker of hypoxia: correlations with tumor oxygen measurements and prognosis in locally advanced carcinoma of the cervix
    Cancer Research 61:6394–6399.
    1. Qian F
    2. He M
    3. Duan W
    4. Mao L
    5. Li Q
    6. Yu Z
    7. Zhou Z
    8. Zhang Y
    (2015)
    Cross regulation between hypoxia-inducible transcription factor-1α (HIF-1α) and transforming growth factor (TGF)-ß1 mediates nickel oxide nanoparticles (NiONPs)-induced pulmonary fibrosis
    American Journal of Translational Research 7:2364–2378.
  1. Book
    1. Russell S
    2. Gatenby RA
    3. Gillies RJ
    4. Ibrahim-Hashim A
    (2017)
    Chapter 4 - Pseudohypoxia: Life at the Edge
    In: Ujvari B, Roche B, Thomas F, editors. Ecology and Evolution of Cancer. Academic Press. pp. 57–68.

Decision letter

  1. Gordana Vunjak-Novakovic
    Reviewing Editor; Columbia University, United States
  2. Paul W Noble
    Senior Editor; Cedars-Sinai Medical Centre, United States
  3. Gianni Carraro
    Reviewer; Cedars-Sinai, United States

Our editorial process produces two outputs: i) public reviews designed to be posted alongside the preprint for the benefit of readers; ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Pseudohypoxic HIF pathway activation dysregulates collagen structure-function in human lung fibrosis" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Paul Noble as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Gianni Carraro (Reviewer #3).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

The reviewers found your manuscript of broad interest to researchers interested in lung biology and especially to those focused on the pathogenesis and therapy of lung fibrosis, and find the data strong and convincing. They recognize that this study builds upon the previous original work of the authors, by identifying a pathway that regulates collagen nanostructure and stiffness in lung fibrosis and demonstrating that this pathway it is independent of pathways regulating collagen synthesis. They also valued the elegant analysis you performed to validate the specificity of experimental finding, and the demonstration that HIF activation is required for the increased tissue stiffness associated with fibrosis. They also raise several important points and constructive suggestions we would like you to address before we can make decision about the suitability of your manuscript for publication in eLife.

Essential revisions:

1) Although the authors consider the fact that LOX and LOXL1 are also increased, and justify their focus on LOXL2 by its correlation with PLOD2 expression, it may not be justified to completely ignore LOX and LOXL1. The reported data do not rule out a role for LOX and LOXL1.

2) The authors need to show direct involvement of LOXL2 and PLOD2, e.g. using siRNA to revert the stiffness phenotype. They also should show co-localization of LOXL2 and PLOD2 in areas of COL1A1 and COL 3A deposition), as there is a synergistic effect between TGF β and HIF pathways in IPF foci.

3) Measurements of the fiber width, length and density quantification would increase the confidence into the morphological differences shown in picrosirius red staining images.

4) The authors should show criteria and characterization of the healthy and diseased donors of primary lung fibroblasts used in the experiments.

5) The mRNA data should be confirmed with immunostaining/WB as detailed in the reviews.

Reviewer #2 (Recommendations for the authors):

Figure 1h. PLOD 2 should be counterstained with LOXL2 in tissues with IPF foci and show high expression level versus non foci area in same samples and healthy lung tissue samples. Lower magnification images (together with high magnification quadrants) should be presented to show that overexpression is more than in few isolated spots.

Figure 2. b, c, e show some discrepancies in levels of LOXL2. Panel b has the highest level of LOXL2 mRNA at 72hrs, while the western blot in c showed only minimal increase of LOXL2 protein at same time point (72 hrs) compared to other conditions. What happens at 24 and 48 hrs at protein level? In panel E, the immunostaining for LOXL2 at 24hrs shows a notable difference with controls. This is not what would be expected based on the WB data at 72hrs. Authors needs to show WB and immunostaining data at same time point to support the quality of their data using different methodologies and support their conclusions. PLOD 2 immunostaining needs to be included in this figure to support the role of both PLOD2 and LOXL2 in the fibrotic phenotype.

Figure 3. If HIF levels are supposedly higher in IPF fibroblasts, authors should compare them to the levels of HIF in healthy fibroblasts and show how much they can be reduced in both samples (healthy versus IPF) by siRNA.

Figure 4A- B. These experiments show a significant synergistic effect of TGFbeta with DMOG and IOX2 on the protein levels of PLOD2, and none on LOXL2 levels. The data don't support the same magnitude effects shown in Figure 2 C. The authors need to do the same experiment as Figure 4 A and 4B using healthy and IPF fibroblasts (3 healthyversus 3 IPF donors) and compare them at 72 hrs as in Figure 2C, and clarify why 48hrs time point was used here. TGFbeta + DMOG appear to decrease the protein levels of LOXL2 compared to DMOG alone. These data are in disagreement with mRNA levels in Figure 4C. The authors need to explain in the text this difference and why.

Figure 5. To show a direct effect of PLOD2 and LOXL2 on pyridoline cross-linking and ECM stiffness, authors can do the same experiment as Figure 5 A and B with siRNA for PLOD2 and LOXL2 after IOX2 treatment and see whether this can revert the phenotype.

Figure 7. The claim of HIF/LOXL2/PLOD 2 colocalization needs to be supported by more images (quality and quantity). PLOD/LOXL2 constating is the same of Figure 2. Authors should provide co-staining of PLOD2/LOXL2 and HIF/CAIX in the same slide. If hard to obtain, PLOD2/LOXL2 with HIF and PLOD2/LOXL2 with CAIX costaining and show lower and higher magnification images and no few single events as in the actual figure.

Reviewer #3 (Recommendations for the authors):

Overall this manuscript will be an important contribution to the field of lung fibrosis. The experiments are well executed and include adequate number of biological replicates for statistical analysis. One aspect I would consider modifying is the interpretation that the identified effect of HIF is acting in non-hypoxic conditions. At least from the data presented it doesn't look obvious that this is the case. If there are no compelling evidence for this claim I would modify the narrative and include this as a possibility in the discussion.

Below are some points for improvement of the manuscript:

Figure 1: The quantification of the correlation of expression of PLOD2 and LOXL2 is properly executed and include comparison with other cross-linking enzyme to show specificity of the result. Figure 1h corroborate this result by showing co-localization by ISH, and it would be helpful to add an high magnification of a negative region to show specificity of the staining.

Figure 7: Try to show co-localization from serial sections. This does not allows to determine if the same or adjacent cells were probed. A more definitive way to achieve this would be to perform ISH with multiple fluorescent probes, using the same technique (RNAscope) that here was used with a chromogenic approach that limit the co-localization to two transcripts. In alternative RNAscope + immunofluorescence could also be combined for the same purpose.

The single cell data displayed in Figure 8 could also be used to corroborate cellular co-expression of PLOD2/LOXL2 and HIF1a in addition to the signature score that are provided.

Figure 8: From the code description it seems that a dimensional reduction was created first, followed by dataset subsetting to include only 'Control' and 'IPF' in the final figure. Did the authors tried to create a dimensional reduction after subsetting the 'Kropski' dataset and see if this can improve the segregation of the clusters of interest?

The color used in the tSNE plots in Figure 8a, 8c, S6a may be difficult to distinguish for color blind people. I suggest the use of alternative color palettes. For R, several packages are available for this purpose. The 'viridis' package for example use color blind accessible color scales.

The 'Cell Type' naming on top of the tSNE map is difficult to read, I suggest to adjust the position or to have it as a side legend.

https://doi.org/10.7554/eLife.69348.sa1

Author response

Essential revisions:

1) Although the authors consider the fact that LOX and LOXL1 are also increased, and justify their focus on LOXL2 by its correlation with PLOD2 expression, it may not be justified to completely ignore LOX and LOXL1. The reported data do not rule out a role for LOX and LOXL1.

We thank the reviewer for highlighting this comment which we have now further considered within the discussion (Page 14 Line 302 – Page 15 Line 315). The LOX(L) enzymes play key roles in the process of fibrillar collagen production and are tightly regulated in normal development and under physiological conditions (Trackman PC. Expert Opinion on Therapeutic Targets 2016;20:935–945). In our LCMD RNA-Seq analyses LOXL2 was the most highly expressed LOX/LOXL family member as well as the only LOX/LOXL member which correlated with PLOD2 expression, whilst in our previous work investigating collagen structure-function dysregulation in human lung fibrosis (eLife 2018;7:e36354) we identified that gene expression of LOXL2 was significantly increased in IPF tissue when compared to age-matched control lung tissue. Furthermore, using a small molecule LOXL inhibitor in our 3D model of fibrosis, we previously identified a significant reduction in mature pyridinoline cross-links using a concentration which completely inhibits LOXL2 but has minimal effects on LOX and LOXL1 (eLife 2018;7:e36354). This is consistent with previous reports that LOXL2 has key pathologic roles in cancer and fibrosis (Barker et al., Nature Reviews Cancer 2012;12:540–552; Barry-Hamilton et al., Nature Medicine 2010;16:1009–1017). Therefore, whilst we acknowledge within the discussion that our studies do not exclude a potential role for LOX and LOXL1 in human lung fibrosis, here we focussed upon the LOXL2/PLOD2 relationship.

2) The authors need to show direct involvement of LOXL2 and PLOD2, e.g. using siRNA to revert the stiffness phenotype. They also should show co-localization of LOXL2 and PLOD2 in areas of COL1A1 and COL 3A deposition), as there is a synergistic effect between TGF β and HIF pathways in IPF foci.

As requested, we tested small interfering RNA (siRNA) knockdown against LOXL2 within our long term 6 week 3D model of fibrosis. Unfortunately, LOXL2 knockdown was not sustained throughout the 6 week culture (Author response image 1a), although the knockdown was sufficient to significantly reduce tissue stiffness in the presence of IOX2 (Author response image 1b). This is a finding consistent with our previous study using a small molecule LOXL inhibitor in our 3D model of fibrosis (eLife 2018;7:e36354) where we demonstrated a significant reduction in tissue stiffness using a concentration which completely inhibits LOXL2/3 but not other LOX(L) family members. We did not feel it was appropriate to include these data as we were unable to assess the impact of long-term silencing upon tissue stiffness, therefore we acknowledge within the discussion that an approach such as CRISPR would be required for persistent gene silencing and that this would be a potential area for future study (Page 15 Lines 313-315). If the Reviewers felt that including these data was helpful we would be willing to include them within the Supplementary Results.

Author response image 1

In our previous studies, through human lung tissue analyses and in vitro studies we identified that lung fibrosis tissue stiffness is significantly determined by pyridinoline collagen cross-linking. PLOD2 encodes the lysyl hydroxylase 2 (LH2) enzyme, which is responsible for the hydroxylation of lysine residues in fibrillar collagen telopeptides. Telopeptide hydroxylysines are essential for the hydroxyallysine pathway of cross-linking, and this is required for the production of all mature lysylpyridinoline and hydroxylysylpyridinoline cross-links in extracellular collagen fibrils (van der Slot AJ et al., J Biol Chem. 2003;278:40967–72 and Gistelinck et al., J Bone Miner Res. 2016; 31: 1930–1942), with promotion of PLOD2 expression increasing tissue stiffness and pyridinoline cross-links (Chen et al., J Clin Invest. 2015;125:1147-62). Thus PLOD2 actitivty is essential for the presence of pyridinoline (hydroxylysine aldehyde-based) cross-links which we and others have demonstrated significantly determine tissue stiffness. In view of these considerations and the problems of maintaining silencing of LOXL2, we did not investigate siRNA PLOD2 knockdown within our current studies.To demonstrate co-localization of LOXL2 and PLOD2 in areas of fibrillar collagen (COL1A1 and COL3A1) deposition we performed Masson’s trichrome staining for collagen fibres on serial sections adjacent to those used for LOXL2/PLOD2 RNAScope, clearly identifying that LOXL2 and PLOD2 co-localize within areas of fibrillar collagen deposition (new Figure 1—figure supplement 1h and Page 6, Lines 94-95).

3) Measurements of the fiber width, length and density quantification would increase the confidence into the morphological differences shown in picrosirius red staining images.

As requested we have performed density (surface area quantification) of the Picrosirius red staining, identifying no difference in Picrosirius red surface area between conditions (new Figure 5—figure supplement 1c). In our analyses we performed Picrosirius red for qualitative comparison between conditions, however we did not perform detailed quantitative analyses for metrics such as collagen fiber width and length as, in our experience and those of others (Nazac et al., Microsc Res Tech. 2015;78:723-30. Latouf et al., Histochem Cytochem. 2014 Oct;62:751-8), there are potential limitations to the application of Picrosirius red imaging for quantitative fibrillar collagen analyses depending upon the tissue studied and imaging approach. Whilst picrosirius red staining is a well established histological detection method for binding to cationic collagen fibers, the utility for detailed quantitative assessment of collagen structure is less established and a requirement for validation for any specific tissue is recognised (Drifka et al., J Histochem Cytochem. 2016; 64: 519–529). We therefore utilised an accepted gold standard of transmission electron microscopy for our quantitative fibrillar collagen analyses.

4) The authors should show criteria and characterization of the healthy and diseased donors of primary lung fibroblasts used in the experiments.

As requested we have now included these data (new Supplementary File 1b). Across all experiments we used a total of 7 IPF lung fibroblast lines and 5 healthy lung fibroblast lines which were age, sex, and smoking history matched.

5) The mRNA data should be confirmed with immunostaining/WB as detailed in the reviews.

As requested we have included additional protein expression data. Within our 72 hour time course experiment (Figure 2a-d) of profibrogenic stimuli the 24 and 48hr protein expression levels of PLOD2 and LOXL2 were requested to complement the included mRNA data and 72hr protein expression (Figure 2a,b,c). We have now included western blotting data for LOXL2 and PLOD2 at each point within the time course (24,48 and 72hrs) (new Figure 2—figure supplement 1b) in the presence or absence of TGFβ1 or the hypoxia mimetic DMOG. Together our protein and gene expression data consistently identify that DMOG most strongly upregulates both PLOD2 and LOXL2 expression at each time point studied. To complement the western blot analyses of protein expression we performed additional immunofluorescent staining for PLOD2 and LOXL2 in the presence or absence of TGFb or the hypoxia mimetics IOX2, and DMOG (new Figure 2e), confirming by immunofluorescent staining an increase in intracellular LOXL2 and PLOD2 expression following DMOG or IOX2 treatment in comparison to treatment with TGFβ1.

Reviewer #2 (Recommendations for the authors):

Figure 1h. PLOD 2 should be counterstained with LOXL2 in tissues with IPF foci and show high expression level versus non foci area in same samples and healthy lung tissue samples. Lower magnification images (together with high magnification quadrants) should be presented to show that overexpression is more than in few isolated spots.

Co-localisation is present across fibroblast foci of diagnostic lung tissue for IPF patients (n=7) studied. The pattern of in situ hybridization was assessed by an expert histopathologist (AF) who confirmed consistent co-expression, and we have now additionally included a supplementary table of semi-quantitative analysis of PLOD2 and LOXL2 RNAScope expression within cell types in IPF tissue confirming expression is greatest with fibroblast foci (Supplementary File 1a) Additional examples including lower magnification images as well as high magnification images are also provided (new Figure 1-figure supplement 1f-h).

Figure 2. b, c, e show some discrepancies in levels of LOXL2. Panel b has the highest level of LOXL2 mRNA at 72hrs, while the western blot in c showed only minimal increase of LOXL2 protein at same time point (72 hrs) compared to other conditions. What happens at 24 and 48 hrs at protein level? In panel E, the immunostaining for LOXL2 at 24hrs shows a notable difference with controls. This is not what would be expected based on the WB data at 72hrs. Authors needs to show WB and immunostaining data at same time point to support the quality of their data using different methodologies and support their conclusions. PLOD 2 immunostaining needs to be included in this figure to support the role of both PLOD2 and LOXL2 in the fibrotic phenotype.

Please see the response to Essential Revision 5.

Figure 3. If HIF levels are supposedly higher in IPF fibroblasts, authors should compare them to the levels of HIF in healthy fibroblasts and show how much they can be reduced in both samples (healthy versus IPF) by siRNA.

Elevated levels of HIF1 and HIF2 in IPF fibroblasts under normoxic conditions have recently been reported (Aquino-Galvez, A. et al., Respir Res 20, 130 (2019)), and so in our studies we investigated possible mechanism for this. We focussed upon assessment of FIH activity, identifying using a luciferase reporter system that activity of FIH is decreased in lung fibroblasts from patients with IPF, so identifying a potential mechanism for increased HIF pathway activation in IPF. Consistent with this we identified in publicly available datasets that cultured fibroblasts from patients with a usual interstitial pneumonia pattern of fibrosis have a significantly increased HIF score compared to cultured control fibroblasts.

Figure 4A- B. These experiments show a significant synergistic effect of TGFbeta with DMOG and IOX2 on the protein levels of PLOD2, and none on LOXL2 levels. The data don't support the same magnitude effects shown in Figure 2 C. The authors need to do the same experiment as Figure 4 A and 4B using healthy and IPF fibroblasts (3 healthyversus 3 IPF donors) and compare them at 72 hrs as in Figure 2C, and clarify why 48hrs time point was used here. TGFbeta + DMOG appear to decrease the protein levels of LOXL2 compared to DMOG alone. These data are in disagreement with mRNA levels in Figure 4C. The authors need to explain in the text this difference and why.

As TGFβ1 strongly induced major collagen fibrillogenesis genes whilst HIF pathways most strongly induced PLOD2 and LOXL2 expression, in subsequent experiments we investigated the effects of activating these pathways individually or in combination using lung fibroblasts from patients with IPF. We selected the 48hr time point on the basis of our timecourse experiment (Figure 2), with all experiments performed using a minimum of 3 IPF cell lines. The effect of DMOG in the absence or presence of TGFβ1 upon PLOD2 and LOXL2 induction (Figure 4a-c) was comparable to that identified using normal control lung fibroblasts. When HIF stabilisation (DMOG or IOX2) and TGFβ1 were combined, a synergistic effect upon the induction of PLOD2 gene and protein expression was identified which was greater than either pathway alone (new Figure 4a and c). Whilst the gene expression of LOXL2 was also increased with the combination of HIF stabilisation and TGFβ1 a corresponding increase in protein expression within cell lysates was not apparent. As LOXL2 is processed intracellularly before being extracellularly secreted we therefore investigated whether increased secretion of LOXL2 was occurring, identifying that under conditions with HIF stabilisation in the absence or presence of TGFβ1, LOXL2 secretion was increased (new Figure 4d; new Figure 4—figure supplement 1).

Figure 5. To show a direct effect of PLOD2 and LOXL2 on pyridoline cross-linking and ECM stiffness, authors can do the same experiment as Figure 5 A and B with siRNA for PLOD2 and LOXL2 after IOX2 treatment and see whether this can revert the phenotype.

Please see the response to Essential Revision 3.

Figure 7. The claim of HIF/LOXL2/PLOD 2 colocalization needs to be supported by more images (quality and quantity). PLOD/LOXL2 constating is the same of Figure 2. Authors should provide co-staining of PLOD2/LOXL2 and HIF/CAIX in the same slide. If hard to obtain, PLOD2/LOXL2 with HIF and PLOD2/LOXL2 with CAIX costaining and show lower and higher magnification images and no few single events as in the actual figure.

Our findings were consistent across diagnostic tissue blocks and included independent histopathology review. We have included additional images (new Figure 7—figure supplement 1) demonstrating co-localisation of LOXL2/PLOD2 together with serial sections demonstrating the expression of CAIX and HIF1A.

Reviewer #3 (Recommendations for the authors):

Overall this manuscript will be an important contribution to the field of lung fibrosis. The experiments are well executed and include adequate number of biological replicates for statistical analysis. One aspect I would consider modifying is the interpretation that the identified effect of HIF is acting in non-hypoxic conditions. At least from the data presented it doesn't look obvious that this is the case. If there are no compelling evidence for this claim I would modify the narrative and include this as a possibility in the discussion.

Thank you for this important comment. As described in our comments to the Editor, we now provide additional data demonstrating that selective FIH inhibition is sufficient to promote tissue stiffness.

Below are some points for improvement of the manuscript:

Figure 1: The quantification of the correlation of expression of PLOD2 and LOXL2 is properly executed and include comparison with other cross-linking enzyme to show specificity of the result. Figure 1h corroborate this result by showing co-localization by ISH, and it would be helpful to add an high magnification of a negative region to show specificity of the staining.

As requested we have provided additional images including a high magnification of a negative region to show specificity of the staining (new Figure 1—figure supplement 1f).

Figure 7: Try to show co-localization from serial sections. This does not allows to determine if the same or adjacent cells were probed. A more definitive way to achieve this would be to perform ISH with multiple fluorescent probes, using the same technique (RNAscope) that here was used with a chromogenic approach that limit the co-localization to two transcripts. In alternative RNAscope + immunofluorescence could also be combined for the same purpose.

Thank you for this suggestion. We agree that fluorescent probes could be a helpful alternative, however here we have used IHC as our analysis pathway; this included formal histopathologist review for which the histopathologist requested this approach. We have included additional images (new Figure 7—figure supplement 1) demonstrating co-localisation of LOXL2/PLOD2 together with serial sections of CAIX and HIF1A.

The single cell data displayed in Fig8 could also be used to corroborate cellular co-expression of PLOD2/LOXL2 and HIF1a in addition to the signature score that are provided.

Thank you for this suggestion. As the HIF1a gene is constitutively expressed we felt that the HIF signature score would be an appropriate methodology to investigate this.

Figure 8: From the code description it seems that a dimensional reduction was created first, followed by dataset subsetting to include only 'Control' and 'IPF' in the final figure. Did the authors tried to create a dimensional reduction after subsetting the 'Kropski' dataset and see if this can improve the segregation of the clusters of interest?

Thank you for your suggestion. As suggested dimensional reduction was performed after subsetting the ‘Kropski’ dataset and we have now included this within the final figure and updated the uploaded code (new Figure 8; new Figure 8—figure supplement 1). Visually this did not improve segregation of the clusters of interest.

The color used in the tSNE plots in Figure 8a, 8c, S6a may be difficult to distinguish for color blind people. I suggest the use of alternative color palettes. For R, several packages are available for this purpose. The 'viridis' package for example use color blind accessible color scales.

Thank you for your suggestion. We have changed the plots to use an alternative color palette using the ‘viridis’ package as suggested.

The 'Cell Type' naming on top of the tSNE map is difficult to read, I suggest to adjust the position or to have it as a side legend.

As requested we have modified this to aid readability.

https://doi.org/10.7554/eLife.69348.sa2

Article and author information

Author details

  1. Christopher J Brereton

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Writing – original draft, Writing – review and editing
    Contributed equally with
    Liudi Yao
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8302-702X
  2. Liudi Yao

    Biological Sciences, Faculty of Environmental and Life Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review and editing
    Contributed equally with
    Christopher J Brereton
    Competing interests
    No competing interests declared
  3. Elizabeth R Davies

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    3. Biological Sciences, Faculty of Environmental and Life Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Formal analysis, Investigation, Methodology, Project administration, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8629-8324
  4. Yilu Zhou

    1. Biological Sciences, Faculty of Environmental and Life Sciences, University of Southampton, Southampton, United Kingdom
    2. Institute for Life Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
  5. Milica Vukmirovic

    1. Section of Pulmonary, Critical Care and Sleep Medicine, Department of Medicine, Yale University School of Medicine, New Haven, United States
    2. Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, Canada
    Contribution
    Formal analysis, Investigation, Methodology, Resources, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
  6. Joseph A Bell

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    Contribution
    Formal analysis, Methodology, Resources, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
  7. Siyuan Wang

    Biological Sciences, Faculty of Environmental and Life Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    No competing interests declared
  8. Robert A Ridley

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    Contribution
    Formal analysis, Investigation, Methodology, Project administration, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
  9. Lareb SN Dean

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    Contribution
    Formal analysis, Investigation, Methodology, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8703-9236
  10. Orestis G Andriotis

    Institute of Lightweight Design and Structural Biomechanics, TU Wien, Vienna, Austria
    Contribution
    Investigation, Methodology, Writing – review and editing
    Competing interests
    No competing interests declared
  11. Franco Conforti

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    Contribution
    Investigation, Methodology, Writing – review and editing
    Competing interests
    No competing interests declared
  12. Lennart Brewitz

    Department of Chemistry and the Ineos Oxford Institute for Antimicrobial Research, Chemistry Research Laboratory, Oxford, United Kingdom
    Contribution
    Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  13. Soran Mohammed

    School of Chemistry, University of Southampton, Southampton, United Kingdom
    Contribution
    Investigation, Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3882-6129
  14. Timothy Wallis

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    Contribution
    Investigation, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  15. Ali Tavassoli

    School of Chemistry, University of Southampton, Southampton, United Kingdom
    Contribution
    Investigation, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  16. Rob M Ewing

    1. Biological Sciences, Faculty of Environmental and Life Sciences, University of Southampton, Southampton, United Kingdom
    2. Institute for Life Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  17. Aiman Alzetani

    1. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    2. University Hospital Southampton, Southampton, United Kingdom
    Contribution
    Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  18. Benjamin G Marshall

    1. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    2. University Hospital Southampton, Southampton, United Kingdom
    Contribution
    Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-0946-0399
  19. Sophie V Fletcher

    1. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    2. University Hospital Southampton, Southampton, United Kingdom
    Contribution
    Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  20. Philipp J Thurner

    Institute of Lightweight Design and Structural Biomechanics, TU Wien, Vienna, Austria
    Contribution
    Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  21. Aurelie Fabre

    Department of Histopathology, St. Vincent's University Hospital & UCD School of Medicine, University College Dublin, Dublin, Ireland
    Contribution
    Investigation, Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  22. Naftali Kaminski

    Section of Pulmonary, Critical Care and Sleep Medicine, Department of Medicine, Yale University School of Medicine, New Haven, United States
    Contribution
    Resources, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5917-4601
  23. Luca Richeldi

    1. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    2. Unità Operativa Complessa di Pneumologia, Università Cattolica del Sacro Cuore, Fondazione Policlinico A. Gemelli IRCCS, Rome, Italy
    Contribution
    Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  24. Atul Bhaskar

    Faculty of Engineering and Physical Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  25. Christopher J Schofield

    Department of Chemistry and the Ineos Oxford Institute for Antimicrobial Research, Chemistry Research Laboratory, Oxford, United Kingdom
    Contribution
    Resources, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0290-6565
  26. Matthew Loxham

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    3. Institute for Life Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Investigation, Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6459-538X
  27. Donna E Davies

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    3. Institute for Life Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Supervision, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-5117-2991
  28. Yihua Wang

    1. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    2. Biological Sciences, Faculty of Environmental and Life Sciences, University of Southampton, Southampton, United Kingdom
    3. Institute for Life Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing – original draft, Writing – review and editing
    For correspondence
    yihua.wang@soton.ac.uk
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5561-0648
  29. Mark G Jones

    1. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
    3. Institute for Life Sciences, University of Southampton, Southampton, United Kingdom
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing – original draft, Writing – review and editing
    For correspondence
    mark.jones@soton.ac.uk
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6308-6014

Funding

Wellcome Trust (100638/Z/12/Z)

  • Mark Jones

Medical Research Council (MR/S025480/1)

  • Yihua Wang

Academy of Medical Sciences (SBF002\1038)

  • Yihua Wang

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

This project was supported by Medical Research Council (MR/S025480/1), the Wellcome Trust (100638/Z/12/Z), an Academy of Medical Sciences/the Wellcome Trust Springboard Award [SBF002\1,038], and the AAIR Charity. CJB and LSND acknowledge the support of the NIHR Southampton Biomedical Research Centre. LY was supported by China Scholarship Council. YZ was supported by an Institute for Life Sciences PhD Studentship. FC was supported by Medical Research Foundation [MRF-091–0003-RG-CONFO]. ML was supported by a BBSRC Future Leader Fellowship [BB/PO11365/1] and a NIHR Southampton Biomedical Research Centre Senior Research Fellowship. We thank Carine Fixmer, Maria Lane, Benjamin Johnson, and the nurses of the Southampton Biomedical Research Unit for their help in the collection of human samples, supported by the Wessex Clinical Research Network and the National Institute of Health Research, UK. We also thank Dr. Tammie Bishop (University of Oxford) for her technical support in IHC and Prof Sir Peter Ratcliffe (University of Oxford) for the FIH antibody (mouse monoclonal 162 C) and the UAS-luc/GAL4DBD-HIF1αCAD binary reporter system.

Ethics

Human subjects: All human lung experiments were approved by the Southampton and South West Hampshire and the Mid and South Buckinghamshire Local Research Ethics Committees (ref 07/H0607/73), and all subjects gave written informed consent.

Senior Editor

  1. Paul W Noble, Cedars-Sinai Medical Centre, United States

Reviewing Editor

  1. Gordana Vunjak-Novakovic, Columbia University, United States

Reviewer

  1. Gianni Carraro, Cedars-Sinai, United States

Publication history

  1. Received: May 11, 2021
  2. Preprint posted: May 12, 2021 (view preprint)
  3. Accepted: January 4, 2022
  4. Version of Record published: February 21, 2022 (version 1)

Copyright

© 2022, Brereton et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 784
    Page views
  • 101
    Downloads
  • 1
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Christopher J Brereton
  2. Liudi Yao
  3. Elizabeth R Davies
  4. Yilu Zhou
  5. Milica Vukmirovic
  6. Joseph A Bell
  7. Siyuan Wang
  8. Robert A Ridley
  9. Lareb SN Dean
  10. Orestis G Andriotis
  11. Franco Conforti
  12. Lennart Brewitz
  13. Soran Mohammed
  14. Timothy Wallis
  15. Ali Tavassoli
  16. Rob M Ewing
  17. Aiman Alzetani
  18. Benjamin G Marshall
  19. Sophie V Fletcher
  20. Philipp J Thurner
  21. Aurelie Fabre
  22. Naftali Kaminski
  23. Luca Richeldi
  24. Atul Bhaskar
  25. Christopher J Schofield
  26. Matthew Loxham
  27. Donna E Davies
  28. Yihua Wang
  29. Mark G Jones
(2022)
Pseudohypoxic HIF pathway activation dysregulates collagen structure-function in human lung fibrosis
eLife 11:e69348.
https://doi.org/10.7554/eLife.69348

Further reading

    1. Cell Biology
    Rebecca G Wells
    Insight

    Oxidative stress following a lung injury can alter the structure of collagen, causing it to stiffen and trigger the formation of a fibrotic scar that further hardens the tissue.

    1. Cell Biology
    Dillon Jevon et al.
    Research Article Updated

    A developing understanding suggests that spatial compartmentalisation in pancreatic β cells is critical in controlling insulin secretion. To investigate the mechanisms, we have developed live-cell subcellular imaging methods using the mouse organotypic pancreatic slice. We demonstrate that the organotypic pancreatic slice, when compared with isolated islets, preserves intact β-cell structure, and enhances glucose-dependent Ca2+ responses and insulin secretion. Using the slice technique, we have discovered the essential role of local activation of integrins and the downstream component, focal adhesion kinase (FAK), in regulating β cells. Integrins and FAK are exclusively activated at the β-cell capillary interface and using in situ and in vitro models we show their activation both positions presynaptic scaffold proteins, like ELKS and liprin, and regulates glucose-dependent Ca2+ responses and insulin secretion. We conclude that FAK orchestrates the final steps of glucose-dependent insulin secretion within the restricted domain where β-cell contact the islet capillaries.