Journal of Clinical Laboratory Analysis 6:379-383 (1992)
Automated Microparticle Enzyme lmmunoassay for Neural Thread Protein in Cerebrospinal Fluid From Alzheimer’s Disease Patients Jonathan K. Chong,’ Lois Cantrell,’ Mazhar Husain,* Sue Riesing,’ Barney E. Miller,’ Jack Wands,3 Suzanne de la M ~ n t eand , ~ Hossein A. Ghanbari‘ 7NeuropsychiatricMarkers R&D and 2Depattmentof Applied Immunochemistry, Abbott Diagnostics Division, Abbott Laboratories, Abbott Park, Illinois; 3Molecular Hepatology Laboratory, Cancer Center, Massachusetts Alzheimer’s Disease Research Center and Departments of Medicine and Pathology, Massachusetts General Hospital, Boston, Massachusetts An automated microparticleenzyme immunoassay (MEIA) with the IMxB analyzer for the detection of neural thread protein (NTP) in cerebrospinal fluid (CSF) from Alzheimer’s disease (AD) patients was developed. This assay uses monoclonal antibodies produced against the purified pancreaticform of the protein. The assay employs one monoclonalantibody covalently coupled to the microparticle to capture immunoreactive material in CSF or brain tissue. The second monoclonal antibody was conjugated to alkaline phosphatase and serves as detection antibody. The assay provides results in approximately 45 minutes with a sensitivity of 60 pg/ml (3 fmoles/ml). The titration curve of both normal and AD CSF resulted in a linear relationship with Key words:
0 1992 Wiley-Liss, Inc.
Alzheimer’s disease, cerebrospinalfluid, antemortem Alzheimer’s disease marker, microparticle enzyme immunoassay (MEIA)
INTRODUCTION Elevated levels of pancreatic thread protein (PTP)-like immunoreactivity have been recently reported in brains of patients with Alzheimer’s disease (AD) using a three-sited monoclonal antibody based immunoradiometric assay (IRMA) (1). Further analysis of this pancreatic thread protein-like immunoreactivity in brain and CSF revealed a larger (20 kDa) protein species referred to as neural thread protein (NTP) as compared to the pancreatic form of the protein or PTP (14 kDa) (1). Moreover, mRNA levels were elevated in AD brain when compared to normal brain by using human PTP DNA sequences as a probe (2). Thus, NTP overexpression has been demonstrated in AD brain both at DNA transcription and translation levels. Our laboratory has been interested in developing assays for potential antemortem biochemical markers of AD. Recently, a biochemical enzyme immunoassay (3) that measures Alzheimer’s disease associated protein(s) (ADAP) in brains from AD patients was reported (4). In this communication we report 0 1992 Wiley-Liss, Inc.
respect to the volume of CSF used. A similar relationship was observed when normal and AD brain tissue extracts were serially diluted. The molecular weight of NTP in CSF was approximately 20 kD as determined by gel filtration method under non-denaturing conditions. The recovery for pancreaticthread protein (PTP) spiked in either normal or AD CSF was 104% and 108%, respectively. Intra-, inter-, and total assay CVs (coefficient of variation) for controls were less than 2.9%, 3.3% and 3.0%, respectively. This assay will provide a useful tool in the study of the Alzheimer’s disease and may help research in diagnosis and prognosis of Alzheimer’s disease and related disorders.
an automated enzyme immunoassay for PTP-like immunoreactivity, referred to as IMx NTP assay, using non-radioactive reagents.
MATERIALS AND METHODS Assay Format The assay was performed as follows: About 200 ~1 of CSF or brain tissue extracts or PTP standard solutions were placed into the sample well of an IMx@reaction cell (Abbott Laboratories, Abbott Park, IL). The IMx@NTP reagent pack (comprised of microparticles, conjugate, substrate, diluent), and carousel loaded with the reaction cells were placed in the IMx@ instrument. The assay was initiated by pressing “Run” on
Received April 24, 1992; accepted May 29, 1992. Address reprint requests to Jonathan K . Chong, Neuropsychiatric Markers R&D, D-90J. AP20/4, Abbott Diagnostics Division, Abbott Laboratories, Abbott Park, IL60064.
380
Chong et al.
the instrument panel. During the assay, the IMx@instrument performed the following manipulations: 25 p l aliquot of microparticles and 25 p1 aliquot of diluent were added into the reaction well. After 5 minutes incubation a 150 pl aliquot of specimen was added to the reaction well. The mixture was incubated for 15 minutes and then a 50 p l aliquot of alkaline phosphatase conjugated detection antibody was added to the reaction well. After a 15 minute incubation, a 200 p1 aliquot of the reaction mixture was then aspirated from the reaction well and dispensed onto a glass fiber matrix. The microparticles trapped on the matrix were washed with buffer several times. MUP(4-methyl-umbelliferyl phosphate) substrate was added to the matrix and the rate of development of a fluorescent signal resulting from conversion of the substrate to 4-methylumbelliferone was measured by a surface reflectance reader in the instrument. The results were printed on the tape as pg/ml for each sample. The analysis of 22 Samples takes approximately 45 minutes.
Reagents PTP was purified from human pancreatic tissue as described by the method of Gross et al. (5). Mouse monoclonal antibodies HT110 and HT9 were used for this study (5). Antibodies were purified from tissue culture media using ImmunoPure Plus (A) IgG column using the procedure as suggested by Pierce (Rockford, IL). Carboxymethylated latex microparticles of uniform size (0.285 pm) were puchased from Seradyn (Indianapolis, IN). HT 1 10 mouse monoclonal antibody was covalently coupled to microparticles and used for capturing antigens. To 1 ml reaction volume, 6 mg of microparticles, 0.5 mg of HT110, 5 mM 2-(N-morpholino) ethanesulfonic acid (MES), pH 4.5 and 1 mg of l-ethyl-3-(dimethylaminopropyl)carbodiimide hydrochloride (EDAC) were added. The mixture was incubated for 30 minutes at room temperature on a rotator. The microparticles were blocked with 1% BSA/PBS. HT9 mouse monoclonal antibody was used as detection antibody which was preabsorbed by human IgG-Sepharose immunoaffinity column. The preabsorbed detection antibody was coupled to alkaline phosphatase (Boehringer-Mannheim Biochemicals, Indianapolis, IN). Since human serum albumin (HSA) and gamma globulins constitute most of CSF protein content, an “artificial CSF” matrix comprising of 0.5 mg HSA/ml and 0.1 mg human gamma globulinlml in PBS was used to dilute PTP standards, controls, and CSF samples. Brain tissues from AD and normal control patients were homogenized as described by Ghanbari et al. (3). Low speed supernatants from brain homogenates were used for analysis. AD CSF was fractionated by gel filtration after injecting AD CSF into HPLC gel filtration column (Bio Rad Bio Sil TSK-125, 600 X 7 . 8 mm). All fractions were assayed by IMx@NTP assay. Bio Rad gel filtration molecular weight standard was used to determine the molec-
ular weight of NTP in CSF. Endogenous NTP in CSF was subjected to heat stress test at 45”C, freeze-thaw, room temperature, and 2-8°C to study its stability.
Assay Performance Assay precision was determined by analyzing five replicates of low, medium, and high controls, respectively, for 5 days. The assay precision was calculated by analysis of variance techniques (6). The sensitivity was determined by assaying 0 ng PTP/ml level in replicates of four for 5 days. Two standard deviations above the 0 PTP “artificial CSF” blank was used as the detection limit of the assay. For the recovery study a known amount of exogenous PTP was spiked into each of three normal and AD CSF samples, respectively. The recovery was calcualted by subtracting the endogenous NTP concentrations in unspiked CSF sample from that of corresponding PTP-spiked CSF sample. All CSF samples used in this study were obtained by lumbar puncture of living patients.
Diagnosis of Alzheimer’s Disease (AD) Generally, the clinical diagnosis of AD was based on National Institute of Neurological and Communicative Disorders and Strokes (NINCDS) and the Alzheimer’s Disease and Related Disorders Association (ADRDA) criteria (7): dementia, insidious onset with progressive deteriorating course, and subsequent exclusion of other specific causes of dementia by history, physical examination, and laboratory tests. Dementia was considered to be present if there was evidence of impairment of memory and impairment of either abstract thinking or judgement. The postmortem diagnosis of AD brain was made by pathologic examination such as the presence of neuritic plaques and neurofibrillary tangles in conjunction with a clinical history of dementia (8).
RESULTS AND DISCUSSION Figure 1 shows the relationship between rate and antibodycoated microparticle concentration. The signal approached maximum at 0.25% microparticle. The inhibition effect was observed when excess microparticles were used. Figure 2 shows the effect of BSA blocking in the presence of EDAC after antibody is coupled to the microparticles. BSA-blocked microparticles demonstrated an enhanced signal to noise ratio as compared to non-blocked microparticles. Based on this observation BSA-blocked microparticle was used throughout the study. Figure 3 shows the IMx@PTP calibration curve demonstrating the relationship between PTP (at 0, .25, .5, 1, 2, and 4 ng PTP/ml) and rate of development of fluorescence (counts/s2). Maximum and background rates were 600 and 12 counts/s2,respectively. The curve was calculated by using the four-parameter logistics equation, y = A B/(C xD),
+
+
MElA for NTP in CSF From AD Patients Microparticle
381
titration 600
800 500
400 0
600
z
0
e
:
300
I
C 0 3
400
200
-
2! C
E
100
200 0 0
2
1
PTP
(
0.00
0.02
0 04
0 06
0.08
0 10
C
4
(nglrni)
Fig. 3. Calibration curve for the IMxmNTP assay. A set of calibrators (0, 0.25,0.5, 1, 2 , 4 FTP ng/ml) were assayed with the IMx NTP assay reagents. The line represents the curve fit for concentration of PTP vs. units of fluorescence, drawn with four-parameter logistics. Points represent the average of two determinations.
YO microparticle
Fig. 1. Optimization of microparticle concentration. The numbers show the % (w/v) solid in the reaction well of IMx@reaction cell. Points represent the mean of triplicate determinations.
200
0 0 84
:
3
100
* e
0
C 0
ng PTP
Fig. 2. Effect of BSA blocking on the microparticle in the presence of EDAC. See “Materials and Methods” for details.
where y is fluorescence, x is concentration, A is the maximum rate,B is the negative of the slope of the linear calibration curve plotted in log-logit coordinates, C is the concentration of PTP at the half-maximal rate, and D is the theoretical minimum rate (9). The sensitivity (detection limit) of the assay was 60 pg/ml(3 femtomoles/ml). Figure 4A shows the titration curve of one AD and one normal CSF. A linear titration curve of NTP was observed for both CSF and brain. A measurable level of NTP was observed in the normal CSF, but the NTP level in AD CSF was significantly higher. The titration curve was extrapolated to 0 NTP concentration (0 p-1 of CSF). The intercepts of AD and normal CSF were 0.15 NTP (ng/ml) and 0.07 NTP (ng/ml), respectively. Figure 4B shows the titration curve of NTP from normal and AD brain. The intercepts of AD and normal brain were 0.11 NTP (ng/ml) and - 0.03 NTP (ng/ml), respectively. Figure 5 shows the HPLC gel fractionation of antemortem CSF obtained from a clinically diagnosed probable AD patient. A single peak was observed at approximately 20 kD molecular weight. This molecular weight was similar to what was previously reported (1) as determined by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (SDS-PAGE). Figure 6 shows the antigen stability in CSF. Antigen was stable for at least 14days at 2-8”C, room temperature, - 70°C freeze-thaw, and 45°C conditions.
382
Chong et al. 40
6
0
20
60
40
80
0
0
100
UI ot CSF
20
AD brain
40
60
80
100
UI of low speed supernatant
Fig. 4. A: The titration curve of NTP from AD (90 year old female) and normal (70 year old male) CSF. The correlation coefficient for AD and norma1 CSF was 0.997 and 0.998, respectively. 6:The titration curve of NTP from AD (brain extract pool from 98 year old female and 78 year old male
patients) and normal brain (61 year old male). The correlation coefficient for AD and normal brain was 0.992 and 0.788, respectively. Points represent the mean of triplicate determinations.
Table 1 summarizes the results of the precision study. Five replicates of each control were performed for 5 days. Within-, between-, and total assay CVs (coefficient of variation) for low, medium, and high controls were less than 2.9%, 3.4%, and 3.0%,respectively. Table 2 shows the results of the recovery study of PTP from normal and AD CSF samples. The mean recovery from normal and AD CSF was 104% and 108%, respectively. The mean overall recovery for both normal and AD CSF was 105%.
Stability of NTP in A D CSF
200
U 0 a
50 OJ, 0.16
, , , 4.92
I
, , 9.68
I
, , ,
I
,
1 ' 1
14.45
1921
, , , 23.97
1
,
I
-
NTPRT
-0-
NTP.45oC
---C
NTPFI
J
27 15
Elution Time (rnin.)
Fig. 5. A HPLC gel filtration column (Bio Rad Bio Sil TSK-125, 600 x 7.8 mm) was equilibrated with 0.1 M phosphate/0.3 M NaC1/0.05% NaN3, pH 6.8 at a flow rate of 1 ml/min. After equilibration, 300 ~1 of antemortem CSF obtained from a clinically diagnosed probable AD patient (87 year old female) was injected into the column. All fractions were collected and tested by IMx@NTP assay. Bio Rad gel filtration molecular weight standard (bovine thyroglobulin 670,000, bovine IgC 158,000, ovalabumin [egg chicken] 44,OOO. horse myoglobulin 17,000, and vitamin B12 1,350) was used to determine the molecular weight of NTP in CSF. Solid line is absorbance at 280 nm, and dotted line is NTP concentration as expressed ngiml.
0 0
2
4
6
8
10
12
'
TimeIDays)
Fig. 6. The stability of NTP in CSF. NTP containing CSF samples were stored at 2-8"C, room temperature, and 45°C continually. For freeze-thaw study, CSF was frozen at - 70°C and thawed at room temperature prior to analysis. NTP irnmunoreactivity was measured on an 1Mx@instrument at days 1 , 2 , 3 , 5 , 7 , and 14.
MElA for NTP in CSF From AD Patients
TABLE 1. Precision Study of IMx@NTP Assay“ Control
Number
Mean PTP (ngiml)
Within-
Between-
Total
Low Medium High
25 25 25
0.42 1.27 2.81
2.8 2.8 1.9
0.7 1.7 2.3
2.9 3.3 3 .O
”Low, medium, and high controls were prepared in “artificial CSF.” CV = coefficient of variation.
TABLE 2. Recovery of PTP From Normal and AD CSF PTP fndml) Sample Normal CSF Ras 2 Normal CSF Rie 7 Normal CSF Rie 3 AD CSF Ban H2 414 AD CSF Rie 17 AD CSF Ban H2 491
Age/sex
Expected
Observed”
79F 60F 37F 69F 68F 79F
2.5 2.5 2.5 2.5 2.5 2.5
2.6 2.5 2.6 2.7 2.9 2.5
% recovery
104 100
104 108 116 100
“Determined by IMx@NTP assay. The mean recovery for normal (n = 3) and AD (n = 3) CSF was 2.6 0.06 and 2.7 ? 0.20 FTP (ngiml), respectively. The total recovery was 2.6 2 0.15 PTP ( n g h l ) .
*
CONCLUSION We have developed an automated enzyme imunoassay that reliably measures NTP ( M W 20 kD) in CSF from AD patients. The IMx@NTP assay is rapid and reproducible. This assay will provide a useful tool in the area of Alzheimer’s disease research and diagnostics, and may help further research efforts to achieve better understanding of the disease itself. A clinical study that involves a large number of antemortern CSF samples is being conducted in our laboratory. The specificity of NTP as a potential antemortem AD diagnostic marker is being investigated in collaboration with several Alzheimer’s disease research centers.
ACKNOWLEDGMENTS We are grateful to Dr. Harold Baker, Fertility, Pregnancy and Neurodiagnostics, Abbott Diagnostics Division, Abbot1
383
Laboratories, for the invaluable suggestions during the development of this assay. Cerebrospinal fluid specimens were kindly provided by the following investigators: Csaba M. Banki, M.D., Ph.D., Regional Neuropsychiatric Institute, H-4321 Nagy Kallo, Szechenyi U 29, Hungary; Murray Raskind, M.D., Department of Gerontology, Veterans Medical Center, University of Washington, Seattle; Paavo Riekkinen, M.D., Department of Neurology, University of Kuopio, SF-7021 1 Kuopio, Finland. Brain tissues were kindly provided by Peter Davies, Ph.D., Department of Pathology, Albert Einstein College of Medicine of Yeshiva University, Bronx, New York.
REFERENCES I . Ozturk M, de la Monte SM, Gross J, Wands JR: Elevated levels of an exocrine pancreatic secretory protein in Alzheimer’s disease brain. Proc NatlAcadSci USA 86:419-423, 1989. 2. de la Monte SM, Ozturk M, Wands JR: Enhanced expression of an exocrine pancreatic protein in Alzheimer’s Disease and the developing brain. J Clin Znvest 86: 1004- 1013, 1990. 3. Ghanbari HA, Kozuk T, Miller BE, Riesing S: A sandwich enzyme immunoassay for detecting and measuring Alzheimer’s Disease-Associated Proteins in human brain tissue. J CIin Lab Anal 4: 189-192, 1990. 4. Ghanbari HA, Miller BE, Haigler HJ, Arato M, Bisette G, Davies P, Nemeroff CB, Peny EK, Perry R, Ravid R, Swaab DF, Whetsell WO, Zemlan FP: Biochemical assay of Alzheimer’s Disease-Associated Protein(s) in human brain tissue. JAMA 263:2907-2910, 1990. 5. Gross J, Carlson RI, Brauer AW, Margolies MN, Warshaw AL, Wands JR: Isolation, characterization, and distribution of an unusual pancreatic human secretory protein. JClinInvest 76:2115-2126, 1985. 6. Colton T: Statistics in Medicine. Little, Brown, and Co., Boston, 1974, pp 11-62, 7. McKhann G , Drachman D, Folstein M, Katzman R, Price D, Stadlan EM: Clinical diagnosis of Alzheimer’s disease: Report of the NINCDSADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s disease. Neurology 34:939-944, 1984. 8. Khachaturian ZS: Diagnosis of Alzheimer’s disease. Arch Neurol 42:1097-1105, 1985. 9. De Lean A, Munson PJ, Rodbard D: Simultaneous analysis of families of sigmoidal curves: Application to bioassay, radioligand assay, and physiological dose-response curves. Am J Physiol235:E97-E104, 1978.
Automated Microparticle Enzyme lmmunoassay for Neural Thread Protein in Cerebrospinal Fluid From Alzheimer’s Disease Patients Jonathan K. Chong,’ Lois Cantrell,’ Mazhar Husain,* Sue Riesing,’ Barney E. Miller,’ Jack Wands,3 Suzanne de la M ~ n t eand , ~ Hossein A. Ghanbari‘ 7NeuropsychiatricMarkers R&D and 2Depattmentof Applied Immunochemistry, Abbott Diagnostics Division, Abbott Laboratories, Abbott Park, Illinois; 3Molecular Hepatology Laboratory, Cancer Center, Massachusetts Alzheimer’s Disease Research Center and Departments of Medicine and Pathology, Massachusetts General Hospital, Boston, Massachusetts An automated microparticleenzyme immunoassay (MEIA) with the IMxB analyzer for the detection of neural thread protein (NTP) in cerebrospinal fluid (CSF) from Alzheimer’s disease (AD) patients was developed. This assay uses monoclonal antibodies produced against the purified pancreaticform of the protein. The assay employs one monoclonalantibody covalently coupled to the microparticle to capture immunoreactive material in CSF or brain tissue. The second monoclonal antibody was conjugated to alkaline phosphatase and serves as detection antibody. The assay provides results in approximately 45 minutes with a sensitivity of 60 pg/ml (3 fmoles/ml). The titration curve of both normal and AD CSF resulted in a linear relationship with Key words:
0 1992 Wiley-Liss, Inc.
Alzheimer’s disease, cerebrospinalfluid, antemortem Alzheimer’s disease marker, microparticle enzyme immunoassay (MEIA)
INTRODUCTION Elevated levels of pancreatic thread protein (PTP)-like immunoreactivity have been recently reported in brains of patients with Alzheimer’s disease (AD) using a three-sited monoclonal antibody based immunoradiometric assay (IRMA) (1). Further analysis of this pancreatic thread protein-like immunoreactivity in brain and CSF revealed a larger (20 kDa) protein species referred to as neural thread protein (NTP) as compared to the pancreatic form of the protein or PTP (14 kDa) (1). Moreover, mRNA levels were elevated in AD brain when compared to normal brain by using human PTP DNA sequences as a probe (2). Thus, NTP overexpression has been demonstrated in AD brain both at DNA transcription and translation levels. Our laboratory has been interested in developing assays for potential antemortem biochemical markers of AD. Recently, a biochemical enzyme immunoassay (3) that measures Alzheimer’s disease associated protein(s) (ADAP) in brains from AD patients was reported (4). In this communication we report 0 1992 Wiley-Liss, Inc.
respect to the volume of CSF used. A similar relationship was observed when normal and AD brain tissue extracts were serially diluted. The molecular weight of NTP in CSF was approximately 20 kD as determined by gel filtration method under non-denaturing conditions. The recovery for pancreaticthread protein (PTP) spiked in either normal or AD CSF was 104% and 108%, respectively. Intra-, inter-, and total assay CVs (coefficient of variation) for controls were less than 2.9%, 3.3% and 3.0%, respectively. This assay will provide a useful tool in the study of the Alzheimer’s disease and may help research in diagnosis and prognosis of Alzheimer’s disease and related disorders.
an automated enzyme immunoassay for PTP-like immunoreactivity, referred to as IMx NTP assay, using non-radioactive reagents.
MATERIALS AND METHODS Assay Format The assay was performed as follows: About 200 ~1 of CSF or brain tissue extracts or PTP standard solutions were placed into the sample well of an IMx@reaction cell (Abbott Laboratories, Abbott Park, IL). The IMx@NTP reagent pack (comprised of microparticles, conjugate, substrate, diluent), and carousel loaded with the reaction cells were placed in the IMx@ instrument. The assay was initiated by pressing “Run” on
Received April 24, 1992; accepted May 29, 1992. Address reprint requests to Jonathan K . Chong, Neuropsychiatric Markers R&D, D-90J. AP20/4, Abbott Diagnostics Division, Abbott Laboratories, Abbott Park, IL60064.
380
Chong et al.
the instrument panel. During the assay, the IMx@instrument performed the following manipulations: 25 p l aliquot of microparticles and 25 p1 aliquot of diluent were added into the reaction well. After 5 minutes incubation a 150 pl aliquot of specimen was added to the reaction well. The mixture was incubated for 15 minutes and then a 50 p l aliquot of alkaline phosphatase conjugated detection antibody was added to the reaction well. After a 15 minute incubation, a 200 p1 aliquot of the reaction mixture was then aspirated from the reaction well and dispensed onto a glass fiber matrix. The microparticles trapped on the matrix were washed with buffer several times. MUP(4-methyl-umbelliferyl phosphate) substrate was added to the matrix and the rate of development of a fluorescent signal resulting from conversion of the substrate to 4-methylumbelliferone was measured by a surface reflectance reader in the instrument. The results were printed on the tape as pg/ml for each sample. The analysis of 22 Samples takes approximately 45 minutes.
Reagents PTP was purified from human pancreatic tissue as described by the method of Gross et al. (5). Mouse monoclonal antibodies HT110 and HT9 were used for this study (5). Antibodies were purified from tissue culture media using ImmunoPure Plus (A) IgG column using the procedure as suggested by Pierce (Rockford, IL). Carboxymethylated latex microparticles of uniform size (0.285 pm) were puchased from Seradyn (Indianapolis, IN). HT 1 10 mouse monoclonal antibody was covalently coupled to microparticles and used for capturing antigens. To 1 ml reaction volume, 6 mg of microparticles, 0.5 mg of HT110, 5 mM 2-(N-morpholino) ethanesulfonic acid (MES), pH 4.5 and 1 mg of l-ethyl-3-(dimethylaminopropyl)carbodiimide hydrochloride (EDAC) were added. The mixture was incubated for 30 minutes at room temperature on a rotator. The microparticles were blocked with 1% BSA/PBS. HT9 mouse monoclonal antibody was used as detection antibody which was preabsorbed by human IgG-Sepharose immunoaffinity column. The preabsorbed detection antibody was coupled to alkaline phosphatase (Boehringer-Mannheim Biochemicals, Indianapolis, IN). Since human serum albumin (HSA) and gamma globulins constitute most of CSF protein content, an “artificial CSF” matrix comprising of 0.5 mg HSA/ml and 0.1 mg human gamma globulinlml in PBS was used to dilute PTP standards, controls, and CSF samples. Brain tissues from AD and normal control patients were homogenized as described by Ghanbari et al. (3). Low speed supernatants from brain homogenates were used for analysis. AD CSF was fractionated by gel filtration after injecting AD CSF into HPLC gel filtration column (Bio Rad Bio Sil TSK-125, 600 X 7 . 8 mm). All fractions were assayed by IMx@NTP assay. Bio Rad gel filtration molecular weight standard was used to determine the molec-
ular weight of NTP in CSF. Endogenous NTP in CSF was subjected to heat stress test at 45”C, freeze-thaw, room temperature, and 2-8°C to study its stability.
Assay Performance Assay precision was determined by analyzing five replicates of low, medium, and high controls, respectively, for 5 days. The assay precision was calculated by analysis of variance techniques (6). The sensitivity was determined by assaying 0 ng PTP/ml level in replicates of four for 5 days. Two standard deviations above the 0 PTP “artificial CSF” blank was used as the detection limit of the assay. For the recovery study a known amount of exogenous PTP was spiked into each of three normal and AD CSF samples, respectively. The recovery was calcualted by subtracting the endogenous NTP concentrations in unspiked CSF sample from that of corresponding PTP-spiked CSF sample. All CSF samples used in this study were obtained by lumbar puncture of living patients.
Diagnosis of Alzheimer’s Disease (AD) Generally, the clinical diagnosis of AD was based on National Institute of Neurological and Communicative Disorders and Strokes (NINCDS) and the Alzheimer’s Disease and Related Disorders Association (ADRDA) criteria (7): dementia, insidious onset with progressive deteriorating course, and subsequent exclusion of other specific causes of dementia by history, physical examination, and laboratory tests. Dementia was considered to be present if there was evidence of impairment of memory and impairment of either abstract thinking or judgement. The postmortem diagnosis of AD brain was made by pathologic examination such as the presence of neuritic plaques and neurofibrillary tangles in conjunction with a clinical history of dementia (8).
RESULTS AND DISCUSSION Figure 1 shows the relationship between rate and antibodycoated microparticle concentration. The signal approached maximum at 0.25% microparticle. The inhibition effect was observed when excess microparticles were used. Figure 2 shows the effect of BSA blocking in the presence of EDAC after antibody is coupled to the microparticles. BSA-blocked microparticles demonstrated an enhanced signal to noise ratio as compared to non-blocked microparticles. Based on this observation BSA-blocked microparticle was used throughout the study. Figure 3 shows the IMx@PTP calibration curve demonstrating the relationship between PTP (at 0, .25, .5, 1, 2, and 4 ng PTP/ml) and rate of development of fluorescence (counts/s2). Maximum and background rates were 600 and 12 counts/s2,respectively. The curve was calculated by using the four-parameter logistics equation, y = A B/(C xD),
+
+
MElA for NTP in CSF From AD Patients Microparticle
381
titration 600
800 500
400 0
600
z
0
e
:
300
I
C 0 3
400
200
-
2! C
E
100
200 0 0
2
1
PTP
(
0.00
0.02
0 04
0 06
0.08
0 10
C
4
(nglrni)
Fig. 3. Calibration curve for the IMxmNTP assay. A set of calibrators (0, 0.25,0.5, 1, 2 , 4 FTP ng/ml) were assayed with the IMx NTP assay reagents. The line represents the curve fit for concentration of PTP vs. units of fluorescence, drawn with four-parameter logistics. Points represent the average of two determinations.
YO microparticle
Fig. 1. Optimization of microparticle concentration. The numbers show the % (w/v) solid in the reaction well of IMx@reaction cell. Points represent the mean of triplicate determinations.
200
0 0 84
:
3
100
* e
0
C 0
ng PTP
Fig. 2. Effect of BSA blocking on the microparticle in the presence of EDAC. See “Materials and Methods” for details.
where y is fluorescence, x is concentration, A is the maximum rate,B is the negative of the slope of the linear calibration curve plotted in log-logit coordinates, C is the concentration of PTP at the half-maximal rate, and D is the theoretical minimum rate (9). The sensitivity (detection limit) of the assay was 60 pg/ml(3 femtomoles/ml). Figure 4A shows the titration curve of one AD and one normal CSF. A linear titration curve of NTP was observed for both CSF and brain. A measurable level of NTP was observed in the normal CSF, but the NTP level in AD CSF was significantly higher. The titration curve was extrapolated to 0 NTP concentration (0 p-1 of CSF). The intercepts of AD and normal CSF were 0.15 NTP (ng/ml) and 0.07 NTP (ng/ml), respectively. Figure 4B shows the titration curve of NTP from normal and AD brain. The intercepts of AD and normal brain were 0.11 NTP (ng/ml) and - 0.03 NTP (ng/ml), respectively. Figure 5 shows the HPLC gel fractionation of antemortem CSF obtained from a clinically diagnosed probable AD patient. A single peak was observed at approximately 20 kD molecular weight. This molecular weight was similar to what was previously reported (1) as determined by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (SDS-PAGE). Figure 6 shows the antigen stability in CSF. Antigen was stable for at least 14days at 2-8”C, room temperature, - 70°C freeze-thaw, and 45°C conditions.
382
Chong et al. 40
6
0
20
60
40
80
0
0
100
UI ot CSF
20
AD brain
40
60
80
100
UI of low speed supernatant
Fig. 4. A: The titration curve of NTP from AD (90 year old female) and normal (70 year old male) CSF. The correlation coefficient for AD and norma1 CSF was 0.997 and 0.998, respectively. 6:The titration curve of NTP from AD (brain extract pool from 98 year old female and 78 year old male
patients) and normal brain (61 year old male). The correlation coefficient for AD and normal brain was 0.992 and 0.788, respectively. Points represent the mean of triplicate determinations.
Table 1 summarizes the results of the precision study. Five replicates of each control were performed for 5 days. Within-, between-, and total assay CVs (coefficient of variation) for low, medium, and high controls were less than 2.9%, 3.4%, and 3.0%,respectively. Table 2 shows the results of the recovery study of PTP from normal and AD CSF samples. The mean recovery from normal and AD CSF was 104% and 108%, respectively. The mean overall recovery for both normal and AD CSF was 105%.
Stability of NTP in A D CSF
200
U 0 a
50 OJ, 0.16
, , , 4.92
I
, , 9.68
I
, , ,
I
,
1 ' 1
14.45
1921
, , , 23.97
1
,
I
-
NTPRT
-0-
NTP.45oC
---C
NTPFI
J
27 15
Elution Time (rnin.)
Fig. 5. A HPLC gel filtration column (Bio Rad Bio Sil TSK-125, 600 x 7.8 mm) was equilibrated with 0.1 M phosphate/0.3 M NaC1/0.05% NaN3, pH 6.8 at a flow rate of 1 ml/min. After equilibration, 300 ~1 of antemortem CSF obtained from a clinically diagnosed probable AD patient (87 year old female) was injected into the column. All fractions were collected and tested by IMx@NTP assay. Bio Rad gel filtration molecular weight standard (bovine thyroglobulin 670,000, bovine IgC 158,000, ovalabumin [egg chicken] 44,OOO. horse myoglobulin 17,000, and vitamin B12 1,350) was used to determine the molecular weight of NTP in CSF. Solid line is absorbance at 280 nm, and dotted line is NTP concentration as expressed ngiml.
0 0
2
4
6
8
10
12
'
TimeIDays)
Fig. 6. The stability of NTP in CSF. NTP containing CSF samples were stored at 2-8"C, room temperature, and 45°C continually. For freeze-thaw study, CSF was frozen at - 70°C and thawed at room temperature prior to analysis. NTP irnmunoreactivity was measured on an 1Mx@instrument at days 1 , 2 , 3 , 5 , 7 , and 14.
MElA for NTP in CSF From AD Patients
TABLE 1. Precision Study of IMx@NTP Assay“ Control
Number
Mean PTP (ngiml)
Within-
Between-
Total
Low Medium High
25 25 25
0.42 1.27 2.81
2.8 2.8 1.9
0.7 1.7 2.3
2.9 3.3 3 .O
”Low, medium, and high controls were prepared in “artificial CSF.” CV = coefficient of variation.
TABLE 2. Recovery of PTP From Normal and AD CSF PTP fndml) Sample Normal CSF Ras 2 Normal CSF Rie 7 Normal CSF Rie 3 AD CSF Ban H2 414 AD CSF Rie 17 AD CSF Ban H2 491
Age/sex
Expected
Observed”
79F 60F 37F 69F 68F 79F
2.5 2.5 2.5 2.5 2.5 2.5
2.6 2.5 2.6 2.7 2.9 2.5
% recovery
104 100
104 108 116 100
“Determined by IMx@NTP assay. The mean recovery for normal (n = 3) and AD (n = 3) CSF was 2.6 0.06 and 2.7 ? 0.20 FTP (ngiml), respectively. The total recovery was 2.6 2 0.15 PTP ( n g h l ) .
*
CONCLUSION We have developed an automated enzyme imunoassay that reliably measures NTP ( M W 20 kD) in CSF from AD patients. The IMx@NTP assay is rapid and reproducible. This assay will provide a useful tool in the area of Alzheimer’s disease research and diagnostics, and may help further research efforts to achieve better understanding of the disease itself. A clinical study that involves a large number of antemortern CSF samples is being conducted in our laboratory. The specificity of NTP as a potential antemortem AD diagnostic marker is being investigated in collaboration with several Alzheimer’s disease research centers.
ACKNOWLEDGMENTS We are grateful to Dr. Harold Baker, Fertility, Pregnancy and Neurodiagnostics, Abbott Diagnostics Division, Abbot1
383
Laboratories, for the invaluable suggestions during the development of this assay. Cerebrospinal fluid specimens were kindly provided by the following investigators: Csaba M. Banki, M.D., Ph.D., Regional Neuropsychiatric Institute, H-4321 Nagy Kallo, Szechenyi U 29, Hungary; Murray Raskind, M.D., Department of Gerontology, Veterans Medical Center, University of Washington, Seattle; Paavo Riekkinen, M.D., Department of Neurology, University of Kuopio, SF-7021 1 Kuopio, Finland. Brain tissues were kindly provided by Peter Davies, Ph.D., Department of Pathology, Albert Einstein College of Medicine of Yeshiva University, Bronx, New York.
REFERENCES I . Ozturk M, de la Monte SM, Gross J, Wands JR: Elevated levels of an exocrine pancreatic secretory protein in Alzheimer’s disease brain. Proc NatlAcadSci USA 86:419-423, 1989. 2. de la Monte SM, Ozturk M, Wands JR: Enhanced expression of an exocrine pancreatic protein in Alzheimer’s Disease and the developing brain. J Clin Znvest 86: 1004- 1013, 1990. 3. Ghanbari HA, Kozuk T, Miller BE, Riesing S: A sandwich enzyme immunoassay for detecting and measuring Alzheimer’s Disease-Associated Proteins in human brain tissue. J CIin Lab Anal 4: 189-192, 1990. 4. Ghanbari HA, Miller BE, Haigler HJ, Arato M, Bisette G, Davies P, Nemeroff CB, Peny EK, Perry R, Ravid R, Swaab DF, Whetsell WO, Zemlan FP: Biochemical assay of Alzheimer’s Disease-Associated Protein(s) in human brain tissue. JAMA 263:2907-2910, 1990. 5. Gross J, Carlson RI, Brauer AW, Margolies MN, Warshaw AL, Wands JR: Isolation, characterization, and distribution of an unusual pancreatic human secretory protein. JClinInvest 76:2115-2126, 1985. 6. Colton T: Statistics in Medicine. Little, Brown, and Co., Boston, 1974, pp 11-62, 7. McKhann G , Drachman D, Folstein M, Katzman R, Price D, Stadlan EM: Clinical diagnosis of Alzheimer’s disease: Report of the NINCDSADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s disease. Neurology 34:939-944, 1984. 8. Khachaturian ZS: Diagnosis of Alzheimer’s disease. Arch Neurol 42:1097-1105, 1985. 9. De Lean A, Munson PJ, Rodbard D: Simultaneous analysis of families of sigmoidal curves: Application to bioassay, radioligand assay, and physiological dose-response curves. Am J Physiol235:E97-E104, 1978.
