New Techniques for Biopsy and Culture of Human Olfactory Epithelial Neurons

ORIGINAL ARTICLE New Techniques for Biopsy and Culture of Human Olfactory Epithelial Neurons

Franc¸ois Fe´ron, PhD; Christopher Perry, MBBS; John J. McGrath, MBBS, PhD; Alan Mackay-Sim, PhD

Objective: To improve the success of culturing olfac- lium in a biopsy specimen ranged from 30% to 76%, de- tory neurons from human nasal mucosa by investigat- pending on its location. The dorsoposterior regions of ing the intranasal distribution of the olfactory epithe- the nasal septum and the superior turbinate provided the lium and devising new techniques for growing human highest probability, but, surprisingly, olfactory epithe- olfactory epithelium in vitro. lium was also found anteriorly and ventrally on both septum and turbinates. A new method of culturing the ol- Design: Ninety-sevenbiopsyspecimenswereobtainedfrom factory epithelium was devised. This slice culture 33 individuals, aged 21 to 74 years, collected from 6 regions technique improved the success rate for generating ol- of the nasal cavity. Each biopsy specimen was bisected, and factory neurons from 10% to 90%. 1 piece was processed for immunohistochemistry or electron microscopy while the other piece was dissected fur- Conclusions: This study explains and overcomes most ther for explant culture. Four culture techniques were per- of the variability in the success in observing neurogen- formed,includingwholeexplantsandexplantedbiopsyslices. esis in cultures of adult human olfactory epithelium. The Five days after plating, neuronal differentiation was induced techniques presented here make the human olfactory epi- by means of a medium that contained basic fibroblast growth thelium a useful model for clinical research into certain factor. After another 5 days, cultures were processed for im- olfactory dysfunctions and a model for the causes of neu- munocytochemical analysis. rodevelopmental and neurodegenerative diseases.

Results: The probability of finding olfactory epithe- Arch Otolaryngol Head Neck Surg. 1998;124:861-866

HE OLFACTORY epithelium is, trast, olfactory neurogenesis can be stud- in many ways, similar to the ied in living patients and controls. The neuroepithelium of the em- adult olfactory epithelium is therefore an bryonic neural tube. Within accessible source of neuronal tissue in adult olfactory epithelium which to observe the early stages of neu- areT stem cells that retain the capacity to di- rogenesis and neuronal differentiation and vide and give rise to neuronal precursors the factors that control and regulate these that further divide and differentiate into ma- processes.4,6 ture sensory neurons.1-3 Olfactory neuro- Although Murrell et al4 demon- genesis also occurs in humans: we have strated the potential for the study of neu- shown that human olfactory epithelium re- rogenesis in human patients, there re- tains the capacity for neurogenesis and neu- main problems with the efficacy of the From the School of ronal differentiation at least until the age technique as applied to living patients, Biomolecular and Biomedical of 72 years and that this can be demon- compared with postmortem studies. When Science, Griffith University, strated in tissue collected up to 25 hours nasal biopsy specimens are collected from Nathan, Australia (Drs Fe´ron post mortem.4 Thus, the human olfactory subjects, it is ethically important to col- and Mackay-Sim); Department epithelium has the potential to be used as lect as little tissue as necessary and to avoid of Otolaryngology, Princess a tool to examine certain human disor- rebiopsy. It is therefore vital to maximize Alexandra Hospital, South ders resulting from abnormal develop- the efficacy of the culture technique. Brisbane, Australia (Drs Perry ment of the nervous system. In our previous study, only approxi- and McGrath); and Queensland Neurogenesis has been shown to oc- mately 50% of explants produced viable Centre for Schizophrenia 5 Research, Wolston Park cur in the brain of the adult mouse, but cultures (cultures with at least an out- Hospital, Wacol, Australia this is not confirmed in humans. In any growth of epithelial cells), and of these, (Drs Fe´ron, McGrath, and event, neurogenesis in adult human brain only about 50% produced neurons after Mackay-Sim). could only be studied post mortem. In con- stimulation with basic fibroblast growth

ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 124, AUG 1998 861

©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 SUBJECTS AND METHODS methanol at − 20°C and processed for immunocytochemi- cal analysis. Alternatively, some explants were cultured by a dif- Human nasal mucosa was obtained by biopsy during rou- ferent technique. Instead of fibronectin-coated chamber tine nasal surgery with the patient under general anesthe- slides, some explants were grown on small pieces of fibro- sia, with the use of an ethmoid forceps (No. 8211.551; nectin-coated 0.4-µm nitrocellulose membrane (Milli- Richard Wolf Medical, Hoyland Medical, Camp Hill, Aus- pore, North Ryde, Australia) at the interface of air and me- tralia). Ninety-seven biopsy specimens were obtained from dium.10 The nitrocellulose membrane was placed on a 33 individuals (24 men and 9 women aged 21 to 74 years); transwell insert (Millipore) within a 6-well Costar plate of these, 10 had nasal abnormalities, such as polyps. The (Trace Biosciences, Tingalpa, Queensland, Australia). The patients were undergoing surgery for septoplasty or turbi- well was filled so that the medium just wetted the surface nectomy. All samples were obtained under a protocol ap- of the plastic membrane. With the methylene blue stain- proved by the ethics committees of the hospital and uni- ing used to orient the tissue, some explants were plated lu- versity involved. All biopsy tissues were obtained with the men side down, while others were plated lumen side up. informed consent of the patients, and the studies were car- These cultures were otherwise treated identically to those ried out in accordance with the guidelines of the National grown on chamber slides, except that, after fixation and Health and Medical Research Council of Australia. immunochemistry, the plastic membranes were mounted Six areas of collection were chosen: the dorsomedial and on slides before the coverslip was applied. dorsoposterior areas of the superior turbinate, the dorsoan- For electron microscopy, the tissue pieces were fixed terior and the ventroposterior areas of the middle turbi- for 2 hours in ice-cold fixative composed of 4% glutaral- nate, and the dorsomedial and the dorsoposterior areas of dehyde in 0.1-mol/L sodium cacodylate buffer (pH 7.4). the septum. The approximate location of each biopsy was Tissues were then washed in the same buffer, postfixed in noted and marked on standardized “maps” of the septum 1% osmium tetroxide, washed, dipped in uranyl acetate, and the medial face of the lateral wall of the nasal cavity. To and embedded in Spurr epoxy resin (ProSciTech, Queens- aid localization of specimens collected from the septum, ref- land). Ultrathin sections were cut, stained with uranyl ac- erence points to adjacent landmarks were noted. In some etate and lead citrate, and viewed on an electron micro- cases, the luminal surface of the nasal epithelium was stained scope. For immunhistochemistry, the tissue pieces were with methylene blue before collection to assist in orienta- fixed in Bouin fixative for 2 hours and washed in phosphate- tion of the excised pieces during explant culture. buffered saline (pH 7.4). The tissues were embedded in par- Biopsy specimens were placed on ice in Dulbecco modi- affin, sectioned at 4 µm, and laid on slides coated with 3- fied Eagle medium (Gibco-BRL, Gaithersburg, Md) supple- aminopropyltriethoxy-silane (Sigma, Castle Hill, New South mented with a serum supplement (Monomed, CSL, Mel- Wales, Australia). bourne, Australia), penicillin, streptomycin, gentamicin, and Immunochemistry was performed with peroxidase- nystatin (Gibco-BRL). After 1 to 2 hours the specimens were conjugated secondary antibodies (Sigma) with diamino- bisected: 1 piece was processed for immunohistochemis- benzidine as the chromogen. The following polyclonal or try or electron microscopy, and the other was dissected fur- monoclonal antibodies were used: anti–olfactory marker ther for explant culture. For culture, biopsy specimens were protein (goat polyclonal antibody, gift of Frank Margolis, cut into pieces of approximately 1 to 2 mm2 or, alterna- PhD, University of Maryland, Baltimore), anti–neuron- tively, sliced (200-µm thickness) by means of a McIlwain specific ␤-tubulin type III (mouse monoclonal antibody, tissue chopper (Mickle Laboratory Engineering Company Sigma), and anti–microtubule-associated protein 5 (mouse Ltd, Gomshall, Surrey, England) before being transferred monoclonal antibody, Sigma). Fixed cultures or deparaf- to fibronectin-coated glass chamber slides and cultured in finized sections were incubated overnight at 4°C followed the conditions above, as previously described.4,6 Five days by appropriate species-specific antibodies and peroxidase later the cultures were exposed to FGF2 (50 ng/mL; reagents. Extensive controls were used to exclude or de- Boehringer-Mannheim, South Brisbane, Queensland, termine the extent of nonspecific binding of secondary an- Australia). Another 5 days later, cultures were fixed in tibody and the presence of endogenous peroxidase.

factor (FGF2).4 Thus, the combined probability of ob- opsy location and culture method not be misinterpreted serving neurogenesis in an individual explant was only as an absence of neurogenesis. Therefore, before taking 25%. It is apparent that the proportion of cultures that specimens from patients with neurological disorders, it produce neurons must be significantly improved if ol- is imperative to identify and control the sources of vari- factory epithelium cultures are to be used to study neu- ability in explant growth and to raise the success rate as rogenesis in different patient populations. high as possible. A crude method for increasing the numbers of suc- There are 2 obvious reasons for the low success rate cessful explants would be simply to take more biopsy in producing neurons in our culture system: first, the specimens from each patient. It is routine in other labo- specimens may have contained no olfactory epithelium, ratories to take 4 to 6 specimens from each patient when and second, the explants often failed to attach to the sub- olfactory dysfunction is assessed.7,8 Lanza et al9 showed strate. The aim of the current study was to improve the that olfactory function is not affected after 2 to 5 biopsy rate of success by addressing these 2 issues. Biopsy speci- specimens from each individual. Nevertheless, for ethi- mens were taken from patients undergoing routine na- cal reasons, it is important to keep the numbers of speci- sal surgery. To address the first issue, different regions mens to a minimum. It is also important that an absence of the nasal cavity were surveyed for the rate of success of neurons because of technical problems related to bi- in finding olfactory epithelium assessed histologically and

ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 124, AUG 1998 862

©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 by in vitro explant culture. To address the second issue, a new culture method was developed. A

RESULTS

After collection in operating theater, biopsy specimens were identified by 3 techniques: immunohistochemical analysis, electron microscopy, and in vitro explant culture. To identify the olfactory epithelium inside the specimens, sections were immunostained with 3 neuron- specific antibodies: anti–neuron-specific ␤-tubulin type III, anti–microtubule-associated protein 5, and anti– olfactory marker protein, raised against a protein ex- pressed only by mature olfactory neurons. The olfactory mucosa was easily identified under the light B microscope by the presence of bipolar neurons in the epithelium and axon bundles in the lamina propria (Figure 1). Samples visualized with transmission electron microscopy were screened with the use of 2 main criteria: the presence of tall columnar supporting cells and the shape of cilia-bearing cells. As shown in Figure 2, olfactory dendrites terminate in expanding knobs bearing immotile cilia, while the cells arising of the respiratory epithelium bear cilia whose parallel orientation indicates they are beating synchronously. Biopsy cultures provided complementary informa- tion. Cilia of respiratory cells continued to beat in vitro and were easily identified; thus, we could estimate the proportion of respiratory and olfactory tissues on the C whole explant. Furthermore, using a medium containing FGF2,4 we tested the capability of explants to produce new neurons. Figure 3 shows neurons induced by FGF2 and immunolabeled with antitubulin antibody, 10 days after plating. When “pure” respiratory mucosa (judged by the abundance of beating ciliated cells) was plated, no neurogenesis was observed after FGF2 treatment. In total, 97 biopsy specimens were collected from 33 patients: 15 of these were not cultivated in vitro and only observed by either immunochemistry or electron microscopy. Of the others (82 specimens), 1 piece was plated in a culture dish and induced to differentiate, while the other was processed for either immunohistochemical Figure 1. Sections of human nasal mucosae immunostained with antibodies analysis or electron microscopy. After comparison of the raised against tubulin (A), microtubule-associated protein 5 (B), and olfactory marker protein (C). Arrows indicate bipolar neurons in the histological findings and cultures, the explants were clas- epithelium. In B, a dashed line shows a transition between olfactory and sified into 1 of the 4 following groups: pure olfactory mu- respiratory epithelia. AX indicates axon bundles. Bar in C indicates 10 µm cosa, pure respiratory mucosa, mixed tissue (olfactory (diaminobenzidine, original magnification ϫ100). and respiratory mucosae), and unidentified tissue. The results are summarized in Table 1. In the whole The probability of obtaining olfactory tissue was cal- population under study (33 patients), olfactory tissue culated with only biopsy specimens collected from nor- (alone or with respiratory mucosa) was found in one half mal individuals (71 specimens from 23 patients). of the biopsy specimens (53%). This percentage de- Figure 4 shows that olfactory tissue is more likely to creased to 34% in the subpopulation (10 patients) of be found further back in the nasal cavity on the septum people with nasal abnormalities and increased to 59% in as well as on the turbinates. However, on the posterior the healthy individuals (23 patients). When polyps were septum and on the posterior superior turbinate, pure ol- present, the ratio of pure olfactory mucosa decreased sub- factory mucosa was found in only 1 of 4 cases. In half of stantially while the percentage of unidentified tissue in- the cases, specimens contained interspersed patches of creased dramatically. Furthermore, pure respiratory mu- respiratory epithelium. A greater proportion of pure ol- cosa was found in the dorsoposterior area of the septum, factory mucosa (40%) was detected on the posterior area even in the younger patients. of the medial turbinate.

ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 124, AUG 1998 863

Figure 3. Slices of human olfactory mucosae cultured and immunostained with antitubulin antibody (diaminobenzidine, original magnification ϫ400). A, Neuronal cell body (arrowhead) at the periphery of the slice (dashed line) extends its neurite (arrows) across an epithelioid cell layer. B, Several interlaced neurons outside the epithelioid cell layer (bar indicates 10 µm).

Table 1. Tissues Obtained by Biopsy

No. (%) of Specimens

Patients With All Subjects Normal Subjects Nasal Disease Biopsy Tissue (N = 33) (n = 23) (n = 10) Olfactory 20 (21) 18 (25) 2 (7) Figure 2. Ultrastructural images of human nasal mucosae (osmium tetroxide Olfactory plus 31 (32) 24 (34) 7 (27) and uranyl acetate). A, Respiratory tissue can be recognized by its relatively respiratory flat surface and the parallel orientation of the cilia (bar indicates 2 µm) Respiratory 29 (30) 21 (30) 8 (31) (original magnification ϫ3000). B, Olfactory neurons can be recognized by Unidentified 17 (17) 8 (11) 9 (35) the presence of olfactory knobs and cilia with nonparallel orientation (bar indicates 1 µm) (original magnification ϫ6000). Total 97 (100) 71 (100) 26 (100)

There were no obvious age-related differences in our sample. In fact, examination of the specimens of the 7 gave us the opportunity to demonstrate that the attach- youngest healthy individuals (all younger than 30 years) ment success rate was dependent on the way the ex- allows us to assert that interindividual variations are prob- plant was plated. After the luminal side of the nasal tis- ably as important as age-dependent differences in the rela- sue was stained in the operating theater by means of tive distribution of respiratory and olfactory epithelium. methylene blue, the explant was seeded lumen side– Human biopsy specimens were first plated as pre- down (the epithelium in contact with the insert mem- viously described in the mouse,5 but only a third of them brane) or lumen side–up (the lamina propria in contact remained attached to the substrate (Table 2) whereas with the membrane). In the first case, the attachment ra- the attachment ratio was 90% in mouse experiments (un- tio was very poor, and almost no cellular outgrowth was published observations, July 1996). We attempted to avoid observed. In the second case, the attachment ratio was detachment by using a technique in which explants were 66%. Nevertheless, the neuronal outgrowth ratio re- not able to float in the medium: they were plated on a mained low (Table 2). nitrocellulose membrane and fed with an underlying me- Light microscope observations of the explants sug- dium. The use of large membranes (30-mm diameter) also gested that epithelial cells and neurons were contained

ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 124, AUG 1998 864

©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 mersed slices on nitrocellulose membranes), both can Septum be routinely used. However, culturing on chamber slides provides superior visualization compared with the membrane technique. 76% 40% (n = 17) (n = 5) COMMENT

This study explains some of the variability in the efficacy of observing neurogenesis in explants of human olfactory epithelium. Two key issues were addressed: the location of the nasal region from which the biopsy specimen was taken and the method of explant culture. It is evident that the success rate for observing olfactory neurogenesis in healthy persons can be increased by (1) taking biopsy specimens only from the dorsoposterior re- Lateral Wall Superior gions of the nasal septum and the surface of the superior Turbinate turbinate opposite and (2) culturing the epithelium in 30% 73% thin slices. The application of these methods should re- (n = 10) (n = 15) 50% duce the numbers of biopsy specimens necessary to dem- (n = 12) onstrate olfactory neurogenesis in each individual. Medial In accordance with the accepted location of the ol- 11 58% Turbinate factory epithelium, the highest probability of finding (n = 12) Inferior olfactory epithelium was in the dorsoposterior region. Ac- Turbinate cordingly, we recommend taking biopsy specimens on the most posterior areas of the septum and the superior turbinates only, in contrast to Strahan et al,7 who sug- gest taking 4 specimens from each patient, 1 anterior and 1 posterior from each side of the nasal septum. Our observations confirm previous reports that the Figure 4. Percentage of olfactory tissue in the 71 biopsy specimens collected from 23 healthy individuals. Six areas were screened. The number distribution of the olfactory epithelium in adult hu- of samples containing olfactory epithelium was divided by the number of mans is frequently disrupted with interspersed patches samples collected in each area (n). of respiratory epithelium.12-14 In contrast, the olfactory epithelium in the human fetus has been reported to be uniformly distributed without interruption by respira- Table 2. Comparison of Culture Methods* tory epithelium, extending from the roof of the nasal cavity onto the superior turbinate and to the middle of the Explants Explants Slices in Slices septum in a continuous pattern.13 in Chamber on Chamber on Even in the dorsoposterior regions of the adult na- Slide Insert Slide Insert sal cavity, where the probability of finding olfactory epi- Attachment 33 66 75 80 thelium was highest (Figure 4), only 40% of the speci- ratio, % mens contained olfactory epithelium exclusively. Neuronal 10 33 90 85 outgrowth Considering the small surface area within each biopsy 2 ratio, % specimen (1 mm ), it is apparent that the olfactory and respiratory tissues are intimately dispersed in the adult *Attachment ratio is the number of attached nasal mucosae per number of nasal cavity. Comparison of fetal and adult tissues sug- plated nasal mucosae; the neuronal outgrowth ratio is the number of cultures gests that invasion of respiratory tissue into olfactory epi- with neuronal outgrowth per number of attached olfactory mucosae. thelium increases with age, as suggested in a previous study of adult olfactory epithelium.8 within the explant but were unable to migrate from it. The observations of this study raise issues related The human lamina propria is thicker than that in the ro- to the etiology of the distribution of olfactory tissue. It dent, especially on the turbinates, and the human olfac- is often assumed that the appearance of respiratory epi- tory epithelium is thinner. For this reason, outgrowing thelium within olfactory epithelium represents an “in- cells may have been unable to reach the substrate and vasion” of the former into the latter, or a replacement of were prevented from proliferating and differentiating on the latter by the former after damage to the olfactory epi- the membrane, away from the explant. thelium by environmental insults, such as viruses and To provide direct contact of the epithelium with other diseases or head trauma.15,16 However, the wide dis- the substrate, biopsy specimens were sliced (200 µm) tribution of the olfactory epithelium within the nasal cav- and cultured either in chamber slides or on inserts.10 ity of the adult also raises the possibility that the olfac- In both cases, attachment and neuronal outgrowth tory epithelium may invade, replace, or migrate into ratios increased dramatically. Because no significant regions of respiratory epithelium. This issue is unlikely difference was observed between the 2 culture tech- to be resolved without a systematic mapping of the ol- niques (immersed slices in chamber slides or nonim- factory epithelium throughout the nasal cavity in hu-

ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 124, AUG 1998 865

©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 mans of different ages, particularly of the ages between Reprints: Alan Mackay-Sim, PhD, School of Biomolecu- birth and adulthood, for which there are no data at all. lar and Biomedical Science, Griffith University, Nathan, QLD It is now possible to efficiently cultivate human ol- 4111, Australia (e-mail: [email protected]). factory epithelium. The optimal technique (slice culture, Table 2) differs significantly in efficacy from our previ- REFERENCES ous method (whole explant culture, Table 2)4 by improv-

ing the success rate for attachment of the tissue to the sub- 1. Nagahara Y. Experimentelle Studien uber die histologichen Veranderungen strate and by improving the success rate for the outgrowth des Geruchorgans nach der Olfactoriusdurchschneidung: Beitrage zur Kennt- of neurons. For attachment, the important factor appears nis des Feineren baus des Geruchorgans. Jpn J Med Sci V Pathol. 1940;5: to be that the lamina propria is in contact with the sub- 165-199. strate (here the fibronectin-coated dish or membrane in- 2. Moulton DG, Fink RP. Cell proliferation and migration in the olfactory epithelium. In: Schneider D, ed. Olfaction and Taste IV. Stuttgart, Germany: Wissen- stead of the underlying cartilage). For neuronal out- schaftliche Verlagsgesselschaft; 1972:20-26. growth, the factor appears to be that the migrating neurons 3. Graziadei PPC. Cell dynamics in the olfactory mucosa. Tissue Cell. 1973;5: are able to contact the substrate. Both of these conditions 113-131. are satisfied in the slice preparation. These results differ 4. Murrell W, Bushell GR, Livesey J, et al. Neurogenesis in adult human. Neurore- port. 1996;7:1189-1194. from our observations in the mouse, for which high suc- 5. Weiss S, Reynolds BA, Vescovi AL, Morshead C, Craig CG, Van der Kooy D. Is cess rates for attachment and neuronal outgrowth can be there a neural stem cell in the mammalian forebrain? Trends Neurosci. 1996; achieved when whole explants are cultured, without slic- 19:387-393. ing them. We ascribe these species differences in efficacy 6. MacDonald KPA, Murrell WG, Bartlett P, Bushell GR, Mackay-Sim A. FGF2 pro- to differences in the thickness of the lamina propria. The motes neuronal differentiation in explant cultures of adult and embryonic mouse olfactory epithelium. J Neurosci Res. 1996;44:27-39. human lamina propria is thicker than in mouse and may 7. Strahan RC, Jafek BW, Moran DT. Biopsy of the olfactory neuroepithelium. In: provide a bigger barrier to neuronal migration from the Getchell TV, Doty RL, Bartoshuk LM, Snow JB, eds. Smell and Taste in Health epithelium to the substrate. and Disease. New York, NY: Raven Press; 1991:703-709. The success rates of the biopsy and culture tech- 8. Paik SI, Lehman MN, Seiden AM, Duncan HJ, Smith DV. Human olfactory biopsy: the influence of age and receptor distribution. Arch Otolaryngol Head Neck niques presented here make them particularly useful for Surg. 1992;118:731-738. further research on olfactory disorders as well as com- 9. Lanza DC, Deems DA, Doty RL, et al. The effect of human olfactory biopsy on parative studies on putative neurodevelopmental dis- olfaction: a preliminary report. Laryngoscope. 1994;104:837-840. eases (eg, schizophrenia) or neurodegenerative diseases 10. Stoppini L, Buchs PA, Muller D. A simple method for organotypic cultures of (eg, Alzheimer or Parkinson disease). nervous tissue. J Neurosci Methods. 1991;37:173-182. 11. Nauta WJH, Feirtag M. Fundamental Neuroanatomy. New York, NY: Freeman Press; 1986. Accepted for publication March 4, 1998. 12. Naessen R. The identification and topographical localisation of the olfactory epi- This work was supported by the Garnett Passe and Rod- thelium in man and other mammals. Acta Otolaryngol. 1970;70:51-57. 13. Nakashima T, Kimmelman CP, Snow JB Jr. Structure of human fetal and adult ney Williams Memorial Foundation, Melbourne, Austra- olfactory neuroepithelium. Arch Otolaryngol. 1984;110:641-646. lia; the Rebecca L. Cooper Foundation, Sydney, Australia; 14. Morrison EE, Costanzo RM. Morphology of the human olfactory epithelium. the Theodore and Vera Stanley Foundation, Arlington, Va; J Comp Neurol. 1990;297:1-13. and Queensland Health, Brisbane, Australia. 15. Moran DT, Jafek BW, Eller PM, Rowley JC III. Ultrastructural histopathology of human olfactory dysfunction. Microsc Res Tech. 1992;23:103-110. We thank Wen Hu, PhD, and Joseph Kan for their skill- 16. Talamo BR, Feng WH, Stockmayer M. Human olfactory epithelium: normal pat- ful assistance and Frank Margolis, PhD, for his generous terns and types of lesions found in the general population. Inhalation Toxicol. gift of the antibody to olfactory marker protein. 1994;6:249-275.

ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 124, AUG 1998 866