Generation of rat-derived lung epithelial cells in Fgfr2b-deficient mice retains species-specific development

The lungs are an interface for gas exchange from oxygen to carbon dioxide through respiration, an essential process for life. As pulmonary alveoli do not regenerate from extensive chronic damage, chronic obstructive pulmonary disease (COPD), the third leading cause of death in the world, are progressive and incurable (Li et al., 2022). Although the only fundamental treatment for COPD or end-stage lung disease is lung transplantation, donor shortage is a critical limitation (Valapour et al., 2021). To overcome this problem, biological artificial lungs have been created in vitro using a decellularized matrix scaffold. Decellularized lungs filled with endogenous lung epithelial cells have been successfully transplanted with a life of a few hours, but are yet to offer a long-term solution (Ott et al., 2010; Petersen et al., 2010). Lung epithelial cells differentiated from human induced pluripotent stem cells (iPSCs) can repopulate into the required scaffold (Ghaedi et al., 2018), but to generate a human-scale lung, applying xeno-organs that contain different species as scaffolds has immunological problems (Stahl et al., 2018).

The lungs develop from epithelial tissue derived from the foregut endoderm and from mesenchymal tissue derived from the visceral mesoderm. At embryonic day (E) 9.5 in mouse and E11 in rat, the lung bud bifurcates anteriorly from the ventral foregut endoderm into the mesenchymal tissue (Herriges and Morrisey, 2014). By E16.5 in mouse and E18.5 in rat, the basic structures of the lung are formed, including airways and terminal bronchi. The lung endoderm progenitor cells produce basal, ciliated, secretory and neuroendocrine cells, whereas the lung mesenchymal cells produce smooth muscle cells, chondrocytes, mesothelial cells, myofibroblast and lipofibroblast (Liu et al., 2021). Next, in E16.5-E17.5 in mouse and E18.5-E20 in rat, the surrounding mesenchyme becomes thinner and capillary vessels are actively formed. At this time, type I and II alveolar epithelial cells and lipofibroblasts arise. In the saccular stage [E17.5 to postnatal day (P) 4 in mouse and E21-P4 in rat], alveolar sacs are formed, surfactant protein production begins and capillaries develop (Loering et al., 2019). The overall difference in developmental timing between mice and rats throughout lung development is ∼1-3 days (Schittny, 2017).

Fibroblast growth factor 10 (Fgf10), which is essential for lung bud elongation, is secreted by the mesenchyme surrounding epithelial tissue (Sekine et al., 1999; Yuan et al., 2018) and its signal is received by fibroblast growth factor receptor 2 isoform IIIb (Fgfr2b; also known as Fgfr2) in the lung epithelium (Ohuchi et al., 2000). The interaction of Fgf10 and Fgfr2b is crucial for lung development, and both Fgf10-knockout (KO) and Fgfr2b-KO mice have shown lung agenesis in previous studies (Sekine et al., 1999; De Moerlooze et al., 2000).

To solve the problem of organ shortage, attempts have been made to create donor organs from pluripotent stem cells (PSCs) in the animal body through a process called blastocyst complementation (Chen et al., 1993a; Kobayashi et al., 2010; Isotani et al., 2011; Usui et al., 2012; Yamaguchi et al., 2017; Chang et al., 2018; Hamanaka et al., 2018; Mori et al., 2019; Goto et al., 2019; Kitahara et al., 2020; Ruiz-Estevez et al., 2021). In this method, PSCs such as iPSCs and embryonic stem cells (ESCs) are injected into blastocysts of an organ-deficient model. These PSCs can compensate for the defective organs, and PSC-derived organs are created in the body of the resulting chimeric animal. Using this method, transplantable pancreases and thymi have been successfully produced in interspecies chimeras using mice and rats (Kobayashi et al., 2010; Isotani et al., 2011). Moreover, rat blood vasculature, heart and eyes have been generated in the corresponding organ-deficient mouse model at the embryonic stage (Wang et al., 2020; Wu et al., 2017). Even human endothelium or skeletal muscle have been generated in the embryonic stage of ETV2 or MYF5/MYOD/MYF6-deficient pigs (Das et al., 2020; Maeng et al., 2021). PSC-derived kidneys were also generated in interspecies chimeras using mouse PSCs in a rat kidney-deficient model, but this process was not successful for rat PSCs in a mouse model (Usui et al., 2012; Goto et al., 2019). Functional lungs have been produced in intraspecies chimeras using the Fgf10-KO or Fgfr2b-KO mouse models with mouse PSCs, but are yet to be observed in an interspecies chimera (Mori et al., 2019; Kitahara et al., 2020). This suggests that the combination of blastocyst and PSC species may be crucial, but the exact requirements for successful organogenesis by blastocyst complementation are not yet known. Furthermore, resulting organs are often only partially PSC-derived, even if the model organism exhibits organ-deficient gene dysfunction (Usui et al., 2012; Mori et al., 2019; Goto et al., 2019; Kitahara et al., 2020). Therefore, the evaluation of the organ-deficient model is important for generating fully PSC-derived organs and for the realization of future regenerative medicine applications.

This study aimed to examine the lung-deficient model through a ‘reverse-blastocyst complementation (rBC) method’, which involves the injection of mutant ESCs into wild-type (WT) embryos. This method is capable of obtaining chimeras containing WT and mutant cells, similar to the blastocyst complementation method. The method allows us to efficiently detect mutant cells in the organ and to clarify the conditions for successful lung formation. We then achieved lung formation by rat cell complementation in an Fgfr2b-KO mouse model using a tetraploid-based organ-complementation method. The rat cells in the generated lungs unexpectedly retained their developmental timing in the mouse body.

For an rBC system (Fig. 1A), we used mutant ESCs constitutively expressing Su9-DsRed2 (RFP), such that the contribution of mutant cells in the chimera was easily detected. In E14.5 allogeneic chimeric fetuses, RFP-expressing WT ESC-derived cells were found to have similar contribution rates (chimerism) in various tissues and organs (Fig. S1A-C). The results suggest that it is possible to estimate the chimerism in the target tissue from the chimerism in the non-target tissue. In addition, it is thought that mutant ESCs exhibit reduced contribution to the target tissues in the chimera. Therefore, when estimating the chimerism in the target tissues of the chimeras, we chose to assess the chimerism in non-target tissues, presumed to be unaffected by the gene mutation.

We designed gRNAs on either side of exon 1 on the Fgf10 gene, which contains a start codon, and established the Fgf10-KO ESC lines (Fig. S2A,B). We also designed two gRNAs to remove the IgIIIb domain of Fgfr2 and generate Fgfr2b-KO ESC lines (Fig. S3A,B). Two ESC lines for Fgf10-KO or Fgfr2b-KO, which have different mutations, were used to produce chimeras (Figs S2C, S3C). Chimeric embryos were generated by injecting Fgf10-KO or Fgfr2b-KO ESCs into WT embryo at E2.5 stage, which were dissected at E14.5 (Tables 1 and 2). Flow cytometry analysis of Fgf10-KO chimeras revealed that lung chimerism is correlated with tail chimerism, resembling the pattern observed in chimeras injected with WT ESCs (Fig. S2D). As Fgf10-KO primarily exhibits defects in the limbs and lungs (Sekine et al., 1999), we estimated the overall chimerism of Fgf10-KO chimeras from the tail chimerism, considering it as a tissue unaffected by the impact of Fgf10 deficiency. Chimeras with over 90.7% contribution of Fgf10-KO cells i.e., less than 9.3% WT cells, exhibited defects in lung and limb formation (Fig. 1B,C). However, limb and lung defects were not observed in chimeras with more than 14.7% contribution of WT cells (Fig. 1B,C). These results suggest that a certain number of cells expressing Fgf10 as WT cells is required to rescue the lung and limb development. Analogous to the result of Fgf10-KO chimeras, the lung chimerism observed in the Fgfr2b-KO chimeras displayed a similarity to the tail chimerism. In addition, the distribution of Fgfr2b-KO cells exhibited a highly similar pattern across various tissues (Fig. S3D,E). Therefore, we estimated general chimerism of Fgfr2b-KO chimera from the tail chimerism. Fgfr2b-KO chimeras with less than 3.8% WT cell contribution showed defects in the lungs and forelimbs (Fig. 1B,C). Unlike the Fgfr2b-KO phenotype described in a previous study (De Moerlooze et al., 2000), hindlimb defects were not observed in our model. Our Fgfr2b-KO model detected an epithelium defect in salivary gland but not in thyroid gland in E14.5 embryos (Fig. S3F). Forelimbs and lungs were observed in chimeras with a WT cell contribution of more than 9.8% (Fig. 1B,C), indicating that 10% contribution of WT cells derived from a WT embryo could support forelimb and lung development in an Fgfr2b-KO model.

Next, RFP⁺ cells derived from mutant ESCs and RFP^– cells derived from WT embryos were sorted from chimeric lungs at E14.5, and the Acta2 and E-Cad (Cdh1) expression was examined. Acta2, which is expressed in smooth muscle cells that are differentiated from the lung mesenchyme, was detected in the RFP⁺ group of Fgf10-KO and Fgfr2b-KO derived cells (Fig. 1D,E). This indicated that Fgf10-KO and Fgfr2b-KO cells could differentiate into smooth muscle cells. E-Cad, which is expressed in lung epithelial cells, was also detected in the RFP⁺ group of Fgf10-KO derived cells but not in that of Fgfr2b-KO derived cells (Fig. 1D,E). This suggests that Fgfr2b-KO cells did not contribute to the lung epithelium. To further investigate, sections of Fgf10-KO and WT chimeric lungs were immunostained for mesenchymal cell-derived tissues, such as smooth muscle or vasculature with smooth muscle actin (SMA), or for endomucin antibody, respectively. Fgf10-KO and Fgfr2b-KO cells contributed to the smooth muscle and vasculature cells (Fig. 1F; Figs S2E, S3G). We also immunostained the epithelial tissues using antibodies against E-cadherin. Fgf10-KO cells contributed to the lung epithelium (number of lung epithelium samples without Fgf10-KO cells: 0/16 tubules; n=3 animals) (Fig. 1G; Fig. S2E). These results indicated that the Fgf10-KO cells could contribute to most of the lung tissues. In contrast, Fgfr2b-KO cells did not contribute greatly to lung epithelial cells (number of lung epithelium samples without Fgfr2b-KO cells: 77/81 tubules; n=3 animals) (Fig. 1G; Fig. S3G). However, even in the four cases of lung epithelium to which Fgfr2b-KO cells contributed, only a few Fgfr2b-KO cells were identified (Fig. S3G). The Hematoxylin and Eosin (H&E) staining of lung in Fgfr2b-KO chimera showed normal morphology compared with WT lung (Fig. S3H). These results indicate that the Fgfr2b-KO model provided an organ niche for lung epithelial tissues, and 10% contribution of WT cells into the empty organ niche was necessary for lung development.

To explore a novel organ generation technology, we next generated mutant and WT chimeras from two types of ESCs (mouse Fgfr2b-KO and WT) using the tetraploid-based organ-complementation (TOC) method (Kobayashi et al., 2015). We first injected RFP-expressing mouse Fgfr2b-KO ESCs into tetraploid embryos at the E2.5 stage, followed by GFP-expressing mouse WT (G-mWT) ESCs injected at the E3.5 stage (Fig. 2A). After transplantation of the chimeric embryos, we obtained E14.5 chimeric fetuses, which contain RFP⁺ and GFP⁺ cells (n=3; Table 3). Success of tetraploid complementation was verified by analyzing whether tissue is only composed of RFP⁺ and GFP⁺ cells in various tissues, compared with injection in the diploid (2N) embryo (Fig. S4A,B). Similar to the chimeras obtained using the rBC method, defects in the forelimbs and lung epithelium were complemented by cells derived from G-mWT ESCs (Fig. 2B). We examined whether the lung epithelium was complemented with G-mWT ESCs by immunostaining with an E-cadherin antibody. In the chimeric lungs, the epithelial tissue was found to be complemented by GFP-expressing cells (number of lung epithelium without Fgfr2b-KO cells: 12/12 tubules; n=3 animals), whereas E-Cadherin⁻ tissue was composed of GFP⁺ and RFP⁺ cells (Fig. 2C). To address whether the resulting lung was functional after birth, we performed a Caesarean section at E19.5, and obtained chimeras (n=3; Table 3). Chimeras showed a normal appearance, but the non-chimeras that did not contain GFP⁺ cells showed cyanotic skin color (Fig. 2D). Lungs were not present in the non-chimera, as expected. GFP-expressing lungs were present in chimeras and the lung showed almost exclusively RFP⁺ and GFP⁺ cells (Fig. 2E), whereas the RFP⁻ and GFP⁻ cells were observed in the lungs of the diploid (2N) derived chimeras (Fig. S4C). The reconstituted lung epithelium showed differentiation lung epithelial cell markers such as SftpB, SftpC and Aqp5 (Fig S4D,E). This indicates that the mouse WT ESCs were able to rescue the lung epithelium defect in the Fgfr2b-KO model through the TOC method.

As the mouse WT ESCs ameliorated the lung epithelial defect in the Fgfr2b-KO mouse model, we next examined whether rat lungs could be formed in the Fgfr2b-KO lung-deficient mouse model using the TOC method. We first introduced GFP-expressing rat WT (G-rWT) ESCs into WT mouse embryos at E3.5 and analyzed the chimeras at E14.5 to investigate the extent of rat cell contribution to mouse development and their potential to overcome the lung epithelial defect in the Fgfr2b-KO model. In the mouse-rat interspecies chimeras, rat cells exhibited varying contributions to different tissues, with a notable predilection for lung tissue, as previously reported (Yamaguchi et al., 2018) (Fig. S5A-C). Therefore, we hypothesized that the lung epithelial defect of Fgfr2b-KO mice could be complemented by rat cells. We then injected RFP-expressing mouse Fgfr2b-KO ESCs into tetraploid embryos at the E2.5 stage, followed by G-rWT ESCs injected at the E3.5 stage (Fig. 3A). At E14.5, we observed successful lung formation in the interspecies chimeric fetuses (Fig. 3B; Fig. S6A) (n=3; Table 4), although the lung size was smaller compared with the mouse lung (Fig. S6B). Success of tetraploid complementation with rat ESCs was verified by assessing that tissues were exclusively composed of RFP⁺ and GFP⁺ cells (Fig. S6C). This indicated that the defective lung phenotypes from Fgfr2b-KO mice could be rescued even with rat cells. We also confirmed that most of the tubular structures in interspecies chimeric lungs were complemented by rat cells at E14.5 (number of lung epithelium without Fgfr2b-KO cells: 68/75 tubules; n=3 animals) (Fig. 3C; Fig. S6D). Among the six cases of lung epithelium to which Fgfr2b-KO cells contributed, only a few Fgfr2b-KO cells were identified. In only one case of lung epithelium to which Fgfr2b-KO cells contributed were Fgfr2b-KO cells observed in part of epithelium (Fig. S6D). To examine whether the lung rescued by rat WT cells was functional after birth, we performed a Caesarean section at E19.5 and obtained chimeras (n=3; Table 4). Success of tetraploid complementation was verified by analyzing whether mouse CD45⁺ splenocyte was not composed of non-fluorescent cells (Fig. S7A). Pups with GFP fluorescence (rat chimera) showed cyanotic skin color similar to the pups that did not show GFP signal (non-chimera) (Fig. 3D; Fig. S7B). Pups without GFP died within 21 min, on average, due to respiratory failure, whereas the Fgfr2b-KO chimeras with G-mWT cells survived for more than 5 h (Fig. 3E; Movie 1). The rat chimeras showed postnatal mortality within 13 min on average, even though they exhibited GFP-expressing lungs (Fig. 3E,F; Fig. S7C). The reconstituted lung was a similar size as the mouse lung at P0 (Fig. S7D,E). Similar to the result at E14.5, GFP⁺ rat cells but not Fgfr2b-KO cells could contribute to the Nkx2.1-expressing lung epithelial cells, whereas both rat cells and Fgfr2b-KO cells could contribute to SMA⁺ smooth muscle cells (Fig. S8A,B). We also conducted RT-PCR to investigate whether the lung epithelial cells originated from rat cells. The results showed that mouse epithelial markers such as Epcam, SftpB, SftpC and Aqp5 were barely detected, whereas rat epithelial cell markers were detected in the rat-rescued lung (Fig S8C). On the other hand, mouse Acta2, marker of mesenchyme-derived cells, was detected in the rat-rescued lung (Fig S8C). Furthermore, the rat lung epithelium showed differentiation lung epithelial cell markers such as SftpB, SftpC and Aqp5 (Fig. S8C,D). These results showed that rat WT ESCs did generate lung epithelial cells but not lung mesenchyme cells in the Fgfr2b-KO mouse model with the TOC method; however, the generated lungs were not functional after birth.

As the morphology of the lungs in the rat chimera was not the same as that in the mouse chimera for the Fgfr2b-KO model (Fig. 2E versus Fig. 3F), we further analyzed the generated lungs. Histology of lungs composed of Fgfr2b-KO and G-mWT cells revealed normal saccular expansion and septal thinning, similar to that of the mouse WT control, suggesting that the G-mWT ESCs could fully compensate for the lung dysfunction in the Fgfr2b-KO model (Fig. 4A). In contrast, histological analysis revealed that lungs from Fgfr2b-KO and G-rWT cells showed abnormal alveolar expansion with smaller airspaces and much smaller alveoli, indicating that lung was not matured compared with the intraspecies model, and that no surfactant protein was secreted and the lungs failed to inflate with air (Fig. 4A,B). To examine whether the histological abnormality of the lung from Fgfr2b-KO and G-rWT chimeras was from dissection timing, we dissected WT mouse at 10 min after Caesarean section. However, we did not observe a difference in air space between the WT mouse at 5 h and 10 min after C-section (Fig. 4A,B). This suggested that the abnormality of the lung from Fgfr2b-KO and G-rWT chimeras was not due to dissection timing. As rat development was delayed by 2 days at this stage compared with mouse development, we further investigated the immaturity of the lungs of the Fgfr2b-KO and rWT chimeras. As Sox9, a marker of lung distal epithelial progenitor cells, is mainly expressed in the epithelial progenitor cells from E11.5-E16.5 and is hardly detectable by E18.5 in mouse (Liu and Hogan, 2002; Okubo et al., 2005), we immunostained Sox9 expression in the rat lungs at E18.5-E20.5, which is thought to correspond to E16.5-E18.5 in mice. Similar to in the mouse, we rarely detected Sox9⁺ cells in E20.5 rat lung, but readily detected Sox9⁺ cells in E18.5 and E19.5 rat lung (Fig. S9A). We examined Sox9 expression in the lungs of Fgfr2b-KO and G-mWT chimeras, and could not detect Sox9⁺ progenitor cells in the lungs at P0 (E19.5) (Fig. 4C; Fig. S9B). However, in the lungs of Fgfr2b-KO and G-rWT chimeras, numerous Sox9⁺ progenitor cells were detected in the epithelial cells derived from rat cells at P0 (E19.5) (Fig. 4C; Fig. S9B-D). Furthermore, we explored the lung maturation marker in lung epithelium using the LungDTC database (Du et al., 2017), and found that Scnn1g and Mia expression in rat lung may reflect the developmental timing of lung epithelium (Fig. 4D; Fig. S9E). The analysis of the lung of Fgfr2b-KO and G-rWT chimera showed that rat Scnn1g expression is similar to the expression of E19.5 rat lung but not P0 (E21.5) rat lung (Fig. 4D). In contrast, rat Mia expression is not close to the expression of E19.5 rat lung (Fig. 4D). Furthermore, the expression of mouse Scnn1g and mouse Mia expression was clearly less in the lung of Fgfr2b-KO and G-rWT chimera (Fig. 4D). These results suggest the intriguing possibility that species-specific developmental timelines are selectively preserved within specific cell types or are contingent upon particular genes. Together, the Fgfr2b-KO and rat WT chimeras could not breathe after birth, likely because rat lung epithelial cells preserved their own developmental timing even in the mouse body.

The production of organs by blastocyst complementation has been highlighted as one of the most promising regenerative medicine platforms. One of the remaining problems is the lack of knowledge regarding successful production of PSC-derived organs. In this study, we applied an rBC method to evaluate an organ-deficient model for the production of lungs, establishing an important benchmark that will allow us to evaluate whether a given organ-deficient model can provide an appropriate organ niche. Consequently, we clarified the success conditions for generating lungs through the rBC method and achieved the production of lungs with rat cells in a mouse lung-deficient model even with the TOC method.

Here, we demonstrate the effectiveness of the rBC method for evaluation of organ-deficient model. rBC has been applied in previous studies to analyze whether the abnormalities caused by gene knockout in mutant embryos are intrinsic defects due to their gene functions (cell-autonomous) or extrinsic defects due to the surrounding microenvironment (cell non-autonomous) (Porter et al., 1997; Shawlot et al., 1999; Tremblay et al., 2000; Kanai-Azuma et al., 2002). The development of lymphocytes, hematopoiesis, liver and heart was also assessed using chimeras generated from mutant ESCs and WT embryos (Chen et al., 1993b; Tsai et al., 1994; Sturzu et al., 2015; Bort et al., 2005). Compared with the blastocyst complementation method, the rBC and TOC method provides several advantages for evaluating the organ-deficient model. First, the analysis of organ-deficient WT chimera embryos in the conventional blastocyst complementation method was inefficient owing to the difficulty in genotyping gene-deficient embryos, which occur in only 25% of genetically modified heterozygous mouse crosses due to the presence of WT cells. Moreover, the blastocyst complementation method using the CRISPR/Cas9 genome editing system (Wu et al., 2017) also requires the removal of WT cells for genotyping and, in this case, the possibility of mosaicism must be eliminated. In the rBC and TOC method, only mutant PSCs need to be established to analyze organ-deficient WT chimeras, and organ-deficient WT chimeras can be obtained from all embryos injected with mutant PSCs. In addition, using CRISPR-Cas9 technology, it is possible to obtain mutant PSCs containing multiple and complex mutations with high efficiency (Oji et al., 2016). Second, in blastocyst complementation, if the gene plays a pivotal role in extra-embryonic tissue, injected PSCs cannot compensate for the abnormality. As PSCs only differentiate into the three embryonic germ layers but not extra-embryonic lineage, the abnormality from gene mutation occurs only in the embryonic tissue in the rBC and TOC method. Third, in the blastocyst complementation method, injected PSCs usually express fluorescent proteins (Mori et al., 2019; Kitahara et al., 2020), so their distribution in the tissue can be easily tracked. However, if the contribution of PSC-derived cells is high, it is difficult to determine the contribution of mutant-derived cells in the tissue, unless the host cells also express a fluorescent protein (Usui et al., 2012; Goto et al., 2019; Mori et al., 2019; Kitahara et al., 2020). In contrast, with the rBC and TOC method, the distribution of mutant PSC-derived cells in the tissue can be readily ascertained, even if the mutant cell contribution is low.

We investigated the lungs of Fgf10-KO or Fgfr2b-KO and WT chimeras using the rBC method. Recently, Kitahara et al. (2020) generated lungs in an Fgf10-KO mouse model using blastocyst complementation. Consistent with our results, they showed that Fgf10-KO cells were included in most cell types in the lungs. However, they observed that WT PSC-derived cells were primarily detected in the epithelial cells of the generated lungs, which we did not observe. In the Fgf10-KO chimeras, Fgf10-KO cells were also found in the mesenchymal tissue of the lungs. This may be because the coexisting WT lung mesenchymal cells can secrete Fgf10 and the lung epithelial cells can receive the signal, allowing the lungs to continue to develop and thus Fgf10-KO cells to survive without Fgf10 secretion. Mori et al. (2019) also showed that mouse PSCs could generate functional lungs in an Fgfr2 conditional KO mouse model through the blastocyst complementation method, and the epithelial tissues of generated lungs were highly populated by PSC-derived cells compared with lung mesenchyme tissue. Consistent with their results, we observed that Fgfr2b-KO cells could not contribute to the lung epithelium through the rBC method. In our model, hindlimb defects were not observed. Thus, it may indicate that Fgfr2b-KO model used in this study did not delete sequences important for hindlimb development, which may have been the case in the previous study. In summary, we were able to demonstrate the feasibility of the rBC method to evaluate the contribution of WT cells to organs without generating organ-deficient mice.

We found that WT ESCs make relatively uniform contributions to various tissues of E14.5 embryos. This suggests that the overall chimerism resulting from ESC contribution remains consistent throughout development. This enabled us to make a rough estimation of the chimerism in the chimera by using the chimerism in non-target organs, such as tail. In addition, it was noted that not only WT cells, but also Fgf10-KO and Fgfr2b-KO cells, exhibited comparable contributions to tissues in E14.5 embryos. Considering the inability of Fgfr2b-KO cells to contribute into the lung epithelium, it was rather unexpected to find a comparable level of Fgfr2b-KO cell contribution in the lung, as observed in the tail and other tissues. This suggests that Fgfr2b-KO cells may have a greater propensity to contribute to cell populations other than the lung epithelium. Based on this estimation, the presence of a certain number of normal cells was necessary to overcome the phenotype of lung agenesis in Fgf10-KO or Fgfr2b-KO models, so we concluded that the Fgf signaling in lung development is required in only ∼10% of WT cells. If the presence of a certain number of normal cells is also important for the formation of kidney, then rat cells might not be able to rescue the mouse kidney agenesis model (Usui et al., 2012), because we realized that rat cells cannot contribute much to the mouse kidney in this study. In the future, we will need insight into the percentage of WT cells required to produce the organs of each tissue using the rBC method.

In this study, the tetraploid complementation method was used, which can yield fully ESC-derived organisms (Nagy et al., 1993) to obtain chimeras derived from two different types of PSCs – WT ESCs and KO ESCs – and successfully generated rat lungs in the Fgfr2b-KO mouse model. In the experiment using the tetraploid embryo, we obtained surviving embryos [Fgfr2b-KO ESCs with mouse WT ESCs 8.5% (7/82) or Fgfr2b-KO ESCs with rat WT ESCs 7.0% (18/256)] at E14.5 or full-term pups [Fgfr2b-KO ESCs with mouse WT ESCs 9.4% (6/64) or Fgfr2b-KO ESCs with rat WT ESCs 7.2% (32/446)] at P0, which is a similar success rate to a previous study (Yagi et al., 2017). However, the low frequency of obtaining rat chimeras, similar to that in a previous study (Kobayashi et al., 2010), suggests that further improvements are needed to overcome this problem, such as restricting the contribution of rat PSCs to specific tissues only. Also, the success of tetraploid complementation has thus far only been reported in mice and rats (Li et al., 2017), and further research is deemed necessary to implement the TOC method in other species. Although higher WT contribution in the defective organ was expected to be one of the crucial factors when interspecies chimeras are produced for organ generation (Okumura et al., 2019), we and others (Yamaguchi et al., 2018) found that rat PSCs were likely to contribute to mouse lung tissue. In addition, we realized that only ∼10% of WT cells were required to prevent lung agenesis in the Fgfr2b-KO model in this study. Indeed, we succeeded in overcoming the lung defect in Fgfr2b-KO mice with rat PSCs even through the tetraploid complementation method. The size of lungs complemented by rat PSCs was smaller than that of lungs complemented by mouse PSCs at E14.5. In particular, the lung size was smallest in Fgfr2b-KO and rat WT chimeras, with the lowest percent contribution from rat cells. As lung size is determined by the number of lung epithelial progenitor cells (Sui et al., 2019), this may indicate that lung size is also determined by the number of rat lung epithelial progenitor cells in interspecies chimera. However, the size of rat rescued lung is almost same as the mouse lung at P0, suggesting that the growth of the lung may be complemented during the development.

Intriguingly, in one of the resulting rat lungs at postnatal stage (Fig. 3F; chimera #1), the rat epithelial progenitor cells remained immature, even if the lung was composed almost entirely of rat cells. This result suggests that rat lung epithelial cells retain their own developmental speed despite their presence in a different species, or that rat lung mesenchymal cells, including lung epithelial cells, are delayed compared with mouse development. In another example (Fig. 3F; chimera #2, #3), rat lung epithelial cells were still in an immature state, even though mouse cells were detected in the mesenchymal cells. This result raises the possibility that rat epithelial cells are unable to receive signals from mouse mesenchymal cells. Similar to this observation, rat germ cells, even when present in mouse testes, have been shown to differentiate at the time typical for rats and therefore generate the structural pattern of rat spermatogenesis (Franca et al., 1998). This suggests that some cells exhibit intrinsic regulation of differentiation, even in interspecies environments. However, to control the rat chimerism in the lung, a comprehensive lung cell-deficient mouse model is required, encompassing not only epithelial-derived cells but also mesenchyme-derived cells. Using this model is essential to determine whether developmental timing is intrinsically determined or influenced by external factors. In summary, the generation of functional lungs in interspecies chimeras may require overcoming the barrier of species-specific intrinsic developmental timing.

Although rat lungs were formed in the Fgfr2b-KO mouse model, they were not functional after birth because the rat lung epithelial tissue might be still in the prenatal stage. Balancing the proliferation and differentiation of lung epithelial progenitor cells requires fine control of Sox9 expression levels (Rockich et al., 2013). Thus, modulation of Sox9 may be important for generating functional lung in interspecies blastocyst complementation. Sox9 expression is regulated by Nmyc (Mycn) or Asxl1 (Okubo et al., 2005; Moon et al., 2018). Therefore, it may be important to modulate the expression of Nmyc or Asxl1 to regulate the intrinsic developmental timing of lung epithelial progenitor cells. In this study, we searched the genes that reflect developmental timing of lung epithelial cells from LungDTC database (Du et al., 2017). In the future, it may be necessary to determine the factors that are related to the intrinsic developmental timing of lung epithelial progenitor cells in xenogeneic lung using single cell RNA-seq analysis. Moreover, future studies comparing species with different developmental speeds other than the mouse-rat combination would be useful to better understand species-specific developmental time.

In summary, our analysis provides evidence that the Fgfr2b-KO demonstrates a lung epithelial-deficient model using the rBC method. With the interspecies TOC method, we firstly propose that rat ESCs potentiate the ability to rescue the lung agenesis in mouse. Furthermore, our findings may indicate that respecting species-specific developmental timing is a key point for generating functional lungs in interspecies blastocyst complementation.

All animal experiments were conducted in accordance within the guidelines of the Regulations and By-Laws of Animal Experimentation at the Nara Institute for Science and Technology and were approved by the Animal Experimental Committee at the Nara Institute of Science and Technology (approval numbers 1639 and 2109). The animal experiments in this study were performed in compliance with ARRIVE guidelines (Kilkenny et al., 2010). Institute of Cancer Research (ICR) mice were purchased from Japan SLC.

ICR females aged 8-10 weeks were treated with CARD HyperOva (Kyudo) and human chorionic gonadotropin (hCG; ASKA Animal Health) for superovulation, then were mated with ICR male mice. Two-cell-stage zygotes were collected from female oviducts 42-46 h after hCG injection using the flush-out method. The collected two-cell-stage embryos were incubated in KSOM medium at 37°C under 5% CO₂ conditions until use.

Oligo DNAs of the target sequence were ligated into the BbsI site of the pSpCas9(BB)-2A-Puro (pX459) V2.0 plasmid (Addgene plasmid #62988). The combination of #1 and #2 or #3 and #4 oligos was used to establish Fgf10-KO ESCs, and the combination of #8 and #9 or #10 and #11 oligos was used to establish Fgfr2b-KO ESCs (Table S1). All oligonucleotides were designed using the CrisperDirect website to identify specific target sites (Naito et al., 2015).

Fgf10-KO or Fgfr2b-KO ESCs were established using the R01-09 ESC line, which was newly established from 129X1 and R01 F1 embryos. R01 mouse lines were established through the tetraploid complementation method from R01 ESCs obtained from Dr Masahito Ikawa (Osaka University, Japan), established from 129 and BDF1 F1 embryos and constitutively expressing RFP by the CAG-Su9-DsRed2 transgene, which localizes to the mitochondria. R01-09 ESCs were seeded on mouse embryonic fibroblasts (MEFs) and then transfected with the two designed plasmids (Figs S2A, S3A) using Lipofectamine 3000 (Thermo Fisher Scientific). Transfected cells were selected by transient treatment with 1 μg/ml puromycin for 2 days, and then ESC colonies were subjected to genotyping with PCR and sequencing. ESCs were cultured on gelatin-coated dishes, and MEFs with N2B27 medium supplemented with 3 μM CHIR99021, 1.5 μM CGP77675, and mouse LIF (leukemia inhibitory factor) (N2B27-a2i/L medium) (Choi et al., 2017). Detail cell sources are stated in Table S3.

Genotyping primers for detecting Fgf10-KO and Fgfr2b-KO ESCs are shown in Table S1. DNA fragments were amplified using GoTaq (Promega) for 40 cycles to detect null or WT alleles under the following conditions: 94°C for 30 s, 60°C for 30 s and 72°C for 60 s.

All chimeric embryos were recovered at the E14.5 stage. Tail, kidney, lung, stomach and intestine samples were incubated with 0.25% trypsin for 10 min at 37°C. After pipetting to dissociate the tissue, 10% fetal bovine serum in PBS was added, and samples were filtered through a 37 μm mesh. The FL3 detector on Accuri Flow Cytometer (BD Bioscience) or the ECD detector on the CytoFLEX S Flow Cytometer (Beckman Coulter) were used to detect RFP⁺ populations. Tail samples were used to estimate general chimerism on Fgf10-KO or Fgfr2b-KO chimera. An MA900 Multi-Application Cell Sorter (Sony) was used to sort between the RFP⁺ and RFP⁻ subpopulations. For splenocyte analysis, cells isolated from spleen were exposed to RBC lysis buffer (BioLegend, 420302) to remove red blood cells and stained with APC-anti-mouse CD45 (BioLegend, 103112, 1:100). CD45⁺/RFP⁺ or CD45⁺/RFP⁻ populations were analyzed with the CytoFLEX S Flow Cytometer.

Total RNA was purified using Trizol reagent (Thermo Fisher Scientific) and used for reverse transcription (RT). cDNA was prepared using the SuperScript IV VILO master mix (Thermo Fisher Scientific). RT-PCR was performed using GoTaq (Promega). For quantitative RT-PCR analysis, Luna Universal qPCR Master Mix (New England Biolabs) was used to amplify the DNA fragment, and amplified DNA was detected on a LightCycler 96 (Roche). Species specificity of all primer sets was assessed by the amplification curve and melting curve from the qPCR result. The primers used for RT-PCR are described in Table S2.

Tetraploid embryos were prepared as described previously (Okada et al., 2007; Kishimoto et al., 2021). In brief, ICR two-cell-stage embryos were placed in fusion buffer (0.3 M mannitol, 0.3% bovine serum albumin and 0.1% polyvinylpyrrolidone), and electrofusion was performed using CFB16-HB and LF501PT1-10 electrodes (BEX Co). Tetraploid embryos were incubated in KSOM at 37°C under 5% CO₂ until use. Then, 6-8 cells of Fgfr2b-KO ESCs were injected into tetraploid embryos at E2.5. These embryos were cultured until the E3.5 stage and injected into 6-8 cells of GFP-expressing rat ESCs (rG104) (Isotani et al., 2016), followed by transfer into the uterus of E2.5 pseudopregnant ICR mice. The quality of rat ESCs have been confirmed with teratoma formation assay or germline transmission (Isotani et al., 2016; Isotani et al., 2017). The fetuses were recovered and dissected at E14.5, and offspring were recovered at E19.5 via Caesarean section. The ESC-derived offspring were analyzed using RFP or GFP signal under a fluorescence stereo microscope (MZ FLIII; Leica).

The lungs, salivary glands or thyroid glands at E14.5 or E19.5 were fixed with 4% paraformaldehyde (PFA) in phosphate buffered saline [PBS (−); Nacalai] for 15 min at 25°C or overnight at 4°C. After washing with PBS, the tissues were immersed in 10, 20 and 30% sucrose. The treated tissues were then sunk into Tissue-TeK O.C.T. compound (Sakura Finetek). After sectioning at 10 μm with a cryostat (NX70; Leica), the slides were dried at 25°C, followed by washing with PBS (−). Slides were immersed in ethanol and 4% PFA solution (1:1) for 15 s and washed with ddH₂O, then treated with Mayer's Hematoxylin solution (Wako Chemicals) for 1 min and washed with PBS (−) and ddH₂O. Next, slides were immersed in 0.5% Eosin solution (Wako Chemicals) for 1 min and washed three times with 100% ethanol, then twice more in xylene. Mountquick (Daido Sangyo) was added to the slides and samples were covered with cover glass. The sections were observed under a microscope with 10× objective lens (BX60; Olympus). For immunostaining, slides were treated with 1% bovine serum albumin (BSA; Sigma-Aldrich) for 60 min at 25°C. The primary antibody was incubated overnight at 4°C. The slides were then washed three times for 5 min with PBS (−) and incubated with the secondary antibody for 1 h at 25°C. After washing three times for 5 min with PBS (−) at 25°C, the nuclei were stained with Hoechst 33342 (Dojindo, KV072) diluted to 1:1000 in PBS for 30 min at 25°C before a final PBS (−) wash. The antibodies used included: rabbit anti-E-Cadherin (Cell Signaling Technology, 3195, 1:100), rat anti-E-Cadherin (Takara, M110, 1:100), mouse anti-E-Cadherin (BD Biosciences, 610182, 1:100), rabbit anti-Nkx2.1 (TTF1) (Cell Signaling Technology, 12373S, 1:100), rat anti-Endomucin (Santa Cruz Biotechnology, sc-65495, 1:100), mouse anti-SMA (BioLegend, 904601, 1:100), rabbit anti-Sox9 (Millipore, AB5535, 1:100), rabbit anti-SftpC (Proteintech, 10774-1-AP, 1:100), Alexa647-anti-mouse EpCam (BioLegend, 118212, 1:50), goat Alexa Fluor 488 anti-rabbit IgG (Thermo Fisher Scientific, A11017, 1:1000), goat Alexa Fluor 555 anti-rabbit IgG (Thermo Fisher Scientific, A21430, 1:1000), goat Alexa Fluor 647 anti-rabbit IgG (Thermo Fisher Scientific, A21246, 1:1000), goat Alexa Fluor 488 anti-mouse IgG (Thermo Fisher Scientific, A11017, 1:1000), goat Alexa Fluor 555 anti-mouse IgG (Thermo Fisher Scientific, A21425, 1:1000), goat Alexa Fluor 647 anti-mouse IgG (Thermo Fisher Scientific, A21237, 1:1000), goat Alexa Fluor 488 anti-rat IgG (Thermo Fisher Scientific, A11006, 1:1000) and goat Alexa Fluor 647 anti-rat IgG (Thermo Fisher Scientific, A21247, 1:1000). Immunostained samples were examined using a laser confocal microscope (LSM700, LSM710, LSM980; Zeiss). In the case of separating DsRed and Alexa Fluor 555, linear unmixing method was applied. In the epithelial structures with E-Cad⁺ cells and lumen structure, the number of those with RFP⁺ cells were counted. See Table S3 for details of key reagents and resources.

The air area fraction at E19.5 was measured from lung sections stained with H&E. More than three non-overlapping fields (×10 objective lens) from each lung sample were analyzed. The percentage of air space in the total distal lung area was analyzed using ImageJ software.

To search for marker genes that reflect the developmental stages of the lung epithelium, we sorted genes mainly expressed in the epithelium in Lung-GENS (Du et al., 2015). Then, we compared each gene expressed in the epithelium with E18.5 and P1 in Lung Developmental Time Course Analysis (Lung-DTC) in the LGEA database (Du et al., 2017).

For air space measurement, all values are expressed as mean±s.d. from at least three different regions in each sample. For quantitative RT-PCR data expressed as relative fold changes, all values are expressed as mean±s.d. from at least triplicate experiments. Student's t-test for unpaired comparisons was performed and results at P<0.01 were considered statistically significant.

We thank members of the Isotani lab for helpful discussions. The core facility at the Nara Institute of Science and Technology (NAIST) were also instrumental to this work. We thank Dr Ikawa (Osaka University) for kindly providing R01 ESCs. We also thank Editage (www.editage.com) for English language editing.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.202081.reviewer-comments.pdf