This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Brosnan, R. J. Articles by Sonner, J. M. Search for Related Content PubMed Articles by Brosnan, R. J. Articles by Sonner, J. M. Related Collections Mechanisms Pharmacology

---

ANESTHETIC PHARMACOLOGY

Ammonia Has Anesthetic Properties

Robert J. Brosnan, DVM, PhD^*, Liya Yang, PhD, Pavle S. Milutinovic, MS, Jing Zhao, MD, Michael J. Laster, DVM, Edmond I. Eger, II, MD, and James M. Sonner, MD

From the *Department of Surgical and Radiological Sciences, School of Veterinary Medicine, University of California, Davis, California, Department of Anesthesia and Perioperative Care, The University of California, San Francisco, California, The University of Pittsburg, Pennsylvania, and Peking Union Medical College, Beijing, China.

Address correspondence and reprint requests to James M. Sonner, MD, Department of Anesthesia, S-455, University of California, San Francisco, CA 94143-0464. Address e-mail to sonnerj{at}anesthesia.ucsf.edu.

	Abstract

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
BACKGROUND: A recent theory of anesthesia predicts that some endogenous compounds should have anesthetic properties. This theory raises the possibility that metabolites that are profoundly elevated in disease may also exert anesthetic effects. Because in pathophysiologic concentrations, ammonia reversibly impairs memory, consciousness, and responsiveness to noxious stimuli in a manner similar to anesthetics, we investigated whether ammonia had anesthetic properties.

METHODS: The effect of ammonia was studied on ₁β₂ and ₁β_22s -amino butyric acid type A, ₁ glycine, and NR1/NR2A N-methyl-d-aspartate receptors, and the two-pore domain potassium channel TRESK. Channels were expressed in Xenopus laevis oocytes and studied using two-electrode voltage clamping. The immobilizing effect of ammonia in rats was evaluated by determining the reduction in isoflurane minimum alveolar concentration produced by IV infusion of ammonium chloride. The olive oil-water partition coefficient was measured to determine whether free ammonia (NH₃) followed the Meyer-Overton relation.

RESULTS: Ammonia positively modulated TRESK channels and glycine receptors. No effect was seen on ₁β₂ and ₁β_22s -amino butyric acid type A receptors or NR1/NR2A N-methyl-d-aspartate receptors. Ammonia reversibly decreased the requirement for isoflurane, with a calculated immobilizing EC₅₀ of 1.6 ± 0.1 mM NH₄Cl. The Ostwald olive oil-water partition coefficient for NH₃ was 0.018. At a pH of 7.4, and at the anesthetic EC₅₀, the NH₃ concentration in bulk olive oil is 0.42 µM, approximately five orders of magnitude less than observed by anesthetics that follow the Meyer-Overton relation.

CONCLUSIONS: These findings support the hypothesis that ammonia has anesthetic properties. Bulk oil concentration did not predict the potency of ammonia.

	Introduction

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
A recent theory of anesthesia predicts specific endogenous compounds that may have anesthetic properties (1) via indirect actions on ion channels in high, nonspecific concentrations. A corollary to this theory is the possibility that a variety of other endogenous compounds may, in high concentrations, also have anesthetic properties. A variety of metabolites are profoundly elevated in certain diseases associated with impaired responsiveness, such as liver failure (2), uremia (3), and diabetic ketoacidosis (4); we conjecture that some of these metabolites, prominent among which is ammonia, may be acting as anesthetics.

Ammonia is a product of protein catabolism which, in mammals, is metabolized in the liver to urea by enzymes of the urea cycle. Humans with liver failure (2) or urea cycle disorders (4) develop symptoms ranging from mild confusion, through drowsiness and sleep to coma. Elevated ammonia concentrations are a feature of these conditions and correlate with disease severity (5). In the case of liver failure, therapies are available (e.g., liver transplantation) and these therapies demonstrate that the symptoms are reversible. For these reasons, we hypothesized that ammonia has anesthetic properties. We tested this by demonstrating that ammonia produces the stereotypic features of inhaled anesthetics on ion channels and in animals. Those features include enhancement of currents through inhibitory anesthetic-sensitive ion channels, and a reversible reduction in the requirement for a conventional inhaled anesthetic in rats.

	METHODS

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
Our Institutional Animal Care and Use Committee approved all studies involving animals.

Ion Channel Physiology
Ion channel expression in Xenopus laevis oocytes, and two electrode voltage clamp study of TRESK channels and N-methyl-d-aspartate (NMDA) receptors were performed as described previously (6).

-Aminobutyric acid type A (GABA_A)and glycine receptors were expressed and studied using similar methods (7). Human GABA_{A 1} and _2s and the rat GABA_A β₂ subunits in pCIS II vectors, and human ₁ glycine receptors in a PBK-CMV vector, were a gift from Professor R.A. Harris. Approximately 1 ng total plasmid, consisting of equal parts GABA_{A 1} and β₂ receptor subunits, and a 10-fold excess of _2s, or of the ₁ subunit of the glycine receptor, were injected into the nucleus of frog oocytes and studied 1-4 days later. Expression of GABA_A receptors containing the _2s subunit was tested by application of 10 µM zinc chloride for 1 min followed by coapplication of zinc chloride with GABA, and observing inhibition of currents by 10% or less.

Oocytes were perfused with frog Ringer’s solution for 5 min, and then frog Ringer’s solution containing agonist (0.03 mM GABA or 0.075 mM glycine) for 20 s; this was repeated three times. NH₄Cl in frog Ringer’s solution was perfused for 100 s followed by the same concentration in frog Ringer’s solution plus agonist. Reversibility was verified after washout with frog Ringer’s solution. Solutions were made fresh daily and pH adjusted.

General Anesthetic Effects in Rats
To determine whether ammonia was an anesthetic in animals, we tested whether ammonia could reduce the concentration of a conventional inhaled anesthetic, isoflurane, required to cause immobility in 19 Crl:CD®(SD)Br rats. After determining the anesthetic concentration for isoflurane (i.e., MAC, for minimum alveolar concentration producing immobility in response to a supramaximal stimulus) using previously described methods (8), 1.5 M NH₄Cl was administered IV at approximately 10 mL · kg^–1· h^–1 over 60-75 min in rats inhaling subanesthetic concentrations of isoflurane. One molar NaHCO₃ was coadministered intraperitoneally at the same rate. At the end of the infusions, the presence or absence of movement in response to a supramaximal tail clamp was assessed, inspired isoflurane concentrations were measured, animals were killed and blood was collected for ammonia, blood gas, and electrolyte analysis. Additional animals were studied with isoflurane alone (Table 1). A logistic regression analysis (9) was used to solve the concentration of total ammonia alone that produced anesthesia in rats administered both isoflurane and NH₄Cl and only isoflurane.

Measurement of Ammonia’s Partition Coefficient
Ammonia’s olive oil partition coefficient was measured in quadruplicate by shaking 25 mL 1 M NH₄ at pH 7.4 with 50 mL extra-light olive oil (Bertoli, Unilever, NV) for 2 h at 37°C. The unionized ammonia in this solution was calculated from the Henderson-Hasselbach equation. The emulsion was centrifuged and 40 mL of oil removed and shaken with 40 mL of fresh water at pH 4.0. The ammonium in this water was measured using an ion selective electrode (Omega Engineering, Stamford, CT), and assumed to equal all the ammonia originally dissolved in the oil. The oil-water partition coefficient was the concentration ratio of unionized ammonia (NH₃)in the oil to that in the original aqueous solution.

	RESULTS

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
Effects of Ammonia on Ion Channel Function
We applied ammonium chloride in millimolar concentrations to expressed ion channels, because animal models of hyperammonemia from liver failure report total ammonia (NH₄⁺ + NH₃) concentrations in this range (10); this compares to normal total ammonia concentrations, which are <40 µM (11). We found that at concentrations of 1 mM and larger, the function of strychnine-sensitive, glycine receptors was enhanced (Fig. 1). We also found that NH₄Cl enhanced currents through the anesthetic-sensitive two-pore domain potassium leak channel TRESK (12) in a concentration-dependent manner above 3 mM (Fig. 2). Thus, both glycine receptors and TRESK channels were affected in a manner similar to conventional inhaled anesthetics.

View larger version (28K): [in this window] [in a new window]   Figure 1. Anesthetic-like effects of NH4Cl on currents through strychnine-sensitive glycine receptors. Xenopus laevis oocytes expressing homomeric human 1 glycine receptors were exposed to NH4Cl for 100 s, or perfusates containing an isosmotic concentration of sucrose. Six oocytes were studied at 1 mm, 5 at 2.5 mM, 7 at 5 mM, and 6 at 10 mM NH4Cl. NH4Cl at all concentrations studied enhanced (potentiated) chloride current through this receptor in response to coapplication of 75 µM glycine with NH4Cl for 20 s.

View larger version (30K): [in this window] [in a new window]   Figure 2. Anesthetic-like effects of NH4Cl on currents through TRESK two-pore domain potassium channels. Xenopus laevis oocytes expressing human TRESK channels were studied by two electrode voltage clamping. Six oocytes were studied at each concentration of NH4Cl. Exposure to 6, 9, and 12 mM NH4Cl significantly enhanced currents at +60 mV compared to isosmotic sucrose controls.

There was no effect of NH₄Cl on ₁β₂ or ₁β_22s -aminobutyric acid type A (GABA_A) receptors or NMDA receptors comprising NR1/NR2A subunits, at NH₄Cl concentrations up to 10 mM.

Immobilizing EC₅₀
Applied isoflurane concentrations ranged from 0.3% to 0.9% atm in rats receiving ammonium chloride infusions. Total blood ammonia concentrations were between 0.3 and 2.2 mM in rats receiving ammonium chloride infusions, with an average of 0.9 mM. The NH₄Cl infusion decreased isoflurane MAC by an average of 60%. The calculated EC₅₀ ± se for ammonium chloride was 1.6 ± 0.1 mM, and for isoflurane, 1.52% ± 0.06% atm. At doses larger than those we used, rats developed pulmonary edema; consequently, enough NH₄Cl to achieve anesthesia by itself could not be infused. Blood chemistry was within normal limits, with a mean (±se) arterial blood pH, base deficit, and plasma potassium concentration of 7.39 ± 0.02, –0.6 ± 1.2, and 5.8 ± 0.3 mmol/L, respectively. Arterial Po₂ was 461 ± 15 mm Hg, and all animals were normocapneic. Blood osmotic pressure was only modestly increased at 320 ± 3 mmol/kg.

The reversibility of the NH₄Cl effect was determined in six rats. Isoflurane MAC was determined to be 1.61% ± 0.04% atm (mean ± se) in these animals, but only 0.06% ± 0.07% higher 1 day after NH₄Cl infusion, a value not significantly different than baseline. Hence, the effects of ammonium on isoflurane MAC were reversible.

Oil Solubility
The Ostwald olive oil-water partition coefficient for NH₃ was 0.018. At a pH of 7.4, given that the pK_a of ammonia is 9.24, the bulk olive oil concentration of NH₃ in equilibrium with the 1.6 mM aqueous anesthetic concentration of NH₄Cl is calculated to be 0.42 µM.

	DISCUSSION

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
Our results show that ammonia has anesthetic properties: ammonia had a MAC-sparing effect in rats, and was found to enhance the function of two anesthetic-sensitive ion channels with inhibitory effects, the strychnine-sensitive glycine receptor and the two-pore domain K⁺ channel TRESK.

We chose to study the effect of ammonia on MAC because unresponsiveness to painful stimuli is seen in patients with elevated ammonia levels (13). We studied the effect of ammonia on ion channels which have been demonstrated to be sensitive to anesthetics, because these channels are thought to be molecular targets mediating anesthetic actions. These included glycine receptors, which are pentameric chloride channels in the same receptor superfamily as the GABA_A receptor. These receptors are important because they are the major mediators of inhibitory neurotransmission in the hindbrain and spinal cord, and inhaled anesthetics and ethanol at 1 MAC have been shown to enhance the action of glycine on glycine receptors (14,15). These studies also included the human TRESK channel, which is a two-pore domain potassium leak channel formed from dimers of pore-forming units, each containing four transmembrane segments and two tandem P domains (the putative pore-forming region) (16). TRESK currents are enhanced by volatile anesthetics (12). The enhancement of glycine receptor and TRESK channel function we observed may contribute to the central nervous system depression seen in diseases in which ammonia is elevated, such as fulminant hepatic failure and urea cycle disorders. More generally, our findings suggest that other ion channels or subtypes of channels we have not studied, and which respond to anesthetics, may contribute to the central nervous system depression found in hyperammonemic disorders.

Surprisingly, we did not find an effect of NH₄Cl on ₁β₂ or ₁β_22s GABA_A receptors or NMDA receptors comprising NR1/NR2A subunits, at NH₄Cl concentrations up to 10 mM. This contrasts with previous investigations, which reported enhancement of GABA_A receptor function with 1 mM (NH₄⁺ in dissociated cortical neurons (17), and a NMDA receptor response to ammonia (18). Possibly, different subtypes of receptors than ours were present on neurons in these studies.

We asked whether NH₃ was exerting the observed anesthetic effects in a manner similar to conventional volatile anesthetics, rather than (NH₄⁺ which is the predominant form of ammonia present during infusion. Conventional volatile anesthetics have bulk oil phase concentrations of 25-50 mM at their anesthetizing concentrations (the Meyer-Overton relation (19). We measured the Ostwald olive oil-water partition coefficient of NH₃, finding it to be 0.018. This corresponds to a bulk oil phase concentration of 0.42 µM NH₃ at an anesthetizing concentration of NH₄Cl (1.6 mM; see above) at a physiologic pH of 7.4. This is approximately five orders of magnitude lower than that observed with conventional inhaled anesthetics. This indicates that the anesthetic-like effects we observed were probably not from free ammonia acting like a conventional inhaled anesthetic.

The concentrations of ammonia we used in our studies were in the range that produces hyperammonemic symptoms in animal models (10). This is higher than that found in humans with hepatic failure. However, in fuminant hepatic failure, a wide variety of metabolites are elevated, and these may add to the depressant effect of lower concentrations of ammonia.

Ammonium chloride infusions produced a hyperosmolar state. This however cannot explain our finding of a decreased MAC for isoflurane during ammonium chloride infusion. For example, previous studies of infusion of 12% dextrose in dogs increased osmolarity from 306 to 359, but did not change MAC (20). Administration of 25% dextrose changed osmolarity from 301 to 370, but also did not change MAC significantly. Thus, osmolarity alone does not change MAC. The aforementioned increases were not accompanied by an increase in sodium in the cerebrospinal fluid (CSF). When hypertonic saline or mannitol was infused and sodium in the CSF increased, MAC increased. Thus, if anything, an increased osmolarity through an increased CSF sodium should increase MAC, but we found a decrease in MAC.

In summary, these results are significant because they 1) show that ammonia has anesthetic properties which, in disease, may contribute to central nervous system depression, 2), suggest that ion channels which mediate anesthetic actions may contribute to impaired responsiveness in liver failure and other hyperammonemic disorders, and 3) potentially identify a new class of molecules with anesthetic properties.

	Footnotes

 
Accepted for publication March 6, 2007.

Supported, in part, by NIGMS R01 GM069379 (to J.S.) and T32 GM08440 (to R.B.).

Dr. Eger is a paid consultant to Baxter Healthcare Corp.

	REFERENCES

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 

1. Cantor R. Receptor desensitization by neurotransmitters in membranes: are neurotransmitters the endogenous anesthetics? Biochemistry 2003;42:11891–7.[Medline]
2. Jalan R, Shawcross D, Davies N. The molecular pathogenesis of hepatic encephalopathy. Int J Biochem Cell Biol 2003;35:1175–81.[Web of Science][Medline]
3. Brouns R, De Deyn PP. Neurological complications in renal failure: a review. Clin Neurol Neurosurg 2004;107:1–16.[Web of Science][Medline]
4. Warrell D, Cox T, Firth J, Benz E, eds. Oxford textbook of medicine. 4th ed. Oxford: Oxford University Press, 2003.
5. Ong JP, Aggarwal A, Krieger D, et al. Correlation between ammonia levels and the severity of hepatic encephalopathy. Am J Med 2003;114:188–93.[Web of Science][Medline]
6. Brosnan R, Gong D, Cotten J, et al. Chirality in anesthesia. II. Stereoselective modulation of ion channel function by secondary alcohol enantiomers. Anesth Analg 2006;103:86–91.[Abstract/Free Full Text]
7. Mihic SJ, Ye Q, Wick MJ, et al. Sites of alcohol and volatile anaesthetic action on GABA (A) and glycine receptors. Nature 1997;389:385–9.[Medline]
8. Eger EI II, Xing Y, Laster M, et al. Halothane and isoflurane have additive minimum alveolar concentration (MAC) effects in rats. Anesth Analg 2003;96:1350–3.[Abstract/Free Full Text]
9. Sonner JM. Issues in the design and interpretation of minimum alveolar anesthetic concentration (MAC) studies. Anesth Analg 2002;95:609–14.[Abstract/Free Full Text]
10. Szerb JC, Butterworth RF. Effect of ammonium ions on synaptic transmission in the mammalian central nervous system. Prog Neurobiol 1992;39:135–53.[Web of Science][Medline]
11. Kitaura J, Miki Y, Kato H, et al. Hyperinsulinaemic hypoglycaemia associated with persistent hyperammonaemia. Eur J Pediatr 1999;158:410–13.[Web of Science][Medline]
12. Liu C, Au JD, Zou HL, et al. Potent activation of the human tandem pore domain K channel TRESK with clinical concentrations of volatile anesthetics. Anesth Analg 2004;99:1715–22.[Abstract/Free Full Text]
13. Sass DA, Shakil AO. Fulminant hepatic failure. Liver Transpl 2005;11:594–605.[Web of Science][Medline]
14. Mascia MP, Machu TK, Harris RA. Enhancement of homomeric glycine receptor function by long-chain alcohols and anaesthetics. Br J Pharmacol 1996;119:1331–6.[Web of Science][Medline]
15. Downie DL, Hall AC, Lieb WR, Franks NP. Effects of inhalational general anaesthetics on native glycine receptors in rat medullary neurones and recombinant glycine receptors in Xenopus oocytes. Br J Pharmacol 1996;118:493–502.[Web of Science][Medline]
16. Patel AJ, Honore E. Anesthetic-sensitive 2P domain K⁺ channels. Anesthesiology 2001;95:1013–21.[Web of Science][Medline]
17. Takahashi K, Kameda H, Kataoka M, et al. Ammonia potentiates GABAA response in dissociated rat cortical neurons. Neurosci Lett 1993;151:51–4.[Web of Science][Medline]
18. Marcaida G, Felipo V, Hermenegildo C, et al. Acute ammonia toxicity is mediated by the NMDA type of glutamate receptors. FEBS Lett 1992;296:67–8.[Web of Science][Medline]
19. Taheri S, Halsey MJ, Liu J, et al. What solvent best represents the site of action of inhaled anesthetics in humans, rats, and dogs? Anesth Analg 1991;72:627–34.[Abstract/Free Full Text]
20. Tanifuji Y, Eger EI II. Brain sodium, potassium, and osmolality: effects on anesthetic requirement. Anesth Analg 1978;57:404–10.[Abstract/Free Full Text]

This article has been cited by other articles:

				Y. Weng, T. T. Hsu, J. Zhao, S. Nishimura, G. G. Fuller, and J. M. Sonner Isovaleric, Methylmalonic, and Propionic Acid Decrease Anesthetic EC50 in Tadpoles, Modulate Glycine Receptor Function, and Interact with the Lipid 1,2-Dipalmitoyl-Sn-Glycero-3-Phosphocholine Anesth. Analg., May 1, 2009; 108(5): 1538 - 1545. [Abstract] [Full Text] [PDF]

				E. I. Eger II, D. E. Raines, S. L. Shafer, H. C. Hemmings Jr, and J. M. Sonner Is a New Paradigm Needed to Explain How Inhaled Anesthetics Produce Immobility? Anesth. Analg., September 1, 2008; 107(3): 832 - 848. [Abstract] [Full Text] [PDF]

				J. M. Sonner A Hypothesis on the Origin and Evolution of the Response to Inhaled Anesthetics Anesth. Analg., September 1, 2008; 107(3): 849 - 854. [Abstract] [Full Text] [PDF]

				L. Yang and J. M. Sonner Anesthetic-Like Modulation of Receptor Function by Surfactants: A Test of the Interfacial Theory of Anesthesia Anesth. Analg., September 1, 2008; 107(3): 868 - 874. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Brosnan, R. J. Articles by Sonner, J. M. Search for Related Content PubMed Articles by Brosnan, R. J. Articles by Sonner, J. M. Related Collections Mechanisms Pharmacology