Epithelial Stress and Apoptosis Underlie Hermansky-Pudlak Syndrome–associated Interstitial Pneumonia

Mice

Breeding pairs of HPS monomutant mice and HPS1/6 double-mutant mice were purchased from Jackson Laboratory (Bar Harbor, ME). Breeding pairs of HPS1/2 double-mutant mice were a kind gift from R. Swank (Roswell Park, Buffalo, NY). C57BL/6J was the background strain for the HPS1, HPS2, HPS6, and HPS1/2 mice. HPS1/6 mice were on the B6C3Fe background. All mice were mated and maintained under specific pathogen–free conditions, and five mice per group were killed at the ages of 3 and 9 months and genotyped as described in the online supplement.

Quantification of Lung Collagen

Hydroxyproline levels in murine lungs were determined as described elsewhere (16) and in the online supplement.

Computerized Lung Morphometry

Morphometric analysis was performed on hematoxylin and eosin–stained lung tissue sections by determining the mean interalveolar distance and mean septal thickness (linear intercept measurement). AxioVision software (Carl Zeiss MicroImaging GmbH, Jena Germany) was used for the analysis as outlined in the online supplement.

Isolation of Murine Alveolar Epithelial Type II Cells

Murine AECII from the lungs of 3-month-old HPS1/2 and wild-type (WT) control mice were isolated according to the protocol given in the online supplement.

Cloning and Cell Culture

Full-length murine cathepsin D was cloned into the pcDNA3.1 expression vector and transfected into mouse lung epithelial (MLE)-12 cells, using primers and protocols described in the online supplement. NIH 3T3 mouse fibroblasts were then incubated with conditioned medium from either cathepsin D– or empty vector–transfected MLE-12 cells (24 h posttransfection). WST-1:4-[3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]-1,3-benzene disulphonate or BrdU: 5-bromo-2′-deoxyuridine (Roche Applied Science, Mannheim, Germany) was added as described in the online supplement and absorbance was measured at 450 nm.

Western Blot

Samples were subjected to denaturing sodium dodecyl sulfate–polyacrylamide gel electrophoresis followed by electroblotting on polyvinylidene difluoride membrane and immunostaining for the desired proteins. Protocol and the source of antibodies are given in the online supplement.

Phospholipid Analysis, Isolation of Large Surfactant Aggregates, and Assessment of Surface Activity

Phospholipids (PLs) were extracted from bronchoalveolar lavage fluid (BALF) or tissue extracts according to the Bligh and Dyer method as referred to in the online supplement. Characterization and surface activity of large surfactant aggregates (LAs) was undertaken as outlined in the online supplement.

Lipidomics

Lipids were quantified by electrospray ionization–tandem mass spectrometry in positive ion mode, as described in the online supplement.

Immunohistochemistry and in Situ Apoptosis Assay

ZytoChem Plus (AP) Broad Spectrum (Fast Red) kit (ZytoMed systems, Berlin, Germany) was used with paraffin-embedded lung sections for immunohistochemical localization of proteins according to the manufacturer's instructions. The protocol and source of antibodies are given in the online supplement. The degree of cellular apoptosis was determined by the terminal deoxynucleotidyltransferase dUTP nick end labeling (TUNEL) method, using the in situ cell death detection kit, AP (Roche Applied Science) according to the manufacturer's instructions. All stained sections were analyzed on digital slide scanning, employing a Mirax scanner (Carl Zeiss GmbH, Jena, Germany). Mirax Viewer software (Carl Zeiss GmbH) was used to analyze and make snapshots of the stained tissue sections.

Statistics

Unless indicated otherwise, all data are expressed as means ± SEM of at least five mice. Statistical evaluation was performed by Mann-Whitney U test (comparison of two groups) and by Kruskal-Wallis H test (if more than two groups were analyzed). The level of significance is indicated in the figure legends and footnotes as follows: *^/#P < 0.5, **^/##P < 0.01, ***^/###P < 0.001.

Development of Pulmonary Fibrosis in HPS1/2

In HPS1/2 double-mutant mice, but not in HPS1/6 or HPS1, HPS2, or HPS6 mice, spontaneous development of lung fibrosis was evident at the age of 3 months (Figure 1a, left). Being patchy and subpleural at the beginning, the extent of lung fibrosis progressed over the age of 9 months (Figure 1a, right). Although fibrosis was encountered in all the 9-month-old HPS1/2 mice analyzed in this study, the extent of fibrosis differed somewhat from mouse to mouse. At higher magnification, lymphoplasmacellular infiltration and extracellular matrix deposition, with a dramatic increase in AECII size but roughly preserved lung structure, were evident at 3 months (Figure 1b). In contrast, dense fibrosis with extensive remodeling of the alveolar lung structure was prominent in 9-month-old HPS1/2 mice (Figure 1b, second row), because of which quantitative lung morphometric analysis did not reveal any significant increase in enlarged airspaces in HPS1/2 mice (Figure 2a) but revealed a more than twofold increase in septal thickness, when compared with WT controls (Figures 2c and 2e; and see Figure E1 in the online supplement). As a result of such fibrotic changes, lung hydroxyproline content was significantly elevated (Figure 1c). In HPS1/2 mice at 3 months, HPS1 and HPS2 monomutant mice at 9 months, and even more in HPS1/6 double-mutant mice at 9 months, airspace enlargement was a prominent finding (Figure 1b; and Figures 2b, 2d, and 2f).

Disturbed Composition and Function of the Alveolar Surfactant Pool in HPS1/2 Mice

As compared with WT mice, mature surfactant protein (SP)-B was undetectable in BALF from 3- and 9-month-old HPS1/2 mice and appeared greatly reduced in HPS1/6 mice of the same age. In contrast, BALF levels of mature SP-B remained unchanged in any of the other monomutant HPS mice (Figure 3a). Similarly, significantly reduced levels of mature SP-C were observed in the BALF of HPS1/2 mice at 3 and 9 months, but appeared unchanged in all other HPS mutants (Figure 3b). In addition, the total phospholipid content, but not the relative content of large surfactant aggregates (LAs), was significantly decreased in BALF from HPS1/2 mice (Figure 3c). As compared with WT controls, LA preparations from HPS1/2 BALF, but not from HPS1/6 BALF, showed a significant reduction in surface activity (Figures 3d and 3e). From these data, we concluded that HPS1/2 mice do have a significant defect in surfactant secretion, resulting in inappropriately low alveolar phospholipid and surfactant protein concentrations, thus leading to increased alveolar surface tension.

In addition, BALF cellular profiles from various HPS mice were analyzed. Of interest, many enlarged alveolar macrophages were encountered (data not shown), especially in HPS1/2 mice. In addition, a modest lymphocytic alveolitis was observed in all of the HPS mice analyzed (Table 1), and there was no significant difference between HPS1/2 and the other mice.

Accumulation of Hydrophobic Surfactant Proteins in HPS1/2 Lungs

A dramatic increase in intracellular content of almost all forms of the pro- and mature forms of the hydrophobic surfactant proteins SP-B and SP-C was encountered in lungs of HPS1/2 mice in comparison with their respective WT controls (Figure 4a). At the age of 3 months, especially 21-kD pro–SP-C, mature SP-B, and mature SP-C were elevated. In 9-month-old HPS1/2 mice, a more extensive accumulation of all pro- and mature forms of SP-B/C, with the exception of 42-kD pro–SP-B, was encountered (Figure 4a). Densitometric analysis revealed a 25-fold increase in mature SP-C and a 5-fold increase in mature SP-B in HPS1/2 lung homogenates (Figures 4g and 4h). To confirm that such accumulation would reflect the situation within the AECII, the intracellular SP-B/C content was analyzed in AECII isolated from 3-month-old HPS1/2 mice and WT mice. Accumulation of mature SP-B and SP-C was seen exclusively in the AECII of HPS1/2 mice (Figure 4b), but not in any other HPS mutants (Figure E2). Regarding the proforms, only 19-kD pro–SP-B appeared highly up-regulated in all HPS mice (although significantly higher in HPS1/2), and some minor increases were encountered for 21-kD pro–SP-C and 25-kD pro–SP-B in these mice (Figures 4c–4f; and see Figure E2). Gene expression analysis did not show any differences in the expression of SP-B and SP-C between HPS1/2 and control mice (Figure E3a). Consistent with previous reports on the subcellular distribution of pro- and mature forms of the hydrophobic surfactant proteins in healthy AECII (17, 18) and with the ultrastructural appearance of AECII in humans with HPSIP (10), mature surfactant compounds seem to accumulate in the distal part of the lysosomal transport axis of AECII from HPS1/2, and hence largely in lamellar bodies.

Accumulation of Surfactant Phospholipids in HPS1/2 Lungs

We next analyzed the total phospholipid content of lung homogenates in all HPS mono- and double-mutant mice along with their WT controls. Phospholipid content was increased in all HPS mice, but to various extents. The strongest increase was found in HPS1/2 mice (approximately threefold), which was also significantly higher as compared with all other HPS mice (Figure 5a). It seems noteworthy to mention that, in contrast to the changes in mature SP-B/C, the phospholipid content of HPS1/2 lung homogenates did not further increase beyond the age of 3 months. In AECII, the combinatory loss of HPS1 and HPS2 genes thus does not seem to affect early lysosomal transport processes, as suggested for HPS1 or HPS2 gene products in other cell types (19, 20).

Lipidomic Profiling in HPS Mice

Because we observed differences in total phospholipids in HPS mice, we further performed mass spectrometry–based lipidomics analysis of lung tissues from all 9-month-old HPS and WT control mice. As evident from Table E1 in the online supplement and Figure 5a, extensive alterations of the phospholipid profile were encountered. In particular, HPS1/2 mice showed a significant increase in phosphatidylcholine (PC). To some extent this also held true for HPS1/6 mice. Of note, phosphatidylethanolamine plasmalogens were markedly depressed in HPS1/2 versus the other mice (Table E1). Moreover, a significant increase in saturated PC species was encountered in HPS1/2 mice and, to some extent, also in HPS1/6 mice (Figure 5b), which could be ascribed almost entirely to a marked increase in dipalmitoylated phosphatidylcholine (DPPC, PC-32:0; Figure 5c). Another interesting observation was a significant increase in glucosylceramides (GlcCer) in 9-month-old HPS1/2 mice, with a ceramide/GlcCer ratio of 3.9 in HPS1/2 mice compared with 7.6 in controls (see also Table E2). Because the expression level of the synthesizing enzyme (glucosylceramide synthase) remained unchanged (data not shown), the observed accumulation of GlcCer in HPS1/2 mice might be due to their decreased clearance.

AECII Apoptosis in HPS1/2 Mice

Accumulation and impaired secretion of pulmonary surfactant may have resulted in chronic cell stress and apoptosis of AECII in lungs of HPS1/2 mice. To identify apoptotic AECII, we performed immunohistochemistry for cleaved, activated caspase-3, along with pro–SP-C on serial sections of HPS1/2 and control mouse lungs. It was found that the pro–SP-C–positive cells (AECII) also stained positive for cleaved caspase-3 in HPS1/2 tissue sections, whereas no signal for cleaved caspase-3 was detected in any WT control (Figure 6a). Western blot analysis for caspase-3 produced a similar result (Figure 6b). Alternatively, we performed TUNEL staining, along with immunohistochemistry, for the AECII-specific marker pro–SP-C on serial sections from all HPS and control mice. As depicted in Figure E4a in the online supplement, numerous TUNEL-positive AECII were already evident only in the lung tissue of HPS1/2 mice at an age of 3 months (but not in other mutants), indicating that AECII apoptosis is an early event in these mice. Thus, TUNEL data fully confirmed the cleaved caspase-3 immunohistochemistry and again suggest that AECII undergo extensive apoptosis in HPS1/2 mice, but not in other HPS or control mice.

Lysosomal and Endoplasmic Reticulum Stress Underlie AECII Apoptosis in HPS1/2 Mice

Observations to this point indicated that intracellular accumulation of surfactant was especially prominent in HPS1/2 mice and possibly related to the increased apoptosis of AECII, thus favoring fibrosis development in HPS1/2 double-mutants. To further study the pathways that may interconnect surfactant accumulation and AECII apoptosis in HPS1/2 mice, it was hypothesized that such surfactant accumulation may cause primarily lysosomal stress in these mice. Therefore, a well-known lysosomal aspartyl protease, cathepsin D, which has been shown to induce apoptosis by several mechanisms (21, 22), was analyzed by Western blotting. Increased levels of procathepsin D (∼44 kD), its glycosylated form (∼54 kD), and subsequent cleavage products were observed in lung homogenates of 3- and 9-month-old HPS1/2 mice and in the AECII of 3-month-old HPS1/2 mice when compared with respective controls (Figures 6c and 6d). Interestingly, cathepsin D levels in lung homogenates of all other HPS mice were found to be virtually identical (see Figures E5a–E5c in the online supplement). Cathepsin D gene expression in HPS1/2 mice, however, did not differ from controls (Figure E5d). In colocalization studies employing immunohistochemistry for cathepsin D, cleaved caspase-3, and pro–SP-C on serial sections, we identified all three proteins on identical regions in HPS1/2 lungs (Figure 6e) with a pan-cytoplasmic distribution pattern for cathepsin D (inset in Figure 6e).

Apart from the lysosomal stress response, the possible existence of an endoplasmic reticulum (ER) stress response was also assessed in HPS mice. For this purpose, the late ER stress marker C/EBP homologous protein (CHOP; GADD153), which represents an obligatory and important proapoptotic ER stress compound, as well as activating transcription factor-4 (ATF4), which serves as a direct transcription factor for CHOP, were studied. Western blot analysis using lung homogenates revealed an up-regulation of both proteins in HPS1/2 mice, at the age of 9 months (Figure 6f) but not at an age of 3 months (data not shown). To check whether the ATF4 and CHOP signals of lung homogenates originate from AECII, immunohistochemistry of ATF4, CHOP, and pro–SP-C was performed on serial sections. Interestingly, these stainings illustrated that pro–SP-C–positive cells stained positive for ATF4 and CHOP in lung sections from 9-month-old HPS1/2 mice, whereas control mouse sections showed no positivity (Figure 6g), thus indicating that AECII from HPS1/2 mice undergo ER stress during a later stage of the disease.

AECII Apoptosis Due to Lysosomal and ER Stress Is Also a Prominent Finding in Human HPSIP

Lysosomal and ER stress pathways were characterized in human HPSIP. Lack of frozen lung material from transplanted human patients with HPS restricted the current work to available serial lung sections from paraffin-embedded lung tissues from only two patients with HPS, on which immunohistochemistry was performed. Interestingly, AECII from patients with HPSIP not only showed much more pronounced reactivity for pro–SP-C as compared with donor lungs, but also showed positive staining for both cleaved caspase-3 and cathepsin D (Figure 7a), thus fully confirming the data observed in the elder and fibrotic HPS1/2 mice. Diffuse cytoplasmic staining for cathepsin D within AECII was also observed in lung sections of HPSIP (inset in Figure 7a). Of interest, in healthy lung tissues from mice (Figure 6e) and humans (Figure 7a) cathepsin D staining was observed primarily in alveolar macrophages but not in AECII, again underscoring the significance of the observed increase in cathepsin D in AECII under conditions of HPSIP. In addition, we investigated the expression of the proapoptotic ER stress factor CHOP and its transcription factor ATF4 in human HPSIP. As depicted in Figure 7b, immunoreactivity for CHOP and ATF4 was found in AECII of HPSIP lungs, but not of donor lungs, again providing evidence for a proapoptotic ER stress response in the AECII of these patients.

Cathepsin D Promotes Apoptosis of Alveolar Epithelial Cell Line and Fibroblast Proliferation

In addition, to assess the causative role of cathepsin D overexpression in alveolar epithelial cell apoptosis and its role in fibroblast proliferation, we overexpressed full-length mouse cathepsin D in AECII-like MLE-12 cells. By analysis of cleaved caspase-3, we observed significant cell death in cathepsin D–transfected cells, as compared with empty vector–transfected cells (Figure 8a). These cells were largely detached from the culture plate, 24 hours posttransfection. Conditioned medium from cathepsin D–transfected cells was then applied on NIH 3T3 fibroblasts and this resulted in increased viability and proliferation, as measured by WST-1 assay (Figure 8b) and BrdU incorporation (Figure 8c), respectively. Incubation of NIH 3T3 fibroblasts with conditioned medium from empty vector–transfected cells resulted in a significantly depressed proliferative signal.