Journal of Chemical Neuroanatomy, Vol. 5:51 62 (1992)

Carnosine, Nerve Growth Factor Receptor and Tyrosine Hydroxylase Expression during the Ontogeny of the Rat Olfactory System Stefano

Biffo, Elisa Mark and Aldo Fasolo*

Dipartimento di Biologia Animale, Universifft di Torino, V. Accademia Albertina 17, 10123 Torino, Italy

ABSTRACT The localizations ofcarnosine, nerve growth factor (NGF) receptor and tyrosine hydroxylase (TH) were studied in the embryonic and postnatal rat olfactory bulb and epithelium by means of single- and double-immunostaining methods. Tyrosine hydroxylase ontogeny was also evaluated at the mRNA level by in situ hybridization. All these molecules were expressed in the olfactory bulb but with different developmental patterns and cellular localization: carnosine immunoreactivity is seen from embryonic day 17 in primary olfactory neurons scattered in the nasal cavity and in fibres projecting from them to the olfactory bulb. Nerve growth factor-receptor immunoreactivity associated with small glial-like cells is visible in some glomeruli starting from the second day of postnatal life. At postnatal day 10 NGF-receptor immunoreactivity is extended to all glomeruli, Periglomerular neurons expressing TH mRNA and protein are present prenatally and their number sharply increases during the early postnatal development. Doublestaining methods show that TH and NGF-receptor immunoreactivity do not overlap in cell bodies and processes. In addition, NGF-receptor immunoreactivity is not colocalized with carnosine. These findings definitely exclude NGF-receptor expression in periglomerular and primary olfactory neurons, suggesting that at least part of NGF-receptor expression in the olfactory bulb is associated with glial cells. In addition, they provide the first immunohistochemical data on carnosine ontogeny and confirm at the mRNA level previous studies on the ontogeny of TH protein. KEYWORDS: in situhybridization

Immunohistochemistry

INTRODUCTION The olfactory system is a remarkable model in which to study, in vivo, basic phenomena of neuronal differentiation such as neurogenesis, targetmediated survival and trans-synaptic regulation. All of these processes occur normally in the olfactory epithelium and bulb during development and in adulthood and can be experimentally manipulated (for review, see Margolis and Getchell, 1988). Primary chemosensory olfactory neurons are located in the nasal cavity and project their axons to the glomerular layer of the olfactory bulb where they synapse with target neurons, i,e. mitral, periglomerular and tufted cells (Shepherd and Greer, 1990). During embryonic development, olfactory neurons arise from the olfactory placode, beginning from embryonic day (E) 10 in the mouse (Brunjes and Frazier, 1986) and before the formation of the anlage of the olfactory bulb (Rugh, 1968; Cuschieri and Bannister, 1975). Shortly after their generation *Correspondence should be addressed to: Dr Aldo Fasolo, Dipartimento di Biologia Animale, Universit/~ di Torino, V. Accademia Albertina 17, 10123Torino, ltaly. 0891-0618/92/01005 l - 12 $06.00 © 1992 by John Wiley and Sons Ltd

they send axons to the developing olfactory bulb: the early differentiation of olfactory neurons is associated with the expression of proteins that are involved in neuronal axonal growth as B-50/GAP43 (Verhaagen et al., 1989, 1990a; Biffo et al., 1990c) or cell adhesion as embryonic cell adhesion molecules (N-CAM; Miragall et al., 1988). The final stage of differentiation, correlated with the contact of olfactory axons with the olfactory bulb, is accompanied by the expression of the olfactory marker protein (OMP; Margolis, 1972; Monti-Graziadei et al., 1980; Farbman and Margolis, 1980; Miragall and Monti-Graziadei, 1982; Verhaagen et al., 1990b) and the appearance of dendritic cilia (Graziadei, 1973; Menco and Farbman, 1985: Costanzo and Morrison, 1989). Early anatomical reports suggested an inductive role for the olfactory neurons to the olfactory bulb based on the evidence that olfactory placode explants induce ectopic formation of glomeruli in brain (Graziadei and Samanen, 1980: Graziadei and Monti-Graziadei, 1986). The ability of olfactory neurons to influence their target cells in the olfactory bulb has been well demonstrated in the adult

52

S. Biffo, E. Marti and A. Fasolo

animal, where olfactory peripheral denervation results in dramatic morphologic changes of the olfactory bulb (Margolis, 1974; Margolis et al., 1974) and strikingly, it causes the specific disappearance of the dopaminergic phenotype in the periglomerular neurons (Kawano and Margolis, 1982; Baker et al., 1983) due to the loss of tyrosine hydroxylase (TH) mRNA (Erlich et aL, 1990). There is evidence also that the initiation of TH expression during embryonic development is dependent on signals deriving from the olfactory nerve since TH protein expression in periglomerular neurons is detected only after their migration from the ventricular layer--where they are generated--to the glomerular layer, where they make contact with olfactory axons (McLean and Shipley, 1988). These data indicate that olfactory neurons stimulate TH expression either directly through synaptic activity or indirectly through activity-dependent release of factors. Some lines of evidence seem to indicate a possible involvement of nerve growth factor (NGF; LeviMontalcini, 1987) in the complex interactions that occur between olfactory neurons and their target cells. First, N G F has been found transiently expressed during development in primary olfactory neurons (Williams and Rush, 1988). Second, NGFreceptor immunoreactivity has been found in the glomerular layer of the adult olfactory bulb (Yan and Johnson, 1989; Pioro and Cuello, 1990) though its cellular origin is controversial. Interestingly, a previous report has described NGF-receptor immunoreactivity associated with periglomerular neurons (Gomez-Pinilla et al., 1987). Third, NGFreceptor immunoreactivity is increased following unilateral olfactory deprivation (Gomez-Pinilla et al., 1988). In this study we address two important questions related to the presence of N G F receptor in the olfactory bulb: (1) which cells are NGF-receptor positive in the glomerular layer and, in particular, are some of these cells periglomerular neurons and (2) how is the developmental pattern of NGFreceptor expression related to the maturation of other marker molecules of the olfactory system? We have therefore analysed the patterns of development and distribution of carnosine, a dipeptide selectively expressed in mammalian olfactory neurons (Margolis, 1974; Neidle and Kandera, 1974; Sakai et al., 1987; Biffo et al., 1990b), TH, a marker for periglomerular dopaminergic neurons, and N G F receptor.

MATERIALS AND METHODS

Animals Timed pregnancies were obtained from SpragueDawley female rats. Birth generally occurred

between E21 and E22. After overnight mating, the day on which a vaginal plug was observed was noted as E 1. The number of animals used for each time point ranged from six (embryos) to three (postnatals). Fetuses were removed from the mother at E 15, El7 and E20, their brains and olfactory epithelia were dissected out and fixed by immersion in 4% paraformaldehyde in 0.1 M-phosphate buffer, pH 7.4 (PB), for 12h at 4°C. Pups from birth (postnatal day 1, P1) to P7 were killed daily by decapitation, their olfactory epithelia and brains were dissected out and fixed by immersion as above. Older rats (10-day-, 2-week-, 2-month-old and adults) were anaesthetized with Nembutal and perfused transcardially with phosphate-buffered saline (PBS) followed by 4% paraformaldehyde in PB. Following dissection, the olfactory bulbs and epithelia were postfixed in 4% paraformaldehyde in PB, for 4 h at 4°C. All tissues were cryoprotected in ascending surcrose solutions (7.5, 15 and 30%) in PB until they sank, frozen in isopentane liquid nitrogen cooled and stored at -70°C. Ten micron thick sections from all the levels of the olfactory bulbs and epithelia were cut with a cryostat onto polylysinecoated slides. In situ hybridization

A 35mer oligonucleotide complementary to the rat TH mRNA (nucleotides 1223-1258 from the rat sequence: Grima et al., 1985) was labelled to a specific activity of 3 × 108cpm/lxg by oligotail labelling (Eschenfeldt et al., 1987). Briefly, 100 ng of dephosphorylated oligonucleotide were incubated with 40~tCi of alpha-[3~S] d-CTP (1200mCi/mM specific activity) and 7 units of terminal deoxynucleodidyltransferase m labelling buffer (Promega), for 2h at 37°C. The reaction was stopped by adding ethylenediaminotetra-acetic and sodium dodecylsulphate. Unincorporated nucleotides were removed through a Sephadex G-25 spin column. The flow-through from the column was ethanol precipitated, dissolved in sterile water and stored at - 70°C. Prior to hybridization, sections were sequentially hydrated, acetylated, washed and dehydrated. Incubation with the labelled TH oligonucleotide was performed with 500 000 cpm/slide in hybridization buffer (Biffo et al., 1990c), for 16h at 42°C. Following hybridization, sections were sequentially washed twice in 2X sodium salt citrate (SSC) at room temperature and at 55°C, and once in 1XSSC, 10mM-dithiothreitol for 30min at 50°C. Sections were dehydrated in ascending ethanol containing 0.3 M-ammonium acetate. After drying, slides were coated with NTB2 emulsion (Kodak), exposed for up to 3 weeks, developed, counterstained and mounted as previously described (Biffo et al., 1990c).

NGFr/TH/carnosine in Olfactory Development Controls of specificity included the use of an excess (20-fold) of unlabelled TH oligonucleotide and the pretreatment of sections with Ribonuclease A (5 ~tg/ml, 30 rain, 37°C): no labelling was seen following these treatments.

lmmunohistochemistry: TH and carnosine A rabbit polyclonal antiserum against carnosine and a rabbit polyclonal antiserum against rat TH were used. Their specificity profiles and preparations have been thoroughly described elsewhere. Briefly, anti-carosine specifically recognizes the aminoacylhistidine dipeptide-conjugates of carnosine, homocarnosine and anserine and does not cross-react with any amino acid conjugate (Biffo et al., 1990b). Since the only dipeptide biochemically detectable in the olfactory system is carnosine (Margolis et al., 1985), carnosine immunoractivity can be ascribed exclusively to carnosine and neither to homocarnosine nor anserine. Rabbit polyclonal anti-TH was raised against rat TH purified from a rat phaechromocytoma cell line and has been shown to be specific to TH (Thibault et al., 1981). Single stainings were performed by the avidinbiotin peroxidase technique using the Vectastain (Vector) reagents, as previously described (Biffo et al., 1990a). Sections from unperfused animals were treated for 30min with 0.3% H202 in methanol in order to block the endogenous peroxidase activity. Briefly, sections were rehydrated, methanol-treated, washed and sequentially incubated with normal goat serum (1:100, 30 min, RT), primary antisera (1:800, overnight, RT), biotinylated goat anti-rabbit (1:250, 30min, RT) and avidin-biotin complex. All the dilutions of antibodies were performed in 0.1% Triton-X in PBS. Following each incubation, sections were washed in PBS. The reaction was developed with 0-02% diaminobenzidine, 0.2% ammonium nickel sulphate and 0.025% H202. Control reactions were performed using a normal rabbit serum instead of the primary antiserum or preadsorbing carnosine antiserum with carnosine bovine serum albumin conjugate (Artero et al., 1991): no staining was seen using these procedures.

NGF-receptor histochemistry Since NGF-receptor histochemistry with tissue collected on slides gave relatively low signal levels, all the single-staining experiments with NGFreceptor antibody and most of the double-staining experiments were performed on free-floating sections essentially according to Yan and Johnson (1989). The anti-NGF receptor used here is a mouse monoclonal antibody, know as Mab 192, specific to the rat N G F receptor (Chandler et al., ! 984). Recent studies suggest that it recognizes predominantly or

53

exclusively the low affinity N G F receptor (Radeke et al., 1987; Weskamp and Reichardt, 1991 ). Frozen tissues were cut on a cryostat at a thickness ranging from 15 to 30 ~tm and sections were collected at 0.2% Triton-X-PBS and processed for histochemistry. Thinner sections were used for double-staining experiments. Briefly, sections were incubated with Mab 192 (supernatant from hybridoma cells kindly obtained from Dr L. Aloe, CNR, Rome and originally from Dr E. Johnson) in 0.2% Triton-X in PBS containing 5% normal horse serum, for 24 h at 4'~C. Following washes, sections were incubated in biotinylated horse anti-mouse, for 16 h at 4°C and subsequently in avidin-biotin complex. The reaction was developed as above and sections were either dehydrated (single methods) or glycerol mounted (double methods). Control reactions were performed using unrelated mouse monoclonal antibodies, e.g. antibodies against choline acetyltransferase (Boehringer) and mouse IgG. No staining was seen in the olfactory bulb using these immunochemical reagents except for anti-choline acetyltransferase, where some fibres were seen in the glomerular layer as well in other parts of the main olfactory bulb.

Immunohistochemistry double staining Routinely, double-staining labellings with antiN G F receptor were performed on free-floating sections. The first reaction was run using mouse anti-NGF receptor as described above. Following the development of the first reaction with diaminobenzidine plus ammonium nickel sulphate (blue colour), sections were incubated with either rabbit polyclonal anti-TH or anti-carnosine, for 24 h at 4°C. The second reaction was then revealed by either fluoresceinated anti-rabbit antibodies (24h) or sequentially by biotinylated anti-rabbit antibodies (16h) followed by avidin-biotin complex and a peroxidase reaction using only diaminobenzidine (brown colour) as a chromogen. Controls for crossreaction between secondary antibodies were performed by: (1) the omission of either one of the two primary antibodies and (2) the preadsorption of the primary antiserum with the homologous antigen for anti-carnosine. The omission or preadsorbtion of the primary antibodies in double-staining experiments always resulted in the lack of immunoreactivity for the corresponding antigen. The possibility that the first reaction impeded the full attachment of the antibodies in the second reaction was also checked by inverting the order of the reactions and compring the intensity pattern between sections treated with single staining and double staining: no differences in the intensity and the pattern of TH, and carnosine stainings were seen when the NGFreceptor reaction was omitted or the order of the reactions was exchanged.

54

S. Biffo. E. Marti and A. Fasolo

ob

Fig. 1. Carnosine ontogeny in the olfactory epithelium and olfactory bulb. (A) Embryonic day 17. Low magnification of the nasal region showing immunoreactive fibres in the developing olfactory bulb (ob) and immunopositive neurons in the olfactory epithelium. (B) Higher magnificationof(A); note that carnosine immunoreactivityisnot associated with glomerular-likestructures and the presence ofcarnosinepositive fibres extending to the inner layers of the ob. (C) Carnosine-positiveneurons in the olfactory epithelium. The arrowhead points to an olfactory knob. (D) Postnatal day 5. Carnosine immunoreactivity is organized in glomerular structures. Abbreviations: gl, glomerular layer; epl, external plexiform layer; fl, fibre layer. Scale bar is 150 v.m for (A) and 25 I~m for (B,C,D).

RESULTS

Carnosine, NGF-receptor and TH expression in the developing olfactory system Carnosine ( ~-alanyl-histidine ) This dipeptide is expressed at high levels in the olfactory neurons of adult rats (Sakai et al., 1987; Biffo et al., 1990b), whereas the only available biochemical data seem to indicate the beginning of carnosine synthesis during development from El7 (Margolis et al., 1985). Carnosine immunostaining was not seen at El5. Starting from El7 and more extensively at E20, immunopositive neurons were found scattered in the nasal mucosa, but exclusively in the dorsal and posterior regions, where the olfactory epithelium is located (Fig. IA). Olfactory

neurons were stained both in the dendrite and in the axon (Fig. I C). Thin bundles of carnosineimmunoreactive fibres projected to the olfactory bulb. The first carnosine-immunopositive fibres were seen innervating mainly the ventral part of the olfactory bulb. Carnosine immunostaining in the embryonic olfactory bulb was characterized by the total absence of glomerular structures and the extension of carnosine-immunopositive fibres to the innermost layers (Fig. 1B). At birth, carnosineimmunopositive staining was seen at the periphery of the olfactory bulb in small scatteredglomerularlike structures and some carnosine-immunoreactive fibres extended beyond the glomerular layer deep to the external plexiform layer. During the postnatal development of the olfactory bulb, carnosine

N G F r / T H / c a r n o s i n e i n Olfactory Development

fl

55

c,

epl 2A

Fig. 2. Nerve growth factor (NGF)-receptor ontogeny in the olfactory bulb. All sections were treated as free floating (thickness 30 ~tm). (A) Postnatal day 2. A lightly stained glomerulus (large arrowhead) and an immunopositiveglial-likecell (small arrowhead) are visible. (B) Glomeruli and glial-likecells stained in the postnatal day l0 olfactory bulb. (C) 2-month-old olfactory bulb. NGF-receptor staining intensity has reached the adult level. Abbreviations:gl, glomerular layer; epl, external plexiformlayer. Scale bars are 15 ~tm for (A) and 25 lam for (B,C). immunostaining became progressively limited to the glomerular layer (Fig. 1D). Nerve growth factor receptor In the embryonic rat olfactory epithlium, N G F receptor immunoreactivity was found associated with the trigeminal nerve innervating the nasal cavity but was absent from the carnosine-positive primary olfactory neurons (not shown). No evidence of N G F - r e c e p t o r staining associated with primary olfactory neurons was seen at any stage examined. No staining associated with the glomerular layer of the olfactory bulb was seen prior to, or at birth. A

weak staining of the glomerular layer was first observed starting at P2, associated with some cells and very lightly stained glomeruli. NGF-receptorimmunopositive cells were of 7 9 ~tm diameter, appeared to be stained perimembranally and manifested short processes (Fig. 2A). A progressive increase in the staining intensity and in the number of NGF-receptor-immunopositive glomeruli was then observed in the period from P2-PI0. At P10 N G F - r e c e p t o r immunostaining was present in almost all glomeruli although of varying intensity (Fig. 2B). In 2-month-old rats NGF-receptor staining associated with the glomerular layer was similar to that observed in the adult animal. This

56

S. Biffo, E. Marti and A. Fasolo

Fig. 3. Tyrosine hydroxylase (TH) mRNA and protein ontogeny. (A,B) Show respectively the pattern of in situ and immunohistochemical labelling of the olfactory bulb at birth. Sections are 100 fun apart. (C) 2-week-old olfactory bulb. Several TH mRNA-expressing periglomerular neurons are visible. (D) Dark field image of a postnatal day I olfactory bulb (left side) compared to a 2-week old olfactory bulb (right side). Note the increase in the number ofTH mRNA-expressing neurons. Abbreviations: FL, fibre layer; GL, glomerular layer; EPL, external plexiform layer. Scale bars are 25 Ixm for (A,B,C) and 100 lam for (D).

staining was observed both in the main and in the accessory olfactory bulb (Fig. 2C). NGF-receptorimmunopositive processes were not observed in the fibre layer where axons of olfactory neurons are present. Tyrosine hydroxylase

Few TH-expressing cells were seen as early as E20. At birth both in situ hybridization and immunohistochemistry revealed several TH-expressing neurons (size: 11-16 lam) always located in proximity to the developing gtomerular layer (Fig. 3A,B). TH-expressing neurons were found either in small clusters or occasionally isolated. TH-immunopositive fibres were also present in the glomerular layer of the neonatal olfactory bulb. During postnatal development (P1-P7, P14, 2-month-old and adult) a massive increase in the number of neurons expressing both TH mRNA (Fig. 3C,D) and protein (not shown) was observed. The increase in positive neurons was evident as an increase in the number of

positive cells per glomerulus and in the number of glomeruli.

Double stainings Starting from P2 and very evidently from P 10 and in the adult animal, NGF-receptor and TH immunostaining were both found in the glomeruli. However, glomeruli exhibiting high expression of N G F receptor could present either high or low levels of TH immunostaining and vice versa. In addition, within the same glomerulus, NGF-receptor and TH immunostaining were never seen colocalized in processes in either early (Fig. 4A, B) or late postnatal (not shown) development. The perikarya containing TH immunoreactivity never had NGFreceptor immunostaining nor did the few scattered NGF-receptor-positive cells ever present TH immunoreactivity. Carnosine and NGF-receptor immunoreactivities were both evident in almost all the glomeruli, although glomerular NGF-receptor staining was

NGFr/TH/carnosinein Olfactory Development

F¸

~

~

~

~

~

57

~

J

77!

~y

4C Fig. 4. Nerve growth factor (NGF) receptor, tyrosine hydroxylase (TH) and carnosine immunoreactivity are not colocalized. Arrowheads indicate equivalent spots in the section. (A,B) Two-week-old olfactory bulb: an example of immunolocalization of TH (right side) and NGF receptor (left side). Sections 18 p_mthick treated as free floating. Note that although the two patterns appear in the same region of the olfactory bulb, they are clearly mutally exclusive. (C,D) Show respectively NGF receptor and carnosine immunoreactivity in a 2-week-old olfactory bulb. Note the presence of a carnosine-immunopositive glomerulus in a zone completely devoid of NGF-receptor staining (arrowhead). Abbreviations: gl, glomerular layer; epl, external plexiform layer; fl, fibre layer. Scale bar is 25 p_m.

strikingly seen in carnosine-negative glomeruli (Fig. 4C,D). No colocalization of carnosine and NGF receptor was ever observed. DISCUSSION We have studied the patterns of development and codistribution of carnosine, NGF receptor and TH during the ontogeny and adulthood of the rat olfactory system. This study proves the different cellular localization of all of these molecules: in particular, it rules out the expression of NGF receptor in TH-expressing periglomerular neurons and details the developmental time-course of NGFreceptor expression in the olfactory bulb. Furthermore, it gives the first immunohistochemical

evidence of carnosine expression during embryonic development and confirms at the mRNA level previous data on the early expression of TH protein. The specificity for all the antisera used in this study has been thoroughly documented (TH: Thibault et al., 1981; NGF receptor: Yan and Johnson, 1989; carnosine: Biffo et al., 1990b). Control reactions performed in our hands gave resuits consistent with specific staining. The controls used for the in situ hybridization were the canonical ones (use of an excess of unlabelled oligo, or pretreatment of sections with RNase A) and confirmed the specificity of the probe. The similarity of the data obtained by in situ hybridization and immunohistochemical staining is further proof of the reliability of the reagents used here. The results

58

S. Biffo, E. Marti and A. Fasolo

obtained with double-staining methods were in agreement with single-staining methods and all the specific controls performed (exchange of the order of reactions, omission of the one primary antibody) indicated absence of cross-reacting secondary antibodies and lack of interference between the reactions.

Cellular iocafization of NGF receptor Previous studies have associated NGF-receptor expression with the olfactory bulb of the adult rat: Pioro and Cuello (1990) have described NGFreceptor immunostaining in mitral cells (only following colchicine injection) and aggregates of reaction product that could have represented periglomerular cell bodies, whereas a previous report identified NGF-receptor-immunopositive cells as periglomerular neurons (Gomez-Pinilla et al., 1987). Our study clearly shows the presence of NGF-receptor-positive cells in the glomerular layer of the developing rat olfactory bulb: these cells are barely visible in the adult olfactory bulb since they are masked by the strong NGF-receptor immunoreactivity associated with a network of processes. These cells are perimembranally stained, have relatively thick processes, a cell body size of 7-9 Ixm, and are immunoreactive beginning from the second day of postnatal life. In contrast, no mitral neurons were ever seen immunopositive to N G F receptor during both development and adulthood. At least two hypotheses can be considered to explain the absence of NGF-receptor staining in mitral cells, in our study. The first is that NGFreceptor staining associated with mitral neurons might be due to an artifact deriving from colchicine injection; for instance, colchicine, as well as other stressful stimuli, has been shown to modify gene expression (Cortes et al., 1990). An indirect support to the absence of N G F receptor in the mitral cells of normal animals comes from the fact that the earliest NGF-receptor immunoreactivity in the glomerular layer is delayed with respect to the maturation of mitral cell dendrites and is morphologically dissimilar to the mitral branchings (Pomeroy et al., 1990). The second is that the absence of NGF-receptor immunoreactivity in mitral cell bodies is due to the rapid transport of this protein to the dendritic fields, and the late expression of N G F receptor during development may be related to the low sensitivity of immunohistochemical techniques. Highly sensitive in situ hybridization for N G F receptor could be an effective tool to determine whether mitral cells truly express N G F receptor. In a previous report, NGF-receptor immunoreactivity was linked to the olfactory nerve (Yan and Johnson, 1988). In contrast, our data clearly show the absence of NGF-receptor immunoreactivity from primary olfactory neurons (carnosine

immunopositive), indicating that NGF receptor is not present in the primary olfactory pathway. The appearance of NGF-receptor-immunopositive cells in the developing glomerular layer was puzzling since there is one morphologically homogenous and largely predominant class of small interneurons present in this area, i.e. periglomerular neurons. Most periglomerular neurons express either gamma-aminobutyric acid (GABA) or coexpress TH and GABA (Halasz and Shepherd, 1983; Gall et al., 1987; Baker et al., 1988). It was therefore of great interest to evaluate whether TH-immunopositive neurons were also NGFreceptor immunopositive. The present in situ hybridization and immunohistochemical data show the late expression of N G F receptor as compared to TH in the olfractory bulb and the different morphological appearance between TH and NGF-receptorpositive cells. Although the differential times of expression of N G F receptor and TH as well the different morphology between N G F receptor and THpositive cells seemed to indicate no overlapping between these two cell populations, we could not definitely exclude that at least some TH periglomerular neurons were NGF-receptor positive due to the possible differential sensitivity for the two immunocytochemical reactions. We therefore performed double-staining experiments, thus definitely ruling out the presence of NGF receptor in TH-expressing periglomerular neurons. All these data, taken together, point to two alternative explanations for the cellular origin of NGF-receptor staining in the glomerular layer of the olfactory bulb, i.e. presence in a set of glial cells and/or contribution to the staining by the cholinergic input. These two interpretations are not mutually exclusive. In favour of the first are: the presence of immunopositive small-sized cells in the developing olfactory bulb, the fact that these cells are not periglomerular neurons, the nature of the antigen recognized by the monoclonat used in this study (see below). In addition, several glial cell types are present in the glomerular and fibre cell layer, including one peculiar to the olfactory system (ensheathing cells: for review, see Doucette, 1990). These ensheathing cells have phenotypic features of both Schwann cells and astrocytes and are morphologically similar to NGF-receptor-immunopositive cells. However, the full identification of the glial subtype expressing N G F receptor requires simultaneous double-staining for N G F receptor and glial markers, in vivo. The only marker that we found suitable for this purpose, anti-GFAP (a marker for central nervous system (CNS) astrocytes), identified a rich population of astrocytes in all layers of the olfactory bulb, but it did not stain consistently NGF-receptor-positive cells (S. Biffo, unpublished data). A second source for NGF-receptor immunostaining in the glomerular layer of the olfactory bulb could be related to the cholinergic afferent input to the olfactory bulb (Macrides et al., 1981). However,

NGFr/TH/carnosinein Olfactory Development factors against this interpretation are: (1) the evidence that choline acetyltransferase immunoreactivity, a marker for cholinergic fibres, is much more restricted than NGF-receptor immunoreactivity (Pioro and Cuello, 1990) and, (2) the fact that conversely to the NGF-receptor staining, the cholinergic input is inter-glomerular and not intraglomerular (Shepherd and Greer, 1990). One interesting feature of NGF-receptor staining in the developing glomerular layer of the olfactory bulb is the heterogeneity of staining between different glomeruli: since the olfactory system undergoes a conspicuous postnatal maturation and a continual renewal (Graziadei, 1990), this pattern may be related to the functional status of the single glomeruli (i.e. regenerating or stable) and deserves further analysis.

NGF-receptor epitope Although the trophic effects of NGF on neuronal survival in the CNS are apparently limited to the cholinergiclneurons (Mobley et al., 1985), the rat NGF receptor (Johnson et al., 1986; Radeke et al., 1987) in the CNS is not exclusively localized to cholinergic structures (see, for instance, Taniuchi et al., 1986; Kiss et al., 1988; Yan and Johnson, 1988, 1989; Pioro and Cuello, 1900). NGF receptor exists in two forms pharmacologically defined as high and low affinity (Sutter et al., 1979). Cloning data have shown that the low and high affinity NGF receptor are two separate molecules and that the trk proto-oncogene is the putative high affinity form (Radeke et al., 1987; Kaplan et al., 1991). The biological effects of NGF appear to be mediated by the high affinity form (Green and Greene, 1986), although several roles have also been proposed for the low affinity form (DiStefano and Johnson, 1988), since, for instance, high affinity NGF-receptor binding may require the coexpression of both the trk proto-oncogene and the low affinity NGF receptor (Hempstead et al., 1991). The latter is generally thought to be expressed also in non-neuronal cells (Taniuchi et al., 1988). The anti-NGF receptor used in this study (Mab 192) seems to recognize predominantly the low affinity NGF receptor, although this antibody, when injected in developing rats, mimics NGF deprivation (Johnson et al., 1983). Autoradiography binding data for the high affinity receptor using labelled NGF clearly indicate a more restricted distribution than with Mab 192 (Raivich and Kreutzberg, 1987): accordingly, previous papers have shown that Mab 192 recognizes an epitope present in a subset of microglial cells in the rat neural lobe (Yan et al., 1990), as well in nonneuronal mesenchymal derivatives (Byiers et al., 1990). Taken together, these data seem to imply that the immunoreactivity observed in the olfactory bulb

59

with Mab 192 is conceivably due to the low affinity NGF receptor.

TH and carnosine ontogeny in the olfactory bulb With regard to the onset of TH expression in periglomerular neurons, this study largely confirms at the mRNA level the previous results of McLean and Shipley (1988) at the protein level. The two developmental patterns revealed by in situ hybridization and immunohistochemistry were substantially identical and confirm the notion that although periglomerular neurons are already generated during prenatal development in the ventricular layer of the olfactory bulb (Bayer, 1983), they need to arrive in the proximity of the glomerular layer to begin TH mRNA (this study) and protein (McLean and Shipley, 1988) expression. Since in the adult animal the regulation of dopamine expression in periglomerular neurons relies on the synaptic input by olfactory neurons (Baker, 1990) and it is regulated by a specific effect on TH mRNA levels (Erlich et al., 1990), our data suggest that also during development the beginning of TH mRNA expression is correlated with the contact of periglomerular neurons with olfactory axons. Previous biochemical studies detected carnosine synthetase activity in the nasal region from El6, although carnosine can be synthesized earlier if beta-alanine is exogenously supplied (Margolis et al., 1985). However, since the sensitivity of biochemical analyses may be inadequate to reveal the presence of carnosine expression in small numbers of scattered olfactory neurons, a direct anatomical analysis was needed. We have shown that carnosine-positive neurons appear in the nasal cavity between El5 and El7. We have previously shown, in the mouse, the expression of B 50/ GAP43, a marker for axonal growth, at El 1.5 (Biffo et al., 1990c); this stage corresponds approximately to El2 in the rat. Furthermore, B-50/GAP43 staining in the El5 rat olfactory epithelium is already very extensive (S. Biffo and J. Verhaagen, unpublished data). The fact that the embryonic development of carnosine immunoreactivity occurs later than B-50/GAP43 immunoreactivity indicates that during development the ability of olfactory neurons to express carnosine represents a late stage of cellular differentiation. Since in one of the El5 embryos analysed, few immunoreactive carnosine neurons were seen, it is conceivable that carnosine synthesis is normally beginning between El5 and El6, as reported earlier (Margolis et al., 1985). This result is very similar to that observed for OMP ontogeny, a marker for terminally differentiated olfactory neurons (Farbman and Margolis, 1980; Allen and Akeson, 1985). Our data thus support the view that carnosine is a marker for olfactory mature neurons. So far, nothing is known about the events that trigger carnosine expression.

60

S. Biffo, E. Marti and A. Fasolo

ACKNOWLEDGEMENTS We are indebted to Dr Frank Margolis (Roche Institute of Molecular Biology) for helpful discussions and useful suggestions during this study, and to Dr Aurora Battaglia for her technical support. We also thank Dr Antonio Bertolotto (Clinica Neurologica) for having shared with us his invaluableexperience in glial cell classification. This work was supported by CNR grants. Elisa Marti was recipient of a postdoctoral fellowship from the Spanish Ministry of Education (MEC EX90).

REFERENCES Allen, W. K. and Akeson, R. (1985). Identification of an olfactory receptor neuron subclass: cellular and molecular analysis during development. Devel. Biol. 109, 393-401. Artero, C., Mulatero, B., Biffo, S., Margolis, F. L., Go zzo, S. and Fasolo, A. (1991). Distribution of carnosine like immunoreactivity in the brain of the crested newt. Brain Behav. Evolut. 37, 168-178. Baker, H., Kawano, T., Margolis, F. and Joh, T. H. (1983). Transsynaptic regulation of tyrosine hydroxylase expression in olfactory bulb of mouse and rat. J. Neurosci. 3, 69-78. Baker, H., Towle, A. and Margolis, F. L. (1988). Differential afferent regulation of dopaminergic and GABAergic neurons in the mouse olfactory bulb. Brain Res. 450, 69-80. Baker, H. (1990). Unilateral, neonatal olfactory deprivation alters tyrosine hydroxylase expression but not aromatic amino acid decarboxylase or GABA immunoreactivity. Neuroscience 36, 761-77 I. Bayer, S. (1983). 3-H Thymidine radiographic studies of neurogenesis in the rat olfactory bulb. Exp. Brain. Res. 50, 329-340. Biffo, S., De Lucia, R., Mulatero, B., Margolis, F. and Fasolo, A. (1990a). Carnosine-, caleitonin gene-related peptide- and tyrosine hydroxylase-immunoreactivity in the mouse olfactory bulb following peripheral denervation. Brain Res. 528, 353-357. Biffo, S., Grillo, M. and Margolis, F. (1990b). Cellular localization of carnosine-like immunoreactivities in rodent and avian central nervous system. Neuroscience 35, 637-651. Biffo, S., Verhaagen, J., Schrama, L. H., Schotman, P., Danho, W. and Margolis, F. (1990c). B-50/GAP43 expression correlates with process outgrowth in the embryonic mouse nervous system. Eur. J. Neurosci. 2, 487-499. Brunjes, P. and Frazier, L. (1986). Maturation and plasticity in the olfactory system of vertebrates. Brain Res. Rev. 11, 1-45. Byiers, M. R., Schatteman, G.C. and Bothwell, M. (1990). Multiple functions for NGF receptor in developing, aging and injured rat teeth are suggested by epithelial, mesenchymal and neural immunoreactivity. Development 109, 461-471. Chandler, C., Parsons, L., Hosang, M. and Shooter, E. (1984). A monoclonal antibody modulates the interaction of nerve growth factor with PC12 cells. J. Biol. Chem. 259, 6882-6889. Cortes, R., CeccateUi, S., Schalling, M. and Hokfelt, T. (1990). Differential effects of intracerebroventricular colchicine administration on the expression ofmRNAs

tbr neuropeptides and neurotransmitter enzymes with special emphasis on galanin: an in situ hybridization study. Synapse 6, 369-391. Costanzo, R. M. and Morrison, E. E. (1989). Threedimensional scanning electron microscopy study of the normal hamster olfactory epithelium. J. Neurocyt. 18, 381-391. Cuschieri, A. and Bannister, L. H. (1975). The development of the olfactory mucosa in the mouse: light microscopy. J. Anat. 119, 277-286. DiStefano, P. S. and Johnson, M. (1988). Identification of a truncated form of the nerve growth factor receptor. Proc. Natl. Acad. Sci. USA 85, 270-274. Douchette, R. (1990). Glial influences on axonal growth in the primary olfactory system. Glia 3, 433-449. Ehrlich, M. E., Grillo, M., Joh, T., Margolis, F. and Baker, H. (1990). Transneuronal regulation of neuronal specific gene expression in the mouse olfactory bulb. Mol. Brain Res. 7, 115-122. Eschenfeldt, W., Puskas, R. and Berger, S. (1987). Homopolymeric tailing. In Methods in Enzymology, Vol. 152, pp. 337-342. Academic Press, London. Farbman, A. I. and Margolis, F. L. (1980). Olfactory marker protein during ontogeny: immunohistochemical localization. Dev. Biol. 74, 205-215. Gall, C., Hendry, S. H., Seroogy, K., Jones, E. and Haycock, J. W. (1987). Evidence for coexistence of GABA and dopamine in neurons of the rat olfactory bulb. J. Comp. Neurol. 266, 307-318. Gomez-Pinilla, F., Cotman C. and Nieto-Sampedro, M. (1987). NGF-receptor in rat brain: topographical distribution and response to entorhinal ablation. Neurosci. Lett. 82, 260-266. Gomez-Pinilla, F., Guthrie, K., Leon, M. and NietoSampedro, M. (1988). NGF receptor increase in the olfactory bulb of the rat after early odor deprivation. 48, 161-165. Graziadei, P. P. C. (1973). The ultrastructure of the vertebrate olfactory mucosa. In The Ultrastructure o f Sensory Organs, pp. 267-305. Elsevier, Oxford. Graziadei, P. C. and Samanen, D. W. (1980), Ectopic glomerular structures in the olfactory bulb of neonatal and adult mice. Brain Res. 187, 467-472. Graziadei, P. P. C. and Monti-Graziadei, G. A. (1986). Neuronal changes in the forebrain of mice following penetration by regenerating olfactory axons. J. Comp. Neurol. 247, 344-356. Graziadei, P. P. (1990). Olfactory development. In Development o f Sensory Systems in Mammals (ed. Coleman, R.), pp. 519-566. John Wiley & Sons, New York. Green, S. and Greene, I. (1986). A single Mr. 103000 ~25IB-NGF affinity labeled species represents both the low and the high affinity forms of the nerve growth factor receptor. J. Biol. Chem. 261, 15316-153261 Grima, B., Lamoroux, A., Blanot, F., Biguet, N. F. and Mallet, J. (1985). Complete coding sequence of rat tyrosine hydroxylase mRNA. Proc. Natl. Acad. Sci. USA 82, 617-621. Halasz, N. and Shepherd, G. M. (1983). Neurochemistry of the vertebrate olfactory bulb. Neuroscience 10, 579-619. Hempstead, B., Martin-Zanca, D., Kaplan, D., Parada, L. and Chao, M. (1991). High affinity NGF binding requires coexpression of the trk proto-oncogene and the low affinity NGF receptor. Nature 350, 678-682.

N G F r / T H / c a r n o s i n e in Olfactory Development Johnson, D., Lanahan, A., Buck, R., Seghal, A., Morgan, C., Mercer, E., Bothwell, M. and Chao, M. (1986). Expression and structure of the human NGF-receptor. Cell47, 547 554. Johnson, E. M., Osborne, P. A. and Taniuchi, M, (1989). Destruction of sympathetic and sensory neurons in the developing rat by a monoclonal antibody against the NGF-receptor. Brain Res. 478, 166-170. Kaplan, D., Martin-Zanca, D. and Parada, L. (1991), Tyrosine phosphorylation and tyrosine kinase activity of the trk proto-oncogene product induced by NGF, Nature 350, 158-160. Kawano, T. and Margolis, F. L. (1982). Transsynaptic regulation of olfactory bulb catecholamines in mice and rats. J. Neurochem. 39, 342-348. Kiss, J., McGovern J. and Patel, A. J. (1988). lmmunohistochemical localization of cells containing nerve growth factor receptors in the different regions of the adult forebrain. Neuroscience 27, 731 748. Lcvi-Montalcini, R. (1987). The nerve growth factor: thirty five years later. Eur. Molec. Biol. Org. J. 6, 1145- 1154. Macrides, R., Davis, B., Youngs, W., Nadi, S. and Margolis, F. L. (1981). Cholinergic and catecholaminergic afferents to the olfactory bulb in the hamster: a neuroanatomical, biochemical and histochemical investigation. J, Comp. Neurol. 203, 495 514. Margolis, F. L. (1972). A brain protein unique to the olfactory bulb. Proc. Natl. Acad. Sci. USA 69, 1221 1224. Margolis, F. L. (1974). Carnosine in the primary olfactory pathway. Science 184, 909 911. Margolis, F. L., Roberts, N,, Ferriero, D. and Feldman, J. (1974). Denervation in the primary olfactory pathway of mice: biochemical and morphological features. Brain Res. 81,469 483. Margolis, F. L., Grillo, M., Kawano, T. and Farbman, A. I. (1985). Carnosine synthesis in olfactory tissue during ontogeny: effect of exogeneous beta-alanine. J. Neurochem. 44, 1459 1464. Margolis, F. and Getchell, T. (1988). Molecular Neurobiology ~/ the O[/'actory System. Plenum Publishing Corporation. McLean, J. and Shipley, M. T. (1988). Postmitotic, postmigrational expression of tyrosine hydroxylase in olfactory bulb dopaminergic neurons. J. Neurosci. 8, 3658-3669. Menco, B. P, and Farbman, A. I. (1985). Genesis of cilia and microvilli ot" rat nasal epithelia during prenatal development. 1. Olfactory epithelium, qualitative studies. J. Cell Sci. 78, 283-310. Miragall, F. and Monti-Graziadei, G. A. (1982). Experimental studies on the olfactory marker protein. II Appearance of the olfactory marker protein during differentiation of the olfactory sensory neurons of mouse: an immunohistochemical study. Brain Res. 239, 245 -250. Miragall, F., Kadmon, G., Husmann, M. and Schachner, M. (1988). Expression of cell adhesion molecules in the olfactory system of the adult mouse: presence of the embryonic form of N-CAM. Devel. Biol. 129, 516 531. Mobley, W., Rutkowsky, J. L., Tennekoon, K., Buchanan, K. and Johnston, M. (1985). Choline acetyltransferase activity in striatum of neonatal rats increased by nerve growth factor. Science 229,284 287. Monti-Graziadei, G. A., Stanley, R. S. and Graziadei, P. P. C. (1980). The olfactory marker protein in the

61

olfactory system of mouse during development. Neuroscience fi, 1239-1252. Neidle, A. and Kandera, J. (1974). Carnosine: an olfactory bulb peptide. Brain Res. 80, 359 364. Pioro, E. P. and Cuello, A. (1990). Distribution of nerve growth factor receptor like immunoreactivity in the adult rat central nervous system. Effect of colchicine and correlation with the cholinergic system 1. Forebrain. Neuroscience 34, 57 87. Pomeroy, S. L., La-Mantia, A. S. and Purves, D. (1990). Postnatal construction of neural circuitry in the mouse olfactory bulb. J. Neurosci. 10, 1952 1966. Radeke, M., Misko, T., Hsu, C., Herzenberg, L. and Shooter, E. (1987). Gene transfer and molecular cloning of the rat nerve growth factor receptor. Nature 325, 593-597, Raivich, G. and Kreutzberg, G. W. (1987). The localization and distribution of high affinity beta-nerve growth factor binding sites in the central nervous system of the adult rat. A light microscopy autoradiographic study using 125-I-beta nerve growth factor. Neuroscience 20, 23 -36. Rugh, R. (1968). The Mouse. lts Reproduction and Development. Burgess Publishing Company, Minneapolis. Sakai, M., Yoshida, M., Karasawa, N., Teramura, M., Ueda, H. and Nagatsu, I. (1987). Carnosine like immunoreactivity in the primary olfactory neuron of the rat. Experientia 43, 298 300. Shepherd, G. and Greet, C. (1990). Olfactory bulb. In The Synaptic Organization qf the Brain (ed. Shepherd, G.), pp. 132 170. Oxford University Press, New York. Sutter, A., Riopelle, R. J., Harris-Warrick, R. and Shooter, E. M. (1979). Nerve growth factor receptors: characterization of two distinct classes of binding sites on chick embryo sensory ganglia. J. Biol. Chem. 254, 5972 5982. Taniuchi, M., Schweitzer, J. B. and Johnson, E. M. (1986). Nerve growth factor receptor molecules in rat brain. Proc. Natl. Acad. Sci. USA 83, 1950-1954. Taniuchi, M., Clark, H., Schweitzer, J. and Johnson, E. (1988). Expression of nerve growth factor receptors by Schwann cells of axotomized peripheral nerves; ultrastructural location, suppression by axonal contact and binding properties. J. Neurosci. 8, 664-681. Thibault, A., Vidal, D. and Gros, F. (1981). In vitro translation of mRNA from rat pheochromocytoma tumors, characterization of tyrosine hydroxylase. Biochem. Biophys. Res. Commun. 93, 96 98. Verhaagen, J., Oestreicher, A. B., Gispen, W. H. and Margolis, F. L. (1989). The expression of the growth associated protein B-50/GAP43 in the olfactory system of neonatal and adult rats. J. Neurosci. 9, 683 691. Verhaagen, J., Greer, C. and Margolis, F. L. (1990a) B-50/GAP43 gene expression in the rat olfactory system during postnatal development and aging. Eur. J. Neurosci. 2, 397 407. Verhaagen, J., Oestreicher, A. B., Grillo, M., KhewGoodall, Y.-S., Gispen, W. H. and Margolis, F. L. (1990b). Neuroplasticity in the olfactory system: differential effects of central peripheral lesion of primary olfactory pathway on the expression of B-50/GAP43 and the olfactory marker protein. J. Neurosci. Res. 26, 31 44. Weskamp, G. and Reichardt, L. F. (1991). Evidence that biological activity of N G F is mediated through a novel subclass of high affinity receptors. Neuron 6, 649 663.

62

S. Biffo, E. Marti and A. Fasolo

Williams, R. and Rush, R. A. (1988). Electron microscopic immunocytochemical localization of nerve growth factor in developing mouse olfactory neurons. Brain Res. 463, 21-27. Yan, Q. and Johnson, E. M. (1988). An immunohistochemical study of the nerve growth factor receptor in developing rats. J. Neurosci. 8, 3481-3498. Yan, Q. and Johnson, E. M. (1989). Immunohisto-

chemical localization and biochemical characterization of nerve growth factor receptor in adult rat brain. J. Comp. Neurol. 290, 585-598. Yan, Q., Clark, B. and Johnson, E. M. (1990). Nerve growth factor receptor in neural lobe of rat pituitary gland: immunohistochemical localization, biochemical characterization and regulation. J. Neurocyt. 19, 302-312.

Accepted25 June 1991