Metabolism of Dog Gastric Mucosa I. NUCLEOTIDE LEVELS IN PARIETAL CELLS*

and pyridine nucleotide levels as well as those of phosphate, phosphocreatine, lactate, pyruvate, B-hydroxybutyrate, acetoacetate, glucose, and glycogen were measured in histologically defined parietal and mucous cell sections of biopsies of dog gastric mucosa at rest, and in various secretory states. As a result of stimulation of secretion, there appeared to be no change in adenine nucleotide levels. or phosphocreatine, but there was a rise in inorganic phosphate and a fall in phosphorylation potential. However, there was a marked increase in NADH, but no change in NADPH with onset of acid secretion. The increase in the lactate to pyruvate ratio showed that the increased NADH level occurred in the cytoplasm and these data are discussed with reference to change in cell pH.

Adenine and pyridine nucleotide levels as well as those of phosphate, phosphocreatine, lactate, pyruvate, B-hydroxybutyrate, acetoacetate, glucose, and glycogen were measured in histologically defined parietal and mucous cell sections of biopsies of dog gastric mucosa at rest, and in various secretory states. As a result of stimulation of secretion, there appeared to be no change in adenine nucleotide levels. or phosphocreatine, but there was a rise in inorganic phosphate and a fall in phosphorylation potential. However, there was a marked increase in NADH, but no change in NADPH with onset of acid secretion. The increase in the lactate to pyruvate ratio showed that the increased NADH level occurred in the cytoplasm and these data are discussed with reference to change in cell pH.
Proton transport is an almost universal phenomenon occurring in subcellular organelles such as mitochondria (1) and chromatophores (2), as well as in unicellular organisms such as Escherichia coli (3) or yeast (4), and in plants (5) or animal tissues such as kidney (6), pancreas (7), or stomach (8). The latter tissue is probably the most specialized in this function, the mammalian parietal cells being capable of developing a 106.6-fold concentration gradient of H+. Since it is unlikely that a totally novel mechanism evolved in gastric mucosa, it seems that an understanding of the transport process in this organ would be generally applicable to the problem of H+ transport in most situations.
Measurement of metabolite levels in gastric mucosa therefore has as its main purpose the determination of the primary energy source for H+ secretion. Thus, measurement of adenine nucleotide and phosphocreatine levels may be able to answer the question as to whether ATP is the primary energy source, and extension of these measurements to other metabolites should be able to provide information as to the pathway or pathways stimulated to provide the primary energy for transport.
Logically, the tissue metabolites should be studied before and after stimulation, and a comparison of the steady state levels should indicate physiologically significant changes. Two types of measurements have been made. Nondestructive techniques such as dual or split beam spectroscopy (9, 10) have provided information on the steady state changes in the oxidation-reduction components of the mitochondrial respiratory chain, and destructive techniques such as described here have given data on adenine nucleotide levels (11,12). However, these past experiments have been confined almost exclusively to in vitro studies of frog gastric mucosa where the change in H' rate is limited, and moreover no distinction has been made between the multiple cell types present in the tissues. The work reported in this and succeeding papers deals with measurements in biopsy samples from an intact mammalian mucosa, in what were determined histologically to be parietal cell regions weighing about 1 pg. The fluorimetric cycling procedures developed by Lowry and his collaborators (13) were used to obtain the necessary sensitivity. This paper reports data obtained for adenine and pyridine nucleotides and metabolites related to the "free" pyridine nucleotide levels in cytosol and mitochondria under resting (i.e. nonsecreting), secretory onset, and steady state secretion in dog gastric mucosa.

Gastric
Preparation-Adult beagle dogs were anesthetised with Nembutal.
The abdomen was incised and the gastric mucosa exposed by an incision along the greater curvature.
Resting tissue samples were taken at 15 and 30 min after surgery, at which time histamine infusion was started at a rate of 100 &kg/hour to give maximal secretion.
In some experiments, 100 kg of urecholine were used in addition to histamine to increase secretion further. After 8 to 10 min of infusion, the gastric mucosa reddened, which occurred 1 to 2 min prior to onset of acid. With histamine and urecholine, a faster onset was seen. A sample was taken from the reddened nonsecreting mucosa during this period and labeled "blood flow" sample. Onset of acid secretion was monitored by measuring the fall of pH in 1 ml of saline (0.9% NaCl solution) held on the mucosa by means of a Lashley suction cup (2.27 cm') for 3 min. When pH started to fall, tissue samples were taken at that time (HI) and another when the pH stabilized at about 2.0 in the approximately 10.fold diluted sample (HII). H+ rate was obtained by titration to pH 7.0 ( Fig. 1). The acid rate for histamine alone was one-third that found with histamine and urecholine in combination, as known previously. Precautions were taken to control bleeding, and to keep the mucosa moist and warm at all times. Dogs which showed cyanotic mucosae were rejected as bad experiments.
Sampling Procedure--Samples were taken by precooled rectal biopsy forceps, with scalpel excision of the piece of tissue (-100 mg wet weight).
The tissue sample was frozen immediately (<3 sl in liquid nitrogen, and stored at -80" until use. Holding the tissue at room temperature for a period of 30 s did not change the adenine or pyridine nucleotide levels, hence the rapidity of the sampling procedure was deemed adequate.
Phosphocreatine decreased and lactate increased in this 30-s procedure, but since no change was found in phosphocreatme m the 3-s sample, this was further evidence for adequately rapid sampling.
Treatment of Samples-The samples were placed in a Harris cryostat at -20" and serially sectioned (20 Frn thick) until parietal cell areas were seen, as evidenced by the darker appearance of the tissue. The initial sections consisted exclusively of surface or glandular neck cells. The frozen sections were placed in drilled aluminum blocks at -20" and dried under vacuum at -40". Every fourth frozen section was transferred to a glass slide and stained for succinic dehydrogenase activity using succinate and mtro blue tetrazolium (14); in this manner the parietal cell area was clearly identifiable (Fig. 2 Table III. This is the method used for determination of levels of the pyridine nucleotides, and also for the pyridine nucleotide produced in the other substrate assay methods with, therefore, the omission of the initial step destroying the oxidized or reduced form of the pyridine nucleotide, which step is performed slightly differently for each substrate as detailed in Table I. Also, the amount of cycling enzyme varies for each substrate, as a function of the number of cycles required and this is detailed in Table II. The other reagent levels remain the same. Indicator Step  the NADP' cycle, 6.phosphogluconate was measured directly using 6-phosphogluconate dehydrogenase (Table IV). The sequence of steps is summarized in Fig. 3 for the entire procedure. Materials-All enzymes were obtained from Boehringer-Mannheim

Acid Secretion
As shown in Fig. 1    but at steady state secretion there is apparently a decrease in the level of oxidized nucleotide. The level of this nucleotide is lower than in the mucous cells. NADPH-Corresponding to the values of NADP+ found in the parietal cell region of the mucosa there is no significant change in NADPH levels, during the stages of acid secretion (Table VII). Mucous cells have a slightly lower value of NADPH than parietal cells (Table VI). No change occurs in the mucous cells with secretion.
NADPH/NADP+ Ratio-The value for this ratio does not change during onset of acid secretion, but at the steady state level there is a slight rise in the value, with no change in mucous cells.
Pyruuate- Table  VIII shows that there is a progressive rise in pyruvate from the resting sample to blood flow and to secretion in the parietal cells; the levels in the mucous cells are about the same as in the parietal cells (Table VI), but do not change with the onset of acid secretion.  Lactate-Lactate levels double with the onset of acid secretion and are maintained during steady state secretion (Table VII). The lactate level in the mucous cells is similar to that of resting parietal cells and shows no change (Table VI).
LactatelPyruvate Ratio-The lactate/pyruvate ratio rises with onset of acid secretion to nearly twice its previous value, but does not change in mucous cells.
Acetoacetate- Table  VIII shows that there are no changes observed with this keto acid over the period of observation. The value obtained for gastric mucosa is higher than that observed for control fed rat liver, for example (22), but lower than others have found for fasted liver (23) and higher than in rat brain (24).
&Hydroxybutyrate-The level of the reduced product of acetoacetate, as was found for lactate, shows an increase with onset of acid secretion, which is maintained with continued secretion (Table VIII). The level found, however, is low compared to levels obtained for fasted liver. P-HydryoxybutyratelAcetoacetate Ratio-There is a progressive rise in this value from onset of acid secretion into the steady state H+ secretion sample, the final result being an approximate doubling of the ratio found. It should be noted that this ratio is the inverse of what is found in rat liver (22), under either fed or fasted conditions. Glucose-The level of glucose remained quite constant over the period of observation, showing an insignificant 10% increase (Table IX).
Glycogen-There is a gradual continuing decline in glycogen levels with onset and steady state acid secretion (Table IX), with a net 15% fall.
Znorganic Phosphate-There is a significant rise in levels of inorganic phosphate found in parietal cells with acid secretion, but no change with blood flow (Table IX), the maximal change being about 30%. The rise in this metabolite is unexpected in view of the lack of change of the adenine nucleotides and phosphocreatine, and may be due to entry of phosphate from the blood.

DISCUSSION
Hypotheses as to the energy source for acid secretion have had a major effect on the concepts prevalent in the field. The earliest suggestion postulated that protons could be derived   The values are the means +S.E. of n dogs, at least four measurements having been made on each sample in each dog. Blood flow refers to the state of reddening of the mucosa, HI the sample when the acid secretion was detected in that area, and HII a sample at steady state secretion. Values are given, when appropriate, in millimoles kg-' dry weight.

Substance
Resting  The values are the means +S.E. of n dogs, at least four measurements having been made on each sample in each dog. Blood flow refers to the state of reddening of the mucosa, HI the sample when acid secretion was detected in that area, and HI1 a sample at steady state secretion. Values are given, when appropriate, in millimoles kg-' dry weight. (n = 11) (n = 11) (n = 11) (n = 10) PI 6.42 * 0.14 6.38 i 0.37 8.33 zt 1.30" (n = 5) (n = 5) (n = 5) "p < 0.05.
directly from oxidation-reduction of a substrate with secretion of H+ and donation of electrons to an unspecified acceptor (25). This oxidation-reduction hypothesis was amplified further, based on work in yeast (26, 27) and data obtained from dual-split beam spectroscopy (28) have been interpreted as supporting an oxidation-reduction mechanism.
On the other hand, measurements showing correlations between ATP levels and acid secretion (29, 30) or results obtained from the use of inhibitors (31, 32) have been interpreted as for (29) or against (30) an ATP-based mechanism. The discovery of ATPase in gastric mucosal membranes, either HCO $--or K +-stimulated (33, 34) has also been used to support an ATP hypothesis. Measurement of adenine nucleotide and phosphocreatine levels in frog mucosa during onset or increase of acid rate has also implicated ATP as the primary energy source (35). The conflict of interpretation of the data is reminiscent of the problems in the field of oxidative phosphorylation.
Here the proposal of chemiosmotic coupling (36) whereby the breakdown or synthesis of ATP is linked intimately to the generation or dissipation of an H+ or potential gradient, has in a sense blurred the distinction between ATP-and oxidationreduction-linked mechanisms. Since techniques have not been developed.for flux measurements of adenine and pyridine nucleotides, steady state levels of these components were measured; the hypothesis was that, because of the large change in energy consumption by the tissue, at some point in the transition between rest, tissue perfusion, and varying secretory states, critical steady state changes would be observed.
The primary reaction, that of H+ transport, may be the site of stimulation of metabolism. The largest changes in the tissue should be observed in the acid-secreting cells, the parietal cells, and changes in surface epithelial cells should be minimal. The data obtained showed no change in levels of ATP and other adenine nucleotides, as well as no change in phosphocreatine, for either the parietal or the mucous cells. Moreover there was no significant difference in the levels found when the histamine and histamine + urecholine experiments were compared, in spite of the widely differing acid rates.
There is a variety of possible interpretations. The high mitochondrial content of the parietal cells (> 32%) suggests that this tissue may well be capable of a rapid response to widely varying energy demands, so that there is little change in ATP or phosphocreatine.
In this case, however, since there is no change in total ADP, the primary metabolic stimulus for the increased 0, consumption (37) must be elsewhere in metabolism or due to changes in intramitochondrial ADP. It may be that mitochondrial metabolism is substrate-limited, rather than ADP-limited, i.e. that the mitochondria in gastric mucosa are in State 2 rather than State 4, as has been suggested elsewhere (28).
Compartmentation of the various nucleotides between mito-chondrion and cytoplasm, also occurs, although this does not apply to phosphocreatine. Thus it is possible to calculate or to measure the phosphorylation potential ATP/ADP .P, (38). The calculation assumes cytoplasmic localization and equilibrium operation of lactic dehydrogenase, glyceraldehyde-3-POa dehydrogenase, and phosphoglycerate kinase. where G3P is glyceraldehyde-3-P, 3PGA is 3-P-glycerate, and LDH is lactate dehydrogenase and HPO,'-is taken as 60%) of total P,. Substituting the data from Table VIII, and the levels of glyceraldehyde-3-PO, and 3phosphoglycerate presented in detail in a subsequent paper and in Ref. 39, the results of Table  X are found. It can be seen that the cytoplasmic and total phosphorylation potential fall by 25 to 350/c, but that the cytoplasmic potential is twice as large as the total. This may be due to the altered energy state of the mitochondria.
A possible artifact to be considered is a change in parietal cell pH, which has often been suggested, but rarely measured (28). Measurements in amphibian mucosa have suggested increases ranging from 0.3 to 1 pH unit.' A fall in cell [H+] will result in activation of phosphofructokinase as well as a decrease in the NAD+/NADH, hence [pyruvate]/ [lactate] ratios. Since pgruvate is the oxidized form, and glyceraldehyde-%phosphate is the reduced form of the two oxidation-reduction couples involved in calculation of the phosphorylation potential, changes in the ratio of these due to a shift in NAD+/NADH due to a change of pH will cancel, assuming equilibrium operation. Data from frog gastric mucosa (35) which show a slight fall in ATP and a rise in ADP may be due to poorer oxygenation of the in vitro mucosa, with also slower removal of HCO,-, with subsequent hypoxia and alkalinization of the parietal cell, the latter consistent with the effects of increased CO, in vitro (40). The increase in phosphocreatine observed (35) was explained by the shift in cell pH. Since, however, the equilibrium constant for creatine kinase is 7.2 x 10m9 at pH 7.4, and 2.98 x 10 ' at pH 9.8 (41), the predicted change as a function of pH would be a fall in phosphocreatine, rather than a rise, or a fall in ADP with a rise in ATP. Hence, the constancy, or increase of phosphocreatine is difficult to reconcile with ADP driven metabolic transition. It should be pointed out that constancy in ATP, and ATP/ADP ratios is also found in muscle during a work load (42). However, associated with this, there is a fall in phosphocreatine contrary to what was observed here, and in frog mucosa (35).
Since the ratio [ATP].
[AMP]/[ADP]' remains fairly constant at about 0.5 and since the equilibrium constant of the reaction, the inverse of the above ratio, is 2.26 at pH 7.4. it can be concluded that gastric parietal cells contain sufficient adenylate kinase to maintain this equilibrium.
In contrast to the constancy of adenine nucleotide level, we observed consistent and significant increases in the NADH level. Although the NAD+/NADH ratio fell from 6.68 to 3.81, the NAD + level was still greater than NADH. Previous work in this area has been confined to spectroscopy where the data show reduction of pyridine nucleotide with stimulation of ' S. J. Hersey, personal communication secretion (10, 28) or at least reduction of the rest of the respiratory chain (39), but some disagreement exists for this finding (9). The spectroscopic data naturally do not distinguish di-or triphosphopyridine nucleotides, nor the cell type or cellular location of the change. NADP+ and NADPH change significantly with acid rate changes, and the NAD +/NADH ratio doubles in the mucous cell, hence the spectroscopic data probably do indeed relate to the parietal cell (oxyntic cell in amphibia).
It is possible to calculate the NADH/NAD + ratio in the cytoplasm and mitochondria (43) on the assumption that (a) lactic dehydrogenase and cu-gly-cerophosphate dehydrogenase: P-hydroxybutyrate dehydrogenase and glutamic dehydrogenase are located in the cytoplasm or mitochondria exclusively, (b) their activity is sufficient to maintain equilibrium, (c) there is no pH gradient between mitochondria and cytoplasm. The equilibrium constants for lactic dehydrogenase is 1.11 x IO-" M, and hence from the equation n' AL) ' [lactate] The cytoplasmic ratio NAD+/NADH at pH 7.4 is calculated to be 620 and to change to 348 with secretion. The ratio changes of a-glycerophosphate and dihydroxyacetone phosphate are similar to the lactate/pyruvate changes (39), hence the assumption that these dehydrogenases are sufficiently active to maintain equilibrium seems justified (Table XI). This may well not be so for the /%HO-butyrate in stomach which has low activity in relation to liver. In the liver, the direction of change of acetoacetatelp-hydroxybutyrate and cu-ketoglutarate NHJglutamate correspond, showing identity of the NAD+/NADH ratio in mitochondrial membrane and matrix. In the stomach, whereas the acetoacetate/butyrate ratio falls, the oc-ketoglutarate ratio increases (39, and subsequent papers) by 50%. This implies that the @-hydroxybutyrate dehydrogenase has too low an activity to maintain equilibrium in parietal cells, and assuming at least no fall in NH,, that the mitochondrial NAD+ pool is relatively oxidized with onset of acid secretion. The ratio for cytoplasm is much larger than the measured ratio, presumably due to greater binding of NADH relative to NAD + (43).
The influence of pH on this equilibrium is large. Thus, a change of pH from 7.4 to 7.7, as has been suggested to occur in in vitro frog mucosa, would be sufficient to induce a Z-fold change in the ratio, as observed here. Since, however, there is no difference between the two maximal secretory rates, where an additional change in cell pH might be expected, and also since there is no change in the NADP+/NADPH ratio which would be equally pH-sensitive, it may be that in uiuo, little change in cell pH occurs. Also the rise in mucous cell NAD+/NADH argues against an alkalinization in this cell type which is coupled electrically to the parietal. Resolution of this problem will have to await direct measurements of pH change in uiuo, and an attempt to distinguish between cytoplasmic and mitochondrial pH. In some dogs a sample was taken after prolonged anesthesia and secretion. In those dogs there was a significant fall in ATP and phosphocreatine, as well as a large rise in lactate and NADH. These data could be mimicked by holding the biopsy sample in the forceps for 3 min prior to freezing. Hence anoxia could be a significant factor in these samples. Maintaining anoxia in a biopsy from resting and secreting mucosa did not give significant differences in ATP fall between rest and secretion. Data obtained for rat gastric mucosa (total thickness) show similar changes with hemorrhagic shock (45). The mucous cells showed little change in any of the metabolites with onset of acid secretion. This shows a lack of effect of histamine on this cell type in the dog.