© 2002 American Thoracic Society
Upper Airway Epithelial Structural Changes in Obstructive Sleep-disordered BreathingInstitute of Anatomy, Christian Albrecht University of Kiel, Kiel; Department of Otolaryngology, Section Rhinology and Rhonchopathology, University of Ulm, Ulm; Institute of Legal Medicine, University of Hamburg, Hamburg; and Department of Otolaryngology, Head and Neck Surgery, Friedrich Schiller University of Jena, Jena, Germany Correspondence and requests for reprints should be addressed to Dr. Friedrich P. Paulsen, Institute of Anatomy, Christian Albrecht University of Kiel, Olshausenstr. 40, D-24098 Kiel, Germany. E-mail: fpaulsen{at}anat.uni-kiel.de
The etiology of upper airway collapsibility in patients with snoring and obstructive sleep apnea (OSA) remains unclear. Structural mucosal changes could be contributory factors. The objective of this study was to determine whether pathologic changes in the epithelium or the epithelial–connective tissue interface are present in patients with snoring and/or OSA by means of scanning electron microscopy and immunohistochemistry. Uvulae were obtained by uvulopalatopharyngoplasty from three patients with habitual snoring and nine patients with mild to severe OSA, as well as by dissection from 43 nonsnoring body donors. Scanning electron microscopy revealed structural changes in the epithelial–connective tissue boundary that significantly differed from age-related changes in the control subjects. The immunohistochemical staining with antibodies against epithelial cytokeratins showed differences in the expression pattern of cytokeratin 13 between patients and control subjects. No differences were found in the distribution pattern of laminin. Analysis of defense cells revealed a significant diffuse infiltration of leukocytes, mainly T cells, inside the lamina propria of the patient group, which was not observed in the control group. In conclusion, these results support the hypothesis that progressive structural changes in the mucosa caused by the trauma of snoring are a possible contributory factor to upper airway collapsibility.
Key Words: obstructive sleep apnea • uvula • soft palate • snoring
In patients with obstructive sleep apnea (OSA), recurrent obstruction of the pharynx during sleep results in frequent episodes of airflow cessation, leading to significant hypoxemia, fragmentation of sleep, and cardiovascular complications (1–3). The estimated prevalence of OSA in the population aged between 30 and 60 years is about 2% in women and 4% in men (1). The pathogenesis of pharyngeal occlusion in this disorder has not been elucidated. So far, several theories emphasizing abnormalities in the control of breathing as well as structural and functional alterations in the pharynx have been proposed (4). The airway mucosa is known to play an important role of modulating lower airway patency (5), but its role in modulating upper airway aperture in patients with OSA is uncertain. A large body of experimental data indicates that the retropalatal region of the oropharynx is a major site of upper airway occlusion during sleep in patients with OSA (6–10). Recognition of the fact that pharyngeal structures may play a role in OSA has stimulated numerous studies of pharyngeal dimensions in patients with OSA, in both wakefulness and sleep, based on radiographic, endoscopic, acoustic, and other diagnostic techniques. Most studies suggest that the patient with OSA has a smaller pharyngeal airway or increased pharyngeal airflow resistance even when examined during wakefulness. However, the explanation for this reduced upper airway size is not clearly understood. It has been speculated that knowledge of the detailed microanatomy of pharyngeal tissue might yield insights into the pathogenesis of OSA. To test this speculation, several studies have been performed in recent years on pathophysiologic changes in the uvulae in patients with OSA undergoing uvulopalatopharyngoplasty (UPPP) (11–16). Sekosan and coworkers (4) demonstrated the presence of inflammation, characterized by plasma cell infiltration and interstitial edema. They concluded that the presence of inflammation in this area may conceivably increase the thickness of the mucosa, resulting in upper airway narrowing. Friberg and coworkers (17) described abnormal afferent nerve endings in the soft palatal mucosa of patients with sleep apnea and habitual snorers, suggesting this to be a contributory factor to the collapse of upper airways during sleep in OSA patients. Woodson and coworkers (13) showed that histologic differences occur in snorers and in patients with OSA compared with nonsnorers, visible as mucous gland hypertrophy, focal atrophy of muscle fibers, and extensive edema of the lamina propria. No study, however, has examined the epithelium and the epithelial–connective tissue boundary of the human oral mucosa in snorers or OSA patients. The epithelial–connective tissue boundary is a highly specialized interface formed by the epithelial cells and cells of the lamina propria, which are separated by a basement membrane (BM). One characteristic is the occurrence of connective tissue papillae (CTP), which are regarded as adaptive structures enlarging the epithelial–connective tissue interface to achieve a broader anchorage for the epithelium and to provide a larger exchange surface for nutritional purposes (18). Mechanical trauma can produce changes in epithelial surfaces under other conditions, such as formation of bullae under the influence of unphysiologic mechanical stress, resulting in detachment of the overlying epithelium from the underlying connective tissue (19, 20), and previous studies have proposed that snoring or apnea-related trauma may produce changes in upper airway tissues in these patients (4, 13, 17). In addition, in light of the fact that the epithelium and the epithelial–connective tissue interface have not been investigated previously in snorers or OSA patients, the purpose of the present study was to determine whether pathologic changes in the epithelium or the epithelial–connective tissue interface are present in patients with snoring and/or OSA.
Subjects Twelve men undergoing pharyngeal surgery were included in the present study. Of these patients, 10 were nonsmokers and 2 were cigarette smokers. In all patients, UPPP (including tonsillectomy) was performed as treatment for habitual snoring (with OSA) and various degrees of excessive daytime sleepiness. The patients had a median age of 50 years (range, 35–62 years) and a median body mass index of 29.3 kg/m2 (range, 22–38.1 kg/m2). Patients with gross abnormalities of the upper airway (i.e., severe tonsillar hyperplasia, micrognathia), or with previous treatment for their snoring/OSA, e.g., continuous positive airway pressure, dental devices, were excluded. To study normal age-related changes, tissue preparations were obtained from 43 selected cadavers. The control subjects had a median age of 50.3 years (range, 11–80 years) and a median body mass index of 24 kg/m2 (range, 16.3–28.2 kg/m2). The 43 cadavers were selected from a total of 86 cadavers whose history was studied and next of kin interviewed regarding snoring; any control subjects who had common colds during the last weeks of life or other diseases that could have affected pharyngeal function were excluded. Of the 86 cadavers, which were autopsy cases at the Institute of Legal Medicine, University Hospital Hamburg Eppendorf, Hamburg, Germany, or body donors at the Institute of Anatomy, Christian Albrecht University of Kiel, Germany, 12 cadavers were rejected due to a history of a malignant tumor, 4 cadavers were rejected on the basis of a history of intubation for life support, and an additional 27 cadavers were rejected because they did not conform to the selection criteria, i.e., because of a history of snoring or suspected apnea. Thus, all cadavers analyzed in the present study were nonsnorers and were free of recent oral trauma, oral or pharyngeal infections, or diseases potentially involving or affecting pharyngeal function. Apart from six selected cadavers using antihypertensive drugs, two using diuretic drugs, and one using a ß-blocker, none was on medication. Of all selected cadavers, 31 were nonsmokers, 10 were cigarette smokers, 1 was a cigar smoker, and 1 was a pipe smoker. For individual data on cadavers, see Table 1.
Respiratory Sleep Recordings All patients underwent full-night polysomnography. From these recordings, the oxygen desaturation index (ODI, average number of  4% desaturations per hour of sleep) and the lowest (nadir) arterial oxygen saturation were calculated. According to Svanborg and coworkers (21), the criterion for OSA is that ODI should be greater than 6. Recordings with ODI less than 2 are considered to be normal. Patients having ODI of 2 or more without fulfilling the criterion for OSA are in a borderline state and were classified as primary snorers. The nine OSA patients (nos. 4–12) had a median ODI of 35.8 (range, 15–63.4) and a median nadir of 68% (64–92%). The three patients (nos. 1–3) who were classified as primary snorers had a median ODI of 4.3 (range, 3–6) and a median nadir of 92.3% (range, 90–95%). The nonsnoring control subjects had not undergone any sleep recordings, but their history was thoroughly studied, and their next of kin were questioned by one of the authors (F.P.P. or P.S.) regarding any history of snoring, which they denied. For individual data of respiratory recordings, see Table 1.
Biopsy Procedures All tissue preparations of cadavers were resected by a technique comparable to UPPP from the soft palate and had sizes comparable to those of the surgically obtained specimens. The tissue was obtained latest by 24 hours after death and was fixated immediately in 4% formalin. After 1 week, all preparations from the patients and cadavers were split longitudinally into right and left halves. The right halves were embedded in paraffin using standard histologic techniques and were subjected to light microscopic and immunohistochemical procedures, whereas the left halves were prepared for scanning electron microscopy (SEM).
Light Microscopic Procedures
Scanning Electron Microscopic Procedures
Morphologic Analysis
Immunohistochemical Procedures
Immunohistochemical Grading of Subepithelial Leukocytes
Statistical Analysis
Morphology Light microscopy. The uvulae of the 43 control specimens were coated with a stratified, nonkeratinized squamous epithelium (Figure 1a) . The epithelium rested on a BM. The superficial layer of the lamina propria formed long, finger-like CTP. Capillaries supplying the CTP with blood arose from the capillary bed in the loose connective tissue of the lamina propria (inset, Figure 1a). Reticular fibers demonstrated by Gomori staining formed a loose network in the lamina propria connecting with the CTP and also with reticular fibers of seromucous glands and smooth musculature located in the center of the uvula (Figure 1a). The distal third of the uvula was free of glands and musculature. This structural arrangement did not change with age except for CTP, which seemed to be reduced in number and height.
The specimens of the entire group of patients exhibited morphologic abnormalities, including acanthosis of the overlying epidermis and a diffuse infiltration of leukocytes inside the lamina propria. Acanthosis was visible as a thickening of the stratum spinosum at irregular intervals along the uvula. It was more prominent in patients with OSA than in snorers. Moreover, there was a loss of CTP, up to a total absence (Figures 1b and 1c), in all patients.
Scanning electron microscopy.
The subepithelial CTP of the control group showed a diffuse distribution (Figure 2a). The shape of the CTP varied from rod-shaped (Figure 2b) to extreme windings (Figure 2a) and was not related to age. With increasing age, the height of the CTP decreased from 199–232 µm in the 10- to 20-year age group down to 34–63 µm in the 51 years-and-above age group (Figures 2a–2d). The density of CTP revealed a clear reduction from high density (++++) in the 10- to 20-year age group down to low density (+) in the 51 years-and-above age group (Figures 2a–2d). The uvulae of patients revealed clear differences between their papillary bodies. Here, areas without CTP (denuded areas) or with only a few CTP were visible (Figures 3a and 3b). These structural changes were already visible in patients below 50 years of age with OSA. In addition, in cases where CTP were present, their shape was different and height was clearly reduced in comparison with that of the CTP in the corresponding age group of asymptomatic individuals (Figures 3a and 3b). If CTP were present, their height varied between 20 and 55 µm. The density of CTP in the patient group varied between low density (+) and no CTP (-). The results of the morphologic classification between height and density of CTP in patients and control subjects are shown in Table 2. Mean values of height of CTP differed significantly (p < 0.001) when comparing the 35- to 62-year group of patients with the control subjects of the same age group (Figure 4) . Also, the density of CTP was found to be reduced in most of the patients to only a few CTP up to denuded areas without CTP. This feature was already visible in two patients with OSA aged 35 and 38 years.
There was neither a significant correlation between the density or height of CTP and ODI, nor between the density or height of CTP and the lowest (nadir) arterial oxygen saturation. To establish the interindividual variation of the scoring of CTP height and density, the results of the two morphologists were compared. They varied at the most with one category in patients (e.g., low density [+] versus no CTP [-]) in three of the specimens. The findings are given in Table 2.
Immunohistochemistry
Antibodies against CK4 and CK14 showed the same reactivity in both the control and the patient groups. Positive reactivity was observed in the epithelium of specimens from individuals aged 10–60 years (Figures 5b and 5c). The staining revealed expression of CK4 by the suprabasal cells (Figure 5b) and expression of CK14 by the basal cells (Figure 5c). CK4 and CK14 were not detected in the uvulae of individuals older than 60 years.
Laminin.
Leukocytes.
In our study, we used light and electron microscopy as well as immunohistochemistry to examine the epithelium and the epithelial–connective tissue boundary of patients with upper airway obstruction and uvulae from asymptomatic individuals. The main findings in this study were (1) reduction of CK13 expression only in the patient group; (2) acanthosis of the overlying epithelium only in the patient group; (3) a significant reduction, up to a total absence, of CTP in patients compared with the nonsnoring control subjects that revealed age-related structural changes of the CTP; and (4) a diffuse infiltration of leukocytes inside the lamina propria of the patient group, which was not observed in the control group. Vibration trauma related to severe snoring has been discussed as the cause of structural changes in the soft palate. Most studies dealing with such trauma address the musculature of the uvula and suggest that the musculus uvulae undergoes hypertrophy in patients with upper airway obstruction (11, 24) and/or atrophy with disruption of muscle bundles (13, 25, 26). Woodson and coworkers (13) also demonstrated mucous gland hypertrophy with ductal dilation and focal squamous metaplasia as well as extensive edema in the lamina propria with vascular dilation. Only two previous studies considered the mucosa and subepithelial connective tissue to be involved in the processes occurring in patients with upper airway obstruction (13, 27). We used antibodies against CK and laminin to study possible structural changes of the epithelium and the subepithelial lamina propria in cases of upper airway obstruction and during the natural aging process. CK are well-known markers for the analysis of epithelial changes (28). They are the intermediate filament proteins of epithelia. Twenty different CK—products of two gene families—are classified according to their molecular weights and isoelectric points (29). Although CK are present almost universally in epithelia, only a limited number are expressed in a given cell type, and the pattern of distribution of CK, as mapped by biochemical separation techniques, immunohistochemistry, and immunoblotting, has revealed a high degree of tissue specificity (30). In the present study, we used four CK related to stratified squamous epithelia: CK10 facilitates characterization of epidermal-type (cornifying) differentiation (31); CK4 and CK13 are suprabasal keratins characteristically expressed by noncornifying stratified squamous epithelia (32); and CK14 is present in the basal layer in mitotically active cells in both cornified and noncornified stratified epithelia (32). In the epithelium of the soft palate, the expression of CK4 and CK14 reveals a clear correlation to age. Both these CK are absent in the epithelia of asymptomatic individuals older than 60 years as well as in patients with upper airway obstruction older than 60 years, whereas the expression of CK13 seems to be reduced in the patient group. The absence of CK10 is not unexpected because the epithelium of the soft palate does not normally cornify. A difference in the CK expression pattern between asymptomatic individuals and patients with upper airway obstruction was observed only for CK13. When we decided to investigate the epithelium using antibodies against CK, we thought we might find increased epithelial stratification and proliferative index or an abnormal differentiation with maintenance of a basal phenotype because the primary function of CK is to protect epithelial cells from mechanical stress (33), and oral pathologies are known to be associated with modifications of intermediate filament organization (34, 35). Despite our expectations, this was not the case. The impact of the observed difference in CK13 expression between patients and control subjects thus remains obscure. However, in contrast to specimens from asymptomatic individuals, an acanthosis of the overlying epidermis of the uvula was observed in specimens from patients with upper airway obstruction. These structural changes seem to be associated with vibration trauma, as they show no correlation to the natural aging process as detected in the control group. Shteyer and coworkers (36) showed that long-standing chronic mechanical irritation applied to the tongues of rats is associated with fibrous hyperplasia, and in particular, epithelial changes including acanthosis and hyperkeratosis. A similar pathomechanism can be assumed for the occurrence of acanthosis in our patient group, although the mechanical irritation is of a somewhat different nature from the experimental irritation in the rat (36), there being no constant mechanical irritation of the epithelial surface in snorers and patients with OSA, but rather intermittent mechanical irritation due to vibration of the soft palate. This may be the reason that we indeed found acanthosis in our patient group but no signs of hyperkeratosis. On the other hand, it is astonishing that there were no marked changes in CK expression. To get further insights into the processes occurring in the mucosa of patients with sleep apnea and habitual snorers, we investigated the epithelial–connective tissue boundary of the soft palate of such patients and compared it with that of asymptomatic control subjects. The epithelial–connective tissue boundary of the human oral mucosa is a highly variable interface, which, owing to the uneven distribution of the projecting CTP, can be extremely irregular in some sites and rather smooth in others (37). Wentz and coworkers (38), Horstmann (37), Shklar (39), Löe and Karring (40), and Karring (41) described variations in connective tissue papillary density related to age and sex. The CTP are regarded as adaptive structures that enlarge the epithelial–connective tissue interface to achieve a broader anchorage for the epithelium and provide a larger exchange surface for nutritional purposes (18, 38, 42–44). Data on the hard palate and gingiva (38) as well as the soft palate (45) indicate that "keratinizing" epithelia are associated with a high papillary density in connective tissue, whereas the density of "nonkeratinizing" epithelia elsewhere in the mouth is much lower. Horstmann (45) argued that this relationship might favor a functional adaptation to external mechanical stimuli. Moreover, the typical papillary architecture, as well as the type of epithelium covering the papillary body, has been demonstrated to result from the action of unknown connective tissue inducers (46, 47). Our data show a clear correlation between age and height as well as density of CTP. The height and density of CTP decrease with increasing age, reaching a plateau in the fifth decade (Figure 2). These age-related changes in the connective tissue boundary of the uvula differ from structural changes induced by upper airway obstruction. Uvulae of such patients reveal an earlier decrease of the height and density of CTP up to a total loss of CTP. One can argue that the large numbers and height of CTP in young subjects could represent a postmortem artifact based on postmortem contraction of subepithelial connective tissue. However, large numbers and such CTP heights are also known from animal studies in which the tissue was obtained freshly (48). Therefore, the prominent structures at the epithelial–connective tissue boundary cannot be consistent with artifacts. With increasing age, CTP are more and more reduced. This may result from the action of unknown connective tissue regulators. Hypothetically, one explanation for the downregulation of such connective tissue inducers would be a reduction of mechanical pressure on the maxillary and mandibular bones if teeth are lost during life. In snorers and patients with OSA, the intermittent mechanical irritation could be a special stimulus for downregulation of connective tissue inducers resulting in an earlier reduction of CTP. Subsequently, the reduction or absence of CTP could be associated with the formation of acanthosis during the intermittently occurring mechanical irritation of the soft palate in snorers and patients with OSA. However, this hypothesis is contradicted by the findings of Hale (49) and Horstmann (37), who showed that the configuration of CTP is already established before birth. Moreover, transplantation experiments have shown that the features of the connective tissue do not change when the entire tissue is relocated to areas in which different mechanical conditions prevail (50, 51). A special role in the epithelial–connective tissue boundary is also played by the BM. Three major functions have been established for BM: (1) physical support, (2) cell attachment, and (3) filtration (52). The presence of BM can be revealed using an antibody against the BM component laminin. Just as in other epithelial areas of the oral mucosa, it was demonstrated that BM is a ubiquitous component of the extracellular matrix at the boundary between the epithelial cells and the connective tissue stroma. In contrast to the changes observed in CTP, we found no disruption or changes in the integrity of BM in patients with upper airway obstruction compared with control specimens. However, this does not exclude the possibility of changes in the BM that are detectable only at the electron microscopic level but do not manifest under immunohistochemical analysis. Friberg and coworkers (17) found an increased density of abnormal nerve endings in the soft palatal mucosa of patients with sleep apnea and habitual snorers, and hypothesized this to be a contributing factor to the collapse of upper airways during sleep in OSA patients. Sekosan and coworkers (4) showed edema and inflammation of the uvular mucosa and proposed that this contributed to upper airway occlusion in OSA by increasing the thickness of the oropharyngeal wall. Patients with such problems have to inspire against higher resistance during sleep, which leads to increased negative intraluminal pressure and upper airway collapse (6, 9). Analyzing the cell types occurring in the inflamed uvular mucosa of patients with OSA, Sekosan and coworkers (4) found higher numbers of lymphocytes as well as significant plasma cell infiltration inside the uvulae of patients with upper airway obstruction. Our findings clearly underline these data. Interestingly, we found T lymphocytes to be the main population of the lymphocytes present. As shown recently, T cells prefer to accumulate, irrespective of the particular disease, in skin lesions under chronic inflammatory conditions (53), where they can secrete several chemokines that may lead to cell activation with matrix degradation of musculature, glands, and the supplying nerves. This finding should be made the subject of further investigations. The structural changes observed in our study show no clear correlation to each other, i.e., inflammatory cell numbers were not related to any of the other changes, and inflammatory cells were not present to a similar extent in each area subepithelially. Only acanthosis was found in individuals in whom a reduction up to a total loss of CTP was also observed. The only plausible explanation available at the moment is that a mixed pattern of structural changes such as that found in our study could be based on the mechanical effects occurring during intermittent vibration trauma in tissues, such as that which occurs in snorers and OSA patients. However, little is known about such effects, and the matter should therefore be addressed in further investigations. In conclusion, our investigations reveal structural changes in the mucosa of the uvula and at the subepithelial–connective tissue boundary of patients with sleep apnea and habitual snorers; these changes differ clearly from age-related changes. It can be hypothesized that vibration trauma leads to activation of unknown connective tissue inducers, possibly initiated by immigration of T cells and plasma cells, leading to changes in the anchorage between the epithelium and the subepithelial connective tissue. These changes may contribute to airway instability, among other things, by loosening the connection between the epithelium and the subepithelial tissue, thus creating conditions favorable to uvular edema as described (4, 26). The actual etiology of these changes and the effect on the development of apnea, snoring, and airway properties requires further investigations. Such studies should elucidate the role of T cell subpopulations, chemokines, growth factors, and matrix metalloproteinases in progressive snorers disease. They are important because they might lead to the development of quantitative criteria in the diagnosis of snoring and sleep apnea by means of examination of the uvula and to the development of a model for a more general understanding of the effects of tissue changes in the oral mucosa.
The authors thank Mrs. Karin Stengel and Mrs. Regine Worm for their technical assistance, Mrs. Heidi Waluk and Mrs. Heide Siebke for their expert photographic work, and Mr. Michael Beall for correcting the English text. Received in original form September 25, 2001; accepted in final form May 27, 2002
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
-defensins 1–3 [1:800] (Bachem, Heidelberg, Germany), human CD3+ T cells [1:100] (DAKO), and human CD20+ B cells [concentrated] (DAKO). They were applied using a standard peroxidase-labeled streptavidin–biotin technique, either with microwave heating pretreatment or by using conventional methods with trypsinization where appropriate. Two negative control sections were used in each case; one was incubated with the secondary antibody only, the other with the primary antibody only. Sections of human skin and spleen were used as positive controls. All slides were examined under a Zeiss Axiophot microscope. 
