Macroamylasemia: variable response of a 7S and a 11S macroamylase to phosphate-benzoic acid buffer during gel filtration

During thin layer or macrocolumn gel filtration with Sephadex G 200, a 7S macroamylase dissociated completely to normal-sized amylase when the eluant was phosphate-benzoic acid buffer. In contrast a 11S macroamylase was not affected by the buffer and amylase activity was associated with 11S proteins. Since phosphate-benzoic acid buffer is used in a thin layer screening method described by Peeters et al. (9), some macroamylases may go undetected. After treatment of the 7S macroamylase with phosphate-benzoic acid buffer, the amylase-free 7S proteins obtained formed macroamylase with added pancreatic amylase. This indicates that the buffer did not inhibit the activity of the macroamylase, but caused a dissociation of the macroamylase during gel filtration. The difference in the effect of phosphate-benzoic acid buffer on the two types of macroamylases may be attributed to the difference in affinity of binding proteins for the amylase molecule.     

Impaired amylase activities caused by binding of abnormal immunoglobulin A in patients with macroamylasemia

Eight of thirty analyzed samples of immunoglobulin A-linked macroamylases showed significantly higher enzyme activity with the low molecular (short-chain amylose) substrate than with the macromolecular (blue starch polymer) substrate. Immunoglobulin A, which had been separated from one macroamylase, was found to bind to normal amylases from different organs resulting in decreased affinity for the macromolecular substrate. These findings suggest that abnormal immunoglobulin A is stoichiometrically associated with a specific site adjacent to the active site of the amylase molecule. In addition, alpha-amylase inhibitor failed to inhibit the enzyme activity of some macroamylases with macromolecular substrate, suggesting that immunoglobulin A blocked the binding site for the inhibitor. We conclude that immunoglobulin A-linked macroamylases have heterogeneous binding sites for immunoglobulin A, resulting in functional abnormality of amylase activity.     

Macroamylase immunoglobulins show high affinity for animal and human amylases

We have examined the affinity shown by the immunoglobulin fraction from each of five sera containing macroamylase for amylases from different sources: human saliva or human, porcine, or ovine pancreas. High affinity constants, 0.4 X 10(10) to 7.2 X 10(10) L/mol, were found in competitive binding experiments with human or porcine pancreatic amylase. All but one serum yielded linear Scatchard plots, indicating that in most sera the amylase-binding immunoglobulins are homogeneous, possibly monoclonal. The immunoglobulin fractions from different sera differed in their specificity: two of them bound all four types of amylases, whereas two bound only one type. Three of the five immunoglobulin fractions showed considerably higher affinity towards one or both of the animal amylases than towards the human ones, and may be primarily directed against some animal amylase.     

A comparison between naturally occurring macroamylasaemia and macroamylasaemia induced by hydroxyethyl-starch

Macroamylasaemia was produced in vitro by incubation of hydroxyethylstarch with serum, and in vivo by intravenous infusion of hydroxyethylstarch. Gel filtration on Sephadex G-100 revealed distinct differences in molecular size distribution between such hydroxyethylstarch-induced macroamylase and the usual form of naturally occurring macroamylase which was observed in a few patients from our hospital. Further studies demonstrated that the gel filtration elution pattern of amylase activity in serum containing hydroxyethylstarch-induced macroamylase is significantly altered with time in vitro and in vivo, probably because of an enzymatic degradation of the hydroxyethylstarch components of the macromolecular complexes. In a healthy volunteer the serum amylase activity was elevated to a maximum of 797 u/l and the renal clearance rate of amylase was diminished to a minimum of 0.3 ml/min after infusion of 500 ml of a 6% solution of hydroxyethylstarch, as compared to 300 u/l, and 0.95 ml/min, respectively, during the pre-infusion period.     

Macroamylase: prevalence, distribution of age, sex, amylase activity, and electrophoretic mobility

The contribution made by macroamylase to the occurrence of hyperamylasemia of unknown origin has not been previously quantitated nor has the distribution of age and sex been identified in a large sample. Knowing the pattern of this enzyme complex when analyzed by an electrophoretic isoenzyme procedure would be of importance in recognizing its presence. During a three-year period we have studied a population of 2 900 patients with elevated serum amylase from whom specimens were sent to our laboratory with a request for amylase isoenzyme analysis. It was assumed that the primary reason for requesting these analyses was to clarify the presentation of hyperamylasemia. Macroamylase occurred in 9.6% of these patients. The total activity of the macroamylase specimens was typically 1 to 4 times normal but ranged to 20 times normal. Comparing these patients with a macroamylase-negative group, the distribution of age and sex was found to be no different. Macroamylase had a characteristically unusual electrophoretic pattern with bands faster than normal pancreatic amylase in 10% of cases, bands between pancreatic and salivary in 30% of cases, and the remaining 60% with atypical anodal bands.     

Electrophoretic separation, detection, and variation of amylase isoenzymes.

An electrophoretic technique for demonstrating amylase isoenzymes is described. After separation in an agarose gel containing a linear polyacrylamide polymer to reduce electroendosmotic flow, the amylase fractions are visualized by incubation with a commercially available dye-starch polymer (Phadebas Amylase Test). The technique detects amylase fractions with activities below 10 U/l. Some characteristic changes in such diseases as acute and chronic pancreatitis, cystic fibrosis of the pancreas, macroamylasemia and inherited variants as well as after maxillofacial surgery are mentioned.     

Variations in amylase isoenzymes and lipase during acute pancreatitis, and in other disorders causing hyperamylasemia

We compare the clinical value of assay of amylase (EC 3.2.1.1) isoenzymes with that of lipase (EC 3.1.1.3) in serum from patients with proven acute pancreatitis or with hyperamylasemia from other causes. In the former group we measured amylase, lipase, and isoamylases daily. Lipase and P(pancreas)-type isoamylases reached the highest mean values on the first day of an attack of acute pancreatitis (day one). Lipase declined rapidly, and by day four its mean activity was about the same as that of amylase and lower than that of the P-type isoamylases. Great inter-individual variations were found among patients with a similar clinical course. Of the 85 samples analyzed, amylase activity for 36 declined to within reference limits, but 18 of the 36 had high lipase activity, 18 had high P-type isoamylases activity, and 31 had P3 isoenzyme, which is not detectable in normal sera. Determination of isoamylases is a more sensitive index to acute pancreatitis than lipase assay and may be particularly useful when pancreatitis is suspected despite a normal total amylase activity. In the group of patients with hyperamylasemia from other origins, three had macroamylasemia, one had mumps, one had abdominal trauma without pancreatic injury, and one had pelvic inflammatory disease. The specific pattern of macroamylase on electrophoresis permitted a precise diagnosis of macroamylasemia; normal lipase had only ruled out pancreatitis. In the three other cases, lipase and isoamylases excluded pancreatic involvement.     

Interaction of immobilized anti-salivary amylase antibody with human macroamylases: implications for use in a pancreatic amylase assay to distinguish macroamylasemia from acute pancreatitis

We examined the ability of an immobilized antibody to salivary amylase (Clin Chem 1985;33:1283-8) to react with amylase in macroamylasemic sera. The antibody removed 50% (SD 23%) of the total amylase activity from 39 macroamylase sera, a percentage indistinguishable (P greater than 0.75) from the percentage removed from concurrently analyzed sera from healthy volunteers (49%, SD 11%). Electrophoretic analysis of 23 macroamylasemic sera revealed that the antibody removed only part of the macroamylase band(s) in 71% of the cases. We conclude that the mean isoenzyme composition of the macroamylase complexes is essentially identical to the mean isoenzyme distribution in normal sera (i.e., about half salivary and half pancreatic amylase). Further, the immobilized antibody can be used to distinguish most patients with macroamylasemia from those with acute pancreatitis, because sera from the latter contain an increased proportion (greater than 80%) of pancreatic amylase.     

Properties of amylase-linked immunoglobulins.

In this report, the properties of eight cases of amylase-linked immunoglobulins were gel filtration using Sephadex G-200 superfine. The class of heavy chain of amylase-linked immunoglobulins was proved to be gamma in four cases and alpha in three cases by immunoelectrophoresis followed by amylase activity staining. In one of the cases, the precipitin line against light chain lambda could be visualized only after treatment with 0.1 M 2-mercaptoethanol. Then, the type of light chain was determined to be exclusively lambda irrespective of heavy chain classes. In two cases, the immunoglobulin complexes were partially dissociated into normal amylase and immunoglobulin G at pH 8.6, and completely dissociated at pH 9.0. In three cases, the complexes were completely dissociated at pH 8.6. The precipitin line of papain treated amylase-linked immunoglobulin G against anti-Fc was not. This fact suggests that the binding site of amylase-linked immunoglobulins G was located in the Fab portion of immunoglobulin molecule and that the complexes are specific antigen-antibody complexes. Treatment with Con-A Sepharose caused the dissociation of the amylase-immunoglobulin complexes. It is suggested that the changes in the conformation of amylase-linked immunoglobulin causes the dissociation of this immuno-complex. Thus, it is elucidated that the complexes of amylase and immunoglobulins are specific antigen-antibody complexes, and that these complexes must be recognized as one of the circulating autoantibodies in plasma, and must be clearly distinguished from the other unknown macromolecular amylase complexes.     

Nonpancreatic-type hyperamylasemia associated with pancreatic cancer.

Isoamylase analysis of the serum and urine of a patient with anaplastic spindel cell carcinoma of the pancreas revealed that virtually all of the serum amylase and almost all of the urine amylase behaved chromatographically as the salivary (S) type. Both the serum and urine amylases were bound by a substance derived from a macroamylase complex which had been shown to bind only salivary amylase and to lack any affinity for pancreatitis (P) type amylase. The ratio of amylase to creatinine clearance was markedly increased (12.5%) without evidence of acute pancreatitis at autopsy and despite the presence of only a minute amount of P-type isoamylase in the serum.     

Macroamylases: differences in activity against various-sized substrates.

Hyperamylasemia caused by macroamylases can lead to the overdiagnosis of acute pancreatitis. We examined whether interference from macroamylase is less in assays that use high-molecular-mass (high-M(r)) substrates rather than oligosaccharide substrates. We hypothesized that high-M(r) substrates would be sterically excluded from macroamylasemic complexes and thus would be hydrolyzed less efficiently. Eighteen macroamylasemic samples were assayed by using red-dyed amylopectin or blue-dyed starch as polysaccharide substrates or by using maltoheptaose or maltotetraose as oligosaccharide substrates. The oligosaccharide substrates gave comparable results (y = 0.81x + 83), but we observed consistently lower activities for amylopectin than for maltotetraose (y = 0.32x + 38). We observed no bias among methods when nonmacroamylasemic specimens were analyzed. The mechanism of this difference was examined by adding antihuman pancreatic amylase antibodies to hyperamylasemic serum samples from patients without macroamylasemia and to purified human pancreatic or salivary isoamylases. In each case, polyclonal and monoclonal antibodies lowered amylase activity more in assays with complex polysaccharides than in those with oligosaccharides. The use of high-M(r) substrates diminishes interference, and detection of suspected macroamylasemia may be possible through comparing activities determined from automated methods that use different substrates.     

A screening technique for macroamylasemia using thin-layer gel filtration

A simple method is presented for screening human sera for macroamylasemia. Thin layers of Sephadex G-200 are equilibrated with a starch containing buffer. Samples of 20 μl each are applied to the plate and eluted for 3 h. The plate is then sprayed with an iodine solution, which stains starch-containing zones deep blue. Zones corresponding to anylase activity remain white. With normal sera one single white spot with a relative migration distance of 0.3 is obtained. The presence of macroamylase is revealed by a fast moving spot at a relative migration distance of 0.9. The technique is simple, sensitive and reliable. It detects and differentiates amylase and macroamylase activities in samples containing as few as 20 Somogyi units per 100 ml. One plate will easily accomodate eight samples in one row. By cutting down the elution time to 1.5 h, two rows of eight samples each might be applied to one plate. In such case the sensitivity is reduced to about 30 Somogyi units per 100 ml and elution should be followed more closely to avoid overlap.     

Detection of macro-creatine kinase and macroamylase by polyethylene glycol precipitation and ultrafiltration methods

BackgroundMacroenzymes may cause elevations in serum enzyme activity. Macroenzymes are not common; however their detection is important because they cause diagnostic confusion and therapeutic errors.MethodsWe analyzed 2 of the most prevalent macroenzymes in the literature, macro-creatine kinase (macro-CK) and macroamylase, using 2 methods for detection, polyethylene glycol (PEG) precipitation and ultrafiltration (UF). Enzyme measurements were made using a Roche Modular Analytics P analyzer. Imprecision was assessed using quality control material. We evaluated 125 samples from apparently healthy subjects to establish reference intervals. For macro-CK comparison, 94 samples with activities > 200 U/l were analyzed with both PEG precipitation and UF and compared to electrophoresis. PEG precipitation and UF were compared for macroamylase detection using 130 samples with amylase activities > 110 U/l.ResultsUF was more precise and demonstrated narrower reference intervals for both analytes. PEG precipitation and UF were able to detect true cases of macro-CK with overall agreement with electrophoresis of 79.8% and 80.9%, respectively. Both methods detected the same number of ‘positive’ macroamylase samples; however PEG precipitation resulted in a greater number of ‘indeterminate’ cases.ConclusionThis is the first report where UF has been shown useful for the detection of both macro-CK and macroamylase.     

Immunoglobulin light-chain immunoblots of urine proteins from patients with tubular and Bence-Jones proteinuria

Immunoglobulin excretion by patients with monoclonal gammopathies and tubular proteinuria has been analysed by agarose gel isoelectricfocussing of untreated urine and immunoblotting. About three-quarters of the Bence-Jones proteins detected occurred as multiple bands on isoelectricfocussing; in about half of these cases the multiple forms were due to polymerisation or fragmentation of the light chains. In specimens with tubular proteinuria, a characteristic light chain pattern of three broad bands covering the pI ranges 7.1-7.3, 7.8-8.0 and 8.3-8.5 was found. This pattern also occurred in 53% of specimens with Bence-Jones proteinuria and was identical to that found in concentrated normal urine. Intact monoclonal immunoglobulin usually appeared as three or more evenly spaced bands of similar intensity whereas polyclonal intact immunoglobulins produced diffuse staining from pI 5.0-8.5. A scheme for the qualitative analysis of urinary immunoglobulin excretion using this technique has been developed.     

The role of gamma-globulin in composition of macroamylase.

A macroamylase complex has been artificially generated by combining purified pancreatic and salivary alpha-amylase solutions with physiological amounts of human gamma-globulin. This complex was only partially dissociable at pH 3.5. Salivary alpha-amylase was shown to have higher binding affinity to gamma-globulin than the pancreatic alpha-amylase. Small amounts of macroamylase complex were also formed after the addition of purified salivary or pancreatic alpha-amylase to normal serum. This effect was more distinct when salivary alpha-amylase was added. Electrophoretic examination of serum proteins from subjects with macroamylasemia showed a significant increase of gamma-globulin fraction. Seromucoid of these subjects was also above normal range. Electrophoresis of serum glycoproteins yielded raised values of the fraction moving with gamma-globulin.     

Characterization of amylases produced by tumors

Amylases were purified and characterized from three amylase-producing human tumors. The relative molecular mass of the amylases was estimated to be 54 000 on sodium dodecyl sulfate/polyacrylamide gel electrophoresis, different from human salivary amylase (61 000 and 64 000) and human pancreatic amylase (60 000). The tumor amylases had completely identical antigenicities with human salivary and pancreatic amylases against antibody to human salivary amylase, while the intensity of the tumor amylases was less than 20% of that of human salivary and pancreatic amylases on single radial immunodiffusion. On isoelectric focusing, two of the three tumor amylases showed a major peak at pH 6.4, which corresponded to a major peak of human salivary amylase, the other showed a major peak at pH 6.4, which corresponded to a minor peak of human salivary amylase. The three tumor amylases showed similar amino acid composition, different from those of human salivary and pancreatic amylases. These findings suggest that tumor amylases have a tertiary structure similar to that of normal human amylases, but differ from them in amino acid composition.     

Electrophoretic pattern of amylase isoenzymes in serum and urine of normal persons.

We separated and measured amylase isoenzymes in the serum and urine of 3036 normal persons by electrophoresis on a thin layer of polyacrylamide gel. We wished to establish the normal pattern of these isoenzymes and to evaluate the usefulness of this method of electrophoresis in clinical diagnosis. Results for patients with hyper- or hypofunctioning pancreas and salivary glands suggested that essentially all the isoamylases in human serum and urine are derived from the salivary glands and the pancreas, and revealed that isoamylases of more than 98% of normal persons consisted of two major isoenzymes and two to three minor ones. Although these observations indicate that data on changes in the proportion of amylase activity of each isoenzyme can be useful in clinical medicine, the following points should be remembered: (a) quantitative differences in the isoenzyme pattern were observed, depending upon the condition of the samples; (b) because the proportion of isoamylase activity in serum of different normal persons differs, seriatim determination of amylase isoenzymes is necessary; and (c) because five different genetically controlled types of isoamylases were observed in normal persons, genetic investigations are also necessary.     

Action pattern of serum amylase using p-nitrophenyl-maltoheptaoside as substrate

Selected serum samples with different isoamylase patterns, as revealed by agarose gel electrophoresis, were analysed by the Phadebas amylase test, and by using p-nitrophenyl-maltoheptaoside as the substrate. The products with the latter substrate were examined by high performance liquid chromatography (HPLC). For some samples, different catalytic activities were registered by the two amylase methods. In these cases, electrophoresis showed a double pancreatic band associated with increased pancreatic amylase catalytic activity, or an abnormal, broad salivary band associated with increased salivary enzyme catalytic activity. The stoichiometric factors for most of these samples, calculated from the amount of p-nitrophenol detected after HPLC, differed from the mean values.     

Pancreatic amylase measured in serum by use of a monoclonal antibody immunochemically immobilized to a solid phase

We studied a method for measuring the pancreatic isoenzyme of amylase (EC 3.2.1.1) by use of a mouse monoclonal antibody against human salivary-type amylase (Clin Chem 1985;31:1283) coupled indirectly to particles of polyvinylidene fluoride via polyclonal goat anti-mouse immunoglobulin. These particles, in 200 microL of a suspension, could remove salivary amylase (activity 2200 U/L) from an equal volume of serum in 5 min. Measurement of amylase activity in the supernatant fluids from treated sera thus provided an assay of pancreatic amylase. Precision studies at three activity concentrations yielded within-run CVs of 1.6% to 1.7% (n = 25) and total CVs of 2.2% to 5.1% (20 days). Salivary amylase added to each of 10 sera was completely (99.8%, SD 1.6%) removed. The new method (y) showed the following regression statistics when compared with an electrophoretic method (x): slope = 0.989 (SD 0.019), intercept = -0.220% (SD 1.48%), SEE 4.0%, n = 51. Similar respective regression values were found for urine samples: slope = 0.934 (SD 0.053), intercept = 2.3 U/L (SD 3.2), SEE 8.4 U/L, n = 26. The following respective values were found when the new method (y) was compared with the previously described immunoprecipitation assay (x): slope = 1.02 (SD 0.02), intercept = 2.2% (SD 1.4%), SEE 3.3%, n = 23 sera. Reference intervals for pancreatic amylase activity in serum were established for three different substrates: maltotetraose, maltopentaose, and p-nitrophenylheptaoside.     

Three methods compared for determination of pancreatic and salivary amylase activity in serum of cystic fibrosis patients

We evaluated three methods for serum amylase (EC 3.2.1.1) isoenzymes to determine whether they are interchangeable and to test their ability to discriminate between cystic fibrosis patients with and without pancreatic insufficiency. One method involved salivary amylase inhibitor (O), and two were polyacrylamide gel electrophoresis separations differing in method of detection--either direct zymogram (G) or gel slicing followed by activity estimates per slice (W). Results for percentage pancreatic amylase differed significantly. Reproducibility for percentage pancreatic amylase was high, moderate, and low (r = 0.95, 0.53, and 0.02) for methods G, O, and W, respectively; moderate (r = 0.60) among the three methods; and moderate between pairs. Therefore, this result for a subject must be considered relative to the method used in its determination. The clinical diagnosis of pancreatic insufficiency was verified by 77.8%, 83.3%, and 94.4% correct classification rates for methods O, W, and G, respectively. Evidently, method G is the most efficient and may be the method of choice for measuring serum amylase isoenzymes in cystic fibrosis.