Postgrad Med J: first published as 10.1136/pgmj.60.709.743 on 1 November 1984. Downloaded from

Postgraduate Medical Journal (November 1984) 60, 743-750

Pepsinogen BASIL I. HIRSCHOWITZ M.D., F.R.C.P., F.R.C.P.E., F.A.C.P. Director, Division of Gastroenterology, University ofAlabama, Birmingham, Alabama, U.S.A.

Introduction via acetylcholine, and . was still a In 1951 life for the student of pepsinogen was quite theoretical secretagogue, and the Zollinger-Ellison simple. We believed we had learned most ofwhat we syndrome was yet to be described. We had only one needed to know in the 115 years since the first specific antagonist, atropine, though we did know discovery and naming of pepsin by Schwann in that the recently discovered antihistamines did not 1836*. In 1880 Heidenhain reported that the gastric antagonize the gastric effects of histamine. Between mucosa contained specialized cells which, respec- those who studied man and those who studied cats tively, secreted and pepsin; and in 1881-1886, and dogs, there was much argument as to whether Langley performed and reported his classical experi- histamine did or did not stimulate pepsin secretion. ments describing the cycle of the cells and Both were right, of course. The measurement of their granules. He also defined and crystallized the uropepsin and the nearly contemporaneous Azure-A precursor ofpepsin, pepsinogen, which he recognized tubeless gastric analysis were seen as providing by copyright. to be the contents of the granules of the peptic cell. diagnostic shortcuts to objectively diagnose upper The first assignment Avery Jones gave me was to gastrointestinal (GI) disease, including gastric cancer study 'uropepsin', first described by Brucke in 1861, and peptic ulcer, and to sort out the ulcer-prone and the subject of a recent revival of interest (Mirsky (Mirsky et al., 1948). et al., 1948). What made it worth restudy was the Since 1951, the study of pepsinogen has benefited availability of a reliable method for measuring from spectacular advances in molecular biology, cell proteolytic activity by the timed of acidified biology, chemistry, knowledge of drugs, haemoglobin (Anson and Mirsky, 1932). This hormones, and receptors, in cell separation, immuno- method could be used to measure with relative ease logy and tissue culture. pepsin in urine, blood and gastric juice as well as in In these intervening years we have learned that http://pmj.bmj.com/ other fluids or tissue extracts; the same method is still there are many major and minor structural and size widely used and is the standard by which the activity variants of pepsinogen; the genes for human, rat and of commercially available pepsin(ogen) is measured swine pepsinogen have been identified, and the (Sigma Chemical Co.). sequences of many pepsinogens are In 1951 we knew that pepsinogen was a protein of known. Pepsinogens have been classified by electro- molecular weight (mol. wt). 43,000 and that it was phoretic mobility and immunological characteristics; activated to pepsin by acidification with the loss of a this classification has been further amplified by number of , totalling about 5,000 mol.wt. cellular and anatomic localization, and by distribu- on September 26, 2021 by guest. Protected (Northrop, Kunitz and Herriot, 1948). At that time tion in blood, semen and urine. These findings have we knew that pepsinogen occurred in all stomachs, been related to disease and heredity. but it was believed to be a molecule of the same Developments in the field of neurohormonal con- weight and structure in all species. The histology of trol of secretion have illustrated the complexity ofthe the peptic cell in mammals and the combined cell in integrative physiology of digestion. The workings of submammals was known from light microscopy. protein-secreting cells are better known; receptors However, the processes whereby protein was synthe- have come under intense and systematic study, and sized, transported in the cell and excreted were new drugs have made such studies easier and more poorly understood, and the structure and function of specific. Finally peptic ulcer disease is better under- DNA and RNA were still unknown. Improved stood and better defined. quantitative methods had begun to define the rates of Studies of pepsin have advanced knowledge on secretion, and we knew only two stimuli: the vagus two parallel tracks. On the one hand, we know more about protein synthesis and secretion in general and *For historical references see Hirschowitz (1957). protein in particular. On the other hand, the Postgrad Med J: first published as 10.1136/pgmj.60.709.743 on 1 November 1984. Downloaded from

744 B. I Hirschowitz application to diagnosis and understanding of upper two groups, pepsinogen I and II (PGI and PGII) GI disease has progressed, perhaps more slowly, against which specific and non-crossing since other technologies in diagnosis and treatment can be developed. PGI comprises Pg 1-5 and PGII, have tended to overshadow pepsinogens. Pgs 6 and 7. What follows is a brief presentation of the present Such a classification appears to be further vali- status of the chemistry, cell biology, physiology and dated by the anatomic distribution in the , pathophysiology of pepsinogen. antral and duodenal mucosa and the 2-3 cm of cardiac glands surrounding the oesophagogastric junction contain only PGII, while the main body of the stomach contains both PGI and II with PGI pre Chemistry dominating. Furthermore, while serum contains both Pepsinogen, the inactive precursor of pepsin, is PGI and PGII, urine contains only PGI (uropepsin) secreted from peptic cells. In an acid medium, and semen only PGII, originating in the prostate. through a multistage process, pepsinogen is con- Before 32 weeks gestation, human amniotic fluid con verted to pepsin by a loss of the N terminal sequence ains PGI, and from 32-40 weeks, both PGI and II. consisting of a variable number of amino Moreover, PGI and II have different bond (Kageyama and Takahashi, 1980a, b); further con- specificities with artificial dipeptide substrates (Sam- formational changes occur in the remaining protein loff, 1983), e.g. I>II for tyrosyl-, but to expose two catalytic sites (Tang, 1977). The II>I for tyrosyl-alanine, -threonine, -, and activation of pepsinogen is slow at pH 6-0, but -serine. The catalytic rate ratio of PGI and PGII for extremely rapid at pH 2 or less. The activation bovine haemoglobin is 3:4, while egg albumin is peptide and analogues of this peptide have some digested by both at a much lower rate in the ratio of anticatalytic activity, but this is of no practical 0-25:0-1. The pH optima for PGI range from 1-5 to consequence in the protein digestive function of 2-0, and for PGII (gastricsin), the pH optimum is 3-2. in the stomach. PGI is more sensitive to denaturation alkali pepsin by (pHby copyright. Once activated, pepsin is liable to irreversible 7-2 vs pH 8-0 for PGII) and less so to heat (PGI 16% denaturation at pH above 7-2, by temperatures above vs PGII 80% at 62°C after 15 min at pH 2-4) (Samloff, 65°C, and by high concentrations ofurea. In contrast, 1983). pepsinogen is not denatured by pH as high as 10 or at Thus, there is much diversity in de- temperatures up to boiling in salt-free solutions. This pending on , pH, temperature, solute and difference in alkali denaturation led Langley a substrate concentration. Some methods for quantita- century ago to distinguish pepsinogen from pepsin. tion of pepsinogen are insensitive and probably Pepsin digests protein with a pH optimum for provide quite inadequate data, e.g. radial diffusion digestion of haemoglobin or albumin of 1-8 to 2-3, (see Ishague and Bardhan, 1978) and studies that use but as has been known for many years, the acid- egg albumin or casein substrate or those that depend active gastric have a wide optimal pH upon milk-clotting. The use of haemoglobin sub- http://pmj.bmj.com/ range. The other most clearly identified acid strate at pH 2-0 is still recommended as an inexpen- with a pH optimum of about 2-3-5 was called sive assay with a long and reliable history (Anson ; and yet another protease, gastricsin, ex- and Mirsky, 1932). hibits a pH optimum of 3-2. However, the pH Heterogeneity of pepsinogens is apparently due to optimum of pepsin for different substrates may vary, substitutions at the N-terminal end (Kageyama and thus milk clotting, another action of this , Takahashi, 1980a, b, c). However, immunological occurs at 5-5. The of of PGI and PGII remains valid across pH chemistry protein hydrolysis grouping on September 26, 2021 by guest. Protected has been extensively reviewed (Tang, 1977). species with differing N-terminal sequences or even Pepsinogen can be electrophoretically separated with pepsinogen of different sizes; thus the unusually into multiple bands representing decreasing electro- low molecular weight frog pepsinogen reacts with negativity or anodal mobility (Richmond et al., 1958; human PGI antiserum (Shugerman et al., 1982). Seijffers, Segal and Miller, 1963; Kushner, Rapp and The original belief that all pepsinogens and pep- Burtin, 1964; Samloff, 1969). There appear to be sins were nearly the same size was disproved by the about seven pepsinogens (Pg 1-7) that conform to the finding from my laboratory of a pepsinogen of mol. operational definition of inactive precursors resistant wt. 29,000 daltons (Shugerman et al., 1982). We also to alkaline pH, but which, upon acidification, convert found that pepsinogens from several species, includ- to acid active proteases with pH optima near 2-0 and ing crustaceans, amphibians, fish and mammals susceptibility to inactivation above pH 7-2. Each of varied from 29,000 to 65,000, unpublished values the seven (or more) pepsinogens apparently give rise which are outside the general range of 35,000 to a unique and equivalent pepsin. These seven to 48,000 daltons described for mammals and pepsinogens may be grouped immunologically into birds (Yasugi and Mizuno, 1981). As well, Yasugi Postgrad Med J: first published as 10.1136/pgmj.60.709.743 on 1 November 1984. Downloaded from

Pepsinogen 745 and Mizumo described a chicken-embryo specific- or fusion of the zymogen granule membrane with the pepsinogen of mol. wt 56,000 and corresponding cell membrane, the reduction in surface tension at the pepsin of 53,000, which resemble adult pepsinogen in site of adherence and the thinning and the ultimate its alkali stability and optimal pH, but differed in opening of the granule interior to the outside of the sensitivity to inhibition by pepstatin and its antigene- cell with discharge of granule contents. Since it is not ity. Human fetal pepsinogen is also different from clear that the granule contents are under pressure, the that in the adult. It is of some interest that a fetal type release of these contents must be an active process, pepsinogen has been isolated from a gastric cancer achieved either by the flow of the rest of the granule (Hirsch-Marie et al., 1976). membrane to the level of the cell membrane (trampo- Pepsinogens from various species have been se- line phenomenon) or an unseen action of contractile quenced (Kageyama and Takahashi 1980a, b, c; elements (slingshot phenomenon). The mechanisms Muto et al., 1980). Even more importantly, the gene whereby the granule approaches the membrane is for human pepsinogen has been identified (Sogawa et unknown; the best bet is again the contractile al., 1983) from a recombinant clone using swine elements of the cytoskeleton, activated via Ca+, pepsinogen cDNA as a probe. It occupies 9.4 calmodulin and protein kinase. I know of no data kilobases of DNA and comprises 9 exons and 8 describing changes in either membrane which pre- introns of various lengths, and produces a 373 amino cede or which would promote fusion. At any rate, the acid protein that is 82% homologous with swine fusion of the zymogen granule and cell membranes pepsinogen. As a matter ofparticular interest, there is under stimulation (exocytosis) begs the question of a 15-amino acid residue at the NH2 terminus which forces, electrostatic or other, normally keep indicating that the pepsin precursor is synthesized, granules from fusing with each other or with the cell like many other , as a pre-pepsinogen. The membrane and how, for example, atropine would two homologous sequences, including the two active cause the cell to become overstuffed with granules site aspartyl residues present in different coding without granule fusion. These granules stay the same that the evolved but increase in even the segments, suggest pepsinogen gene size, greatly number; so, by copyright. by duplication of a shorter ancestral gene. Ichihara, granule and cell membranes still remain unfused and Sogawa and Takahashi (1982) used rat gastric secretion is much diminished (Hirschowitz, O'Leary mucosa RNA to produce rat pre-pepsinogen with a and Marks, 1960). Also unknown is whether, after 16 amino acid extension at the NH2 terminal. They exocytosis, the fused granule membrane is recycled to also reported molecular cloning of complementary new granules or whether all granule membranes are DNA to swine pepsinogen mRNA (Sogawa et al., synthesized de novo. 1981). Since the synthesis of protein depends on the The second major effect of stimulation upon the production and control of mRNA, the key to further secreting cell is to initiate the process of sustained understanding of the action of stimuli of pepsinogen protein secretion. This step is probably more depen- secretion lies in understanding fully how secretion dent upon cAMP. This sequence ofevents is even less and secretagogues affect the synthesis and turnover well understood than exocytosis, because we do not http://pmj.bmj.com/ of mRNA in the peptic cell. know whether the receptors act from the surface or are internalized, whether they act in the cytosol or in of the cell the nucleus. We know from general principles that Biology peptic mRNA is produced and that mRNA in turn migrates The peptic cell shares with all exocytotic cells most to the very abundant rough endoplasmic reticulum ofthe processes leading from stimulation to secretion. (RER) where protein, in this case prepepsinogen, is Thus interact with to and an unknown method to secretagogues specific receptors produced transported by on September 26, 2021 by guest. Protected initiate either acute cationic movements or the the Golgi apparatus where it is encapsulated (Hir- stimulation of adenylcyclase to produce cAMP. A schowitz, 1967a, b). However, during prolonged further series of steps involving calmodulin, protein active stimulation, granules are depleted, suggesting kinases and the cytoskeleton lead ultimately to the that newly synthesized pepsinogen may be rapidly release by exocytosis of products stored in encapsu- transported and secreted without passing through the lated packets. granule or storage stage. The intracellular pathways The steps from the activation of second messengers for such a phenomenon are not well delineated, leading to secretion of protein product are at present though most likely the microtubular system of the not clearly understood. There are probably three cell is involved. major events involved: (1) the prompt release of The uptake from interstitial fluid or medium ofthe stored cell product; (2) stimulation of new synthesis amino acids used in protein synthesis must also be and sustained secretion of product; and (3) the stimulated. We do not know whether there are uptake of amino acids. The first process, which is specific transport mechanisms or requirements, such probably Ca+ + dependent, must involve conjunction as Na+, for the uptake of amino acids and how, after Postgrad Med J: first published as 10.1136/pgmj.60.709.743 on 1 November 1984. Downloaded from

746 B. I. Hirschowitz passing the basolateral membrane, these amino acids trin, which stimulated in vivo, but not in vitro. In are directed or transported to the site of protein isolated vectorially oriented peptic cells from the dog, synthesis. histamine was found to stimulate electrolyte tran- These questions should be much more readily sport, but not pepsinogen secretion (Ayalon et al., answerable now that isolated peptic cells can be 1982). arranged in single cell thickness sheets on a collagen matrix with tight junctions and vectorially oriented and can be shown to exhibit appropriate resistance Physiology and potential difference measurements (Ayalon et al., Pepsinogen secretion in intact animals and man. In 1982). the intact stomach of the mammal, the peptic cells Pepsinogen is secreted by specific gastric mucosal are co-mingled with parietal cells even to the extent cells in mammals, the chief cells of Heidenhain, of forming tight junctions. Though under many, if whereas in birds, fish, amphibia and reptiles, pepsi- not most, conditions, peptic and parietal cells func- nogen is synthesized and secreted in the same cells tion in concert, they are capable of independent that secrete HC1. One interesting exception is the function (Hirschowitz, 1967a, b). Thus, there are well localization of dedicated peptic cells in densely defined circumstances where peptic cells are actively packed glands in the lower oesophagus of frogs. In stimulated while parietal cells are inhibited (and vice the stomach of frogs, however, as in other submam- versa); thus insulin inhibition of acid secretion in the mals, pepsinogen is also secreted in the mixed dog does not affect the simultaneous hypoglycaemic- function acid-secreting oxyntic cells. Pepsinogen II is vagal stimulation of the peptic cells whilst in also secreted in much smaller amounts by antral and vivo reduces acid, while stimulating pepsin in both prostatic cells. man and dog and both acid and pepsin in the chicken Studies in in vitro tissues have begun to identify (Burhol, 1982). In the opposite condition, administra- various stimuli of pepsinogen secretion and to tion of histamine to dogs and cats will strongly suggest second messenger pathways. In isolated stimulate acid secretion yet inhibit pepsin secretion. rabbit and guinea-pig glands, Both actions are mediated via H2 receptors (Hir-by copyright. (CCK), but not secretin (Koelz et al., 1982), is a schowitz and Hutchison, 1977a; Hirschowitz, Rentz potent stimulus, while caerulein and gastrin G-17 are and Molina, 1981). Atropine added to histamine about 1/10th as potent. However, these peptides are further accentuates the discrepancy, while vagal or equipotent with CCK as stimuli of acid secretion cholinergic stimulation during histamine infusion (Hersey, May and Schuberg, 1983). Cholinergic will stimulate pepsin secretion without necessarily stimulation is reported for all preparations (Koelz et further affecting acid secretion (Hirschowitz, 1967a, al., 1982; Hersey et al., 1983; Simpson, Goldenberg b). In birds, such as the chicken, in which acid and and Hirschowitz, 1980) as is isoproterenol (Koelz et pepsin are secreted by one cell, it was also possible to al., 1982; Hersey et al., 1983; Shirakawa et al., 1983), separate responses of the two cell products with some a weaker stimulus. Both are blocked by their specific stimuli (Burhol, 1982), suggesting different messen- http://pmj.bmj.com/ antagonists, atropine (Inoue et al., 1983) and propa- ger pathways specific to each cell product. nolol (Shirakawa et al., 1983). In the frog, bombesin There are clear species differences in the response is also a potent stimulus of pepsinogen secretion of peptic and parietal cells to stimuli. Thus, in man (Shirakawa et al., 1983). Ca + removal partly or histamine stimulates acid and pepsin equally, and the completely inhibits most stimuli (Koelz et al., 1982; specific H2 receptor antagonists equally reverse the Inoue et al., 1983), and dibutyryl cAMP stimulates in stimulation of both. Gastrin also stimulates both all preparations. All stimuli increased cAMP pro- well. While the vagus stimulates both acid and pepsin duction in the presence of the phosphodiesterase via muscarinic pathways, only the acid secretion is on September 26, 2021 by guest. Protected inhibitor IBMX (Shirakawa et al., 1983; Inoue et al., inhibited by H2 antagonists. Since H2 antagonists 1983). The combined evidence suggests that both inhibit acid secretion resulting from all stimuli, but Ca++ and cAMP are involved in stimulation by all pepsin only when stimulated by histamine, it seems active stimuli-CCK in rabbit and guinea-pig, cho- clear that the peptic cell, unlike the , linergic and adrenergic in mammals and the frog, neither has an absolute dependence upon histamine and bombesin in the frog. From the handful of nor is approached by all stimuli via an H2 histamine papers on isolated peptic glands reported so far, it has pathway. The stomachs of humans respond poorly to not been possible to ascribe either/or specificity of the acetylcholine analogues, bethanechol, carbachol second messengers, i.e. either Ca + or cAMP, to any or metacholine, which are potent stimuli in dog and one of the known stimuli of pepsinogen secretion. cat acting via muscarinic M-1 (high affinity for the Correlation of in vitro data with data from the intact novel antimuscarinic drug, pirenzepine (Hirschowitz, animal agrees with respect to cholinergic stimulation, Fong and Molina, 1983) receptors. In dogs H2 but disagrees regarding secretin, histamine and gas- antagonists non-competitively inhibit acid secretion Postgrad Med J: first published as 10.1136/pgmj.60.709.743 on 1 November 1984. Downloaded from

Pepsinogen 747 only, leaving cholinergically stimulated pepsin secre- Combining data from intact systems with those tion intact (Hirschowitz and Molina, 1983). Basal from isolated tissues or cells, we may conclude that secretion of both acid and pepsin in man is choliner- the peptic cell probably has receptors for gic and inhibitable by very small doses of atropine acetylcholine in all species, including the frog. (Hirschowitz, Molina and Ou Tim, 1984). Histamine H2 receptors mediate stimulation ofpepsi- In dogs and cats, histamine stimulates pepsin nogen secretion in man; both stimulation and inhibi- secretion at very low doses via high affinity (low tion in dogs and cats; neither in rabbit or frog. dose) H2 receptors and dose-responsively inhibits Gastrin, for which no good specific antagonist is pepsin secretion via low affinity (high dose) H, available, stimulates peptic cells in man and dog, and receptors (Hirschowitz and Hutchison, 1977a, b; is very sensitive to atropine (Hirschowitz and Hutchi- Hirschowitz et al., 1981). No such inhibition is seen in son, 1977a, b), but stimulates poorly in rabbits and man. In the dog, gastrin stimulation of pepsinogen not at all in frogs. Both histamine and gastrin are secretion is somewhat less potent than acid stimula- good stimuli of the mixed function cells of the tion. In rats, pigs and monkeys, histamine stimulates chicken (Burhol and Hirschowitz, 1971; Burhol, pepsin secretion as in man. 1982). CCK, however, is a weak stimulant in man, One unexplained stimulus for pepsin secretion is dogs and chickens, stronger in the rabbit, but inactive luminal application of HC1 (Bynum and Johnson, or nearly so in frogs. 1975). This phenomenon has been ascribed to an osmotic effect by one group (Puurunen, 1979) and to initiation of local cholinergic reflexes by another Pathophysiology of pepsinogen (Bynum and Johnson, 1975). A change in the luminal In duodenal ulcer (DU) patients as a group, under membrane of the peptic cell promoting fusion with basal or stimulated conditions, pepsin output, like granules is another possibility. However, under many acid output, is on the average 1-5 to 2 times greater conditions pepsin secretion may be very low in the than in controls (Hirschowitz, 1984). DU males presence of high concentrations of acid (Hirschowitz, secrete more acid and pepsin per kg body weight than 1967b). do DU females. The overlap between DU and by copyright. Receptor-spec#ic inhibitors. Specific inhibitors be- controls amounts to perhaps 50% of DU and 85% of have predictably-thus atropine blocks equally the controls. There is no difference between patients with acid and pepsin stimulated by cholinergic agonists. DU or Zollinger-Ellison syndrome and non-ulcer The actions of atropine also define both basal subjects in the relative ratios of acid to pepsin secretion in man (the dog and cat do not have basal secretion (Hirschowitz, 1983; Aly and Emas, 1982), secretion), and direct vagal stimulation of the fundus suggesting that the gastric hyperplasia seen in many, in all species as acting via muscarinic pathways but not all, DU equally affects all cells of the gastric (Hirschowitz et al., 1983, 1984). mucosa. It has been reported that pepsin secretion is Likewise, histamine effects are specifically anta- higher in active than inactive or healed DU (Elder gonized by histamine H2 antagonists. The controls of and Smith, 1975; Achord, 1981). This could not be http://pmj.bmj.com/ acid and pepsin secretion in intact systems are less confirmed (Hirschowitz, 1984). The attempts to find clear because of the cross-over effects of antagonists, consistent or useful patterns of electrophoretically e.g. H2 receptor antagonism of gastrin or vagal/cho- separable pepsins in gastric juice of patients with linergic stimulation (Hirschowitz and Molina, 1983), gastric or duodenal ulcer, gastritis or cancer (Elder atropine inhibition of gastrin (Hirschowitz and and Smith, 1975; Samloff, 1983; Taylor, 1970; Hutchison, 1977a, b) and, to a lesser extent, of Walker and Taylor, 1980; Walt, Roberts and Taylor, histamine stimulation (Hirschowitz, Hutchison and 1979) have not yielded much firm data. Little of Sachs, 1973). practical clinical value has emerged from these on September 26, 2021 by guest. Protected Other inhibitors. inhibits both acid studies. Treatment of DU with pepsin antagonists has and pepsin, most potently when acting against not been found useful, especially in the era of H2 cholinergic stimuli, less so against gastrin and hardly antagonists. at all against histamine (Hirst et al., 1982). Prosta- In gastric ulcer, pepsin secretion, like acid secre- glandin E2 has the reverse order of potency (Mihas, tion, is the same as in controls. After vagotomy for Gibson and Hirschowitz, 1976). How these agents act DU, those without ulcer secrete in the normal range, or by what mechanism they are equally effective in but those with recurrent ulcer are closer to unoper- the disparate cell types is unknown. Presumably, they ated DU, probably because of incomplete or inade- act upon second messenger steps that both cell types quate surgery. In , pepsin and acid have in common. From the pattern of antagonism secretion both decline to the point of virtual or one would expect that somatostatin acts on the Ca+ + complete absence, as in pernicious anaemia. dependent pathways and prostaglandin E2 on the Much has been written about the measurement of cAMP oathwavs. urinary or serum pepsinogen. In the 1940s and 1950s Postgrad Med J: first published as 10.1136/pgmj.60.709.743 on 1 November 1984. Downloaded from

748 B. I. Hirschowitz these measurements involved total proteolytic activ- control (Samloff et al., 1975)-the same proportion ity. No direct relationship could be established as for total pepsinogen (Hirschowitz, 1955). In a few between gastric pepsin secretion and either serum or Zollinger-Ellison cases even higher levels were seen urine pepsinogen. In young army recruits, those with (Samloff et al., 1975). Recurrent ulcers have more a high serum pepsinogen had a much higher later PGI than do successfully operated DU patients incidence of DU than those with normal values (Samloffet al., 1975), again confirming earlier reports (Mirsky et al., 1948), though the incidence was not based on total proteolytic activity (Hirschowitz, high enough to be useful as a general marker. As a 1955). In cases of fundic atrophy, PGI tends to group, patients with duodenal ulcer have signifi- disappear, but, since the antrum remains intact in cantly higher levels of serum and urine pepsinogen most pernicious anaemia patients, PGII persists than controls; 60% of DU had serum pepsinogen undiminished (Samloff, 1982). There are, however, greater than controls (Hirschowitz, 1953, 1955). more direct ways of diagnosing pernicious anaemia. Stimulation and inhibition of gastric secretion was In two families of patients with DU and elevated not reflected by corresponding urine or serum levels serum PGI, 50% of first degree relatives with elevated (Hirschowitz, 1955). In pernicious anaemia or after PGI, distributed as an autosomal dominant, were total gastrectomy, serum and urinary pepsinogen liable to have DU (Rotter et al., 1979), whereas those levels are extremely low but not entirely absent, with normal serum PGI were free of DU. An indicating other sources of pepsinogen, from prostate elevated PGI might thus serve as a marker for DU, and leucocytes. In partially gastrectomized DU though no correlation with any haplotypes, blood without recurrent ulcer, levels were as low as in types or other possible markers has been so far normal, while those with recurrent ulcer had higher clearly established (Walt et al., 1979; Weitkamp et al., levels. Thus, in all respects, urine and serum pepsino- 1975). gen levels were of no greater clinical or diagnostic What, then, is the potential value of measuring value than gastric analysis and fell out of favour, PGI and II in blood and urine? For diagnosis of especially with the rapid development of universal gastric or duodenal disease, the value of serum PGI upper GI fibreoptic endoscopy after 1960. measurement appears limited, especially since theby copyright. With the development of radioimmunoassay for test involves an expensive radioimmunoassy, un- PGI and II, interest has again revived in the possible likely to be performed regularly in any clinical clinical usefulness of measurements of serum pepsi- laboratory. The same is true of pernicious anaemia nogens (Samloff, 1983). The correlation between PGI and Zollinger-Ellison syndrome, for which better and and total proteolytic activity in the serum is good more specific tests are available. Similarly for the and, in fact, would predict that the pattern of diagnosis of duodenal, gastric or recurrent postopera- distribution of PGI values would be very similar to tive peptic ulcer, and of gastric cancer or gastritis, that found with chemical serum pepsinogen measure- fibreoptic endoscopy, though more expensive, is ments 30 years ago (Hirschowitz, 1955). Such was the easier and definitive. For population screening,

case in a large study involving over 900 subjects perhaps for closely defined genetic studies, measure- http://pmj.bmj.com/ (Samloff, Liebman and Panitch, 1975), and the ment of PGs and variants in blood and urine may diagnostic usefulness in individual cases is no greater prove useful in a research context. Certainly measur- now than it was then. However, several interesting ing PGI is not a shortcut to studying gastric secretion findings have emerged from these studies. directly. As mentioned earlier, urinary pepsinogen is all PGI except in premature infants and in cases ofrenal Postscript failure with proteinuria, suggesting that PGI and II I am most grateful to Avery for many things, not are both filtered, but that PGII is reabsorbed least for introducing me into a field which has yielded on September 26, 2021 by guest. Protected preferentially over PGI in the healthy kidney. In me more most persons, PGI in the urine comprises pepsino- than 30 years of unalloyed intellectual gens electrophoretically identified as Pg 1-5. How- pleasure. ever, in some subjects, Pg 5 is missing. This trait, References possibly linked to HL-A loci (Weitkamp, Towner and May, 1975), affects 14% ofnon-ulcer whites as an ACHORD, J.L. (1981) Gastric pepsin and acid secretion in patients autosomal recessive, and 14% with acute and healed duodenal ulcer. Gastroenterology, 81, 15. affects more than of ALY, A. & EMAS, S. (1982) Sensitivity ofthe oxyntic and peptic cells blacks, but is not found in those of East Asian to in duodenal ulcer patients and healthy subjects descent (Samloff, 1983). with similar secretory capacity. Digestion, 25, 88. Serum pepsinogen comprises six times more PGI ANSON, M.L. & MIRSKY, A.E. (1932) The estimation of pepsin with than PGII. The relative abundance of hemoglobin. Journal of General Physiology, 16, 59. PGII is not AYALON, A., SANDERS, M.J., THOMAS, L.P., AMIRIAN, D.A. & surprising since PGII is not excreted in the urine. SOLL, A.H. (1982) Electrical effects of histamine on monolayers About 2/3 of DU have serum PGI values above formed in culture from enriched canine gastric chief cells. Postgrad Med J: first published as 10.1136/pgmj.60.709.743 on 1 November 1984. Downloaded from

Pepsinogen 749

Proceedings of the National Academy of Sciences of the United AMP mediation of muscarinic stimulation of pepsinogen secretion States of America, 79, 7009. in the frog. American Journal of Physiology (In press). BURHOL, P.G. (1982) Regulation of gastric secretion in the chicken. ISHAGUE, M. & BARDHAN, K.D. (1978) An assessment of the radial Scandinavian Journal of Gastroenterology, 17, 321. diffusion method for the measurement of pepsin in gastric BURHOL, P.G. & HIRSCHOWITZ, B.I. (1971) Gastric stimulation by secretion and its comparison with the haemoglobin substrate subcutaneous infusion of histamine in fistula chickens. Scandina- colorimetric method. Clinical Chimica Acta, 87, 259. vian Journal of Gastroenterology, 6, (suppl. II) 15. KAGEYAMA, T. & TAKAHASHI, K. (1980a) Monkey pepsinogen and BYNUM, T.E. & JOHNSON, L.R. (1975) Stimulation of human pepsin pepsins IV. The amino acid sequence of the activation peptide output by tropical . American Journal of segment of Japanese monkey pepsinogen. Journal of Biochemistry Digestive Diseases, 20, 607. (Tokyo), 88, 9. ELDER, J.B. & SMITH, I.S. (1975) Gastric-acid output, pepsin output, KAGEYAMA, T. & TAKAHASHI, K. (1980b) Isolation of an activation and lean body mass in normal and duodenal-ulcer subjects. intermediate and determination of the amino acid sequence of the Lancet, i, 100. activation segment of human pepsinogen A. Journal of Biochemis- HERSEY, S.J., MAY, D., SCHYBERG, D. (1983) Stimulation of try (Tokyo), 88, 571. pepsinogen release from isolated by cholecystok- KAGEYAMA, T. & TAKAHASHI, K. [c] (1980) Monkey pepsinogens inin-like peptides. American Journal of Physiology, 244, G 192. and pepsins V. purification, characterization and amino-terminal HIRSCH-MARIE, H., LOISILLIER, F., ToUBOUL, J.P. & BURTIN, P. sequence. Determination of crab-eating monkey pepsinogens and (1976) Immunochemical study and cellular localization of human pepsins. Journal of Biochemistry (Tokvo), 88, 635. pepsinogens during ontogenetics and in gastric cancers. Labora- KOELZ, H.R., HERSEY, S.J., SACHS, G. & CHEW, C.S. (1982) tory Investigation, 34, 623. Pepsinogen release from isolated gastric glands. American Journal HIRSCHOWITz, B.I. (1953) Urinary excretion of pepsinogen in of Physiology, 243, G218. gastroduodenal ulceration. Lancet, i, 66. KUSHNER, I., RAPP, W. & BURTIN, P. (1964) Electrophoretic and HIRSCHOWITz, B.I. (1955) Pepsinogen in the blood. Journal of immunochemical demonstration of the existence of four human Laboratori' and Clinical Medicine. 46, 568. pepsinogens. Journal of Clinical Investigation, 43, 1983. HIRSCHOWITZ, B.I. (1957) Pepsinogen: Its origins, secretion and MIHAS, A.A., GIBSON, R.G. & HIRSCHOWITz, B.I. (1976) Inhibition excretion. Physiological Reviews, 37, 475. of gastric secretion in the dog by 16, 16-dimethyl prostaglandin HIRSC HOWITz, B.I. (1967a) The secretion of pepsinogen. In Hand- E2. American Journal of PhVsiology, 230, 351. book of phvsiologv, Alimentartn Canal II, Chapter 50. American MIRSKY, l.A., BLOCK, S., OSHER, S. & BROH-KAHN, R.H. (1948) Phi siological Societv, pp. 889. Uropepsin excretion by man: 1. The source, properties and assay HIRSCHOWITZ, B.l. (1967b) The control of pepsinogen secretion. of uropepsin. Journal of Clinical Investigation, 27, 818. Annals of the Newt York Academy of Sciences, 140, 709. MUTO, N., MURAYAMA, K., AKAHANE, K. & TANI, S. (1980) HIRSCHOWITZ. B.I. (1984) Apparent and intrinsic sensitivity to Biochemical and immunological characterization of rat pepsino- by copyright. pentagastrin of acid and pepsin secretion in peptic ulcer. gen and pepsins. Journal of Biochemistry (Tokyo), 87, 717. Gastroenterology. 86, 843. NORTHROP, J.H., KUNITZ, M. & HERRIOT, R.M. (1948) Crstalline HIRSCHOWITZ, B.1., FONG, J. & MOLINA, E. (1983) Effects of enzymes. 2nd Edition, Columbia, New York. pirenzepine and atropine on vagal and cholinergic gastric PUURUNEN, J. (1979) Effect of intragastric acid on pepsinogen secretion and gastrin release and on heart rate in the dog. Journal secretion in the rat. European Journal of Pharmacology, 55, 207. of Pharmacologv and Experimental Therapeutics, 225, 263. RICHMOND, V., TANG, J., WOLFE, S., TRUCCO, R.E. & CAPUTTO, R. HIRSCHOWITz, B.1. & HUTCHISON, G.A. (1977a) Evidence for a (1958) Chromatographic isolation of gastricsin, the proteolytic histamine H-2 receptor which inhibits pepsin secretion in the dog. enzyme from gastric juice with pH optimum 3.2. Biochimica et American Journal of Phvsiologv, 233, E225. biophysica acta, 29, 453. HIRSCHOWITz, B.I. & HUTCHISON, G.A. (1977b) Kinetics of ROTTER, J.1., SONES, J.Q., SAMLOFF, I.M., RICHARDSON, C.T., atropine inhibition of pentagastrin-stimulated H +, electrolyte, and GURSKY, J.M., WALSH, J.H. & RiMOIN, D.L. (1979) Duodenal- pepsin secretion in the dog. American Journal ofDigestive Disease. ulcer disease associated with elevated serum pepsinogen I: An 22, 99. inherited autosomal dominant disorder. New England Journal of http://pmj.bmj.com/ HIRSCHOWITz, B.I., HUTCHISON, G.A. & SACHS, G. (1973) Kinetics Medicine, 300, 63. of atropine and Diamox inhibition of histamine-stimulated gastric SAMLOFF, I.M. (1969) Slow moving protease and the seven secretion. Scandinavian Journal of Gastroenterologv, 8, 555. pepsinogens. Electrophoretic demonstration of the existence of HIRSCHOWITz, B.1. & MOLINA, E. (1983) Effects of four H2 eight proteolytic fractions in human . Gastroentero- histamine antagonists on bethanechol-stimulated acid and pepsin logy, 57, 659. secretion in the dog. Journal of Pharmacology and Experimental SAMLOFF, I.M. (1982) Pepsinogens I and II: Purification from gastric Therapeutics, 224, 341. mucosa and radioimmunoassay in serum. Gastroenterology, 82, 26. HIRSCHOWITZ, B.I., MOLINA, E. & Ou TIM L. (1984) Effects of very SAMLOFF, I.M. (1983) Pepsinogens and pepsins in gastrointestinal low doses of atropine on basal and pepsin secretion, gastrin and disease. Smith, Kline and French International Co., England, p. on September 26, 2021 by guest. Protected heart rate in normals and DU. Digestive Diseases and Sciences (in 60. press). SAMLOFF, I.M., LIEBMAN, W.M. & PANITCH, N.M. (1975) Serum HIRSCHOWITz, B.I., O'LEARY, D.K. & MARKS, I.N. (1960) Effects of group I pepsinogens by radioimmunoassay in control subjects and atropine on synthesis and secretion of pepsinogen in the rat. patients with peptic ulcer. Gastroenterology, 69, 83. American Journal of Physiology, 198, 108. SEIJFFERS, M.J., SEGAL, H.L. & MILLER, L.L. (1963) Separation of pepsin I, pepsin II A, pepsin, II B, and pepsin III from human HIRSCHOWITz, B.1., RENTZ, J. & MOLINA, E. (1981) Histamine H-2 gastric mucosa. American Journal of Physiology, 205, 1099. receptor stimulation and inhibition of pepsin secretion in the dog. SHIRAKAWA, T., SHAH, G., INOUE, M., HELDMAN, A. & HIRSCHOW- Journal of Pharmacology and Experimental Therapeutics, 218,676. ITZ, B.I. (1983) Stimulants of pepsinogen secretion from frog HIRST, B.H., CONLON, J.M., COY, D.H. & HOLLAND J, SHAW (1982) esophageal glands. American Journal of Physiology (submitted). Comparison of the gastric exocrine inhibitory activities and SHUGERMAN, R.P., HIRSCHOWITz, B.I., BHOWN, A.S., SCHRONHEN- plasma kinetics of somatostatin-28 and somatostatin-14 in cats. LOHER, R.E. & SPENNEY, J.G. (1982) A unique "mini" pepsinogen Regulatorty Peptides, 4, 227. isolated from bullfrog esophageal glands. Journal of Biological ICHIHARA, Y., SOGAWA, K. & TAKAHASHi, K. (1982) Rat gastric Chemistry, 257, 795. prepepsinogen: in vitro synthesis and partial amino-terminal SIMPSON, L., GOLDENBERG, D. & HIRSCHOWITz, B.I. (1980) signal sequence. Journal of Biochemistry (Tokyo), 92, 603. Pepsinogen secretion by the frog in vitro. American INOUE, M., FONG, J., SHAH, G., HIRSCHOWITZ, B.1. (1984) Cyclic Journal of Physiology, 238, G79. Postgrad Med J: first published as 10.1136/pgmj.60.709.743 on 1 November 1984. Downloaded from

750 B. I. Hirschowitz

SOGAWA, K., FuJiI-KURIYAMA, Y., MIZUKAMI, Y., ISCHIHARA, Y. aspects of duodenal ulceration: Pepsin I secretion in relation to & TAKAHASHI, K. (1983) Primary structure of human pepsinogen ABO blood groups and ABH secretory status. Journal ofMedical gene. Journal of Biological Chemistry, 258, 5306. Genetics, 16, 423. SOGAWA, K., ICHIHARA, Y., TAKAHASHI, Y., FuJII-KURIYAMA, Y. WALKER, V. & TAYLOR, W.H. (1980) Pepsin I secretion in chronic & MURAMATSU, M. (1981) Molecular cloning of complementary peptic ulceration. Gut, 21, 766. DNA to swine pepsinogen mRNA. Journal of Biological Chemis- WEITKAMP, L.R., TOWNES, P.L. & MAY, A.G. (1975) Probable try, 256, 1256. genetic linkage between a locus for human urinary pepsinogen TANG, J. (1977) Acid proteases: Structures, function and biology. and the HL-A loci. American Journal of Human Genetics, 27, 486. Advances in Experimental Medicine and Biology, 95, Ch. I-I 1. YASUGI, S. & MIZUNo. T. (1981) Purification and characterization TAYLOR, W.H. (1970) Pepsins of patients with peptic ulcer. Nature, of embryonic chicken pepsinogen, a unique pepsinogen, with 227, 76. large molecular weight. Journal of Biochemistry (Tokyo), 89, 31 1. WALT, L., ROBERTS, N.B. & TAYLOR, W.H. (1979) Hereditary by copyright. http://pmj.bmj.com/ on September 26, 2021 by guest. Protected