Ratio of miRNA-29 to miRNA-199 expression coordinates mesenchymal stem cell repair of bleomycin-induced pulmonary injury

Abstract

Our previous work demonstrated the anti-fibrotic effects of infusion of adipose-derived mesenchymal stem cells (ASCs) to prevent or repair bleomycin (BLM)-induced lung injury. The present study investigates mechanisms driving these anti-fibrotic effects. Pulmonary fibrosis developed at day 12 in 22-month-old C57BL/6 male mice after intratracheal BLM instillation. There was a decrease in indices of pulmonary fibrosis, including collagen content, AKT activation, collagen types I and III, α_V-integrin, tumor necrosis factor alpha, and transforming growth factor β mRNA after infusion of ASCs 12 days post-BLM treatment compared to BLM alone. Infusion of ASCs increased the population of alveolar types I and II epithelial cells that had been reduced after BLM treatment. miRNAscope technology and reverse-transcription polymerase chain reaction revealed that ASC-treated mice demonstrated increased miR-29a, decreased miR-199, and increased telomere length, telomerase RNA component, and telomerase reverse transcriptase compared to BLM alone. In vitro and ex vivo experiments using double-transfected mouse or human myofibroblasts (miR-29 mimic, and miR-199 inhibitor) confirmed that alterations of these miRNAs regulate downstream effectors of fibrosis. These data suggest that alteration of the ratio of anti-fibrotic to fibrotic miRNAs and increase in telomere length are critical mechanisms of ASC-mediated repair of BLM-induced pulmonary fibrosis.

Graphical abstract

Glassberg and colleagues found that adipose-derived mesenchymal stem cells (ASCs) decreased the extent of lung fibrosis in aged mice in part through the regulation of microRNAs (miR)-29a and -199a. These effects were demonstrated by changes in alveolar epithelial cell expression favoring repair, elongation of telomeres, and reduction of pro-fibrotic pathways.

Introduction

Approximately 60,000 adults aged 60 years and over are diagnosed annually with idiopathic pulmonary fibrosis (IPF), the most fatal and irreversible form,¹ with over 40,000 IPF deaths each year.^2,3 A meta-analysis reported 0.09–1.30 incidence and 0.33–4.51 prevalence per 10,000 individuals worldwide,⁴ Older age, history of smoking, and male sex are associated with both a higher incidence of IPF and a shorter survival time.^2,5,6 The current hypothesis for the pathogenesis of IPF focuses on a primary chronic injury to epithelial cells in genetically susceptible individuals, resulting in a lung with aberrant wound healing and activation of fibroblasts, a depleted population of type I epithelial cells, and a senescent population of type II epithelial cells.^7,8,9 The triggers for this response are poorly understood, but they may include cigarette smoke, as well as other environmental exposures, infection, or aspiration of gastric acid.⁷ Despite recent advances in treatment, no curative therapy is available for IPF. Mesenchymal stem cells (MSCs) represent a potential therapy for IPF, given their immunomodulatory properties and ability to contribute to epithelial repair.^10,11

Pre-clinical studies using the murine model of bleomycin (BLM)-induced pulmonary fibrosis (PF)¹² have shown that MSCs can prevent the development of PF when administered during the initial inflammatory phase.^{13,14,15,16,17} We demonstrated that adipose-derived mesenchymal stem cells (ASCs) from young mouse donors have the ability to prevent BLM-induced PF in aged mice.¹⁷ We and others have also found that ASCs can rescue established BLM-induced PF as quantified by decreased Ashcroft scores and hydroxyproline levels and changes in transforming growth factor β (TGF-β) signaling.^18,19 The greatest limitation of these pre-clinical studies where the BLM treatment is given at a set time point and the course of injury has an established timeline in aged mice is the lack of direct applicability to PF in humans. The onset of PF in most individuals cannot be determined at the time of clinical diagnosis. The burden of disease at that point is already significant based on lung tissue biopsies.^16,20,21,22

Since these studies, we have focused on understanding the potential mechanisms of action of ASCs. We wanted to determine whether ASC infusion at day 12 post-BLM with subsequent modulation of the ratio of microRNA (miRNA; miR)-29/miR-199 expression would impact the fibrotic milieu of BLM-treated lungs and our aforementioned previous findings.^{22,23,24,25,26} We also investigated the mean density of alveolar type I and type II epithelial cell (AEC) expression after ASC infusion, as increasing evidence suggests that AECII cells undergo senescence in IPF, promoting proinflammatory and profibrotic mediators.^{27,28,29,30,31} Finally, we studied effects of ASC infusion on telomere shortening as a recognized finding in individuals with IPF.^{32,33,34,35,36}

Results

BLM-induced PF

BLM-induced lung injury was confirmed prior to ASC administration in the aged mouse model by interval weight change and in vivo lung imaging with micro-computed tomography (μCT) 7 days following BLM administration. Mice treated with intratracheal BLM lost significantly more weight than did the saline controls as has been reported previously (Table 1).³⁷ There was no difference in interval weight loss between mice in the BLM-only vs. BLM+ASC group on day 7 post-BLM (Table 1). However, at 21-day sacrifice, mice treated with BLM+ASC at day 12 weighed more than the BLM-only group (Table 1, p < 0.05).

Table 1.

BLM-induced PF

Weight changes	Interval weight change at day 7 (mean ± SEM)	Overall weight change at day 21 (mean ± SEM)
Saline	−2.2% ± 0.68% (n = 2)	−5.4% ± 1.1% (n = 6)
BLM	−14.2% ± 1.7% (n = 5)	−27.3% ± 3.7% (n = 6)a
BLM+day 12 ASCs	−11.2% ± 2.4% (n = 4)	−15.0% ± 2.6% (n = 6)b,c

^ap < 0.01 vs. saline.

^bp < 0.05 vs. saline.

^cp < 0.05 vs. BLM.

Baseline chest μCT prior to BLM administration demonstrated well-aerated lungs without evidence of pulmonary edema or increased tissue density (Figure 1A, baseline). By day 7 post-BLM, chest μCT demonstrated increased attenuation over the lung fields and loss of aeration, indicating lung injury in mice given intratracheal BLM. Control mice treated with intratracheal saline showed no significant changes or evidence of lung injury compared to their baseline chest μCT (Figure 1B).

Figure 1.

Bleomycin-induced lung injury, which is reduced after adipose-derived mesenchymal stem cell instillation

(A) Representative μCT transverse and coronal lung sections acquired from aged (22-month-old) male C57BL/6 mice at baseline (left) and 7 days following intratracheal bleomycin (BLM, 2.0 U/kg) administration (right) demonstrating increased lung density and loss of airspaces. (B) Saline treatment did not result in evidence of lung injury on μCT scan at baseline (left) or 7 days post-instillation (right). Histological sections of lung tissue collected at day 21 post-BLM were stained with Masson’s trichrome as described in materials and methods. (C and D) Representative photomicrographs (20× and 40× magnifications) of lung sections from saline-treated control mice (C) and BLM-treated mice (D). (E) Infusion of adipose-derived mesenchymal stem cells (ASCs) 12 days post-BLM instillation resulted in reduced severity of pulmonary fibrosis (PF). (F) Degree of PF on histological sections was measured by semi-quantitative Ashcroft score as described in materials and methods. BLM-induced lung injury resulted in increased Ashcroft score compared to saline controls. Infusion with ASCs 12 days post-BLM injury resulted in decreased Ashcroft score. (G) Intratracheal BLM instillation increased lung collagen content as measured by hydroxyproline assays as described in materials and methods. Mice treated with ASCs on day 12 post-BLM had decreased lung collagen content compared to BLM-only controls. Each data point represents an individual biological replicate (mouse); n = 6–10 mice/group. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. (H) Ratio of pAKT to AKT protein expression in lung tissue of mice was quantified by western blot analysis at day 21 post-BLM sacrifice. Aged C57BL/6 mice treated with intratracheal BLM demonstrated increased pAKT/AKT protein expression compared to saline-treated controls. Lungs from mice treated with intravenous infusion of ASCs 12 days post-BLM-induced injury demonstrated decreased expression of pAKT/AKT compared to BLM-only group. Inset shows a representative western blot and β-actin loading control. Data are graphed as individual biological replicates (n = 6–8 mice/group); ∗p < 0.05.

Day 12 ASC infusion post-BLM injury decreases lung fibrosis in aged mice

At day 21 post-BLM, lungs from BLM-treated mice exhibited interstitial fibrosis with increased collagen deposition, alveolar wall thickening, and distortion of alveolar architecture (Figure 1D). Saline controls did not demonstrate evidence of PF (Figure 1C). Compared to BLM-only mice, those mice treated with ASCs on day 12 post-BLM had lower Ashcroft scores on histological sections, indicating less severe fibrosis (Figures 1E and 1F). Lung collagen content (hydroxyproline), another indirect quantification of PF, was increased in the BLM-only group compared to saline controls (Figure 1G). Treatment with ASCs 12 days post-BLM injury resulted in a significant decrease in lung collagen content compared to the BLM-only group (Figure 1G).

Day 12 ASC infusion decreases mRNA expression of established molecular markers of fibrosis and inflammation at day 21 post-BLM sacrifice

Day 12 infusion of ASCs post-BLM injury resulted in a significant decrease in markers associated with BLM-induced pulmonary injury. TNF (tumor necrosis factor α), a marker of inflammation, was increased in BLM-treated mice compared to saline controls (Table 2; p < 0.05). Expression of ITGAV (α_v-integrin) mRNA, a transmembrane cell adhesion molecule that modulates tissue fibrosis,³⁸ and COL1A1 (type I collagen) were also increased in BLM-treated mice compared to saline controls (Table 2; p < 0.05). Day 12 infusion of ASCs resulted in decreased mRNA expression of α_v-integrin and collagen and TNF-α (Table 2; p < 0.05). The TGF-β signaling pathway was also dampened by treatment with ASCs (Table 2).

Table 2.

Effect of day 12 ASC infusion on markers of fibrosis and inflammation after BLM-induced lung injury

mRNA/18S (% of BLM, n = 5–12/group)	αv-Integrin	Collagen type 1α1	TNF-α
Saline	45 ± 8.9a	31 ± 14a	37 ± 13.6b
BLM	100 ± 9.3	116 ± 35	100 ± 15.93
BLM+day 12 ASCs	43 ± 19a	33 ± 15.2a	34 ± 21.3a
mRNA/18S (% of BLM, n = 7-10/group)	TGF-β	SMAD2	SMAD3
Saline	13 ± 2.7a	38 ± 14.5a	44 ± 18.45a
BLM	100 ± 11.9	101 ± 9.5	109 ± 10.7
BLM+day 12 ASCs	57 ± 11.6a	41 ± 15.7a	56 ± 16.3a

ASCs, adipose-derived mesenchymal stem cells; BLM, bleomycin.

^ap < 0.01.

^bp < 0.001 vs. BLM.

Day 12 ASC infusion decreases BLM-induced lung AKT activation

At day 21 post-BLM, mice that were treated with intratracheal BLM had a significantly higher phosphorylated AKT (pAKT)/AKT protein expression ratio compared to saline controls (Figure 1H, p < 0.05). Day 12 infusion of ASCs post-BLM resulted in decreased activation of AKT (pAKT:AKT ratio) compared to BLM-only mice (Figure 1H, p < 0.05).

Immunohistochemistry of mouse lungs

To determine whether infusion of ASCs day 12 post-BLM treatment increased the population of AECI and/or AECII cells as reflected in an increase in aquaporin 5 (AQP5) staining (red color) or surfactant protein C (SPC) staining (purple color, Figure 2), we performed immunofluorescent analysis of lung tissue. The mean density of positive AQP5 and SPC expression, which was decreased after BLM injury, increased after ASC infusion (Figures 2A and 2B).

Figure 2.

BLM-induced decreased AQP5 and SPC staining

Day 12 ASC infusion increased AQP5 and SPC fluorescent staining, representing increased AECI and AECII expression, respectively. Data are graphed as (A) integrated density of AQP5 or (B) SPC staining/integrated density of DAPI and expressed as percentage of BLM (n = 3–9 slides/group). ∗p < 0.05, ∗∗∗∗p< 0.0001 Each point represents a single mouse. Representative images (20×); insets are delineated on original slides with white box; scale bar represents 250 μm. Saline-treated mouse lung, BLM, BLM+day 12 ASC infusion; DAPI, blue; αSMA, green; SPC, purple; AQP5, red.

Day 12 ASC infusion increases miR-29a and decreases miR-199 in vivo at day 21 post-BLM sacrifice

miR-29a, a well-characterized anti-fibrotic mediator in several diseases including PF,^39,40 was significantly decreased in the lungs of BLM-treated mice compared to saline control lungs (Figure 3; p < 0.01). Mice treated with ASCs 12 days post-BLM injury demonstrated increased expression of miR-29a compared to BLM-only mice (Figure 3A; p < 0.05). In contrast, miR-199 expression, reported to be upregulated in lung, kidney, and liver fibrosis,⁴¹ was increased in BLM-injured lungs. ASC treatment downregulated miR-199 expression (Figure 3B; p < 0.05). To better visualize and semi-quantitate the expression change in miR-29 and -199, we performed fluorescence in situ hybridization (FISH) on lung tissue isolated from the lungs of BLM-treated, ASC-treated, and saline-treated mice in the study. Mean density of FISH for miRNAs was divided by density of DAPI staining. A ratio was calculated for each mouse lung section. There was an increase in miR-29 and a decrease in miR-199 in lung sections after mice were treated with BLM and ASCs (Figures 3C and 3D). As previously shown,¹⁷ activity of matrix metalloproteinase-2 (MMP-2), a downstream target of miR-29, was increased in the lungs of mice that had received BLM only (Figure 3E, p < 0.05). Infusion of ASCs 12 days post-BLM injury resulted in a decrease in MMP-2 activity compared to the BLM-only group (Figure 3E, p < 0.05). Caveolin-1 (CAV-1), a molecule known to be regulated by miR-199, is an important structural molecule that is downregulated in fibrotic lung disease.^22,42 Western blot analysis showed that ASC treatment at day 12 resulted in an increase in CAV-1 (Figure 3F; p < 0.05).

Figure 3.

Lung miR-29a-5p expression is increased and miR-199a-3p expression is decreased after day 12 infusion of ASCs

Aged C57BL/6 mouse lung expression of miR-29a is decreased at day 21 post-BLM sacrifice in response to BLM lung injury compared to saline controls. Infusion with ASCs on day 12 post-BLM-induced lung injury resulted in higher expression of miR-29a vs. BLM-only controls. (A and B) miR-29a (A) and miR-199a (B) expressions were measured by reverse-transcriptase polymerase chain reaction from total lung RNA, as described in materials and methods. U6 expression was used as a control. Data are graphed as data from individual biological replicates (n = 6–8/group). ∗p < 0.05; ∗∗p < 0.01. (C) The ratio of miR-29:miR-199 is increased following day 12 ASC infusion to BLM-treated mice. Data are graphed as a ratio of mean integrated density of miRNA staining/DAPI staining of individual biological replicates (n = 3–6/group). ∗p < 0.05; ∗∗p < 0.01. (D) Representative images of miRNAscope staining for the expression of miR-29 and miR-199 in mouse lung tissue (20×). Yellow square denotes where smaller images are enlarged (red, miRNA; cyan, SPC), scale bar represents 1,000 μm. (E) BLM lung injury results in increased MMP-2 activity in lungs of aged C56BL/6 mice measured at day 21 post-BLM sacrifice. Intravenous infusion of ASCs on day 12 post-BLM injury resulted in decreased MMP-2 activity compared to BLM-only control mice. Zymography was performed on protein extracts from lung tissue from saline, BLM, or BLM+ASC-treated mice exposure to measure MMP-2 activity. Insert is representative zymogram of two mice per group. Data are graphed as individual biological replicates (n = 4–6/group); ∗p < 0.05. (F) Cav-1 protein expression was determined by western blot analysis as described in materials and methods. Inset is representative western blot of two mice per group and β-actin loading control. Data are graphed as individual biological replicates of n = 3–6/group; ∗p < 0.05, ∗∗p < 0.01.

In vitro double transfection with miR-29a mimic and miR-199a inhibitor directly regulates MMP-2 and CAV-1

To confirm that ASC-induced changes in MMP-2 and CAV-1 expression were direct effects of miRNA changes, we transfected myofibroblasts isolated from human lung tissue of individuals with IPF with a mimic of miR-29 (knockin) and an inhibitor of miR-199 (knockout) (29KI/199KO). Upregulation of miR-29 and downregulation of lung miR-199 expression, confirmed by RT-PCR (Table 3), correlated with decreased MMP-2 activity (Figure 4A, arrow) and increased CAV-1 expression (Figure 4B).

Table 3.

miRNA and mRNA expression

Myofibroblast miRNA expression percent increase over scrambled control transfected cells
miR-29a expression (percent increase)	miR-199a expression (percent decrease)
350	1,083
546	464
151	834

3D ex vivo explant expression mRNA expression/18S	αv-integrin (n = 3)	collagen type 1 (n = 3)	collagen type III (n = 3)	TGF-β (n = 2)
(% of PBS control) Scrambled control myofibroblasts	477.9 ± 92.1	866 ± 517	467 ± 184	201 ± 7.4
(% of scrambled control) 29KI/199KO myofibroblasts	61.8 ± 2	38.8 ± 3.12	45.9 ± 6.2	81.9 ± 10.6

Figure 4.

Transfection of miR-29 mimic and miR-199 inhibitor regulates downstream targets MMP-2 and CAV-1

In vitro double transfection of miR-29 mimic and miR-199 inhibitor (29KI/199KO) to myofibroblasts isolated from lungs of individuals with IPF recapitulates the effects of in vivo ASC infusion. Human myofibroblasts were transfected with scrambled controls (“C”) or miR-29 and -199 (29/199). (A) Representative zymogram for MMP-2 activity (as indicated by arrow). (B) Western blot for CAV-1 expression and β-actin loading control. (n = 2 unique cell lines). Lung 3D ex vivo explants from aging male C57BL/6 mice were prepared as described in materials and methods. Human 29KI/199KO cells or control scrambled plasmids were injected into the explant at a concentration of 10⁵ cells/explant. Explants were assessed for MMP-2 activity (C) and CAV-1 protein expression (D). Insets are representative zymography and western blot, respectively. Data are graphed as biological replicates n = 3/group; ∗p < 0.05. Experimental replicates of n = 3/2 unique patient myofibroblast cell lines. (E) Ashcroft scoring materials and methods was performed on n = 4 individual explants/group, ∗∗∗p < 0.002; ∗∗∗∗p < 0.0001. (F) Representative images of trichrome staining of lung 3D ex vivo explants (4× and 20×; scale bars, 200 and 50 μm, respectively). (G) Immunofluorescent staining of lung 3D ex vivo explants with α-SMA (green), AQP5 (purple), SPC (red), and DAPI (blue).

Three-dimensional (3D) ex vivo mouse lung explants (3D lung tissue cultures): to further demonstrate the relevance of decreased miR-29 and increased miR-199 in myofibroblasts isolated from IPF lungs, we utilized 3D ex vivo aging mouse lung explants.^43,44 The histology was unchanged by the injection of control transfected myofibroblasts or phosphate-buffered saline (PBS) as reviewed by our pulmonary pathologist. Ashcroft analyses (n = 4/group; Figure 4E) of explants revealed that injection with myofibroblast control transfected cells induced fibrosis (Ashcroft score = 3.1 ± 0.08) compared to 29KI/199KO cells (Ashcroft score = 2.3 ± 0.09; Figure 4D, p < 0.001). The 29KI/199KO cells did not provoke a fibrotic response compared to those with PBS control (Ashcroft score = 2.3 ± 0.07). In those explants receiving myofibroblast control cells, there was patchy thickening of interalveolar space, increased multinucleated AECII cells, and some alveolar macrophages. This was in contrast to those explants injected with 29KI/199KO cells, which exhibited no to few inflammatory cells, occasional macrophages, and increased AECI and AECII cells with preserved structure.

Increased mRNA expression of selected fibrotic markers: α_v-integrin, TGF-β, and collagen types I and III were evident in the explant injected with control transfected myofibroblasts (2- to 7-fold). In contrast, explants injected with 29KI/199KO cells did not increase α_v-integrin, TGF-β, and collagen types I and III mRNA expression (Table 3).

Lung explant expression of MMP-2 activity increased (Figure 4C), while CAV-1 decreased (Figure 4D) after treatment with control-transfected myofibroblasts. Those explants injected with 29KI/199KO cells showed increased CAV-1 and decreased MMP-2 activity compared to control transfected cells (Figures 4C and 4D). Ashcroft score (Figure 4E) and histology (Figure 4F) of explants showed normal architecture after treatment with 29KI/199KO cells as compared to those explants infused with control transfected myofibroblasts. Immunofluorescent staining of explants revealed decreased AQP5⁺ cells (AECI) and decreased SPC⁺ cells (AECII) after treatment with control transfected myofibroblasts compared to treatment with 29KI/199KO cells (Figure 4G).

ASCs repair the BLM-induced telomere shortening and decreased TERT/TERC expression

The administration of BLM to aging mice reduced lung telomere length per chromosome (106 ± 10.66 kb) as compared to saline-treated control mice (153 ± 11.69 kb, p < 0.05). The telomere length of lung DNA isolated from mice receiving BLM+ASC treatment (146 ± 11 kb) was comparable to that of the saline control mice and greater than mice treated with BLM alone (Table 4, p < 0.05). Telomerase reverse transcriptase (TERT) and telomerase RNA component (TERC) mRNA expression increased in the lungs of mice after BLM+ASC treatment compared to BLM alone (Table 4).

Table 4.

Telomere length and mRNA of TERT and TERC expression

	Telomere length/chromosome, kb	TERT	TERC
Saline (n = 8)	152 ± 11.69a	NA	NA
BLM (n = 5)	105 ± 10.66	0.075 ± 0.12	0.022 ± 0.012
BLM+day 12 ASC (n = 8)	146 ± 11.09	0.95 ± 0.42	0.023 ± 0.06a

NA, not assessed.

^ap < 0.05 compared to BLM

Discussion

Our previous studies showed that ASC infusion 24 h post-BLM injury prevents development of PF in aged mice¹⁷. Additionally, a single infusion of ASCs in the fibrotic phase of BLM-induced pulmonary injury in mice can alter established fibrosis.¹⁸ In this study, we demonstrate that miRNAs play a critical role in the anti-fibrotic effects of ASCs. We confirmed that changes in lung imaging (μCT) correlated with histological changes following BLM administration, although μCT cannot differentiate between inflammation and fibrosis at this early stage.^45,46 Due to higher anesthesia risk in older mice, we relied on interval weight loss from the time of μCT on days 7 and 12 to confirm BLM injury prior to ASC treatment.

Several studies have shown that MSCs, injected after the first week of BLM injury, reduce fibrosis.^{14,15,19,47,48} Huleihel et al. studied day 7 post-BLM infusion of MSCs modified to overexpress miR let-7d, an miRNA decreased in the lung tissue of individuals with IPF.⁴⁹ It is likely that at this early time point post-BLM, MSCs could elicit anti-inflammatory effects.⁵⁰ A recent publication by Franzen et al. selected time points post-BLM administration to address the acute inflammatory and early fibrogenic phase (day 7), as well as the established fibrotic stage (day 21) in young mice.⁵¹ We chose day 12 post-BLM administration since it is a well-established time point for fibrosis in old mice.⁵²

Moodley et al. found amnion-derived MSCs reduced fibrosis when administered 10 days post-BLM instillation, while human bone marrow-derived MSCs (BM-MSCs) prevented fibrosis.⁴⁸ Garcia et al. reported similar results using amniotic fluid-derived MSCs injected at day 14 post-BLM in young mice.¹⁴ Zakaria et al. found both an improvement of lung architecture and pulmonary function with a single BM-MSC infusion 28 days post-BLM injury,⁵³ and a study conducted by Chu et al. demonstrated that intratracheal administration of Wharton’s jelly-derived MSCs at 21 days post-BLM injury caused a reduction in fibroblast activation and extracellular matrix (ECM) accumulation in the lungs.⁵⁴ Lung tissue obtained at day 14 post-BLM showed a slight decrease in collagen type I expression, but no change in either Ashcroft score or hydroxyproline levels.⁴⁹ Lee et al. used a long-term, repeated low-dose BLM injury model to show that biweekly administration of human-derived ASCs starting 8 weeks after the first of eight BLM doses resulted in attenuation of PF in young mice.¹⁵ A more recent study administered MSCs day 14 post-BLM to 8-week-old male or female mice, with an improvement in Ashcroft score and reversal of prominent fibrosis markers.¹⁹ Although the results are encouraging, the observation that young C57BL/6 mice have been shown to experience spontaneous resolution of BLM-induced PF after 4 months challenges the interpretation of these studies and applicability to older mice.⁵⁵

Our study confirmed that ASC infusion on day 12 post-BLM injury reduced activation of the AKT and TGF-β signaling pathways, consistent with earlier findings.^17,18 The dysregulation of ECM turnover linked to activation of the AKT and TGF-β signaling pathways, is a key driver of fibrosis in IPF^24,56 and BLM models.^17,57

Rana et al. showed that TGF-β1 induced AECII senescence, resulting in the stimulation of profibrotic and proinflammatory mediators of lung fibrosis.³⁰ Senescent AECII may contribute to failure of alveolar repair after injury since mechanisms necessary for restoring the damaged alveolar epithelial layer depend on the proliferation and differentiation of AECII into AECI.⁵⁸ We found that lung tissue from ASC-treated mice had increased AECII and in parallel increased AECI density. These data suggest that ASCs may protect against AECII cell senescence and apoptosis^59,60 and stimulate differentiation of AECII cells.⁶¹ ASC treatment induced an increase in AECII density in lung tissue to the level found in the lungs of saline-treated mice, although this was not the case for AECI density. It is likely that the complex process of repair induced by ASCs on the epithelium may be ongoing, and the regeneration timeline is longer than the experimental timeline. Further studies on these mechanisms of ASC-induced repair are ongoing.

Cellular senescence results in the stimulation of profibrotic and proinflammatory mediators of lung fibrosis,³⁰ leading to the senescence-associated secretory phenotype (SASP).⁶² The SASP, regulated by transcription factors and non-coding RNAs like miRNAs,^63,64 contributes to fibrosis by promoting fibroblast proliferation and ECM deposition. We therefore determined how day 12 ASC treatment affected the regulation of miR-29 and -199, two miRNAs found to be dysregulated in the lungs of individuals with IPF²⁶ and regulated in our previous studies by ASC infusion.^22,65 Using PCR and miRNAscope technology,⁶⁶ we detected a ratio change in the expression of miR-29 to miR-199 (increased miR-29, decreased miR-199) in the lung tissue of mice infused with ASCs on day 12 post-BLM injury. This correlated with the expected regulation of lung expression of known miR-29 targets: MMP-2, collagen type I α1 (Col1α1), TGF-β/SMAD2, and SMAD3 mRNA in ASC-treated mice.^39,67 AKT activation was also decreased in lung tissue from ASC-infused mice, another expected result given that increased miR-29 has previously been implicated in dampening TGF-β-induced AKT activation.⁶⁸ Montgomery et al. showed that administration of a pharmacological miR-29 mimic attenuated BLM-induced PF in C57BL/6 mice, even when administered 10–17 days post-BLM.²⁵ As mentioned above, Cav-1, highly expressed in AECI^69,70 and important for the regulation of AEC apoptosis⁷¹ and fibroblast activation,⁷² is decreased in the lungs of individuals with IPF.⁷³ In our study, inhibition of miR-199 expression upregulated Cav-1 protein. Of note, Cav-1-scaffolding-protein-derived peptide isbeing evaluated in clinical trials as a potential anti-fibrotic therapeutic agent in individuals with IPF (this study was registered at ClinicalTrials.gov: NCT04233814).⁷⁴ These findings strongly suggest that the effects of MSCs in PF may be carried out in part via gene expression regulation by miR-29 and miR-199 in conjunction with other factors induced by the administration of ASCs.^75,76

We performed further experiments with double-transfected lung myofibroblasts isolated from individuals with IPF. Control transfected IPF myofibroblasts injected into 3D ex vivo lung explants conferred a pro-fibrotic phenotype in the explant. Administration of double-transfected myofibroblasts partially prevented the development of the profibrotic phenotype induced by control transfected myofibroblasts. These results mirror our in vivo findings and suggest a causal relationship between miRNA modulation, ASC infusion, and the subsequent rescue of the lung.

Finally, we explored the impact of ASCs on telomere length, as telomere shortening is associated with a worse prognosis when found in individuals with familial and sporadic IPF.^32,77 ASC infusion restored telomere length and upregulated telomerase components (TERT and TERC)⁷⁸ in BLM-injured lungs, suggesting a protective effect. Telomere-related gene mutations have been found in approximately 30% of individuals with familial IPF and 1%–3% of those with sporadic disease.⁷⁹ In addition, individuals with IPF without mutations have shorter telomeres when compared to age-matched healthy controls.^33,80 Furthermore, induction of telomerase deficiency or telomere dysfunction in murine models is associated with a propensity for the subsequent development of PF in synergy with a low dose of BLM,⁸¹ while treatment of wild-type and telomerase-deficient mice with telomerase gene therapy mitigates the onset of profibrotic pulmonary pathologies.⁸² BLM reduces telomerase activity and, accordingly, telomere length in epithelial tissues.^83,84 Multiple studies suggest that telomere pathology is a risk factor for AECII senescence and fibrosis,²⁹ while other investigations suggest that the interplay between telomere shortening and TGF-β/Smad signaling may be important in the development of fibrosis.^85,86 Our data suggest that ASCs may induce the restoration of telomere length by virtue of upregulated telomerase function.

Although others have shown no direct effect of BLM treatment on lung telomere length in young mice,^87,88 these contradictory findings may be due to the fact that aged mice, such as those used for our study, have greater susceptibility to tissue DNA damage and telomere shortening at baseline compared to young mice.^89,90 We recognize that rodents are known to have longer telomeres at baseline, so further results on telomere biology in human lung fibrosis would be best conducted with human fibrotic lung tissue samples.³²

In summary, our study suggests that ASC infusion during the fibrotic phase of BLM-induced PF in aged mice can rescue the ongoing fibrotic process in part through dampening of TGF-β signaling and modulation of miRNA expression. ASCs may act as a “factory” of reparative miRNAs targeting multiple fibrotic pathways and counteracting the SASP in the fibrotic lung. Future studies are ongoing to investigate the paracrine effects of ASCs and the use of MSC products including exosomes.

Materials and methods

Animals

Male C57BL/6 mice were obtained from The Jackson Laboratories (Bar Harbor, ME). We used 22-month-old male mice for all experiments (n = 6–8/group). Four-month-old male C57BL/6 were used for the isolation of ASCs. Animals were housed under pathogen-free conditions with food and water ad libitum. All experiments and procedures were approved by the Institutional Animal Care and Use Committee at the University of Miami (protocol no. 16-041), a facility accredited by the American Association for the Accreditation of Laboratory Animal Care.

BLM-induced lung injury

After induction of anesthesia with ketamine, BLM sulfate (Sigma-Aldrich, St. Louis, MO) dissolved in 50 μL sterile saline was administered by direct intratracheal instillation (2.0 U/kg). Control mice received 50 μL intratracheal sterile saline. Mice were weighed at baseline, day 7 post-BLM, and at sacrifice. Mice were sacrificed 21 days following BLM or saline administration.

ASC isolation from young male mice or adult male human adipose tissue

Donor ASCs were isolated from the subcutaneous adipose pads of 4-month-old male C57BL/6 mice, as previously described.¹⁷ Mice were anesthetized with ketamine (200 mg/kg) and xylazine (10 mg/kg) injected intraperitoneally. Subcutaneous adipose tissue was excised, washed in PBS without Ca²⁺ and Mg²⁺ containing 30% GIBCO penicillin-streptomycin (Life Technologies, Grand Island, NY), and digested in media containing 0.75% type II collagenase (Sigma-Aldrich). The suspension was centrifuged to separate floating adipocytes from the stem vascular fraction. The resultant pellet was resuspended and cultured in ADSC Growth Medium (Lonza Group, Basel, Switzerland). Cells were expanded in plastic Nunc Cell Culture Treated Flasks with Filter Caps (Thermo Fisher Scientific, Waltham, MA). After a 24-h incubation period, non-adherent cells were removed. When the adherent cells became confluent, they were trypsinized, expanded for two to three passages, and cryopreserved in Recovery Cell Culture Freezing Medium (Life Technologies). Characterization of ASCs was performed as previously described.¹⁷ Briefly, ASCs were incubated with fluorescence-labeled antibodies and analyzed by fluorescence-activated cell sorting Canto II (BD Biosciences, San Jose, CA). For mesenchymal differentiation potential, the Mouse Mesenchymal Stem Cell Functional Identification Kit (R&D Systems, Minneapolis, MN) was used according to the manufacturer’s instructions. Multipotency was assessed via osteogenic and adipogenic differentiation.¹³ Human ASCs were isolated and characterized as previously described.⁹¹

Lung μCT

Mice underwent thoracic imaging by μCT (SkyScan microCT, Bruker, Kontich, Belgium) at baseline and 7 days following BLM or saline administration. Scan parameters used were according to a previously validated protocol.⁹² Mice were lightly anesthetized by intraperitoneal ketamine injection. Respiratory-gated μCT images were acquired with the following image parameters: 50 kVp X-ray source, 500 μA current, and 193-ms exposure time per projection, with 0.7° increments, and 0.5-mm aluminum filter. Total scan time was approximately 9 min per mouse. The images obtained were reconstructed using manufacturer’s software (SkyScan NRecon, Bruker) with the following settings: image smoothing 5, beam-hardening correction 31%, ring artifact reduction 6, and histogram dynamic range 0–0.03 attenuation values. Since aging mice are frail, they were unable to be anesthetized and scanned after day 7 post-BLM.

Day 12 ASC infusion

Young adult donor-derived ASCs (passage two or three) were thawed in a 37°C water bath and washed two times in PBS to remove the cell culture freezing solution prior to injection. ASCs were then passed through a 70-μm cell strainer to remove cell clumps. Cells were counted and resuspended in PBS immediately prior to injection. At 12 days post-BLM injury, mice were administered 5 × 10⁵ human or mouse ASCs in 200 μL PBS by tail-vein injection over 1 min. Control mice received 200 μL PBS by tail-vein injection over 1 min.

Histological analysis and Ashcroft scoring

Right lung lobes were inflated with 10% neutral-buffered formalin (NBF) under 25 cm H₂O constant pressure. Lung tissue and 3D ex vivo explants were embedded in paraffin, and 4-μm sections were taken for H&E and Masson’s trichrome staining. PF was assessed by a pathologist⁹³ blinded to the experimental groups using the numerical Ashcroft scale⁹⁴ on Masson’s trichrome-stained slides at 20× magnification. Individual fields were assessed by systematically moving over a 32-square grid; each field was assessed for severity of fibrosis and assigned a score of 0 (normal lung) to 8 (total fibrosis of the field). Mean ± SEM values are reported.

Collagen content as measured by hydroxyproline assay

Left lung lobes were harvested for tissue analyses. Lung hydroxyproline assay was performed according to the manufacturer’s instructions (Hydroxyproline Assay Kit, Sigma-Aldrich). Briefly, 2 mg lung fragments were weighed and homogenized in 100 μL distilled water. An equal volume of 10 M HCl was added to the samples before drying at 49°C for 3 h. We loaded 50 μL of sample onto the plate and incubated it overnight at 37°C. A hydroxyproline standard curve was prepared according to a standard solution (between 0 and 1 μg/well). Absorbance was measured at 557 nm using SoftMax Pro software (Molecular Devices, Sunnyvale, CA). Lung collagen content per milligram of tissue was calculated from the hydroxyproline measurement by dividing by a factor of 13.5%, as previously described.⁹⁵

Immunofluorescence

We cut 4-μm-thick sections from paraffin-embedded tissue blocks. Dewaxed sections on glass slides were subjected to epitope retrieval by standard methods using citrate buffer (pH 6) for 30 min at 110°C. Cooled slides were rinsed quickly with deionized water and incubated singly with either of the following antibodies (diluted in 10% donkey serum, 1% BSA in Tris-buffered saline with 0.1% Tween 20 detergent [TBS-T] buffer) overnight at 4°C: PECAM1/CD31 (goat immunoglobulin G [IgG], R&D Systems, catalog no. AF3628), 1:250; surfactant protein C (SFTPC) (rabbit IgG, Millipore [Burlington, MA], catalog no. AB3786), 1:500; AQP5 (rabbit IgG, Abcam [Cambridge, UK]), 1:200; and ACTA2 (mouse IgG-Alexa Fluor 488 conjugate, Invitrogen [Waltham, MA], catalog no. 53-9760-82); 1:50. After triple rinsing with TBS-T, slides were treated with 0.5% H₂O₂/PBS for 15 min at room temperature to inactivate endogenous peroxidase.

Slides were then incubated with a donkey anti-goat IgG-horseradish peroxidase (HRP) polymer (R&D Systems, catalog no. VC004-050) for 30 min at room temperature, followed by the HRP substrate VIVID570 (for PECAM1/CD31; Bio-Techne [Minneapolis, MN], catalog no. 7526) diluted 1:1,500 in tyramide signal amplification (TSA) buffer for 15 min at room temperature. The HRP was inactivated by incubating the slides in 0.2% NaN₃/0.3% H₂O₂/PBS for 15 min at room temperature. Slides were incubated with goat anti-rabbit IgG-HRP polymer (Biocare Medical [Pacheco, CA], catalog no. M2U522L) for 30 min at room temperature followed by the HRP substrate VIVID650 (for SFTPC; Bio-Techne, catalog no. 7527) diluted 1:1,500 in TSA buffer for 15 min at room temperature. For AQP5, slides were exposed to MACH2 goat anti-rabbit HRP polymer conjugate (Biocare Medical, catalog no. RALP525L). Slides were mounted with coverslips using Vectashield Vibrance mounting medium with DAPI (Vector Laboratories, Newark, CA, catalog no. H-1800). The mean integrated density of signal was obtained using Fiji for each color. Data were normalized to DAPI mean integrated density.

Isolation of myofibroblasts from human lungs

After receiving signed informed consent forms, we obtained lung samples at the time of lung biopsy at the University of Miami Affiliated Hospitals from individuals with IPF. This study was approved by the institutional review board at the University of Miami Leonard M. Miller School of Medicine (IRB no. 20060249) and was conducted in compliance with Health Insurance Portability and Accountability Act regulations. Human lung was cut into small pieces and plated in separate 6-well plates (Nunc, Thermo Fisher Scientific) for 30 min prior to adding media. Cells were allowed to grow and were transferred to separate T25 flasks when confluent. A portion of the cells was placed on a chamber slide, and myofibroblasts were identified by positive staining for α-smooth muscle actin (α-SMA) (Abcam) and vimentin (Abcam). Cells were used for experiments between two and four passages.

Western blot analysis

AKT, pAKT, and CAV-1 protein expression in lung tissue and myofibroblasts were measured by western blot, as previously described.⁹⁶ Lung tissue and myofibroblasts were homogenized, and lysates were collected for western blot analyses, as previously described.⁹⁶ For AKT and pAKT, 10 and 25 μg protein lysate, respectively, were loaded onto 10% polyacrylamide gels. Goat anti-AKT (1:1,000) and rabbit anti-pAKT (1:1,000) were used to detect protein expression (Santa Cruz Biotechnology, Dallas, TX). For CAV-1, 15 μg protein was loaded. Immunoreactive bands were determined by exposing nitrocellulose blots to a chemiluminescence solution (Denville Scientific, Metuchen, NJ), followed by exposure to Amersham Hyperfilm ECL (GE Healthcare, Chalfont St Giles, UK). ImageJ version 1.48 (National Institutes of Health, Bethesda, MD) was used to determine the relative density of bands. β-Actin expression was determined using mouse anti-β-actin (1:10,000). All values were corrected for corresponding β-actin bands.

Isolation of RNA and real-time PCR

Total RNA was extracted from 5 mg lung tissue or 100,000 myofibroblast cell homogenates. Amplification and measurement of target RNA was performed on the Step 1 real-time PCR system, as previously described.⁹⁷ α_v-Integrin, collagen type I α1 (Col1α1), collagen type III α1 (Col3α1), TNFα, TGF-β, SMAD2, SMAD3, SMAD7 TERT and TERC mRNA expression were measured. TaqMan probes and primers for amplification of the specific transcripts were designed using Primer Express 1.5 from Applied Biosystems (Foster City, CA) (catalog numbers are supplied in the supplemental information). TaqMan rRNA control reagents (Life Technologies) designed to detect 18S rRNA were used as an endogenous control to normalize for variations in the isolated RNA amount. For miR-29a and -199a-3p analyses, cDNA was generated using the microDNA cDNA Synthesis Kit (TaqMan Advanced miRNA cDNA kit, ThermoFisher) according to the manufacturer’s instructions. Amplification of miR-29a and -199-3p was performed using specific primers (ThermoFisher probes, see supplemental information) and Applied Biosystems TaqMan™ Fast Advanced Master Mix for qPCR (ThermoFisher) U6 expression was used as a control for miRNA analyses, and relative expression was calculated using the comparative C(T) method.⁹⁸

Lung miR-29a and -199a expression by FISH (miRNAscope)

Paraffin lung tissue blocks were sectioned at 5-μm thickness. Sections were placed on silanized (3-aminopropyltriethoxysilane) glass slides and incubated in 10% NBF overnight. Slides were then washed twice in 100% ethanol, treated with 3% H₂O₂ in ethanol for 10 min at room temperature, and incubated at 110°C for 15 min in Co-Detection Target Retrieval Reagent (catalog no. 323165, Advanced Cell Diagnostics [ACD], Newark, CA). Following cooling to room temperature, slides were rinsed twice in 100% ethanol, dried for 5 min at 60°C, and marked with an ImmEdge Hydrophobic Barrier pen (catalog no. H-4000, Vector Laboratories). Slides were treated with RNAscope Protease Plus (catalog no. 322381, ACD) for 30 min at 40°C and rinsed twice in distilled water.

Subsequently, slides were processed for in situ hybridization with the miRNAscope HD (Red) Assay kit (catalog no. 324500, ACD) for murine miR-29a (catalog no. 895241-S1, ACD) and miR-199a (catalog no. 1226031-S1, ACD). After detection of miRNAscope probes using FAST Red chromogen, slides were blocked in co-detection blocker (catalog no. 323180, ACD) and incubated overnight with primary antibodies against SFTPC (catalog no. AB3786, Millipore; dilution 1:500) and ACTA2 (Alexa Fluor 488 conjugate; catalog no. 53-9760-82; Invitrogen; dilution 1:100) in Co-Detection Antibody Diluent (catalog no. 323180, ACD). The following day, the slides were rinsed and incubated for 30 min with MACH2 goat anti-rabbit IgG-HRP polymer conjugate (catalog no. RHRP520, Biocare Medical), followed by the TSA fluorophore VIVID 650 (catalog no. 7527, Bio-Techne; dilution 1:1,000) for 15 min. Slides were mounted in Vectashield VIBRANCE Plus DAPI (catalog no. H-1800, Vector Laboratories).

Digitizing was performed at the Center for Advanced Microscopy at Northwestern University. Slides were imaged using a TissueFAXS System (TissueGnostics, Vienna, Austria) on a Zeiss Axio Imager.Z1 stand with a 20× EC Plan-Neofluar lens and an Andor Zyla complementary metal oxide semiconductor camera. Image acquisition was performed using the TissueFAXS System (TissueGnostics). Tissue sections were outlined, and integrated density was calculated using Fiji. RNU6 was used as a positive control, and Scramble S1 was used as a negative control.

Double transfection of myofibroblasts isolated from lungs of individuals with IPF

Myofibroblasts were isolated from lung tissue obtained by biopsy from individuals with IPF. Myofibroblasts expressed positive histochemical staining for α-SMA.⁹⁹ Inhibitors and mimic plasmids were commercially synthesized (Exiqon, Germantown, MD). Cells were plated in 6-well plates and transfected with TransIT-LT1 Transfection Reagent (Mirus Bio, Madison WI) with the miR-29 mimic (100 nM) and the inhibitor of miR-199 when 80% confluent in Opti-MEM Reduced Serum Medium. Mutated reporter plasmids were used as controls. Media was changed to 0.1% BSA and was collected at 24, 48, and 72 h to perform a time course of miRNA expression. Protein was subsequently collected 48 h post-transfection to measure MMP-2 activity and CAV-1 expression.

Zymography

MMP-2 activity was measured in lung tissue and myofibroblast homogenates, as previously described.¹⁷ Briefly, samples and standards (Chemicon) were loaded onto 10% zymogram gels (Life Technologies). Following electrophoresis, gels were incubated for 24 h at 37°C in a gelatinase solution to allow for the determination of MMP-2 proteolytic activity without interference from associated tissue inhibitors. Relative MMP-2 activity was measured by densitometry using ImageJ version 1.48 (National Institutes of Health).

Ex vivo lung explants (3D lung tissue cultures)

Lungs from 16- to 18-month-old male mice were cannulated via a visible bronchus and filled with warm agarose (3%, Sigma-Aldrich) immediately after excision. Lung segments were cooled on ice for 30 min to allow solidification of the agarose and punched to a diameter of 4 mm using a biopsy punch.^44,100 Explants were transferred to an air-liquid interface with phenol red free MEM without serum, with antibiotics, and injected with 10⁵ control transfected or double-transfected IPF myofibroblasts or PBS control in a volume of 100 μL with 29G needles. The 100-μL volume of injection was divided equally into four injection sites on the explants. Following injections, explants were incubated at 37°C in a humidified atmosphere of 5% CO₂ for 4 days. Some explants were fixed in 10% formalin (Sigma-Aldrich), processed for paraffin embedding, and stained with trichrome to assess histology and immunohistochemical analysis. Parallel explants from the same experiment were prepared for RNA and protein purification.

Quantification of telomere length

Total DNA was extracted from sections of human lung tissue using the SpeeDNA isolation kit. Whole lung tissue was lysed, and DNA was isolated by spin column extraction. The length of telomeres was quantified by real-time PCR analysis with TaqGreen tagging according to the manufacturer’s directions (ScienCell Research Laboratories, Carlsbad, CA).

Statistical analysis

Data were analyzed using GraphPad Prism 10.0 (San Diego, CA). All values were expressed as mean ± SEM. The overall significance of differences within the experimental groups was determined using the Kruskal-Wallis test and Mann-Whitney test. p Values less than 0.05 were considered statistically significant.

Data and code availability

All datasets are available upon request.

Acknowledgments

All authors have read the journal’s authorship agreement and policy on disclosure of potential conflicts of interest and would like to thank the Lester and Sue Smith Foundation and the Samrick Foundation for their generous support in research funding. The authors also thank Dr. Guy Howard and his group for assistance using the μCT machine. This paper is dedicated to Lester Smith (1942–2018) and David Samrick (1944–2021) and all the wonderful patients who we have cared for over the years. Histology and immunofluorescence services were provided by the Northwestern University Mouse Histology and Phenotyping Laboratory, which is supported by National Cancer Institute grant no. P30-CA060553, awarded to the Robert H. Lurie Comprehensive Cancer Center.

Author contributions

M.K.G. and S.J.E. wrote the manuscript and determined the study design. G.A.R., D.A.-T., X.X., S.P.-S., P.C., and S.S. were responsible for data collection, analysis, and manuscript editing. B.R., G.C., and E.S.H. were responsible for data interpretation and figure design. C.A. performed the imaging.

Declaration of interests

The authors declare no competing interests.