Background
Preterm infants experience a range of debilitating health issues primarily resulting from lung immaturity, and few treatment options are available [
1‐
3]. Owing to lung underdevelopment, preterm infants often require mechanical ventilation with hyperoxic gas in order to survive [
4,
5]. However, high levels of oxygen or prolonged use of ventilators can damage the lungs and interrupt normal alveolar and bronchiolar development, which may lead to chronic lung diseases known as bronchopulmonary dysplasia (BPD) [
6]. It is common, that preterm infants that are born at less than 32 weeks of gestation have increased progression of other short and long-term respiratory illnesses, such as asthma and chronic obstructive pulmonary disease (COPD) [
7].
Experimental studies have attempted to reduce the negative effects of hyperoxia on the developing lungs; however, clinical translation has been disappointing to date. Consequently, there is still no effective treatment for BPD in preterm infants [
8‐
10]. BPD is associated with the inflammation of lungs, which typically involves the recruitment of monocytes that differentiate into alveolar macrophages [
11]. Macrophages are a heterogeneous cell type that can be broadly categorised into two groups: “classically activated” M1 macrophages that have pro-inflammatory functions and M2 macrophages are “alternatively activated” cells that play a reparative or regulatory role [
12‐
14].
Human mesenchymal stem cells (hMSCs) have the ability to alter macrophage phenotype from an inflammatory to anti-inflammatory phenotype that may be therapeutically beneficial to treat injured lungs [
15]. There have been completed or ongoing clinical trials involving cell therapies that delay the progression or reverse a variety of immune and non-immune diseases in the lung and other organs [
16]. Preclinical and clinical data support the use of cell therapy to treat COPD, acute adult lung injury and severe chronic asthma [
17]. Recently, the protective effects of stem cells in lung diseases, for example BPD, have been demonstrated [
18]. Studies have shown that stem cells protect the newborn injured lung from BPD and aid in endogenous repair of the injured lung tissue, either by differentiation into lung parenchymal cells [
19], or into type II alveolar epithelial cells [
20], via the secretion of factors [
21]. However, some limitations to the studies have also been reported such as risks associated with possible immune reactions against hMSCs [
22]. Understanding the therapeutic benefit of hMSCs in neonatal hyperoxic lung injury in mice remaining largely unclear, specifically, the ability to alter the lung immune cells population/s.
This study aimed to determine the effectiveness of human bone marrow-derived hMSC therapy on neonatal hyperoxia, through the modulation of pulmonary immune cells and lung injury–induced collagen deposition in mice, based on the alteration of macrophage phenotype. We showed that hMSCs tracked to the injured lung following intratracheal injection where they modulated macrophage phenotype leading to reduced collagen deposition, thus improving lung structure.
Discussion
The administration of hMSCs to the neonatal lung ameliorated the hyperoxia-induced injuries, including reducing collagen deposition. Additionally, the hMSC administration was found to effectively reduce the hyperoxia-induced infiltration and phenotype of sub-populations of macrophages into the damaged lung.
To study the effects of hMSC therapy in the neonatal lung, a mouse model was used which mimics the effects of neonatal lung hyperoxia in human preterm babies [
1]. This model has provided significant insights into the lung pathology induced by exposing the developing lungs to hyperoxic gas [
30]. This study provides the first evidence that a non-surgical, intra-airway route of administration in mice can effectively deliver hMSCs to the neonatal lung as early as one-hour post-injection where they remained elevated for 24 h. Confirmation of MSCs in damaged lungs has been difficult to ascertain due to the entrapment of hMSCs in lung capillaries when delivered intravenously [
31].
The exposure of neonatal mice to 90% O
2 induced lung injury by postnatal day 14, where there was an accumulation of interstitial collagen which is consistent with a previous report [
32]. This finding was associated with pathological changes to the lungs, namely alveolar wall thickening and pathological changes indicative of emphysema [
30].
The current study used polychromatic flow cytometry analysis to identify and compare granulocytes and macrophage phenotypes in the normoxic lung to the inflammatory lung following hyperoxic injury.
Our hypothesis on the lung response to hyperoxic injury was assessed by quantifying the inflammatory cells, a method we could study by using the flow cytometric assay.
In this study, CD45
+ leukocytes were used to quantify the two subsets of granulocytes and macrophages without reference to the other leukocyte subsets (natural killer cells, invariant natural killer T-cell, T-helper cell, Cytotoxic T-cell, Dendritic cell, monocytes) [
33].
The use of separate markers against Ly6C and Ly6G allowed a defined demarcation of granulocytes from other CD45 myeloid cells populations [
34]. Furthermore, we used the standard approach of staining with both CD11b and CD11c markers to differentiate macrophages from other myeloid cell populations [
33,
34]. The increased proportion of granulocytes indicated an inflammatory response following four days of exposure to hyperoxia, which was consistent with studies showing that granulocytes are the predominant cell type that infiltrates the lung tissue following an injury [
35,
36]. The inflammatory environment may provide important cues leading to the infiltration of other inflammatory cells, including blood monocytes, that have the propensity to differentiate into M1 and M2 macrophages [
37].
We have shown that hMSC attenuates the increase in the total number of CD45
+ leukocyte (
P < 0.05) at day 7 in the neonatal lung following 4 days of exposure to hyperoxia. The elevation of the leukocytes occurred as a result of primary granulocytes recruitment into the alveolar spaces and pulmonary interstitial parenchyma, as defined previously in different lung injury literature [
38].
In the present study, the cell count of Ly6C
+Ly6G
+ granulocytes in lung tissue was elevated at day 4 in the mice exposed to hyperoxia and reduced thereafter. Moreover, the administration of hMSCs was shown to have no effect of granulocyte number in hyperoxic lungs at day 7 and day 14. However, other studies suggest that the granulocytes might raise as a result of releasing granulocyte colony stimulating factor from the stem cells [
39].
This study showed that the CD11b CD11c expressed macrophages were relatively high from day zero, although the level of CD11c expression was low in comparison with day 4 and 7 as shown in a representative FACS plots in Fig.
2. It is hard to confirm whether these macrophages are monocytes derived macrophages or foetal monocytes-macrophages, and immature-macrophages, with significant overlap in expression of marker sets. These interstitial macrophages were derived from the yolk sac [
12] which initially derived from mesoderm cells that develop during gastrulation from the primitive streak [
40].
Throughout the study, we demonstrated that CD11b
+CD11c
+ macrophages increased in response to hyperoxia by day 14, including an increased population of F4/80
lowCD206
low inflammatory ‘M1’ macrophages and decreased F4/80
highCD206
high anti-inflammatory ‘M2’ macrophages. In other studies, M1 macrophage infiltration has been confirmed as well, together with myeloid differentiation and the alteration of relative function at the site of inflammation [
41,
42]. These changes in phenotype result in the activation of monocytes at different maturation stages leading to mature macrophages of distinctive functional states [
43]. Because macrophages are necessary for the phagocytosis of apoptotic neutrophils, the exposure to neonatal hyperoxia may lead to a reduced number of macrophages proceeding to necrosis, leading to the expansion of damaged alveoli [
11]. This conclusion contradicts studies suggesting that macrophages might secrete certain chemokines, which could influence neutrophil infiltration and recruitment [
11].
Our data confirms the beneficial effect of hMSC treatment on the phenotype and function of macrophages after neonatal hyperoxia. The administration of MSCs to mice with hyperoxia-induced acute lung injury was shown alteration of the macrophages proportion and phenotypes. These findings indicated the presence of an inflammatory environment. These results have also been demonstrated in an experimental model of asthma after bone marrow-derived hMSC administration [
44].
Following the intra-airway administration of hMSCs, an increase in M2 macrophages was observed in the injured lungs of mice. It is possible that the reduction in lung injury in hMSC-treated mice was due to downregulation of proinflammatory factors and the upregulation of anti-inflammatory chemokines. In contrast, the advantageous effects of hMSC treatment were displayed by reducing eosinophil infiltration in mouse models with allergic lung injury [
45]. These findings and other supporting studies suggested that hMSCs are able to alter the balance of macrophage phenotype and function that occur during injury to promote repair [
46]. We showed that the protective effect of hMSCs in hyperoxia-induced lung injury could alter macrophage phenotype. Recent studies have shown the in vivo interaction between hMSCs and macrophages to promote M2 polarisation. The
in vitro co-culture of hMSCs and macrophages resulted in an alternatively activated macrophage phenotype characterised as mannose receptor (MR) high, IL-10high, IL-6high, TNF-αlow and IL-12low which exhibits enhanced phagocytic activity, increased secretion of IL-10 and VEGF and decreased secretion of pro-inflammatory cytokines [
44]. Further investigation to identify the polarisation mechanism will be important for the understanding how hMSCs alter the host response following therapeutic delivery, in the context not only the inflammatory response to hyperoxia, but also lung disorders such as allergic asthma and COPD. Apart from the alteration of lung myeloid cells, lung fibrosis is the final common pathway of lung injury regardless of aetiology. In examining the effect of hMSCs therapy on lung injury induced by 90% O
2, we chose an end-point of delivery of oxygen of four days as published data has shown that delivery of hMSCs may ameliorate established lung fibrosis after 14 days of hyperoxia- induced injury [
47]. We found that the treatment with hMSCs induced decreased lung collagen accumulation following hyperoxia. We propose that hMSCs are able to create a more favourable environment, leading to less tissue damage and fibrosis. In this environment, hMSCs may possibly not only alter lung immune cells but may also have an increased capacity to facilitate their antifibrotic and reparative effects, resulting in a greater reduction in fibrosis [
48]. The administration of hMSCs to the neonatal lung ameliorated the hyperoxia-induced structural injuries, including reducing collagen deposition. Additionally, the intratracheal delivery of hMSC was found to effectively reduce the hyperoxia-induced infiltration of myeloid cells, including sub-populations of macrophages, into the damaged lung tissue. This indicates that hMSCs can modulate the inflammatory environment in the lung to reduce the development of fibrotic damage and structural injury.
Using hMSCs in a wide range of settings has shown impressive treatment responses as these cells release anti-inflammatory factors providing a positive effect by modulating the inflammatory environments, hence improving tissue healing [
46]. Our study has shown the effect of the administration of hMSCs immediately, after exposure to hyperoxia, on myeloid cell sub-populations in the lung. Given our findings, it is possible that future clinical trials may incorporate the use of hMSCs therapy to reduce the lung injury including injury that may developed following in preterm birth hyperoxia and other lung diseases that results from pathologic fibrosis. This study was designed to test the efficacy of hMSCs in limiting hyperoxia-induced lung injury. In order to establish a ‘proof of principle’ we used 90% oxygen, which is a higher concentration of oxygen that is used clinically. It is recognised that our hyperoxic lung injury mouse model is likely to induce a more severe form of BPD than in preterm infants in normal physiological condition. In the future, it will be critical to assess the effectiveness of hMSC in lung injury models that more closely replicate clinical conditions. Additionally, the exposure to 90% O
2 is likely to cause hyperoxic injury in organs other than the lungs. As hMSCs administered intratracheally will likely home to injured tissues, it will be important study the systemic effects of hMSC in future studies.