Journal of Neuroendocrinology, 2000, Vol. 12, 355–359

Guinea-Pig Gonadotropin-Releasing Hormone:Immunoreactivity and Biological Activity1

C. -Q. Gao, J. Van den Saffele, M. Giri and J. -M. KaufmanDepartment of Endocrinology and Heymans Institute of Pharmacology, University of Ghent, Ghent, Belgium.

Key words: Guinea-pig, gonadotropin-releasing hormone, hypothalamus, immunoreactivity, bioactivity.

Abstract

The DNA sequence of the encoding gene predicts a unique structure for guinea-pig gonadotropin-releasing hormone (GnRH). We assessed the immunoreactivity of synthetic mammalian GnRH, of asynthetic peptide with predicted guinea-pig GnRH structure, and of extracts from rat and guinea-pighypothalami, using two different RIA systems. Whereas immunoreactivity of mammalian and guinea-pig GnRH was rather similar when using an antiserum with conformational specificity for mammalianGnRH (Root RR-5 antiserum), binding of both peptides to an antiserum with sequential specificity(Kelch R-13 antiserum) was markedly different. The findings for GnRH extracted from rat andguinea-pig hypothalami were similar to those for the corresponding synthetic peptides. Assessmentof in-vivo biological activity of synthetic mammalian and guinea-pig GnRH in the intact male guinea-pig showed that both peptides stimulate LH secretion dose-dependently, the response tomammalian GnRH being slightly greater at low dose. This study confirms that the GnRH expressedin the brain of the adult guinea-pig differs from mammalian GnRH and indicates that mammalian andguinea-pig GnRH display conformational similarity and that both can stimulate guinea-pig luteinizinghormone secretion.

Amongst small laboratory animals, the rat has been the most cursor molecule predicts a unique structure for the guinea-pig GnRH, which is different from all known forms ofextensively used to study the neuroendocrine regulation ofGnRH. This is unexpected: the structure of the predominantreproduction. The guinea-pig is, however, a potentiallyform of GnRH in the hypothalamus has proved to beattractive alternative model, because several aspects of itsidentical in all other mammalian species studied to date.reproductive function are more similar to that of primates,Moreover, from immunological studies following high-e.g. the existence of a true luteal phase as a component of itsperformance liquid chromatography (HPLC), Kelsall et al.reproductive cycle and a more similar anatomical distribution(11) have previously concluded that the dominant form ofof hypothalamic gonadotropin-releasing hormone (GnRH)GnRH in Guinea-pig is mammalian GnRH.neurones as compared to that in rat or mice (1–3). On the

We are presenting here evidence from immunological stud-other hand, the guinea-pig also has a number of remarkableies confirming the uniqueness of the major form of GnRHanomalies in its endocrine axes and metabolic pathways,expressed in the guinea-pig hypothalamus and data on thewhich by comparison with other species may provide usefulbiological activity of this peptide and mammalian GnRH innew insights (4). Using medial basal hypothalamus explantsthe guinea-pig.for the in-vitro study of the regulation of GnRH secretion,

we observed that while several regulatory mechanisms opera-tional in the guinea-pig hypothalamus are in line with data

Materials and methodsfor other species (5, 6), a number of findings diverge fromobservations for the rat (7–9). GnRH peptides

Recently, Jimenez-Linan et al. (10) reported that the DNA Mammalian GnRH (pGlu-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-GlyNH2) waspurchased from either Peninsula Laboratories (Belmont, CA, USA) for thesequence of the gene encoding the guinea-pig GnRH pre-

1This work was supported by a grant from the Belgian Queen Elisabeth Medical Foundation and the Fund For Scientific Research Flanders(grant 1.5176.96).

Correspondence to: Professor Dr Jean M. Kaufman, Department of Endocrinology 9K12, De Pintelaan 185, Ghent, B-9000, Belgium(e-mail: jean.kaufman@rug.ac.be).

© 2000 Blackwell Science Ltd

356 Further characterization of guinea-pig GnRH

Bin

din

g %

120

100

80

60

40

0 2.5 10

1:8 1:4 1:2 1:1 extract dilution 1:8 1:4 1:2 1:1 extract dilution

Mam-Syn GP-Syn GP-extr RAT-extr Mam-Syn GP-Syn GP-extr RAT-extr

40

Synthetic GnRH (pg/ml) Synthetic GnRH (pg/ml)

160 640 0 1 4 16 64 256

20

0

Bin

din

g %

120

Root RR-5 antibody Kelch R-13 antibody

100

80

60

40

20

0B

ind

ing

%

120

100

80

60

40

20

0

F. 1. Immunoreactivity of synthetic mammalian (Mam-Syn) and guinea-pig (GP-Syn) GnRH peptides and extracts of guinea-pig hypothalami(GP-extr) and rat hypothalami (Rat-extr), as tested in two different RIA systems using the Root RR-5 antiserum (right panel ) and Kelch R-13antiserum ( left panel ), respectively.

in-vitro experiments or from Wyeth Laboratoria B.V. (Hoofdorp, design. Samples were placed at 4 °C overnight, then serum was separated byNetherlands) for in-vivo administration. The GnRH-peptide with sequence as centrifugation and stored at −30 °C until assayed.predicted by the guinea-pig GnRH gene sequence (‘guinea-pig GnRH’: pGlu-Tyr-Trp-Ser-Tyr-Gly-Val-Arg-Pro-GlyNH2 (10)),was custom synthesized by

Hypothalamus extractsGenosys Biotechnologies Ltd (Cambridge, UK), >95% purity by HPLC.

The hypothalami of male guinea-pigs (BW 550–980 g) or rats (BW 200–Animals 280 g) were isolated and immediately weighed and homogenized in 1 ml ice-

cold 2 acetic acid. The homogenates were neutralized and centrifuged atMale Hartley-Dunn guinea-pigs (B & K, N. Humberside, UK) were housed1800 g for 30 min. The supernatants were collected and stored at −30 °Cunder conditions of constant temperature (21 °C) and light5dark cycle ( lightsuntil radioimmunoassay of GnRH.on, 08.00–20.00 h), with free access to guinea-pig chow and water, supple-

mented with vitamin C and straw. Male Wistar rats (Iffa credo, St Aubin-les-Elbeuf, France) were housed under similar conditions. All experiments GnRH RIA methodswere carried out in accordance with the Belgian Regulations on the Protections

The immunoreactivity of the synthetic GnRH peptides and extracts of ratof Experimental Animals.and guinea-pig hypothalami were studied in two RIA systems using different

Serum collection antibodies raised in rabbits against synthetic mammalian GnRH, Root RR-5antibody (a gift from A.Root, Department of Paediatrics, University of SouthOne week after arrival, guinea-pigs were anaesthetized with sodium pentobar-Florida) and Kelch R-13 antibody (a gift from V. Padmanabhan, Departmentbital (25 mg/kg BW ) and a polyethylene catheter (i.d.. 0.86 mm, O.D.of Paediatrics, University of Michigan). For both antibodies, previously1.27 mm) was introduced into the right atrium through the external jugularvalidated competitive RIA procedures were applied. RIA methods for thevein, the free end of the catheter being passed s.c. to emerge at the back ofRoot RR-5 antiserum were as previously described (7). In brief, 24 hthe neck. Between periods of sampling, the catheter was filled with 0.9%preincubation with primary antibody is followed by 24 h incubation afterNaCL containing 50 units/ml heparin and plugged. The guinea-pigs wereaddition of 125I-GnRH (NEN-Dupont, MA, USA) and subsequent separationhoused in individual cages and handled regularly by the operator so as toof bound and free with use of a sheep-antirabbit decanting agent (Pharmaciaenable blood sampling procedures to be carried out from conscious, unres-Upjohn, Sweden). The assay sensitivity is 0.28 pg/tube and the intra-assaytrained guinea-pigs, sampling being initiated not earlier than 48 h followingcoefficient of variation is <11%. RIA methods for the Kelch R-13 antiserumsurgery. At time of sampling, the guinea-pigs had a body weight of 800 g towere minor modifications from a previously validated protocol (J. Hotchkiss,980 g. Blood samples (0.3 ml ) were obtained via the catheter and replaced byHouston, Texas, personal communication). In brief, preincubation withequal volumes of saline without heparin. After obtaining two basal samplesprimary antibody for 96 h was followed by 24 h incubation with 125I-GnRHat 15-min intervals, five more samples were collected over a 140-min periodand subsequent separation of bound and free with use of a goat-antirabbitfollowing intracardiac bolus injection of the tested GnRH-peptide. Theantibody (Antibodies Inc, Davis, CA, USA) and absolute ethanol; assayresponse to an identical dose of the two tested peptides was studied in the

same guinea-pigs with at least a 1-week interval, according to a crossover sensitivity is 0.07 pg/tube and intra-assay coefficient of variation is <12%.

© 2000 Blackwell Science Ltd, Journal of Neuroendocrinology, 12, 355–359

Further characterization of guinea-pig GnRH 357

pg

/hyp

oth

alam

us

pg

/hyp

oth

alam

us

2500

2000

1500

1000

500

2000

1600

1200

800

400

0

01

1 2 3 4 5 6 7

2 3 4

Antibody Root RR-5 Antibody Kelch R-13

Antibody Root RR-5 Antibody Kelch R-13

5 6 7

Guinea-pig no.

Rat no.

8 9 10 11 12

F. 2. Apparent hypothalamic GnRH content, expressed in equivalentsof synthetic mammalian GnRH, as assessed in two different RIA systemsfor extracts of guinea-pig hypothalami (n=12; upper panel ) and rathypothalami (n=7; lower panel ); P=0.0016 and P>0.5 for differencein GnRH content as measured with the two antibodies for guinea-pigand rat hypothalami, respectively.

GnRH-Mam-Syn 0.1mg GnRH-GP-Syn 0.1mg n=5

Un

its

Un

its

Un

its

9876543

9

9876543210

876543210

210

GnRH-Mam-Syn 1mg GnRH-GP-Syn 1mg n=3

GnRH-Mam-Syn 10mg GnRH-GP-Syn 10mg n=6

_15 0 5 20

Min

45 80 140

_15 0 5 20

Min

45 80 140

_15 0 5 20

Min

45 80 140

F. 3. In-vivo response of serum LH to intracardiac bolus injection of0.1 mg, 1 mg and 10 mg of the synthetic mammalian (Mam-Syn) and

In-vitro bioassay for luteinizing hormone synthetic guinea-pig (GP-Syn) GnRH in intact adult male guinea-pigs.The data for a same dosage of both peptides are paired observations inSerum bioactive LH levels before and after i.v. injection of GnRH peptidethe same animals. The results (mean±SEM) are expressed in arbitrarywere determined by an in-vitro Leydig cell testosterone bioassay for guinea-units of an in-house guinea-pig LH standard.pig luteinizing hormone (LH), modified from the methods previously

described by Vam Damme et al. (12) and Dufau et al. (13), with use ofguinea-pig Leydig cells as the major distinctive feature. In brief, immediately

the two tested GnRH-peptides in a same animal were assayed in a singlefollowing the guinea-pigs being killed (BW 550–850 g), the testes werebioassay run, such as to eliminate interassay variation. Area under the curvedecapsulated and dispersed by incubation in medium 199 (6.2 ml for two(AUC) for the LH response in function of time was calculated according totestis) containing 40 U (0.28 mg)/ml collagenase type III (Sigma, St Louis,the trapezoidal method; concentrations below the assay limits of detectionMO, USA) and 0.14 U/ml dispase grade II (Boehringer Mannheim, Germany)were assigned as zero.at 370 °C for 20 min. The suspension was then diluted with equal volume of

medium and filtered through a double layer of nylon mesh (100 mm, Becton StatisticsDickinson, Franklin Lakes, NJ, USA). Ficoll at a final concentration of 13%

Statistical comparisons were performed using the Student’s paired t-test.was added to the filtrate which was then centrifuged at 1500 g for 10 min.The sediment containing the Leydig cells was resuspended in a mixture ofmedium 199 diluted with PBS buffer (351; 4 ml/testis), containing 0.125 m

ResultsMIX (3-isobutyl-1-methylxanthine, Aldrich-Europe, Beerse, Belgium), 0.5 mforskoline (Sigma), 0.5 g% BSA and 2 vol% foetal calf serum (LifeTechnologies, Merelbeke, Belgium). The concentration of viable Leydig cells Immunoreactivitywas assessed using the Tryptan blue exclusion method. Hundred ml of theLeydig cell suspension was then incubated with 50 ml of unknown sample A typical example of the binding-inhibition curve obtaineddiluted 1510 in PBS buffer, or 50 ml of guinea-pig LH standard solution in for synthetic mammalian GnRH, synthetic guinea-pig GnRHPBS-buffer containing an equal concentration of LH-free serum, for 4 h at and serial dilutions of rat and guinea-pig hypothalamic370 °C in 95% O2 and 5% CO2. The reaction was terminated by addition of

extracts, with use of the Root RR-5 and Kelch R-13 antimam-1.5 ml PBS-gelatin buffer containing 0.01% merthiolate. Following furthermalian GnRH antisera, respectively, are shown in Fig. 1. Indilution with PBS-gelatin buffer, testosterone concentration in the incubate

was measured directly using an in-house testosterone RIA. The LH standard the Root RR-5 RIA system, binding-inhibition curves forused in the bioassay has been prepared by extraction of guinea-pig pituitaries mammalian GnRH and guinea-pig GnRH show good paral-(14) and the results for unknown samples are expressed as arbitrary units of lelism with a slight shift to the right for the guinea-pig GnRH.this in house guinea-pig LH standard. The assay sensitivity is 0.15 arbitrary-

Serial dilutions of rat and guinea-pig hypothalamic extractsunits/ml, the intra-assay coefficient of variation is 9–15%. All samples wereassayed in triplicate; samples corresponding to the study of a same dose of show good parallelism with the binding-inhibition curve for

© 2000 Blackwell Science Ltd, Journal of Neuroendocrinology, 12, 355–359

358 Further characterization of guinea-pig GnRH

GnRH is illustrated in Figs 3 and 4. Both peptides clearlystimulated LH secretion, synthetic mammalian GnRH beingslightly more potent at low dose (P=0.047 for 0.1 mg dose;Fig. 4), with a somewhat prolonged apparent duration ofaction (Fig. 3).

Discussion

Our results show a distinct pattern of immunoreactivity forGnRH extracted from guinea-pig hypothalami as comparedto either synthetic mammalian GnRH or mammalian GnRHextracted from rat hypothalami. This is in accordance withthe data of Jimenez-Linan et al. (10), who predicted fromthe DNA sequence of the cloned gene encoding the guinea-pig GnRH precursor molecule, that guinea-pig hypothalamicGnRH has a unique structure with Tyr2 and Val7 assubstitutions.

We observed a similar immunoreactivity for mammalianand guinea-pig GnRH when tested with the Root RR-5antibody, raised against synthetic mammalian GnRH conjug-ated to bovine serum albumin by the carbodiimide conjuga-tion method. The RR-5 antiserum has previously beencharacterized as having conformational specificity (15),requiring the entire GnRH decapeptide sequence for binding,with lack of binding of any GnRH fragment. Monosub-stitutions in position 2 or 7, other than those in guinea-pigGnRH, affect binding to the RR-5 antiserum variably,

AU

C (

un

its.

min

)

450

400

350

300

250

200

150

100

50

N=5

N=3

N=6

0.1 mg 1 mg 10 mg0

GnRH-mam-Syn GnRH-Guinea pig-Syn

some of these analogues showing full cross-reactivity com-F. 4. Area under the curves (AUC; mean±SEM) for the response ofserum LH to 0.1 mg, 1 mg and 10 mg of the synthetic mammalian and pared to native mammalian GnRH. The similar binding ofguinea-pig GnRH. Paired observations for a same dose of the two mammalian and guinea-pig GnRH to the RR-5 antiserumpeptides are in the same animal (see Fig. 3); P=0.047 for 0.1 mg dose.

thus suggests conformational similarity of both peptides.Synthetic mammalian GnRH or rat hypothalamic GnRH,both synthetic peptides. By contrast, in the Kelch R-13 RIA

on the one hand, and GnRH extracted from guinea-pigsystem there is no parallelism between the binding-inhibitionhypothalami or synthetic guinea-pig GnRH on the othercurves for both synthetic peptides, and guinea-pig GnRHhand, displayed markedly different patterns of immunoreac-displays a clearly lower immunoactivity than mammaliantivity when tested with the Kelch R-13 antiserum. The latterGnRH. In the latter RIA system, there is also no parallelismantiserum, raised against synthetic mammalian GnRH con-for serial dilutions of rat and guinea-pig hypothalamicjugated to bovine serum albumin by a variant of the carbodii-extracts, but good parallelism between the binding-inhibitionmide conjugation method (16), is specific for the C-terminalcurves for rat hypothalamic extract and synthetic mammalianamide end of the GnRH molecule, amino acid residues 4–10GnRH, on the one hand, and between those for guinea-pigof the mammalian GnRH being required for binding to thehypothalamic extract and synthetic guinea-pig GnRH, on theantiserum (17).other hand.

Kelsall et al. (11) identified in extracts of adult guinea-pigThe apparent GnRH recovery from extracts of sevenbrain tissue a single dominant form of GnRH that eluted indifferent rat and 12 different guinea-pig hypothalami asthe same high pressure liquid chromatography fraction asmeasured in the two different RIA systems is shown in Fig. 2.synthetic mammalian GnRH and was recognised by the B-6For rat hypothalamic extracts, the findings with both RIAantiserum raised against a conjugate of mammalian syntheticsystems are consistent, whereas large and systematic discrep-GnRH. This latter antiserum appears to require amino acidancies are observed for guinea-pig extracts.residues 4–10 of mammalian GnRH for binding, data onThese results thus show a different pattern of immunoactiv-cross-activity of GnRH analogues with B-6 suggesting thatity for rat and guinea-pig hypothalamic GnRH as well as forsubstitutions in position 8 or in positions 7 and 8 greatlysynthetic mammalian GnRH and a synthetic peptide corres-reduce or abolish immunoreactivity. It seems plausible thatponding to the predicted structure of guinea-pig GnRH, andthe dominant form of GnRH observed by Kelsall et al. ina good concordance for the pattern of immunoactivityguinea-pig brain extracts may correspond to the guinea-pigbetween the latter peptide and extracts of guinea-pig hypo-GnRH form predicted by Jimenez-linan et al. (10). Indeed,thalamic tissue.one can predict that binding to antiserum B-6 would not begreatly affected by the substitution of tyrosine by histidine in

Biological activity position 2, while the data available do not allow to concludethat binding would be abolished by the conservative substitu-The response of serum bioactive LH to bolus injection of

increasing doses of synthetic mammalian and guinea-pig tion of leucine by valine in position 7.

© 2000 Blackwell Science Ltd, Journal of Neuroendocrinology, 12, 355–359

Further characterization of guinea-pig GnRH 359

expression in luteinizing hormone-releasing hormone neurons of guineaIn this context we can also mention that, when using LR-1pigs, with knife cuts separating the preoptic area and the hypothalamus,antiserum (a gift from Robert Benoit, Department ofdemonstrating luteinizing hormone surges. Biol-Reprod 1998; 58:

Endocrinology, The Montreal General Hospital, Canada) 323–329.raised against a conjugate of [-lys6 ] GnRH, an antibody 4 Keightley MC, Fuller PJ. Anomalies in the endocrine axes of the guinea

pig: relevance to human physiology and disease. Endocr Rev 1996;that has been used for immunostaining of GnRH neurones17: 30–44.in a variety of mammalian species (18), we obtained a pattern

5 Giri M, Kaufman JM. Effects of long term orchidectomy on in vitroof immunostaining in guinea-pig brain slices which was in pulsatile gonadotropin-releasing hormone release from the medial basalgood correspondence with previous descriptions (1, 3), hypothalamus of the adult guinea pig. Endocrinology 1994; 134:

1621–1626.immunostaining being completely abolished by preabsorption6 Giri M, Kaufman JM. Opioidergic modulation of in vitro pulsatileof the LR-1 antiserum with the synthetic peptide correspond-

gonadotropin-releasing hormone release from the isolated medial basaling to the predicted (10) guinea-pig GnRH (data not shown). hypothalamus of the male guinea pig. Endocrinology 1994; 135:Following intracardiac bolus injection, both synthetic mam- 2137–2143.

malian and guinea-pig GnRH dose-dependently stimulated 7 Giri M, Kaufman JM. In vitro GnRH release from the isolated medialbasal hypothalamus of the male guinea pig: evidence for the existence ofLH secretion in the intact male guinea-pig. Although in-vitrotwo pools of releasable GnRH. Brain Res 1994; 648: 270–280.binding studies have not been performed, it seems plausible

8 Giri M, Kaufman JM. Involvement of neuroexcitatory amino acids inthat this observation reflects that both peptides bind the the control of gonadotropin-releasing hormone release from the hypothal-guinea-pig pituitary GnRH receptor, which has not yet been amus of the adult male guinea pig: predominantly inhibitory action of

N-Methyl--Aspartate-mediated neurotransmission and its reversal aftercloned. That both peptides bind and activate the guinea-pigorchidectomy. Endocrinology 1995; 136: 2404–2411.GnRH receptor is in agreement with the suggestion of

9 Giri M, Gao CQ, Kaufman JM. The N-Methyl--Aspartate-mediatedconformational similarity indicated by comparable binding inhibitory control of gonadotropin-releasing hormone release in theto the Root RR-5 antiserum. Interestingly, the more potent hypothalamus of the adult male guinea pig is expressed through opioid-

ergic systems. Endocrinology 1996; 137: 1468–1473.biological activity (Fig. 4) and slightly prolonged duration of10 Jimenez-linan M, Rubin BS, King JC. Examination of guinea pigaction (Fig. 3) for mammalian GnRH suggests that it might

luteinizing hormone-releasing hormone gene reveals a unique decapeptidehave a higher affinity for the guinea-pig GnRH receptor than and existence of two transcripts in the brain. Endocrinology 1997; 138:guinea-pig GnRH. However, other mechanisms, such as a 4123–4130.

11 Kelsall R, Coe IR, Sherwood NM. Phylogeny and ontogeny of gonad-slower enzymatic degradation, could also underlie the greaterotropin-releasing hormone: comparison of guinea pig, rat, and a proto-response to the mammalian GnRH.chordate. Gen Comp Endocrinol 1990; 78: 479–494.

12 Van Damme MP, Robertson DM, Diczfalusy E. An improved in vitrobioasay method for measuring luteinizing hormone (LH) activity, usingAcknowledgementsmouse Leydig cell preparations. Acta Endocrinol 1974; 77: 655–671.

13 Dufau ML, Mendelson CR, Catt KJ. A highly sensitive in vitro bioassayWe would like to thank Mrs Kathelijne Mertens for her skilful technicalfor LH and hCG. Testosterone production by dispersed Leydig cells.assistance. We are greatly indebted to Drs A. Root (Department of Paediatrics,J Clin Endocrinol Metab 1974; 39: 610–613.University of South Florida), V. Padmanabhan (Department of Paediatrics,

14 Parlow AF. Importance of differential, quantitative bio-assays for pituit-University of Michigan) and R. Benoit (Department of Endocrinology, theary gonadotrophins in the rat. Endocrinology 1964; 74: 138–141.Montreal General Hospital, Canada) for their generous gifts of GnRH

15 Copeland KC, Aubert ML, Rivier J, Sizonenko PC. Luteining hormone-antisera and to Dr J. Hotchkiss (Houston, Texas, USA) for her helpful advicereleasing hormone: seguential versus conformational specificity of anti-in establishing the RIA and bioassay.luteinizing hormone-releasing hormone sera. Endocrinology 1979; 104:1504–1512.

Accepted 22 October 1999 16 Koch Y, Wilchek M, Fridkin M, Chobsieng P, Zor V, Lindner HR.Production and characterization of an antiserum to synthetic gonadotro-pin-releasing hormone. Biochem Biophys Res Comm 1973; 55: 616–622.

17 Hauger RL, Kelch RP, Chen YDI, Payne AH. Testicular receptors forReferencesluteinizing hormone after immunoneuralization of gonadotropin releasing

1 Silverman AJ. Distribution of luteinizing hormone-releasing hormone hormone in the male rat. J Endocrinol 1977; 75: 23–32.(LHRH) in the guinea pig brain. Endocrinology 1975; 99: 30–41. 18 Silverman AJ, Witkin JW, Miller RP. Light and electron microscopic

2 King JC, Anthony EL. LHRH neurons and their projections in humans immunocytochemical analysis of antibodies directed against GnRH andand other mammals: species comparisons. Peptides 1984; 5: 195–207. its precursor in hypothalamic neurons. J Histochem Cytochem 1990;

38: 803–813.3 King JC, Ronsheim P, Liu E, Powers L, Slonimski M, Rubin BS. Fos

© 2000 Blackwell Science Ltd, Journal of Neuroendocrinology, 12, 355–359