Keywords
teriparatide -
in vitro bioassay - cAMP - time-resolved fluoroimmunoassay - method verification
Introduction
Teriparatide (PTH1–34) was the first anabolic drug to be approved by the U.S. Food
and Drug Administration (FDA) for bone reconstruction.[1]
[2] Following the expiration of the patents on the original drug, many generics were
extensively developed and have received approval worldwide. Bioactivity is a very
important attribute of polypeptide drugs, including PTH1–34. Therefore, establishing
a sensitive, efficient, and stable method to demonstrate the similarity of biosimilars
or generic drugs with pharmacopeia standards and reference medicinal products is required.
The current edition of the United States Pharmacopeia (USP, 2021) includes an in vitro cell-based biological activity assay to evaluate the potency of PTH1–34. According
to this pharmacopeia, this method requires the seeding of samples 24 hours in advance,
a 24-hour starvation period, and drug detection, which takes 2 to 3 hours, which is
a long time. The required test reagents must also be ordered 1 year in advance, which
is expensive and unsuitable for high-throughput testing. Because of the high cost
of this testing method, challenges in sourcing the reagents, and the significant time
and cost required for testing, there is an urgent need to improve the in vitro activity testing methodology.
In this study, a novel cell-based in vitro assay was established to measure the activity of PTH1–34. The N-terminus of PTH1–34
binds to the near-membrane structure of PTH1R and activates G protein signaling.[3]
[4]
[5] When UMR-106 cells are coincubated with PTH1–34, cyclic adenosine monophosphate
(cAMP) signals are generated through the G protein-coupled receptor signaling pathway.[5]
[6] In the present study, UMR-106 cells were coincubated with PTH1–34, and a time-resolved
fluoroimmunoassay was used to quantify intracellular cAMP levels to assess the activity
of PTH1–34.[7] The principle is based on the fact that the labeled cAMP tracer competes with the
PTH1–34-induced cAMP in cells through binding to labeled cAMP antibodies. PTH1–34-induced
cAMP levels can be determined by measuring the concentration of the labeled cAMP when
the labeled antibody binds to the labeled cAMP. We comprehensively validated this
method and demonstrated its feasibility as a high-throughput assay for PTH1–34 biological
activity.
Materials and Methods
Instruments and Reagents
The instruments used in this study were an electric constant temperature sink (model
HH-420, Changzhou Zhiborui Instrument Manufacturing Co., Ltd., Changzhou, China),
ultrapure water meter (model Millipore/Milli-Q Advantage A10, Merck, Germany), CO2 incubator (model CCL-1708-8, Esco, Singapore), inverted microscope (model CKX53,
Olympus, Japan), type A2 biosafety cabinet (model AC2-4S1, Esco, Singapore), benchtop
centrifuge (model TD6, Lu Xiangyi, Shanghai, China), microplate thermostatic oscillator
(model BE-9008, Haimen Qilinbell Instrument, HaiMen, China), multifunctional microplate
reader (model SpectraMax i3X, MD, United States), and electronic balance (model SQP,
Sartorius, Germany).
The materials used in this study were Dulbecco's Modified Eagle Medium (DMEM) (lot
number 12800-017, Gibco, United States), penicillin-streptomycin (lot number 15140-122,
Gibco, United States), bovine albumin (batch number 69003435, Sinopharm, Shanghai,
China), fetal bovine serum/FBS (batch number 10099-141, Gibco, United States), trypan
blue (lot number 2101673, Gibco, United States), LANCE Ultra cAMP Kit (lot number
TRF0263, PerkinElmer, United States), white 384-well plate (lot number 6008280, PerkinElmer,
United States), and recombinant teriparatide (lot number 410019-201801, National Institutes
for Food and Drug Control, China). PTH1–34 API, PTH1–34 preparation, PTH1–30 impurity,
and adrenocorticotropic hormone (ACTH) were purchased from Shanghai Duomirui Biotechnology
Ltd., (Shanghai, China). The UMR-106 cell line was obtained from the Chinese Academy
of Sciences Cell Bank (lot number TCR11, Shanghai, China).
Reagent Preparation
The analysis media were prepared under sterile conditions. DMEM basal medium (99 mL,
containing 3.7 g/L sodium bicarbonate) was supplemented with bovine serum albumin
(0.1 g) and penicillin-streptomycin (1 mL). The PTH1–34 reference and samples were
diluted to specific concentrations using this media.
Bioactivity Assay
UMR-106 cells were cultured in DMEM medium with 10% FBS and 1% penicillin–streptomycin
in the presence of 3.7 g/L sodium bicarbonate at 37°C in a 5% CO2 atmosphere.
Viable cells were stained with trypan blue, and counted under an inverted microscope.
A total of 103 cells/well in 5 µL of analysis medium were added to each well of a 384-well plate.
The PTH1–34 reference and samples were at an initial concentration of 4,000 ng/mL
and 3-fold serial dilutions were prepared in the analysis medium. Then, 5 µL of serially
diluted PTH1–34 reference and the samples were added to each well. After a 30-minute
incubation period at 25°C in the dark and the addition of 10 µL of LANCE Ultra cAMP
reagent, the samples were incubated for 60 minutes at 25°C in the dark. The 384-well
plate was placed on an MD microplate reader for TR-FRET detection and the detection
procedure is shown in [Table 1].
Table 1
SpectraMax i3X program settings
|
Detect items
|
Set parameters
|
|
Cartridges, read modes, read type
|
HTRF, TR-FRET, endpoint
|
|
Wavelengths
|
Excitation: 340 nm
Emission: 615/665 nm
|
|
PMT and optics
|
Integration time: 0.2 ms
Excitation time: 0.05 ms
Number of pulses: 30
Measurement delay: 0.02 ms
Read from top
Read height: 8.25 mm
|
|
Shake
|
No shake
|
|
Read order
|
Row
|
|
Show optimizer
|
On
|
Abbreviation: HTRF, homogeneous time-resolved fluorescence; TR-FRET, time-resolved
fluorescence resonance energy transfer.
Method Optimization
Dilution Rate
Using PTH1–34 API at 4,000 ng/mL as the initial concentration, 2-fold (1:1), 3-fold
(1:2), and 5-fold (1:4) gradient dilutions were prepared (9 gradients each) and the
potency levels of PTH1–34 was detected according to “Bioactivity Assay.”
Cell Seeding Density
Using PTH1–34 API at 4,000 ng/mL as the initial concentration and a 3-fold dilution,
the cell density was adjusted to obtain 250, 500, 1,000, and 2,000 cells/well and
the potency levels of PTH1–34 was detected according to “Bioactivity Assay.”
Method Verification
Specificity
Recombinant teriparatide, PTH1–34 API, PTH1–34 preparation, PTH1–30 impurity, ACTH,
and the prepared blank excipient were diluted to 4,000 ng/mL as an initial concentration,
and a 3-fold gradient dilution was established and the potency levels of PTH1–34 was
detected according to “Bioactivity Assay.”
Relative Accuracy, Intermediate Precision, Linearity, and Range
PTH1–34 API (10 mg) was diluted to 10 mg/mL with analysis medium (a total of six parts
were diluted, one of which was used as a standard solution and the other five as a
test solution), and diluted to 4,000 ng/mL as the standard solution in analysis medium
and the mass concentration of 2,000, 2,800, 4,000, 5,200, and 6,000 ng/mL were established
as test solutions. The relative potency levels of five analyte solutions were 50,
70, 100, 130, and 150%, respectively, and the activity of PTH1–34 was detected according
to “Bioactivity Assay.” Each potency level was assessed by two analysts using three
cell generations on three different dates, and the data are presented as measured
potency. Measured potency = R/X × 100%, where R is the EC50 value of the standard solution and X is the EC50 value of the test solution.
Robustness
UMR-106 cells were cultured in DMEM medium containing 10% FBS and 1% penicillin–streptomycin
in the presence of 3.7 g/L sodium bicarbonate at 37°C in a 5% CO2 atmosphere and passaged 1:1 every other day. The cells were passaged to the 4th,
11th, and 25th generations, and the effect of each cell generation on the 100% potency
level was determined according to the “Bioactivity Assay.”
Data Analysis and Statistics
All statistical analyses were performed using SoftMaxPro version 7.1.2 and GraphPad
Prism 9.5 software. Taking the sample concentration as the x-axis and the corresponding
fluorescence ratio as the y-axis, a four-parameter equation was selected for fitting,
and the dose–response curves of the active standard and the test product were generated
to obtain the concentration for 50% of the maximal effect (EC50) value, which was calculated by the following formula: fluorescence ratio = EM665 nm/EM615 nm; EM, emission.
Results
Dilution Ratio
The bioassay conditions were optimized by evaluating two parameters: dilution ratio
and cell density. The first optimized parameter was the dilution ratio, where the
fluorescence ratios were measured for different concentrations of PTH1–34 API with
a dilution ratio of 1:1, 1:2, and 1:4, respectively. As shown in [Fig. 1], the data point distribution of the four-parameter curve obtained under the 1:1
dilution ratio was not good and not considered, and the data point distribution of
the four-parameter curve obtained under the 1:2 dilution ratio was more uniform than
that obtained at 1:4 dilution ratio. The logarithmic period data points were greater
than those of the 1:4 dilution ratio, which can truly reflect the data fitting situation.
The calculated results were more reliable; thus, the optimal dilution ratio for this
method was 1:2.
Fig. 1 Optimization of the dilution ratio (n = 3).
Cell Density
The second optimized parameter was cell density, where the fluorescence ratios were
measured for different concentrations of PTH1–34 API using cell densities of 250,
500, 1,000, and 2,000 cells/well, respectively. As shown in [Fig. 2], the four cell densities yielded a typical S-type curve, and the R squared (R
2) values were 0.867, 0.945, 0.984, and 0.979, respectively. The window of the four-parameter
fitting curve for 250 and 500 cells/well was very narrow. The R
2 values for 1,000 and 2,000 cells/well were very close, and the window range of the
four-parameter fitting curve was close; however, the 1,000 cells/well density dropped
more in the inflection point value of the four-parameter fitting curve. Thus, a cell
density of 1,000 cells/well was considered optimal.
Fig. 2 Optimization of the cell density (n = 3).
Method Verification
The optimized conditions were applied for further validation according to the 2020
edition of the Chinese Pharmacopeia (ChP). The validation characteristics are described
below.
Specificity
Specificity refers to the ability of an assay to correctly measure an analyte without
interference from other components, such as impurities, degradation products, and
matrices. We selected recombinant teriparatide, PTH1–34 API, PTH1–34 preparations,
PTH1–30 impurities, ACTH, preparation blank excipients, and drug dilution solution
(analysis medium) for the treatment of cells to study the specificity of the method.
As shown in [Fig. 3], the same concentration of PTH1–30 impurity, ACTH, PTH1–34 preparation control (blank
excipients), and drug dilution solution could not stimulate the production of cAMP
in UMR-106 cells; however, the same concentration of recombinant teriparatide, PTH1–34
API, and PTH1–34 preparation stimulated the production of cAMP in UMR-106 cells. They
yielded a good dose–response relationship, indicating that the biological activity
detection method of PTH1–34 was unique.
Fig. 3 Specificity curve (n = 2).
Relative Accuracy, Intermediate Precision, Linearity, and Range
The method was evaluated for relative accuracy, intermediate precision, linearity,
and range, and the four indices were verified as a combined design. The potency was
measured using five levels (50, 70, 100, 130, and 150%) of prediluted initial working
concentrations of PTH1–34 API (i.e., 4,000 ng/mL). Each potency level was determined
by two analysts using three cell generations on three different days. The results
of six different assays are shown in [Table 2]. The results of the following four validation indexes are based on the results of
these six measurements.
Table 2
Results for six different assays
|
Expected potency (%)
|
Measured potency
|
|
Analyst 1
|
Analyst 2
|
|
Day 1 (%)
|
Day 2 (%)
|
Day 3 (%)
|
Day 1 (%)
|
Day 2 (%)
|
Day 3 (%)
|
|
150
|
148.1
|
147.2
|
149.2
|
152.0
|
148.2
|
152.6
|
|
130
|
133.2
|
126.4
|
125.5
|
133.1
|
130.9
|
132.0
|
|
100
|
98.1
|
99.8
|
102.3
|
99.7
|
102.2
|
97.8
|
|
70
|
67.3
|
70.5
|
73.5
|
70.2
|
66.7
|
68.2
|
|
50
|
47.7
|
50.7
|
55.8
|
52.2
|
50.6
|
47.2
|
Linearity and Range
To obtain a linear relationship between the measured relative valence and the true
value or reference value within the design range, a linear regression equation was
generated from the results of the six biological activities measured in [Table 2]. As shown in [Fig. 4], the regression equation (y = 0.9951x + 0.4542, R
2 = 0.9953) exhibited a linear correlation.
Fig. 4 Linearity and range (n = 6).
Relative Accuracy
Relative accuracy refers to the degree to which the measured relative potency is close
to the true value or reference value within a specified range, generally using relative
bias (RB, %) for assessment. In this work, the RBs and confidence intervals (CIs)
at each potency level were calculated according to the ChP 2020 General Rule 9401
([Table 3]). The results showed that the average RB of the relative potency measurements at
the five potency levels ranged from −0.8 to 1.4%, and the RB was within the range
of ± 12%, which complies with the pharmacopeia standard.
Table 3
Relative bias and confidence intervals for relative potency measures at different
potency levels
|
Theoretical potency level (%)
|
Number of trials
|
Logarithmic potency
|
Measured potency
|
RB
|
|
Average value
|
LCI
|
UCI
|
Average value (%)
|
LCI (%)
|
UCI (%)
|
Average value (%)
|
LCI (%)
|
UCI (%)
|
|
150
|
6
|
0.402
|
0.390
|
0.415
|
149.6
|
147.7
|
151.4
|
−0.3
|
−1.5
|
0.9
|
|
130
|
6
|
0.263
|
0.242
|
0.285
|
130.2
|
127.4
|
133.0
|
0.1
|
−2.0
|
2.3
|
|
100
|
6
|
0
|
−0.016
|
0.016
|
100.0
|
98.4
|
101.6
|
0
|
−1.6
|
1.6
|
|
70
|
6
|
−0.366
|
−0.395
|
−0.336
|
69.4
|
67.3
|
71.5
|
−0.8
|
−3.8
|
2.1
|
|
50
|
6
|
−0.681
|
−0.732
|
−0.630
|
50.7
|
48.1
|
53.3
|
1.4
|
−3.8
|
6.5
|
Abbreviations: CI, confidence interval; LCI, lower confidence limit; RB, relative
bias; SD, standard deviation; UCI, upper confidence limit.
Notes: Logarithmic potency = Ln (measured potency), the base for logarithmic conversion
in the table is e.
Relative bias = (measured potency/theoretical potency level − 1) ×100%.
CI = average ± tdf × SD/n
1/2; LCI = average − tdf × SD/n
1/2; UCI = average + tdf ×SD/n
1/2, where the average is the logarithmic mean of the potency values for each potency
level, SD is the logarithmic standard deviation of the potency determination at each
potency level, n is the number of potency measurements for each potency level, tdf is the value of the t-test table when the degree of freedom is df, and df is the degree of freedom and
equal to the number of measured values for each potency level minus 1.
Intermediate Precision
Intermedia precision refers to the proximity between the results of the same homogeneous
test sample measured by multiple samples under the specified conditions. Based on
the ChP 2020 General Rule 9401, the geometric standard deviation and geometric coefficient
of variation (GCV) of the five relative potencies were measured for each potency level.
The intermediate precision was evaluated using the GCV for each potency level. The
results are listed in [Table 4]. Our data showed that the GCV range was 2.0 to 3.5% and the GCV at the five potency
levels was less than 25%. The relative accuracy, intermediate precision, linearity,
and range verification results of the five potency levels indicated that the test
results of this method meet the requirements of General Rule 9401 in the range of
50 to 150%. Thus, the relative potency of the method was 50 to 150%.
Table 4
Geometric standard deviation and geometric coefficient of variation of relative potency
measured at different potency levels
|
Potency level (%)
|
Number of trials
|
GSD
|
GCV (%)
|
|
150
|
6
|
1.0227
|
2.3
|
|
130
|
6
|
1.0347
|
3.5
|
|
100
|
6
|
1.0197
|
2.0
|
|
70
|
6
|
1.0254
|
2.5
|
|
50
|
6
|
1.0321
|
3.2
|
Abbreviations: GCV, geometric coefficient of variation; GSD, geometric standard deviation.
Note: GSD = antilog (SD), where SD is the logarithmic standard deviation of the potency
determination at each potency level; GCV = (GSD − 1) ×100%.
Robustness
As shown in [Fig. 5] and [Table 5], the S-type dose–response curves could be obtained for all three generations of
cells, with actual potency measurements of 114.2, 97.3, and 77.5%, respectively. Fluctuations
in the range of 70 to 130% indicated that they were not affected and that the potency
values of the three generations were very close. Given the above, the cell generation
at the 4th to 25th passage had no significant effect on the measured biological activity results.
Fig. 5 Cell generation robustness (n = 2). (A) 4th generation. (B) 11th generation. (C) 25th generation.
Discussion
PTH1–34 is an anabolic (bone-forming) agent that is effective in postmenopausal women
with osteoporosis and patients with glucocorticoid-induced osteoporosis.[8] Abaloparatide, engineered from PTHrP (1–34), was approved by the FDA in April 2017
for the treatment of osteoporosis in postmenopausal women.[9] Abaloparatide binds to PTH1R. As a prominent new drug, PTH1R-related targets have
again captured the interest of researchers developing potent and safer drugs to promote
bone health.[10] Although physical and chemical analysis of macromolecular drugs have become more
advanced over time, even if they are detailed, they cannot always be determined from
their high-order structures, they need to be inferred from the biological activity
of the product. This indicates the importance of biological activity analyses. Regulatory
agencies in various countries expect biological activity analyses to be based on the
mechanism of action of the drugs; thus, cell-based biological activity methods are
widely used for the release and stability analysis of macromolecule drugs.[11]
In the USP rhPTH1–34 biological activity detection method, cells should be pretreated
for approximately 48 hours, plated 24 hours in advance, and starved for 24 hours.
This lengthy precell treatment may affect the cell state and the working solutions,
which must be transferred twice. This undoubtedly increases experimental error and
reduces the accuracy of the resulting data.[12] In the present study, UMR-106 cells directly interacted with rhPTH1–34 for 30 minutes
in a one-step method, which greatly reduced cell preparation time and improved the
efficiency of the assay.
Reporter gene methods have been widely used for cell viability assays in recent years.[13] Reporter gene assays, which are the mechanism of action related, less variable,
accurate, precise, and time-saving, are becoming increasingly recognized and adopted
for quality control.[14] Therefore, plasmids containing luciferase and overexpressing the PTHR1 gene may
be used to transfer HEK293 and other cells to reduce the experimental period and improve
the stability of methods through a reaction with luciferin substrates. The method
described here will accelerate the progress of current drug development.
Conclusion
In this study, drug dilution gradient and cell density were optimized. Moreover, the
methodology was verified according to the USP and the Pharmacopeia of the People's
Republic of China biological activity guidelines. The results indicate that the assay
exhibits good precision, accuracy, specificity, and good linear relationship in a
range of 50 to 150%. The results of passage stability studies show that the cells
exhibited a good dose–response curve when passed up to 25 generations.
Table 5
100% potency results for different cell generations
|
Generation
|
Theoretical potency (%)
|
Actual potency (%)
|
Average potency (%)
|
|
4th
|
100
|
114.2
|
96.3
|
|
11th
|
97.3
|
|
25th
|
77.5
|
