CLIN. CHEM.
34/11,
2235-2238
(1988)
Liquid-ChromatographicMeasurement of p-AminobenzoicAcid and Its Metabolitesin Serum
Unda L Y. Yung-Jato,’
P. R. Dude,’ and Steven J. SoldIn2
This is a high-performance liquid-chromatographic method
for measuring p-aminobenzoicacid (PABA) and its metabolites in plasma or serum. Samples are deproteinized,then
extracted with organicsolventsbefore chromatography.For
quantification,the peak heightof the individualcompoundis
comparedwiththat of the internalstandard.Analyticalrecoveries ranged from 41% to 100%, depending on the compound studied. Comparison of patients’ samples after oral
administrationof either N-benzoyl-L-tyrosyl-p-aminobenzoic
acid or free PABA revealed that PABA is extensivelymetabolized and conjugatedto either p-acetamidobenzoicacid, paminohippuricacid, or p-acetamidohippuricacid. PABA concentrationsin serum as measuredwith the Bratton-Marshall
ultravioletspectrophotometricprocedurewould appear predominantlyto reflectmeasurementsof metabolites,withonly
a minorcontributionfrom PABA itself.
AddItIonalKeyphraaes:cystic fibrosis
cy
pediatric cMemist,y
pared
.
pancreatic insufficien-
Bratton-Marshall procedure
com-
There are no simple, practicable diagnostic
tests to determine exocrine pancreatic
function. Current tests involve
gastrointestinal
intubation followed by the stimulation
of
exocrine pancreatic secretion with hormones (secretin, socretin-eholecystokin)
or a test meal (e.g., the Lundh meal)
and subsequent measurement of bicarbonate concentration
and pancreatic enzyme output. These invasive
direct pancreatic-function
tests can be performed only in specialized
hospital departments and usually are expensive and time
consuming. The need for a simple, accurate, indirect test of
pancreatic
function has led to efforts to develop an oral
procedure that is more affordable, convenient, and acceptable to patients, especially children (1-3).
In 1972, Imondi et al. (4) first described reliable results for
a pancreatic function test in animals, which involved the
oral administration
of a synthetic peptide, N-benzoyl-ir
tyrosyl-p-axninobenzoic
acid (Bz-ty-PABA) (3). In the small
bowel, this peptide is specifically cleaved to N-benzoyl-Ltyrosine and p-aminobenzoic
acid (PABA) by the pancreatic
endopeptidase,
chymotrypsin
(EC 3.4.21.1).
The PABA is
rapidly absorbed from the gut-probably
by passive diffusion-and undergoes conjugation and further metabolism in
the liver to p-acetamidobenzoic
acid (PAABA), p-aminohippuric acid (PAHA), and p-acetamidohippuric
acid (i)
(Figure 1) before being excreted in the urine. The amount of
PABA and its metabolites
found in serum or plasma and
urine can be used as an index of the intraluminal
activity of
‘Research Institute, and Departments of Biochemistry and Pediatrics, Hospital for Sick Children, Toronto, Canada.
2Department of Laboratory Medicine, Children’s Hospital National Medical Center, 111 Michigan Ave., N.W., Washington, DC
20010 (address for correspondence); and Departments of Pharmacology, Pathology, and Child Health & Development., George Washington University, Washington, DC.
Received June 9, 1988; accepted July 22, 1988.
chymotrypsin and thus provides an indirect measurement
of
exocrine pancreatic function (5, 6).
Shwachman’s syndrome and cystic fibrosis are the major
causes of exocrine pancreatic insufficiency
in children. Studies on cystic-fibrosis patients evaluated the diagnostic importance of this test in cases of exocrine pancreatic insufficiency, by means of a nonspecific spectrophotometric
procedure (7-10). The PABA recovered in urine was significantly
lower in cystic fibrosis patients than in a normal control
group (7-9), and there was a good correlation between PABA
recovery, fecal chymotrypsin
activity,
and coefficient of fat
absorption.
Evidently,
PABA recovery is significantly
decreased in patients with cystic fibrosis and steatorrhea,
and
thus PABA might be a reliable marker of pancreatic insufficiency in other patients as well.
At present, in most published procedures the PABA collected in a timed urine specimen is measured. Determinations
in plasma or serum would, however, provide a more direct
evaluation of the enzymatic hydrolysis of the peptide (10,
11), and blood samples are much easier to obtain from
-NH
-H-CO-+-NH--/--COOH
CH2
OH
COOH
CONHCH2COOH
NH2
p-
aminobenzoic
NH2
acid
(PP.BA)
p-ainohippric
acid
CPAMA)
CONHCH2COOH
OOH
‘H
HNCOCH3
HNCOCH3
p-acetaiidobenzoic
acid
(PAABA)
p-acetamidoh
ippuric
acid
(PMk{A}
Fig. 1. (Tc) Chemical formula of the synthetic peptide N-benzo1-Ltyrosyl-p-aminobenzoic acid (the peptide bond cleaved by chymotrypsin Is Indicated by a broken line);(Bottom)Structures of p-aminobenzoic acid and its metabolites
CLINICAL
CHEMISTRY,
Vol. 34, No. 11, 1988
2235
children
specimens. The amount of
at a certain time after
administration
of the compound should allow assessment of
both the exocrine pancreatic function and the conjugating
ability of the liver. As measured with a nonspecific spectrophotometric
procedure (Bratton-Marshall
method), the purported concentration
of PABA in plasma or serum of patients
with exocrine
pancreatic
insufficiency
was significantly
lower than in healthy controls (12-14). Some reports also
indicate that the determinations
of PABA in plasma or serum
are a more sensitive
indicator of pancreatic
insufficiency
than is steatorrhea,
and more specific than the results from
determination
only in urine (12-17).
Our goal in this study was to develop a micro-scale HPLC
procedure for measuring PABA and its metabolites in serum
or plasma, to obtain greater and more reliable information
than is possible with the conventional Bratton-Marshall
method (18, 19).
PABA
than
are timed
and its metabolites
urine
present
MaterIals and Methods
Reagents. Throughout these studies, we used PABA supplied from J. P. Baker Chemical Co., via Johns Scientific
Co., Toronto, Canada, and PAHA, PAABA, and m-hydroxybenzoic acid from Sigma Chemical Co., St. Lauis, MO.
PA.ARA was synthesized
in house as follows: add 5 g of
PAHA to a mixture of 60 mL water and 12.5 mL concentrated
hydrochloric acid. Dissolve 3 mL of acetic anhydride and
3.85 g of sodium acetate in 12 mL water, and add this to the
PkHA-HC1 mixture. Stir at 25 #{176}C
for about 10 mm until the
precipitation of white crystals is complete. Wash the crystals with cold distilled water and air-dry. PirA?s
melting
point is 236-238 #{176}C.
Phosphoric
acid was obtained from Fisher Scientific,
Toronto, Canada. Methanol, “Accusolv,” and dichloromethane were from Anachemia,
Mississauga,
Ontario, Canada.
Sulfosalicylic acid, 50 g per liter of methanol/water
(50/50 by
vol), was obtained from Abbott Diagnostics, Toronto, Canada. Ethyl acetate-Omnisolv
was purchased
from BDH
Chemicals, Toronto, Canada.
Chromatographic
system. The HPLC system consisted of a
Model U-6-K injector, Model 6000 pump, a Lambda-Max LC
spectrophotometer
481, a Servogor 120 recorder, and an LC
pre-column ifiter, all from Waters Associates, Milford, MA.
The column, selected from several C18 columns of various
lengths and particle sizes, was a 4.6 mm (i.d.) 25-cm Altex
Ultrasphere
column (Beckman Instruments,
Inc., Fullerton,
CA), packed with C15 material bonded to 5-tim silica particles. This column best resolved PABA and its metabolites
from the solvent front and from the other peaks in the
chromatogram.
The chromatography
was routinely
performed at room temperature
(20-25 #{176}C).
The mobile phase was 15/85 (by vol) methanol/phosphate
buffer (10 mmol/L, pH 3.0). Solvent delivery was set to a
flow rate of 2.0 mL/min. Absorbance of the column effluent
was monitored at 290 nm and a sensitivity setting of 0.05 A
full scale. The Servogor 120 recorder was set at a chart
speed of 0.5 cm/nun.
Extraction procedure. Combine, then vortex-mix for 15s,
100 uL each of serum or plasma and sulfosalicylic
acid
containing the internal standard, m-hydroxybenzoic
acid, to
precipitate
protein. Extract PABA and its metabolites
by
combining the protein-free
filtrate with 400 /AL of ethyl
acetate, vortex-mixing
for 1 mm, and centrifuging for 2 mm.
Remove the organic (top) phase and evaporate it under a
stream of nitrogen. Add 80 i.L of doubly distilled water and
2236
CLINICALCHEMISTRY, Vol. 34, No. 11, 1988
200 4, of dichloromethane,
vortex-mix
for 30 s, then
centrifuge for 6 miii. Inject 204, of the aqueous extract (top
layer) onto the column.
Calculations.
Quantify the analytes by comparing the
peak-height
ratio for the compound of interest and the
internal standard with that obtained for three standards of
known concentrations
processed through
the entire procedure.
To verify the identification of the compounds we compared
the retention times of the peaks in the chromatograms
of
serum samples with those obtained by direct injection of
standards.
Results
Figure 2 depicts typical chromatograuns
for a blank serum
sample and a patient’s sample.
The order in which the compounds eluted was pH dependent over the pH range 3.0-7.0, the retention time of PABA
and its metabolites increasing at lower pH. Resolution was
optimal at pH 3.0.
Linearity. We prepared
PABA, PAABA, PAAHA, and mhydroxybenzoic acid standards to cover a wide concentration
range and analyzed them as described. The standard curves
were linear up to 80 mg/L in all cases, well exceeding the
highest concentration
likely to be encountered
in patient’s
samples.
Precision. Between-day
precision of the method was assessed by repeated analyses of serum controls containing
various concentrations
of PABA, PAABA, PAHA, and PAAHA.
The results are shown in Table 1.
Recovery. Known concentrations
of PABA and its metabolites were added to serum, the extractions and assays were
performed as described, and the peak heights obtained were
compared with those obtained by direct injection of aqueous
standards. Table 2 summarizes the results.
Interference. We obtained 20 plasma samples from the
blood bank (normal population) and 20 serum samples from
patients
receiving various medications
(e.g., digoxin, anticonvulsanta,
ticarcillin,
theophylline,
chloramphenicol).
0.001 Au
0001 Au
_LLJLU
L
64
2220186141210
2
0
222018161412108
TIME (mm)
6
j
4
L
2
0
TIME (mm)
FIg. 2. Chromatogramsof (left) a plasma blank, (i7ght) a patients
sample
p,
1; PAB.A, 2;
3; m-hydroxybenzosc
acid, 4; and
5
Table 1. Between-Day PrecIsion (n
Control 1
Control 2
Control 3
Mean SD
Mean SD
10)
=
Mean SD
mg/L
CV, %
mg/L
CV, %
PAHA
0.57 0.07
11.9
1.8 0.11
6.4
4.7
0.6
PABA
0.50 0.03
5.3
1.9
0.15
PMPIA
0.55 0.04
0.51 0.02
7.4
3.8
2.0
2.0
0.09
0.04
7.7
4.5
1.9
5.3
4.7
4.9
0.4
0.35
0.24
Compound
PAAw
mg/L
CV, %
12.8
7.9
7.4
4.9
These were analyzed
as described. We saw no co-elution
with the compounds of interest.
Preliminary clinical studies. A standard dose of Bz-tyPABA containing
the equivalent of 5 mg of free PABA per
kilogram of body weight was administered
to two normal
subjects and to a patient with cystic fibrosis. In a second
study, free PABA, 5 mg/kg, was administered
to a normal
volunteer
and another
patient with cystic fibrosis. Table 3
shows the concentrations
of PABA and its metabolites found
60 and 90 miii after the dose. The }IPLC results are
compared with those obtained by the Bratton-Marshall
procedure.
DIscussIon
To evaluate exocrune pancreatic
function, most PABA
have been done on urine samples by the BrattonMarshall spectrophotometric
method, which does not differentiate between PABA and its metabolites
(20-23). Later
analyses
Table 2. Analytical-Recovery Studies
Mean recovery ,
Concn,
mg/L
2
5
10
%
(and CV, %)
n7-lIydroxybenzolc
acId
PAHA
PABA
PAAHA
49.7
40.6
(7.0)
93.6
(1.6)
(3.9)
(2.7)
43.4
73.4
95.8
(4.6)
40.9
(12.3)
45.3
(11.5)
PAABA
104
100
(4.8)
(2.0)
(3.2)
83
(3.0)
46.5
76.4
77.2
(4.8)
(2.3)
88.4
(2.9)
(3.4)
n = 5 each.
Table 3. Metabolltes In Serum of Two Normal Persons
and a Patient with Cystic Fibrosis Compared After an
Oral Dose of Bz-Ty-PABA or Free PABA
“PABA”
by
Bratton-
Time
Marshall
reports described HPLC techniques
using ion-pairing
reagents to detect PABA and its metabolites in urine (24,25). A
method for measuring PABA in serum was described based
on the procedure of Bratton-Marshall
(12-14, 17) with the
disadvantage
that it again does not differentiate
between
PABA and its metabolites, and the result produced is presumably a combination of PABA and metabolite concentrations
(Table 3). The present HPLC method has many advantages
over the classical Bratton-Marshall
procedure in that it
allows for individual quantification
of PABA and each of its
metabolites
and is not interfered
with by those commonly
prescribed drugs that we tested.
It is difficult to establish
conditions that will produce
optimal recoveries for all of the PABA metabolites
in serum.
Although the recovery figures for i’n
and PABA are
relatively
low compared with those for the other metabolites, the recoveries are consistent and provide adequate
sensitivity for the assay.
After an oral dose of Bz-ty-PABA, a patient with cystic
fibrosis and pancreatic
insufficiency
showed no detectable
PABA, or any of its metabolites, in serum, whereas all these
compounds were seen in patients with normal pancreatic
function (Table 3). In contrast, a dose of PABA produced
measurable concentrations
of PABA and (or) its metabolites
in both the normal subject and the patient
with cystic
fibrosis (Table 3).
Evidently,
very little of the “PASA” measured
in the
Bratton-Marshall
procedure is in fact PABA. Rather, the
measurement
primarily is of PABA metabolites,
and possibly
of other nonspecific substances.
Whether the ability to
quantify PABA and its metabolites separately by HPLC will
be more helpful than PABA measurement
by nonspecific
spectrophotometric
procedures and allow for improved clinical assessment of the patient’s condition is the subject of
another study. We will assess the diagnostic capability of
this test by analyzing plasma PABA and its metabolites in
both patients with cystic fibrosis and those with liver
disease.
References
1. Arvanitakis C, Greenberger NJ. The use of BT-PABA (PFT) in
chronic and other diseases. In: Masuda M, ed. Pancreatic function
diagnostant. Tokyo-New York: Igaku-Shoin, 1979:60-S.
2. Dockter G, Nacu I, Kohiherger E. Determination of proteasecleaved PABA in serum after oral administration of N-benzoyl-Ltryosyl-aminobenzoic acid in children. J Pediatr 1981;135:277-9.
3. Sacher M, Kosba A, Schmerling DH. PABA screening test for
exocrine pancreatic function in infants and children. Arch Dis Child
After Bz-Ty-PABA
60
Normal (no. 1)
0.80
90
0.40
0.12
0.08
1.70
0.80
1.40
1.17
2.10
2.30
60
90
0.54
0.65
-
1.50
1.20
0.78
5.00
4.10
1978;53:639-41.
4. Imondi AR, Stradley RP, Wolgemuth RL. Synthetic peptides in
the diagnosis of exocrine pancreatic insufficiency in animals. Gut
1972;13:726-31.
5. Arvanitakis C, Greenberger NJ. Diagnosis of pancreatic diseases
by a synthetic peptide. Lancet 1976;i:663-6.
6. Gyr K, Stalder GA, Schiffman Z, et al. Oral administration of a
chymotrypsun labile function in man (PFr). Gut 1976;17:27-32.
Nousia-Arvanitakis
5, Ai-vanitakis C, Death N, et al. Diagnosis
1 60
1 70
of exocrine pancreatic insufficiency in cystic fibrosis by the synthetic
1978;92:734-7.
peptide N-benzoyl-i.-tyrosyl-p.aminobenzoic
acid. J Pediatr
after
Subject
Normal (no. 2)
Cystic-fibrosis
patient
Afterfree PABA
Normal (no. 1)
Cystic-fibrosis
patient
PABA
dos&
mln
PAAHA
PAABA
procedure
Concn, mg/L
1.00
60
-
-
-
-
90
-
-
-
-
60
90
0.93
0.73
0.32
0.17
2.80
60
90
0.89
0.46
-
2.80
1.40
1.60
1.20
0.67
5.80
0.56
0.41
2.80
3.90
4.30
8. Koran G, Weiznian Z, Forstner G, et al. Altered PABA pharmacokinetics in cystic fibrosis-implications
for bentiromide test. Dig
DisSci 1985;30:928-32.
9. Weizman Z, Forstner G, Gaskin LU, et al. The bentiromide test
for assessing pancreatic dysfunction using analysis of PABA in
plasma and urine: studies in cystic fibrosis and Shwachman’s
syndrome. Gastroenterology 1985;89:596-604.
CLINICALCHEMISTRY,Vol. 34, No. 11, 1988 2237
10. Lang C, Cyr K, Tonko I, et al. Value of serum PABA as a
pancreatic function test. Gut 1894;25:508-12.
11. Sterchi EE, Green JR, Lentze M. Nonpancreatic hydrolysis of
Bz-Ty-PABA in the human intestine. Cliii Sci 1982;62:557-60.
12. DelchierJC, Soule JC. BT-PABA test with plasma PABA measurements: evaluation of sensitivity
and specificity. Gut 198324:31825.
13. Gharbo SA, Boyer JA, Williamson MJ. Colorimetric plasma
assay for the bentiromide test (BT-PABA) for exocrine pancreatic
insufficiency. .Anal Biochexn 1985;148:228-32.
14 Morita M, Kozu T, Watanabe S, et al. Fundamental and clinical
evaluation of pancreatic function test. Op. cit. (ref. 1), 70-7.
15. Bellentani S, Gaisendi A, Manenti F. Serum PABA test in
chronic pancreatitis.
Gut 1985;26:537-8.
16. Mitchell C.J. Oral pancreatic function tests. In: Mitchell CJ,
Kelleher J, eds. Pancreatic disease in clinical practice. London:
Pitman, 1981;149-65.
17. Noda A, Hayakawa T, Kondo T, et al. Clinical evaluation of
pancreatic excretion test with dimethadione and oral BT-PABA teat
in chronic pancreatitis.
Dig Dis Sci 1983;28:230-5.
18 Bratton AC, Marshall GK. A new coupling component for
sulfanilaxnide determination.
J Biol Chem 1939;128:537-50.
19. Smith MW, Finkelstein N, Alumunosa L, Crawford B, Graber
2238
CLINICAL
CHEMISTRY,
Vol.
34, No. 11, 1988
M. Renal clearances of substituted
hippuric acid derivatives and
other aromatic acids in dog and man. J Clin Invest 1945;24:388404.
20. Karnes HT, Batch RP, Winefordner JD, et al. Determination of
PABA in urine by room temperature phosphorimetry, with application to the bentiromide test for pancreatic function. Clin Chem
1984;30:1565-7.
21. Van der Bergh FAJTM, Klein Elhorst JT. Rapid determination
of the PABA excretion index in urine without use of radioactivity.
Cliii Chem 1986;32:234-6.
22. Cavallini G, Piubello W, Brocco G, et al. Serum PABA and
fluorescein in the course of Bz-Ty-PABA and pancreolauryl test as an
index of exocrine pancreatic insufficiency. Dig Dis Sci 1985;30:65562.
23. Ito 5, Maruta K, Imai Y, et al. Urinary p-aminobenzoic acid
determined in the pancreatic function test by liquid chromatography with electrochemical detection. Clin Chem 1982;28:323-6.
24 Karnea lIT, Riley CM, Curry S, et al. Analysis ofN-Bz-Ty-PABA
metabolites in urine by ion-pair high performance liquid chromatography. J Chromatogr 1985;338:377-88.
25. Schicker AM, Zoest R, Hung CT, et al. Determination
of PABA
and its metabolites in urine using ion-pair HPLC in the NBT-i’ABA
pancreatic function test. Dunedin, NZ: Proceedings, Univ of Ontago
Med School, 1986:63(3):75-6.