BMP1 is not required for lung fibrosis in mice

Bone morphogenetic protein 1 (BMP1) belongs to the astacin/BMP1/tolloid-like family of zinc metalloproteinases, which play a fundamental role in the development and formation of extracellular matrix (ECM). BMP1 mediates the cleavage of carboxyl terminal (C-term) propeptides from procollagens, a crucial step in fibrillar collagen fiber formation. Blocking BMP1 by small molecule or antibody inhibitors has been linked to anti-fibrotic activity in the preclinical models of skin, kidney and liver fibrosis. Therefore, we reason that BMP1 may be important for the pathogenesis of lung fibrosis and BMP1 could be a potential therapeutic target for progressive fibrotic disease such as idiopathic pulmonary fibrosis (IPF). Here, we observed the increased expression of BMP1 in both human IPF lungs and mouse fibrotic lungs induced by bleomycin. Furthermore, we developed an inducible Bmp1 conditional knockout (cKO) mouse strain. We found that Bmp1 deletion does not protect mice from lung fibrosis triggered by bleomycin. Moreover, we found no significant impact of BMP1 deficiency upon C-term propeptide of type I procollagen (CICP) production in the fibrotic mouse lungs. Based on these results, we propose that BMP1 is not required for lung fibrosis in mice and BMP1 may not be considered a candidate therapeutic target for IPF.

www.nature.com/scientificreports/ progressive fibrotic diseases such as IPF. Upon analyses of fibrotic lung samples from human and mouse, we found BMP1 expression is increased in fibrotic lesions. Next, we developed a mouse model where efficient deletion of both Bmp1 alleles can induced by tamoxifen treatment. However, when these Bmp1 conditional knockout (cKO) mice were tested in the bleomycin model of lung fibrosis, we observed that BMP1 deficiency does not prevent lung fibrosis in mice. Further examination of CICP production also revealed that BMP1 is not necessary for type I procollagen C-term processing and maturation during lung fibrosis. Taken together, we provided the first genetic evidence that BMP1 is not required for the development of lung fibrosis in mice, thus arguing against the drug target candidacy of BMP1 for IPF.

Results
BMP1 expression is increased in IPF lungs. We first assessed the protein level of BMP1 by immunohistochemical staining in lung samples collected from normal control and IPF patients. We only observed a minimal signal of BMP1 protein in normal lung. On the contrary, the immunoreactivity of BMP1 was strongly detected in IPF lungs. Specifically, BMP1 protein appears to be enriched in the fibroblastic foci (FF), a histological feature of IPF, and the epithelium lining ( Fig. 1).

BMP1 expression is increased in the bleomycin model of lung fibrosis.
To complement our observation in human samples, we sought to determine whether BMP1 expression is increased in the mouse model of lung fibrosis. Intratracheal (IT) instillation of bleomycin model has been widely used as an animal model to induce experimental lung fibrosis. Mechanistically, bleomycin damages lung epithelium, which results in an immune response and fibroblasts activation, differentiation and ECM deposition 35 . Through analysis of a previously published transcriptomics data in the bleomycin model 36 , we found that Bmp1 expression was trending slightly lower, though not significant, in the bleomycin-challenged moue lungs at early time points (day 2 and 7) but was significantly increased at the advanced fibrotic stage of the model (day 14-28) ( Fig. 2A). We also examined the expression of other procollagen C-term proteinases including tolloid-like 1 (Tll1), tolloid-like 2 (Tll2), meprin A (Mep1a) and meprin β (Mep1b) in the same dataset. The expression of these genes was not significantly increased in the bleomycin injured mouse lungs throughout the experiment except that Mep1a expression was significantly increased only at day 14 ( Fig. 2A). Mep1b expression was below the detection limit (data not shown). Consistent to the results from transcriptomic analyses, we found that BMP1 (the longer spliced version of Bmp1, also named as mTLD, ~ 120KD) protein level was dramatically elevated in mouse lungs 24 days after the challenge of the bleomycin while the shorter spliced version of BMP1 (~ 90KD) appears to be below the detection limit (Fig. 2B). These results thus suggest that the bleomycin model may be suitable to investigate the in vivo role of BMP1 in lung fibrosis.
Inducible and efficient deletion of Bmp1 alleles in Bmp1 cKO mice. Bmp1 constitutional knockout mice die soon after birth from the failure of ventral body wall closure due to abnormal collagen fibrillogenesis 31 .
To bypass this postnatal lethality, we generated mice homozygous for the floxed (cKO) or wild-type (WT) Bmp1 alleles and hemizygous for a Cre-ERT2 fusion transgene driven by the human Rosa26 promoter. Thus, both Bmp1 alleles can be deleted when these cKO mice are treated with tamoxifen. In order to confirm the efficient www.nature.com/scientificreports/ deletion of Bmp1 alleles, we administered tamoxifen to Bmp1 cKO and WT mice through intraperitoneal (IP) injection a week prior to IT injection of bleomycin or saline (Fig. 3A). Lung tissues were then harvested for gene expression analyses on day 24 after the initial bleomycin challenge (Fig. 3A). The data showed that tamoxifen induced more than 90% loss of Bmp1 transcript in lungs in the presence or absence of bleomycin when compared to WT controls (Fig. 3B). In addition, we observed that Tll1, Tll2 and Mep1a expression was not significantly affected by Bmp1 deletion, suggesting that the loss of BMP1 is unlikely compensated by the increased expression of other C-proteinases as examined.
Bmp1 cKO mice are not protected from bleomycin-induced lung fibrosis. Next, we investigated whether inducible deletion of Bmp1 protects mice from the development of lung fibrosis in the bleomycin model. To this end, we first preformed the histological analysis to assess the severity of fibrotic lesions in the lungs of Bmp1 cKO and WT mice. The data indicate there is no difference of the severity of fibrotic lesions in the lungs between Bmp1 WT and cKO mice (Fig. 4A,B). Furthermore, using hydroxyproline assay, we quantified the overall collagen deposition in lungs and found no significant impact of Bmp1 deletion on total hydroxyproline level in lungs (Fig. 4C,D). We also measured the newly synthesized hydroxyproline and found that Bmp1 deletion does not affect this readout either, suggesting that BMP1 is unlikely to contribute to active collagen synthesis www.nature.com/scientificreports/ and production. Consistent to the histology and hydroxyproline data, we observed that the representative profibrotic gene (Col1a1, Col3a1, Serpine1 and Timp1) expression was not significantly impacted by Bmp1 deletion (Fig. 5). Collectively, these results suggest that BMP1 does not contribute to the pathogenesis of lung fibrosis in the bleomycin model.

Bmp1 deletion does not reduce CICP production during bleomycin-induced lung fibrosis.
The lack of anti-fibrotic effects in Bmp1 cKO mice prompted us to explore whether BMP1 modulates procollagen, particularly type I procollagen, C-term processing during the development of lung fibrosis in the bleomycin model. We performed western blot (WB) in lung extracts from a satellite group of Bmp1 WT and cKO mice treated with bleomycin or saline to assess the proα1(I), pCα1(I), fully processed COL1 and CICP production. While the BMP1 protein level appears to be efficiently depleted in the lungs of cKO mice, we found no significant change of proα1(I), pCα1(I), fully processed COL1 and CICP levels upon BMP1 depletion while their production were substantially increase in the bleomycin injured lungs (Fig. 6A-E). Therefore, these results suggest that BMP1 is not necessary for type I procollagen C-term processing and maturation during lung fibrosis in mice, which may explain the redundant role of BMP1 in lung fibrosis. www.nature.com/scientificreports/

Discussion
As IPF has the worst prognosis among all fibrotic diseases and most IPF patients survive less than 3 years from diagnosis, this fatal and progressive fibrotic lung disease placed a huge burden on the global public health system 1-3 . However, very few experimental therapies exhibited clinical benefits in human patients and the current standard of care medications, pirfenidone and nintedanib, only slow down the decline of lung function without significantly improving patient survival 37,38 . New treatment options are therefore urgently needed to prevent or even reverse the progression of IPF toward disease mortality 4,5 .
Since its discovery as the first C-term proteinase in early 90s 14,15 , BMP1 has long been proposed to be an appealing therapeutic target for progressive fibrotic diseases 39 . Within the past few decades, several potent hydroxamic acid-containing small molecule BMP1 inhibitors have been reported, but no advancement to clinical trial has been documented 29,40-42 possibly due to poor selectivity and/or poor metabolic stability 43 . More recently, BMP1 antibody inhibitors have been reported 31,34 and one of them exhibited anti-fibrotic efficacy in preclinical models of kidney and liver fibrosis 31,33 . However, to the best of our knowledge, there are no genetics studies that have ever validated the role of BMP1 in the pathogenies of tissue fibrosis.
In this end, we developed a Bmp1 cKO mouse model where both Bmp1 alleles can be deleted in the adulthood of the animals. This unique mouse model allows us to fully employ a genetic approach to determine the contribution of BMP1 to lung fibrosis in animals. To our surprise, we found that BMP1 is clearly not required for lung fibrosis in the bleomycin model and BMP1 does not contribute to type I procollagen C-term processing. These findings not only suggest that BMP1 may not be a suitable drug target for progressive lung fibrotic diseases but also indicate that BMP1 is not the major C-term proteinase for procollagen in this particular model. These results are somewhat contradictory, but not mutually exclusive, with previous reports of BMP1's dominant role in collagen deposition or procollagen C-term processing using BMP1 inhibitors and in vitro cell culture system 17,18,[29][30][31]33,34 . Herein, we suggest that there could be three explanations to reconcile these discrepancies.
First of all, a few non-BMP1 proteinases are also able to cleave C-term propeptide of procollagen. For instance, due to the similar domain structures, TLL1 and TLL2 were identified as BMP1-like proteinases and all these BMP1/tolloid-like proteinases (BTPs) can process C-term procollagen efficiently 12,13 . MEP1A and MEP1B have also been identified as procollagen C-term proteinases 16 . Hence, the functional redundancy between these www.nature.com/scientificreports/ C-proteinases may well compensate the depletion of BMP1 in vivo. In support of this notion, genetic deletion of Bmp1 or Bmp1/Tll1 in mouse embryonic fibroblasts (MEFs) suggested that both BMP1 and TLL1 contribute to C-term procollagen processing 17 . On the other hand, as tissue fibrosis ensues from a pathological microenvironment, multiple cell types can be involved in this process. Thus, it is possible that proteinase activities from non-fibroblastic cell types may contribute to C-term processing of procollagens as well. Therefore, future studies using genetic models or unbiased biochemical analyses are needed to identify which proteinases are truly responsible for procollagen C-term processing and maturation during lung fibrosis in vivo. In addition to the functional redundancy, we cannot ignore the possibility that BMP1 may play a tissue specific role in the pathogenesis of fibrosis. It is noteworthy that, while previous studies tested BMP1 inhibitors in the animal models of skin 29,30 , kidney 31,32 and liver fibrosis 33 , the anti-fibrotic efficacy of BMP1 inhibitors in lung fibrosis models has not been documented. Thus, it is plausible that BMP1 may play a much more important role in skin, kidney and liver fibrosis than in lung fibrosis. This possibility can be further supported by human genetics data that the collagen abnormality caused by BMP1 loss-of-function mutations appears to be largely restricted to bone tissue [19][20][21][22][23][24][25][26][27][28] . Nonetheless, additional studies on Bmp1 cKO mice in the fibrosis models of other tissues are mandatory to fully test this hypothesis.
Thirdly, as the selectivity for both small molecule and antibody BMP1 inhibitors has not been fully characterized, it may be reasonable to speculate that these molecules may hit targets beyond BMP1 in vivo, if the anti-fibrotic effects of these BMP1 inhibitors can be independently reproduced. Given the high sequence and structure similarity between BTP family members 44 , it is highly likely that BMP1 inhibitors may target the entire BTP family to achieve their anti-fibrotic effects. However, as the CICP level was not determined in the previous pharmacology studies in animal models, we cannot exclude another possibility that the anti-fibrotic effects of these inhibitors may, at least partially, come from non-BTP targets, which may be functionally relevant to tissue fibrosis.
Overall, our study provided the first genetic evidence that BMP1 is not required for lung fibrosis and BMP1 is not the major proteinase to cleave C-term of type I procollagen in the bleomycin model. Based on these findings, we thus propose that BMP1 should not be considered as a candidate therapeutic target for fibrotic lung diseases like IPF. More importantly, our study exemplifies the necessity of leveraging genetic models to validate drug targets when possible and suggests that the role of BMP1 in the pathological tissue fibrosis may need to be re-evaluated using the similar genetic approaches. www.nature.com/scientificreports/

Materials and methods
Human samples. Explanted lung tissues were obtained from patients with a pathological diagnosis of usual interstitial pneumonia and a consensus clinical diagnosis of IPF assigned by multidisciplinary discussion and review of clinical materials. Written informed consent was obtained from all subjects and the study was approved by the University of California, San Francisco (UCSF) institutional review board. Human lungs not used by the Northern California Transplant Donor Network were used as controls; studies indicate that these lungs are physiologically and pathologically normal 45 . All procedures were carried out according to the Helsinki Declaration and its later amendments.
Immunohistochemistry staining. 4 μm sections of formalin-fixed and paraffin-embedded specimens were deparaffinated followed by antigen retrieval using Target Retrieval (Dako #S1700) and quenching endogenous peroxidase activity. Bleomycin model of lung fibrosis. The study was carried out in compliance with the ARRIVE guidelines.
Adult mice (> 12 weeks) were randomized based on pre-study weights to minimize variance between experimental and control groups. For intratracheal (i.t.) dosing, all mice were lightly anesthetized with isoflurane in an induction chamber. Once anesthetized, the animals were removed from the chamber, manually restrained, the mouth of the animal was opened and the tongue set aside. A 1 ml syringe with 50 µl of sterile injectable isotonic saline or bleomycin (0.72 U/kg (DNC# 0703-3155-01; TEVA) in 50 μl sterile isotonic saline) was connected to a 24 gauge gavage needle. The gavage needle was inserted into the trachea and a dose of either vehicle or bleomycin was delivered intratracheally. After delivery, animals were monitored continuously until fully awake and ambulatory. Mouse whole lungs were harvested at day 24 after bleomycin administration, n = 5 in the saline control group, n = 30 in the bleomycin group in Bmp1-cKO and WT control littermates respectively.
Hydroxyproline and measurement. Total hydroxyproline was analyzed as previously described 48 and mass spectrometry and analysis were performed by Metabolic Solutions. Deuterated water labeling was used to assess new collagen synthesis. In brief, mice were injected with deuterated water (DLM-4-99.8-1000; Cambridge Isotope Laboratories) intraperitoneally two weeks prior to the end of the study at 35 mL/kg in 2 divided doses 4 h apart. Afterward, 8% deuterated water in drinking water was provided ad lib in water bottle until the end of the study. Mass spectrometry and analysis were performed by Metabolic Solutions.
Complementary DNA synthesis was carried out with iScript RT Supermix (Bio-Rad). Quantitative PCR was performed in technical triplicates using SYBR Green reagent (Bio-Rad). The relative standard curve method was used for quantitation and expression levels were calculated by normalization to Hrpt. The sequences of primers are listed in Supplementary Table 1.

Western blot.
Western blot was carried out in total protein extracts as previously described 49 . Equal amounts of protein lysates were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and subjected to immunoblotting analysis using the following primary antibodies: BMP1 and CICP antibodies were www.nature.com/scientificreports/