Beyond Upper Airway Involvement: Evidence of Intrinsic Lung Disease in a Mouse Model of Mucopolysaccharidosis I

ABSTRACT

Almost all patients with mucopolysaccharidosis (MPS) develop respiratory dysfunction of varying severity during disease progression. While respiratory disease in MPS has traditionally been attributed to upper airway obstruction caused by glycosaminoglycan (GAG) accumulation in the trachea and bronchi, involvement of the intrapulmonary conducting airways and lung parenchyma remains poorly defined. Here, we characterised lung disease in a mouse model of MPS I using a combination of non‐invasive X‐ray Velocimetry (XV) functional lung imaging and gold‐standard flexiVent respiratory mechanics testing, complemented by lung volume measurements and histological analysis. XV provides regional ventilation information across the entire lung during tidal breathing. MPS I mice demonstrated reduced mean specific ventilation (the average regional expansion of lung tissue across the respiratory cycle), driven predominantly by reduced ventilation in the inner (mediastinal‐adjacent) lung regions, with evidence of spatially heterogeneous ventilation distribution. Lung mechanics testing showed increased conducting airway resistance, increased respiratory system compliance and reduced tissue elastance, consistent with impaired elastic recoil and expiratory flow limitation. Lung volume analysis revealed reduced opening pressure following degassing together with increased residual volume, functional residual capacity and vital capacity. Histological analysis demonstrated heterogeneous parenchymal architecture with regions of enlarged airspaces. Together, these findings demonstrate that respiratory dysfunction in MPS I is not limited to upper airway obstruction but also involves intrinsic abnormalities of the intrapulmonary conducting airways and lung parenchyma. This intrinsic pulmonary pathology likely contributes to obstructive lung disease and may underlie the susceptibility to respiratory failure observed in patients with MPS I.

1. Introduction

Almost all Mucopolysaccharidosis (MPS) patients have respiratory dysfunction with varying symptoms and increasing severity during disease progression [1]. Substrate accumulation in the lungs and surrounding tissues is presumed to have direct effects on lung structure and functional capacity [2, 3, 4, 5], which can ultimately result in respiratory failure. However, the molecular mechanism and impact of undegraded glycosaminoglycans (GAG) on lower airway disease and lung function have yet to be fully investigated.

Respiratory pathology in MPS patients is multifactorial and has mainly been attributed to extrinsic factors such as the gross anatomical changes associated with the tongue, nasal cavities, adenoids, tonsils, larynx and thoracic cavity [6], leading to frequent obstructive airway disease. In this manuscript, we use the following anatomical definitions: the upper airways refer to the nasal cavity, pharynx and larynx; central airways refer to the trachea and main extrapulmonary bronchi; intrapulmonary conducting airways refer to the bronchial tree within the lung parenchyma down to the terminal bronchioles; and the peripheral lung refers to the distal bronchioles, alveolar ducts and alveolar parenchyma. We define extrinsic pathology as abnormalities external to the lung parenchyma that mechanically influence airflow (e.g., craniofacial anatomy, thoracic cage, mediastinal structures) and intrinsic pathology as disease processes affecting the airway wall, lung parenchyma, or alveolar structure itself. Excessively thick secretions throughout the upper and lower airways can worsen airway obstruction, increasing the risk and incidence of obstructive sleep apnoea [7]. Importantly, intrinsic pathology—including glycosaminoglycan accumulation within the airway wall, altered parenchymal architecture and changes in pulmonary surfactant function—can involve the large conducting airways and result in tracheal and bronchial airway narrowing [1, 2]. Dyspnoea and airway disease involving the trachea is common in MPS patients, particularly in MPS I (alpha‐L iduronidase; Hurler syndrome) and MPS II (iduronate‐2‐sulfatase; Hunter syndrome) patients, where it presents as tracheomalacia resulting from tracheal deformity, stenosis and tortuosity [8]. Tracheal deformity is proposed to be the result of weakened supporting cartilage, due to GAG storage [9]. In addition, there is evidence of changes in adenotonsillar tissue in patients with MPS, with altered extracellular matrix remodelling involving increases in Collagen IV and laminin ⍺‐5 [10].

In MPS I and MPS II patients there is a high incidence of difficult and failed intubation during surgical interventions and frequent problems with upper airway obstruction due to respiratory pathogenesis [8]. In an adult MPS II patient cohort, pulmonary function tests together with inspiratory and expiratory static chest CT scans, showed extensive tracheobronchial collapse and a reduction in central airway diameter [9]. A recent case study in Malaysia reported a 40‐year‐old man who presented with chronic obstructive pulmonary disease (COPD) including unusual tracheal and bilateral bronchial stenosis who was subsequently diagnosed with Hunter syndrome [11]. While lung pathology in MPS patients has not been a primary research focus in the MPS field, both MPS I and MPS II patient cohorts display obstructive and restrictive airway disease either individually or together. Therefore, lung pathology is now being recognised as critical in MPS disease and importantly appears to be refractory to therapeutic intervention for example, enzyme replacement therapy [9, 10, 12]. Despite the deleterious clinical impact of respiratory issues in the MPS, there are very few preclinical studies in animal models.

To date, studies focussed on respiratory pathology in MPS patients have predominantly utilised traditional spirometry and static CT scans, which are limited to investigating global lung function and pathology in the upper airways. In the clinic, spirometry is commonly used for assessing lung function and it measures the volume of air exhaled at specific time points during forced expiration (FE). In rodents, this test can be achieved by rapidly applying a negative pressure to the airways to create a forced expiratory flow [13]. This technique is referred to as negative pressure‐driven forced expiration (NPFE) and can be performed alongside the forced oscillation technique (FOT). The Scireq flexiVent NPFE extension enables parameters such as the forced expiratory volume (FEV), forced vital capacity (FVC), Tiffeneau index (FEV/FVC) and peak expiratory flow (PEF) to be measured. Spirometry and spirometry‐like NPFE assessments do not enable the origin of airflow abnormalities to be identified, necessitating the development of additional technology for more detailed regional lung function assessment.

There is a need for MPS lung studies to utilise additional technology to further define the lung pathogenesis. X‐ray Velocimetry (XV) is a relatively new functional lung imaging method that tracks the motion of lung tissue during breathing to measure regional ventilation. XV breaks the analysis of the lung down into ~6000 specific regions and in each of those calculates a quantity termed specific ventilation; the change in volume of a region relative to its size at peak exhalation (ΔV/V ₀). From this a visual ventilation map can be created, along with metrics such as the mean specific ventilation (MSV), ventilation defect percentage (VDP; for XV the VDP is defined as the percentage of lung with ventilation below 60% of the MSV) and the ventilation heterogeneity (VH; the spread of the ventilation histogram; the interquartile range divided by the MSV) [14]. XV technology has been validated in animal models including the ꞵ‐ENaC cystic fibrosis mouse model, in which XV visualised the patchy lung disease and enabled regions of reduced airflow to be identified [15, 16, 17, 18] and in rat models of cystic fibrosis [19, 20].

While respiratory disease in the MPS has been attributed to extrinsic pulmonary pathology and upper airway obstruction, we hypothesised that the accumulation of GAG in the respiratory system of MPS patients leads to intrinsic conducting airway and peripheral parenchymal pathology affecting the conducting airways and parenchymal tissue. This could compromise lung function and increase the risk of acute respiratory failure. Here we quantified the lung phenotype in MPS I mice [21] using a unique combination of X‐ray Velocimetry (XV) functional lung imaging and flexiVent lung mechanics tests [19] including lung volume measurements [22]. Histology provided confirmatory evidence of lung structural changes that likely contribute to the observed alterations in lung function.

2. Methods

2.1. Animals

All animal procedures were approved by the University of South Australia (UniSA) animal ethics committee under applications U14‐22 (lung imaging and function), U36‐23 (lung volumes) and U22‐21 (histology) and were performed in accordance with the ARRIVE guidelines [23]. MPS I (IDUA−/−) knockout mice on a C57Bl/6 background [21] were provided by the UniSA Core Animal Facility. All animals were genotyped by an automated genotyping service (Transnetyx, Memphis, USA). XV and flexiVent studies were performed in n = 14 control (normal/heterozygote) and n = 15 MPS I age‐matched (~21 weeks) female mice. Females were selected for lung function testing to minimise sex‐related variability in respiratory mechanics and to maximise statistical power within feasible cohort sizes. Lung volume measurements were performed in n = 13 control and n = 11 MPS I female mice. Histological analyses were performed in n = 6 control and n = 8 MPS I mice of both sexes. Throughout their life mice were maintained in individually ventilated cages (IVC) with a 12‐h light/dark cycle. Food and water were provided ad libitum. All XV imaging was performed at the SAHMRI Preclinical Imaging and Research Laboratories (PIRL, South Australia).

2.2. X‐Ray Velocimetry (XV) Imaging

XV imaging was performed as described previously [18, 24]. Mice were anaesthetised with an intraperitoneal injection of a mixture of 75 mg/kg of ketamine (Ceva, Australia) and 1 mg/kg medetomidine (Ilium, Australia). Once fully anaesthetised, the mice were prepared as previously described by performing a tracheostomy followed by cannulation with a short endotracheal tube (ET; 18 Ga BD Insyte plastic cannula bevel‐cut to 15 mm length). Mice were then restrained head‐high in a custom 3D printed holder and placed onto the translation/rotation stage of a Permetium preclinical XV scanner (4DMedical, Australia) [19], configured with a 1700 mm sample to detector distance. Mice were connected to a pressure‐controlled small animal ventilator (4DMedical Accuvent 200, Australia) set to a peak inspiratory pressure of 14 cmH₂O, a positive end‐expiratory pressure of 2 cmH₂O and a rate of 150 breaths/min (200 ms inspiration and 200 ms expiration; I:E ratio of 1:1). A single 4D XV scan was acquired at a framerate of 25 Hz, with 10 phase‐points per breath and 600 projections per phase‐point. XV scan acquisition took less than 5 min.

All XV ventilation maps were produced by 4DMedical, using a proprietary three‐dimensional cross‐correlation algorithm that quantifies tissue displacement throughout the breath to calculate specific ventilation, MSV, VH, VDP and tidal volume (V_T). Specific ventilation represents the fractional change in volume of a lung region during a breath relative to its volume at end‐expiration; mean specific ventilation (MSV) therefore reflects the average regional expansion of lung tissue across the respiratory cycle. The 4DMedical software generates a ventilation report featuring lung slices with 2D maps of local ventilation. Specific ventilation data is colour‐coded: green for average ventilation, red for below average and blue for above average. Additionally, volumetric datasets of lung structure and specific ventilation were provided for each animal, enabling voxel‐based regional analysis (refer to the data repository for ventilation reports and 3D volumes). For this study, we report the normalised ventilation defect percentage (nVDP), which was calculated using 60% of the MSV of the control population to define a threshold for defective regions, rather than 60% of the MSV for each individual animal, as described by Reyne et al. [18, 20]. The total volume of the lungs at peak exhalation (lung tissue and air) was calculated from the size of the 3D volume produced by XV imaging and was reported as mean ± SD.

To visualise and quantify regional variations in ventilation, XV parameters were also calculated on subsets of the lung volume in a similar manner to that first proposed by Phipps et al. [25], and more recently applied to XV by Reyne et al. [18] and Karmali et al. [26]. A closing filter of kernel size 9 × 9 × 9 voxels (4.5 × 4.5 × 4.5 mm) was applied to a binary lung mask, followed by an erosion filter with size 3 × 3 × 3 voxels (1.5 × 1.5 × 1.5 mm). This created a mask excluding all regions of the lungs less than 4.5 mm from the surface in the areas near the rib cage, while including the inner surfaces around the heart cavity [18].

2.3. flexiVent Respiratory Mechanics

After XV scan acquisition, lung function assessments were performed using a flexiVent FX small animal ventilator (SCIREQ, Montreal, Canada) fitted with a FX2 mouse module and NPFE forced expiration extension and operated by flexiWare v8.0 software. Mice were placed in a supine position and integrated to the flexiVent via the 18 Ga cannula and ventilatory parameters set at a respiratory rate of 150 breaths/min, inspiratory to expiratory ratio of 2:3, tidal volume 10 mL/kg and positive end expiratory pressure (PEEP) of 3 cmH₂O. Lung mechanics measurements were made using a SCIREQ mouse mechanics scan script, with the automated algorithms set to default values [27].

A Deep Inflation manoeuvre was used to recruit closed lung areas, standardise lung volume history and calculate the inspiratory capacity (IC) by maximally inflating the lungs to a pressure of 30 cmH₂O over a period of 3 s. A single frequency (SnapShot‐150) perturbation that matched the mouse respiratory rate and tidal volume was then performed and the data was fitted to the single‐compartment model to measure the respiratory system resistance (R_rs) and compliance (C_rs). A broadband perturbation (Quick‐Prime 3) forced oscillation perturbation using a range of frequencies above and below the respiratory rate was then applied and the software then fitted the constant‐phase model of respiratory system impedance (Z_rs) to calculate Newtonian resistance (R_n; primarily the resistance of the large conducting airways), tissue damping (G; related to the tissue resistance) and tissue elastance (H; characterises the elastic energy storage within the tissues) [28]. The parameter Eta called hysteresivity was calculated by dividing G by H and used as a measure of heterogeneity. Pressure‐controlled stepwise pressure‐volume loops (PVs‐P) were generated and the Salazar‐Knowles equation automatically fitted to the expiratory data to obtain the maximal lung capacity (A), curvature of the deflating PV loop (K) and quasi‐static compliance (C _st). The area between the inspiratory and expiratory arms of the PV loop was also calculated. Finally, a NPFE perturbation was performed to calculate a flow‐volume curve and a range of parameters including forced expiratory volume (FEV), forced vital capacity (FVC), FEV to FVC ratio, forced expiratory flow (FEF), peak expiratory flow (PEF) and time to peak expiratory flow (TPEF). For each parameter three measurements were made per mouse and these were averaged. Data was excluded if the coefficient of determination (a measure of model fit) was less than 0.9 for each model.

Lung volume measurements were made in a second cohort of female animals by degassing the lung to allow inflation to start at zero volume, followed by construction of a full‐range PV curve covering the subject's entire lung volume range [22]. Animals were prepared in the same manner as described above, including anaesthesia and tracheostomy with a short endotracheal tube. They were then connected to the flexiVent, ventilated on 100% oxygen and paralysed using a 6 mg/kg intramuscular dose of vecuronium bromide (Viatris, Australia). A Deep Inflation was performed to inflate the lungs to 35 cmH₂O. Ventilation was then stopped and degassing was allowed to occur for 5 min with the ventilator valves closed to seal the system and allow oxygen within the alveoli to be absorbed. Then the lung was slowly inflated to TLC (35 cmH₂O) and then deflated to RV (−10 cmH₂O). Two additional inflation/deflation cycles were performed. The flexiVent software calculated the opening pressure (P_op), total lung capacity (TLC), vital capacity (VC), reserve volume (RV), functional residual capacity (FRC), expiratory reserve capacity (ERV), static compliance (C_st), specific compliance (C_s; static compliance normalised to FRC) and shape of the deflation limb (V₁₀_TLC; the volume at 10 cmH₂O normalised to TLC).

2.4. Histological Analysis

To determine any changes in tissue structure, a third cohort of age‐matched MPS I and control mice were humanely killed with an intraperitoneal injection of sodium pentobarbitone (~200 mg/kg). The trachea was cannulated and the lungs were instillation‐fixed in situ under gravity at a constant fluid pressure of 25 cmH₂O with 4% paraformaldehyde. The lungs and heart were removed and suspended in 4% paraformaldehyde overnight. The left and right lungs were separated and then processed using a Leica Histocore PEARL automatic tissue processor and embedded in paraffin. Tissue sections (2 μm) were cut from the formalin‐fixed paraffin‐embedded left lung blocks. To ensure sections were taken from the same region of the lung between samples, six domains of 13 sections each were cut and sections from domain 5 of each animal were chosen to stain. Following deparaffinisation and rehydration with xylene (2× 5 min) and a series of ethanol washes (1 min each in 2× 100%, 2× 90%), the tissue sections were stained with Ehrlich's Haematoxylin (5 min), differentiated in acid alcohol (1% HCl in 70% ethanol) and blued in 0.4% ammonia water (1 min). After washing in water, the sections were counterstained with Eosin (2 min) and dehydrated in a series of ethanol (1 min each in 1× 90%, 2× 100%) and xylene (2× 2 min). All slides were imaged in brightfield with an Axio Scan Z.1 slide scanner (Zeiss, Germany) with a plan‐achromat 20× objective.

The airspace (assigned as negative space) was determined in the lung tissue sections. To reduce bias, a non‐histologist chose seven random areas within each de‐identified tissue section, which were then extracted at 5× magnification (3400 pixels square). Images were processed using custom Matlab (The Mathworks, Natick, USA) code that converted images to greyscale and then to black and white using the imbinarize function (based on Otsu's method [29]). A distance transform was performed using the bwdist function to determine the size of the air spaces within each field. The output of this was averaged for each image and then for all of the seven randomly chosen fields per animal.

2.5. Statistics

All statistical analyses were performed in R version 4.3.1 [30]. Separate t‐tests were used to determine whether the mean age and weight of the control and MPS I mice were significantly different. For the remainder of the study, the statistical findings were expressed in terms of estimated marginal means and confidence intervals returned from linear models fitted to the data. For every flexiVent and XV parameter, a standard linear regression model was fitted using the ‘lm’ function with a fixed effect of genotype. When the XV ventilation data was split into inner and outer regions (see results), a linear mixed effects model was fitted using the ‘lme’ function with fixed effects of genotype and lung region and a random effect of mouse ID. Post hoc pairwise comparisons for the fitted models were carried out using the ‘emmeans’ package [31]. All results are graphically represented as box plots. Table S1 shows the estimated mean of each XV (full lung) and flexiVent parameter for the control and MPS animals, their 95% confidence intervals and the actual p‐value, derived from fitted linear models. The Pearson correlation coefficient was calculated to measure the relationship strength between each of the flexiVent and XV parameters. A t‐test was used to determine whether the airspace sizes were significantly different between MPS I and control.

3. Results

Data was collected from age‐matched female control (n = 14) and MPS I (n = 15) mice and there was no significant difference in the weight of the two groups. Compared to the control animals, the tracheal tissue in the MPS animals lacked structural integrity and in two MPS animals the tracheal tissue tore when the ET tube was inserted, likely due to the weakened supporting cartilage caused by GAG accumulation [9]. These animals were humanely killed prior to XV imaging.

3.1. X‐Ray Velocimetry Imaging

XV produces a three‐dimensional map of the specific ventilation at around 6000 voxels (0.5 × 0.5 × 0.5 mm analysis regions) across the lung. There were no statistically significant differences in the tidal volume (V_T; p = 0.92) or total lung volume (p = 0.087) measured by XV between the MPS I and control mice. Previous studies have shown that analysis of the spatial distributions of the ventilation variability within this 3D volume may give additional insight [17, 20]. Here, a visual inspection of the ventilation maps from all the animals suggested that the MPS I mice had more areas of low specific ventilation (i.e., regions that inflate less than the average) in the inner (mediastinal‐adjacent) regions of the lungs near to the heart than the control mice (Figure 1).

FIGURE 1.

Example slices from XV ventilation maps. The 3D ventilation map produced by XV from (A) control and (B) MPS I mice in the axial and coronal planes. Green indicates average ventilation, red below average ventilation and blue above average ventilation. MPS I mice appeared to have more areas of low specific‐ventilation (red) located within the central areas of the lung. (C and D) show the matching specific ventilation histograms. The nVDP region is shaded in red, the MSV is shown as a dashed line and the interquartile range (IQR; where VH = IQR/MSV) is shown as a solid line.

To visualise and quantify these regional variations, the lungs were split into an outer cortex and inner core region for analysis, with the outer regions containing 72% of the lung by volume on average and the inner region containing the remaining 28% of the volume (Figure 2A). The global parameter mean specific ventilation (MSV; the amount that each voxel expands during inspiration) was significantly lower in MPS I mice compared to control (estimated means 0.349 vs. 0.377; p = 0.023; Table S1) and this was driven by a large reduction in the inner lung region MSV (Figure 2B). MSV in the outer region of the lung was not significantly different between MPS I and control mice. The ratio MSV_Outer:MSV_Inner was greater in the MPS I animals than the control animals (p < 0.001). There were no statistically significant differences in ventilation heterogeneity between genotypes when looking at the full‐lung, inner‐lung or outer‐lung (Figure 2C). The inner lung normalised ventilation defect percentage was significantly greater in the MPS I mice compared to control (estimated means 26.2 vs. 14.5; p = 0.0004), despite the full‐lung nVDP not being significantly different (Figure 2D). As with the other parameters, nVDP in the outer lung was not significantly different between the two genotypes.

FIGURE 2.

X‐ray velocimetry. The XV ventilation volume was segmented into (A) inner (black) and outer (grey) regions, which showed (B) lower mean specific ventilation (MSV) in the full lung and inner lung analyses in MPS (red) I mice compared to control (blue), (C) no significant differences in ventilation heterogeneity and (D) increased normalised ventilation defect percentage in the inner lung region in MPS I mice compared to control.

The XV scan data also enables the thoracic structure to be examined. All MPS I mice were noted to have a much more curved spine, protruding sternum and barrel‐shaped chest than the control mice (Figure 3).

FIGURE 3.

Slices from CT scan of control (top) and MPS (bottom) mice showing the difference in bone structure and torso shape.

3.2. flexiVent Respiratory Mechanics

Selected results from the flexiVent testing are shown in Figure 4 and Table S1. The deep inflation manoeuvre showed that the MPS I mice had a higher inspiratory capacity compared to controls (Figure 4A). The single frequency forced oscillation (single compartment model) demonstrated higher total respiratory system resistance (R_rs) and total respiratory system compliance (C_rs) in MPS I mice compared to control (Figure 4B,C). The broadband forced oscillation (constant phase model) demonstrated that Newtonian resistance (R_n), which reflects resistance of the central conducting airways, is higher and the tissue elastance (H) is lower in MPS I than control mice (Figure 4D,E). Tissue hysteresivity (G:H) was also significantly higher in the MPS I mice (Figure 4F), but tissue damping (G) was not significantly different. Average pressure volume loops constructed from mean data showed an upward shift in the pressure‐volume relationship of MPS I mice compared to control—the shape parameter K was higher in the MPS I mice, but the PV loop Area was not significantly different (Figure 4G–I). The negative pressure forced expiratory test showed significantly decreased FEV0.05, FEV0.05/FVC ratio and PEF and increased FEF0.05 in MPS I mice compared to control (Figure 4J–L). FVC and TPEF were not significantly different.

FIGURE 4.

Overall mechanics of the respiratory system are significantly altered in MPS mice. (A) inspiratory capacity from the deep inflation manoeuvre. (B, C) respiratory system resistance and compliance from the single‐compartment model. (D–F) Central airway resistance, tissue elastance and tissue hysteresivity from the forced oscillation technique. (G–I) The stepwise pressure‐volume loop, upper deflation limb shape parameter K and pressure‐volume loop area. (J–L) Flow‐volume plot and selected flow parameters from the negative pressure forced expiratory test.

Lung volume analysis was performed on n = 13 control and n = 11 MPS I animals. Average lung volume curves constructed from mean data showed an upward shift in the pressure‐volume relationship of MPS I mice, compared to control (Figure 5A). In the lung volume manoeuvre the opening pressure (P_op), or the pressure needed to open the collapsed lung, was significantly reduced in the MPS I mice compared to control. TLC and VC were not significantly different, but MPS I animals also had significantly higher RV, FRC and ERV (see Figure 5 and Table S1). The static compliance (C_st) and shape of the deflation limb (V₁₀_TLC) were significantly higher in MPS I animals and the specific compliance (C_s) was significantly lower in the MPS I animals.

FIGURE 5.

Lung volumes are significantly altered in MPS mice. (A) Full range pressure volume curves, (B) opening pressure, (C) total lung capacity, (D) reserve volume, (E) functional residual capacity, (F) expiratory reserve volume, (G) static compliance, (H) shape of the deflation limb and (I) the specific compliance.

The relationships between the XV and flexiVent metrics were examined and there was a statistically significant correlation between the MSV measured from XV and hysteresivity (G:H) from the Quick‐Prime 3 perturbation (see Figure S1). Interestingly, when the correlations were altered to use the XV data from the inner lung only, five of the flexiVent parameters were significantly correlated with nVDP_Inner and six were significantly correlated with MSV_Inner (see Figure S2). The correlation coefficients between the XV parameters in the outer lung region and the flexiVent parameters were not statistically significant for any parameter.

3.3. Lung Tissue Histology

Histological analysis was performed on lung tissue from n = 6 control and n = 8 MPS I animals. Tissue sections from all control animals exhibited normal alveolar structure (Figure 6A (a–g)). In contrast, in most MPS I tissue sections the alveolar structure was heterogeneous, with isolated regions of disrupted alveolar tissue showing increased airspace (Figure 6B (a–g)). The amount of airspace was quantified using a Matlab script and found to be significantly higher in the MPS I lung tissue than the control tissue (Figure 7; p = 0.0035).

FIGURE 6.

Disrupted alveolar structure in MPS I mouse lung tissue. Representative images from (A) Control and (B) MPS I mouse lung tissue sections. (a–g) Seven randomly chosen regions in control and MPS I tissue sections showing regions of normal and disrupted alveolar structure.

FIGURE 7.

Mean airspace distance was significantly greater in the MPS I mice compared to control. Each data point consists of the mean of seven separate measures for the fields of view per whole lung slice per animal (as shown for the representative animals in Figure 6).

4. Discussion

Respiratory complications significantly contribute to morbidity and mortality in the MPS and have traditionally been attributed to upper airway obstruction. In this study, we used novel X‐ray Velocimetry functional lung imaging to dynamically visualise differences in regional airflow and structure at every point in the MPS I mouse lung, throughout the breath. This was complemented with traditional measures of global lung function and lung volume measurements using the flexiVent system and histological assessments of lung structure. We showed that respiratory disease in MPS I is not solely an upper airway pathology but can be attributed to an intrinsic aetiology, as pathology also involves the conducting airways, lung parenchyma and surfactant system. MPS I mice showed a reduction in the mean specific ventilation, suggesting that regions of the MPS I lungs did not expand uniformly and to the same extent as the control lungs. Specifically, there was reduced ventilation in the inner region of the lung. In addition, an increase in lung heterogeneity (G:H) suggested patchy ventilation across the lung. MPS I mice also exhibited an increase in lung compliance and reduced tissue elastance, reflecting diminished elastic recoil, as well as reduced airflow. Consistent with these findings, following lung collapse, MPS I lungs required a reduced opening pressure to inflate and all lung volumes, including TLC, VC, FRC, ERV and RV, were increased. Together, these results are consistent with the presence of obstructive lung disease with emphysema‐like structural and functional features, that is not only caused by narrowing of the upper airways, but likely also due to intrinsic pulmonary alterations involving the conducting airways and parenchymal component of the peripheral lung. Lung histology supported the presence of obstructive disease with isolated regions of disrupted alveolar tissue showing increased airspace. This integrated perspective may help explain the variability in respiratory symptoms observed in the MPS and guide the introduction of more effective therapeutic strategies.

MPS patients exhibit heterogeneous oxygenation that is attributed to sleep apnoea and upper airway obstruction [32]. X‐ray Velocimetry revealed a significant decrease in the whole‐lung mean specific ventilation in MPS I mice compared to controls. As the other XV metrics (VH and nVDP) were not significantly different across the whole lung, and since the Accuvent PIP and PEEP are fixed for all subjects, this suggests that the lung tissue expands less on average in the MPS I animals. For clarity, the regional analyses presented here are based on volumetric segmentation of the lung into inner (mediastinal‐adjacent) and outer regions rather than lobular anatomy, as X‐ray Velocimetry provides voxel‐based functional data that do not map directly onto individual lung lobes. When the XV data were split spatially into inner lung and outer lung regions, the MSV was significantly lower and the nVDP significantly higher in the inner, that is, mediastinal lung region of the MPS I animals. The ventilation histogram was shifted to the left, but not dramatically changed in shape (as VH was not different). For the outer lung region, there were no differences between the MPS I and control mouse XV parameters, indicating that MPS I mice have a small degree of regional variability in lung ventilation, specifically indicating a decrease in airflow in the inner (mediastinal‐adjacent) lung region. In a healthy lung, due to the presence of the heart and other mediastinal structures, the lung tends to expand less near the mediastinum compared to the outer lung. In MPS I patients, cardiac changes, including cardiac hypertrophy [33] may contribute to reduced tissue expansion in the inner lung region. These intrinsic lung alterations may also contribute to variable oxygenation and lung function. Regional imaging could inform targeted physiotherapy or pulmonary rehabilitation strategies for MPS I patients.

Despite the reduced tissue expansion observed by XV‐velocimetry, spirometry as measured by the single compartment model (SnapShot‐150 perturbation) demonstrated a higher inspiratory capacity combined with greater respiratory system compliance. This indicates that the MPS I lungs are more compliant than the controls, exhibiting a lower elasticity or elastic recoil. Simultaneously, MPS I mice also have greater conducting airway or chest wall resistance, but the single compartment model provides no information about where these changes originate. The changes in compliance and resistance were consistent with those previously observed in MPS IIIA mice [34]. Moreover, in that study, upon opening the chest wall the increase in resistance in MPS IIIA mice was exacerbated, specifically attributing this obstructive phenotype to the conducting airways as opposed to the chest wall [34]. Notably, skeletal involvement has traditionally been ascribed as the cause of restrictive respiratory pathology in the MPS [1], therefore, likely masking the obstructive phenotype of the distal lung.

To further resolve the complexity of respiratory pathology and to separate the contribution of central and peripheral airways, application of the constant phase model of respiratory mechanics in conjunction with the QuickPrime‐3 perturbation was used to produce deeper penetration into the lung than the single frequency SnapShot‐150 perturbation. This showed that MPS I mice have higher central conducting airway resistance and hence obstruction than the controls. This could be due to narrowing of the conducting airways, because of airway wall thickening, other changes in geometry, or airways that are more contorted. We did not note any dynamic airway collapse. The QuickPrime‐3 perturbation also demonstrated reduced tissue elastance (H), which matches the increased compliance detected by the single compartment model. Furthermore, the pressure volume loops were shifted to the left, with the MPS I mice having a higher static compliance than the controls, matching the compliance data from the other perturbations. These results are also consistent with the changes in compliance and elastance in MPS IIIA mice [34]. On the other hand, the tissue damping parameter G was not different in MPS I mice, suggesting that there was no difference in tissue resistance or resistance of the peripheral, smaller airways deep in the lung. However, in MPS IIIA mice, there was a reduction in tissue damping relative to control mice and this was exacerbated by opening the chest wall [34]. Given that MPS I disease is associated with a more pronounced musculoskeletal phenotype relative to MPS IIIA, subtle changes in the lung parenchyma may have been masked in the closed chest measurements undertaken in the present MPS I study. Together, these findings suggest that future studies should perform open chest measurements in MPS I mice to fully elucidate the relative contributions of parenchymal tissue, airways and the chest wall to respiratory pathology in the MPS.

In obstructive airway diseases such as emphysema, decreased tissue damping is attributed to disruption of alveolar architecture and altered parenchymal mechanics, which can influence resistance of small conducting airways through changes in alveolar tethering and airway–parenchymal interdependence [27, 35]. In the present study, histological analysis of MPS I lung tissue showed areas of normal lung structure as well as regions with enlarged air spaces. The presence and degree of heterogeneous parenchymal structure differed between individual MPS I animals, which is likely due to the variability of disease penetrance and may contribute to the patchy ventilation across the lung and variable respiratory clinical manifestations between patients.

Clinical application of spirometry requires patient compliance as it is an effort‐dependent test and hence is not well tolerated in a very unwell paediatric cohort. In preclinical studies, the flexiVent NPFE extension applies a large negative pressure to the mouse airways, so the test is not forced, and in some animal models of lung disease the large negative pressure can reduce flow rates if there is airway collapse. There were marked differences in the flow volume curves in the MPS I mice compared to the controls, clearly demonstrating that there are flow limitations in these mice, with fixed airway obstruction likely driven by the increased conducting airway resistance. We did not see a ‘scooped out’ pattern that is characteristic of an inspiratory flow limitation, but the FEV1/FVC ratio was decreased in MPS I mice compared to the controls. These changes further reinforce an obstructive respiratory phenotype, which is consistent with what we know about tracheal obstruction and tracheobronchomalacia in MPS I patients.

The lung volume perturbations revealed statistically significant differences in the lung volumes and capacities between the MPS I and control mice. Across the whole lung, MSV is equivalent to the ratio V_T:FRC, so a reduction in MSV in the MPS I mice could be driven by an increase in FRC (or RV), or a reduction in V_T. While there was no change in V_T in the MPS I mice, our lung volume data, which showed both an increase in FRC and RV in MPS I mice, suggests that it is these lung volumes that are driving the alterations in MSV. Both the flexiVent pressure volume loops and the lung volume analysis involving lung collapse, indicated increased compliance in the MPS I mice. Pathological increases in compliance, for example, in emphysema, can lead to air trapping and increased FRC, causing individuals to breathe at a higher lung volume and thereby reducing MSV in the central airways. Reduced elastance indicates impaired elastic recoil of the lung parenchyma, which may reduce radial traction on conducting airways and increase their susceptibility to expiratory narrowing and flow limitation, particularly in the presence of altered airway wall structure. Together, this obstructive disease in the upper airways also likely contributed to the development of the barrel‐shaped chest that we noted on the CT scans of the MPS I mice (Figure 3). Indeed, Hurler syndrome patients exhibit significant skeletal deformities (dysostosis multiplex), which include abnormalities in bone shape and structure [36].

In addition to changes in lung volumes and capacities, the lung volume perturbations also revealed differences in the opening pressure, P_op. Changes in lung P_op from the collapsed and degassed state in the lung volume manoeuvre can relate to changes in pulmonary surfactant function and/or to changes in lung ultrastructure. MPS I mice exhibited a significant decrease in P_op, which appears counterintuitive, as in emphysema patients an increase in airway opening pressure is often observed [37]. However, a similar reduction in P_op has been observed in elastase‐treated mice and in the spontaneously emphysematous LTC4S‐KO model [38], which was speculated to be due to altered pulmonary surfactant properties. Indeed, we have previously described a dysfunction in pulmonary surfactant in MPS IIIA mice [4]. A surfactant impairment may contribute to ventilation heterogeneity and reduced lung recoil and such a finding could open avenues for adjunctive surfactant‐targeted therapies or monitoring of surfactant biochemistry in patients.

In the current MPS I study, the lung function, volume and tissue histology all point to emphysema‐like structural and functional features with possible changes in pulmonary surfactant. The concomitant increases in RV, FRC and ERV, coupled with the decrease in P_op in MPS I mice, are each consistent with such a conclusion. Indeed, evidence of subpleural bullous changes, consistent with emphysema, has been reported in some adults with MPS I [39].

5. Limitations

In the present study, lung function was determined in MPS I mice with a closed chest only. Given the potential impact of thoracic cage structure on lung function, future studies should examine the separate contributions of the lung and chest wall to the disease pattern, in mice with an open chest. Indeed, opening the chest in MPS IIIA mice showed an exacerbated increase in resistance and an exacerbated reduction in damping, suggesting that changes in the composition or structure of the airways contributed to altered lung function and that this was partially masked by the musculoskeletal changes [34]. Lung parenchymal architecture is also a significant determinant of lung function. While there were no quantitative differences in alveolar structure of MPS IIIA mice [34], the histological analysis in the present study showed a significant increase in alveolar regions with enlarged airspaces in MPS I mice, supporting a finding of altered alveolar structure. However, quantifying structural changes in the parenchyma across the entire lung in three dimensions needs to be definitively confirmed using gold‐standard design‐based stereology. A final limitation is that three separate cohorts of animals were used in this study. It would have been advantageous to perform all tests with a single cohort to determine if there were any correlations between the metrics (lung function, lung volumes, XV and histology) that might have been caused by variability in disease severity. However, it was thought that over inflation of the lung and/or degassing of the lung to collapse it completely would compromise lung structure and skew the histology, so we chose to utilise separate cohorts. Moreover, at the time of performing the XV and lung mechanics assessments, the methodology to determine the lung volumes and capacities was not developed.

6. Conclusion

We have performed detailed analysis of lung function in a mouse model of MPS I using revolutionary non‐invasive in vivo XV lung function imaging in conjunction with gold‐standard flexiVent whole‐lung mechanics, including lung volume analysis; the first study of its kind. This was complemented with histological analysis of the lung. Consistent with the presence of obstructive lung disease in some MPS I patients, our findings in the MPS I mouse model reveal significant respiratory dysfunction, including reduced ventilation in the inner lung regions and spatially heterogeneous ventilation during a single breath. Our results suggest that respiratory impairment in MPS I is multifactorial, arising not only from upper airway obstruction and thoracic cavity changes, but also from intrinsic pulmonary abnormalities involving the conducting airways and the peripheral parenchyma. This highlights the importance of recognising lower airway and parenchymal contributions to respiratory disease in MPS I, with this intrinsic peripheral pathology likely contributing to the rapid and acute respiratory failure seen in some patients. Future human paediatric and adult XV clinical trials, similar to those being performed in children with cystic fibrosis [40], could reveal more about the clinical implications of MPS I lung pathology and aid the management of the disease [41].

Author Contributions

Martin Donnelley: conceptualisation, data curation, data interpretation, figure generation, formal analysis, funding acquisition, investigation, methodology, supervision, visualisation, writing – original draft. Ronan Smith: data curation, data interpretation, figure generation, formal analysis, investigation, methodology, software, visualisation, writing – original draft. Patricia Cmielewski, Kate Barratt, Jessica Logan, Nicole Reyne: investigation, writing – review and editing. Ben Ung: figure generation, formal analysis, writing – review and editing. Piraveen Pirakalathanan, Nina Eikelis, Kris Nilsen, Jennie Louise, Jessica Logan: data interpretation, formal analysis, writing – review and editing. David Parsons, Doug Brooks: conceptualisation, data interpretation, funding acquisition, methodology, writing – review and editing. Sandra Orgeig, Emma Parkinson‐Lawrence: conceptualisation, data interpretation, funding acquisition, investigation, project administration, methodology, supervision, writing – original draft.

Funding

This work was supported by the National MPS Society (1673), Medical Research Future Fund (RFRHPSI000013), Cystic Fibrosis Foundation (DONNEL21GO) and Australia's Economic Accelerator (Innovate grant IV240100090).

Disclosure

M.D. and D.P. have personally purchased shares in 4DMedical. N.E., K.N. and P.P. are employed by 4DMedical.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Summary of XV and flexiVent measurements. Data presented as estimated marginal means and 95% confidence intervals for the control and MPS I animals, plus the actual p value for each parameter tested (bolded for p < 0.05)

Figure S1: Scatter plot showing the relationship between flexiVent and XV measurements. The only correlation between XV and flexiVent parameters we found was a weak negative correlation between MSV and hysterisivity (G:H) calculated from the constant phase model.

Figure S2: Scatter plots showing the relationship between flexiVent and XV measurements in the inner lung regions. (A) Inner lung normalised ventilation defect percentage. (B) Inner lung mean specific ventilation.

Donnelley M., Smith R., Cmielewski P., et al., “Beyond Upper Airway Involvement: Evidence of Intrinsic Lung Disease in a Mouse Model of Mucopolysaccharidosis I,” Journal of Inherited Metabolic Disease 49, no. 4 (2026): e70219, 10.1002/jimd.70219.

Data Availability Statement

The data generated in this study is available at DOI: 10.25909/28192334.