Association of ecto-5?-nucleotidase with specific cell types in the adult and developing rat olfactory organ

Association of Ecto-58-Nucleotidase WithSpecific Cell Types in the Adult and

Developing Rat Olfactory Organ

NORBERT BRAUN AND HERBERT ZIMMERMANN*Biozentrum der J.W. Goethe-Universitat, AK Neurochemie,

D-60439 Frankfurt am Main, Germany

ABSTRACTA unique feature of the olfactory epithelium is its ability to give rise to new sensory

neurons throughout life and also following injury. Cells at the basal side of the epitheliumserve as neurogenic progenitor cells. The enzyme ecto-58-nucleotidase is expressed at thesurface of developing nerve cells and is regarded as a marker of neural development. To studythe expression pattern of the enzyme, we analyzed its distribution in the adult and developingrat olfactory organ. Labeling is restricted to specific cell types and varies between the epitheliainvestigated. At the basal side of the olfactory epithelium, activity of 58-nucleotidase isassociated specifically with the dark/horizontal basal cells. Neither the light/globose basalcells, which are the immediate precursors of the sensory receptor cells, nor subsets ofpotentially immature olfactory receptor cells are labeled. On the other hand, microvillar cellsdispersed at the lumenal side of the epithelium contain 58-nucleotidase activity. The enzymeis also present at the inner lining of the ducts of Bowman’s glands as they traverse theepithelium. Within the respiratory epithelium, activity of 58-nucleotidase is associated withbasal cells as well as with the epithelial surface. During development, 58-nucleotidase isinitially limited to the respiratory epithelium, including its basal cells. Dark/horizontal basalcells of the olfactory epithelium, which are positive for 58-nucleotidase, first appear at theborder of the respiratory epithelium, suggesting that they might originate from immigratingbasal cells of the respiratory epithelium. Within the vomeronasal organ, labeling is largelyrestricted to the receptor-free epithelium. Although the functional role of 58-nucleotidasein the olfactory system needs to be further defined, the distribution of the enzyme can beused successfully as a marker for defined cell types. J. Comp. Neurol. 393:528–537, 1998.r 1998 Wiley-Liss, Inc.

Indexing terms: basal cells; development; microvillar cells; olfactory epithelium; respiratory

epithelium; vomeronasal organ

Rodents have a dual olfactory system with two differenttypes of sensory epithelia. One is represented by the regioolfactoria, the sensory epithelium of the main olfactorysystem. It is located in distinct regions of the nasal septumand the nasal turbinates. The other type is the neuroepithe-lium of the smaller vomeronasal organ (VNO), which issequestered at the base of the nasal septum (Farbman,1992; Mendoza, 1993). The structure of the olfactoryepithelium of the main olfactory system is characterizedby several laminar zones, representing the distribution ofdistinct cell types. Three major cell types may be differen-tiated. The most superficial zone is formed by the nuclei ofsustentacular or supporting cells. These cells extend pro-cesses from the mucosal surface to the basal lamina. Thenuclei of the olfactory sensory cells are located in themiddle region, and those of the basal cells are located at

the basal portion of the epithelium. Olfactory cells can bedistinguished easily from the microvilli-containing susten-tacular cells. They possess a slim dendrite. The dendriticknob carrying ciliary extensions extends above the epithe-lial surface. The epithelium is traversed by the ducts ofBowman’s glands.

Grant sponsor: Deutsche Forschungsgemeinschaft; Grant number: SFB269, A4.

*Correspondence to: Prof. Dr. Herbert Zimmermann, Biozentrum derJ.W. Goethe-Universitat, AK Neurochemie, Zoologisches Institut, Marie-Curie-Str. 9, D-60439 Frankfurt am Main, Germany.E-mail: [email protected]

Received 20 August 1997; Revised 4 December 1997; Accepted 4 Decem-ber 1997

THE JOURNAL OF COMPARATIVE NEUROLOGY 393:528–537 (1998)

r 1998 WILEY-LISS, INC.

Two types of basal cells can be differentiated on the basisof their morphological and biochemical phenotypes (Grazia-dei and Monti Graziadei, 1979; Hunter et al., 1994; Huardand Schwob, 1995). The dark/horizontal basal cells are indirect contact with the basal lamina. They can be identi-fied by their immunostaining for cytokeratins (Vollrath etal., 1985; Suzuki and Takeda, 1991; Holbrook et al., 1995).The light/globose basal cells are situated superficially tothe dark/horizontal basal cells, carry rounded nuclei, andstain only lightly with toluidine blue. They can be labeledspecifically by the monoclonal antibody GBC-1 (Goldsteinand Schwob, 1996). At a density considerably lower thanthat of the primary sensory cells, the superficial layercontains an additional and distinct cell type that is de-scribed as bipolar or flask-shaped. It reaches to the muco-sal surface with a knob-like and microvilli-extending struc-ture and is referred to as a microvillar cell (Moran et al.,1982). A potential sensory function of microvillar cells iscontroversially discussed (Rowley et al., 1989; Carr et al.,1991; Morrison and Costanzo, 1992; Yamagashi et al.,1992; Miller et al., 1995).

The less intensely investigated neuroepithelium of theVNO is also composed of several cell types, includingsupporting cells, primary sensory cells, and basal cells. Incontrast to the olfactory sensory cells, the dendritic end-ings of the sensory cells within the VNO contain numerousand long microvilli. The neuroepithelium is penetrated byblood vessels (Eisthen, 1992; Garrosa et al., 1992; Men-doza, 1993).

A unique feature of the olfactory epithelium is its abilityto give rise to new sensory neurons throughout life andalso following injury (Moulton et al., 1970; Graziadei andMonti Graziadei, 1979; Schwob et al., 1995). Basal cellsproliferate and can transform into sensory cells (Graziadeiand Monti Graziadei, 1979). Data from retrovirus lineagestudies (Caggiano et al., 1994; Hunter et al., 1994; Schwobet al., 1994) as well as studies of cellular kinetics (Calofand Chikaraishi, 1989; Suzuki and Takeda, 1991; De-Hamer et al., 1994; Huard and Schwob, 1995) support thenotion that the globose basal cells are the immediateneuronal progenitor cells. Recent investigations demon-strate that globose basal cells themselves represent aheterogeneous population (Calof et al., 1996; Goldsteinand Schwob, 1996).

The surface-located enzyme ecto-58-nucleotidase is ex-pressed on reactive glial and neural cells (Schoen et al.,1992; Schoen and Kreutzberg, 1994; Braun et al., 1997).Because it is transiently present at the surface of nervecells during development, it is regarded as a marker ofneural development (Schoen et al., 1988, 1991, 1993). Theenzyme catalyzes the extracellular hydrolysis of adenosine58-monophosphate (AMP), resulting in the formation ofextracellular adenosine (Zimmermann, 1992, 1996). Ecto-58-nucleotidase can also carry the cell-adhesion carbohy-drate epitope HNK-1 (Vogel et al., 1993), and it interactswith extracellular matrix proteins, such as laminin andfibronectin (Zimmermann, 1992). Experiments with cul-tured neural cells demonstrate that surface-located 58-nucleotidase can promote neurite outgrowth (Heilbronnand Zimmermann, 1995). The enzyme is essential forneuritic differentiation and survival of neural cells inculture (Heilbronn et al., 1995). In the olfactory bulb,where turnover of synaptic contacts continues into adult-hood, activity of ecto-58-nucleotidase is located at glomeru-lar and mitral cell synapses (Schoen and Kreutzberg,

1995). To probe for ecto-58-nucleotidase as a potentialmarker of specific cellular elements, we analyzed itsdistribution in the adult and developing rat olfactoryorgan.

MATERIALS AND METHODS

Animals

Mated female albino Wistar rats were obtained fromCharles River Wiga (Sulzfeld, Germany). Postnatal albinoWistar rats (nine animals) were obtained from Hoechst AG(Frankfurt, Germany). All protocols have been approvedby the appropriate care committee of the university. Ani-mals were anesthetized with CO2 and decapitated. Atembryonic day 12 (E12), E18, and E20, the uteri wereexcised and transferred into ice-cold phosphate-bufferedsaline (PBS; 137 mM NaCl, 3 mM KCl, 15 mM Na1/K1-phosphate buffer, pH 7.4). Then, the embryos (six animals)were removed from the uteri. The whole embryo or thehead only was mounted with Tissue Tek (Plano, Marburg,Germany) and quick-frozen in isopentane (280°C). Headsof postnatal animals (postnatal day 0 [P0], P8, P14, andP21) were dissected free of skin and muscle. Subsequently,the nose was removed from the skull and quick frozen. Thefrozen tissues were stored at 280°C until sectioning.

For electron microscopy, 3-week-old rats were anesthe-tized with sodium pentobarbital (Nembutal) and fixed byintracardial perfusion. After preperfusion with PBS con-taining heparin (0.5 mg/ml), 150 ml of fixative (4% parafor-maldehyde, 0.05 M cacodylate buffer, pH 7.4) was applied(4°C). After perfusion, the nose was opened, and theethmoturbinates were removed and immersed in fixativefor 1 hour at 4°C.

Light microscopic 58-nucleotidasehistochemistry

For light microscopic localization of 58-nucleotidase activ-ity in the olfactory epithelium, a lead-phosphate method(Braun et al., 1997) was applied. Frozen sections (10 µm or20 µm) were deposited on 3-aminopropyltriethoxy-silane(Sigma, Deisenhofen, Germany)-coated slides and allowedto dry for 1 hour. The dry tissue sections were stored at280°C until further processing. Sections warmed to roomtemperature were fixed with 0.05 M cacodylate-buffered4% paraformaldehyde, pH 7.4, for 12 minutes at roomtemperature and subsequently washed with 0.05 M caco-dylate buffer, pH 7.4, containing 0.25 M sucrose. Theenzyme reaction was carried out in a Tris-maleate-buffered substrate solution (1 mM 58-AMP, 2 mM Pb[NO3]2,5 mM MnCl2, 0.25 M sucrose, 50 mM Tris-maleate, pH 7.4)for 2.4 hours at room temperature. After washing thesections with demineralized water and incubating sectionsin an aqueous solution of ammonium sulfide (2% v/v), thelead orthophosphate precipitated as a result of 58-nucleotidase activity was visualized as a brown deposit. Tocontrol for nonspecific phosphatase activity, 58-AMP wassubstituted by p-nitrophenyl phosphate (Sigma). To con-trol for nonspecific lead precipitation, 58-AMP was omittedfrom the incubation medium. After dehydration in gradedethanol, sections were mounted with Permount.

Electron microscopic 58-nucleotidasehistochemistry

For electron microscopic analysis, fixed ethmoturbi-nates were washed for 1 hour (4°C) with 0.05 M cacodylate

ECTO-58-NUCLEOTIDASE IN THE RAT OLFACTORY ORGAN 529

buffer, pH 7.4, containing 0.25 M sucrose. The enzymereaction was carried out in the substrate solution de-scribed above for 1 hour at room temperature. Afterwashing in cacodylate-sucrose buffer, the reaction productwas visualized by incubating tissue blocks in 0.2 MTris/HCl, pH 6.8, containing ammonium sulfide (2% v/v).Tissue blocks were washed in cacodylate-sucrose bufferand fixed with 2.5% glutaraldehyde in PBS. Postfixationwas performed with 1% OsO4 in PBS. After several washeswith PBS, tissue blocks were transferred into 70% ethanolfollowed by staining with 2% uranyl acetate in 70%ethanol. After dehydration, blocks were embedded in Epon.Semithin sections were investigated with the light micro-scope, and selected sections were reembedded. Ultrathinsections were examined by using contrast enhancement,filtering out inelastic scattered electrons (EM 902 electronmicroscope; Zeiss, Oberkochen, Germany).

Morphometric analysis

The number of 58-nucleotidase1 microvillar cell pro-cesses was counted on 20-µm-thick coronal sections alongthe olfactory epithelium of turbinates and septum. Thelength of the epithelium analyzed was determined byusing an MCID image-analysis system (Imaging Research,St. Catharines, Ontario, Canada), and the average densityof cellular processes 6 S.E.M. per 400 µm was calculated.

RESULTS

58-Nucleotidase histochemistry

On cryosections of 3-week-old animals, activity of 58-nucleotidase can be detected at defined cellular elementswithin the entire olfactory and respiratory epithelia. Theseinclude basal cells, spindle-shaped cells above the layer ofthe sustentacular cell nuclei, ducts of Bowman’s glands,and the lumenal surface of the respiratory epithelium. Inaddition, the turbinal bone is strongly labeled (Fig. 1).Olfactory receptor cells and sustentacular cells of theolfactory epithelium remain unlabeled. The reaction of thebony tissue is nonspecific, because it is detectable also incontrol incubations without 58-AMP as substrate. In theabsence of 58-AMP or in the presence of p-nitrophenylphosphate as a substrate for nonspecific phosphatases,nasal epithelia reveal no labeling.

Basal cells carry the strongest enzyme activity. Incryosections, they form a single row lining the basallamina of olfactory (Fig. 1A–D) and respiratory epithelia(Fig. 1D,E). An analysis of shape and position of the58-nucleotidase1 basal cells within the olfactory epithe-lium at the electron microscopic level suggests that theycorrespond to the dark/horizontal type (Fig. 2A). Thesecells represent the deepest layer of the olfactory epithe-lium and have a flat or often a pyramidal shape. Theirnuclei are flat and of irregular shape. At the apical side,they carry numerous strongly labeled membrane invagina-tions. Lead deposit is situated entirely at the cell surface.At sites of contact with the basal lamina, the extracellulardeposits are less intense. Processes of dark/horizontal cellsare found in close contact with individual axons or withaxon bundles. These sites stain positive for 58-nucleotidase(Fig. 2A, inset).

A distinct population of cells with a head structureextending toward the epithelial surface is labeled at theapical side of the olfactory epithelium (Fig. 1A–C). Label-ing is most intense at their apical portion, situated superfi-

cially to the layer of sustentacular cell nuclei. To furtheridentify this cell type, we carried out a morphometric andalso an electron microscopic analysis. In 20-µm-thicksections, the average density of the cells is 14.8 6 3.4 per400 µm of olfactory epithelium. A comparison of thedensity of labeled cellular processes between the septaland turbinate olfactory epithelia revealed no significantdifference. The cells tend to be grouped within smallclusters (Fig. 1A). In semithin sections (Fig. 1C) and alsoat the ultrastructural level (Fig. 3A,B), the labeled cellshave a spindle-like appearance. Reaction product for 58-nucleotidase activity is situated at the cell surface. Thenuclei have an oval or irregular shape and are located justabove the layer of the nuclei of sustentacular cells. Byusing lead staining of ultrathin sections, their chromatinand nucleomatrix appear dark compared with the nuclei ofsustentacular cells (not shown). The apical process of thecells terminates in a flat, apical microvilli-carrying expan-sion, suggesting that they represent microvillar cells. Theminimal diameter below the apical knob of microvillar cellprocesses is approximately twice that of the diameter ofciliated olfactory receptor cell dendrites. Their cytoplasmis more electron dense than the cytoplasm of olfactoryreceptor cell dendrites. It has approximately the samedensity as the cytoplasm of sustentacular cells. In thecytoplasm, there are numerous mitochondria, but they arelargely excluded from the area immediately subjacent tothe surface of the epithelium (Fig. 3A,B). Reaction productfor 58-nucleotidase is distributed asymmetrically over thecell surface. It is intense at the apical microvilli, but itgradually decreases toward the cell body (Fig. 3B). Thesurface staining also outlines the plasma membrane invagi-nations along the apical process. A cross section throughthe surface of the olfactory epithelium demonstrates theregular arrangement of the microvilli that are surroundedby dense enzyme reaction product (Fig. 3A, inset).

The ducts of Bowman’s gland extending through theolfactory epithelium are only slightly stained in cryostatsections (Fig. 1B). However, at the ultrastructural level,reaction product is clearly detectable at the entire lumenalsurface of the cells that build the terminal portion of theglandular ducts (Fig. 2B). The branched tubular glandslocated within the lamina propria are unlabeled.

In contrast to the olfactory epithelium, activity of 58-nucleotidase is detectable at the apical surface of therespiratory epithelium (Fig. 1D,E). The reaction productcovers the entire surface as a broad band, corresponding tothe lumenal extension of the apical cilia. At the transitionbetween respiratory and olfactory epithelia, surface stain-ing is terminated abruptly (Fig. 1D).

Developmental analysis

58-Nucleotidase histochemistry was applied to assessthe appearance of enzyme activity in the developing nasalepithelia. Sections taken from rats at E12, E18, E20, P0,P8, and P14 were analyzed. No 58-nucleotidase activity isdetectable in the olfactory placode at E12 (not shown). AtE18 staining, is confined to the respiratory epithelium(Fig. 4A) and subepithelial glands. Basal cells as well asthe apical surface of the epithelium are labeled. Reactionproduct is largely restricted to the respiratory epitheliumalso at E20, and staining of the olfactory epitheliumremains very weak (Fig. 4B). At this stage, cellular struc-tures within the olfactory epithelium staining for 58-nucleotidase activity can hardly be identified. The basal

530 N. BRAUN AND H. ZIMMERMANN

staining is very faint. Labeled structures extending throughthe olfactory epithelium presumably represent ducts ofBowman’s gland. In contrast to 3-week-old rats, thesestructures also extend into the lamina propria. Occasion-

ally, faintly labeled cellular processes appear at the apicalsurface of the epithelium, but they cannot be assigned withcertainty to microvillar cells. After birth (P0), labeledcellular structures become discernible more easily (Fig. 4C).

Fig. 1. Localization of activity of 58-nucleotidase in olfactory andrespiratory epithelia (OE and RE, respectively). A: Cryosection includ-ing the dorsal lamella of the endoturbinale II. The bony tissue in thecenter of the lamella is intensely labeled by nonspecific lead precipi-tate. The lamina propria (LP) is free of reaction product. Dark/horizontal basal cells (HBC) are stained throughout the entire olfac-tory epithelium. At the apical surface and above the layer ofsustentacular cell nuclei, apical processes of microvillar cells arepositive for 58-nucleotidase (arrows). B: At higher magnification,faintly stained ducts of Bowman’s gland (arrowhead) become appar-ent. Arrows indicate apical processes of microvillar cells. Dark/horizontal basal cells (HBC) are also labeled. C: In semithin sections,

the surfaces of the apical processes of microvillar cells (arrows) and thedark/horizontal basal cells are outlined by lead precipitate. D: Stronglylabeled basal cells (BC) line the basal lamina throughout the entirenasal epithelia, including the respiratory (RE) and olfactory epithelia(OE; cryosection). Arrows indicate the microvillar cell processes in theolfactory epithelium. The staining of the apical surface of the respira-tory epithelium (arrowheads) terminates at the transition to theolfactory epithelium. E: Semithin section of the respiratory epithe-lium. 58-Nucleotidase activity occurs at the lumenal surface (arrow-head) and at the surface of basal cells. Scale bars 5 100 µm in A, 50 µmin B,C,E, 25 µm in D.


Stained apical processes of microvillar cells make theirappearance, and individual dark/horizontal basal cells canbe identified. Dark/horizontal basal cells at the border tothe respiratory epithelium are intensely labeled. Labelingceases with increasing distance to this border line. At P8,the entire epithelium is labeled, and the number of 58-nucleotidase1 microvillar cells is increased (Fig. 4D). Also,dark/horizontal basal cells and ducts of Bowman’s glandscarry reaction product. At P14, the staining pattern corre-sponds largely to that of the adult (Fig. 4E).

VNO

The distribution of 58-nucleotidase was examined also inthe adult and developing VNO. The neuroepithelium isalmost free of reaction product (P21; Fig. 5A). No labeledbasal cells can be detected. Only intraepithelial bloodvessels express the enzyme, and the apical surface occasion-ally reveals some very faint staining. Labeling of thereceptor-free epithelium is similar to that obtained in thenasal respiratory epithelium. Basal cells and the apicalsurface are intensely labeled. Vomeronasal glands are alsostained. Labeling of the receptor-free epithelium wasdetectable at the earliest developmental stage investi-gated (E18; not shown). At P0, reaction product is spreadover the entire epithelium (Fig. 5B). The thickness of theepithelium is increased at P8, and the apical superficialstaining can be differentiated from the labeling of thebasal cell layer (not shown). At all stages investigated, theneuroepithelium was essentially free of reaction product.

DISCUSSION

58-Nucleotidase is a markerof dark/horizontal basal cells

Within the adult olfactory epithelium, the dark/horizon-tal basal cells represent the most conspicuous cell typelabeled for 58-nucleotidase. The distribution of 58-nucleotid-ase resembles that of cytokeratin and of the receptor forepidermal growth factor (Holbrook et al., 1995). Dark/horizontal basal cells are heavily labeled also by a monoclo-nal antibody that binds to phosphotyrosine, and theyexpress binding sites for the isolectin B4 (BS-I-B4) fromBandeiraea simplicifolia. Globose basal cells are free ofthese markers and express the neural cell adhesion mol-ecule (N-CAM), a feature they share with the olfactoryreceptor cells (Miragall and Dermietzel, 1992). The dark/horizontal basal cells line the bottom of the epithelium andform envelopes around the bundles of olfactory axonsbefore these leave the epithelium. This intimate relation-ship might serve the exchange of signals between dark/horizontal basal cells and axons (Holbrook et al., 1995).

The function of the dark/horizontal basal cells needs tobe elucidated. The association of ecto-58-nucleotidase withthese cells can have several functional implications. ATP isreleased in many tissues, where it acts as an extracellularsignaling substance via receptor-mediated mechanisms. Itis then sequentially catabolized to adenosine. The latter istaken up by adjacent cells via high-affinity transportmechanisms serving purine salvage (Zimmermann, 1996).At the apical side of the epithelium, tight junctions pre-

Fig. 2. Ultrastructural localization of 58-nucleotidase at dark/horizontal basal cells (HBC) and glandular duct cells. A: Flat orpyramidal-shaped dark/horizontal basal cells are in contact with thebasal lamina (BL). The cells carry numerous, strongly labeled, thinprocesses. The surface of olfactory axons (arrowhead) in lacunaeensheathed by dark/horizontal basal cells is also stained. The inset

shows labeled olfactory axons at higher magnification (arrowhead)that are in contact with dark/horizontal basal cells (HBC) near thebasal lamina. B: Reaction product for 58-nucleotidase covers thelumenal surface (arrowheads) of the glandular ducts inside the olfac-tory epithelium. Scale bars 5 5 µm in A,B, 0.5 µm in inset.


vent the diffusion of the catabolite adenosine. At the basalside, activity of 58-nucleotidase could reduce the loss of theintermediate catabolite AMP by hydrolyzing it to adeno-sine. The same functional role might also apply to theenzyme at the basal surface of the respiratory epithelia.Labeling for 58-nucleotidase is intense at sites of contactwith globose basal cells and axons of receptor cells. Fromstudies on the interaction of lymphocytes with vascularendothelial cells, it has been suggested that 58-nucleotid-ase can function in cell adhesion. Lymphocytes readilyadhere to COS cells (SV40 transformed African greenmonkey kidney cells) transfected with 58-nucleotidase

(Airas et al., 1995). Thus, it is possible that the expressionof 58-nucleotidase at the surface of dark/horizontal basalcells also plays a role in the specific interaction with theaxonal processes leaving the epithelium.

Ontogenetic relationship between basal cellsof the respiratory epithelium

and the dark/horizontal basal cellsof the olfactory epithelium?

Little is known about the origin of the dark/horizontalbasal cells. After lesioning of the olfactory epithelium, they

Fig. 3. Ultrastructural localization of activity of 58-nucleotidase atmicrovillar cells. A: A 58-nucleotidase1 microvillar cell (MC) located inthe layer of sustentacular cells. Its nucleus (NU) is situated above thesustentacular cell nuclei. Reaction product covers the surface of themicrovillar cell process (arrows), including the apical microvilli (MV).A cross section demonstrates the regular array of the apical microvilli

and the presence of dense enzyme reaction product (inset). B: Thesurface staining (arrows) also outlines invaginations (arrowheads)along the MC surface. Staining includes the MV and is limited to theapical process of the cell. OC, olfactory sensory cell; star, dendriticknob of olfactory sensory cell. Scale bars 5 2 µm.


Fig. 4. Localization of 58-nucleotidase activity during pre- andpostnatal development (cryosections). A: Horizontal section throughthe right nasal cavity at embryonic day 18 (E18). The olfactoryepithelium (OE) does not express any 58-nucleotidase activity. Label-ing is restricted to the developing respiratory epithelium (RE) and theglands underneath it (arrowheads). SE, septum. B: Horizontal sectionof the right nasal cavity at E20. Apical and basal labeling is prominentat the respiratory epithelium and at associated glands (arrowheads).The olfactory epithelium is stained only slightly. C–E: Cross sectionsthrough the nasal epithelia at postnatal stages. C: After birth(postnatal day 0; P0), 58-nucleotidase activity persists in the respira-

tory epithelium. Basal cells and the apical surface (arrowheads) arelabeled. In the olfactory epithelium (OE), the dark/horizontal basalcells (HBC) and the microvillar cell processes (arrows) become discern-ible. 58-Nucleotidase1 basal cells of the olfactory epithelium firstappear at the boundary to the respiratory epithelium. Basal cells at agreater distance to the boundary are unstained. D,E: At P8 and P14,58-nucleotidase1 microvillar cells (arrows), dark/horizontal basal cells(HBC), and ducts of Bowman’s glands (arrowhead) are presentthroughout the entire olfactory epithelium. LP, lamina propria. Scalebars 5 500 µm in A,B, 50 µm in C–E.

presumably arise by transdifferentiation from another celltype during recovery of the epithelium. Labeling of horizon-tal/basal cells with the globose cell-specific monoclonalantibody GBC-1 supports the notion that, during epithe-lial reconstitution, globose cells can give rise not only toneurons but also to horizontal basal cells (Schwob et al.,1995; Goldstein and Schwob, 1996). It remains to beestablished whether this also applies to normal develop-ment. Our results demonstrate that, during development,staining for 58-nucleotidase of basal cells in the respiratoryepithelium by far precedes that of the olfactory epithelialbasal cells. Whereas staining is prominent at E18 and E20in the respiratory epithelium, positive basal cells in theolfactory epithelium can be identified only at experimentalstage P0. The first labeled dark/horizontal basal cellsmake their appearance at the border to the respiratoryepithelium. This phenomenon is very reminiscent of thedevelopment of immunostaining for cytokeratins or epider-mal growth factor and staining for BS-I-B4 (Suzuki andTakeda, 1991; Holbrook et al., 1995). Could basal cells ofthe respiratory epithelium migrate into the olfactory epi-thelium to form dark/horizontal basal cells?

In this respect, it is of interest that the neuroepitheliumof the VNO lacks 58-nucleotidase1 basal cells, although thecells at the basal side of the receptor-free epithelium carrythe enzyme. On [3H]thymidine labeling of the VNO, apopulation of dividing cells could be identified in theregions of the neuroepithelium adjacent to the boundarieswith the receptor-free epithelium. It has been suggestedthat these cells can become both neurosensory and support-ing cells. The stem cells of the VNO would thus not besituated in the basal layer but at the margins of theepithelial sheet (Barber and Raisman, 1978).

It is generally accepted that the globose basal cellsrepresent the reservoir of progenitor cells for the immedi-ate replacement of lost olfactory receptor cells (Farbman,1992; Hunter et al., 1994; Schwob et al., 1995; Calof et al.,

1996). Dark/horizontal basal cells, however, do not appearto simply represent a static element within the epithelium.Not only globose cells but also dark/horizontal basal cellsincorporate [3H]thymidine, suggesting that they can alsoproliferate in the mature epithelium (Graziadei and MontiGraziadei, 1979; Mackay-Sim and Kittel, 1991; Schwartz-Levey et al., 1991). They also respond to epithelial damageand may have the potential to act as stem cells (Holbrooket al. 1995; Schwob et al., 1995). In the unperturbed andadult animal, the turnover of olfactory sensory cells and,thus, the need for replacement may be less pronounced.Proliferation density declines rapidly with age, suggestingthat the life span of individual olfactory sensory cellsincreases with age (Weiler and Farbman, 1997). At pre-sent, it is uncertain whether the presence of 58-nucleotid-ase at dark/horizontal cells can be interpreted as a charac-teristic of their immature differentiation status andwhether the paradigm for ecto-58-nucleotidase as a neuralmaturation marker can be applied to the olfactory system.In the central nervous system, the enzyme was found to betransiently associated with neurons in both the developingand the regenerating adult nervous system in a number ofcases. These include migrating nerve cells and transientclimbing fiber synapses during postnatal ontogeny of therat cerebellum or also synapses of the developing visualcortex of the cat. Within the main olfactory bulb of theadult rat, the enzyme is located at glomerular and mitralsynapses that are prone to synaptic turnover even atmaturity (Schoen and Kreutzberg, 1995).

Microvillar cells carry ecto-58-nucleotidaseactivity

Ecto-58-nucleotidase is a marker also of microvillar cellsof the olfactory epithelium. Labeling is confined to theapical portion of the cell and includes its microvillarprocesses. This type of cell is absent from the VNO,

Fig. 5. Localization of 58-nucleotidase activity in the vomeronasalorgan (cryosections). A: The reaction product is confined mainly to thereceptor-free epithelium (RFE). Staining of basal cells (BC) is discern-ible from that of the apical surface (arrows). In the neuroepithelium(NE), intraepithelial blood vessels are labeled (arrowhead), but the

basal side of the epithelium (BC) is free of reaction product. B: Also,immediately after birth (P0), the reaction product is confined to thereceptor free epithelium (RFE). The neuroepithelium (NE) is unla-beled. Scale bar 5 50 µm.


although olfactory receptors in this epithelium carry micro-villi rather than cilia (Farbman, 1992). Microvillar cellscan also be labeled selectively with the monoclonal anti-body 1A-6, which has been raised against rat olfactoryepithelia (Carr et al., 1991), or with antibodies directedagainst the calcium-binding proteins spot-35 and calbin-din (Yamagashi et al., 1993). In this respect, they differfrom the olfactory receptor cells that may be labeled forolfactory marker protein (Margolis, 1972) or also forneuron-specific enolase (Yamagashi et al., 1992). Themorphology of microvillar cells varies to some extentbetween species. The cells revealing activity of 58-nucleotidase in our experiments are spindle-shaped ratherthan flask-shaped and are similar in shape and positionwithin the epithelium to the cells that react positively forthe spot-35 protein (Yamagashi et al., 1992, 1993). The58-nucleotidase1 microvillar cells cannot be identifiedclearly before birth, in agreement with the late develop-ment of 1A-61 cells (Carr et al., 1991). The number oflabeled cells per length of epithelium corresponds to thatobserved for 1A-6 immunopositive cells. The functionalrole of 58-nucleotidase activity at the surface of these cellsremains elusive. Apparently, it is selectively required atthe apical portion of the cell, where it needs to be trans-ported by a cell domain-specific sorting process. Numerousinvaginations of the plasma membrane increase the reac-tive surface of the cell below the tight-junctional barrier.Would these cells respond to extracellular nucleotides?

Staining of the epithelial surface

The surface of the respiratory epithelium stains for58-nucleotidase. The reactive material appears to be associ-ated with the apical cilia. In contrast, the surface of theolfactory epithelium remains negative. It appears conceiv-able that surface-located 58-nucleotidase can contribute tothe inactivation and degradation of nucleotides releasedinto the mucus. Human respiratory nasal epithelia appar-ently possess receptors for ATP and have been shown toreact to the application of extracellular ATP by elevatinglevels of intracellular Ca21 (Paradiso et al., 1995).

Ducts of Bowman’s glands within theolfactory epithelium

Bowman’s glands contribute most to the composition ofthe mucus covering the lumenal surface of the olfactoryepithelium. Their acini are situated in the lamina propria,but the ducts extend through the epithelium. The innerlining of the ducts stains for activity of 58-nucleotidase, butneither the glands themselves nor the proximal parts ofthe ducts within the lamina propria are reactive. Studieson lesioned epithelium suggest that Bowman’s glands andducts contribute to the recovery of sustentacular cells(Schwob et al., 1995). The enzyme activity at the innerlining of the ducts might serve the salvage of purinecompounds after hydrolysis to the nucleoside and reuptakeinto the duct cells. It also suggests that 58-nucleotidase canbe targeted selectively to the lumenal surface domain ofthe duct cells.

CONCLUSIONS

58-Nucleotidase can be used successfully as a marker fordefined cell types of the nasal epithelia. It is a prominentmarker of the dark/horizontal basal cells in the olfactoryepithelium, of the basal cells of the respiratory epithelia in

the main olfactory organ and VNO, and of the microvillarcells of the olfactory epithelium. Labeling also includes thelining of the ducts of Bowman’s glands as they traverse theepithelium and the surface of the respiratory epithelia.Because 58-nucleotidase appears to function not only as anenzyme but also can be involved in cell/matrix and cell/cellinteractions, its functional role might differ at the variouslocations. Purine salvage by producing extracellular aden-osine for reuptake could be an important function at theepithelial boundaries. At the same time, the protein mightbe involved in the interaction between cells, as in the caseof dark/horizontal basal cells and axons of the sensorycells. Neither olfactory neurons nor their immediate pro-genitors carry the enzyme. This excludes 58-nucleotidaseas a marker for immediate neural development in theolfactory organ.

ACKNOWLEDGMENTS

We thank Klaus Hammer for excellent technical sup-port.

LITERATURE CITED

Airas, L., J. Hellman, M. Salmi, P. Bono, T. Puurunen, D.J. Smith, and S.Jalkanen (1995) CD73 is involved in lymphocyte binding to theendothelium: Characterization of lymphocyte vascular adhesion pro-tein 2 identifies it as CD73. J. Exp. Med. 182:1603–1608.

Barber, P.C. and G. Raisman (1978) Cell division in the vomeronasal organof the adult mouse. Brain Res. 141:57–66.

Braun, N., C. Lenz, F. Gillardon, M. Zimmermann, and H. Zimmermann(1997) Focal cerebral ischemia enhances glial expression of ecto-5-nucleotidase. Brain Res. 766:213–226.

Caggiano, M., J.S. Kauer, and D.D. Hunter (1994) Globose basal cells areneuronal progenitors in the olfactory epithelium: A lineage analysisusing a replication-incompetent retrovirus. Neuron 13:339–352.

Calof, A.L. and D.M. Chikaraishi (1989) Analysis of neurogenesis in amammalian neuroepithelium: Proliferation and differentiation of anolfactory neuron precursor in vitro. Neuron 3:115–127.

Calof, A.L., N. Hagowara, J.D. Holcomb, J.S. Mumm, and J. Shou (1996)Neurogenesis and cell death in olfactory epithelium. J. Neurobiol.30:67–81.

Carr, V.M., A.I. Farbman, L.M. Colletti, and J.I. Morgan (1991) Identifica-tion of a new nonneuronal cell type in rat olfactory epithelium.Neuroscience 45:433–449.

DeHamer, M.K., J.L. Guerva, K. Hannon, B.P. Olwin, and A.L. Calof (1994)Genesis of olfactory receptor neurons in vitro: Regulation of progenitorcell divisions by fibroblast growth factors. Neuron 13:1083–1097.

Eisthen, H.L. (1992) Phylogeny of the vomeronasal system and of receptorcell types in the olfactory and vomeronasal epithelia of vertebrates.Microsc. Res. Technol. 23:1–21.

Farbman, A.F. (1992) Cell Biology of Olfaction. New York: Press Syndicateof the University of Cambridge.

Garrosa, M., C. Iniguez, J.M. Fernandez, and M.J. Gayoso (1992) Develop-mental stages of the vomeronasal organ in the rat: A light and electronmicroscopic study. J. Hirnforsch. 33:123–132.

Goldstein, B.J. and J.E. Schwob (1996) Analysis of the globose basal cellcompartment in rat olfactory epithelium using GBC-1, a new monoclo-nal antibody against globose basal cells. J. Neurosci. 16:4005–4016.

Graziadei, P.P.C. and G.A. Monti Graziadei (1979) Neurogenesis andneuron regeneration in the olfactory system of mammals. I. Morphologi-cal aspects of differentiation and structural organization of the olfactorysensory neurons. J. Neurocytol. 8:1–18.

Heilbronn, A. and H. Zimmermann (1995) 58-Nucleotidase activates and aninhibitory antibody prevents neuritic differentiation of PC12 cells. Eur.J. Neurosci. 7:1172–1179.

Heilbronn, A., V. Maienschein, C. Carstensen, W. Gann, and H. Zimmer-mann (1995) Crucial role of 58-nucleotidase in differentiation andsurvival of developing neural cells. Neuroreport 7:257–261.

Holbrook, E.H., K.E. Szumowski, and J.E. Schwob (1995) An immunochemi-cal, ultrastructural, and developmental characterization of the horizon-


tal basal cells of rat olfactory epithelium. J. Comp. Neurol. 363:129–146.

Huard, J.M. and J.E. Schwob (1995) Cell cycle of globose basal cells in ratolfactory epithelium. Dev. Dynam. 203:17–26.

Hunter, D.D., M. Caggiano, and J.S. Kauer (1994) Lineage analysis of theolfactory epithelium using a replication-incompetent retrovirus. Chem.Sens. 19:683–693.

Mackay-Sim, A. and P. Kittel (1991) Cell dynamics in the adult mouseolfactory epithelium: A quantitative autoradiographic study. J. Neuro-sci. 11:979–984.

Margolis, F.L. (1972) A brain protein unique to the olfactory bulb. Proc.Natl. Acad. Sci. USA 69:1221–1224.

Mendoza, A.S. (1993) Morphological studies on the rodent main andaccessory olfactory systems: The regio olfactoria and vomeronasalorgan. Ann. Anat. 175:425–446.

Miller, M.L., A. Andringa, J.E. Evans, and L. Hastings (1995) Microvillarcells of the olfactory epithelium: Morphology and regeneration follow-ing exposure to toxic compounds. Brain Res. 669:1–9.

Miragall, F. and R. Dermietzel (1992) Immunocytochemical localization ofcell adhesion molecules in the developing and mature olfactory system.Microsc. Res. Technol. 23:157–172.

Moran, D.T., J.C. Rowley, III, and B.W. Jafek (1982) Electron microscopy ofhuman olfactory epithelium reveals a new cell type: The microvillarcell. Brain Res. 253:39–46.

Morrison, E.E. and R.M. Costanzo (1992) Morphology of olfactory epithe-lium in humans and other vertebrates. Microsc. Res. Technol. 23:49–61.

Moulton, D.G., G. Celebi, and R.P. Fink (1970) Olfaction in mammals—Twoaspects; Proliferation of cells in the olfactory epithelium and sensitivityto odours. In G.E.W. Wostenholme and J. Knight (eds): Ciba FoundationSymposium on Taste and Smell in Vertebrates. London: Churchill, pp.227–250.

Paradiso, A., S.J. Mason, R. Lazorowski, and R.C. Boucher (1995) Mem-brane-restricted regulation of Ca21 release and influx in polarizedepithelia. Nature 377:643–646.

Rowley, J.C., III, D.T. Moran, and B.W. Jafek (1989) Peroxidase backfillssuggest the mammalian olfactory epithelium contains a second morpho-logically distinct class of bipolar sensory neuron: The microvillar cell.Brain Res. 502:387–400.

Schoen, S.W. and G.W. Kreutzberg (1994) Synaptic 58-nucleotidase activityreflects lesion-induced sprouting within the adult dentate gyrus. Exp.Neurol. 127:106–118.

Schoen, S.W. and G.W. Kreutzberg (1995) Evidence that 58-nucleotidase isassociated with malleable synapses—An enzyme cytochemical investi-gation of the olfactory bulb of adult rats. Neuroscience 65:37–50.

Schoen, S.W., M.B. Graeber, L. Toth, and G.W. Kreutzberg (1988) 58-Nucleotidase in postnatal ontogeny of rat cerebellum: A marker formigrating nerve cells? Dev. Brain Res. 39:125–136.

Schoen, S.W., M.B. Graeber, L. Toth, and G.W. Kreutzberg (1991) Synaptic58-nucleotidase is transient and indicative of climbing fiber plasticityduring the postnatal development of rat cerebellum. Dev. Brain Res.61:125–138.

Schoen, S.W., M.B. Graeber, and G.W. Kreutzberg (1992) 58-Nucleotidaseimmunoreactivity of perineuronal microglia responding to rat facialnerve axotomy. Glia 6:314–317.

Schoen, S.W., G.W. Kreutzberg, and W. Singer (1993) Cytochemical redistri-bution of 58-nucleotidase in the developing cat visual cortex. Eur. J.Neurosci. 5:210–222.

Schwartz-Levey, M.A., D.M. Chikaraishi, and J.S. Kauer (1991) Character-ization of potential precursor populations in the mouse olfactoryepithelium using immunocytochemistry and autoradiography. J. Neuro-sci. 11:3556–3564.

Schwob, J.E., J.M. Huard, M.B. Luskin, and S.L. Youngentob (1994)Retroviral lineage studies of the rat olfactory epithelium. Chem. Sens.19:671–682.

Schwob, J.E., S.L. Youngentob, and R.C. Mezza (1995) Reconstitution of therat olfactory epithelium after methyl bromide-induced lesion. J. Comp.Neurol. 359:15–37.

Suzuki, Y. and M. Takeda (1991) Keratins in the developing olfactoryepithelia. Dev. Brain Res. 59:171–178.

Vogel, M., H. Zimmermann, and W. Singer (1993) Transient association ofthe HNK-1 epitope with 58-nucleotidase during development of the catvisual cortex. Eur. J. Neurosci. 5:1423–1425.

Vollrath, M., M. Altmannsberger, K. Weber, and M. Osborn (1985) Anultrastructural and immunohistological study of the rat olfactoryepithelium: Unique properties of olfactory sensory cells. Differentiation29:243–253.

Weiler, E. and A.I. Farbman (1997) Proliferation in the rat olfactoryepithelium: Age-dependent changes. J. Neurosci. 17:3610–2622.

Yamagashi, M., H. Nakamura, Y. Nakano, and R. Kuwano (1992) Immuno-histochemical study of the fourth cell type in the olfactory epithelium inthe guinea pigs and in a patient. ORL 54:85–90.

Yamagashi, M., Y. Ishizuka, M. Fujiwara, H. Nakamura, S. Igarashi, Y.Nakano, and R. Kuwano (1993) Distribution of calcium binding proteinsin sensory organs of the ear, nose and throat. Acta Otolaryngol.(Stockholm) 506(Suppl.):85–89.

Zimmermann, H. (1992) 58-Nucleotidase—Molecular structure and func-tional aspects. Biochem. J. 285:345–365.

Zimmermann, H. (1996) Biochemistry, localization and functional roles ofecto-nucleotidases in the nervous system. Progr. Neurobiol. 49:589–618.


Documents

Association of ecto-5?-nucleotidase with specific cell types in the adult and developing rat olfactory organ