BackgroundGrowth cone–associated protein (GAP43) is found in growing axons and we hypothesized that systemic treatment with antineoplastic agents should disrupt regeneration of olfactory receptor cells. Disruption of regeneration should be evidenced by decreased presence of growing axons in the olfactory bulb.ObjectiveTo evaluate GAP43 in human olfactory bulb in normal controls and in individuals receiving treatment for neoplasms.DesignImmunocytochemical studies were performed on autopsied human olfactory bulbs to identify both GAP43 and olfactory marker protein immunoreactivity. The former recognizes growing axons and the latter is a definitive marker of adult olfactory nerve.SubjectsTwenty-seven subjects were evaluated. Seven had received either antineoplastic agents and/or x-irradiation of the whole head. Four subjects were young, untreated controls, 10 were age matched to the treated group, and 2 had neoplasms but did not receive antineoplastic agents or irradiation of the head. In addition, 3 subjects with end-stage renal disease were immunostained.ResultsSubjects treated with antineoplastic agents or x-irradiation of the whole head displayed no statistically significant loss of olfactory bulb glomeruli, but GAP43 immunoreactivity was markedly reduced in all but 1 subject (P<.32). The subjects with end-stage kidney disease showed frank loss of both GAP43 immunoreactivity and olfactory glomeruli.ConclusionsTreatment with antineoplastic agents apparently does not damage olfactory epithelium directly but inhibits growth of new axons into the olfactory bulb. This observation suggests that the quality of olfactory experience may change during the course of treatment with antineoplastic agents because the olfactory nerve is not replaced.THE RECEPTOR cells of the olfactory mucosa are regenerated throughout life from basal cells.Regenerated olfactory receptors grow an axon that terminates in the olfactory bulb glomeruli. Growing axons contain growth cone–associated protein (B50/GAP43).We hypothesized that receptor regeneration should be inhibited by either systemic antimitotic agents or irradiation that included the olfactory epithelium. Inhibition of regeneration should be reflected by decreased GAP43 expression in olfactory bulb glomeruli.To test this hypothesis, we evaluated the olfactory bulbs from 10 age-matched controls and 7 subjects who had received antineoplastic agents or x-irradiation of the whole head. Compared with the control group, GAP43 staining around glomeruli was significantly decreased in treated subjects . No unequivocal evidence of loss of glomeruli in the same subjects was observed. These data suggest that inhibition of regeneration of olfactory receptors may occur during antineoplastic therapy.SUBJECTS AND METHODSOlfactory bulbs were collected at autopsy from 26 subjects. Seven subjects received chemotherapy and/or x-irradiation of the whole head (Table 1). Different regimens and antineoplastic agents had been used; however, the interval between the last cycle and autopsy was 5 months or less with the exception of 1 subject who underwent autopsy 13 months after treatment. Nineteen untreated subjects were evaluated. Four subjects ranged in age from 1 to 26 years at death. Ten controls were age matched to the 7 subjects from the antineoplastic and/or irradiation group (Table 1). Age-matched controls did not have a history (by chart review) of exposure to toxins or medical conditions that might disrupt olfactory function. Five additional subjects were evaluated; 3 subjects with end-stage renal disease, which is reported to disrupt olfactory function,and 2 who had received body-only x-irradiation.Subject Characteristics*See table graphicSections were cut at 10 µm, either in a transaxial plane (perpendicular to the long axis) or parallel to the long axis of the olfactory bulb. Sections were mounted on chrome alum-subbed slides and stained with cresyl violet combined with periodic acid–Schiff (PAS) and luxol fast blue. For immunocytochemistry, sections were deparaffinized, incubated with 0.5% hydrogen peroxide in 50% methanol, and then with 5% nonfat dry milk (to decrease background) and 10% Triton×100 in 0.1 mol/L of phosphate-buffered solution (pH 7.4) for 1 hour each. They were then incubated with antisera to either olfactory marker protein (OMP) or GAP43. Olfactory marker protein, which immunostains olfactory nerves,was a goat-raised antisera used at a 1:400 dilution. Growth cone–associated protein was a rabbit antisera used at 1:250 dilution.The secondary antibody was the appropriate species antisera (1:100; Cappel-ICN, Costa Mesa, Calif) in buffer, 0.1% Triton×100 and dried milk applied for 1 hour, followed by 1:100 peroxidase-antiperoxidase complex (Sternberger-Meyer, Baltimore, Md) of the appropriate species for 1 hour. Sections were developed with diaminobenzidine tetrahydrochloride (0.05%) and hydrogen peroxide (0.01%) for 15 to 30 minutes. Control sections for antisera specificity were incubated with normal serum (in place of the primary sera) that eliminated staining.The antineoplastic- and/or head irradiation–treated patients were compared with the 10 age-matched controls. Glomeruli ratings were performed on the cresyl violet combined with PAS and luxol fast blue sections to obtain an estimate of whether treatment of neoplasms reduced the number of glomeruli. Three to 4 sections through the extent of the olfactory bulb were rated by an experimenter blinded to the treatment condition. The PAS-stained glomeruli were rated as either normal (N) or of decreased (D) density within the glomerular zone when compared with both younger and age-matched controls. Growth cone–associated protein immunostaining in olfactory nerve was rated as absent (0), weak (1), or present in normal amounts (2) compared with 4 subjects younger than 26 years and the control group. Statistical significance was tested by χ2statistic for independent samples comparing the age-matched group with the chemotherapy- and/or head irradiation–treated group.RESULTSComparison of the cresyl violet combined with PAS and luxol fast blue–stained sections with the OMP-immunostained sections confirmed that the PAS-stained, acellular structures contained OMP immunoreactivity, which identifies these structures as glomeruli (Figure 1). Therefore, we easily identified glomeruli with the PAS stain. χ2Testing of PAS-stained glomeruli disclosed no statistical difference between the controls and chemotherapy-exposed patients (P<.32).Figure 1.Glomerular staining in the olfactory bulb. A, The typical glomeruli (G) staining seen with combined cresyl violet and periodic acid–Schiff staining. Glomeruli are clearly obvious between the olfactory nerve layer (ON) and the external plexiform layer (EPL). B, Olfactory marker protein immunoreactivity is shown in G. In addition, a bundle of olfactory marker protein immunoreactive nerves is obvious (arrowheads). The olfactory marker protein sections were counterstained with thionin for identifying laminae (B: bar, 100 µm).In all normal controls GAP43 immunoreactivity was found both in the granule cell layer and in axons in the olfactory nerve and around glomeruli (Figure 2, A and B). This pattern is comparable with that described for the rodent olfactory bulb.In contrast to glomerular number, density of the nerve and glomerular immunoreactivity was markedly decreased in the treated patients (Figure 2, C). This difference was significant (χ22,10.19; P<.006). In the chemotherapy cases we also noted glomeruli (identified with both OMP and PAS) located in the external plexiform and, in 1 subject, granule cell layer (not shown). In the 3 subjects with kidney disease, GAP43, OMP, or PAS staining of glomeruli was conspicuously decreased (1 subject) or absent (2 subjects). Body-only radiation treatment did not appear to have an effect on either the distribution and number of glomeruli or GAP43 immunoreactivity.Figure 2.Growth cone–associated protein (GAP43) immunoreactivity in a control (untreated) olfactory bulb. A, Normal GAP43 immunoreactivity in the human olfactory bulb is comparable with immunoreactivity described in rodents. The glomerular layer (G) shows GAP43 immunoreactivity around glomeruli and in the olfactory nerve at the periphery of the olfactory bulb. The external plexiform layer (EPL), between the glomerular zone and the mitral cell layer (M), is less intensely immunoreactive than either the glomerular zone or the granule cell layer (GCL). The anterior olfactory nucleus (AON) is also well labeled (arrow points to the glomerulus shown in Figure 2, B; A: bar, 500 µm). B (arrow in A), A glomerulus with GAP43 immunoreactivity at the periphery of the glomerulus. A second, smaller glomerulus is just to the left of the labeled glomerulus. The olfactory nerve (ON) is GAP43 immunoreactive. C, A subject with a score of 0 on glomerular GAP43 staining. The external limit of the bulb (arrowheads) is not associated with GAP43 immunostaining of olfactory nerve as seen in A or B. The glomerulus (G) is not associated with GAP43 immunoreactivity. Growth cone–associated protein immunoreactivity is in small processes diffusely distributed in this subject. To increase contrast for photomicrography, the substage condenser was stopped-down.COMMENTOur observations of OMP and GAP43 immunostaining in the human olfactory bulb are compatible with previous observations of the rodent olfactory bulb. The OMP staining of the human olfactory bulb nerve and glomeruli has been previously described and our results are similar.The distribution of GAP43 immunoreactivity in the olfactory bulb has been previously described for rat and mouse.Both the granule cell and glomerular layer of the bulb are immunoreactive for GAP43. Growth cone–associated protein immunoreactivity is predominantly in the nerve or periphery of the glomerulus that is consistent with GAP43 in growing axons.This study confirms the comparability between the rodent and human olfactory bulb.We anticipated that treatments designed to disrupt cell regeneration should have effects on olfactory mucosa and this hypothesis was supported. Growth cone–associated protein immunostaining in the olfactory nerve and around glomeruli decreased in subjects who had received either antineoplastic chemotherapy or irradiation of the whole head. It is noteworthy that the only subject who underwent chemotherapy with a normal pattern of GAP43 staining was last treated 13 months before autopsy.However, a decrease in GAP43 staining was not associated with a frank loss of glomeruli. Although there was a suggestion of glomerular decrease, it was not statistically significant. This observation suggests that chemotherapy or irradiation may not be toxic to adult olfactory receptors but the glomeruli number might decrease secondary to disruption of neurogenesis in olfactory epithelium. Neither GAP43 nor OMP immunoreactivity was notably different from controls in 2 patients who received radiation treatment, suggesting that neither neoplastic disease nor the free radicals proposed to be liberated in radiation treatment cause olfactory nerve or bulb damage. However, this sample size was small and subtle abnormalities could be missed. We noted that glomeruli were found in laminae that did not normally contain glomeruli. The cause of heterotopic glomeruli is not clear, but it could be that antineoplastic treatments disrupt processes that maintain normal lamination in the olfactory bulb. Finally, renal disease resulted in obvious damage, including frank loss of glomeruli, which is complementary to previous observations of olfactory dysfunction in patients with renal disease.These observations suggest that some clinical olfactory abnormalities noted in cancer therapycould be related to disruption of regeneration of olfactory receptors. Our data also suggest that mature fibers (as represented by OMP-positive glomeruli) are not grossly affected by treatment; hence, relative preservation of olfactory ability will probably still exist. Clinical abnormalities of flavor, subsequent to antineoplastic chemotherapy regimens, may represent disruption of normal replacement of olfactory receptor neurons. One obvious prediction is that the perceived quality of an odorant may change as regeneration of fibers is inhibited.PPCGraziadeiGAMonti GraziadeiNeurogenesis and neuron recognition in the olfactory system of mammals, I: morphological aspects of differentiation and structural organization of the olfactory sensory neurons.J Neurocytol.1979;8:1-18.JHardingPPCGraziadeiGAMonti GraziadeiFLMargolisDenervation in the primary olfactory pathway of mice, IV: biochemical and morphological evidence for neuronal replacement following nerve section.Brain Res.1977;132:11-28.SBiffoJVerhaagenLHSchramaPSchotmanWDanhoFLMargolisB50/GAP43 expression correlates with process outgrowth in the embryonic mouse nervous system.Eur J Neurosci.1990;2:487-499.GJRamakersJVerhaagenABOestreicherImmunolocalization of B-50 (GAP-43) in the mouse olfactory bulb: predominant presence in preterminal axons.Neurocytology.1992;21:853-869.SSSchiffmanMLNashCDackisReduced olfactory discrimination in patients on chronic hemodialysis.Physiol Behav.1978;21:239-242.SSSchiffmanTaste and smell in disease, I.N Engl J Med.1983;308:1275-1279.FMargolisA brain protein unique to the olfactory bulb.Proc Natl Acad Sci U S A.1972;69:1221-1224.JVerhaagenABOestreicherMGrilloY-SKhew-DoodallWSGispenFLMargolisNeuroplasticity in the olfactory system: differential effects of central and peripheral lesions of the primary olfactory pathway on the expression of B-50/GAP-43 and the olfactory marker protein.J Neurosci Res.1990;26:31-44.TNakashimaCPKimmelmanJBSnow JrOlfactory marker protein in the human olfactory pathway.Arch Otolaryngol.1985;111:294-297.LMBartoshukChemosensory alterations and cancer therapies.Natl Cancer Inst Monogr.1990;9:179-184.Accepted for publication April 23, 1998.This project was supported by a grant from the Southern Illinois University School of Medicine, Central Research Committee, Springfield (Dr Struble).Presented in part in abstract form at the Annual Meeting of the American Association of Neuropathologists, Salt Lake City, Utah, June 10-13, 1993.Both antisera used by us were generously supplied by Frank Margolis, PhD. We thank Sarah Murphy for valuable technical support of this project.Reprints: Robert G. Struble, PhD, Center for Alzheimer Disease and Related Disorders, Southern Illinois University School of Medicine, PO Box 19230, Springfield, IL 62794-1413 (e-mail: firstname.lastname@example.org).
JAMA Otolaryngology - Head & Neck Surgery – American Medical Association
Published: Aug 1, 1998