Bombesin-like peptides modulate alveolarization and angiogenesis in bronchopulmonary dysplasia

Abstract

Rationale: The incidence of bronchopulmonary dysplasia (BPD), a chronic lung disease of newborns, is paradoxically rising despite medical advances. We demonstrated elevated bombesin-like peptide levels in infants that later developed BPD. In the 140-day hyperoxic baboon model of BPD, anti-bombesin antibody 2A11 abrogated lung injury.

Objectives: To test the hypothesis that bombesin-like peptides mediate BPD in extremely premature baboons (born at Gestational Day 125 and given oxygen pro re nata [PRN], called the 125-day PRN model), similar to "modern-day BPD."

Methods: The 125-day animals were treated with 2A11 on Postnatal Day 1 (P1), P3, and P6. On P14 and P21, lungs were inflation-fixed for histopathologic analyses of alveolarization. Regulation of angiogenesis by bombesin was evaluated using cultured pulmonary microvascular endothelial cells.

Measurements and main results: In 125-day PRN animals, urine bombesin-like peptide levels at P2-3 are directly correlated with impaired lung function at P14. Gastrin-releasing peptide (the major pulmonary bombesin-like peptide) mRNA was elevated eightfold at P1 and remained high thereafter. At P14, 2A11 reduced alveolar wall thickness and increased the percentage of secondary septa containing endothelial cells. At P21, 2A11-treated 125-day PRN animals had improved alveolarization according to mean linear intercepts and number of branch points per millimeter squared. Bombesin promoted tubulogenesis of cultured pulmonary microvascular endothelial cells, but cocultured fetal lung mesenchymal cells abrogated this effect.

Conclusions: Early bombesin-like peptide overproduction in 125-day PRN animals predicted alveolarization defects weeks later. Bombesin-like peptide blockade improved septation, with the greatest effects at P21. This could have implications for preventing BPD in premature infants.

Figure 1.

Elevated urine bombesin-like peptide levels in 125-day PRN (pro re nata) baboons. The 24-hour urine bombesin-like peptide levels at 24 to 48 hours are directly correlated with the mean oxygenation index at 13 to 14 days in 125-day PRN animals. The 24-hour urine specimens were collected at 24 to 48 hours from 125-day PRN animals without intervention (n = 7). A linear correlation was the best fit for the points, with R² = 0.789, R = 0.89, P < 0.0075.

Figure 2.

Increased gastrin-releasing peptide (GRP)–immunopositive pulmonary neuroendocrine cells in 125-day PRN baboons. In the 125-day baboon model of bronchopulmonary dysplasia, there is marked hyperplasia of pulmonary neuroendocrine cells. (a) Immunostaining for bombesin-like peptide demonstrated numerous GRP-positive neuroendocrine cells, which were mainly localized to small bronchioles and alveolar ducts. A small GRP-positive cluster of neuroendocrine cells is demonstrated in a bronchiole. Bar in lower left hand corner = 25 μm. (b) A serial section immediately adjacent to that shown in (a) was immunostained in parallel with the same anti-bombesin antibody (2A11) preabsorbed with human GRP as described in the online supplement Methods. Note that the immunostaining shown in (a) was abrogated, indicating the antigen specificity of this staining (arrow). Bar in lower left hand corner = 25 μm. (c) In 125-day PRN baboons, numerous GRP-positive cells are observed scattered throughout the small bronchioles and alveolar ducts (arrows). Bar in lower left hand corner = 50 μm. (d) The dotted-line box in the upper right hand corner of (c) is shown enlarged in (d), with multiple nuclei and abundant cytoplasm observed in one representative cluster. Bar in lower left hand corner = 25 μm. (e) In contrast, in a 2A11-treated 125-day PRN baboon, GRP-positive cells are only infrequently identified in small bronchioles (arrow). Bar in lower left hand corner = 25 μm. (f) The majority of the GRP-positive positive neuroendocrine cells in 2A11-treated baboons occur in larger bronchioles and in cartilaginous airways (arrows). Bar in lower left hand corner = 50 μm. Cartilage is visible just below the scale bar. (g) GRP-positive pulmonary neuroendocrine cells were counted in entire tissue sections using computerized image analysis with Scion Image 6.2 (Scion Corp.), normalized for the area of lung tissue present because cells were scattered in small bronchioles and alveolar ducts in all groups. Results are expressed as the number of pulmonary neuroendocrine cells per millimeter squared of lung tissue. Numbers of animals quantified were: 125-day gestational controls (GCs; 125GC), n = 6; 125-day PRN, n = 6; 125-day PRN+MOPC21, n = 4; 125-day PRN+2A11, n = 5; 140-day GCs (140GC), n = 5. PNECs = pulmonary neuroendocrine cells. *P < 0.0002 compared with 125-day or 140-day GCs; ^†P < 0.0003 compared with 125-day PRN without any intervention; **P < 0.02 compared with 125-day GCs, 140-day GCs, or 125-day PRN +2A11. There was no significant difference between the 2A11 group and 125- or 140-day GCs (P > 0.27).

Figure 3.

Elevated gastrin-releasing peptide (GRP) mRNA levels in lungs of 125-day PRN baboons. Lung GRP mRNA levels in different groups, as determined by mRNA microarray analyses: 125GC treated with mechanical ventilation PRN for 1 day (n = 3), 125GC treated with mechanical ventilation PRN for 2 days (n = 3), 125GC treated with mechanical ventilation PRN for 6 days (n = 3), 125GC treated with mechanical ventilation PRN for 14 days (n =3), and 140GC (n = 3). *P < 0.005, **P < 0.01, ***P < 0.001, ^¥P < 0.05.

Figure 4.

Alveolar wall thickness in 125-day PRN baboons maintained for 14 days without or with 2A11 or MOPC21. See Figure E4 for a color version of this figure. (a) Section of 125-day PRN animal maintained on mechanical ventilator for 14 days without pharmacological intervention. Lines were drawn perpendicular to alveolar septa. Scale bar in lower left corner, 40 μm. (b) Section of 125-day PRN animal maintained on mechanical ventilator for 14 days treated with 2A11. Lines were drawn perpendicular to alveolar septa. Scale bar in lower left corner, 40 μm. (c) Results of morphometry of alveolar septal wall thickness. Number of animals per group were: 125-day gestational controls (125GC), n = 5; 125-day PRN without intervention, n = 9; 125-day PRN+MOPC21, n = 6; 125-day PRN+2A11, n = 5; 140-day gestational controls (140GC), n = 5. *P < 0.002 compared with 125-day PRN+MOPC21; ^†P < 0.01 compared with 125-day PRN without intervention; **P < 0.0001 compared with 125-day PRN+MOPC21 or 125-day PRN without intervention; ^††P < 0.02 compared with 125-day PRN+2A11.

Figure 5.

α-Smooth muscle actin (SMA)–positive immunostaining of secondary alveolar septal tips. (a) SMA immunostaining of lung sections from a 125-day PRN animal treated with MOPC21 (negative control IgG1). Myofibroblasts (SMA positive) are mostly in the interstitium (representative SMA-positive cells are indicated by arrowheads). Scale bar, 20 μm. (b) SMA immunostaining of lung from a 2A11-treated 125-day PRN animal. Myofibroblasts were concentrated at alveolar septal tips (arrowheads). Scale bar, 20 μm. (c) Pooled morphometric analyses of percentage of SMA-positive secondary septa in: 125-day gestational controls (125GC) (n = 5); 125-day PRN, n = 9; 125-day PRN+MOPC21, n = 6; 125-day PRN+2A11, n = 5; 140-day gestational controls (140GC), n = 5. *P < 0.0002 compared with 140GC; ^†P < 0.001 compared with 125GC, PRN, or PRN +MOPC21.

Figure 6.

CD31 immunostaining in developing secondary septa. (a) CD31 immunostaining of lung from a 125-day PRN animal treated with MOPC21. Secondary septa were mostly CD31 negative (red arrows), with occasional CD31-positive septa (black arrow). Scale bar, 30 μm. Asterisk indicates pleural surface; L = airway lumen. (b) Lung from a 2A11-treated 125-day PRN animal has CD31-positive cells concentrated at alveolar septal tips. (arrows). Scale bar, 30 μm. Asterisk indicates pleural surface. (c) Numbers of CD31-positive secondary alveolar septal tips was determined as given in Methods. Pooled analyses of percentage of CD31-positive secondary septa are shown for: 125-day gestational controls (125GC), n = 5; 125-day PRN without intervention, n = 9; 125-day PRN+MOPC21, n = 6; 125-day PRN+2A11, n = 5; 140-day gestational controls (140GC), n = 5. *P < 0.0002 compared with 125GC, 125-day PRN, or 125-day PRN+MOPC21.

Figure 7.

Bombesin regulates capillary tubulogenesis. (a) Pulmonary microvascular endothelial cells alone on Matrigel (BD Biosciences, San Jose, CA) for 24 hours had scant tubule formation with few branch points. (b) Pulmonary microvascular endothelial cells alone + 1 nM bombesin. Note increased total length of tubules and many branch points. (c) Pulmonary microvascular endothelial cells on Matrigel with fetal lung mesenchymal cells were similar to (a). (d) Pulmonary microvascular endothelial cells on Matrigel with fetal lung mesenchymal cells + 1 nM bombesin had no significant response to bombesin. (e, f) Pooled results of three experiments with pulmonary microvascular endothelial cells (EC) alone or pulmonary microvascular endothelial cells + fetal lung mesenchymal cells (EC + Mes). Bombesin induced greater than fivefold increased capillary tubule length; *P < 0.003 and **P < 0.01 compared with negative control (“Neg” in figure). Bombesin increased branch points per millimeter by more than 50% (f); *P < 0.008 and **P < 0.0005 compared with negative control. All bombesin-induced responses of EC were abrogated by cocultured fetal lung mesenchymal cells (EC + Mes). (g) Pooled results of three assays with pulmonary microvascular endothelial cells plus mesenchymal cells (EC + Mes) ± 2A11 or MOPC21. Values obtained with 2A11 are normalized for the mean values with MOPC21 for each experiment. Numbers of capillary tubule branch points per millimeter of tubule length were increased more than fourfold (**P < 0.0001) and tubule length was increased 40% (*P < 0.05). (h) Results of semiquantitative reverse transcriptase–polymerase chain reaction of pulmonary microvascular endothelial cells (EC) and fetal lung mesenchymal cells (Mes) for gastrin-releasing peptide receptor (GRPR), α-smooth muscle actin (SMA), and glyceraldehyde phosphate dehydogenase (GAPDH).

Figure 8.

Morphometric analyses of alveolar development in 125-day PRN baboons after 14 or 21 days without, or with, 2A11 or MOPC21. See Figure E8 for a color version of this figure. (a) Hematoxylin-and-eosin–stained lung section from a 125-day PRN × 21-day baboon without any intervention. Bar in lower left hand corner = 100 μm. (b) Hematoxylin-and-eosin–stained lung section from a 125-day PRN × 21 days + 2A11. Bar in lower left hand corner = 100 μm. Asterisks (*) indicate pleural surface. (c) Summary of mean linear intercepts for: 125-day gestational controls (125GC), n = 5; 125-day PRN × 14 days without any intervention (14d PRN), n = 9; 125-day PRN × 14 days + MOPC21 (14d MOPC21), n = 6; 125-day PRN × 14 days + 2A11 (14d 2A11), n = 5; 125-day PRN × 21 days without any intervention (21d PRN), n = 7; 125-d PRN × 21 days + 2A11 (21d 2A11), n = 6; 140-day gestational controls (140GC), n = 5. *P < 0.01 compared with 14-day MOPC21; **P < 0.001 compared with 21-day PRN. (d) Pooled data giving the number of branch points per millimeter squared of alveolar surface area for the following groups: 21-day PRN (n = 9), 21-day 2A11 (n = 6), and 146-day GC (age-matched control) (n = 5). *P < 0.05 compared with 21-day PRN. There was no significant difference between 21-day 2A11 and 146-day GC; ^†P < 0.005 compared with 146-day GC. (e) Pooled data for surface density of primary septa (cm²/cm³) for 21-day PRN (n = 9), 21-day 2A11 (n = 6), and 146-day GC (age-matched control) (n = 5). *P = 0.004 compared with 21-day PRN; ^†P < 0.0001 compared with 146-day GC. ^¥P < 0.01 compared with 146-day GC.