Validation of an ELISA kit to measure allopregnanolone in human and equine hair

Abstract

In humans, allopregnanolone plays important roles in a number of different neurodegenerative disorders, and it has been proposed for use in some therapies. Horses are commonly used as animal models for human neurodegenerative diseases, mental and behavioral disorders, and neuropsychiatric disorders, and there is interest in using hair as a biological sample to study hormones in these conditions. We validated the use of a commercial ELISA kit (DetectX allopregnanolone kit; Arbor Assays), which was designed for serum, plasma, feces, urine, and tissue samples, to assess allopregnanolone in hair samples from 30 humans and 63 horses. The ELISA kit had good precision (intra- and inter-assay CVs: 6.4% and 11.0% for equine hair; 7.3% and 11.0% for human hair, respectively), sensitivity (50.4 pg/mL for both equine and human hair), and accuracy (parallelism and recovery tests) in determining allopregnanolone concentrations in hair from both species. The allopregnanolone concentrations in human hair were 7.3–79.1 pg/mg; the concentrations were 286 ± 141 pg/mg (x̄ ± SD) in mares on the day of parturition and 16.9 ± 5.5 pg/mg in nonpregnant mares. The DetectX ELISA kit offered a simple and accessible analysis of allopregnanolone in human and equine hair samples.

Allopregnanolone is a neurosteroid and an important metabolite of progesterone (P4) that is synthesized both in the brain and peripherally by the adrenal glands, ovaries, and testes. ¹³ Allopregnanolone has anxiolytic, anticonvulsant, sedative, analgesic, and anesthetic properties. ¹⁷ In humans, the concentrations of allopregnanolone typically decrease under chronic stress, causing dysregulation of the hypothalamic–pituitary–adrenal (HPA) axis. During acute stress, the concentrations of allopregnanolone instead increase to normalize (dampen) the hyperactivated HPA axis and regain homeostasis. ² Allopregnanolone also modulates fear and aggressiveness, ¹⁶ has a neuroprotective action in different situations, ³ and is a proneurogenic agent that has been proposed for use in some therapies. ¹² Allopregnanolone may have a place in equine clinical medicine given that progestogens, such as allopregnanolone, have been reported to have sedative effects in healthy foals. ¹⁵

During pregnancy in both humans ³ and horses, ⁴ allopregnanolone protects the fetus from the action of glucocorticoids, the release of which is triggered when the HPA axis is stimulated. Near the end of pregnancy, the reflex mechanisms that trigger oxytocin secretion for birth become more excitable, and premature stimulation of the secretion of oxytocin can lead to preterm birth; however, allopregnanolone has important actions that prevent this premature activation. ³ Additionally, the quiescence of the equine uterus is maintained during late gestation by P4 metabolites, ⁸ with statistically significant declines in allopregnanolone concentrations starting from 5 d before delivery. ¹⁴ As well, the ability of allopregnanolone to inhibit muscle (such as uterine muscle) activity has been reported in humans. ⁶

Allopregnanolone has been studied in plasma samples from both humans ¹⁰ and horses, ⁷ but there is interest in using hair as a biological sample to study hormones in depressive disorders, ¹¹ neurodegenerative conditions, ¹⁹ and stress reactivity. ²⁰ Horses are commonly used as animal models for these conditions. ⁹ Hair is a non-conventional matrix that has significant advantages, including the following: sampling is simple, can be performed by a single operator, and is not invasive or painful for the individual that is being sampled; sample storage can be in a paper envelope, which must be kept away from direct light and at room temperature (RT); and hair is not influenced by systemic hormonal fluctuations. Furthermore, hair is a retrospective and cumulative matrix, meaning that it can provide historical information about hormone concentrations with a lag time of ~2 wk. ¹⁸

Our aim was to validate the use of a commercial competitive ELISA kit (K061-H5, DetectX allopregnanolone enzyme immunoassay kit; Arbor Assays), which was designed for serum, plasma, feces, urine, and tissue samples, to assess allopregnanolone in hair, an alternative matrix that requires a highly sensitive analytical method. We tested the kit on human and equine hair because the latter could be useful as an animal model and in translational studies for the therapeutic use of allopregnanolone.

Materials and methods

Sample type and sampling

The collection of equine hair was carried out in accordance with EU Directive 2010/63/EU and Italian legislation on animal care (DL 26; 04/03/2014); human hair samples were collected from volunteers in a previous study that was carried out in accordance with the Declaration of Helsinki and EU Directive 2016/679, which covers the collection and use of data.

The equine hair samples were collected from healthy Italian Trotter mares (7–10-y-old) on a Standardbred farm: 35 pregnant mares (PMs) and 28 non-pregnant mares (NPMs). Hair (200 mg) was cut from the withers region close to the skin using an electric shaver.

Human hair samples were collected from 30 healthy volunteers (22–90-y-old, 47 ± 19-y-old; x̄ ± SD; 12 women, 18 men). We excluded smokers, pregnant women, people with adrenocorticoid dysfunctions and those who used oral contraceptives, hair dyes, anti-dandruff shampoos, and frequent washing or other hair treatments. The human hair samples (50 mg) were obtained from the proximal 3 cm of scalp hair in the posterior-to-vertex position using commercial vacuum hair clippers. Hair samples were stored in paper envelopes and kept in the dark and at RT until processed. They were analyzed 1 wk after collection.

Hair washing and extraction

Each collected sample (200 mg of equine hair; 50 mg of human hair) was placed in a polypropylene tube, covered with 5 mL of ultrapure water, and mixed gently for 3 min at RT; this washing was repeated twice and then twice after covering the sample with 5 mL of isopropanol (Merck). The samples were then air dried. This washing procedure minimized the risk of extracting steroids that may coat the outer cuticle from sebaceous and eccrine secretions.

Subsequently, hair samples (10 mg of equine hair, 25 mg of human hair) were held in microperforated stainless steel cylinders (height: 20 mm; diameter: 0.5 mm) held with a thread, and the cylinder was placed in a glass vial with 3 mL of ≥ 99.8% methanol (Merck) for incubation at 37°C for 16 h to extract steroids. Then, the cylinder was elevated by the thread and blocked by the vial’s cap. Centrifugation of the vial at 274 × g for 15 min allowed the methanol containing the extracted steroids to be separated from the hair specimens. The methanolic extract in the vial was evaporated at 37°C under an air-stream suction hood. The expected hormonal concentrations were not the same for human and equine hair; therefore, we dissolved the dry residue from the equine and human hair extracts in 1.2 mL or 0.4 mL, respectively, of the allopregnanolone enzyme immunoassay kit assay buffer (previously diluted following the manufacturer’s instructions), which allowed us to work within the range of the standard curve provided with the kit.

Allopregnanolone enzyme immunoassay characteristics

We evaluated the allopregnanolone in the equine and human hair samples using the DetectX competitive ELISA allopregnanolone kit. The kit included a clear plastic 96-well microplate that was coated with goat anti-mouse IgG used to attach the capture antibody (mouse IgG anti-allopregnanolone) to the wall of wells; sample allopregnanolone (endogenous) and labeled allopregnanolone (allopregnanolone–peroxidase conjugate, or tracer) competed for the capture antibody. The assay is highly specific for allopregnanolone, with minimal cross-reactivity with other steroids (e.g., maximum cross-reactivity is with tetrahydrodeoxycorticosterone at 3%).

Briefly, the method was as follows: 100,000 pg/mL of allopregnanolone standard was mixed with a special stabilizing solution to generate the standard curve and diluted (as indicated in the kit instructions) to obtain 7 standards: 1 (10,000 pg/mL), 2 (5,000 pg/mL), 3 (2,500 pg/mL), 4 (1,250 pg/mL), 5 (625 pg/mL), 6 (312 pg/mL), and 7 (156 pg/mL). Furthermore, 2 controls from extracted human serum were added to the plate: one had a low allopregnanolone concentration (549 pg/mL) and the other had a high allopregnanolone concentration (1,720 pg/mL).

Hair allopregnanolone ELISA analysis

Standards, controls, and samples (50 µL of each) were pipetted in duplicate into different wells of a 96-well microplate. Then, 25 µL of the DetectX allopregnanolone conjugate and 25 µL of the DetectX allopregnanolone antibody were added to each well. After 15 min of shaking at RT to ensure adequate mixing of the reagents, plates were incubated for 16–18 h at 4°C to allow for the competitive binding between the antigen (allopregnanolone) and the allopregnanolone antibody. The microplate was then washed with wash buffer, and 100 μL of 3,3′,5,5′-tetramethylbenzidine (TMB) was added to each well. This step was followed by incubation at RT for 30 min to allow for the reaction between the TMB substrate and the horseradish peroxidase that was conjugated to the DetectX allopregnanolone in the wells. The stop solution was then added, and the microplate was read at 450 nm (Ensight multimode plate reader; Perkin-Elmer). As recommended by the manufacturer, the temperature was kept under strict control for the duration of the analysis. Standards, controls, samples, and reagents were pipetted carefully using the same calibrated pipettes to minimize errors.

Hair allopregnanolone ELISA validation

All of the validation tests for the assay that we used for the hair allopregnanolone measurements were conducted in the same way for the equine and human specimens; for each species we used 2 different pools of samples (one for parallelism and one for recovery), each comprising 5 hair extracts.

We performed a parallelism test to identify any deviations from the standard curve among the series of hair extracts, which contained known amounts of allopregnanolone. The test was prepared using serial dilutions of the extracts from PMs and the extracts from women in assay buffer, which had high concentrations of allopregnanolone (85.6 pg/mg for equine hair; 51.7 pg/mg for human hair). Each dilution was pipetted in quintuplicate. Linear regression was used to determine if hair extracts and the standard allopregnanolone curve deviated from parallelism.

We conducted a recovery test to evaluate the system response to increasing amounts of allopregnanolone standard, which were added to hair extracts with low allopregnanolone concentrations (NPMs extract pool, and women extract pool). Each spiked or non-spiked sample was pipetted in quintuplicate. The percentage of recovery was determined as follows: ([measured allopregnanolone in a spiked sample]/[measured allopregnanolone in a non-spiked sample + added allopregnanolone] × 100%).

We determined the analytical sensitivity of the assay as the hormone concentration that resulted in the displacement of the labeled hormone by at least 2 SD from maximal binding (as calculated by RiaSmart; Canberra-Packard).

We estimated the precision of the assay by assessing intra- and inter-assay variation, which was expressed as the CV. The intra-assay precision was evaluated by testing 20 replicates of a sample with a known medium allopregnanolone concentration (356 and 10.5 pg/mg for equine and human hair, respectively) in the same assay. The inter-assay precision was evaluated by running the same horse and human hair in 20 assays in duplicate over 10 d.

Statistical analysis

The dose-response curve (4PL curve) was fitted using the statistical software for the RiaSmart immunoassay data. The software was used to interpolate the analyte concentrations in the unknown samples. The arithmetic x̄ and SDs, along with the CVs and linear regression, were calculated in a spreadsheet using routine descriptive and predictive statistical analysis procedures (Excel 2019; Microsoft).

Results

The parallelism test between the hair dilution curves and the standard curve indicated that the hair allopregnanolone and allopregnanolone standard reacted identically to the antibodies (Figs. 1, 2). A high correlation (r = 0.99) was observed between the obtained and expected concentrations. In the recovery test used to evaluate the response of the system to increasing amounts of allopregnanolone standard, the recovery rate was 88.9 ± 2.3% and 86.7 ± 2.4% (x̄ ± SD) for horse and human, respectively (Table 1). The assay analytical sensitivity was 50.4 pg/mL. The equine hair extracts (356 ± 23 pg/mg; x̄ ± SD), in repeated determinations, had intra- and inter-assay CVs of 6.4% and 11.0%, respectively. The human hair extracts (10.5 ± 0.8 pg/mg; x̄ ± SD), in repeated determinations, had intra- and inter-assay CVs of 7.3% and 11.0%, respectively.

Figure 1.

Graphical plot for the parallelism obtained by the DetectX ELISA kit (Arbor Assays) on equine hair. The relationship between equine hair allopregnanolone concentrations and the standard allopregnanolone curve is given by the equation y = 0.96x + 1.16.

Figure 2.

Graphical plot for the parallelism obtained by the DetectX ELISA kit (Arbor Assays) on human hair. The relationship between human hair allopregnanolone concentrations and the standard allopregnanolone curve is given by the equation y = 1.05x–1.13.

Table 1.

Allopregnanolone recovery study: 4 known quantities of allopregnanolone (+31.2, +62.5, +125, and +250 pg/tube) were added to equine (nonpregnant mares; 58.4 pg/tube) and human (women extract pool; 4.7 pg/tube) hair pool extracts.

Sample hair pool	Total, pg/tube	Observed, pg/tube	Recovery, %
Equine	58.4	58.3 ± 2.4	100
	89.6	80.7 ± 5.7	90
	121	108 ± 3.8	89
	183	157 ± 2.6	85
	308	278 ± 7.1	90
Human	4.7	4.7 ± 0.8	100
	35.9	32.4 ± 3.4	90
	67.2	57.5 ± 1.8	85
	130	110 ± 3.7	84
	255	220 ± 3.5	86

The average concentration (x̄ ± SD) is shown. Percentage recovery was calculated on observed versus expected.

The allopregnanolone concentrations (x̄ ± SD) in the hair samples from PMs and NPMs were 286 ± 141 and 16.9 ± 5.5 pg/mg, respectively. The allopregnanolone concentrations in the human hair samples were 7.3–79.1 pg/mg.

Discussion

The DetectX ELISA kit, which has been validated for equine plasma, ⁵ had good precision, sensitivity, and accuracy (by parallelism and recovery tests) for determining allopregnanolone in both human and equine hair. The DetectX allopregnanolone assay allowed us to discriminate between equine hair samples with known high concentrations (PMs) and samples with lower concentrations (NPMs). The allopregnanolone concentrations in human hair were found to be lower than those in equine hair. However, these results need to be confirmed using larger numbers of samples.

We found the DetectX ELISA to be a simple and accessible assay to study allopregnanolone in human and equine hair samples. Hair is the only matrix with properties that allow for retrospective analysis of allopregnanolone, which can provide medium- and long-term information. Given that neurodegenerative diseases usually develop gradually over many years, allopregnanolone concentrations in hair could help to evaluate their development over time in equine disease models using noninvasively collected samples. Allopregnanolone concentrations in hair could help to monitor the performance of allopregnanolone as a therapeutic tool for clinical application in equine medicine and to study the complex mechanisms in which it is involved during gestation and birth. The DetectX ELISA kit, which can use hair as a matrix, might also be used in other reproductive studies (e.g., allopregnanolone plays an important role in ovarian pathophysiology). ¹