|Home | About | Journals | Submit | Contact Us | Français|
To identify the pathways involved in adult lung regeneration, we have employed left unilateral pneumonectomy (PNX) model that promotes regenerative alveolarization in the remaining intact right lung lobes. Here, we show that PNX stimulates pulmonary capillary endothelial cells (PCECs) to produce paracrine (angiocrine) growth factors that induce proliferation of epithelial progenitor cells supporting alveologenesis. After PNX, endothelial-specific inducible genetic ablation of Vegfr2 and Fgfr1 in mice inhibited production of MMP14 impairing alveolarization. MMP14 via unmasking cryptic EGF-like ectodomain and activation of EGF-receptor (EGFR) expands epithelial progenitor cells. Neutralization of MMP14 impaired EGFR-ligand mediated alveolar regeneration. By contrast, administration of recombinant EGF, or intravascular transplantation of MMP14+ PCECs from wild-type mice, into pneumonectomized Vegfr2/Fgfr1 deficient mice restored alveologenesis and lung inspiratory volume and compliance function. This study shows that VEGFR2 and FGFR1 activation in PCECs by increasing MMP14-dependent bioavailability of EGFR-ligands initiates and sustains alveologenesis and holds promise to develop therapeutic strategies to promote lung regeneration.
Defining the cellular and molecular mechanisms that modulate adult lung regeneration is essential to develop strategies to treat respiratory disorders (Beers and Morrisey, 2011; Chapman, 2011; Metzger et al., 2008; Morrisey and Hogan, 2010; Sheppard, 2003; Warburton et al., 2010; Whitsett et al., 2010). To identify the regulatory mechanisms involved in adult lung regeneration, we employed a model in which surgical removal of the left lung, known as left unilateral pneumonectomy (PNX), induces the expansion of mass and volume in the intact lobes of the remaining right lungs (Cowan and Crystal, 1975; Nolen-Walston et al., 2008). This regenerative process is driven by alveologenesis, a process dependent on proliferation of lung epithelial progenitor cells (Cardoso, 2001; Kotton and Fine, 2008; Rock and Hogan, 2011; Stripp and Reynolds, 2008), which comprise of subsets of alveolar epithelial cells (AECs) (Chapman et al., 2011; Liu et al., 2011) and presumably bronchioalveolar stem cells (BASCs) (Kim et al., 2005; Zhang et al., 2008). However, the precise mechanism(s) by which PNX initiates and sustains regenerative alveologenesis is unknown.
During lung development, the vascular plexus (capillary) sprouts in parallel with the alveolar budding (Metzger et al., 2008; White et al., 2007; Yamamoto et al., 2007), raising the possibility that regeneration of the pulmonary capillary plays a key role in orchestrating regenerative alveolarization. As a unique organ that facilitates gas exchange, the lung alveolus is highly vascularized, with pulmonary capillary endothelial cells (PCECs) lining all alveoli and residing in cellular proximity to AECs (Bhattacharya, 2005; Komarova and Malik, 2010; Petrache et al., 2005). The reconstitution of the alveolar-capillary interface is pivotal for pulmonary gas exchange function (Giordano et al., 2008; Huh et al., 2010; Petersen et al., 2010; Vaporciyan et al., 1993). However, the role of PCECs, as a specialized capillary vasculature in guiding alveolarization (Bhattacharya, 2005; DeLisser et al., 2006; Ding et al., 2003; Oh et al., 2007), in particular during regenerative alveolar remodeling (Metzger et al., 2008), remains unknown.
Capillary endothelial cells (ECs) that form the building blocks of the microvasculature of individual organs are endowed with unique organ-specific phenotypic and functional attributes (Aird, 2007; Carmeliet, 2005; Red-Horse et al., 2007; Ruoslahti and Rajotte, 2000). Capillary ECs are not solely passive conduits for the delivery of oxygen or nutrients, but also through elaboration of tissue-specific paracrine growth factors, defined as angiocrine factors (Butler et al., 2010a; Butler et al., 2010b), support organ development (Lammert et al., 2001; Matsumoto et al., 2001; Sakaguchi et al., 2008) and adult organ regeneration.
For example, sinusoidal endothelial cells (SECs) within liver and bone marrow comprise of phenotypically and functionally discreet populations of organ-specific endothelial cells. We have shown that after partial hepatectomy, liver SECs (LSECs) through a process of “inductive angiogenesis” that via angiocrine production of hepatocyte growth factor and Wnt2 stimulates hepatocyte proliferation (Ding et al., 2010). Subsequently, LSECs undergo “proliferative (sprouting) angiogenesis” to meet the incremental demand in the blood supply to the regenerating liver tissue. Similarly, after chemotherapy and irradiation activated bone marrow SECs reconstitute hematopoiesis by angiocrine expression of Notch ligands and IGFBPs (Butler et al., 2010a, Kobayashi et al., 2010). Conditional selective deletion of VEGF-A receptor-2 (VEGFR2) in either LSECs (Ding et al., 2010) or SECs (Hooper et al., 2009) of the adult mice by impairing the production of angiocrine factors inhibited liver or bone marrow regeneration, underscoring the physiological importance of endothelial-derived instructive signals in the adult organ regeneration. These findings have raised the possibility that PCECs also compose of a functionally unique population of specialized organ-specific ECs, which by production of specific angiocrine factors, could also induce regenerative alveolarization.
During development, PCECs undergo extensive sprouting angiogenesis to vascularize alveoli within the expanding lung tissue (Alvarez et al., 2008; Bhattacharya, 2005). In addition to their enduring capacity to undergo proliferative angiogenesis, PCECs by production of paracrine factors specify endoderm and mesoderm progenitors into primitive lung epithelial and vascular precursor cells (Del Moral et al., 2006; Healy et al., 2000; Shu et al., 2002; Voelkel et al., 2006; Yamamoto et al., 2007). These studies suggest that PCECs have the potential of producing as yet unrecognized angiocrine growth signals that support alveologenesis. Whether PCEC-derived instructive signals could also trigger regenerative alveolarization in the adult lungs has not been studied. Indeed, the paucity of mouse lung regenerative genetic models and lack of operational definition of PCECs have handicapped studies of PCECs in guiding alveolar regeneration in adult lungs.
In this study, we have defined the phenotypic and operational markers of mouse PCEC population as VE-cadherin+VEGFR2+FGFR1+CD34+ endothelial cells. Then, we have employed a unilateral PNX model to investigate the role of PCECs in supporting alveolar regeneration. In a variety of species, surgical resection of the left lung, which does not perturb the vascular integrity of the remaining right lobes, induces dramatic regrowth of these residual lobes. Here, we demonstrate that PNX through activation of VEGFR2 and FGFR1 induces PCECs of the remaining right lobe to produce the angiocrine factor, MMP14. In turn, MMP14 by unmasking cryptic epidermal growth factor (EGF)-like ligands promotes regenerative alveolarization through stimulating the proliferation of epithelial progenitor cells. These data suggest that PCECs could be therapeutically exploited for the treatment of lung disorders.
Within 15 days after PNX, there is a dramatic regeneration in both the mass and volume of the remaining lobes of right lungs (Fig. 1A, B). Lung epithelial progenitor cells, including subsets of BASC population identified phenotypically by Clara cell secreted protein (CCSP)+ and pro-surfactant protein C (SPC)+Sca-1+ (CCSP+SPC+Sca-1+ cells) and type II AECs (AECIIs) by SPC+E-cadherin+ cells have been shown to contribute to alveolar epithelialization (Beers and Morrisey, 2011). Therefore, to determine the contribution of epithelial progenitor cells to lung regeneration immediately after PNX, we introduced BrdU in the drinking water to detect slow cycling cells at day 1 to 7 after PNX. On day 3 after PNX, we observed amplification of BrdU+CCSP+ cells at bronchioalveolar duct junction (BADJ) (Fig. 1C). To track the expansion of BrdU+CCSP+ cells after PNX, we used reporter transgenic mice in which CCSP and SPC promoters drive the expression of YFP (CCSP-YFP and SPC-YFP mice) (Perl et al., 2002) (Fig. 1D, E). We performed polyvariate flow cytometric analysis of all mononuclear cell populations within the regenerating lungs on day 3 after PNX. The CCSP+BrdU+ cells localized to the BADJ region were found to be CCSP+SPC+Sca-1+VE-cadherin−CD31− cells, a phenotypic signature observed on BASCs (Kim et al., 2005). At this early time point, we did not detect any significant proliferation of SPC+Sca-1−CCSP− AECIIs or VE-cadherin+CD31+ PCECs.Therefore, PNX induces expansion of slow cycling CCSP+SPC+Sca-1+ BASC-like cells in early phases of lung regeneration, when there is minimal proliferation of AECs and PCECs.
To identify time points after PNX whereby AECs and PCECs undergo significant proliferation, we examined the kinetics of pulmonary incorporation of intraperitoneally injected BrdU, which revealed a global appearance of transit amplifying cells (TACs) in the remaining lung lobes that peaked at day 7 after PNX (Fig. 2A., Fig. S1). In the sham-operated mouse lung, there was little uptake of BrdU. To characterize the cell types among TACs on day 7 after PNX, we performed PNX on SPC-YFP reporter mice. There was increased proliferation of SPC+ cells that co-express pro-surfactant protein D (SPD) and E-cadherin, markers representing AECIIs (Beers et al., 1994) (Fig.2B). The remaining SPC− TAC population consists of VE-cadherin+ PCECs and a small fraction of CCSP+ airway Clara cells (Rawlins et al., 2009). These data indicate that after PNX, BASC-like cells expand at early time point (day 3), while AECIIs proliferate at later time points.
Analysis of BrdU incorporation showed that on day 7 after PNX, proliferating VE-cadherin+ CD34+FGFR1+VEGFR2+CD45− PCEC saccounted for 7% of lung mononuclear cells (Fig.2C) that were detected in the cellular vicinity to SPC+ AECIIs (Fig. 2D). Using SPC+E-cadherin+ and VE-cadherin+CD34+ as operational markers for AECIIs and PCECs, respectively, we found that on day 15 after PNX there was a ~3-fold increase in the population of both AECIIs and PCECs (Fig. 2E). Therefore, after PNX the increase in lung mass and volume is primarily due to proliferation of epithelial progenitor cells and PCECs.
One mechanism by which PNX initiates lung regeneration could be mediated through activation of PCECs to produce epithelial-active angiocrine factors. As VEGFR2, the principal tyrosine kinase receptor of VEGF-A, plays a critical role in induction of angiocrine factors (Ding et al., 2010; Hooper et al., 2009), we analyzed the activation of VEGFR2 in PCECs after PNX. Although VEGFR2 protein expression in PCECs is unaltered, after PNX there is a significant increase in the level of phosphorylated VEGFR2, indicating activation of this VEGF-A receptor (Fig. 3A).
Since FGFR1 is expressed in PCECs and can reciprocally modulate the expression and activation state of VEGFR2 to drive angiocrine factor production (Murakami et al., 2011; White et al., 2007), we also studied the expression of FGFR1 on PCECs. After PNX, FGFR1 protein was upregulated in a time-dependent manner. This finding suggests that while in early phases of lung regeneration, activation of VEGFR2 in PCECs plays a key role in inducing alveologenesis, at later phases, co-activation of FGFR1 might synergize with VEGFR2 to sustain epithelialization in the regenerating lungs.
To elucidate the endothelial-specific function of VEGFR2 and FGFR1 in the lungs, we employed an inducible knockout strategy to selectively delete Vegfr2 gene in adult mouse endothelial cells (Fig. 3B), using transgenic mice in which VE-cadherin promoter-drives the expression of tamoxifen-responsive Cre (VE-Cad-CreERT2) (Wang et al., 2010). Tamoxifen treatment selectively deletes Vegfr2 in endothelial cells (Vegfr2iΔEC/iΔEC mice). To account for the off-target toxicity by CreERT2, we used heterozygous Vegfr2 deficient (Vegfr2iΔEC/+) mice as control. We also generated mice in which both Vegfr2 and Fgfr1 were deleted in endothelial cells. However, these mice because of vascular instability could not tolerate surgical procedures. Therefore, we investigated the role of co-activation of FGFR1 and VEGFR2 in mediating alveologenesis by inducible Vegfr2 and partial Fgfr1 deletion in endothelial cells (Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice).
Before PNX, both Vegfr2iΔEC/iΔEC and Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice didn’t manifest alteration in lung mass and function (Fig. S2). By contrast, on day 3 after PNX, proliferation of CCSP+Sca1+ BASC-like cells was abolished in Vegfr2iΔEC/iΔEC mice (Fig. 3C, Fig. S2), while there was no further inhibition in expansion of these cells in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice after PNX. These data establish the critical role of VEGFR2 activation in supporting epithelial cell expansion at the early phases of lung regeneration.
We then studied the role of VEGFR2 and FGFR1 activation in the amplification of PCECs and AECIIs during lung regeneration. Co-staining of regenerating lungs with BrdU, VE-cadherin, and SPC at day 7 after PNX indicated that endothelial-specific knockdown of Vegfr2 in mice (Vegfr2iΔEC/iΔEC) abrogated propagation of both PCECs and AECIIs (Fig. 3D, E). Notably, endothelial-specific knockdown of Vegfr2 and Fgfr1 (Vegfr2iΔEC/iΔECFgfr1iΔEC/+) further abolished proliferation of PCECs and AECIIs at day 7 after PNX, suggesting that FGFR1 synergizes with VEGFR2 in stimulating PCECs to support AECII amplification and neo-angiogenesis.
To determine whether co-activation of VEGFR2 and FGFR1 plays a role in improving lung function, we examined inspiratory volume and static compliance in Vegfr2iΔEC/iΔECFgfr1iΔEC/+and control mice before and after PNX. These parameters of pulmonary function provide physiologically relevant indexes of respiratory capacity of the lung. The restoration of pulmonary function after PNX was significantly impaired in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice at the time point when control mice exhibited complete recovery (Fig. 3F). Similarly, restoration of lung mass, volume, and cell expansion after PNX were all drastically impaired in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice (Fig. 3G). These data indicate that after PNX, non-proliferating VE-cadherin+ ECs induce the early expansion of BASC-like cells via VEGFR2 activation. At later phases of PNX, upregulation of FGFR1 in conjunction with VEGFR2 activate PCECs to promote regenerative epithelialization and vascular sprouting, restoring lung respiratory capacity (Fig. 3H). Thus, PCECs by producing angiocrine factors foster neo-alveolarization into functional respiratory alveolar units.
To identify the inductive angiocrine cue that initiates epithelialization, we compared the gene expression profiles of the regenerative lungs by microarray (table S1), and found that among alveologenic factors, membrane-type 1 matrix metalloproteinase (MMP14) was specifically upregulated in the PCECs of wild-type, but not Vegfr2iΔEC/iΔEC or Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice (Fig. S3). Western analysis of MMP14 protein in the pneumonectomized lungs revealed its temporal upregulation that peaks at day 7 and levels off afterwards (Fig. 4A). Immunostaining and flow cytometric analysis illustrated the PCEC-specific localization of MMP14 after PNX, which was diminished in the Vegfr2iΔEC/iΔECFgfr1iΔEC/+ lungs (Fig. 4B, C). MMP14 was not upregulated in other vascular rich organs, including liver, heart, spleen and kidney (Fig. S3), indicating that after PNX MMP14 is selectively upregulated in the VEGFR2 and FGFR1 activated PCECs.
To define the mechanism by which angiocrine expression of MMP14 promotes the propagation of epithelial progenitor cells, we used endothelial coculture with AECIIs and BASCs isolated from SPC-YFP and CCSP-YFP mice, respectively. YFP expression was utilized to track their fate during coculture period. Since MMP14 was upregulated in primary MAPKinase-activated ECs (MAPK-ECs) (Kobayashi et al., 2010), we cocultured MAPK-ECs with AECIIs/BASCs in 3-dimensional (3D) spheroid assay. Coculture with MAPK-ECs led to the most significant expansion of SPC+ AECIIs and CCSP+Sca-1+CD31− BASCs (Fig. 4D–G, Fig. S3), resulting in formation of 3D spheroid structures that resemble the capillary-alveolar sacs. MMP14 knockdown in MAPK-ECs abolished expansion of BACSs and AECIIs (Fig. 4D, F). Conditioned medium (CM) from MAPK-ECs showed negligible effect in promoting AECII and BASC propagation, underscoring the requirement for cell-cell contact between endothelial coculture with epithelial cells (Fig. 4E, G). Therefore, resection of the left lung activates VEGFR2 and FGFR1 on PCECs triggering MMP14 production, which in turn stimulates propagation of epithelial progenitor cells.
To determine the physiological significance of MMP14 in modulating alveologenesis, we injected WT mice with neutralizing monoclonal antibody (mAb) to MMP14. After PNX, mAb to MMP14 blunted the increase of mass and volume of the remaining lungs in WT but not Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice, indicating that MMP14 is primarily derived from VEGFR2 and FGFR1 activated PCECs (Fig. 5A, Fig. S4). MMP14 inhibition blocked expansion of E-cadherin+ AECs without impairing reconstitution of the VE-cadherin+ PCECs (Fig. 5B). The mismatched expansion of AECs and PCECs after MMP14 inhibition indicates that MMP14 primarily induces propagation of AECs (inductive angiogenesis), rather than promoting PCEC proliferation (proliferative angiogenesis).
The reduced expansion of AECs, but not PCECs, by MMP14 neutralization was further demonstrated by flow cytometric analysis (Fig. 5C, D). Furthermore, in mice injected with mAb to MMP14, morphological examination revealed inhibition in alveolar regrowth, as evidenced by decrease in alveolar number and increase in alveolar size measured by mean alveolar intercept (Fig. 5E, F). Collagen synthesis remained unchanged in mice injected with mAb to MMP14 (Fig. S4). Therefore, PCEC-derived MMP14 stimulates neoalveolarization, forming alveolar sacs reminiscent of normal adult alveoli.
We next sought to unravel the mechanism by which MMP14 regulates regenerative alveolarization. MMP14 has been shown to shed the ectodomain of heparin binding EGF-like growth factor (HB-EGF) (Koshikawa et al., 2010; Stratman et al., 2011). In addition, MMP14 cleaves laminin5 γ2 chain, which generates the EGF-like fragment that activates EGF receptor (EGFR) (Schenk et al., 2003). We found that at day 3 and 7 after PNX, HB-EGF in the bronchioalveolar lavage fluid (BALF) is increased (Fig. 6A, B). The cleaved fragment of laminin5 γ2 chain appeared in the regenerating lungs at day 7 after PNX (Fig. 6C, Fig. S5). However, the level of these EGFR ligands was decreased in both control mice treated with mAb to MMP14 and Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice, in which there is diminished expression of MMP14. Knockdown of MMP14 in MAPK-ECs in the 3D endothelial coculture with BASCs and AECs also abrogated the release of EGFR ligands into the culture supernatants (Fig. S5). Hence, after PNX, activation of VEGFR2 and FGFR1 in PCECs leads to the angiocrine production of MMP14, which in turn unmasks cryptic EGFR ligands stimulating alveolar regeneration.
Both shedded HB-EGF and cleaved laminin5 γ2 chain induce activation of EGFR (Schenk et al., 2003) that is indispensible for epithelial proliferation and morphogenesis (Knox et al., 2010; Maretzky et al., 2011). Our findings suggest that after PNX, impaired lung alveolarization in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice is due to a decrease in the bioavailability of EGFR ligands. This raises the possibility that injection of EGF might restore alveolarization in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice by enhancing epithelialization. Thus, we tested the effect of recombinant EGF in restoring the defective alveolar regeneration in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice. Intravenous injection of EGF restored lung mass and volume in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice and in mice treated with mAb to MMP14 (Fig. 6D, Fig. S5). Compared to the intravenous injection of EGF, direct introduction of EGF into bronchioalveolar epithelium via intratracheal injection showed a similar effect in rescuing alveolar regeneration (Fig. S5). Therefore, the defective regeneration of AECs in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice (Fig. 3) is due to diminished MMP14 production by PCECs that leads to attenuation in the bioavailability of EGFR ligands.
Notably, in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice injected with EGF, there was a profound enhancement in the cellular association of E-cadherin+ AECs with VE-cadherin+ PCECs (Fig. 6E, Fig. S5) restoring pulmonary function (Fig. 6F). EGF injection into Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice stimulated regeneration of AECs, but not PCECs (Fig. S5), suggesting that EGF has a minor effect in stimulating sprouting angiogenesis, while being more effective in inducing epithelialization. To test this hypothesis, we analyzed the effect of EGF administration on cell amplification at day 7 after PNX. Injection of EGF led to enhanced EGFR phosphorylation in the Vegfr2iΔEC/iΔECFgfr1iΔEC/+ lung (Fig. 6G). BrdU incorporation analysis revealed that EGF restored the proliferation of AECIIs, but not PCECs in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice (Fig. 6H, I). This finding indicates that the alveologenic defect in Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice is due to the impaired generation of the epithelial-active angiocrine factors, rather than a compromise in vascular perfusion to the regenerating lung.
In our study, the tamoxifen-responsive Cre driven by pan-endothelial VE-cadherin could delete Vegfr2 and Fgfr1 in endothelial cells (ECs) of other vascular beds. Thus, to investigate the contribution of activated PCECs to specifically drive lung regeneration, we designed a lung endothelial cell (EC) transplantation model. ECs were purified from either lung or liver of the WT littermate pneumonectomized mice and infused into the jugular vein of Vegfr2iΔEC/iΔEC and Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice (Fig. 7A). Plasma was also collected from pneumonectomized WT mice and injected into the recipient knockout mice to interrogate the contribution of systemic soluble growth factors to lung regeneration after PNX.
Transplanted GFP+ ECs incorporated into ~26% of the pulmonary capillaries of the recipient mice (Fig 7B, Fig. S6). Importantly, the engrafted ECs obtained from the pneumonectomized lungs, but not the liver, restored the amplification of epithelial cells (Fig. 7C–F, Fig. S6). Proliferating BrdU+CCSP+ BASC-like cells and BrdU+SPC+ AECIIs were positioned in the proximity of the transplanted GFP+ PCECs, indicating that the inductive signals derived from the infused WT PCECs restore lung regeneration. Accordingly, the pulmonary function was also improved by transplantation of PCECs but not injection of plasma procured from pneumonectomized WT mice (Fig. 7G). Therefore, PNX induces a lung-specific activation of PCECs to elaborate angiocrine factors that support regenerative lung alveolarization (Fig. 7H).
We have employed PNX-induced alveolar regeneration model, endothelial-specific knock down of Vegfr2 and Fgfr1, and a 3D endothelial-epithelial coculture model, to establish the essential role of the PCECs in promoting regenerative alveologenesis. We have uncovered the angiocrine role of MMP14, which by shedding HB-EGF and generating EGF-like fragment from laminin5 γ2 chain, stimulates the amplification of lung epithelial progenitor cells, including subsets of BASCs and AECs, supporting alveolarization. The role of MMP14/EGFR activation in promoting alveologenesis was borne out in studies in which EGF administration into Vegfr2iΔEC/iΔECFgfr1iΔEC/+ mice restored alveolar regeneration after PNX. Moreover, we established a lung PCEC transplantation model to define the essential role of functionally incorporated PCECs in restoring epithelialization in mice with impaired capacity to undergo neoalveolarization. Hence, we have demonstrated that after PNX, PCECs orchestrate regenerative alveolarization by formation of new vessels as well as through instructive production of epithelial-active angiocrine factors.
PNX induced alveolar regeneration via amplification of epithelial progenitor cells. At early phases (day 0–3), PNX induces expansion of CCSP+SPC+Sca-1+CD31−VE-cadherin− BASC-like cells localized at BADJ. At the later phases of lung regeneration (day 7–15 post PNX) SPC+E-cadherin+ AECIIs and PCECs expand, re-establishing functional alveolar-capillary units. Upon MMP14 inhibition, the loss of alveolar coverage of not only cuboidal SPC+E-cadherin+, but also squamous SPC−E-cadherin+ AEC implicates that transiently amplified SPC+E-cadherin+ AECIIs generate SPC−E-cadherin+ type I AECs (Beers and Morrisey, 2011; Morrisey and Hogan, 2010; Rock and Hogan, 2011), leading to full reconstitution of alveolar surface after PNX. Therefore, activated PCECs drive regeneration of specialized lung epithelial cells that collectively reconstitute functional alveolar-capillary sacs.
We show that PCEC-derived MMP14 is required for the expansion of epithelial cells as well as restoration of alveolar structure and pulmonary function. In the developing mouse fetal lung, MMP14 regulates alveolar formation (Atkinson et al., 2005; Irie et al., 2005; Oblander et al., 2005), presumably by stimulating epithelial proliferation and migration (Chun et al., 2006; Hiraoka et al., 1998; Koshikawa et al., 2010; Stratman et al., 2009). Postnatally, MMP14 deficient mice exhibit defective alveolarization and abnormal sacculation, manifested by impaired vascular integration with AECs. This suggests that MMP14 plays a critical role in mediating alveolar-capillary crosstalk by as yet undefined mechanism (Li et al., 2002; Morris et al., 2003; Page-McCaw et al., 2007; Yana et al., 2007). Here, we show that after PNX, inhibition of MMP14 interfered with the alveolar regrowth but not endothelial proliferation, leading to enlarged alveolar size. This suggests that MMP14 is dispensable for proliferative angiogenesis, but plays a key role in inducing regenerative alveolarization. The mechanism by which MMP14 modulate alveologenesis is mediated by shedding of HB-EGF into alveolar space and generating EGF-like fragment from laminin5 γ2 chain. Subsequently, increase in the bioavailable EGFR-ligands fosters regeneration of epithelial progenitors. In this regard, MMP14 performs as a PCEC-specific angiocrine cue that drives regenerative alveolarization.
Each organ is vascularized by specialized population of capillary endothelial cells identified by unique phenotypic, functional and structural attributes (Butler et al., 2010a; Butler et al., 2010b). We have shown that bone marrow and liver SECs (Butler et al., 2010a; Butler et al., 2010b), which are demarcated by VEGFR2+VEGFR3+VE-cadherin+ vessels, express defined set of angiocrine factors driving organ regeneration. After partial hepatectomy, VEGFR2 and Id1 activated liver SECs produce HGF and WNT2 (Ding et al., 2010), while bone marrow VEGFR2 activated SECs express Notchligands and IGFBPs (Butler et al., 2010a; Butler et al., 2010b) to induce reconstitution of hepatocytes and hematopoietic cells, respectively.
Similarly, PCECs have distinct phenotypic signature and could be identified as VEGFR2+FGFR1+CD34+VE-cadherin+ vessels. Remarkably, after PNX, the production of MMP14 is restricted to VEGFR2 and FGFR1 activated PCECs but not other vascular rich organs, such as liver, spleen, heart and kidneys, highlighting a unique functional signature of PCECs in alveolar regeneration. The negligible effect of plasma obtained from pneumonectomized WT mice in restoring alveologenesis also demonstrated the minimal contribution of systemic soluble growth factor(s) from non-pulmonary vasculature in mediating alveologenesis. These data clearly set forth the notion that PNX turns on a PCEC-specific program to promote alveolar regeneration.
The mechanism by which PNX specifically upregulates MMP14 in PCECs could be regulated by microenvironmental cues and/or unique inherent programming of PCECs. For example, surgical removal of left lung lobe leads to local activation of PCECs in the remaining right lungs, without affecting other vascular beds (Fig. 1A). Alternatively, PCECs, but not other organ-specific capillaries, may be developmentally predetermined to express MMP14 in response to regenerative signals. Notwithstanding to the potential developmental or microenvironmental cues that endow PCECs with their unique functional attributes, our findings consolidate the concept that angiocrine heterogeneity plays a key role in orchestrating organ regeneration.
The mechanism by which PCECs are induced to express MMP14 after PNX is mediated by hierarchical activation and upregulation of VEGFR2 and FGFR1. At the early phase of PNX, expansion of BASC-like cells is largely dependent on the activation of VEGFR2 in PCECs, which causes upregulation of MMP14 without inducing endothelial proliferation. In contrast to the early activation and stable expression of VEGFR2, FGFR1 expression level is induced thereafter peaking day 7 after PNX. We show that FGFR1 synergizes with VEGFR2 in augmenting MMP14 generation thereby sustaining alveolar regeneration. Therefore, sequential activation of VEGFR2 and FGFR1 in PCECs induces MMP14 production, fostering regeneration of the functional alveolar-capillary units.
The development of therapeutic strategies to repair the respiratory capacity in patients with pulmonary disorders, is handicapped by lack of understanding of lung regeneration mechanisms (Jiang et al., 2005; Kajstura et al., 2011; Matthay and Zemans, 2011; Morris et al., 2003; Petrache et al., 2005). We have set forth the concept that after PNX activated PCECs play a central role in restoring respiratory capacity, as measured by inspiratory volume and static compliance. Notably, administration of EGF or transplantation of activated WT PCECs improved respiratory function in mice. It is plausible that transplantation of properly activated PCECs or injection of lung-specific angiocrine mediators could improve lung function in subset of patients with pulmonary disorders.
In conclusion, we have introduced the concept that PCECs not only form passive vascular conduits to fulfil the metabolic demands of the regenerating lungs, but also by relaying inductive angiocrine growth signals, such as MMP14, orchestrates regenerative alveologenesis. Therefore, selective activation of VEGFR2 and FGFR1 or increase in the bioavailability of MMP14 in PCECs might facilitate lung alveolarization thereby improving hypoxemia in patients with debilitating lung diseases.
C57BL/6J mice were obtained from Jackson Laboratories (Bar Harbor, ME). Generation of endothelial-specific Vegfr2 and Fgfr1 inducible knockout mice was carried out as previously described (Hooper et al., 2009; Wang et al., 2010). Briefly, Vegfr2loxP/loxP mice were bred with VE-cadherin-CreERT2 transgenic mice to establish the VE-cadherin-CreERT2+Vegfr2LoxP/LoxP and VE-cadherin-CreERT2+Vegfr2LoxP/+ mice. Fgfr1loxP/loxP mice were further crossed with VE-cadherin-CreERT2+Vegfr2loxP/loxP line to generate compound mutant of VE-cadherin-CreERT2+Vegfr2loxP/loxPFgfr1loxP/+ mice. These mice were i.p. treated with tamoxifen at a dose of 250 mg/kg in sunflower oil for 6 days (interrupted for 3 days after the third dose), leading to endothelial-specific deletion of Vegfr2 and Fgfr1. After 3 weeks of tamoxifen treatment, knock down of the Vegfr2 and Fgfr1 in PCECs was quantified by quantitative PCR and immunostaining.
Mice bearing SPC and CCSP promoter-driven rtTA (SPC-rtTA, CCSP-rtTA) and (tetO)7CMV-driven cre ((tetO)7-cre) (Perl et al., 2002) crossed with Rosa26R–eYFP mice carrying transgene constructs in which YFP expression is blocked by upstream loxP-flanked stop codon (Jackson Labs), resulting in SPC or CCSP-rtTA/(tetO)7-cre/Rosa26R–eYFP mice. Treatment of these mice with tetracycline generated SPC-YFP and CCSP-YFP reporter mice. All animal experiments were carried out under guidelines set by Institutional Animal Care and Use Committee, using age/weight/strain matched littermate animals.
PNX procedure was adapted as described (Nolen-Walston et al., 2008). Briefly, mice were anesthetized and orotracheal intubation was performed. Mice were mechanically ventilated (Minivent 845) at 150– 200 tidal breaths of 0.3 ml of room air per minute. The skin and pleura muscle were incised between the 5th and 6th intercostal ribs, and the left lung lobe was gently lifted while a 5-0 silk suture was tied around the hilum. After the left lobe distal to the suture was resected, the chest wall was reapproximated, and the skin was closed. Mice were extubated at the onset of vigorous spontaneous breathing. Sham mice underwent thoracotomy without lobe resection. At different time points after PNX, lung mass and volume were measured (Nolen-Walston et al., 2008) and normalized to body weight. Isolation of PCECs and examination of phosphorylation and protein level of VEGFR2 and FGFR1 was carried out (Murakami et al., 2011). Briefly, perfused mouse lung was digested with collagenase/dispase and dispersed mechanically into single-cell suspension. PCECs were purified from cell suspension using sequential positive selections with anti-CD31, ICAM-2 and VE-cadherin antibodies conjugated to Dynabeads (Invitrogen). Hematopoietic cells are further excluded by a negative depletion with anti-CD45 Dynabeads.
Inspiratory capacity was determined between the plateau pressure measurements of the total lung capacity (TLC) and functional residual volume (FRC) using the Flexivent software (Scireq). Static compliance was determined from pressure-volume curves that were generated by sequential delivery of air between FRC and TLC and calculated during the expiratory phase of the pressure-volume loop.
Mice were subjected to PNX or sham operation, and tissues were cryoprotected. For IF studies, the lung sections were incubated in primary Abs: anti-VE-cadherin Ab (2 µg/ml, R&D), anti-CD34 mAb (5µg/ml, BD), anti-E-cadherin (2µg/ml, eBiosciences), and anti-SPC polyclonal Ab (5 µg/ml, Abcam). After incubation in fluorophore-conjugated secondary antibodies (2.5 µg/ml, Jackson Immuno Research). Lung cell proliferation in vivo was measured by BrdU uptake. Briefly, mice received injection of BrdU (Sigma) intraperitoneally 60 min before death (50 mg/kg). To track proliferating BASC-like cells, BrdU (1 mg/ml) was introduced in drinking water on day 0–7 after PNX. Cryosections were stained using the BrdU Detection System, and IF images were captured on AxioVert LSM710 confocal microscope (Zeiss). Lung morphological analysis of alveolar number and mean linear intercept was performed (Fehrenbach et al., 2008).
Total lung mononuclear cells were isolated (Bortnick et al., 2003; Kim et al., 2005) and analyzed on LSRII-SORP (BD). Conjugation of purified mAbs, exclusion of cell doublets and procession of data were performed (Ding et al., 2010). Antibodies used: VE-cadherin (BV13, ImClone); VEGFR2 (DC101, ImClone); CD45 (30-F11, BD), and CD34 (14–0341, eBioscience). AECs and PCECs were quantified by staining with conjugated antibodies against SPC + E-cadherin and VE-cadherin + CD34, respectively.
Mice were injected with mAb to mouse MMP14 (MMP14 mAb, 50 mg/kg, Abcam) and IgG control 12 hours before PNX and every other day. To determine the role of recombinant EGF in alveolar regeneration, mice were intravenously injected with 500 µg/kg recombinant mouse EGF (Abcam) on daily basis after PNX for 14 days. Mice were also intratracheally injected with 100 µg/kg EGF (in 50 µl volume) every other day to examine the local effect of EGF.
MAPkinase- and Akt- activated primary endothelial cells (MAPK-ECs, Akt-ECs) (Kobayashi et al., 2010) were cocultured with AECIIs and BASCs isolated from SPC and CCSP-YFP mice (Bortnick et al., 2003; Kim et al., 2005). Mmp14 or scrambled shRNA was used to selectively knockdown Mmp14 in MAPK-ECs or AECs (Ding et al., 2010). For co-culture studies, 50,000 isolated SPC+ AECIIs and 2,000 BASCs were plated in non-adherent dish, seeded with 10-fold more MAPK-ECs. Conditioned medium from MAPK-ECs was added to AECs (Ding et al., 2010). After coculture, AECIIs and BASCs were quantified by flow cytometric analysis and comparing the number of retrieved cells to initially seeded number.
At Day 7 after PNX, total RNA was isolated from the lung of Vegfr2iΔEC/iΔECFgfr1iΔEC/+ and Vegfr2iΔEC/+ mice using RNeasy (Qiagen) and converted to cDNA using Superscript II (Invitrogen). qPCR was carried out using Taqman gene expression systems for mouse MMP14 (Applied Biosystems). Bronchioalveolar lavage fluid (BALF) and lung total protein were obtained (Kim et al., 2005), and concentration of HB-EGF in BALF was examined by sandwich ELISA and immunoblot using anti-HB-EGF antibodies (Santa Cruz), and the cleavage of laminin5 γ2 chain in the lung was tested with antibody against mouse γ2 chain (Santa Cruz).
All data are presented as the Mean ± sem of at least three separate experiments. Differences between groups were tested for statistical significance using Student’s t-test or analysis of variance (ANOVA). Statistical significance was set at P < 0.05.
B.-S. D. is a Druckenmiller Fellow of New York Stem Cell Foundation. S.R. is supported by the Ansary Stem Cell Institute; Howard Hughes Medical Institute; Empire State Stem Cell Board and the New York State Department of Health grants (C024180, C026438, C026878); National Heart Lung and Blood Institute; Qatar National Priorities Research Foundation NPRP08-663-3-140; Anbinder and Newmans Own Foundation. T.N.S is supported by Takeda Science Foundation, The Uehara Memorial Foundation, JSPS (Kiban S). M.S. is supported by R01 HL53793. VE-cadherin-ERT2 mice were gift from Dr. Ralf H. Adams (Max Planck Institute). SPC and CCSP-rtTA and (tetO)-cre mice were generously provided by Drs. Jeffrey A. Whitsett and Anne-Karina T. Perl (Cincinnati Children's Hospital Medical Centre). Fgfr1loxP/loxP mice were kindly offered by Drs. Michael Simons and Masahiro Murakami (Yale University). The authors are grateful to Ms. Biin Sung for her assistance in lung mechanics measurement.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.