Development of a novel humanized mouse model to study bronchopulmonary dysplasia

Abstract

Rationale: The role of circulating fetal monocytes in bronchopulmonary dysplasia is not known. We utilized a humanized mouse model that supports human progenitor cell engraftment (MISTRG) to test the hypothesis that prenatal monocyte programming alters early lung development and response to hyperoxia.

Methods: Cord blood-derived monocytes from 10 human infants were adoptively transferred into newborn MISTRG mice at p0 (1 × 10⁶ cells/mouse, intrahepatic injection) followed by normoxia versus hyperoxia (85% oxygen × 14 days). Lungs were harvested at p14 for alveolar histology (alveolar count, perimeter and area) and vascular parameters (vWF staining for microvessel density, Fulton's index). Human CD45 staining was conducted to compare presence of hematopoietic cells. Murine lung parameters were compared among placebo and monocyte-injected groups. The individual profiles of the 10 patients were further considered, including gestational age (GA; n = 2 term, n = 3 moderate/late preterm, and n = 5 very preterm infants) and preeclampsia (n = 4 patients). To explore the monocyte microenvironment of these patients, 30 cytokines/chemokines were measured in corresponding human plasma by multiplex immunoassay.

Results: Across the majority of patients and corresponding mice, MISTRG alveolarization was simplified and microvessel density was decreased following hyperoxia. Hyperoxia-induced changes were seen in both placebo (PBS) and monocyte-injected mice. Under normoxic conditions, alveolar development was altered modestly by monocytes as compared with placebo (P < 0.05). Monocyte injection was associated with increased microvessel density at P14 as compared with placebo (26.7 ± 0.73 vs. 18.8 ± 1.7 vessels per lung field; P < 0.001). Pooled analysis of patients revealed that injection of monocytes from births complicated by lower GA and preeclampsia was associated with changes in alveolarization and vascularization under normoxic conditions. These differences were modified by hyperoxia. CD45+ cell count was positively correlated with plasma monocyte chemoattractant protein-1 (P < 0.001) and macrophage inflammatory protein-1β (P < 0.01). Immunohistochemical staining for human CD206 and mouse F4/80 confirmed absence of macrophages in MISTRG lungs at P14.

Conclusions: Despite the inherent absence of macrophages in early stages of lung development, immunodeficient MISTRG mice revealed changes in alveolar and microvascular development induced by human monocytes. MISTRG mice exposed to neonatal hyperoxia may serve as a novel model to study isolated effects of human monocytes on alveolar and pulmonary vascular development.

Keywords: chorioamnionitis; fetal monocytes; hematopoietic stem cells; intrauterine inflammation; neonatal lung disease; preeclampsia; preterm birth.

Figure 1

Overview and timeline of experiments. Cord blood monocytes isolated through the workflows and protocols depicted on the far left from 10 individual patients were injected into P0 mice via intrahepatic injection. Mice were immediately placed in hyperoxia (85%) or normoxia (21%) for 14 days. Characteristics of the 10 patients are shown in the table (far right). For 7 of the 10 patients, at least 6–8 pups were injected with patient-specific monocytes. For patients #4, 5, and 7, only 2 pups injected with these monocytes were available for the normoxia group due to smaller paired litters available and lower than expected monocyte counts at the time of sample thaw. In addition, for patients #8 and 9, at least 3 pups were injected for normoxia and hyperoxia groups but all pups died in hyperoxia and 3 total died in hyperoxia before P14, leaving only 2 pups each in the hyperoxia group for lung parameter analyses at P14.

Figure 2

Lung alveolar morphometry and representative images at P14 according to monocyte versus placebo treatment on p0 and normoxia versus hyperoxia exposure (p0–p14). (A–C) Summary of alveolar count, perimeter and area, respectively. Individual datapoints represent each mouse: triangles represent PBS-injected and dots represent monocyte-injected pups. Blue indicates normoxia and red indicates hyperoxia exposure. At p14, mean alveolar area (C) was decreased in mice treated with human monocytes and exposed to hyperoxia. Mean alveolar perimeter (B) but not alveolar count (A) was increased with monocyte transplantation followed by room air. *P < 0.05, **P < 0.01 and ***P < 0.001. (D–E) Representative images of mouse lungs after 14 days of normoxia (D) versus hyperoxia showing typical alveolar simplification (E). Images stained for human CD45 with counterstaining. Images taken at 10× with 20× inlay.

Figure 3

Lung vascular parameters and representative images at P14 according to monocyte versus placebo treatments and normoxia versus hyperoxia exposures. Individual datapoints represent each mouse: triangles represent PBS-injected and dots represent monocyte-injected pups. Blue indicates normoxia and red indicates hyperoxia exposure. (Panel A) small vessel density (measured by vonWillebrand factor stain) was increased in room air-exposed mice treated with monocytes. (Panel B) Fulton's index (ratio of right ventricular weight/left ventricle + septum) was decreased in monocyte-injected group with hyperoxia as compared with normoxia. *P < 0.05, **P < 0.01 and ***P < 0.001.

Figure 4

Representative immunofluorescent images with vWF identifying microvessel density. Note the relative paucity of microvessels (vessel diameter <100 um) in mice treated with placebo (A–B) versus intrahepatic injection of monocytes (C–D).

Figure 5

Lung alveolar morphometry data at P14 according to the 10 individual patients (1-10) and placebo (PBS). Individual datapoints represent mean/median values from 6 non-overlapping fields from each mouse for: alveolar count (A–B), alveolar perimeter (C–D) and alveolar area (E–F). Blue dots indicate normoxia and red dots indicate hyperoxia exposure. Data were analyzed using ANOVA with Bonferroni correction and post hoc analysis to identify patient-specific differences among groups. *P < 0.05.

Figure 6

Lung vascular parameters at P14 according to the 10 individual patients (1–10) and placebo (PBS). Individual datapoints represent each mouse for: mean microvessel count (<100um) from 6-8 images (A–B), and Fulton's index (C–D). Blue dots indicate normoxia and red dots indicate hyperoxia exposure. Data were analyzed using ANOVA with Bonferroni correction and post hoc analysis to identify patient-specific differences among groups. *P < 0.05; **P < 0.01.

Figure 7

Comparison of human CD45 cell counts in MISTRG lung and liver at P14. Individual datapoints represent each mouse: blue indicates normoxia and red indicates hyperoxia exposure. Anti-human CD45 antibody staining was used to quantify human CD45+ hematopoietic cells circulating in the lungs at p14. (Panel A,B) Overall, median human cell count was decreased with hyperoxia exposure in mice treated with the human monocytes. Analysis of liver cell counts (Panel C,D) and ratio of lung to lung + liver counts per each mouse (E,F) revealed no differences in presence of human cells within the 10 patient groups, and no differences between hyperoxia versus normoxia. Missing data in normoxia (blue) for patients 8 and 9 indicate that all pups died prior to p14 and could not be harvested. *P < 0.05.

Figure 8

Representative lung and liver tissue sections with anti-human CD45 antibody staining. (A,B) Lung tissue sections at P14 demonstrating changes in alveolar structure and distribution of human CD45+ stained cells after normoxia (A) versus hyperoxia (B); images taken at 10× with 20× inlay. (C,D) Liver tissue section at P14 after normoxia (C) versus hyperoxia (D); images taken at 10×.

Figure 9

Anti-human CD206 stained MISTRG mouse lung slides showing (A) absence of alveolar macrophages at p14 as compared with (B) positive human control. (C) At p28, there was emergence of positively-stained cells with prominent nuclei similar to human controls, as indicated by red arrows.

Figure 10

Representative images of p14 lungs with F4/80 immunostaining. Lungs harvested at p14 were stained with F4/80 to identify murine lung macrophages. Note the absence of positive staining in both normoxia-exposed (A) and hyperoxia-exposed (B) lungs as compared with positive control lung from an adult wildtype non-humanized mouse lung tissue (C57BL/6J, compliments of Dr. Eniko Sajti) (C). For all slides, biotin labelled secondary antibody used, with counterstain incubation time 30 s Antibody dilution for both primary and secondary was 1:200. Streptavidin-HRP dilution 1:1500. Images taken at 20×.

Figure 11

Cytospin images (40×) of hCD14⁺/hCD206⁺ lung cells isolated from p56 MISTRG lungs. (Panel A,B) Flow cytometry gating by HLA-DR⁺/mouseCD45⁻ followed by FACS for isolation of human CD14⁺/human CD206⁺ cells from lungs exposed to room air or hyperoxia ×14 days followed by recovery in room air until 8 weeks of age. (Panel C,D) Cell surface markers and morphology (inlay) confirm the presence of human alveolar macrophages at p56 in MISTRG mice exposed to room air and early hyperoxia.

Figure 12

RT-qPCR results from MISTRG lungs at p14 showing gene expression changes with hyperoxia. According to plotted Ct values (y-axis) VEGF-A for normoxia and hyperoxia were not remarkably different. Whereas, MCP-1 and MIP-1B lower Ct values were observed in hyperoxia as compared to normoxia, indicating relatively higher levels of RNA (Panel A). Relative to p14, there was down-regulation (as noted by higher Ct value) of mouse MCP-1 and MIP-1B but not mouse VEGF-A after recovery in normoxia at p21 (Panel B). There was emergence of human MCP-1 and MIP-1B expression at p21, which was not seen at p14 (Panel C).