Carcinogenesls vol.9 no. 1 pp. 179-181, 1988

SHORT COMMUNICATION

Mono-, di- and trimethylamine in human gastric fluid: potentialsubstrates for nitrosodimethylamine formation

Steven H.Zeisel, Kerry A.daCosta and J.ThomasLaMont1

Nutrient Metabolism Laboratory, Departments of Pathology and Pediatrics,and 'Section of Gastroenterology, Department of Medicine, UniversityHospital, Boston University School of Medicine, 85 East Newton Street,Room M1002, Boston, MA 02118, USA

Nitrosodimethylamine (NDMA) is a potent carcinogen in awide variety of animal species. In experimental animals, di-methylamine and nitrite, precursors of NDMA, are found ingastric fluid where the acidic conditions are suitable for for-mation of nitrosamines. In this study we measured the con-centrations of mono-, di- and trimethylamine (MMA, DMAand TMA) in gastric Quid from humans, rats, dogs and fer-rets, as well as in saliva, blood and urine from humans.Human gastric fluid contained 3.7 ± 0.4 (SEM) nmol/mlMMA, 12.6 ± 1 . 4 nmol/ml DMA and 2.0 ± 0.4 nmol/mlTMA. MMA, DMA and TMA concentrations in humangastric fluid were similar to those present in human salivaand blood, but were much lower than those present in humanurine. The concentrations of these amines in human gastricfluid were lower than those measured in gastric fluid fromexperimental animals. When we added sodium nitrite tohuman gastric fluid, NDMA was formed. We have shown thatDMA is normally present in human gastric fluid, and thatit can be nitrosated to form NDMA.

Nitrosodirnethylamine (NDMA*) is a potent carcinogen in a widevariety of animal species, and there is no reason to assume thathumans are resistant (1—3). The in vivo formation of NDMAprobably occurs in the normal human stomach, as the acidic con-ditions favor formation of nitrous anhydride and nitrosyl com-pounds which nitrosate amines to form nitrosamines (4). Di-methylamine (DMA), a precursor of NDMA, is ingested in ourfood (5), formed within our intestine by normal bacterial flora(6) and synthesized endogenously by mammals (7). DMA is effi-ciently transported from blood into gastric fluid in experimentalanimals (8). Monomethylamine (MMA) is a widespread com-ponent of fish and vegetables, and it is rapidly nitrosated in thestomach (9). Trimethylamine (TMA), found in high concentra-tions in fish, is also an excellent substrate for nitrosation, form-ing NDMA (10). There is little information about the primaryaliphatic amine content of human gastric fluid; Walters el al. (11)reported that DMA is present and that TMA is not. Nitrite, theother precursor for NDMA formation, is also found in gastricfluid, and is derived from dietary nitrite and nitrate, salivarynitrite and from endogenous synthesis of nitrite (2,12—19). Wemeasured the concentrations of various methylamines in gastricfluid from fasting humans, and compared these concentrationswith those found in other human biological fluids and in gastric

'Abbreviations: NDMA, nitrosodimethylamine; NDPA, nitrosodipropylamine;DMA, dimethylamine; MMA, monomethylamine; TMA, trimethylamine;SEM, standard error of mean.

fluid from fasting experimental animals. We also determinedwhether these endogenous amounts of DMA can be nitrosatedto form NDMA in human gastric fluid.

Gastric fluid was obtained from nine fasting (overnight) patientsundergoing endoscopy to rule out gastric pathology [three epi-gastric pain, one abdominal distension, one gastric lymphoma,one esophageal diverticulum, two reflux esophagitis, oneesophageal hernia; five males, four females; age range 22—95years, age 52 ± 7.4 years (mean ± SEM)]. Gastric fluid wasalso collected from three fasting (overnight) healthy control sub-jects via nasogastric tube (two males, one female; ages 22, 24and 42). All samples were placed in a sterile container and frozenat -95°C until used. Fisher rats (male, 150 g; Charles RiverBreeding Laboratories, Wilmington, MA) were fasted overnightand then were killed by cervical dislocation and gastric fluid ob-tained by surgically opening the stomach. Gastric fluid was col-lected from dogs using an indwelling cannula, and from ferretsusing nasogastric intubation, as described by Zeisel et al. (8).Blood, urine and saliva were collected from nine fasting (over-night) control subjects. Blood was immediately injected into asealed tube containing an equal volume of 10% trichloroaceticacid, this was subjected to centrifugation at 3000 g for 15 minat 4°C, and the acidic supernatant saved. Saliva was collectedusing a Curby cup (20) from a single parotid gland and was im-mediately acidified with 3 N HC1 so that the final concentrationof HC1 was 0.1 N. Urine was collected into vessels containing3 N HC1 so that the final concentration of HC1 was 0.1 N. Theacidified biological fluids were then used for amine assay. MMA,DMA and TMA were measured using a GLC method with nitro-gen—phosphorus detection (7). Recovery of aliphatic amine stan-dards added to gastric juice was 70% for MMA, 93% for DMA,97% for TMA and 92% for isopropylamine. Thiocyanate ingastric fluid was assayed by reaction with ferric nitrate andmeasurement of absorbance at 470 run (21).

For experiments in which nitrite was added to samples, ali-quots of human gastric fluids were placed into sealed vials(Teflon™ lined WISP septum; Waters Instruments) and sodiumnitrite (1.45 M) was injected into the vial to achieve the desirednitrite concentration. Samples were incubated at 37 °C for 0-120minutes as indicated. Nitrosation reactions were stopped by add-ing an equal volume of 2 N sodium hydroxide. At this point,nitrosodipropylamine (NDPA; 5 p.p.m. in dichloromethane) wasadded as an internal standard. Nitrosamines were then extractedinto dichloromethane and were measured using a gas chromato-graph (model 5890, Hewlett-Packard, Andover, MA) interfacedto a TEA analyzer (Thermo Electron, Waltham, MA). Sampleswere injected onto a fused silica capillary column (OV-225bonded phase, 25 m X 0.32 mm ID; Quadrax, New Haven, CT;injector temperature 200°C) and were eluted with nitrogen(6.5 ml/min) and a temperature gradient (50°C for 3 min, ris-ing at 25°C/min to 120°C, 120°C for 30 s, rising at 25°C/minto 200°C, and then holding at 200°C for 1 min). The TEAanalyzer oven pyrolyzer temperature was 35O°C (22). NDMA

© IRL Press Limited, Oxford, England 179

at University of C

alifornia, San Diego on Septem

ber 15, 2014http://carcin.oxfordjournals.org/

Dow

nloaded from

S.H.Zeisel, K.A.daCosU and J.T.LaMont

TaWe I. Mono-, di- and tnmethylamine in biologic fluids from humans

Fluid* No. ofsubjects

MMAb DMAb TMAb

GastricSalivaBloodUrine

fluid 129

119

3.75.03.8

156.4

±±±±

0.41.40.621.0

12.68.2

15.7288.2

±±±±

1.41.61.942.8

2.03.1

12.612.1

±±±±

0.41.01.51.5

"Gastric fluid, blood, urine and saliva were obtained from humans andamines were measured as described in text.bData (nmol/ml) are expressed as mean ± SEM. All measurements wereperformed in duplicate.

Table n . Mono-, di- and trimethylamine in gastric fluid of humans, rats,dogs and ferrets

Species No. ofsubjects

MMA1 DMA1 TMA"

HumanRatDogFerret

12346

3.723.111.817.4

±±±±

0.41.81.11.8

12.696.236.935.3

±±±±

1.44.54.46.0

2.06.29.23.4

±±±±

0.43.52.30.7

Gastric fluid was obtained from humans and animals, and amines weremeasured as described in text."Data (nmol/ml) are expressed as mean ± SEM. All measurements wereperformed in duplicate.

1.0

"S

pH<5(rt-8)

PH-5J(D-I)

30 60 90Time (min)

120

Fig. 1. Formation of NDMA after nitrite was added to human gastric fluid.Gastric fluid was obtained via endoscopy of patients as described in the text.The pH of eight samples was < 5 , one sample's pH was 5.5. AJiquots ofgastric fluid were incubated with 58 mM sodium nitrite for the timesindicated, and NDMA was assayed as described in text. Data are expressedas mean nmol NDMA formed/ml (±SEM). "P < 0.01 different from timezero by one-way ANOVA and Dunnett's test (23).

eluted at 2.32 min; NDPA eluted at 7.80 min. Recovery ofNDPA added to gastric fluid was 105 ± 4.2% (SEM).

Human gastric fluid in = 12) contained 3.7 ± 0.4 (SEM)nmol/ml MMA, 12.6 ± 1.4 nmol/ml DMA and 2.0 ± 0.4nmol/ml TMA. There was no significant difference betweengastric fluid methylamine concentrations in samples obtained us-ing endoscopy or nasogastric tube (Mest; 23). Thiocyanate is acatalyst for nitrosation of DMA (18); we found that human gastricfluid contained 18.7 ± 3.5 /xg/ml (two of the 12 subjects werecigarette smokers). Gastric fluid pH varied between 0.5 and 5.5;we observed no correlation of pH with MMA, DMA or TMAconcentrations (linear regression analyses, data not shown). Theconcentrations of MMA, DMA and TMA in human gastric fluidwere similar to those present in human sahva and blood, but weremuch lower than those present in human urine (Table I). Con-centrations of these amines in human gastric fluid were lowerthan those that we observed in dog, ferret or rat gastric fluid

(Table II). We could not detect NDMA in gastric juice samplesbefore adding nitrite. When we added nitrite, we observed pro-duction of NDMA (Figure 1). The rate of formation of NDMA,after addition of 58 mM sodium nitrite, was linear for the firsthour of incubation at 37°C (0.6 nmol/ml/h) in all samples withpH < 5 (pH = 1.57 ± 0.30). In one individual's gastric fluid(with a pH of 5.5, containing 3.2 nmol/ml MMA, 8.9 nmol/mlDMA, 0.1 nmol/ml TMA and 15 /ig/ml thiocyanate) NDMAformation was especially slow. When smaller amounts of nitritewere added to gastric juice, less NDMA was formed; after ad-dition of 7.3 mM nitrite, 0.34 ± 0.06 nmol NDMA were formedduring a 1-h incubation of a 1 ml of gastric fluid (P < 0.01 dif-ferent form 0 and 58 mM nitrite by one-way ANOVA andScheffe's test; 23).

We have shown that methylamines (MMA, DMA and TMA)are normally present in human gastric fluid (Table I). The con-centrations of these amines in gastric fluid are similar to thosepresent in blood, and we have previously shown that DMA canbe transported from blood into gastric fluid (8). Fed individualsmight have even more DMA and TMA in their gastric fluid thanwe report, as fish consumed in the diet can contain large amountsof DMA and TMA (5). Renal excretion is the major route forexcretion of methylamines, and we found that human urine con-tained higher concentrations of these amines than did gastric fluid.Rats, dogs and ferrets had considerably more MMA, DMA andTMA in the gastric fluid (Table II). This may reflect differencesin the metabolism of these amines, or in the amine content ofthe diet.

We were unable to detect endogenous levels of NDMA inhuman gastric juice that had not been treated with nitrite, howeverwe were only analyzing the content of 1 ml of fluid. Reed etal. (24), analyzing gastric fluid from control humans, observedtotal nitrosamine concentrations of 0.1 nmol/ml (if the pH wasin the range 1 — 1.5) or 1.2 nmol/ml (if pH > 6.5). Eisenbrandet al. (25) examined untreated gastric fluid from 25 individualsand found no volatile nitrosamines. However, Schlag et al. detect-ed significant volatile nitrosamines in gastric fluid from 26 con-trol subjects (26). Our tests, and those of Eisenbrand et al. (25)and Reed et al. (24) were performed on gastric fluid from fastinghumans, and thus may underestimate the nitrosamine that couldbe formed from nitrate and amines in foods. When we addedlarge amounts of nitrite we were able to show that NDMA ac-cumulated (Figure 1). Walters et al. (11) treated human gastricjuice with 5-15 g/1 nitrous acid at pH 2 for periods of 12—72 h.The predominant volatile nitrosamines formed were NDMA andnitrosopyrrolidine at concentrations of 0.05 and 0.005 /iM respec-tively. We believe that the endogenous DMA we have shownis present in gastric fluid was the precursor of the NDMA form-ed in both our and Walters et al. 's experiments, but TMA mayalso have been nitrosated. The maximal rate of NDMA forma-tion that we report (0.6 nmol/ml/h) compares favorably withpublished estimates of nitrosamine formation in gastric juice.Lane and Bailey (27) added 100 ftg/ml DMA and 100 /xg/mlnitrite to human gastric juice and formed between 9 and 34 ng/mlNDMA; we calculate that the rate of formation they observedwas ~ 0.1 -0.45 nmol/ml/h, about the same rate that we reportfor the nitrosation of endogenous DMA. Mirvish (28) added0.02 M DMA to hydrochloric acid (pH 3.4) and converted1 -20% of the DMA to NDMA over 3 h (we calculate a rate offormation of 0.7-13 /imol NDMA/ml/h; ~ 1000 times fasterthan we observed). NDMA formation in hydrochloric acid couldbe faster than in gastric juice due to lack of competition fornitrosation by other amines and phenols (11).

180

at University of C

alifornia, San Diego on Septem

ber 15, 2014http://carcin.oxfordjournals.org/

Dow

nloaded from

Mono-, di- and trimethytamine within human gastric fluid

In summary, we showed that human gastric fluid containsMM A, DMA and TMA. We also showed that the acidic condi-tions present in gastric fluid are suitable for nitrosation of thisendogenous DMA. Our data suggest that NDMA could be formedwithin human gastric fluid from substrates (DMA and nitrite)which are normally present. We do not know whether the smallamounts of NDMA that might be made could cause cancer inhumans.

AcknowledgementsThis work was supported by a grant from the National Institutes of Health(CA-26731).

References1. Bartsch.H. and Montesano.R. (1984) Relevance of nitrosamines to human

cancer. Carcinogenesis, 5, 1381-1398.2. Correa.P., Hacnszel.W., Cuello.C, Tannenbaum.S.R. and Archer.M.A.

(1975) A model for gastric cancer epidemiology. Lancet, ii, 58-60.3. Kakizoe.T., Wang,T.-T., Eng.V.W.S., Furrer.R., Dk>n,P., and Bruce.W.R.

(1979) Volatile A'-nitrosamines in the urine of normal donors and of bladdercancer patients. Cancer Res., 39, 829-832.

4. Mirvish,S.S. (1975) Formation of A'-nitroso compounds: chemistry, kineticsand in vivo occurence. Toxicol. Appl. Pharmacol., 31, 325—351.

5. Zeisel.S.H. and daCosta,K.A. (1986) Increase in human exposure to methyl-amine precursors of A'-nitrosamines after eating fish. Cancer Res., 46,6136-6138.

6. Zeisel.S.H., WishnokJ.S. and BlusztajnJ.K. (1983) Formation of methyl-amines from ingested choline and lecithin. J. Pharmacol. Exp. Ther., 225,320-324.

7. Zeisel.S.H., daCosta.K.A. and FoxJ. (1985) Endogenous formation of di-methylamine. Biochem. J., 232, 403-408.

8. Zeisel.S.H., daCosta.K.A., Edrise.B.M. and FoxJ. (1986) Transport of di-methylamine, a precursor of nitrosodimcthylamine, into the stomach of ferretand dog. Carcinogenesis, 7, 775—778.

9. HuberJC.W. and Lutz,W.K. (1984) Methylation of DNA in stomach and smallintestine of rats after oral administration of methylamine and nitrite. Car-cinogenesis, 5, 1729-1732.

10. Lijinsky.W., Keefer.L., Conrad.E. and Van de Bogart.R. (1972) Nitrosa-tkm of tertiary amines and some biological implications. J. Natl. Cancer Inst.,49, 1239-1249.

11. Walters.C.L., Smith,P.L.R. and Reed,P.I. (1982) Endogenous amines inhuman gastric juice. In Magee.P.N. (ed.), Banbury Report 12: Nitrosaminesin Human Cancer. Cold Spring Harbor Laboratory Press, NY, pp. 445-452.

12. Stuehr.DJ. and Marletta,M.A. (1985) Mammalian nitrate biosynthesis: mousemacrophages produce nitrite and nitrate in response to Escherichia coli lipo-polysaccharide. Proc. Nail. Acad. Sri. USA, 82, 7738-7742.

13. Ruddell.W.SJ., Bone.E.S., Hill.MJ., Blcndis.L.M. and Walters.C.L. (1976)Gastric-juice nitrite. A risk factor for cancer in the hypochlorohydric stomach?Lancet, ii, 1037-1039.

14. Tannenbaum.S.R., Moran.D., Ramd.W., Cuello.C. and Correa.P. (1979)Gastric cancer in Columbia. Nitrite and other ions in gastric contents froma high risk area. / . Natl. Cancer Inst., 62, 9-12 .

15. Harada.M., Nakamura.Y. and Tanimura.A. (1972) Nitrosamines in foods.9. Distribution of nitrite in various foods. J. Fd Hyg. Soc. Japan, 13, 36-40.

16. Tannenbaum.S.R., Weisman.M. and Fett.D. (1976) The effect of nitrate in-take on nitrite formation in human saliva. Fd. Cosmet. ToxicoL, 14, 549—552.

17. WhrteJ.W. (1975) Relative significance of dietary sauces of nitrate and nitrite.Agric. Food Chem., 23, 886-891.

18. Kawabata.T., Oshima.H., Uibu,J., Nakamura.M., Matsui.M. andHamano.M. (1979) Occurrence, formation and precursors of A'-nitroso com-pounds in Japanese diet. In Miller.E.C. (ed.), Naturally OccurringCarcinogens-mutagens and Modulators of Carcinogenesis. University Press,Baltimore, MD, pp. 195-209.

19. Tannenbaum.S.R., Moran.D., Falchuk.K.R., Correa,P. and Cuello.C. (1981)Nitrite stability and nitrosation potential in human gastric juice. Cancer Lett.,14, 131-136.

20. Curby.W.A. (1953) Device for collection of human parotid saliva. / . Lab.Clin. Med., 41, 493-496.

21. Bowler.R.G. (1944) The determination of thiocyanate in blood serum.Biochem. J., 38, 385-388.

22.Hansen,T.J., Archer.M.C. and Tannenbaum.S.R. (1979) Characterizationof pyrolysis conditions and interference by other compounds in the chemo-luminesccnce detection of nitrosamines. Anal. Chem., 51, 1526—1528.

23. Bruning J.L. and Kintz.B.L. (1977) Computational Handbook of Statistics.Scott, Foresman, Glenview, IL.

24. Reed.P.I., Smith.P.L.R., Haines.K., Housc.F.R. and Walters.C.L. (1981)Gastric juice A'-nitrosamines in health and gastrointestinal disease. Lancet,ii, 550-552.

25. Eisenbrand.G., Spiegelhakler.B. and Preussmaim.R. (1981) Analysis of humanbiological specimens for nitrosamine contents. In Bruce.W.R., Correa.P.,Uplrin.M., Tannenbaum.S.R. and Wilkins.T.D. (eds), Banbury Report 7:Gastrointestinal Cancer: Endogenous Factors. Cold Spring Harbor LaboratoryPress, NY, pp. 275-283.

26. Schlag,P., Bcckkr,R., Ulrich,H., Peter.M., Merkle.P. and Herferth.C. (1980)Are nitrite and n-nhroso compounds in gastric juke risk factors for carcinomain the operated stomach? Lancet, I, 727-729.

27. Lanc.R.P. and Bailey,M.E. (1973) The effect of pH on dimcthylnitrosamineformation in human gastric juice. Fd. Cosmet. Toxicol., 11, 851—854.

28. Mirvish.S.S. (1970) Kinetics of dimethylamine nitrosation in relation to nitros-amine carcinogenesis. / . Natl. Cancer Inst., 44, 633—639.

Received on July 16, 1987; revised on October 6, 1987, accepted onOctober Ii. 1987

181

at University of C

alifornia, San Diego on Septem

ber 15, 2014http://carcin.oxfordjournals.org/

Dow

nloaded from