A hybrid IA-LC-MS/MS method for adrenocorticotropic hormone(1–24) to support interpretation of low-dose cosyntropin-stimulation test

ABSTRACT

Adrenocorticotropic hormone 1–24 (ACTH[1–24]) has a similar effect as endogenous ACTH(1–39) to generate cortisol by targeting the MC2R receptor on the adrenal gland. A new investigational ACTH receptor antagonist drug is being developed to treat diseases of ACTH excess (e.g., Cushing's disease) by binding to the MC2R receptor. Administration of ACTH(1–24) was used in a Phase I clinical study to assess the ability of this drug candidate to suppress the cortisol response to ACTH stimulation. A hybrid immunoaffinity-LCMS assay measuring ACTH(1–24) with a concentration range of 10 to 400 pg/ml was developed to support the study. Consistent and acceptable A&P results were achieved. The assay development and qualification will be discussed.

Plain language summary

Article highlights.

Background

• Mechanistic PK/PD modeling of hypothalamic–pituitary–adrenocortical axis with an adrenocorticotropic hormone (ACTH) competitive antagonist drug benefits from distinguishing endogenous ACTH(1–39) and exogenous ACTH(1–24) in the context of ACTH stimulation test.

• Immunoassays able to measure ACTH(1–24) without cross-reactivity to ACTH(1–39) were not found.

• A highly sensitive method was needed due to the low-dose ACTH(1–24) administration and rapid clearance from circulation.

Experimental

• A fit-for-purpose, bead-based, hybrid immunoaffinity-LC-MS/MS method was developed for intact measurement of ACTH(1–24) in human plasma.

• Pre-conjugation of capture antibody to magnetic beads provides tolerance to high levels of endogenous biotin, a problem reported for clinical ACTH(1–39) assays.

• A surrogate matrix calibration strategy is employed, which enables potential future multiplexing of endogenous ACTH counterparts.

Results & discussion

• Suitable sensitivity was achieved with an LLOQ of 10 pg/ml.

• Quantitation of ACTH(1–24) was accurate and precise in the presence of physiological levels of ACTH(1–39).

• The stability of ACTH(1–24) needs further investigation.

Conclusion

• The assay is considered suitable for measurement of ACTH(1–24) in human plasma samples from the clinical study.

ACTH(1–39): Adrenocorticotropic hormone (ACTH) is a 39 amino acid peptide hormone produced by the pituitary gland. It regulates adrenocortical cortisol, aldosterone, and androgen production through its interaction with melanocortin 2 receptor (MC2R) on the adrenal glands.

ACTH(1–24): Also known as cosyntropin or synacthen, ACTH(1–24) is a synthetic version of adrenocorticotropic hormone containing the first 24 amino acids of the full length endogenous human version. It is used in the cosyntropin-stimulation test (CST).

Bead conjugation: The process of affixing an affinity reagent (antibody, aptamer, etc.) to beads, in this case magnetic beads. There are many ways to approach this conjugation, including via direct chemical conjugation and biotin-streptavidin interaction. For biotin-streptavidin approach, the conjugation can occur either before incubation with sample, or afterward.

Cosyntropin-stimulation test A dynamic endocrine test that evaluates adrenocortical function. A synthetic form of ACTH (cosyntropin, synacthen, ACTH[1–24]) is administered intramuscularly at baseline followed by sequential cortisol measurements. The CST is also known as the ACTH-stimulation test and short Synacthen test.

Fit-for-purpose: Validation of an analytical method where the degree of assay evaluation is dependent on the intended use of the generated data. The degree of rigor applied is driven by scientific judgement, rather than regulatory guidance.

Hybrid IA-LC-MS: Hybrid immunoaffinity (IA)-liquid chromatography/mass spectrometry combines affinity capture with liquid chromatography separation and mass spectrometry detection. As ‘affinity capture’ is analogous to immunoassays, a ‘hybrid’ platform relates to the contribution from both LC-MS and ligand-binding assays.

1. Background

Adrenocorticotropic hormone (ACTH) is a 39 amino acid peptide belonging to the group of neuropeptide hormones (melanocortins) that are ligands of melanocortin receptors. All melanocortins are produced by proteolysis of the precursor proopiomelanocortin (POMC). ACTH is produced by corticotropic cells in the anterior pituitary gland in response to stress, and it is the primary regulator of adrenal cortisol production via interaction with the melanocortin type 2 receptor (MC2R) [1]. Measurement of plasma ACTH is an important tool for diagnosis and subtyping of adrenal insufficiency and Cushing's syndrome [2].

Leveraging the specificity of the ACTH interaction with MC2R combined with availability of quality assays for quantitation of cortisol in human blood products allows for interrogation of adrenal function. Broader assessments of the integrated hypothalamic–pituitary–adrenocortical (HPA) axis can be evaluated by dynamic endocrine testing. For example, the insulin tolerance test (ITT) was developed in the 1960s to elicit the hormonal counterregulatory response to insulin-induced hypoglycemia that results in activation of the hypothalamic (corticotropin-releasing hormone [CRH])-pituitary (ACTH) and adrenal (cortisol) axis [3,4]. Other dynamic tests of the HPA axis include the overnight single-dose metyrapone test [5] and CRH stimulation test [6]. Alternatively, the cosyntropin-stimulation test (CST) is frequently utilized in clinical practice to assess the adrenocortical hormonal response to ACTH stimulation [7]. Here, an injection of synthetic ACTH (ACTH[1–24], cosyntropin) is administered and blood-timed cortisol levels are measured in response to this challenge, as ACTH binding to the MC2 receptor at the adrenal gland stimulates a steroid cascade to produce cortisol, aldosterone and androgens. The CST, or ACTH stimulation test, thus interrogates adrenal function without the discomfort of ITT-induced hypoglycemia.

The diagnostic accuracy of the ACTH stimulation test in the evaluation of adrenal insufficiency is inferior to the ITT which is considered the gold standard reference. Cunningham et al. [8] concluded that the integrity of the HPA axis is best evaluated via ITT or the metyrapone test, rather than the CST, when excluding secondary adrenal insufficiency of pituitary etiology. In addition, the typical (high-dose) CST test utilizes 250 μg of ACTH(1–24) resulting in supraphysiological levels of ACTH that clearly exceed values reached in the insulin tolerance test. Additional research revealed that a low-dose cosyntropin stimulation test, at 1 μg dose, is sufficient to reach physiologically relevant levels of ACTH and therefore a more sensitive test of the adrenal function compared with higher ACTH(1–24) doses [9–11].

In the drug-development space, the CST offers a useful tool to evaluate dynamic endocrine function in diseases that present with excess production of ACTH (e.g., Cushing's disease, congenital adrenal hyperplasia). Thus, as an example, the test can be used to demonstrate the proof of concept of an experimental oral ACTH antagonist in normal healthy subjects. In addition to the cortisol readout, we also desired a measurement of the in vivo ACTH(1–24) concentration as well as, but separately measured from endogenous human ACTH(1–39). The ACTH immunoassay measurement utilized in our work, not reported here, was selective to ACTH(1–39) with negligible cross-reactivity to ACTH(1–24) observed. We did not identify an immunoassay for ACTH(1–24) without cross-reactivity to ACTH(1–39). In work reported by Darmon et al. to probe the kinetics of the low-dose stimulation test, the authors used a radioimmunoassay employing a polyclonal antibody with some cross-reactivity to ACTH(1–39) for their measurement of ACTH(1–24) [10]. We desired a similar assessment of ACTH dynamics as part of this test but with improved specificity of the separately measured peptides in the context of evaluating the influence of our candidate drug on the biological pathway.

Because of the low-dose and the short half-life of ACTH(1–24) we required a highly sensitive assay for measurement of circulating concentrations at practical timepoints for the CST. With sensitivity improvements in LC-MS technologies as well as the incorporation of antibody-based purification strategies with mass spectrometric detection for endogenous proteins and peptides [12], we have opportunity to improve the selectivity of these peptide measurements in an otherwise immunoanalytical space that is complicated by cross-reactivity of the critical reagents to similar circulating species. This capability has been demonstrated by Shi et al. where a hybrid immunoaffinity-LC-MS/MS method was developed for ACTH(1–39) to address discordant results obtained by two different clinical ACTH immunoassays [13]. In this study we describe the development and fit-for-purpose validation of a hybrid IA-LC-MS/MS assay, modeled after the Shi et al. format, to selectively measure ACTH(1–24) in human plasma in the presence of endogenous human ACTH(1–39).

2. Experimental

2.1. Materials & reagents

Synthetic peptides ACTH(1–24) [4011958] and the internal standard murine ACTH(1–39) [4030325] were purchased from Bachem. The capture antibody and streptavidin coated magnetic beads used for sample purification were obtained from the Elecsys ACTH Kit, Roche Diagnostics. Human ACTH(1–39) synthetic peptide was purchased from Tocris.

The following materials were obtained from the indicated commercial sources: HPLC-grade deionized water (Fisher), LC-MS-grade acetonitrile (Honeywell), tris-buffered saline (VWR) or equivalent, >88% purity ACS Reagent Grade formic acid (Fisher), Triton X-100 (VWR), ACS Reagent Grade Isopropanol (VWR), hydrochloric acid (Sigma Aldrich), sodium chloride (Sigma Aldrich), potassium chloride (Sigma Aldrich). PBS (1x) (Corning), Protein Lo-Bind tubes and plates (Eppendorf), Agilent BioTek Microplate Washer Flat Magnets (Fisher Scientific) and plasma derived from K₂EDTA human whole blood was obtained from BioIVT.

2.2. Calibration & QC samples

A surrogate matrix calibration curve strategy was adopted for this assay. The surrogate matrix was prepared by mixing 250 ul of Triton X-100 with 500 ml of phosphate-buffered saline (PBS) (1x, pH 7.4) and stored at room temperature for up to 3 months. The calibration standards were prepared in surrogate matrix from ACTH(1–24) spiking solutions in 5 ml Protein Lo-bind tubes at six concentrations spanning 5–400 pg/ml. Quality control (QC) samples were also prepared both in surrogate matrix and plasma at 10, 30, 80 and 300 pg/ml. Prepared calibration standards and QCs were stored at -70°C as single use subaliquots.

2.3. Bead conjugation procedure

A total of 1.6 ml of biotinylated anti-ACTH mouse antibody (concentration: 0.3 mg/l) from the Elecsys ACTH kit and 1.3 ml of streptavidin-coated magnetic microparticles (concentration: 0.72 mg/ml) were combined in a polypropylene tube. The resulting mixture was subjected to incubation at 4°C with rotation mixing for a minimum of 2 h. Subsequently, a magnet (Fisher Scientific) was employed to immobilize the magnetic beads to one side of the tube, allowing for the removal of the supernatant. This step was followed by the addition of 2.9 ml of Tris buffer saline (TBS) and vortex mixing. The washing process was repeated twice more. After the completion of the wash procedure, the isolated magnetic bead complex was reconstituted with 1 ml of 0.05% Triton-X100 in PBS solution and stored at 4°C until ready for the sample preparation procedure, for up to 1 week.

2.4. Sample preparation

To prepare the surrogate matrix calibrators/QCs and plasma samples for LC-MS analysis, a bead-based immunoprecipitation (IP) procedure was employed. Initially, 500 μl of each sample was aliquoted into a 2 ml 96-well extraction plate. Subsequently, 25 μl of the internal standard working solution (5000 pg/ml murine ACTH[1–39] in surrogate matrix) was spiked into each well, except for double blanks in which the same volume of surrogate matrix was added to the wells. Next, 100 μl of the Roche anti-ACTH capture antibody magnetic beads complex reagent was added to each well, and the plate was incubated at 4°C for 2 h on a shaker at 920 rpm, with continuous agitation to prevent bead settling.

Following incubation, the plate was placed on a magnet to allow the beads to settle over a 5-min period. This facilitated the transfer of the supernatant to waste, followed by two wash steps using 200 μl of TBS, with gentle bead mixing during each wash.

For elution, 60 μl of elution buffer (water:acetonitrile:formic acid 69:30:1) was added to each well, and the plate was mixed on a shaker at 900 rpm for 10 min. Then, 140 μl of deionized water was added to each well. The plate was again placed on the magnet, and the entire volume was transferred to a new 1 ml 96-well plate. The plate was covered, gently mixed for 1 min, and stored at refrigerated temperature prior to LC-MS analysis.

2.5. Liquid chromatography

For chromatographic separation, a Phenomenex Kinetex Biphenyl 2.6 μm, 100 A, 2.1 × 50 mm analytical column was used with a Shimadzu LC30AD HPLC system with Shimadzu AD multiplate autosampler. Mobile phase A consisted of 0.1% formic acid in water and mobile phase B consisted of 0.1% formic acid in acetonitrile. 50 μl of the extracted sample was injected onto the system under 93% mobile phase A conditions at a flow rate of 500 μl/min for loading to the column, which was maintained at 50°C. Table 1 shows the gradient conditions for the 6-min LC program.

Table 1.

Chromatographic gradient conditions for the ACTH(1–24) assay.

Time (min)	Mobile Phase B
0.00	7
0.50	7
0.60	15
2.00	30
2.10	95
2.90	95
3.00	7
4.00	95
4.90	95
5.00	7
6.00	7

2.6. Mass spectrometry

The amino acid sequences of ACTH(1–24) and murine ACTH(1–39) are shown in Figure 1 by the single letter symbols. ACTH(1–24) is a synthetic peptide containing the first 24 amino acids of human ACTH(1–39) and has a molecular weight of 2933.4. Murine ACTH (MW = 4582.3) has a similar sequence as human ACTH, except for a valine at position 26 (glycine in human) and asparagine at position 29 (aspartic acid in human). For mass spectrometric analysis, ACTH(1–24) and murine ACTH(1–39) were measured using a Sciex 6500 triple quadrupole mass spectrometer operating in positive ion electrospray ionization mode. Source conditions consisted of electrospray voltage of 4400 V, temperature of 500°C, and gas 1 and gas 2 settings were 55 and 40, respectively. A 5+ charge state (observed m/z 587.5) precursor ion was selected for ACTH(1–24) and a 6+ charge state precursor ion (observed m/z 764.4) was selected for the murine ACTH(1–39) internal standard, both isolated under a high-resolution setting. Collision energy (CE) of 27 V (22 V for IS) and collision-assisted dissociation (CAD) gas setting of 10 were used to fragment the molecular ions. A 4+ charge state product ion (m/z 671.6) was monitored for ACTH(1–24) and a 5+ charge state product ion (m/z 884.1) was followed for the IS, both under high resolution setting for the SRM acquisition. The product ion spectra are presented in Figure 2. The dwell time for ACTH(1–24) was 400 ms, whereas the dwell time for the internal standard transition was 100 ms. Voltages of the ion optics (DP, IE, CE, CXP) were optimized as part of tuning prior to acquisition and were not fixed.

Figure 1.

Amino acid sequences of adrenocorticotropic hormone (ACTH)(1–24), human ACTH(1–39), and the internal standard, murine ACTH(1–39). The first 24 amino acids are the same between the three peptides. The amino acids that are different between murine and human ACTH are underlined.

Figure 2.

Product ion spectra of adrenocorticotropic hormone (ACTH)(1–24) is shown in (A) and murine ACTH(1–39) in (B). The precursor ion and selected quantitative product ion are shown for both peptides.

2.7. Regression & data handling

Analyst 1.7 was the controlling program used to acquire the API 6500 data for this study. The collected chromatographic peaks (analyte and internal standard) were integrated to obtain peak areas for quantitation. The integration results were imported to Watson LIMS (Thermo) version 7.5. Regression of the standard curve was performed using peak area ratio of analyte to internal standard in Watson, with linear regression, y = ax + b (where y was peak area ratio of the analyte to the internal standard and x was concentration of the analyte) and weighting factor of one over x squared (1/x²). The measured analyte concentrations were then calculated against the calibration curve.

3. Results

3.1. Chromatography

Good chromatographic peak shape was observed in both surrogate matrix and human plasma samples. Example chromatograms are shown in Figure 3. For the lowest measured concentration at 5 pg/ml, the chromatographic peak was approximately 6 s wide. The total dwell time was 500 ms, and thus approximately 12 scans were collected across the peak at this concentration. ACTH(1–24) eluted around 1.2 min and murine ACTH(1–39) eluted around 1.7 min. No peaks were observed at these retention times in neither plasma nor surrogate matrix extracted blanks. Carryover was well-controlled, with response in the blank sample following calibration curve less than 20% relative to the 5 pg/ml calibration standard response.

Figure 3.

Example chromatograms. (A) The analyte response in unspiked control human plasma, (B) the 5 pg/ml calibration standard #1 prepared in surrogate matrix, (C) the analyte response in a human plasma spiked with 300 pg/ml ACTH(1–24) and (D) the internal standard response.

3.2. Surrogate matrix

The surrogate matrix used was composed of 0.05% Triton X in PBS (1x, pH 7.4). This surfactant was added to prevent non-specific binding (NSB) of low concentration peptides to container surfaces. The linearity of the calibration line (Figure 4) and the performance of plasma QCs against surrogate calibrators (Table 2) support the absence of NSB effects in the surrogate matrix samples. The protein content between surrogate and authentic matrices was not matched though, and slight differences in analyte responses were observed as shown in Figure 5, suggesting either recovery differences or differing suppressive effects between the two matrices. The internal standard was added at the beginning of the workflow (captured by the antibody reagent) and any differences in IP recovery or ion suppression between matrices were suitably normalized by the IS as evidenced by the statistical performance of plasma QCs against the surrogate matrix curve (Table 2).

Figure 4.

Example surrogate matrix calibration curve with duplicate standard curve points. The mean peak area ratio data (n = 2) are used for regression. The mean peak area (n = 6) for each plasma QC sample in the experiment are also shown.

Table 2.

Inter-day precision and accuracy statistics from the ACTH(1–24) fit-for-purpose method validation.

QC samples prepared in plasma
Sample	LLOQQC	LQC	MQC	HQC
Nominal concentration (pg/ml)	10	30	80	300
Mean concentration found (pg/ml)	11.0	30.6	75.7	305.4
Inter-run %CV	11.8	13.7	13.3	12.8
Inter-run %Bias	9.6	2.0	-5.3	1.8
n	21	22	22	22

Four separate batches on four separate days comprise the inter-assay analysis.

Figure 5.

The mean peak area responses (n = 6) for plasma and surrogate matrix QCs are compared for one analytical batch (A). For ease of comparison across the concentration range, the mean measured peak area concentrations are normalized to the nominal concentration of the QC level in (B).

3.3. QC performance

The precision and accuracy of the assay was evaluated in four separate batches analyzed on separate days, with five to six individual replicates of spiked plasma samples each at four different QC levels spanning the assay range. The lower limit of quantitation QC (LLOQQC) was prepared at 10 pg/ml and was considered the lower limit of quantitation of the assay despite the 5 pg/ml standard curve level. The low QC (LQC) was prepared at 30 pg/ml, the mid QC (MQC) at 80 pg/ml, and the high QC (HQC) was prepared at 300 pg/ml). As shown in Table 2, the inter-assay %RE was less than 20% and %CV was less than 20% for each level, which were the intended limits of acceptance and indicated that the assay was suitably precise and accurate for measurement of our samples.

3.4. Assay interference

The accuracy and precision of ACTH(1–24) was evaluated in the presence of 100 pg/ml ACTH(1–39). This was anticipated to be a conservative upper limit of the ACTH(1–39) levels endogenously present in healthy individuals that were given an investigational ACTH antagonist drug under the doses administered. As shown in Table 3, acceptable accuracy and precision of ACTH(1–24) was observed when measured in the presence of 100 pg/ml ACTH(1–39).

Table 3.

Performance of ACTH(1–24) plasma QCs prepared in the presence of 100 pg/ml ACTH(1–39).

Sample	LQC	HQC
ACTH(1–24) Conc. (pg/ml)	30	300
Mean Conc. Found (pg/ml)	26.8	323
%CV	11.7	4.1
%Bias	-10.7	7.7
n	5	5

3.5. Stability

Instability was observed (data not shown) for plasma samples fortified with ACTH(1–24) when stored at the bench at room temperature or on wet ice for 3 h, after two freeze–thaw cycles and after frozen storage (-70°C) for 30 days. We observed a negative bias approximately -20% to -25% for stressed conditions when compared with calculated concentration of samples measured directly after preparation. The observed biases were largely independent of concentration, thawing method (room temperature or wet-ice thawing) or storage temperature. Similar negative bias was observed for these experiments in surrogate matrix as well. Further experiments on the stability of ACTH(1–24) in biological and surrogate matrix must be conducted, and options for preserving sample stability explored if determined to be necessary.

4. Discussion

To achieve the required assay sensitivity to support the low-dose ACTH stimulation test, a large sample aliquot (500 μl) was required. The need for large sample volume drove the need for a rigorous sample purification approach, such as immunoprecipitation. The format of the hybrid immunoaffinity-LC-MS/MS assay for ACTH(1–24) was modeled after the LC-MS/MS assay for ACTH(1–39) described by Shi et.al. [13]. The epitope assignment reported in their paper indicated that the capture antibody from the Roche assay kit would also be a good choice as capture antibody for ACTH(1–24). That the capture antibody reacts with human ACTH(1–39), murine ACTH(1–39), and ACTH(1–24) highlights the value of the hybrid IA-LC-MS/MS approach. Only one antibody reagent was required for the assay, and the multiple dimensions of separation afforded by the analytical platform ensured selective measurement of our target peptide. Here, the immunoaffinity aspect of the hybrid approach was a purification strategy, concentrating our target analyte while reducing the complexity of the matrix, which improved the signal to noise ratio of measurement by reduction of suppressive matrix effects and background response.

An important aspect in the development of the bead-based IA-LC-MS/MS assay was the optimization of antibody reagent as well as the type and amount of magnetic beads to use for capture. For this work, a streptavidin/biotin conjugated strategy was employed. Rather than incubate the plasma sample with biotinylated antibody followed by short incubation with streptavidin-coated magnetic beads [14], we first conjugated the antibody to the magnetic beads in bulk [15]. The benefits of this approach were twofold. First, a pre-conjugation strategy should be more tolerant to high levels of endogenous biotin in samples. This has been reported as a problem with clinical ACTH assays [16]. High levels of free biotin can bind to the streptavidin coated magnetic beads and compete with the biotinylated capture antibody for binding sites. While an excess of beads can be added to the samples in a hybrid IA-LC-MS/MS assay format to counter potentially high levels of biotin, the levels of biotin in study samples are not measured and the impact to any one sample would be unknown. Further, excess beads offer more surface area for unintended matrix components to non-specifically adhere and may pollute the final extract. Pre-conjugating the biotinylated antibody to the streptavidin-coated beads removed this uncertainty, as the antibody was already conjugated to the bead before being added to the sample. The second benefit of the pre-conjugation approach was control over the antibody/bead stoichiometry. The biotinylated antibody was added in modest excess to the beads, and any excess was washed away prior to use in the assay. This ensured an appropriate match of antibody and beads during extraction. The pre-conjugation approach can also be accomplished using approaches other than the streptavidin/biotin interaction, such as direct bead conjugation using tosyl chemistry [17]. A benefit of direct conjugation techniques is that the need for characterization of the biotin load for the labeled antibody and potential optimization via challenge ratio may be avoided. For this work, we leveraged streptavidin/biotin interaction because the capture antibody provided in the Elecsys kit was already biotinylated.

Figure 4 shows the response linearity of both peak area ratio (PAR) as well as peak area responses of the calibration standards. The linearity of the analyte peak area responses without internal standard normalization indicates that we were not approaching an antibody binding capacity limit with our bead-based extraction procedure. With IA-LC-MS/MS workflows that involve addition of internal standard before immunoprecipitation step, non-linear recovery of analyte due to antibody binding capacity limit can be masked by internal standard normalization. This was not observed here and indicates that the capture antibody was added in suitable excess.

Murine ACTH(1–39) was chosen as the internal standard because of the similarity in sequence. It is identical with the human ACTH sequence for the first 25 residues. Thus, murine ACTH is a good choice as internal standard should human ACTH(1–39) be added to the assay as an analyte in the future. Because the internal standard is 15 amino acids longer than ACTH(1–24), there were tradeoffs in its use for the work presented here. As shown in Figure 3, the analyte and IS do not coelute. Further, due to larger size of murine ACTH(1–39), a higher charge state precursor was selected (6+) compared with ACTH(1–24) (5+). Ideally, the internal standard would co-elute with the analyte and have a similar transition, but murine ACTH(1–39) gave suitable performance as internal standard for the needs of the assay. Should a further optimized assay be required, a stable-isotope labeled version of ACTH(1–24) may be procured and implemented.

It was important to measure ACTH(1–24) in the intact form. Enhanced sensitivity can potentially be achieved via enzymatic digestion as with trypsin. With trypsin digestion, a peptide is cleaved at the C-terminal end of lysine and arginine residues, with some exceptions. This yields primarily doubly charged tryptic peptides, which serve to limit the diluting effect of multiple, higher charge states presented with electrospray ionization of the intact peptide. But with digestion, we lose the selectivity advantage over immunoassay approaches gained from LC and MS separation. As evident from Figure 1, there is significant similarity between murine and human ACTH, with identical sequence for the 24 amino acids that comprise ACTH(1–24). Selectivity of any assay strategy leveraging digestion would be challenged by such sequence homology. As the capture antibody itself was not able to capture ACTH(1–24) without also capturing ACTH(1–39), a digestion-based approach could not meet the needs of the study. Whereas digestion can be an indispensable tool for protein quantitation by LC-MS, great care must be taken to use this to improve sensitivity for low abundance circulating peptide quantitation. Digestion of circulating peptides can, however, be a useful way to coalesce disparate peptides into one measurement. For example, as a total, combined measurement of several peptide isoforms by quantitation of a common surrogate peptide [18]. For this work, key steps in the IA-LC-MS/MS assay were optimized to drive sensitivity with measurement of the intact peptide.

Although ACTH(1–24) was not measurable in naive human plasma, a surrogate matrix approach was chosen for this assay. This was done both for consistency with the established ACTH(1–39) IA-LC-MS/MS assay [13] and to allow for future potential to multiplex with ACTH(1–39) or other POMC-derived peptides without changing the calibration design. The lack of ACTH(1–24) response in control human plasma simplified the quality control strategy, as the spiked analyte concentration served as the nominal concentration for %RE calculation and was not complicated by presence of endogenous response. For assessment of small molecule ‘parallelism’ [19], or the equivalence of slopes between surrogate and biological matrices [20], straightforward %RE and %CV assessments of plasma QCs interpolated from surrogate matrix calibration lines served the purpose due to this lack of endogenous response.

Our reason for quantifying ACTH(1–24) was to understand the total bioactive ACTH in clinical study subject samples taken during the CST. The sum of ACTH(1–39) and ACTH(1–24) concentrations comprises total bioactive ACTH, for our purposes. An immunoassay was used to measure ACTH(1–39), and the assay described here measured ACTH(1–24). Because the total bioactive ACTH was the sum of both measurements, it's important that one analyte does not interfere with the measurement of the other. The demonstrated accurate measurement of ACTH(1–24) in presence of 100 pg/ml ACTH(1–39) supports assertion that the IA-LC-MS/MS method selectively measures ACTH(1–24).

5. Future perspective

The quantitation of low-concentration peptide hormones has historically been dominated by immunoanalytical techniques. Despite the selectivity challenges experienced by the ligand binding assays for endogenous peptides, adoption of LC-MS platforms for these measurements has been slow, mostly due to sensitivity challenges. The incorporation of affinity capture into LC-MS workflows has alleviated some of the challenges around sensitivity, by reducing matrix effects with a purification mode that is orthogonal to the separation chemistries typically used with LC-MS approaches. The selectivity inherent to LC-MS reduces the specificity required of the critical reagents compared with when they are employed in an immunoassay. As shown here, an antibody from an ACTH(1–39) kit is used for capture of ACTH(1–24). We predict that adoption of LC-MS for quantitative measurement of low-level peptide biomarkers and therapeutics will continue to increase, enabled by both incorporation of orthogonal purification/separation strategies (including affinity purification) as well as advancements in mass spectrometry and liquid chromatography technologies. However, the selectivity of the MS platform will not always be a desirable quality. In clinical diagnostics of ACTH(1–39), for example, cross-reactivity to ACTH precursors POMC and pro-ACTH are required. In the ectopic ACTH precursor syndrome, the ectopic tumors secrete primarily precursors, also bioactive at the MC2 receptor. A clinical assay that does not co-measure precursors may miss presentation of an ectopic tumor. Nonetheless, as the sensitivity of LC-MS improves, the enhanced selectivity combined with high multiplexing capacity will be leveraged to better understand the biology and pathology surrounding hormones such that meaningful diagnostic tests can be designed in a way that tells us more about human health than we can currently achieve.

6. Conclusion

The mechanistic PK/PD modeling of the HPA axis with an investigational ACTH competitive antagonist drug benefits from distinguishing endogenous ACTH and exogenous ACTH(1–24) in the context of the cosyntropin stimulation test. Toward this, an IA-LC-MS/MS method for the quantitative determination of ACTH(1–24) in human plasma samples was successfully developed with a lower limit of quantitation of 10 pg/ml and validated to the degree of rigor required for the application. The developed assay benefited from the previous work of Shi et al. on ACTH(1–39) and those authors' comparison to clinical ACTH assays. It also benefitted from the cross-reactivity of the capture antibody from the Roche Elecysys ACTH kit to ACTH(1–24). The application of an ACTH(1–39) assay format to the analysis of a closely related peptide, combining the selectivity of both liquid chromatography and mass spectrometry with the purification of affinity capture, underlines the power of the hybrid technique where the two peptides could not be selectively and separately quantified using immunoanalytical platforms alone. The method described here was shown to be suitable for the use of the generated data, and it is a solid starting point for further optimization should this assay be required for subsequent work where more validation rigor is needed. The use of a stable isotope labeled (SIL) internal standard would add robustness to the measurement by co-elution with unlabeled ACTH(1–24) and by matching of charge states between analyte and internal standard. The observed instability of ACTH(1–24) in both surrogate matrix and human plasma requires more investigation. Sensitivity remains a limiting factor for this assay, and a large amount of plasma was required (500 μl) to reach the desired LLOQ (10 pg/ml). Multiplexing the ACTH(1–24) measurement with ACTH(1–39) in the same assay would be ideal for the intended use and the reported method does have this potential, however further optimization may be required to drive additional sensitivity for both analytes, such as the use of supercharging reagents to alter predominant charge states [21,22] or perhaps more effectively by the use of microflow or nanoflow liquid chromatography [12,17,23] for improved ionization efficiency and ion sampling.