Low-dose administration of bleomycin leads to early alterations in lung mechanics
Funding information:
This work was funded by a Discovery Award of the UTHealth Pulmonary Center of Excellence (H.K.-Q.).
Edited by: Jeremy Ward
Abstract
New Findings
-
What is the central question of this study?
When do alterations in pulmonary mechanics occur following chronic low-dose administration of bleomycin?
-
What is the main finding and its importance?
Remarkably, we report changes in lung mechanics as early as day 7 that corresponded to parameters determined from single-frequency forced oscillation manoeuvres and pressure–volume loops. These changes preceded substantial histological changes or changes in gene expression levels. These findings are significant to refine drug discovery in idiopathic pulmonary fibrosis, where preclinical studies using lung function parameters would enhance the translational potential of drug candidates where lung function readouts are routinely performed in the clinic.
Idiopathic pulmonary fibrosis (IPF) is the most widespread form of interstitial lung disease and, currently, there are only limited treatment options available. In preclinical animal models of lung fibrosis, the effectiveness of experimental therapeutics is often deemed successful via reductions in collagen deposition and expression of profibrotic genes in the lung. However, in clinical studies, improvements in lung function are primarily used to gauge the success of therapeutics directed towards IPF. Therefore, we examined whether changes in respiratory system mechanics in the early stages of an experimental model of lung fibrosis can be used to refine drug discovery approaches for IPF. C57BL/6J mice were administered bleomycin (BLM) or a vehicle control i.p. twice a week for 4 weeks. At 7, 14, 21, 28 and 33 days into the BLM treatment regimen, indices of respiratory system mechanics and pressure–volume relationships were measured. Concomitant with these measurements, histological and gene analyses relevant to lung fibrosis were performed. Alterations in respiratory system mechanics and pressure–volume relationships were observed as early as 7 days after the start of BLM administration. Changes in respiratory system mechanics preceded the appearance of histological and molecular indices of lung fibrosis. Administration of BLM leads to early changes in respiratory system mechanics that coincide with the appearance of representative histological and molecular indices of lung fibrosis. Consequently, these data suggest that dampening the early changes in respiratory system mechanics might be used to assess the effectiveness of experimental therapeutics in preclinical animal models of lung fibrosis.
1 INTRODUCTION
Interstitial lung diseases are a heterogeneous group of restrictive lung disorders that are characterized by the deposition of fibrotic tissue in the lungs, which progressively reduces gas exchange and, ultimately, leads to death from respiratory failure. One of the most widespread forms of lung fibrosis is idiopathic pulmonary fibrosis (IPF; Lederer & Martinez, 2018). Features of IPF include varying degrees of inflammation, aberrant fibroblast proliferation and extracellular matrix deposition, resulting in loss of lung function (Lederer & Martinez, 2018). Idiopathic pulmonary fibrosis is more prevalent in males than in females in their fifth or sixth decade of life (King, Pardo, & Selman, 2011; Lederer & Martinez, 2018; Raghu, Weycker, Edelsberg, Bradford, & Oster, 2006). This may be because of the effect of oestrogens on collagen metabolism protecting premenopausal women, or some X-linked genes that attenuate the development of fibrosis (Tatler et al., 2016). A history of smoking and inhalation exposure to environmental toxicants, such as mining dust, have been associated with an increased risk of IPF (Costabel, 2015; Lederer & Martinez, 2018; Ley & Collard, 2013). In addition to these environmental risk factors, genetic factors have been associated with IPF, the most prevalent being variations of the MUC5B gene (Dickey & Whitsett, 2017; Lederer & Martinez, 2018; Seibold et al., 2011). Despite the emergence of two new compounds for the treatment of IPF, pirfenidone and nintendanib (Martinez et al., 2017), lung transplantation remains as the only ‘curative’ option (Lederer & Martinez, 2018), despite a low 5 year mean survival rate after transplantation (Trulock et al., 2006). As the risk for IPF increases with age, this affects the selection criteria for transplantation, because other co-morbidities, such as hypertension and diabetes, may exclude patients from being suitable for transplantation (Davis & Garrity, 2007). The prognosis is poor for individuals with IPF, because as many as 80% of patients could die within 6 years of diagnosis, which highlights the urgent need for new therapies (Selman, King, & Pardo, 2001). One of the most challenging aspects of drug discovery for IPF is the identification of new targets/molecules using preclinical animal models of lung fibrosis that can be translated effectively to the clinic (Carrington, Jordan, Pitchford, & Page, 2018).
There is an extensive variety of experimental models of lung fibrosis, including transgenic mice or exposure to bleomycin (BLM), silica, asbestos, paraquat or fluorescein isothiocyanate (FITC) and gene overexpression models that have been reviewed extensively (Carrington et al., 2018; Degryse & Lawson, 2011; Jenkins et al., 2017; Moore et al., 2013). Despite this array of models, very few compounds that have shown efficacy in animal models have been successful in human clinical trials (Jenkins et al., 2017). This discrepancy may result from the failure of animal models to recapitulate fully the histopathophysiology of human IPF (Rabeyrin et al., 2015; Raghu et al., 2011) and flaws in the design of preclinical animal models (Carrington et al., 2018; Jenkins et al., 2017), where new compounds or therapies are often tested prophylactically but not therapeutically (Carrington et al., 2018; Jenkins et al., 2017). As a result, the translational potential of these new therapies is limited. Furthermore, the success of experimental therapeutics in animal models of lung fibrosis often relies on the attenuation of histological and molecular indices of lung fibrosis (e.g. collagen content assays, histological scoring and fibrotic gene expression). In humans, disease progression in IPF patients is often determined using imaging and lung function analyses (Martinez & Flaherty, 2006; Trawinska, Rupesinghe, & Hart, 2016; Zisman et al., 2000). Consequently, there is a need to examine more clinically relevant parameters of animal models of lung fibrosis (Carrington et al., 2018). A recent review that discusses animal models of lung fibrosis proposed the use of imaging modalities and lung function analysis as a way forward in preclinical studies of lung fibrosis to provide a better assessment of their translational potential to humans (Carrington et al., 2018).
The most commonly used model of lung fibrosis is the intratracheal (i.t.) instillation of BLM (Carrington et al., 2018; Degryse & Lawson, 2011; Jenkins et al., 2017; Moore et al., 2013). Bleomycin causes double- and single-strand DNA breaks through the production of DNA-cleaving superoxide and hydroxyl free radicals (Claussen & Long, 1999). Lower levels of bleomycin hydrolase in the lungs account for its fibrotic effect in this organ (Sebti, Mignano, Jani, Srimatkandada, & Lazo, 1989). In the i.t. model of BLM-induced lung fibrosis, a single dose of BLM is instilled and results in an inflammatory response that persists for 7–10 days that ultimately leads to lung fibrosis 14–21 days after BLM administration (Carrington et al., 2018; Degryse & Lawson, 2011; Jenkins et al., 2017; Moore et al., 2013). Although this experimental model was used to show the effectiveness of both nintedanib and pirfenidone (Liu, Lu, Kang, Wang, & Wang, 2017; Oku et al., 2008; Wollin et al., 2015), there are several imperfections in this model that could affect future approaches to IPF drug development (Carrington et al., 2018; Degryse & Lawson, 2011; Jenkins et al., 2017; Moore et al., 2013). First and most notably, fibrosis completely resolves over time in these mice, which is in contrast to the progressive and non-resolving nature of fibrosis in IPF. Second, in the BLM-induced animal model of lung fibrosis, peribronchiolar fibrotic deposition is usually peribronchial, which contrasts with the subpleural presentation in IPF patients (Jenkins et al., 2017; Moore et al., 2013). Third, the initial BLM-induced inflammatory reaction observed in mice is not observed in the majority of IPF patients (Jenkins et al., 2017; Moore et al., 2013), perhaps with the exception of patients suffering from the fibroproliferative phase of acute respiratory distress syndrome (Beitler, Schoenfeld, & Thompson, 2014; Matthay & Zemans, 2011; Meduri, 1995).
Several studies have demonstrated changes in respiratory system mechanics after i.t. BLM exposure in rodents (Manali et al., 2011; Manitsopoulos et al., 2018; Phillips et al., 2012; Pinart, Faffe, & Romero, 2012). These studies show changes in lung mechanics that coincide with the presence of lung inflammation and fibrotic deposition in the lungs (Manali et al., 2011; Phillips et al., 2012).
An increasingly used animal model of lung fibrosis is the intraperitoneal (i.p.) model of BLM-induced lung fibrosis (Baran et al., 2007; Zhou et al., 2011), in which low doses of BLM are administered i.p. over 4 weeks and mice are assessed 33 days after the first i.p. injection of bleomycin. Fibrotic deposition in the i.p. BLM model is observed in the subpleural areas of the lung that is also observed in IPF patients and is thus more clinically relevant (Degryse & Lawson, 2011; Della Latta, Cecchettini, Del Ry, & Morales, 2015; Moore et al., 2013; Srour & Thébaud, 2015). Chronic i.p. administration of BLM causes a more persistent fibrotic response that is typically non-reversible (Ask et al., 2008; Degryse & Lawson, 2011; Pardo et al., 2005). Using this experimental model of lung fibrosis that more closely resembles the clinical presentation of IPF (Degryse & Lawson, 2011; Della Latta et al., 2015; Moore et al., 2013; Srour & Thébaud, 2015), our aim was to identify temporal changes in respiratory system mechanics that are also observed in IPF patients. Identification of early changes in lung mechanics in this model that are more closely aligned with physiological measurements in patients (Carrington et al., 2018) will be essential to guide future drug development efforts.
Based on our previous assessment of lung function 33 days after a single i.p. injection of BLM (Karmouty-Quintana et al., 2012, 2015), we hypothesized that mechanical changes in the lung parenchyma would closely follow the appearance of histological and molecular indices of lung fibrosis. Surprisingly, in the absence of any substantial changes in histological or molecular indices of lung fibrosis, we observed significant alterations in respiratory system mechanics and pressure–volume (P–V) relationships as early as 7 days after the first i.p. BLM administration. These results demonstrate that dampening alterations in respiratory system mechanics might be a physiologically relevant tool to assess the translational effectiveness of potential IPF therapeutics.
2 METHODS
2.1 Animals and ethical approval
All studies were reviewed and approved by The University of Texas Health Science Center at Houston Animal Welfare Committee (protocol no. AWC 16-0060). Ninety 4- to 5-week-old male C57BL/6J mice, weighing between 20 and 22 g, were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Male mice were selected for these experiments because IPF is more prevalent in men than in women (King et al., 2011; Lederer & Martinez, 2018; Raghu et al., 2006). Male mice also have a greater response to BLM than female mice irrespective of age (Redente et al., 2011). In addition, the American Thoracic Society's recommendations on use of animal models for the preclinical assessment of potential therapies for pulmonary fibrosis state that initial studies should be performed on male mice (Jenkins et al., 2017). All mice were housed in ventilated cages equipped with a microisolator lid, kept at an ambient temperature of 22°C, and exposed to a 12 h–12 h light–dark cycle with food and water ad libitum. Animal care was in accordance with institutional and National Institutes of Health guidelines. A minimum of four mice per group was used for histological studies and five per group for all functional studies. Animals were randomized into treatment groups by cage, with all littermates receiving either PBS (vehicle control) or BLM. Sample sizes were selected based on a power analysis identifying the minimal number required to conduct statistical analysis.
2.2 Induction of fibrosis
Mice were administered 0.035 U g−1 i.p. of BLM (APP Pharmaceuticals, Schaumburg, IL, USA) or vehicle (PBS; Invitrogen, Carlsbad, CA, USA) twice a week for 4 weeks. On days 7, 14, 21, 28 and 33 after the first i.p. administration, respiratory system mechanics and P–V relationships were measured. Furthermore, after the respiratory system mechanics measurements, lung tissue was collected for histological and gene expression analyses. Given that mice were examined at different times points during BLM or PBS exposure, the mice did not all receive the same number of injections. For example, mice examined on days 7, 14, 21, 28 and 33 received, respectively, two, four, six and eight PBS or BLM injections (Figure 1a). All analyses were conducted in a manner in which the investigators were blinded to the respective treatment groups.
2.3 Lung function analysis
On days 7, 14, 21, 28 and 33 of the bleomycin treatment regimen, respiratory system mechanics and P–V relationships were measured using the flexiVent (SCIREQ Inc., Montreal, QC, Canada) as previously described (Dahm et al., 2014; Karmouty-Quintana et al., 2012, 2015; Malik et al., 2017). In brief, the mice were anaesthetized with pentobarbital sodium (50 mg kg−1, i.p., Oak Pharmaceuticals, Inc.; Lake Forest, IL, USA) and xylazine hydrochloride (7 mg kg−1, i.p., Vedco Inc.; Saint Joseph, MO, USA). Once surgical anaesthesia was achieved, the mouse was tracheotomized using a 19-gauge metal cannula (Brico), connected via an endotracheal cannula to the flexiVent and ventilated at a respiratory rate of 150 breaths min−1 and tidal volume of 10 ml kg−1 against a positive end-expiratory pressure of 3 cmH2O. flexiVent software version 5.3 was used to perform the forced oscillation manoeuvres and generate P–V loops.
The linear single-frequency forced oscillation technique (FOT) was used to assess total respiratory system resistance (R), compliance (C) and elastance (E) with the Snapshot-150 perturbation. The broadband FOT was used to determine Newtonian resistance (Rn), tissue elastance (H) and tissue dampening (G) with the Quicktime-3 perturbation. Volume-driven P–V loops were formed from incrementally inflating the lungs to 40 ml kg−1 from functional residual capacity, which was defined as 3 cmH2O. After the delivery of each volume increment, the airway opening pressure was recorded. The area of the P–V curve and variables from the Salazar–Knowles equation were calculated on flexiVent software to provide data for quantitative analysis of the elastic properties of the lung. All measurements of respiratory system mechanics were conducted in mice with intact chest walls. Upon completion of the measurements, anaesthetized animals were killed by cervical dislocation.
2.4 Histological and gene expression analyses
On day 7, 14, 21, 28 and 33, separate cohorts of mice were weighed and killed with Avertin (500 mg kg−1). Avertin was prepared in house and was a mixture of tert-amyl alcohol (EMD Millipore corporation, Billerica, MA, USA) and 2,2,2 tribromoethanol (Sigma-Aldrich, St Louis, MO, USA). When the mice were sufficiently anaesthetized the neck was cervically dislocated and the lungs were inflated with 10% phosphate-buffered formalin at 25 cm H2O and fixed at 4°C overnight. Lungs were dehydrated in ethanol gradients and embedded in paraffin, and 5-μm-thick tissue sections were placed on microscope slides and stained with Masson's Trichrome (EM Science, Gibbstown, NJ, USA) according to the manufacturer's instructions. Next, two investigators blinded to treatment groups scored the lungs using a modified Ashcroft score (Collum et al., 2017; Karmouty-Quintana et al., 2012, 2015). Finally, RT-qPCR primers and analyses were performed by using our previously published methods (Collum et al., 2017).
2.5 Data presentation and statistics
A two-way ANOVA with the two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli was performed. Statistical significance was defined as P ≤ 0.05 by use of GraphPad Prism version 7 (GraphPad Software, La Jolla, CA, USA). Results are expressed as means ± SD. The data and statistical analysis comply where possible with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2015).
3 RESULTS
3.1 Single-frequency FOT
We first performed a single-frequency oscillation manoeuvre using a sinusoidal waveform. During the execution of a sinusoidal forcing function, the flexiVent calculates the following: (i) volume of gas delivered to the animal; (ii) flow of gas at the airway opening; and (iii) pressure at the airway opening. From these variables, dynamic respiratory system resistance (R), compliance (C) and elastance (E) are calculated using the equation of motion of a single-compartment linear model of respiratory system mechanics (Wald, Jason, Murphy, & Mazzia, 1969). The volume, pressure and flow signals from the snapshot are fitted to the single-compartment model using linear regression to obtain R and E. Compliance is simply the reciprocal of the elastance (C = 1/E) and vice versa.
When compared with vehicle (PBS)-treated animals, BLM administration caused a significant decrease in C as early as day 7 after the first BLM administration (Figure 1b). In addition, C was also significantly less in BLM- compared with vehicle-treated animals on days 14, 21, 28 and 33 after the first BLM administration (Figure 1b). No changes in C were detected between vehicle-treated mice at different time points; however, C on day 33 was significantly less than on day 7 in BLM-treated mice (Figure 1b).
Bleomycin treatment also led to changes in E. Bleomycin increased E as early as day 14 of the BLM treatment regimen and remained greater in BLM- compared with vehicle-treated mice on on days 21, 28 and 33 (Figure 1c). Elastance continued to increase over time as the number of BLM administrations increased such that E was significantly greater in BLM-treated mice on days 28 and 33 compared with day 7 (Figure 1c).
Resistance, which primarily represents resistance changes in the throrax, was greater in BLM- compared with vehicle-treated mice only on day 33 (Figure 1d). No significant differences in R were detected between PBS-exposed mice. However, in BLM-treated mice, R on day 33 was significantly elevated compared with day 7 (Figure 1d).
Taken together, these results demonstrate that low-dose systemic administration of BLM leads to significant early increases and decreases, respectively, in C and E, which is consistent with pathological changes in the lung parenchyma (i.e. increased collagen deposition). These data also demonstrate that indices of lung recoil worsen as the number of BLM administrations increases.
3.2 Broadband FOT
Although the single-frequency FOT is capable of identifying changes in C, E and R, it cannot precisely differentiate between loci within the lung that lead to the changes in lung mechanisms as a whole. In order to identify changes pertaining to the central and conducting airways as opposed to the lung parenchyma, the broadband FOT was performed. The broadband FOT uses the constant phase model to discriminate between changes in the conducting airways and the lung parenchyma. In these studies, we demonstrate that BLM treatment leads to an increase in tissue elastance (H) on day 28 that continues until day 33 (Figure 2a), which is consistent with changes in dynamic compliance at these same time intervals. Among BLM-treated mice, significant differences in H were observed between days 7 and 28 and betweens days 7 and 33. There were no differences in H among PBS-treated mice (Figure 2a).
Lung tissue damping or tissue resistance (G), which measures resistance over a broad range of frequencies (1–2.5 Hz), was also increased in a manner similar to H (Figure 2b). No significant differences in G were seen among PBS-treated mice; however, significant differences in G were observed between days 7 and 33 among BLM-exposed mice (Figure 2b).
No significant changes in Newtonian resistance (Rn), a parameter that determines the resistance in conducting airways, were identified between PBS- and BLM-treated mice (Figure 2c).
Taken together, these results demonstrate that low-dose i.p. administration of BLM alters the mechanical properties of the lung parenchyma and not the conducting airways, an observation that is seen in IPF patients.
3.3 Quasi-static mechanical properties of the respiratory system
In order to capture changes in the quasi-static mechanical properties of the lung after BLM exposure, P–V loops were obtained from PBS- and BLM-treated mice on days 7, 14, 21, 28 and 33 after the first BLM administration. Data from the P–V loops, when fitted to the Salazar–Knowles equation, revealed a reduction in quasi-static compliance as early as day 7 after BLM treatment that remained attenuated in BLM-treated mice for the duration of the study (Figure 3a). Among BLM-treated mice mice, CST was significantly lower in mice on days 28 and 33 compared with day 7 (Figure 3a). However, no differences in CST were observed among PBS-treated mice (Figure 3a).
Interestingly, quasi-static elastance was significantly different between PBS- and BLM-treated mice only on days 28 and 33 (Figure 3b). Among BLM-treated animals, quasi-static elastance was greater on day 33 than on day 7, and no differences in quasi-static elastance were observed among PBS-treated mice (Figure 3b).
The parameter K, which represents the slope of the deflationary limb of the P–V curve, was reduced in BLM- compared with vehicle treated mice on days 28 and 33 (Figure 3c), which is consistent with the presence of lung fibrosis. Within the different cohorts of BLM-treated mice, significant differences in K were seen between days 7 and 28 and between days 7 and 33, whereas no significant changes were seen among PBS-treated mice (Figure 3c).
Bleomycin significantly increased the area of the P–V curve at all time intervals examined (Figure 3d). No significant changes in area were seen among PBS-treated mice, yet within BLM-treated mice significant changes in area were seen between mice on days 7, 21 and 28 (Figure 3d). These changes are consistent with a rightward shift of P–V loops as seen for days 14 and 33 for BLM-treated mice (Figure 3e,f). In line with these changes, we report significant decreases in A, an estimate of inspiratory capacity determined by the Salazar–Knowles equation, as early as day 7 after the first BLM administration, that remains attenuated during the progression of disease (Figure 3g).
3.4 Histological analysis
In order to equate changes in lung mechanics with fibrotic deposition in the lungs, we stained lung sections obtained from PBS- and BLM-treated mice with Masson's Trichrome to determine the extent of fibrotic deposition (Figure 4a). Our data from Ashcroft scores demonstrate increased fibrotic changes as early as day 7 of BLM treatment that progressively increased as the number of BLM administrations increased (Figure 4b). In line with this, we report significant difference among BLM-treated mice between day 7 versus 21, 28 and 33, but no changes among PBS-treated mice (Figure 4b).
These changes were consistent with markers of gene expression for fibrosis that revealed a significant increase in fibronectin, Fn, on days 14 and 28 of BLM treatment (Figure 4c). Expression levels of collagen 3a1 (Col3a1) and hyaluronan synthase 2 (Has2) were increased on day 28 of BLM treatment and remained elevated on day 33 of BLM treatment (Figure 4d,e). Within BLM-treated mice, significant increases in Col3a1 and Has2 were observed between days 28 and 33, and no changes were seen among PBS-treated mice. Taken together, these results demonstrate a progressive fibroproliferative deposition in the lungs after BLM exposure that is apparent both functionally and histologically on day 7, preceding gene expression, in which the earliest changes are seen on day 14 after BLM.
4 DISCUSSION
In this study, we undertook a systematic assessment of lung mechanical forces over time in the i.p. BLM model to identify the time course of functional alterations in the development of experimental fibrosis. Remarkably, we identified early changes in parameters derived from the single-frequency FOT and from P–V loops that encompass the entire respiratory system (i.e. lung and chest wall), whereas parameters derived from the broadband FOT that are specific to the lung parenchyma (G and H) were altered much later in time. The early changes in C (day 7) and P–V loop data (CST and area, day 7) were accompanied by histological evidence of fibrotic deposition and gene expression that were apparent as early as days 7 and 14, respectively. These results are significant because they demonstrate that low-dose systemic exposure to i.p. BLM is able to induce physiological and morphological changes as early as day 7 in C57BL/6 mice and indicate that lung function is a robust and physiologically relevant tool to guide future efforts in drug discovery for IPF.
The i.p. BLM model, in which alterations in lung function are evident as early as day 7, might represent subclinical injury and a critical window for therapeutic intervention. Thus, one could envision administering an agent as early as day 7 of i.p. BLM treatment and assessing improvements in lung function as early as day 15. This approach would demonstrate the capacity of new agents to halt the progression of lung fibrosis, once fibrotic matrix deposition is present. The use of imaging techniques, such as magnetic resonance imaging, in the i.p. BLM model would refine drug discovery approaches further, because the same mouse could be imaged before treatment with BLM and at subsequent time points thereafter (Karmouty-Quintana et al., 2007). Indeed, this approach has been taken successfully in both mouse and rat models of lung fibrosis where single or repeated administration of BLM to the airways was performed (Babin et al., 2011; Egger et al., 2013, 2014) and the effect of therapeutic agents was tested (Egger et al., 2014).
Our results using the single-frequency FOT are consistent with previous experiments, in which BLM administration in both the i.t. and the i.p. models resulted in reduced C and elevated E (Karmouty-Quintana et al., 2012, 2015; Vanoirbeek et al., 2010). Consistent with our previous results, R was elevated on day 33 after the first i.p. BLM administration (Karmouty-Quintana et al., 2012, 2015), which did not appear to be affected in the i.t. BLM model (Vanoirbeek et al., 2010). Although our data show changes in C, and elevated E on day 33, we report changes in these parameters as early as days 7 (C) and 14 (E) that worsen during the course of BLM exposure. Although C is the reciprocal of E, no significant changes were seen on day 7 for E. A possible explanation is the fact the reciprocal number might result in a larger spread of data for E, and these values are at the verge of significance (PBS versus BLM P = 0.0675) versus C, where P = 0.0021 for PBS versus BLM on day 7. These findings are of significance because they demonstrate that the single-frequency FOT can be used effectively to track the progression of fibrosis physiologically in mice.
More commonly used parameters to assess changes in lung mechanics resulting from fibrosis are G and H, which are derived from the constant phase model from the broadband FOT (Thamrin, Janosi, Collins, Sly, & Hantos, 2004). These parameters have also been shown to correlate significantly with histological findings and hydroxyproline content in BLM-treated mice (Manali et al., 2011). The constant phase model uses the broadband FOT to separate out signals from the conducting airways (Rn) or the lung parenchyma (G and H). In line with previous results using both the i.t. and the i.p. models of lung fibrosis (Devos et al., 2017; Karmouty-Quintana et al., 2012, 2015; Manali et al., 2011; Vanoirbeek et al., 2010), we report elevated that BLM increased G and H but had no effect on Rn. These changes are consistent with those of a restrictive respiratory disease, in which chronic lung injury results in changes in the parenchyma but spares conducting airways (Vanoirbeek et al., 2010). Although fibrotic deposition is observed histologically as early as days 7 and 14, we report changes in G and H starting on day 28 that are still present on day 33.
These results are surprising when compared with the data from the single-frequency FOT because the single-frequency FOT takes into account the entire thorax, not only the alveolar compartment where fibrotic scarring is most prevalent. Thus, it would be expected that parameters from the broadband FOT would reflect more closely the progression of fibrosis than data from the single-frequency FOT. This observation is also supported by other investigators using the i.t. BLM model, in which increased fibrotic deposition was observed on day 7 but no changes in G or H were detected at this time point (Manali et al., 2011).
Collectively, these results demonstrate that the single-frequency FOT is more sensitive than the broadband FOT at detecting early functional changes in the BLM model. Given the results obtained with the single-frequency FOT and those from the lung tissue histology, these observations appear to suggest, as previously reported (Bates, Davis, Majumdar, Butnor, & Suki, 2007), that the three-dimensional organization of the lesions could also play a role in the functional manifestations of the disease and that some outcome parameters, either by their nature or by the way they are assessed, might be more sensitive than others at detecting the early signs of respiratory dysfunction associated with this i.p. BLM model. Being induced by repeated i.p. injections, it is possible that, in this respiratory disease model, extrapulmonary organs might be affected after BLM exposure. Interestingly, reductions in glycogen content in the heart and diaphragm have been identified after bleomycin- or paraquat-induced lung injury (Borges et al., 2014). These changes in glycogen are associated with reduced muscle function that might contribute to the early alterations in dynamic mechanics in our BLM-exposed mice. Additionally, in the early stage of the induced disease model, a larger volume amplitude could be needed during the FOT measurements to stretch the tissue fibres sufficiently that impairment becomes apparent and can be detected (Suki, Stamenovic, & Hubmayr, 2011). The volume of air delivered during an FOT measurement can vary, and there are differences between the single-frequency (amplitude set at 10 ml kg−1) and the broadband (amplitude set at 3 ml kg−1) FOT used in the present study. Thus, it is possible that a larger volume of air delivered in the single-frequency FOT might detect smaller fibrotic changes, because the air delivered might be able to expand fibrotic areas that are not affected by the lower volume delivered by the broadband FOT, which can dissipate into areas of lesser resistance. In addition, changes in lung surfactant after BLM administration might also contribute to changes in lung mechanics (Guillamat-Prats, Gay-Jordi, Xaubet, Peinado, & Serrano-Mollar, 2014; Pinart et al., 2012) and reflect alterations in IPF patients (Martinez et al., 2017; Selman et al., 2001).
The changes we found in P–V loops are similar to the changes observed clinically. Herein, we report reduced quasi-static compliance, which reflects the intrinsic elastic properties of the respiratory system (lung and chest wall) at rest, over the entire inspiratory capacity, as early as day 7, and continuing to decline as disease progresses. In line with the loss of elastic properties of the lung, we report increased quasi-static elastance, a measure of elastic stiffness of the lung at rest, on days 28 and 33 of BLM exposure, that are line with a reduced elasticity index (K) derived from the Salazar–Knowles equation. The value of K has also been found to be altered in certain respiratory diseases. In a study in patients with pulmonary fibrosis, a moderate but significant correlation was found between K and the severity of lung fibrosis from patient biopsies, suggesting that the shape of the P–V curve predicts the degree of lung fibrosis and could be useful in clinical assessment (Sansores et al., 1996). In the same study, no correlation was found between the level of fibrosis and conventional pulmonary function tests. This supports the findings in vivo because the majority of patients had >50% fibrotic deposition, thus P–V loop data might be representative of earlier stages of the disease.
Furthermore, our data also reflect early changes in the area of the P–V loops, starting on day 7 of BLM treatment and persisting and increasing as the disease progressed. These changes in area denote lung hysteresis. The changes in area of the P–V loop are strongly associated with indicators of fibrosis and restrictive airway diseases identified by increases in the area of the P–V loop (denoting areas of atelectasis) and a rightward shift in the sigmoidal curve of the expiratory slope. In the present study, assessment of the area of the P–V loop identified significant changes from day 7, increasing to day 33, which demonstrated an increased area of atelectasis and rightward shift in the curve. These changes are a result of multiple functional features, where a reduction in compliance results in higher pressures required to achieve the same volume, resulting in a rightward shift in the loop, sometimes culminating in a ‘beaked’ tip of the loop, which is indicative of alveolar distension. To overcome this, the lungs may compensate and reduce total volume, resulting in a shorter sigmoidal curve, which is indicative of hypoventilation and hypercapnia. In line with these changes, we also found a reduction in inspiratory capacity, consistent with previous studies following i.t. and i.p. BLM treatment (De Langhe et al., 2012; Devos et al., 2017; Li et al., 2017). Interestingly, we report changes in A, an estimate of inspiratory capacity, as early as day 7 of BLM treatment.
Our data derived from P–V loops are in agreement with previously published data in both the i.t. and the i.p. BLM models (Devos et al., 2017; Karmouty-Quintana et al., 2012, 2015; Manali et al., 2011). However, a major finding of our study is that both CST and the area of the P–V loop are altered as early as day 7 of BLM exposure. Similar to our data from the single-frequency FOT, the P–V loop data appeared more sensitive than the broadband FOT in detecting evidence of initial fibrotic injury in the BLM model. Taken together, these results highlight the single-frequency FOT and P–V loops as the most sensitive parameters to measure changes in lung function in experimental models of lung fibrosis. This is conceivable, because the P–V loop is measured over the entire range of the inspiratory capacity and is thus able to detect changes in the non-linear areas of the lung.
As opposed to FOT measurements, P–V loops are large-amplitude manoeuvres, designed to fully the alveoli recruit and to create a significant stretch of the tissue fibres so that the linear and non-linear aspects of the P–V relationship of the entire respiratory system can be studied over the inspiratory capacity range. In the present study, an inflation volume of 40 ml kg−1 was used to construct the P–V curves, which is a much larger volume amplitude than that used for FOT measurements. The structure of the lung is complex (Suki et al., 2011), and the detection of indices of respiratory dysfunction can require various technical approaches (Devos et al., 2017). Taken together, these results suggest that, during the early stages of the disease, high transpulmonary pressures need to be reached to probe for impairment of respiratory function with the greatest measurement sensitivity so that pathological structure–function changes can be linked.
An important aspect of lung injury is inflammation, which plays a central role in the development of experimental lung fibrosis (Karmouty-Quintana et al., 2015; Le et al., 2014; Muller et al., 2017; Pedroza et al., 2011; Philip et al., 2017; Zhou et al., 2011). Our group has previously identified temporal changes in inflammatory markers, such as interleukin-6, ATP and adenosine, that contribute to the recruitment and activation of neutrophils and macrophages, which promote fibrotic processes (Karmouty-Quintana et al., 2015; Le et al., 2014; Muller et al., 2017; Pedroza et al., 2011; Philip et al., 2017; Zhou et al., 2011). In these studies, although increased macrophages are observed as early as day 10 (Zhou et al., 2011), most inflammatory changes do not peak until days 17–21 or later (Karmouty-Quintana et al., 2015); thus, it possible that inflammation might not play a central role in early alterations of lung function in BLM-exposed mice. However, some of these inflammatory mediators are also capable of affecting lung function directly, with an obvious candidate being adenosine (Karmouty Quintana, Mazzoni, & Fozard, 2005), which is increased over time in our model (Karmouty-Quintana et al., 2015)
Our findings are in line with previous studies in which repeated doses of BLM have been administered (Babin et al., 2011; Degryse et al., 2010; Egger et al., 2013) and evidence of lung injury was reported as early as day 7 of BLM treatment. However, it is worth noting that in these studies BLM was administered directly to the airways, and thus the initial fibrotic response appears to be peribronchial, which differs from the subpleural presentation in IPF (Jenkins et al., 2017; Moore et al., 2013). In addition, despite early changes, these could be attributed to the inflammatory response and not necessarily to fibrotic deposition, given the known acute inflammatory effect of BLM when administered directly to the airways (Jenkins et al., 2017; Karmouty-Quintana et al., 2007; Moore et al., 2013; Zhou et al., 2011).
Interestingly, our data show increased Fn expression preceding Col3a1 and Has2 expression, peaking on days 14 and 28, whereas Col3a1 and Has2 peak on days 28 and 33. These differences might be attributable to the fact that during extracellular matrix remodelling, Fn plays a crucial role in matrix assembly, whereby initial Fn deposition is required for subsequent extracellular matrix deposition (Dallas et al., 2005; Sottile et al., 2007).
In conclusion, our study demonstrates that systemic low-dose exposure to BLM leads to early changes in lung function that are more readily captured by single-frequency FOT and P–V loops rather than by broadband FOT manoeuvres. These results are significant because they demonstrate that lung function is altered initially upon insult, when little fibrotic deposition is apparent histologically and before robust changes in gene expression or inflammation. These results also point at lung function as a robust and reproducible tool that can be used to identify new drug candidates that might be successful in the clinic, where lung function is almost exclusively used to guide assessment of disease progression.
ACKNOWLEDGEMENTS
Kelly Volcik, PhD helped to revise and proof-read the manuscript. Annette Robichaud, PhD helped us to interpret data from the flexiVent and in critically reviewing the manuscript.
AUTHOR CONTRIBUTIONS
L.H. made substantial contributions to the design, acquisition of data and analysis, in addition to drafting and revising the manuscript. W.B. made a substantial contribution to the acquisition and analysis of data. C.W. provided assistance in interpreting the data. M.C. contributed to the analysis and interpretation of data. T.D. and A.M.H. made a substantial contribution to the acquisition of data. T.M. was involved in conception and design, analysis and interpretation of data. S.C. made a substantial contribution to the acquisition and interpretation of data. S.R. helped with interpretation of results. S.R.S. and R.A.J. helped with initial study design and interpretation of results. H.K.-Q. made substantial contributions to the conception, design and analysis of the data, co-wrote and revised the manuscript. All authors have revised the manuscript critically for important intellectual content. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.
COMPETING INTERESTS
Stephanie Rosenbaum and Annette Robichaud work for SCIREQ, an Emka technologies company that manufactures and sells the flexiVent.
