See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/46035445

Development of a second generation anti-

Mullerian hormone (AMH) ELISA

Article in Journal of immunological methods · October 2010

DOI: 10.1016/j.jim.2010.08.011 · Source: PubMed

CITATIONS READS

97 612

5 authors, including:

Ajay Kumar Amita Patel

University of Bologna Veer Narmad South Gujarat University
23 PUBLICATIONS 1,153 CITATIONS 3 PUBLICATIONS 116 CITATIONS

SEE PROFILE SEE PROFILE

William Roudebush
University of South Carolina School of Medicin…
126 PUBLICATIONS 1,300 CITATIONS

SEE PROFILE

All content following this page was uploaded by William Roudebush on 21 December 2016.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
1 Development of a Second Generation Anti-Müllerian Hormone (AMH) ELISA
2
3 Ajay Kumar1*, Bhanu Kalra1, Amita Patel1, Lauren McDavid1, William E.
4 Roudebush2
5
6 1. Beckman Coulter, Inc., Webster, TX.
7 2. Beckman Coulter, Inc., Chaska, MN
8
9 *Author for correspondence. 445 Medical Center Blvd., Webster, TX 77598; Ph
10 281-554-0551; Fax 281-338-1895; e-mail Ajay.kumar@beckman.com
11
12 Key Words: Anti-müllerian hormone, Müllerian Inhibiting substance, infertility,
13 gametogenesis, ovarian reserve, In Vitro Fertilization
14
15 Abstract:
16 AMH is a glycoprotein dimer composed of two 72kDa monomers linked by
17 disulfide bridges. It belongs to the transforming growth factor-β family. AMH
18 performs various physiological functions. In males, AMH is secreted by the
19 Sertoli cells of the testis. During embryonic development, AMH is responsible for
20 Müllerian duct regression. AMH continues to be produced by the testicles until
21 puberty and then decreases slowly to residual post-puberty values. In females,
22 AMH is produced in small amounts by ovarian granulosa cells after birth, until
23 menopause, and then becomes undetectable. A two-step, sandwich-type
24 enzymatic microplate assay has been developed to measure AMH levels in 20
25 µL of sample in less than 3 hours. AMH calibrators range from 0.2-28 ng/mL.
26 The antibodies used in the assay bind to the mature region of AMH, which is
27 more stable to proteolysis compared to pro-hormone region. The AMH Gen II
28 assay (Beckman Coulter, Inc., Webster, Texas) was standardized to the
29 Immunotech (IOT, Beckman Coulter, Inc., Marseilles, France) AMH assay. AMH
30 Gen II, when compared to IOT using 120 serum samples in the range of 0-20.4
31 ng/mL yielded a correlation coefficient of 0.98 and a slope of 1.0. Total
32 imprecision, calculated on four samples over 40 runs, four replicates per run,
33 between two lots using CLSI EP5-A guidelines, was 5.7% at 3.8 ng/mL, 7.7% at
34 4.4 ng/mL, 5.8% at 14 ng/mL and 5.3% at 16.4 ng/mL. The average analytical
35 sensitivity calculated by the interpolation of the mean plus two standard
36 deviations of 16 replicates of the zero calibrator on two independent lots was
37 0.08 ng/mL. Dilution and spiking studies showed an average recovery of 91-
38 110%. Lot-to-lot comparison of two independent lots testing 38 serum samples
39 (1.5-33 ng/mL range) yielded a slope of 1.01, intercept of -0.08 ng/mL and r of
40 0.99. When potential interferents (hemoglobin, triglycerides, and bilirubin) were
41 added at two times the physiological concentrations, AMH concentrations were
42 within ±10% of the control. A highly specific and reproducible microplate AMH
43 Gen II assay has been developed to standardize the measurement of AMH
44 between methods. The performance of the AMH Gen II assay is ideal for
45 investigation into the physiologic roles of AMH in men and women.
46
47 Anti-Müllerian hormone (AMH), also called Müllerian inhibiting substance (MIS),
48 is a 140-kDa dimeric glycoprotein hormone and belongs to the transforming
49 growth factor- (TGF-) family. In contrast to other TGF- superfamily members,
50 it is believed that AMH requires the N-terminal domain to potentiate activity of the
51 C-terminal domain to attain full bioactivity (1). Between 5-20% of AMH is cleaved
52 at a specific site between the N-terminal domain and the C-terminal domain of
53 the 72-kDa monomer during cytoplasmic transit, to form two polypeptides of 58-
54 kDa (pro region) and 12-kDa (mature region). These two parts of the molecule
55 remain in non-covalent attachment. The human gene coding for AMH has been
56 sequenced and isolated, and is located on the short arm of chromosome 19 (2).
57 The structure of the specific receptor for AMH has been isolated and
58 characterized (3, 4). Across species, AMH retains 11 to 12 conserved cystine
59 residues, of which seven are located in the mature region. The mature region
60 also demonstrates the greatest degree of homology between species, with 108 of
61 the last 112 residues being conserved between rat and human sequences (5).
62 AMH performs various physiological functions. In males, AMH is secreted by the
63 Sertoli cells of the testes. During embryonic development in males, secretion of
64 AMH from testicular Sertoli cells is responsible for the regression of the Müllerian
65 duct, and thus the normal development of the male reproductive tract (6). In the
66 male, secretion of AMH by the Sertoli cells commences during embryogenesis
67 and continues throughout life. AMH continues to be produced by the testicles
68 until puberty and then decreases slowly to residual post-puberty values (7). In
69 females, AMH is produced in small amounts by ovarian granulosa cells after birth
70 until menopause, and then becomes undetectable.
71 Several clinical applications for measuring serum and plasma AMH in humans
72 have been published. Since AMH secretion is not dependent on other hormones,
73 particularly the gonadotropins, and is expressed at a constant level independent
74 of the cycle, it makes AMH very attractive as a direct measurement of ovarian
75 reserve (8-11). It has been suggested that AMH is the single best predictor of
76 poor response for assisted reproductive technology (ART) (12). AMH not only
77 has demonstrated strong relationship to antral follicle count (AFC), but has
 78 proven best as compared to other typical biomarkers with AFC (8). AMH has
79 also been studied to diagnose intersex disorders in children (13), precocious
80 puberty and the delayed onset of puberty, cryptorchidism, anorchidism, male
81 gonadal function (7) and monitoring of granulosa cell cancer patients (14).
82 The need for specific and sensitive AMH assay for humans and animals has led
83 to the development of various AMH assays (14-16). The first reported AMH
84 immunoassay, developed by Hudson et al (15), used a pair of monoclonal
85 antibodies raised against human recombinant AMH, both directed to epitopes in
86 the pro region. However, the use of this assay by Lee et al (17) showed results
87 with variability in AMH concentration due to storage and freeze-thaw instability.
88 The variability in the AMH concentrations observed may be due to the extent of
89 processing of the AMH protein in vivo. A second assay (14) uses a pair of
90 monoclonal antibodies, one directed to the pro region and the other to the mature
91 region of the AMH molecule. A third assay (16) developed more recently, uses a
92 monoclonal antibody pair directed to epitopes in the pro region. Careful
93 attention to sample collection and storage may be required if reliable results are
94 to be obtained from these assays.
95 The AMH Gen II assay uses a pair of monoclonal antibodies directed to epitopes
96 in the mature region of AMH (18). The AMH measured by this assay is not
97 affected by proteolysis of AMH in the sample as the mature region is more stable
98 against proteolysis compared to the pro region, in part due to its multiple cystine
99 residues. In addition, the Gen II assay measures AMH in human, monkey,
100 bovine, other mammalian species with improved performance and the assay
101 correlates well with the commercially available IOT assay.
102
103 Materials and Methods
104 COATING OF MICROTITER PLATES
105 The immunization protocol, development of monoclonal antibodies to human
106 AMH and the screening procedure for the selection of the antibody to the mature
107 region of AMH has been published (18). The detailed coating protocol is
108 available in the supporting Information.
109
110 BIOTINYLATION OF ANTI-AMH MONOCLONAL ANTIBODY.
111 Monoclonal antibody F2B/7A/A was biotinylated with NHS-LC-Biotin and the
112 detailed conjugation procedure is available in the supporting information.
113
114 CALIBRATION CURVE
115 AMH Gen II calibrators were made in heat-inactivated bovine calf serum
116 (Equitech-bio, Inc. Kerrville, TX, cat # SBCU-30) using native AMH and
117 calibration curve details are available in the supporting information.
118
119 CONTROLS
120 Fresh serum samples (PromedDx, Norton, MA, USA, www.promedDx.com)
121 containing various levels of AMH were used as Quality controls. The nominal
122 concentrations of the control samples were established by analyzing the samples
123 in the AMH Gen II ELISA.
124
125 ASSAY PROCEDURE
126 AMH Gen II ELISA is an enzymatically amplified two-site immunoassay the
127 detailed procedure is available in the supporting information.
128
129 SAMPLE TYPE
130 All samples were collected under an institutional review board (IRB) approved
131 protocol (PromedDx, Norton, MA, USA, www.PromedDx.com). Matched blood
132 samples from 120 volunteers (60 male and 60 female, between the ages of 20-
133 50 years) were drawn using serum SST tubes (BD Plastic, Tube 367987) and
134 lithium heparin plasma tubes (BD Plastic, Tube 367880). One hundred and six
135 cycle day-3 female serum samples and 35 randomly selected from pediatric
136 males and females samples with median ages 4.8 and 5.0 years, respectively,
137 were studied. Forty-five post-menopausal serum samples from women with
138 median age of 71 years of age were used to study false positives in the new
139 assay. The AMH concentration in 20 first trimester maternal serum samples
140 were evaluated and correlated to the sex of fetuses (10 male and 10 female).
141
142 STATISTICAL METHOD
143 The raw data were reduced using the log-log cubic regression curve fit as
144 described above. Data analyses were performed using Excel 2003 (Microsoft
145 Corp.), Analyse-it® version 1.73 (Analyse-it® Software limited, Microsoft Corp.).
146 Descriptive data are presented as the mean and SD unless otherwise specified.
147 Passing & Bablok method conversion and Spearman Rank Correlation methods
148 were used to compare data between methods unless otherwise specified.
149
150 VALIDATION PROCEDURES
151 The validation of the AMH Gen II assay was carried out following the
152 manufacturer’s (DSL- Beckman Coulter, Inc., Webster, TX, USA, AMH Gen II,
153 cat # A73818 & A 73819, Research use only) recommendations for preparation
154 and storage of reagents, calibrators, and controls and for running the protocol.
155
156 SENSITIVITY
157 LIMIT OF BLANK
158 The study was designed following CLSI EP 17 guidelines (20). One post
159 menopausal female serum sample with no AMH was run in duplicate over 20
160 inter-assay runs. The concentration of the sample (no analyte) observed at 95%
161 CI was calculated.
162
163 LIMIT OF DETECTION
164 The study was designed following CLSI EP17 guidelines (20). Six human serum
165 samples with AMH concentrations of 0.0, 0.10, 0.16, 0.20, 0.57 and 1.40 ng/mL
166 were run in duplicate with seven point calibration curve and controls. Two assay
167 runs were performed per day over 10 days with each sample being run in
168 duplicate, as unknowns, in each run. The lowest amount of the AMH in a sample
169 that can be detected with a 95% probability was calculated. The individual SD’s
170 of the five samples over 20 runs with ≤1.5 ng/mL were calculated and applied to
171 the formula below for the best estimate of LoD.
172 LoD= LoB (measured from sample with no analyte) + 1.647*SD’s.
SD’s sqrt[{(SD1* SD1)+ (SD2* SD2)+ (SD3* SD3)+ (SD4* SD4) + (SD5*
= SD5)}/5]
173
174 LIMIT OF QUANTITATION
175 Limit of quantification (LoQ) for AMH Gen II assay was determined as the lowest
176 concentration that can be measured with a total imprecision of ≤20%. The study
177 was designed following CLSI EP17 guidelines (20). Eight human serum samples
178 with AMH concentrations of 0.10, 0.16, 0.20, 0.57, 2.60, 3.40 and 14.3 ng/mL
179 were run with seven point calibration curve and controls. Two assay runs were
180 performed per day over 10 days and each sample run in duplicate, as unknowns.
181 The average, S.D. and % C.V. was calculated and AMH concentration (ng/mL)
182 vs. % C.V. curve was plotted. The estimated minimum dose achieved at 20%
183 total imprecision was then interpolated.
184
185 LINEARITY OF DILUTION
186 The study was designed following CLSI EP6 guidelines (21). Four samples with
187 AMH concentrations in the range of 5-13 ng/mL were diluted in the sample
188 diluent. Multiple dilutions of these samples were then assayed against the
189 calibration curve and the observed results were plotted against the expected
190 values. Expected values were calculated by dividing the concentration from the
191 undiluted sample by the dilution factor used. Percentage recovery was
192 calculated by dividing the observed values by the expected values and
193 multiplying by 100.
194
195 SPIKE RECOVERY
196 Recovery was determined by spiking known amounts of AMH into four human
197 serum samples containing different levels of endogenous AMH. The
198 concentration of AMH in the sample was determined before (endogenous) and
199 after (observed concentration) the addition of exogenous AMH. Expected
200 concentration was calculated as: Expected concentration = [(Endogenous
201 concentration x Volume of sample added) + (Spike concentration x Volume of
202 spike added)/ Volume of sample added + Volume of spike added]. The percent
203 recovery was calculated as: %Recovery= (Observed concentration/Expected
204 concentration)*100
205
206 IMPRECISION
207 The study was designed following CLSI EP5-A guidelines (22). Imprecision of
208 the assay was determined on two serum samples and the kit controls with AMH
209 concentration of 4.42, 14.03, 3.82 and 16.45 ng/mL, respectively. These
210 samples and controls were run in quadruplets in two assay runs per day per lot,
211 for 10 days over two lots. Precision was expressed as percent coefficient of
212 variation (%CV) for within run, between run and total assay (n=160) variability.
213
214 SELECTIVITY/SPECIFICITY
215 AMH in human serum samples and pooled newborn bovine samples were
216 assayed as unknowns to study the specificity of the assay for human and bovine
217 species. Structurally related proteins and some of the other members of the
218 TGF- super family were tested in the AMH Gen II at more than two times the
219 physiological concentrations to determine their cross-reactivity. Inhibin A (DSL,
220 Webster, USA, Cell line 1.203.2600), activin A (DSL, Webster, USA, Cell line
221 BA83.6.2), follicle-stimulating hormone (DSL, Webster, USA, DSL-4706) and
222 luteinizing hormone (Scripps, San Diego, CA, USA, cat # L0813). All the
223 substances were assayed against the standard curve as unknowns. Percent
224 cross-reactivity was calculated as shown in the equation below.
225 %Cross-reactivity = (Observed concentration/ Estimated concentration)*100
226
227
228 INTERFERENCE
229 The study was designed following CLSI EP7-P guidelines (23). Triglyceride
230 stock (1000 mg/mL, for preparation see supporting information) and Hemoglobin
231 from bovine blood (Sigma, St. Louis, MO, USA, cat # H 2500) were studied at 2
232 mg/mL and 20 mg/mL, respectively. Bilirubin (Sigma, St. Louis, MO, USA, cat %
233 B 4126), and human serum albumin, 99% fatty acid free, globulin free (Sigma, St.
234 Louis, MO, USA, cat # A3782), were studied at 0.6mg/mL and 60mg/mL,
235 respectively. Hemoglobin and triglyceride were dissolved in the kit matrix (27
236 mg/mL & 100 mg/mL, respectively) and spiked (15 uL and 30 uL, respectively)
237 into samples (135 uL and 120 uL, respectively). The control sample was spiked
238 with the equivalent amount of the same matrix, and both were assayed as
239 unknowns. Bilirubin was dissolved in solvent (14 mg/mL) and spiked (18.75 uL)
240 into the sample (131.25 uL), the control sample was spiked with the equivalent
241 amount of solvent, and both were assayed as unknowns. Sixty mgs of human
242 serum albumin powder was spiked directly to 1 mL of the reference samples (60
243 mg/mL) and assayed. The % difference to control was calculated for all
244 interfering substances using the equation below.
245
246 %Difference to Reference= (Spiked sample conc.- Conc. of reference sample) *100
247 __________________________________
248 Concentration of reference sample
249
250 METHOD COMPARISON
251 The study was designed following CLSI EP9-A guidelines (24). The Beckman
252 Coulter AMH Gen II assay (cat # A73818 & A73819) was evaluated against a
253 commercial AMH assay from Immunotech (cat # A16507) in our laboratory using
254 60 male and 60 female serum and Li-heparin plasma samples, ranging in age
255 from 20-50 years. The studies were carried out following the manufacturer’s
256 recommendations for preparation and storage of reagents, calibrators, controls
257 and for running the protocol. Serum aliquots were kept frozen, thawed once and
258 analyzed within three hours of thawing. Samples with %CV 15 were repeated
259 and recorded. Serum and Li-heparin plasma samples were compared between
260 AMH Gen II ELISA and IOT MIS/AMH EIA using Passing and Bablok method
261 conversion and Spearman Rank Correlation (Analyse-it®). One hundred and six
262 day-3 female serum samples were studied using Access 2 FSH (Beckman
263 Coulter, Fullerton, CA, USA, cat # 33520 and 33525) and AMH Gen II ELISA.
264 Spearman Rank Correlation between these methods was plotted.
265
266 STABILITY STUDIES
267 SAMPLE STABILITY/ FREEZE THAW STABILITY
268 Blood from 20 patients were drawn by venipuncture into Vaccutainers at the
269 vendor’s site and processed as 10 serums and 10 lithium heparin plasmas
270 (PromedDx, Norton, MA, USA, www.PromedDx.com) per manufacturer’s
271 instructions. The specimens were shipped overnight on cool packs to BCI. The
272 samples were aliquoted upon arrival and stored at 2-8ºC and -20ºC until use.
273 The samples were tested for stability following manufacturer’s instructions after
274 storage at 2-8ºC and -20ºC for seven days and three freeze thaws.
275
276 Results
277 CALIBRATION
278 AMH Gen II assay has been calibrated to IOT A16507 MIS/AMH EIA using
279 sample value transfer. AMH concentrations were assigned to 120 serum
280 samples using the IOT assay. These samples were then used to calibrate the
281 AMH Gen II assay and assign individual calibrators with AMH concentrations of
282 0.16, 0.4, 1.2, 4.0, 10 and 22.5 ng/mL prepared in bovine calf serum. The
283 samples were then re-run on the AMH Gen II kit resulting in a slope of 1.00x and
284 r=0.98, P< 0.0001.
285
286 LIMIT OF BLANK
287 The highest AMH concentration measured using one post-menopausal serum in
288 20 inter-assay runs at 95% CI was reported as zero ng/mL.
289
290 LIMIT OF DETECTION
291 Six serum samples in the range of 0.0-1.4 ng/mL were evaluated against the
292 AMH Gen II calibration curve over 20 inter-assay runs in duplicates (n=40). The
293 observed means for the five samples were 0.106, 0.189, 0.191, 0.548 and
294 1.457ng/mL; SD’s were 0.310, 0.026, 0.024, 0.043 and 0.084, respectively. The
295 lowest amount of the AMH in a sample that can be detected with a 95%
296 probability was calculated to be 0.08 ng/mL.
297
298 LIMIT OF QUANTITATION
299 Eight serum samples in the range of 0.10 - 15 ng/mL were evaluated against
300 AMH Gen II calibration curves over 10 days; 20 runs in duplicate (n=40).
301 The %CV of the seven samples were 29.5% at 0.110 pg/mL, 13.9% at 0.189
302 ng/mL, 12.5% at 0.191 ng/mL, 7.9% at 0.548 ng/mL, 7.4% at 2.438 ng/mL, 7.4%
303 at 3.660 ng/mL, 4.530 at 4.530 and 5.15% at 15.600 ng/mL. The estimated
304 minimum dose achieved at 20% CV with respect to AMH concentration was 1.16
305 ng/mL.
306
307 LINEARITY OF DILUTION
308 Serum samples in the range of 5 - 13 ng/mL AMH diluted in the sample diluent
309 showed linear results across the dynamic range of the assay (Figure 1); the
310 average recovery on dilution for the samples at 12.86, 10.34, 5.47 and 4.94
311 ng/mL were 95, 99, 104 and 101%, respectively.
312
313 SPIKE RECOVERY
314 Different volumes (7.5, 15 and 22.5 L) of a 22.5 ng/mL native AMH solution was
315 added to 150 L of four different serum samples containing endogenous AMH
316 concentrations in the range of 0.67-2.21 ng/mL. The AMH concentrations of the
317 samples before and after spiking exogenous AMH were recorded. The average
318 spiking recovery for individual samples with endogenous AMH concentrations of
319 0.67, 1.16, 2.21 and 1.47 ng/mL, respectively spiked with exogenous AMH doses
320 of 1.30, 2.49 and 3.57 ng/mL was 102, 106, 104 and 102%, respectively.
321
322 IMPRECISION
323 Reproducibility of the AMH Gen II assay was determined using two human serum
324 pools (QC1 and QC2) at 4.42 and 14.03 ng/mL and two kit controls (C1 and C2)
325 at 3.82 and 16.45 ng/mL using two reagent lots. The total imprecision calculated
326 on QC1, QC2 and C1, C2 over 40 assays (n=160) were 7.7 & 5.8 % and 5.7 &
327 5.3%, respectively.
328
329 SELECTIVITY/SPECIFICITY
330 The antibody specificity for human, monkey, mouse, rat, bovine and horse
331 samples has been published (18). Inhibin A (10,000 pg/mL), activin A (10,000
332 pg/mL), follicle-stimulating hormone (450 mIU/mL) and luteinizing hormone (100
333 IU/mL) were spiked individually in to the calibrator matrix (AMH, 0 ng/mL) and
334 assayed as unknowns. All were undetectable and reported as zero percent
335 cross-reactivity. Post-menopausal (PM) serum samples from 45 women with a
336 median age of 71 years were run in the AMH Gen II assay to check for false
337 positives. All 45 PM samples resulted in concentrations below the detection limit
338 of the assay.
339
340 INTERFERENCE
341 The percent (%) difference to reference samples when spiked with hemoglobin
342 from bovine blood (2 mg/mL), triglycerides (20 mg/mL), bilirubin (0.6mg/mL),
343 human serum albumin (60 mg/mL) individually were at 3.9, -6.9, 0.5 and 2.8 %,
344 respectively.
345
346 METHOD COMPARISON
347 The AMH Gen II assay was compared to IOT MIS/AMH EIA assay using
348 matched (60 male and 60 female ranging in age from 20-50 years) serum and
349 lithium heparin plasma samples. Both serum and plasma samples showed
350 significant positive linear correlations (Spearman Rank Correlation) to IOT assay
351 (r=0.98, P<0.0001 and r=0.98, P<0.0001), respectively. Linear regression
352 analysis (Passing Bablok method comparison) results observed were: AMH Gen
353 II serum =1.0 (IOT) serum (Figure 2) and AMH Gen II lithium heparin plasma =
354 0.97 (IOT) plasma, respectively (Figure not shown here). Linear regression
355 analysis of 120 matched serum and lithium heparin plasma specimens showed
356 AMH Gen II (serum) = 0.95 (AMH Gen II lithium heparin plasma) with a
357 correlation coefficient of r=0.99 and P<0.0001 (Figure 3). One hundred and six
358 female day-3 serum samples were assayed for FSH (Beckman Coulter Access)
359 and AMH ELISA. AMH Gen II ELISA and FSH resulted in correlations of r= -0.74,
360 P<0.0001 (Figure 4). Testing samples from thirty-five randomly selected
361 pediatric males between the ages of one and thirteen years of age resulted in
362 non-parametric AMH dose estimation at 90% limit of 3.8-159.8 ng/mL with a
363 median of 56.3 ng/mL. Thirty-five random samples from females between the
364 ages of one and thirteen years resulted in non-parametric AMH dose estimation
365 at 90% limit of 0-8.9 ng/mL with a median of 1.3 ng/mL. Twenty first-trimester
366 pregnancy samples evaluated for AMH concentration in serum resulted in a
367 range from 0.1 - 6.7 ng/mL. These maternal serums AMH concentration did not
368 correlate to the sex of fetus as shown in Figure 5. One hundred and seventy-six
369 female samples between the ages of 20 and 87 years studied for the AMH
370 concentration and age resulted in the range of 1.0 -14.8 ng/mL and showed a
371 decline of AMH concentration with age shown in Figure 6.
372
373 SAMPLE STABILITY
374 Aliquots of 10 serum and 10 lithium heparin plasma samples were stored at 2-8
375 ºC and at -20ºC for 1, 4 and 7 days and then analysed for AMH concentration
376 using comparative descriptive analysis (Analyse-it®). The observed median AMH
377 values for 10 serum and 10 lithium heparin plasma samples stored at 2-8°C on
378 days 0, 1, 4 and 7 were 3.39, 3.18, 3.28, 3.52 ng/mL and 4.23, 3.48, 4.32, 4.01
379 respectively. The average variation between fresh samples and those stored for
380 seven days is approximately 4%. The individual sample variations are shown in
381 figure 7 supporting information. The observed median AMH values for 10 serum
382 and 10 lithium heparin plasma fresh specimens (day 0) and aliquots stored at -
383 20°C day 1, 4 and 7 were 3.39, 3.45, 3.50, 3.74 ng/mL and 4.23, 4.29, 4.59, 4.28
384 respectively. The average variation between fresh samples and those stored for
385 seven days is <1%. Plasma specimens showed more variation than serum
386 samples under stress conditions. Freeze/thaw studies were conducted on 10
387 aliquots of serum and 10 aliquots of lithium heparin plasma samples describe
388 above. The median AMH values for the 10 serum and 10 lithium heparin plasma
389 samples for the three freeze thaw cycles were 3.45, 3.79, 3.89 ng/mL and 4.29,
390 4.55, 4.45, respectively. Results show an average variation of approximately 6%
391 between fresh and frozen serum/Li-hep plasma specimens. The individual
392 sample variation is shown in figure 8 supporting information.
393
394 Discussion
395 The AMH assays from Hudson et al (15) and Immunotech or Beckman Coulter
396 have been highly cited for the measurement of AMH in serum. Both assays have
397 provided significant clinical information related to various normal and disease
398 states. These AMH assays measure human AMH but cannot be used to measure
399 AMH in bovine or rodent samples, probably due to the variability of the amino
400 acid sequence in the pro-region of these species.The AMH Gen II assay has
401 been developed using two monoclonal antibodies that bind to the mature region
402 of the AMH molecule to measure AMH in mammals. In addition to possible
403 human clinical applications, the AMH Gen II assay will be useful for conducting
404 research in animal models. In addition, the AMH Gen II assay provides stable
405 and accurate AMH measurements unaffected by proteolysis of AMH.
406
407 Challenges can arise when different methods produce different results for the
408 same analyte. In an effort to minimize the variation between the measurements
409 of AMH methods, the AMH Gen II assay was standardized to the highly cited IOT
410 assay. The AMH Gen II assay has a significant positive linear correlations to IOT
411 assays for both serum and lithium heparin plasma samples (r=0.98, P<0.0001
412 and r=0.98, P<0.0001, respectively) and a slope of 1.0 for serum samples and a
413 slope of 0.97 for lithium heparin plasma samples. Since there is no WHO
414 preparation for AMH, a set of internal standards were prepared and stored for
415 future reference. Both serum and Lithium heparin plasma can be used as
416 sample type in the AMH Gen II assay as their linear regression analysis showed
417 a slope of 0.95 and a correlation of r=0.99 and P<0.0001. The AMH Gen II
418 assay has a two fold greater sensitivity (0.078 ng/mL) than IOT assay
419 (0.14ng/mL) and the cross-reactivity of Inhibin A, activin A, follicle-stimulating
420 hormone and luteinizing hormone were below the detection limit of the assay.
421 Human anti-species antibodies have been reported to be found in anywhere from
422 1 to 80% of a population (34). These antibodies can cause interference in
423 immunoassay, by generating false positive results. Heterophilic blockers have
424 been used in the assay to minimize the false positives. Individual AMH
425 concentrations of all the PM samples were below the detection limit of the assay.
426 Pediatric males and females were well- distinguished with a median AMH dose of
427 56.3 and 1.3ng/mL, respectively. Interference due to hemoglobin, triglyceride,
428 bilirubin and human serum albumin was within 7% of the reference AMH
429 measurement when the interferents were spiked with at least twice their
430 physiological concentrations. Human serum and lithium heparin plasma samples
431 serially diluted in sample diluent, showed good recovery. Similarly, a good
432 recovery was observed (102-106%) when exogenous AMH was spiked to human
433 samples. The total imprecision of the AMH Gen II assay was <8% across the
434 range of the assay. AMH in human serum was stable for at least seven days at
435 2-8ºC and -20°C. Lithium heparin plasma samples stored at 2-8ºC for seven
436 days showed average variation of approximately 10%. Serum and lithium
437 heparin plasma specimens in general showed a negative bias of approximately
438 5% at one thaw cycle and a positive bias of approximately 5% after two and three
439 thaw cycles. Sample stability studies with a statistically significant number of
440 samples are required to be conclusive. The performance of the AMH Gen II
441 assay and the stability of the serum and plasma AMH concentrations are highly
442 reproducible and compatible with complete automation or even large scale
443 manual sample analysis.
444
445 As expected, negative correlation (r= -0.74, P<0.0001) between AMH and FSH
446 was obtained on day-3 serum samples using the AMH Gen II assay. These
447 results suggest the utility of AMH in IVF clinics. AMH Concentrations in first-
448 trimester serum specimens from pregnant women were in the established range
449 of non-pregnant women and showed random distribution with respect to fetal sex.
450 The result suggests AMH serum concentrations in first trimester pregnant women
451 are not dependent on fetal sex. Decline in AMH concentration was observed
452 with female age. A steep decline was observed after 40 years of age and
453 showed undetectable AMH concentration in females over 50 years of age. The
454 decline in AMH concentration parallels the decline in follicle number with age.
455 Since AMH is produced during specific stages of follicular development rather
456 than follicular stimulation, AMH measurements combined with Inhibin B
457 measurements could serve as a direct marker of ovarian reserve. The two
458 biomarkers will receive additional visibility with the more robust and reliable
459 second generation assays.
460
461 In summary, we have described the development of second generation AMH
462 ELISA for the measurement of AMH in serum and Li-heparin plasma. The assay
463 is a significant improvement over the reported methodologies and is suitable for
464 automation and/or timely manual determination of AMH levels. The antibody pair
465 used in the assay is highly specific, measurement of AMH is unaffected by
466 proteolysis of the AMH molecule and is standardized to IOT assay. The
467 performance of the assay is ideal for investigation into the physiologic role of
468 AMH in both men and women.
469
470 Acknowledgements
471 We thank Beckman Coulter Inc. for the unconditional support to the project. We
472 also thank Dennis E. Smith, Susan L. Retka and Sheila Hanna for their scientific
473 contribution and help.
474
475 REFERENCES
476 1. Wilson C, Clemente N, Ehrenfels C, Pepinsky R, Josso N, Vigier B, Cate
477 R. Mullerian inhibiting substance requires its N-terminal domain for
478 maintenance of biological activity, a novel finding within the transforming
479 growth-factor-beta superfamily. Mol. Endocrinol. 1993; 7: 247-257.
480 2. Picard JY, Benaroust R, Guerrier D, Josso N, Khan A. Cloning and
481 expression of cDNA for anti-Mullerian hormone. Proc. Natl. Acad. Sci.
482 1986; 83: 5464-5468.
483 3. Di Clemente N, Wilson C, Faure E, Boussin L, Carmillo P, Tizard R,
484 Picard JY, Vigier B, Josso N, Cate R. Cloning, expression, and alternative
485 splicing of the receptor for anti-Mullerian hormone. Mol. Endocrinol. 1994;
486 8: 1006-1020.
487 4. Baarends WM, Uilenbroek JT, Kramer P, Hoogerbrugge JW, van
488 Leeuwen, EC, Themmen AP, Grootegoed JA. Anti-mullerian hormone and
489 anti-mullerian hormone type II receptor messenger ribonucleic acid
490 expression in rat ovaries during postnatal development, the estrous cycle,
491 and gonadotropin-induced follicle growth. Endocrinology, 1995; 136: 4951-
492 4962.
493 5. Lee M, Donahoe P. Mullerian inhibiting substance: a gonadal hormone
494 with multiple functions. Endocrine Review 1993; 142: 152-164.
495 6. Picon R. Action du testicule foetal sur le de- velopment in vitro des canaux
496 de Muller ches let rat. Arch. Anat. Micros. Morphol. Exp. 1969; 58:1-19.
497 7. Teixeira J. Maheswaran S, Donahoe PK. Mullerian inhibiting substance:
498 an instructive developmental hormone with diagnostic and possible
499 therapeutic applications. Endocrine Reviews 2001; 22: 657-674.
500 8. Feyereisen E, Mendez Lozano DH, Taieb J, Hester L, Frydman R,
501 Fanchin R. Anti-Mullerian hormone: clinical insights into a promising
502 biomarker of ovarian follicular status. Reprod. Biomed. Online 2006; 12(6):
503 695-703.
504 9. Hehenkamp WJ, Looman CW, Themmen AP, de Jong FH, TeVelde ER,
505 Broekmans FJ. Anti-Mullerian hormone levels in the spontaneous
506 menstrual cycle do not show substantial fluctuation. J Clin. Endocrinol
507 Metab. 2006; 91(10):4057-63.
508 10. LaMarca A, Stabile G, Artenisio AC, Volpe A. Serum anti-Mullerian
509 hormone throughout the human menstrual cycle. Human Reprod. 2006;
510 21(12):3103-7.
511 11. LaMarca A, Giulini S, Tirelli A, Bertucci E, Marsella T, Xella S, Volpe A.
512 Anti-Mullerian hormone measurement on any day of the menstrual cycle
513 strongly predicts ovarian response in assisted reproductive technology.
514 Human Reprod. 2007; 22(3):766-71.
515 12. Muttukrishna S, McGarrigle H, Wakim R, Khadum I, Ranieri DM, Serhal P.
516 Antral follicle count, anti mullerian hormone and inhibin B: predictors of
517 ovarian response in assisted reproductive technology? BJOG. 2005; 112
518 (10): 1384-1390.
519 13. Lee M, Misra M, Donahoe P, MacLaughlin D. MIS/AMH in the assessment
520 of cryptorchidism and intersex conditions. Molecular and Cellular
521 Endocrinology 2003; 211: 91-98.
522 14. Long WQ, Ranchin V, Pautier P, Belville C, Denizot P, Cailla H, Lhomme
523 C, Picard JY, Bidart JM, Rey R. Detection of minimal levels of serum anti
524 mullerian hormone during follow-up of patients with ovarian granulose cell
525 tumor by means of a highly sensitive enzyme-linked immunosorbent assay.
526 Journal of Clinical Endocrinology Metabolism 2000; 85:540-544.
527 15. Hudson P, Douglas I, Donahoe P, Cate R, Epstein J, Pepinsky RB,
528 MacLaughlin D. An immunoassay to detect human Mullerian inhibiting
529 substance in males and females during normal development. 1990; 70:16-
530 22.
531 16. Al-Qahtani A, Muttukrishna S, Appasamy M, Johns J, Cranfield M, Visser
532 JA, Themmen APN, Groome NP. Development of a sensitive enzyme
533 immunoassay for anti-Mullerian Hormone (AMH) and the evaluation of
534 potential clinical applications in males and females. Clin. Endocrinol. 2005;
535 63(3): 267-73.
536 17. Lee M, Donahoe P, Hasegawa T, Silverman B, Crist G, Best S, Hasegawa
537 Y, Noto R, Schoenfeld D, MacLaughlin D. Mullerian inhibiting substance in
538 human: normal levels from infancy to adulthood. Journal of Clinical
539 Endocrinology and Metabolism. 1996; 81:571-575.
540 18. Groome NP, Cranfield M, Themmen APN, Savjani GV, Mehta K.
541 Immunological assay and antibodies for anti-Mullerian hormone. US
542 patent 2006; 0275850A1
543 19. Harlow, E, Lane D. Immunoaffinity purification. Antibodies: a laboratory
544 manual. Cold Spring Harbour Laboratory Press, New York, 1988; 511.
545 20. Protocol for Determination of Limits of Detection and Limits of
546 Quantitation; Clinical and Laboratory Standards Institute: Approved
547 Guideline. EP17-A, V0l 24 No 34.
548 21. National Committee for Clinical Laboratory Standards. Evaluation of the
549 linearity of quantitative analytical methods; proposed guideline EP-6.
550 Wayne, PA: NCCLS, 1986
551 22. National Committee for Clinical Laboratory Standards. Evaluation of
552 precision performance of clinical chemistry devices; approved guideline
553 EP5-A. Wayne, PA: NCCLS, 1999
554 23. National Committee for Clinical Laboratory Standards. Interference
555 testing in clinical chemistry; proposed guideline EP7-P. Wayne, PA:
556 NCCLS, 1986
557 24. National Committee for Clinical Laboratory Standards. Method
558 comparison and bias estimation using patient samples; approved
559 guideline EP9-A. Wayne, PA: NCCLS, 1995
560
561 Tables Legend
562 Table-1 Experimental design to study AMH concentration towards the stability of
563 samples
564
565 Figures Legend
566 Figure-1: Linearity of dilution of serum samples in AMH Gen II assay
567 Figure-2: Passing Bablok regression analysis between AMH concentrations
568 (ng/mL) obtained with Immunotech AMH and AMH Gen II ELISA for 120 serum
569 samples. Linear regression analysis results were as follows: r= 0.99; p<0.0001;
570 y=1.0x
571 Figure-3: Passing Bablok regression analysis between AMH concentrations
572 (ng/mL) obtained with 120 matched serum and lithium heparin plasma
573 specimens in AMH Gen II ELISA. Linear regression analysis results were as
574 follows: r= 0.99; p<0.0001; y=0.95 x.
575 Figure-4: Spearman Rank correlation analysis between FSH concentrations
576 (mIU/mL) and AMH concentration (ng/mL) on one hundred and six female day
577 three serum samples using Beckman Coulter FSH and AMH Gen II ELISA.
578 Correlation observed was as follows: r= -0.74; p<0.0001.
579 Figure-5: Distribution of AMH concentration (ng/ml) in maternal serum during first
580 trimester pregnancy with male fetus vs. maternal serum with female fetus.
581 Figure-6: Distribution of serum AMH concentrations (ng/ml) with random female
582 maternal age.
583
584
585 Table -1
Day 0 Fresh 2-8ºC
Day 1 2-8ºC, -20ºC, 1F/T
Day 2 2-8ºC, -20ºC, 2F/T
Day 7 2-8ºC, -20ºC, 3F/T
586
587 Figure-1

14
Sample 1
13
12 Sample 2
Observed AMH in ng/mL

11 Sample 3
10 Sample 4
9
Absolute line of linearity
8
7
6
5
4
3
2
1
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Expected AMH in ng/mL
588
589 Figure-2:

25

20
AMH Gen II (Serum, ng/mL)

15

10

y = 1.0018x
0
0 5 10 15 20 25
Im m unotech AMH (Serum , ng/m L)
590
591 Figure-3:

AMH Gen II Li Hep Plasma (ng/mL) 25

20

15

10

y = 0.9506x + 2E-15
0
0 5 10 15 20 25
AMH Gen II Serum (ng/m L)
592
593
594
595 Figure-4:

18

13
AMH ng/mL

-2
0 10 20 30 40 50
FSH m IU/m L
596
597
598 Figure-5:

5
AMH (ng/mL)

0
(Females Fetus)
S1
Maternal Serum (Male Fetus)

599
600
601 Figure-6:

5
AMH (ng/mL)

0
20-25 26-30 31-35 36-40 41-45 46-50 51-83
Age (Years)

602
603
604
605 SUPPORTING INFORMATION
606
607 COATING OF MICROTITER PLATES
608 The F2B 12H/E monoclonal antibody used for coating was purified by Protein G
609 (Millipore, Billerica, Massachusetts, USA) affinity chromatography (19) at Bioserv,
610 Inc Ltd. UK. Microtiter plates from Greiner bio-one (Maybachstr. D-72636
611 Frickenhausen, Germany, cat: 705071) were coated with 100 L/well of 7.5
612 g/mL F2B 12H/E in 50mM sodium borate buffer, overnight at room temperature.
613 Excess antibody was removed, and then the plates were washed once using 300
614 L/well of 10 mM phosphate buffer saline (PBS) containing 0.01% Triton X-100
615 (Lab Chem Inc., Pittsburgh, PA, USA CAT# LC26280-1). The plates were
616 blocked with 200 L/well of 10 mM PBS containing 0.5% protease free BSA,
617 2.5% sucrose (Amresco, Solon, Ohio, USA, cat # E588, 0335, 0777,
618 respectively) and 0.05% proclin (Supelco, Bellefonte, PA, USA, cat # 16823-
619 0048) for 16-24 hrs at room temperature in 85% humidity. The blocking solution
620 was aspirated and the plates were dried at 34°C for 4-5 hrs. The plates were
621 then packed in foil pouches with desiccant, labeled and stored at 2-8°C.
622
623 BIOTINYLATION OF ANTI-AMH MONOCLONAL ANTIBODY.
624 Monoclonal antibody F2B/7A/A was purified by Protein G (Millipore, Billerica,
625 Massachusetts, USA) affinity chromatography (19) at Bioserv, Inc Ltd. UK,
626 dialyzed in 0.1 M sodium borate buffer, pH 8.8, biotinylated with 25 mmols of
627 NHS-LC-Biotin (Sigma, St. Louis, MO, USA, cat # B2643) to one mmol of
628 antibody and incubated at room temperature for 2 hours with gentle mixing. The
629 biotin was prepared immediately prior to use by adding 10 mg to 1.0 mL of
630 dimethylsulfoxide (Fisher ChemAlert guide, cat # B231-1). After 2 hours, the
631 reaction mixture was dialyzed against 1L of 10 mM PBS buffer five times at 2-8º
632 C to remove excess biotin.
633
634 CALIBRATION CURVE
635 The assay uses a seven point calibration (with blank subtracted). The log of
636 AMH concentration is plotted on the X-axis, the log of matched optical density on
637 the Y-axis, and the curve is fit using cubic regression (Opsys MR, Dynex
638 technologies, USA, Revelation Quicklink version 4.25). The AMH Gen II assay is
639 calibrated to IOT MIS/AMH EIA.
640
641 ASSAY PROCEDURE
642 AMH Gen II ELISA is an enzymatically amplified two-site immunoassay. In the
643 first step, 20 L of calibrators (seven vials, containing concentrations of
644 approximately 0, 0.16, 0.4, 1.2, 4.0, 10 and 22.5 ng/mL of AMH in bovine calf
645 serum), controls and unknown samples, and 100 L of assay buffer (protein-
646 based trizma-maleate buffer) were added to microtitration wells coated with anti-
647 AMH antibody. The wells were incubated with shaking on an orbital microplate
648 shaker (Lab-line Instrument Inc. USA) for one hour at room temperature (~25°C).
649 Then the plate was washed five times with a Bio-Rad model 1575 microplate
650 washer using phosphate buffer saline containing a nonionic detergent (wash
651 solution).
652
653 In the second step, 100 L of the AMH antibody-biotin conjugate (biotinylated
654 anti-AMH antibody in protein-based trizma-maleate buffer) was added to each
655 well and then incubated on an orbital shaker for one hour at room temperature.
656 The wells were then washed five times as described above. In the third step,
657 100 L of streptavidin labelled, horseradish peroxidase enzyme conjugates
658 (Prozyme, Hayward, CA, USA cat # CJ30H) in protein-based MES-TRIS buffer
659 was added to each well. The wells were incubated on an orbital shaker for an
660 additional 30 minutes at room temperature and washed five times as described
661 above. After the final wash step, 100 L of tetramethylbenzidine (TMB) substrate
662 solution (Neogen Corp., Lexington, KY, USA, K- Blue Aqueous TMB Substrate,
663 cat # 331199) was added to each well and then incubated on an orbital shaker
664 for 10-12 minutes. The color formation was stopped by addition of 100 L
665 stopping solution (0.2M H2SO4) to each well. The degree of enzymatic turnover
666 of the TMB was determined by dual wavelength absorbance measurement on a
667 Dynex Opsys MR microplate ELISA reader (Dynex Technologies Ltd, Virginia,
668 USA) at 450 nm as primary test filter and 630 nm as primary reference filter. The
669 absorbance measured was directly proportional to the concentration of AMH in
670 the samples. Calibrators were used to plot a log-log cubic regression calibration
671 curve of absorbance versus AMH concentration. The AMH concentrations in the
672 samples were then interpolated from the calibration curve.
673
674 INTERFERENCE
675 Triglyceride stock 1000 mg/mL was made by mixing 1000 mg Tricaprin (Sigma,
676 St. Louis, MO, USA, cat # T 7517), 1000 mg Tributyrin 98% (Aldrich, St. Louis,
677 MO, USA, cat # 113026), 1000 mg Triacetin, (Aldrich, St. Louis, MO, USA cat #
678 240885), 1000 mg Tricaproin, (Sigma, St. Louis, MO, USA, cat # T0888) and
679 1000 mg Glyceryl Trioctonoate (Aldrich, St. Louis, MO, USA cat # T9126 )
680 together.
681
682
683 Figure-7: Change in AMH concentrations on individual serum and plasma sample
684 studied for stability at 2-8°C for seven days. Serum samples are labeled as
685 sample 1-10 and plasma samples are labeled as sample 11-20.
686

14 Fresh 2-8ºC-Day 1 2-8ºC-Day 4 2-8ºC-Day 7

12

10
AMH (ng/mL)

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Sample number
687
688
689
 690 Figure-8: Change in AMH concentrations on individual serum and plasma sample
691 studied for freeze thaw stability for up to four thaw cycles. Serum samples are
692 labeled as sample 1-10 and plasma samples are labeled as sample 11-20.
693

15 Fresh 1 F/T 2 F/T 3 F/T

13

11
AMH (ng/mL)

-1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Sample
694

View publication stats