Increasing activity of H2-metabolizing microbes lowers decompression sickness risk in pigs during H2 dives
http://www.100md.com
《应用生理学杂志》
1 Naval Medical Research Center, Silver Spring, Maryland 20910-7500;
2 Department of Microbiology, University of Georgia, Athens, Georgia 30602
ABSTRACT
The risk of decompression sickness (DCS) was modulated by varying the biochemical activity used to eliminate some of the hydrogen (H2) stored in the tissues of pigs (19.4 ± 0.2 kg) during hyperbaric exposures to H2. Treated pigs (n = 16) received intestinal injections of Methanobrevibacter smithii, a microbe that metabolizes H2 to water and CH4. Surgical controls (n = 10) received intestinal injections of saline, and an additional control group (n = 10) was untreated. Pigs were placed in a chamber and compressed to 24 atm abs (20.6-22.9 atm H2). After 3 h, the pigs were decompressed and observed for symptoms of DCS for 1 h. Pigs with M. smithii had a significantly lower (P < 0.05) incidence of DCS (44%; 7/16) than all controls (80%; 16/20). The DCS risk decreased with increasing activity of microbes injected (logistic regression, P < 0.05). Thus the supplemental tissue washout of the diluent gas by microbial metabolism was inversely correlated with DCS risk in a dose-dependent manner in this pig model.
keywords:Methanobrevibacter smithii; methanogens; biochemical decompression; decompression illness; hydrogen diving
INTRODUCTION
DECOMPRESSION SICKNESS (DCS) can be the penalty for failing to eliminate some portion of the tissue load of gas acquired while breathing under hyperbaric conditions (3). We report here studies on a novel approach to facilitating gas washout from tissues, thereby reducing the risk of DCS in an animal model. This approach, which we call biochemical decompression, uses microbes to metabolize hydrogen (H2) inside animals during a hyperbaric exposure with H2 as the primary gas in the breathing mixture. The conditions of the experiments are intended to simulate deep diving to 20-60 atm (700-2,000 feet of seawater). For such dives, H2 may be more appropriate than helium or nitrogen as the diluent to O2. Because of its lower density, H2 requires less lung ventilatory effort at high pressures (4). H2 has the additional advantage that it is less narcotic than N2 at high pressures and that some degree of H2 narcosis suppresses high-pressure neurological syndrome (4).
The concept of H2 biochemical decompression has been demonstrated in a rat model (19). Cultures of Methanobrevibacter smithii were injected into the proximal end of the large intestines of the animals (19). This microbe, which is native to the human intestinal flora (21), metabolizes H2 as
(1)
The microbes convert H2 to water and to a readily traceable end product (methane; CH4) that is derived exclusively from this reaction (17, 21). Rats supplied with M. smithii had a reduction in DCS risk by approximately half from a chosen compression and decompression sequence in H2 (19). Likewise, pigs that released detectable quantities of CH4 were found to have a lower risk of DCS from a chosen exposure to hyperbaric H2 compared with pigs that did not release CH4; this CH4 was generated entirely by microbes (of unknown species) native to the animals' intestinal flora (18). Thus these data demonstrated that the concept of biochemical decompression is valid.
Although most people on a western diet are expected to have M. smithii in their gut flora (21), it would not be acceptable to depend on the activity of this highly variable population for all biochemical decompression in human H2 diving. In order for H2 biochemical decompression to be useful for humans, the gut flora will need to be supplemented with M. smithii, probably by means of enteric-coated capsules of these microbes taken orally.
To advance this work further toward human use, we needed to determine the relationship between the in vitro activity of M. smithii delivered, the amount of H2 these microbes eliminated in vivo, and DCS incidence. In the present study with pigs, intestinal injections of M. smithii were made under surgery. We report here the beneficial effect of this microbial activity on DCS incidence after a chosen compression and decompression sequence in H2.
MATERIALS AND METHODS
Animals and training. Pigs (Sus scrofa, neutered male Yorkshires, n = 36, mean body mass ± 1 SE = 19.4 ± 0.2 kg; Table 1) were used for all experiments. The pigs were housed before experiments in an accredited animal care facility and had ad libitum access to water. The pigs were fed once daily with laboratory animal chow (Harlan Teklad, Madison, WI; 2% by body wt). All procedures were approved by an Animal Care and Use Committee. The experiments reported here were conducted according to the principles presented in the Guide for the Care and Use of Laboratory Animals (National Research Council, 1996).
Animals were trained to walk at a moderate pace (50-60 m/min) for intervals of 5 min on a treadmill in the laboratory, separated by intervals of 5 min rest, to acclimate them to a treadmill and to verify that this workload was not excessively strenuous. They were then trained to walk at a slower pace (30 m/min) for intervals of 5 min inside the compression chamber in 1 atm air, on a treadmill that constituted the floor of the chamber. The pigs were acclimated to being left confined and unattended in the chamber, where they were free to lie down. They were trained to rise and walk whenever the treadmill was activated. Treadmill performance was subsequently used to help evaluate an animal's gait for signs of DCS.
Preparation of M. smithii. A sample culture of M. smithii (strain PS) was obtained from Dr. Terry Miller (Wadsworth Center for Laboratories and Research, Albany, NY) and was grown in an atmosphere of H2-CO2 (80:20 vol/vol; 3 atm) at 37°C at the University of Georgia. Stock cultures were maintained in a complex medium similar to medium 1 (1) except for the following modifications: 3 g/l of yeast extract (Difco, Sparks, MD), 3 g/l of sodium formate, and 3 g/l of sodium acetate · 3 hydrate, and 7 g/l trypticase (BBL, Sparks, MD); lipoic acid and folic acid were omitted from the vitamin solution; and Fe(NH)2(SO4) · 6 H2O was substituted for FeSO4, and AlK(SO4)2 was omitted from the trace mineral solution.
Growth of M. smithii took place in a 14-liter fermentor using 11 liters of the modified medium 1 as described above. The fermentor was sparged with H2-CO2. At least 1 h before inoculation, the temperature was set to 37°C, and 10 ml of Na2S · 9 H2O (20% wt/vol) were added. The inoculum size was ~1% vol/vol of the total medium volume. During growth, the gas pressure was maintained at 2.4 atm, with a flow rate of 0.75 l/h of H2-CO2. A stirring rate of 240 rpm was used on the first day and 300 rpm on subsequent days. An additional 10 ml of sterile Na2S · 9 H2O were added twice daily until harvesting. Before harvesting, the rate of CH4 production ranged from at least 7 to 16 μmol CH4 · min1 · ml culture1, and the cell absorbance ranged from 2.0 to 3.1 OD600.
Cells were harvested with a Sharples continuous-flow centrifuge (model T-1P; Alfa Lavel, Warminster, PA) at 23,000 rpm. The cell paste was transferred to bottles that were flushed with N2, and the cells were resuspended with 2.3 ml of a buffer of 5 mM dithiothreitol and 5% wt/vol NaHCO3 per gram of cells. The bottles were flushed with H2-CO2 for 15 min before pressurization to 3 atm (absolute pressure). The resuspended cells were stored at 0°C and shipped with gel refrigerant to the Naval Medical Research Center within 12-24 h of harvesting. On arrival, the cell suspension bottles were flushed again with H2-CO2 and stored in a refrigerator for use within the next 24-48 h.
Before use, the cultures were assayed for their methanogenic activity by placing 0.2 ml culture and 0.2 ml of resuspension buffer in a 20-ml bottle with 3 atm H2-CO2. The bottle was incubated in a 37°C water bath with agitation at 200 rpm. Samples (100 μl) of the gas in the head space were taken every 12 min for 1-1.5 h and analyzed by gas chromatography for CH4 concentration.
Surgery. Animals were divided randomly into three groups (Table 1): those that were to undergo surgery to be treated with methanogens (treated; n = 16), those that were to undergo the same surgical procedure but given intestinal injections of saline (surgical controls; n = 10), and untreated animals (controls; n = 10). There were no significant differences in the body masses among these three groups of animals (Table 1).
Animals were given an extra meal late in the afternoon of the day before an experiment and were left unfed on the morning of the experiment. Treated and surgical control animals were prepared for surgery by preanesthetizing them with injections of ketamine HCl (20 mg/kg im; Fort Dodge Laboratories, Fort Dodge, IA) and xylazine (Rompun 2 mg/kg im; Bayer, Shawnee Mission, KS). Animals were then kept at a surgical plane of anesthesia with inhaled isoflurane (Abbott Laboratories, N. Chicago, IL) and O2. With use of aseptic technique, a midline incision of ~10 cm was made in the abdomen, and the cecum and spiral colon were exteriorized.
For the treated animals, 1-4 injections of M. smithii culture were made, depending on the volume of the microbial culture to be used; this volume ranged from 12 to 116 ml. The injectate was distributed between the cecum and upper, middle and lower spiral colon, with not more than 30 ml in each injection site. Total activity injected ranged from 200 to nearly 2,200 μmol CH4/min. For the surgical controls, saline that had been deoxygenated by bubbling it with CO2 was used as the injectate. One to three injections were made, of 60 ml total volume, also distributed between the cecum and various locations in the spiral colon. All needle puncture sites were sealed with a drop of surgical cement (Vetbond, 3M, St. Paul, MN). Postexperiment necropsy never revealed any visible leakage from these punctures. The intestines were moistened externally with saline and returned to the abdomen, and the incision was closed with sutures. As the animal recovered from the anesthesia, yohimbine (2 mg iv; Lloyd Laboratories, Shenandoah, IA) was injected into an ear vein to act as an antagonist to the xylazine. Animals appeared to be fully recovered from the anesthesia in 1-2 h. They were then placed in the compression chamber to commence the experiment. Food and water were freely available in the chamber.
Dive protocol. The compression chamber (5.7 m3 internal volume, WSF Industries, Buffalo, NY) and the facility in which it stood were specially designed for safe handling of high pressures of H2 (5). View ports on the chamber, as well as a video camera aimed through a view port, allowed ad libitum viewing of the animal.
For each experiment, one pig was placed in the compression chamber, standing on the treadmill. A stream of gas flowed continuously from the chamber to a gas chromatograph (GC; Model 5890A Series II, Hewlett-Packard, Wilmington, DE). Automated analysis occurred every 12 min. The gases analyzed were O2, N2, He, H2, and CH4. Calibrations were performed before each experiment. The thermal conductivity detector of the GC was calibrated with a certified gas mixture of 2% O2, 2% N2, 4% He, and 92% H2, a mixture that closely approximated the final gas composition in the chamber (ca. 2% O2, 1% N2, 9% He, 88% H2). The flame-ionizing detector of the GC was calibrated with 20 ppm CH4 in H2, a concentration that accommodated the full range of CH4 concentrations measured in the chamber (2-8 ppm). Calibrated values did not vary by more than 5% from standards. A calibration check at the end of each experiment confirmed that the machine did not drift more than 5%. The commercially purchased supplies of O2, He, and H2 were certified and confirmed to be below 0.1 ppm CH4. A gas blower was used to keep the gases in the chamber well mixed, as confirmed by analyzing samples drawn from two spatially separated locations in the chamber and finding not more than 5% difference.
The chamber was pressurized with He to an absolute pressure of 11 atm, adding O2 as necessary to replace O2 consumed by the pig and O2 lost by the ventilation of the chamber. Initial pressurization rate was selected by appearance of comfort for the animal's ears (ca. 0.15 atm/min for the first 2-3 atm, up to 0.45 atm/min at greater pressures). The chamber was then flushed with H2 while the chamber was maintained at a constant pressure of 11 atm, to a concentration of 60-75% H2, with O2 added as necessary to maintain normoxia (0.2 atm PO2). This initial pressurization with He, followed by replacement with H2, is necessary for avoiding explosive mixtures of H2 and O2 (5).
The chamber was further pressurized with H2 and O2 to a final pressure of 24 atm absolute pressure. Animals remained at this pressure for 3 h. The O2 partial pressure was maintained essentially constant throughout this portion of the dive, at the slightly elevated levels (0.3-0.5 atm PO2) that are customary for respiration under hyperbaric conditions (8). However, the percent H2 slowly increased over the period of hours spent at maximal pressure by a few percentage points, as chamber gases containing traces of He and N2 were replaced with H2 during chamber ventilation. The partial pressures of H2 during the final 36 min at maximal chamber pressure were 20.6-22.9 atm H2 and were not different among the three groups (Table 1).
After 2.5 h, animals were made to walk for 5 min on the chamber treadmill to observe their gait and to check that they appeared to be normal while compressed.
Three hours (± a few seconds) after arrival at maximal pressure, decompression to 11 atm occurred at a rate of 0.9 atm/min, while the animals were observed closely. When the chamber pressure arrived at 11 atm, animals were made to walk at 5-min intervals on the chamber treadmill for up to 1 h or until signs of severe DCS were noted and agreed on by at least three observers. Time of diagnosis was recorded to the nearest minute. The signs of DCS were primarily neurological and included falling, difficulty standing or righting after falling, and seizures. Some animals had labored breathing, which may have indicated cardiopulmonary DCS in addition to neurological DCS. Many animals were also observed to have signs of skin DCS (conspicuous lavender to dark purple mottling of the skin, with or without itching), but these signs alone did not warrant a diagnosis of severe DCS. Mild, transient behavioral changes (agitation or lethargy) were also not considered sufficient for a diagnosis of severe DCS. Once the diagnosis was made, or the hour had passed without evidence of DCS, the animal was euthanized quickly by asphyxiation with He. The chamber was returned to 1 atm after the animal was dead.
Throughout the dive, chamber temperature was thermostatically controlled to values that appeared to be comfortable for the animal (ca. 30-31°C at 11 atm and 32-34°C at 24 atm). Comfort was judged on the basis of absence of shivering with blanched skin or panting with dark pink skin. These temperatures are warmer than one would expect for comfort in 1 atm air, because of the high thermal conductivity of compressed H2 (10).
Analysis of CH4 release rate. An accurate estimate of chamber ventilation rate was needed to convert the CH4 concentration (ppm) in the chamber to a CH4 release rate (CH4; μmol CH4/min) from the animals. The flowmeter for analyzing chamber ventilation rate was calibrated with pure He, H2, O2, and N2, and with air, 2% O2 in He, and 2% O2 in H2, using a water spirometer. Ventilation rate estimates were reproducible within 2-4%. Linear interpolation based on relative percentages of gases was used to estimate the ventilation rates of all other gas mixtures. The mean ventilation rate used in these experiments at final chamber gas composition was 115 l/min (STP).
The total volume of gases in the pressurized chamber (136,800 liters STP) was very large compared with the chamber ventilation rate for reasons of economy and safety with H2. The rate at which an animal was releasing CH4 in the chamber was therefore not in equilibrium with CH4 sampling from the chamber in the time intervals we examined. This equilibrium would be reached under the conditions of this experiment only over a period of days (2, 6). Mathematical corrections had to be introduced to estimate the CH4, or actual rate of release of CH4 from the pigs, after extrapolating to equilibrium conditions. The approach of Bartholomew et al. (2) allows for this correction. The fractional concentration of CH4 in chamber gases at equilibrium (Xeq) was computed from CH4 concentration at two time points during the experiment (Xi and Xi1) separated by a known time interval (t) as
(2)
where is chamber ventilation rate and V is chamber effective volume (chamber physical volume times total pressure). The CH4 was then computed by multiplying Xeq by . This method was extensively calibrated and verified in our chamber, as described in detail by Fahlman (6). The CH4 values during the compression phase of the experiment (cCH4) were computed for each animal from a mean of five pairs of gas chromatographic measurements of CH4 concentration, with one member of the pair from the first hour at constant chamber pressure and the other member from the third hour (t = 2 h in each pair).
Statistical analysis. All mean values reported are ± 1 SE.
Multivariate logistic regression techniques (11) were used to determine the probability of DCS, using DCS outcome as the dependent variable and four experimental variables [body mass, chamber partial pressure of H2 (PH2), mean cCH4, and injected activity of microbes] as independent variables. Initially, a univariate analysis on each independent variable was performed; only those variables with a P value <0.20 (Wald test) were then included in a multivariate analysis. Exclusion of a variable from the multivariate analysis was based on the log-likelihood ratio test at the P = 0.05 level (11).
RESULTS
The concentration of CH4 within the chamber steadily increased throughout the 3-h duration of the compression phase of the experiment for all animals (Fig. 1). This does not, however, necessarily indicate that the animals were detectably increasing the rate at which they were releasing CH4 in this 3-h time frame. The progressive buildup of CH4 within the chamber was a consequence of the vast difference between the total volume of gases in the chamber and the sampling rate, as the system slowly approached equilibrium. When the data were mathematically corrected to equilibrium conditions, the cCH4 was best approximated by a single value for the entire 3-h period for each animal. This value was on average higher (P < 0.01, Mann-Whitney test) for animals treated with methanogens (92 ± 9 μmol CH4/min) compared with surgical and untreated controls combined (35 ± 4 μmol CH4/min; Table 1, Fig. 2).
There was a statistically significant correlation (P < 0.001, R = 0.70, slope = 0.046; least squares linear regression) between microbial activity injected and cCH4 (Fig. 2). This regression was minimally influenced by seemingly outlying values from two treated animals (P < 0.001, R = 0.74, slope = 0.039 with these 2 values omitted). The cCH4 was 6-12% of the in vitro activity injected (Fig. 2).
With the chosen compression and decompression sequence, 90% (9/10) of untreated control animals and 70% (7/10) of surgical control animals had symptoms of DCS (Fig. 3). The outcomes of these two groups are not significantly different from each other (2-tailed 2 test, P > 0.50). However, 44% (7/16) of animals receiving intestinal injections of M. smithii had symptoms of DCS (Fig. 3). This is significantly different from the DCS incidence for the two control groups combined (2-tailed 2 test, P < 0.05) and significantly lower than the DCS incidence of the surgical control group alone (1-tailed 2 test, P < 0.05). The mean time from arrival at 11 atm until the onset of DCS symptoms was 11.9 ± 2.1 min and did not differ significantly among the three groups (Table 1).
Logistic regression was used to determine whether there was a statistical correlation between the incidence of DCS in these animals and one or more variables within the experiments. These variables included body mass, PH2, cCH4, and injected activity of methanogens. Neither PH2 nor cCH4 was a significant predictor of DCS (P > 0.40; Table 2). Body mass and injected activity met the Wald test criteria (P < 0.20) for inclusion in a multivariate analysis. Results of this analysis indicated that injected activity was the only one of these variables that was a significant predictor of DCS (P < 0.05; Table 2, Fig. 4).
DISCUSSION
Animals receiving intestinal injections of H2-metabolizing microbes had only roughly half as many cases of DCS as untreated animals after a chosen sequence of compression and decompression with H2 (Fig. 3). This result is unlikely to be an artifact of the surgical procedure for injecting the microbes, on the basis of similarity in DCS incidence between untreated control and surgical control animals (Fig. 3). It was particularly important to test for effects of surgery, because elicitation of an immune reaction (which could potentially be triggered by a breach in asepsis during the surgical procedure) shortly before decompression has been implicated as a means of reducing DCS risk (16, 26).
There is evidence from these experiments that the reduction in DCS incidence is directly attributable to increased removal of H2 from within the animals by the injected microbes. The microbial activity given to an animal was a significant predictor of DCS (Table 2, Fig. 4). We confirmed that neither the body mass of the animal, which is a known risk factor in DCS studies with animals (20), nor the small variations between dives in the PH2 were a significant confounding factor in these experiments (Tables 1 and 2). Chamber temperature throughout the experiment, which was controlled within a narrow band, was also recorded and found not to be correlated with DCS incidence (6). Thus these experiments demonstrate that the benefits of microbial treatments in biochemical decompression respond in a dose-dependent manner in a pig model (Fig. 4).
Under other circumstances, H2 would be referred to as the inert gas component of the breathing mixture for these hyperbaric exposures. In this case, it is specifically the fact that H2, although inert to metabolism by mammalian cells (17, 22), is a highly energetic substrate for metabolism by some microbes that makes biochemical decompression possible. The selected microbe, M. smithii, converts some of the H2 to CH4 (Eq. 1), and there is no other source of CH4 in the chamber. Thus the release of CH4 from the animals is a means of tracking the rate of H2 scrubbing that is occurring within the animal during the hyperbaric H2 exposure.
Animals given the microbial injections generally released more CH4 than untreated animals throughout the experiment (Fig. 2). Surprisingly, the cCH4 was not a significant predictor of DCS risk in this study (Table 2), despite a significant correlation between microbial activity injected and cCH4 (Fig. 2). The cCH4 does, however, reach statistical significance as a predictor variable for DCS when we included data from over 100 experiments using other compression and decompression sequences than presented here (7). We interpret this as indicating that, although cCH4 is an invaluable noninvasive index of microbial activity within the animal, it is less than perfect. We do not know the kinetics associated with generating a given molecule of CH4 within the intestine of a pig vs. sampling that molecule by the gas chromatograph. There are uncertainties in the chromatographic analysis and in the mathematical corrections to equilibrium conditions, which are known to amplify any sampling errors (2). There is also the possibility that, within the complex microbial ecosystem of the large intestine, we have not accounted for all of the H2 metabolism solely by following CH4 release. Although methanogenesis is the primary pathway for microbial H2 consumption in the intestine (15, 21), there are microbes in the intestinal flora that can metabolize H2 to other end products such as acetate and sulfide (14).
We were particularly attracted to the two cases in which animals had unusually high cCH4 (Fig. 2). The two animals that released over 130 μmol CH4/min received microbial injections of 83 and 116 ml, respectively, with the latter the largest volume injected in these experiments. This leads us to speculate that distributing the microbial material over a larger volume of intestinal inner surface may increase the access of the microbes to H2, thereby increasing in vivo microbial activity. Greater microbial activity places animals at a lower risk of DCS (Fig. 4); nevertheless, these two animals displayed signs of DCS.
In contrast, there was an animal that received an injection of the second highest activity of the study (1,574 μmol CH4/min) but had a cCH4 that was similar to that of some control animals (Fig. 2). On examining our notes, we discovered that we had failed to give this animal the usual supplemental late afternoon meal on the day before the experiment. We speculate that this animal had a smaller volume of intestinal contents than others in the study, which in turn may have reduced intestinal blood perfusion and therefore H2 supply to the methanogens. Alternatively, the low intestinal content volume may have altered the internal intestinal environment in a manner that diminished methanogenic metabolism for some reason other than H2 availability. Low methanogenic activity places animals at a higher risk of DCS (Fig. 4); nevertheless this animal did not display signs of DCS.
This reminds us of a recurring theme in DCS research: some portion of DCS risk always manages to defy attribution to a specific risk factor and remains mathematically approachable only as a random event (28). It is significant that the present research has identified a controllable physiological factor that is predictive of DCS risk (Fig. 4), because DCS research is plagued by disputed attributions of risk to such factors as gender, adiposity, age, and body temperature (9, 13, 30). Consequently, most recent modeling efforts in DCS risk have included only physical aspects of the dive and parameter estimates for the gas kinetics of one or more compartments within animals; these compartments are mathematical constructs rather than physiological or anatomical entities (23, 24, 25).
Three additional experiments were performed in which animals were injected with M. smithii and remained at 24 atm in the chamber for 24 h (6). Within this extended time period, the cCH4 increased over time (Fig. 5). After 24 h, the mean cCH4 in each animal was severalfold higher than it had been in the first 3 h (6). We presume that this increasing cCH4 reflects increasing in vivo methanogenic activity over time and is attributable to a combination of microbial reproduction and microbial distribution throughout a greater portion of the large intestine. It should be noted that there was no appearance of discomfort to the animals due to intestinal gases either at the 3-h point or at these higher rates of CH4 release after 24 h. Only one of these three animals had subsequent symptoms of DCS, but a sample size of three is of course minimally instructive.
Some calculations may be of benefit in analyzing the magnitude of H2 removed by the microbes compared with the body burden of H2 dissolved in these animals. If we make the simplifying assumptions that a pig is approximately saturated with H2 after 3 h (6) and that H2 solubility throughout the pig can be estimated from the solubility of H2 in aqueous tissues (0.02 ml H2 · g1 tissues · atm1 at 37°C) (27), then a 20-kg pig breathing 21.7 atm PH2 (Table 1) for 3 h contains 8.7 liters of H2, or 340 mmol H2. The treated animals were releasing an average of roughly 90 μmol CH4/min for 3 h (Table 1, Fig. 2). This indicated that the methanogens were consuming at least four times that volume of H2 (Eq. 1). Thus nearly 65 mmol H2 were consumed in this process, i.e., ~19% (65/340) of the total volume dissolved in the animal.
Our laboratory reported previously that a 50% reduction in DCS incidence was achieved by biochemical elimination of an estimated 5% of the total body burden of H2 in rats (19). An even greater reduction in DCS incidence was found for pigs with native intestinal methanogens that eliminated an estimated 4-17% of their H2 load (18). It is difficult to make direct comparisons among these experiments and the present one, given the differences in compression and decompression sequences as well as body size. However, the message appears to be one consistent with findings from human studies: small differences in estimated tissue gas loads have a surprisingly large impact on DCS risk (12, 29).
Our laboratory created a mathematical model (7), using the data from the present study and from numerous other H2 exposures with pigs, that quantifies this relationship between tissue gas loss via biochemical decompression and DCS risk. Qualitatively, we envision that H2 biochemical decompression works by placing a sink for H2 within the body to supplement the H2 lost passively to the environment by diffusion during conventional decompression. So long as the H2 eliminated in the sink is of some threshold magnitude, it should not matter where that sink is located within the body. A lowering of mixed venous PH2 as a function of perfusing a sink tissue bed should subsequently reduce the alveolar PH2 and therefore reduce PH2 in all perfused tissues. Quantifying the rate or magnitude of gas elimination needed in this sink to have an impact on PH2 in the tissues most critical to DCS risk is precisely the ultimate goal of our research.
We conclude that the benefits of microbial treatments in H2 biochemical decompression are dose dependent in a pig model. It will be exciting but nontrivial to extend this work to practice with humans, because many factors in dosage, gas kinetics, and optimal microbial conditions remain unknown.
ACKNOWLEDGEMENTS
The quality and quantity of work on this project are directly attributable to the dedication, professionalism and friendship of our support team members, Richard Ayres, Jerry Morris, Roland Ramsey and Chief Anthony Ruopoli, U.S. Navy. Animal care, surgical instruction, and support were expertly provided by Colonel George McNamee, Veterinary Corps and Specialist Jesus Juarez, U.S. Army. We are grateful to Diana Temple for editorial assistance and critical reading of this manuscript.
FOOTNOTES
This work was funded by the Naval Medical Research and Development Command Work Unit no. 61153N MR04101.00D-1103 and the Office of Naval Research no. 603706N 00096 133 A0023. The opinions and assertions contained herein are the private ones of the authors and are not to be construed as official or reflecting the views of the Navy Department and the naval service at large.
All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee according to the principles set forth in the Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, National Research Council, National Academy Press, 1996.
Address for reprint requests and other correspondence: S. R. Kayar, Naval Medical Research Center, 503 Robert Grant Ave., Silver Spring, MD 20910-7500.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 30 January 2001; accepted in final form 4 April 2001.
REFERENCES
1.Balch, WE, Fox GE, Magrum LJ, Woese CR, and Wolfe RS. Methanogens: reevaluation of a unique biological group. Microbiol Rev 43: 260-296, 1979.
2.Bartholomew, GA, Vleck DL, and Vleck CM. Instantaneous measurements of oxygen consumption during pre-flight warm-up and post-flight cooling in sphingid and saturniid moths. J Exp Biol 90: 17-32, 1981.
3.Boycott, AE, Damant GCC, and Haldane JS. The prevention of compressed-air illness. J Hygiene (Lond) 8: 342-443, 1908.
4.Brauer, RW, and Naquet R. Contributions of animal experimentation to the establishment of appropriate conditions for ethical use of hydrogen as a component of deep diving atmospheres. In: Hydrogen as a Diving Gas, edited by Brauer R. W.. Bethesda, MD: Undersea and Hyperbaric Medical Society, 1987, p. 3-12. (33rd Undersea and Hyperbaric Medical Society Workshop)
5.Eichert, H, and Fischer M. Combustion-related safety aspects of hydrogen in energy applications. Int J Hydr Energy 11: 117-124, 1986.
6.Fahlman, A. On The Physiology of Hydrogen Diving and Its Implication for Hydrogen Biochemical Decompression (PhD thesis). Ottawa, Ontario, Canada: Carleton University, 2000.
7.Fahlman, A, Kayar SR, Tikuisis P, Lin WC, and Whitman WB. Predicting decompression sickness risk from H2 dives using conventional vs. biochemical decompression in pigs (Abstract). FASEB J 14: A612, 2000.
8.Fontanari, P, Badier M, Guillot C, Tomei C, Burnet H, Gardette B, and Jammes Y. Changes in maximal performance of inspiratory and skeletal muscles during and after the 7.1-MPa Hydra 10 record human dive. Eur J Appl Physiol 81: 325-328, 2000.
9. Francis TJR, Dutka AJ, and Hallenbeck JM. Pathophysiology of decompression sickness. In: Diving Medicine, edited by Bove AA and Davis JC. Philadelphia, PA: Saunders, chapt. 15, p. 170-187, 1990.
10.Homer, LD, and Kayar SR. Density, Heat Capacity, Viscosity, and Thermal Conductivity of Mixtures of CO2, He, H2, H2O, N2, and O2. Bethesda, MD: Naval Medical Research Institute, 1984, p. 3. (Technical Report NMRI 94-51)
11.Hosmer, DW, and Lemeshow S. Applied Logistic Regression. New York: Wiley, 1989.
12.Jacobsen, G, Jacobsen JE, Peterson RE, McLellan JH, Brooke ST, Nome T, and Brubakk AO. Decompression sickness from saturation diving: a case control study of some diving exposure characteristics. Undersea Hyperb Med 24: 73-80, 1997.
13.Jain, KK. Textbook of Hyperbaric Medicine. Lewiston, NY: Hogrefe and Huber, 1990, p. 128.
14.Jensen, BB, and Jorgensen H. Effect of dietary fiber on microbial activity and microbial gas production in various regions of the gastrointestinal tract of pigs. Appl Envir Microbiol 60: 1897-1904, 1994.
15. Jones WJ. Diversity and physiology of methanogens. In: Microbial Production and Consumption of Greenhouse Gases: Methane, Nitrogen Oxides and Halomethanes, edited by Rogers JE and Whitman WB. Washington, DC: Am. Soc. Microbiol., 1991, chapt. 3, p. 39-54.
16.Kayar, SR, Aukhert EO, Axley MJ, Homer LD, and Harabin AL. Lower decompression sickness risk in rats by intravenous injection of foreign protein. Undersea Hyperb Med 24: 329-335, 1997.
17.Kayar, SR, Axley MJ, Homer LD, and Harabin AL. Hydrogen gas is not oxidized by mammalian tissues under hyperbaric conditions. Undersea Hyperb Med 21: 265-275, 1994.
18.Kayar, SR, and Fahlman A. Decompression sickness risk reduced by H2 metabolism of native intestinal flora in pigs during H2 dives (Abstract). FASEB J 13: A408, 1999.
19.Kayar, SR, Miller TL, Wolin MJ, Aukhert EO, Axley MJ, and Kiesow LA. Decompression sickness risk in rats by microbial removal of dissolved gas. Am J Physiol Regulatory Integrative Comp Physiol 275: R677-R682, 1998.
20.Lillo, RS. Effect of N2-He-O2 on decompression outcome in rats after variable time-at-depth dives. J Appl Physiol 64: 2042-2052, 1988.
21.Miller, TL. Biogenic sources of methane. In: Microbial Production and Consumption of Greenhouse Gases: Methane, Nitrogen Oxides, and Halomethanes, edited by Rogers JE, and Whitman WB.. Washington, DC: Am. Soc. Microbiol., 1991, p. 298.
22. Séguin A and Lavoisier A. Premier mémoire sur la réspiration des animaux. Mém Acad Sci (Paris), 566-584, 1789.
23.Srinivasan, RS, Gerth WA, and Powell MR. Mathematical models of diffusion-limited gas bubble dynamics in tissue. J Appl Physiol 86: 732-741, 1999.
24.Thalmann, ED, Parker EC, Survanshi SS, and Weathersby PK. Improved probabilistic decompression model risk predictions using linear-exponential kinetics. Undersea Hyperb Med 24: 255-274, 1997.
25.Tikuisis, P, Gault KA, and Nishi RY. Prediction of decompression illness using bubble models. Undersea Hyperb Med 21: 129-143, 1994.
26.Ward, CA, McCullough D, and Fraser WD. Relation between complement activation and susceptibility to decompression sickness. J Appl Physiol 62: 1160-1166, 1987.
27.Weathersby, PK, and Homer LD. Solubility of inert gases in biological fluids and tissues: a review. Undersea Biomed Res 7: 277-296, 1980.
28.Weathersby, PK, Homer LD, and Flynn ET. On the likelihood of decompression sickness. J Appl Physiol 57: 815-825, 1984.
29.Weathersby, PK, Survanshi SS, Hays JR, and MacCallum ME. Statistically Based Decompression Tables III: Comparative Risk Using U. S. Navy, British, and Canadian Standard Air Schedules. Bethesda, MD: Naval Medical Research Institute, 1986, p. 2. (Technical Report NMRI 86-50)
30.Zwingelberg, KM, Knight MA, and Biles JB. Decompression sickness in women divers. Undersea Biomed Res 14: 311-317, 1987.(S. R. Kayar, A. Fahlman, )
2 Department of Microbiology, University of Georgia, Athens, Georgia 30602
ABSTRACT
The risk of decompression sickness (DCS) was modulated by varying the biochemical activity used to eliminate some of the hydrogen (H2) stored in the tissues of pigs (19.4 ± 0.2 kg) during hyperbaric exposures to H2. Treated pigs (n = 16) received intestinal injections of Methanobrevibacter smithii, a microbe that metabolizes H2 to water and CH4. Surgical controls (n = 10) received intestinal injections of saline, and an additional control group (n = 10) was untreated. Pigs were placed in a chamber and compressed to 24 atm abs (20.6-22.9 atm H2). After 3 h, the pigs were decompressed and observed for symptoms of DCS for 1 h. Pigs with M. smithii had a significantly lower (P < 0.05) incidence of DCS (44%; 7/16) than all controls (80%; 16/20). The DCS risk decreased with increasing activity of microbes injected (logistic regression, P < 0.05). Thus the supplemental tissue washout of the diluent gas by microbial metabolism was inversely correlated with DCS risk in a dose-dependent manner in this pig model.
keywords:Methanobrevibacter smithii; methanogens; biochemical decompression; decompression illness; hydrogen diving
INTRODUCTION
DECOMPRESSION SICKNESS (DCS) can be the penalty for failing to eliminate some portion of the tissue load of gas acquired while breathing under hyperbaric conditions (3). We report here studies on a novel approach to facilitating gas washout from tissues, thereby reducing the risk of DCS in an animal model. This approach, which we call biochemical decompression, uses microbes to metabolize hydrogen (H2) inside animals during a hyperbaric exposure with H2 as the primary gas in the breathing mixture. The conditions of the experiments are intended to simulate deep diving to 20-60 atm (700-2,000 feet of seawater). For such dives, H2 may be more appropriate than helium or nitrogen as the diluent to O2. Because of its lower density, H2 requires less lung ventilatory effort at high pressures (4). H2 has the additional advantage that it is less narcotic than N2 at high pressures and that some degree of H2 narcosis suppresses high-pressure neurological syndrome (4).
The concept of H2 biochemical decompression has been demonstrated in a rat model (19). Cultures of Methanobrevibacter smithii were injected into the proximal end of the large intestines of the animals (19). This microbe, which is native to the human intestinal flora (21), metabolizes H2 as
(1)
The microbes convert H2 to water and to a readily traceable end product (methane; CH4) that is derived exclusively from this reaction (17, 21). Rats supplied with M. smithii had a reduction in DCS risk by approximately half from a chosen compression and decompression sequence in H2 (19). Likewise, pigs that released detectable quantities of CH4 were found to have a lower risk of DCS from a chosen exposure to hyperbaric H2 compared with pigs that did not release CH4; this CH4 was generated entirely by microbes (of unknown species) native to the animals' intestinal flora (18). Thus these data demonstrated that the concept of biochemical decompression is valid.
Although most people on a western diet are expected to have M. smithii in their gut flora (21), it would not be acceptable to depend on the activity of this highly variable population for all biochemical decompression in human H2 diving. In order for H2 biochemical decompression to be useful for humans, the gut flora will need to be supplemented with M. smithii, probably by means of enteric-coated capsules of these microbes taken orally.
To advance this work further toward human use, we needed to determine the relationship between the in vitro activity of M. smithii delivered, the amount of H2 these microbes eliminated in vivo, and DCS incidence. In the present study with pigs, intestinal injections of M. smithii were made under surgery. We report here the beneficial effect of this microbial activity on DCS incidence after a chosen compression and decompression sequence in H2.
MATERIALS AND METHODS
Animals and training. Pigs (Sus scrofa, neutered male Yorkshires, n = 36, mean body mass ± 1 SE = 19.4 ± 0.2 kg; Table 1) were used for all experiments. The pigs were housed before experiments in an accredited animal care facility and had ad libitum access to water. The pigs were fed once daily with laboratory animal chow (Harlan Teklad, Madison, WI; 2% by body wt). All procedures were approved by an Animal Care and Use Committee. The experiments reported here were conducted according to the principles presented in the Guide for the Care and Use of Laboratory Animals (National Research Council, 1996).
Animals were trained to walk at a moderate pace (50-60 m/min) for intervals of 5 min on a treadmill in the laboratory, separated by intervals of 5 min rest, to acclimate them to a treadmill and to verify that this workload was not excessively strenuous. They were then trained to walk at a slower pace (30 m/min) for intervals of 5 min inside the compression chamber in 1 atm air, on a treadmill that constituted the floor of the chamber. The pigs were acclimated to being left confined and unattended in the chamber, where they were free to lie down. They were trained to rise and walk whenever the treadmill was activated. Treadmill performance was subsequently used to help evaluate an animal's gait for signs of DCS.
Preparation of M. smithii. A sample culture of M. smithii (strain PS) was obtained from Dr. Terry Miller (Wadsworth Center for Laboratories and Research, Albany, NY) and was grown in an atmosphere of H2-CO2 (80:20 vol/vol; 3 atm) at 37°C at the University of Georgia. Stock cultures were maintained in a complex medium similar to medium 1 (1) except for the following modifications: 3 g/l of yeast extract (Difco, Sparks, MD), 3 g/l of sodium formate, and 3 g/l of sodium acetate · 3 hydrate, and 7 g/l trypticase (BBL, Sparks, MD); lipoic acid and folic acid were omitted from the vitamin solution; and Fe(NH)2(SO4) · 6 H2O was substituted for FeSO4, and AlK(SO4)2 was omitted from the trace mineral solution.
Growth of M. smithii took place in a 14-liter fermentor using 11 liters of the modified medium 1 as described above. The fermentor was sparged with H2-CO2. At least 1 h before inoculation, the temperature was set to 37°C, and 10 ml of Na2S · 9 H2O (20% wt/vol) were added. The inoculum size was ~1% vol/vol of the total medium volume. During growth, the gas pressure was maintained at 2.4 atm, with a flow rate of 0.75 l/h of H2-CO2. A stirring rate of 240 rpm was used on the first day and 300 rpm on subsequent days. An additional 10 ml of sterile Na2S · 9 H2O were added twice daily until harvesting. Before harvesting, the rate of CH4 production ranged from at least 7 to 16 μmol CH4 · min1 · ml culture1, and the cell absorbance ranged from 2.0 to 3.1 OD600.
Cells were harvested with a Sharples continuous-flow centrifuge (model T-1P; Alfa Lavel, Warminster, PA) at 23,000 rpm. The cell paste was transferred to bottles that were flushed with N2, and the cells were resuspended with 2.3 ml of a buffer of 5 mM dithiothreitol and 5% wt/vol NaHCO3 per gram of cells. The bottles were flushed with H2-CO2 for 15 min before pressurization to 3 atm (absolute pressure). The resuspended cells were stored at 0°C and shipped with gel refrigerant to the Naval Medical Research Center within 12-24 h of harvesting. On arrival, the cell suspension bottles were flushed again with H2-CO2 and stored in a refrigerator for use within the next 24-48 h.
Before use, the cultures were assayed for their methanogenic activity by placing 0.2 ml culture and 0.2 ml of resuspension buffer in a 20-ml bottle with 3 atm H2-CO2. The bottle was incubated in a 37°C water bath with agitation at 200 rpm. Samples (100 μl) of the gas in the head space were taken every 12 min for 1-1.5 h and analyzed by gas chromatography for CH4 concentration.
Surgery. Animals were divided randomly into three groups (Table 1): those that were to undergo surgery to be treated with methanogens (treated; n = 16), those that were to undergo the same surgical procedure but given intestinal injections of saline (surgical controls; n = 10), and untreated animals (controls; n = 10). There were no significant differences in the body masses among these three groups of animals (Table 1).
Animals were given an extra meal late in the afternoon of the day before an experiment and were left unfed on the morning of the experiment. Treated and surgical control animals were prepared for surgery by preanesthetizing them with injections of ketamine HCl (20 mg/kg im; Fort Dodge Laboratories, Fort Dodge, IA) and xylazine (Rompun 2 mg/kg im; Bayer, Shawnee Mission, KS). Animals were then kept at a surgical plane of anesthesia with inhaled isoflurane (Abbott Laboratories, N. Chicago, IL) and O2. With use of aseptic technique, a midline incision of ~10 cm was made in the abdomen, and the cecum and spiral colon were exteriorized.
For the treated animals, 1-4 injections of M. smithii culture were made, depending on the volume of the microbial culture to be used; this volume ranged from 12 to 116 ml. The injectate was distributed between the cecum and upper, middle and lower spiral colon, with not more than 30 ml in each injection site. Total activity injected ranged from 200 to nearly 2,200 μmol CH4/min. For the surgical controls, saline that had been deoxygenated by bubbling it with CO2 was used as the injectate. One to three injections were made, of 60 ml total volume, also distributed between the cecum and various locations in the spiral colon. All needle puncture sites were sealed with a drop of surgical cement (Vetbond, 3M, St. Paul, MN). Postexperiment necropsy never revealed any visible leakage from these punctures. The intestines were moistened externally with saline and returned to the abdomen, and the incision was closed with sutures. As the animal recovered from the anesthesia, yohimbine (2 mg iv; Lloyd Laboratories, Shenandoah, IA) was injected into an ear vein to act as an antagonist to the xylazine. Animals appeared to be fully recovered from the anesthesia in 1-2 h. They were then placed in the compression chamber to commence the experiment. Food and water were freely available in the chamber.
Dive protocol. The compression chamber (5.7 m3 internal volume, WSF Industries, Buffalo, NY) and the facility in which it stood were specially designed for safe handling of high pressures of H2 (5). View ports on the chamber, as well as a video camera aimed through a view port, allowed ad libitum viewing of the animal.
For each experiment, one pig was placed in the compression chamber, standing on the treadmill. A stream of gas flowed continuously from the chamber to a gas chromatograph (GC; Model 5890A Series II, Hewlett-Packard, Wilmington, DE). Automated analysis occurred every 12 min. The gases analyzed were O2, N2, He, H2, and CH4. Calibrations were performed before each experiment. The thermal conductivity detector of the GC was calibrated with a certified gas mixture of 2% O2, 2% N2, 4% He, and 92% H2, a mixture that closely approximated the final gas composition in the chamber (ca. 2% O2, 1% N2, 9% He, 88% H2). The flame-ionizing detector of the GC was calibrated with 20 ppm CH4 in H2, a concentration that accommodated the full range of CH4 concentrations measured in the chamber (2-8 ppm). Calibrated values did not vary by more than 5% from standards. A calibration check at the end of each experiment confirmed that the machine did not drift more than 5%. The commercially purchased supplies of O2, He, and H2 were certified and confirmed to be below 0.1 ppm CH4. A gas blower was used to keep the gases in the chamber well mixed, as confirmed by analyzing samples drawn from two spatially separated locations in the chamber and finding not more than 5% difference.
The chamber was pressurized with He to an absolute pressure of 11 atm, adding O2 as necessary to replace O2 consumed by the pig and O2 lost by the ventilation of the chamber. Initial pressurization rate was selected by appearance of comfort for the animal's ears (ca. 0.15 atm/min for the first 2-3 atm, up to 0.45 atm/min at greater pressures). The chamber was then flushed with H2 while the chamber was maintained at a constant pressure of 11 atm, to a concentration of 60-75% H2, with O2 added as necessary to maintain normoxia (0.2 atm PO2). This initial pressurization with He, followed by replacement with H2, is necessary for avoiding explosive mixtures of H2 and O2 (5).
The chamber was further pressurized with H2 and O2 to a final pressure of 24 atm absolute pressure. Animals remained at this pressure for 3 h. The O2 partial pressure was maintained essentially constant throughout this portion of the dive, at the slightly elevated levels (0.3-0.5 atm PO2) that are customary for respiration under hyperbaric conditions (8). However, the percent H2 slowly increased over the period of hours spent at maximal pressure by a few percentage points, as chamber gases containing traces of He and N2 were replaced with H2 during chamber ventilation. The partial pressures of H2 during the final 36 min at maximal chamber pressure were 20.6-22.9 atm H2 and were not different among the three groups (Table 1).
After 2.5 h, animals were made to walk for 5 min on the chamber treadmill to observe their gait and to check that they appeared to be normal while compressed.
Three hours (± a few seconds) after arrival at maximal pressure, decompression to 11 atm occurred at a rate of 0.9 atm/min, while the animals were observed closely. When the chamber pressure arrived at 11 atm, animals were made to walk at 5-min intervals on the chamber treadmill for up to 1 h or until signs of severe DCS were noted and agreed on by at least three observers. Time of diagnosis was recorded to the nearest minute. The signs of DCS were primarily neurological and included falling, difficulty standing or righting after falling, and seizures. Some animals had labored breathing, which may have indicated cardiopulmonary DCS in addition to neurological DCS. Many animals were also observed to have signs of skin DCS (conspicuous lavender to dark purple mottling of the skin, with or without itching), but these signs alone did not warrant a diagnosis of severe DCS. Mild, transient behavioral changes (agitation or lethargy) were also not considered sufficient for a diagnosis of severe DCS. Once the diagnosis was made, or the hour had passed without evidence of DCS, the animal was euthanized quickly by asphyxiation with He. The chamber was returned to 1 atm after the animal was dead.
Throughout the dive, chamber temperature was thermostatically controlled to values that appeared to be comfortable for the animal (ca. 30-31°C at 11 atm and 32-34°C at 24 atm). Comfort was judged on the basis of absence of shivering with blanched skin or panting with dark pink skin. These temperatures are warmer than one would expect for comfort in 1 atm air, because of the high thermal conductivity of compressed H2 (10).
Analysis of CH4 release rate. An accurate estimate of chamber ventilation rate was needed to convert the CH4 concentration (ppm) in the chamber to a CH4 release rate (CH4; μmol CH4/min) from the animals. The flowmeter for analyzing chamber ventilation rate was calibrated with pure He, H2, O2, and N2, and with air, 2% O2 in He, and 2% O2 in H2, using a water spirometer. Ventilation rate estimates were reproducible within 2-4%. Linear interpolation based on relative percentages of gases was used to estimate the ventilation rates of all other gas mixtures. The mean ventilation rate used in these experiments at final chamber gas composition was 115 l/min (STP).
The total volume of gases in the pressurized chamber (136,800 liters STP) was very large compared with the chamber ventilation rate for reasons of economy and safety with H2. The rate at which an animal was releasing CH4 in the chamber was therefore not in equilibrium with CH4 sampling from the chamber in the time intervals we examined. This equilibrium would be reached under the conditions of this experiment only over a period of days (2, 6). Mathematical corrections had to be introduced to estimate the CH4, or actual rate of release of CH4 from the pigs, after extrapolating to equilibrium conditions. The approach of Bartholomew et al. (2) allows for this correction. The fractional concentration of CH4 in chamber gases at equilibrium (Xeq) was computed from CH4 concentration at two time points during the experiment (Xi and Xi1) separated by a known time interval (t) as
(2)
where is chamber ventilation rate and V is chamber effective volume (chamber physical volume times total pressure). The CH4 was then computed by multiplying Xeq by . This method was extensively calibrated and verified in our chamber, as described in detail by Fahlman (6). The CH4 values during the compression phase of the experiment (cCH4) were computed for each animal from a mean of five pairs of gas chromatographic measurements of CH4 concentration, with one member of the pair from the first hour at constant chamber pressure and the other member from the third hour (t = 2 h in each pair).
Statistical analysis. All mean values reported are ± 1 SE.
Multivariate logistic regression techniques (11) were used to determine the probability of DCS, using DCS outcome as the dependent variable and four experimental variables [body mass, chamber partial pressure of H2 (PH2), mean cCH4, and injected activity of microbes] as independent variables. Initially, a univariate analysis on each independent variable was performed; only those variables with a P value <0.20 (Wald test) were then included in a multivariate analysis. Exclusion of a variable from the multivariate analysis was based on the log-likelihood ratio test at the P = 0.05 level (11).
RESULTS
The concentration of CH4 within the chamber steadily increased throughout the 3-h duration of the compression phase of the experiment for all animals (Fig. 1). This does not, however, necessarily indicate that the animals were detectably increasing the rate at which they were releasing CH4 in this 3-h time frame. The progressive buildup of CH4 within the chamber was a consequence of the vast difference between the total volume of gases in the chamber and the sampling rate, as the system slowly approached equilibrium. When the data were mathematically corrected to equilibrium conditions, the cCH4 was best approximated by a single value for the entire 3-h period for each animal. This value was on average higher (P < 0.01, Mann-Whitney test) for animals treated with methanogens (92 ± 9 μmol CH4/min) compared with surgical and untreated controls combined (35 ± 4 μmol CH4/min; Table 1, Fig. 2).
There was a statistically significant correlation (P < 0.001, R = 0.70, slope = 0.046; least squares linear regression) between microbial activity injected and cCH4 (Fig. 2). This regression was minimally influenced by seemingly outlying values from two treated animals (P < 0.001, R = 0.74, slope = 0.039 with these 2 values omitted). The cCH4 was 6-12% of the in vitro activity injected (Fig. 2).
With the chosen compression and decompression sequence, 90% (9/10) of untreated control animals and 70% (7/10) of surgical control animals had symptoms of DCS (Fig. 3). The outcomes of these two groups are not significantly different from each other (2-tailed 2 test, P > 0.50). However, 44% (7/16) of animals receiving intestinal injections of M. smithii had symptoms of DCS (Fig. 3). This is significantly different from the DCS incidence for the two control groups combined (2-tailed 2 test, P < 0.05) and significantly lower than the DCS incidence of the surgical control group alone (1-tailed 2 test, P < 0.05). The mean time from arrival at 11 atm until the onset of DCS symptoms was 11.9 ± 2.1 min and did not differ significantly among the three groups (Table 1).
Logistic regression was used to determine whether there was a statistical correlation between the incidence of DCS in these animals and one or more variables within the experiments. These variables included body mass, PH2, cCH4, and injected activity of methanogens. Neither PH2 nor cCH4 was a significant predictor of DCS (P > 0.40; Table 2). Body mass and injected activity met the Wald test criteria (P < 0.20) for inclusion in a multivariate analysis. Results of this analysis indicated that injected activity was the only one of these variables that was a significant predictor of DCS (P < 0.05; Table 2, Fig. 4).
DISCUSSION
Animals receiving intestinal injections of H2-metabolizing microbes had only roughly half as many cases of DCS as untreated animals after a chosen sequence of compression and decompression with H2 (Fig. 3). This result is unlikely to be an artifact of the surgical procedure for injecting the microbes, on the basis of similarity in DCS incidence between untreated control and surgical control animals (Fig. 3). It was particularly important to test for effects of surgery, because elicitation of an immune reaction (which could potentially be triggered by a breach in asepsis during the surgical procedure) shortly before decompression has been implicated as a means of reducing DCS risk (16, 26).
There is evidence from these experiments that the reduction in DCS incidence is directly attributable to increased removal of H2 from within the animals by the injected microbes. The microbial activity given to an animal was a significant predictor of DCS (Table 2, Fig. 4). We confirmed that neither the body mass of the animal, which is a known risk factor in DCS studies with animals (20), nor the small variations between dives in the PH2 were a significant confounding factor in these experiments (Tables 1 and 2). Chamber temperature throughout the experiment, which was controlled within a narrow band, was also recorded and found not to be correlated with DCS incidence (6). Thus these experiments demonstrate that the benefits of microbial treatments in biochemical decompression respond in a dose-dependent manner in a pig model (Fig. 4).
Under other circumstances, H2 would be referred to as the inert gas component of the breathing mixture for these hyperbaric exposures. In this case, it is specifically the fact that H2, although inert to metabolism by mammalian cells (17, 22), is a highly energetic substrate for metabolism by some microbes that makes biochemical decompression possible. The selected microbe, M. smithii, converts some of the H2 to CH4 (Eq. 1), and there is no other source of CH4 in the chamber. Thus the release of CH4 from the animals is a means of tracking the rate of H2 scrubbing that is occurring within the animal during the hyperbaric H2 exposure.
Animals given the microbial injections generally released more CH4 than untreated animals throughout the experiment (Fig. 2). Surprisingly, the cCH4 was not a significant predictor of DCS risk in this study (Table 2), despite a significant correlation between microbial activity injected and cCH4 (Fig. 2). The cCH4 does, however, reach statistical significance as a predictor variable for DCS when we included data from over 100 experiments using other compression and decompression sequences than presented here (7). We interpret this as indicating that, although cCH4 is an invaluable noninvasive index of microbial activity within the animal, it is less than perfect. We do not know the kinetics associated with generating a given molecule of CH4 within the intestine of a pig vs. sampling that molecule by the gas chromatograph. There are uncertainties in the chromatographic analysis and in the mathematical corrections to equilibrium conditions, which are known to amplify any sampling errors (2). There is also the possibility that, within the complex microbial ecosystem of the large intestine, we have not accounted for all of the H2 metabolism solely by following CH4 release. Although methanogenesis is the primary pathway for microbial H2 consumption in the intestine (15, 21), there are microbes in the intestinal flora that can metabolize H2 to other end products such as acetate and sulfide (14).
We were particularly attracted to the two cases in which animals had unusually high cCH4 (Fig. 2). The two animals that released over 130 μmol CH4/min received microbial injections of 83 and 116 ml, respectively, with the latter the largest volume injected in these experiments. This leads us to speculate that distributing the microbial material over a larger volume of intestinal inner surface may increase the access of the microbes to H2, thereby increasing in vivo microbial activity. Greater microbial activity places animals at a lower risk of DCS (Fig. 4); nevertheless, these two animals displayed signs of DCS.
In contrast, there was an animal that received an injection of the second highest activity of the study (1,574 μmol CH4/min) but had a cCH4 that was similar to that of some control animals (Fig. 2). On examining our notes, we discovered that we had failed to give this animal the usual supplemental late afternoon meal on the day before the experiment. We speculate that this animal had a smaller volume of intestinal contents than others in the study, which in turn may have reduced intestinal blood perfusion and therefore H2 supply to the methanogens. Alternatively, the low intestinal content volume may have altered the internal intestinal environment in a manner that diminished methanogenic metabolism for some reason other than H2 availability. Low methanogenic activity places animals at a higher risk of DCS (Fig. 4); nevertheless this animal did not display signs of DCS.
This reminds us of a recurring theme in DCS research: some portion of DCS risk always manages to defy attribution to a specific risk factor and remains mathematically approachable only as a random event (28). It is significant that the present research has identified a controllable physiological factor that is predictive of DCS risk (Fig. 4), because DCS research is plagued by disputed attributions of risk to such factors as gender, adiposity, age, and body temperature (9, 13, 30). Consequently, most recent modeling efforts in DCS risk have included only physical aspects of the dive and parameter estimates for the gas kinetics of one or more compartments within animals; these compartments are mathematical constructs rather than physiological or anatomical entities (23, 24, 25).
Three additional experiments were performed in which animals were injected with M. smithii and remained at 24 atm in the chamber for 24 h (6). Within this extended time period, the cCH4 increased over time (Fig. 5). After 24 h, the mean cCH4 in each animal was severalfold higher than it had been in the first 3 h (6). We presume that this increasing cCH4 reflects increasing in vivo methanogenic activity over time and is attributable to a combination of microbial reproduction and microbial distribution throughout a greater portion of the large intestine. It should be noted that there was no appearance of discomfort to the animals due to intestinal gases either at the 3-h point or at these higher rates of CH4 release after 24 h. Only one of these three animals had subsequent symptoms of DCS, but a sample size of three is of course minimally instructive.
Some calculations may be of benefit in analyzing the magnitude of H2 removed by the microbes compared with the body burden of H2 dissolved in these animals. If we make the simplifying assumptions that a pig is approximately saturated with H2 after 3 h (6) and that H2 solubility throughout the pig can be estimated from the solubility of H2 in aqueous tissues (0.02 ml H2 · g1 tissues · atm1 at 37°C) (27), then a 20-kg pig breathing 21.7 atm PH2 (Table 1) for 3 h contains 8.7 liters of H2, or 340 mmol H2. The treated animals were releasing an average of roughly 90 μmol CH4/min for 3 h (Table 1, Fig. 2). This indicated that the methanogens were consuming at least four times that volume of H2 (Eq. 1). Thus nearly 65 mmol H2 were consumed in this process, i.e., ~19% (65/340) of the total volume dissolved in the animal.
Our laboratory reported previously that a 50% reduction in DCS incidence was achieved by biochemical elimination of an estimated 5% of the total body burden of H2 in rats (19). An even greater reduction in DCS incidence was found for pigs with native intestinal methanogens that eliminated an estimated 4-17% of their H2 load (18). It is difficult to make direct comparisons among these experiments and the present one, given the differences in compression and decompression sequences as well as body size. However, the message appears to be one consistent with findings from human studies: small differences in estimated tissue gas loads have a surprisingly large impact on DCS risk (12, 29).
Our laboratory created a mathematical model (7), using the data from the present study and from numerous other H2 exposures with pigs, that quantifies this relationship between tissue gas loss via biochemical decompression and DCS risk. Qualitatively, we envision that H2 biochemical decompression works by placing a sink for H2 within the body to supplement the H2 lost passively to the environment by diffusion during conventional decompression. So long as the H2 eliminated in the sink is of some threshold magnitude, it should not matter where that sink is located within the body. A lowering of mixed venous PH2 as a function of perfusing a sink tissue bed should subsequently reduce the alveolar PH2 and therefore reduce PH2 in all perfused tissues. Quantifying the rate or magnitude of gas elimination needed in this sink to have an impact on PH2 in the tissues most critical to DCS risk is precisely the ultimate goal of our research.
We conclude that the benefits of microbial treatments in H2 biochemical decompression are dose dependent in a pig model. It will be exciting but nontrivial to extend this work to practice with humans, because many factors in dosage, gas kinetics, and optimal microbial conditions remain unknown.
ACKNOWLEDGEMENTS
The quality and quantity of work on this project are directly attributable to the dedication, professionalism and friendship of our support team members, Richard Ayres, Jerry Morris, Roland Ramsey and Chief Anthony Ruopoli, U.S. Navy. Animal care, surgical instruction, and support were expertly provided by Colonel George McNamee, Veterinary Corps and Specialist Jesus Juarez, U.S. Army. We are grateful to Diana Temple for editorial assistance and critical reading of this manuscript.
FOOTNOTES
This work was funded by the Naval Medical Research and Development Command Work Unit no. 61153N MR04101.00D-1103 and the Office of Naval Research no. 603706N 00096 133 A0023. The opinions and assertions contained herein are the private ones of the authors and are not to be construed as official or reflecting the views of the Navy Department and the naval service at large.
All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee according to the principles set forth in the Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, National Research Council, National Academy Press, 1996.
Address for reprint requests and other correspondence: S. R. Kayar, Naval Medical Research Center, 503 Robert Grant Ave., Silver Spring, MD 20910-7500.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 30 January 2001; accepted in final form 4 April 2001.
REFERENCES
1.Balch, WE, Fox GE, Magrum LJ, Woese CR, and Wolfe RS. Methanogens: reevaluation of a unique biological group. Microbiol Rev 43: 260-296, 1979.
2.Bartholomew, GA, Vleck DL, and Vleck CM. Instantaneous measurements of oxygen consumption during pre-flight warm-up and post-flight cooling in sphingid and saturniid moths. J Exp Biol 90: 17-32, 1981.
3.Boycott, AE, Damant GCC, and Haldane JS. The prevention of compressed-air illness. J Hygiene (Lond) 8: 342-443, 1908.
4.Brauer, RW, and Naquet R. Contributions of animal experimentation to the establishment of appropriate conditions for ethical use of hydrogen as a component of deep diving atmospheres. In: Hydrogen as a Diving Gas, edited by Brauer R. W.. Bethesda, MD: Undersea and Hyperbaric Medical Society, 1987, p. 3-12. (33rd Undersea and Hyperbaric Medical Society Workshop)
5.Eichert, H, and Fischer M. Combustion-related safety aspects of hydrogen in energy applications. Int J Hydr Energy 11: 117-124, 1986.
6.Fahlman, A. On The Physiology of Hydrogen Diving and Its Implication for Hydrogen Biochemical Decompression (PhD thesis). Ottawa, Ontario, Canada: Carleton University, 2000.
7.Fahlman, A, Kayar SR, Tikuisis P, Lin WC, and Whitman WB. Predicting decompression sickness risk from H2 dives using conventional vs. biochemical decompression in pigs (Abstract). FASEB J 14: A612, 2000.
8.Fontanari, P, Badier M, Guillot C, Tomei C, Burnet H, Gardette B, and Jammes Y. Changes in maximal performance of inspiratory and skeletal muscles during and after the 7.1-MPa Hydra 10 record human dive. Eur J Appl Physiol 81: 325-328, 2000.
9. Francis TJR, Dutka AJ, and Hallenbeck JM. Pathophysiology of decompression sickness. In: Diving Medicine, edited by Bove AA and Davis JC. Philadelphia, PA: Saunders, chapt. 15, p. 170-187, 1990.
10.Homer, LD, and Kayar SR. Density, Heat Capacity, Viscosity, and Thermal Conductivity of Mixtures of CO2, He, H2, H2O, N2, and O2. Bethesda, MD: Naval Medical Research Institute, 1984, p. 3. (Technical Report NMRI 94-51)
11.Hosmer, DW, and Lemeshow S. Applied Logistic Regression. New York: Wiley, 1989.
12.Jacobsen, G, Jacobsen JE, Peterson RE, McLellan JH, Brooke ST, Nome T, and Brubakk AO. Decompression sickness from saturation diving: a case control study of some diving exposure characteristics. Undersea Hyperb Med 24: 73-80, 1997.
13.Jain, KK. Textbook of Hyperbaric Medicine. Lewiston, NY: Hogrefe and Huber, 1990, p. 128.
14.Jensen, BB, and Jorgensen H. Effect of dietary fiber on microbial activity and microbial gas production in various regions of the gastrointestinal tract of pigs. Appl Envir Microbiol 60: 1897-1904, 1994.
15. Jones WJ. Diversity and physiology of methanogens. In: Microbial Production and Consumption of Greenhouse Gases: Methane, Nitrogen Oxides and Halomethanes, edited by Rogers JE and Whitman WB. Washington, DC: Am. Soc. Microbiol., 1991, chapt. 3, p. 39-54.
16.Kayar, SR, Aukhert EO, Axley MJ, Homer LD, and Harabin AL. Lower decompression sickness risk in rats by intravenous injection of foreign protein. Undersea Hyperb Med 24: 329-335, 1997.
17.Kayar, SR, Axley MJ, Homer LD, and Harabin AL. Hydrogen gas is not oxidized by mammalian tissues under hyperbaric conditions. Undersea Hyperb Med 21: 265-275, 1994.
18.Kayar, SR, and Fahlman A. Decompression sickness risk reduced by H2 metabolism of native intestinal flora in pigs during H2 dives (Abstract). FASEB J 13: A408, 1999.
19.Kayar, SR, Miller TL, Wolin MJ, Aukhert EO, Axley MJ, and Kiesow LA. Decompression sickness risk in rats by microbial removal of dissolved gas. Am J Physiol Regulatory Integrative Comp Physiol 275: R677-R682, 1998.
20.Lillo, RS. Effect of N2-He-O2 on decompression outcome in rats after variable time-at-depth dives. J Appl Physiol 64: 2042-2052, 1988.
21.Miller, TL. Biogenic sources of methane. In: Microbial Production and Consumption of Greenhouse Gases: Methane, Nitrogen Oxides, and Halomethanes, edited by Rogers JE, and Whitman WB.. Washington, DC: Am. Soc. Microbiol., 1991, p. 298.
22. Séguin A and Lavoisier A. Premier mémoire sur la réspiration des animaux. Mém Acad Sci (Paris), 566-584, 1789.
23.Srinivasan, RS, Gerth WA, and Powell MR. Mathematical models of diffusion-limited gas bubble dynamics in tissue. J Appl Physiol 86: 732-741, 1999.
24.Thalmann, ED, Parker EC, Survanshi SS, and Weathersby PK. Improved probabilistic decompression model risk predictions using linear-exponential kinetics. Undersea Hyperb Med 24: 255-274, 1997.
25.Tikuisis, P, Gault KA, and Nishi RY. Prediction of decompression illness using bubble models. Undersea Hyperb Med 21: 129-143, 1994.
26.Ward, CA, McCullough D, and Fraser WD. Relation between complement activation and susceptibility to decompression sickness. J Appl Physiol 62: 1160-1166, 1987.
27.Weathersby, PK, and Homer LD. Solubility of inert gases in biological fluids and tissues: a review. Undersea Biomed Res 7: 277-296, 1980.
28.Weathersby, PK, Homer LD, and Flynn ET. On the likelihood of decompression sickness. J Appl Physiol 57: 815-825, 1984.
29.Weathersby, PK, Survanshi SS, Hays JR, and MacCallum ME. Statistically Based Decompression Tables III: Comparative Risk Using U. S. Navy, British, and Canadian Standard Air Schedules. Bethesda, MD: Naval Medical Research Institute, 1986, p. 2. (Technical Report NMRI 86-50)
30.Zwingelberg, KM, Knight MA, and Biles JB. Decompression sickness in women divers. Undersea Biomed Res 14: 311-317, 1987.(S. R. Kayar, A. Fahlman, )