Stability of sample extracts of vitamin D3 metabolites after chemical derivatization for LC–MS/MS analysis

Liquid chromatography/tandem mass spectrometry (LC–MS/MS) is widely used to determine vitamin D3 metabolites in biological samples. The ionization efficiencies of these metabolites, however, are poor under electrospray ionization conditions. Moreover, the chromatographic separation of multiple vitamin D metabolites and their epimers can be challenging. For these reasons, chemical derivatization reagents are often used to improve sensitivity and selectivity of analysis. While the derivatization schemes have been proven to be very effective, one missing aspect is the investigation of the stability of the chemical derivatization products in stored sample extracts. In this study, we investigated the long-term stability of several vitamin D3 metabolites after 1 and 3 months of storage at − 20 °C. Five vitamin D3 metabolites were examined after derivatization with seven different derivatization reagents. Generally, Amplifex products were the most stable in the long term in our study with 11–20% degraded after 1 month of storage and 14–35% after 3 months. The stabilities for some of the metabolites′ 4-[2-(6,7-dimethoxy-4-methyl-3-oxo-3,4-dihydroquinoxalyl)ethyl]-1,2,4-triazoline-3,5-dione (DMEQ-TAD), 2-fluoro-1-methylpyridinium p-toluenesulfonate (FMP-TS), isonicotinoyl chloride (INC) and 4-phenyl-1,2,4-triazoline-3,5-dione acetylated (PTAD-Ac) products were also acceptable after 1 month of storage. Other derivatized metabolites, however, degraded extensively already after 1 month of storage, such as 4-phenyl-1,2,4-triazoline-3,5-dione (PTAD) (54–72% degradation) and 2-nitrosopyridine (PyrNO) (32–100% degradation). Importantly, for every metabolite, there was an optimum derivatization reagent that met the criteria of stability proposed by international regulatory bodies after 1 month of storage. Some derivatives were stable for even up to 3 months of storage, with degradation of less than 15%. Graphical abstract Supplementary Information The online version contains supplementary material available at 10.1007/s00216-022-04409-5.


Introduction
Measurement of vitamin D metabolites requires selective and sensitive assays that enable accurate measurement of very low concentrations of multiple species in biological sample matrices, which presently can only be achieved by liquid chromatography/tandem mass spectrometry (LC-MS/ MS) [1]. Usually, vitamin D measurements are performed in blood-based matrices such as serum or plasma, but there is a variety of other biological samples (urine, saliva, tissues, etc.), where vitamin D metabolites are investigated [2]. A wide range of different sample preparation techniques are employed prior to instrumental analysis [3] and sometimes a chemical derivatization step is incorporated because it can enhance the assay's sensitivity and selectivity, which is important for quantification of low abundant metabolites. Furthermore, chemical derivatization can also improve the chromatographic separation. Obviously, these chemical derivatization reactions are ideally fast and inexpensive and preferably result in only one stable product with high yields.
The stability of the analytes in the sample is an important factor, which is investigated during the assay validation. Guidelines for the investigation of stability are available from various regulatory bodies such as the Food and Drug Administration (FDA), US Pharmacopeia (USP), US Environmental Protection Agency (USEPA), American Association of Official Analytical Chemists (AOAC), European Medicines Agency (EMA) and Eurachem [4,5]. Stability ensures that the concentration of the investigated compounds 1 3 is not affected during sample preparation and sample analysis as well as the storage conditions of the sample or the extracted sample. There are subtle differences between these guidelines, but all of them aim to guarantee the quality of the analytical method [6][7][8]. FDA recommends the investigation of (1) autosampler stability, (2) bench-top stability, (3) extract (or processed sample) stability, (4) freeze-thaw stability, (5) long term stability and (6) stock solution stability [9]. According to EMA, for the stability of the sample, the following stability tests should be evaluated [10]: (1) stability of the stock solutions and working solutions of the analytes and internal standard, (2) freeze and thaw stability, (3) short-term stability of the analyte in matrix and (4) longterm stability of the analytes in matrix stored in the freezer. Additionally, for the stability of the extracted sample, (1) autosampler stability of the processed sample and (2) stability of the processed sample under the storage conditions to be used during the study until the sample will be analyzed should be evaluated.
If sample preparation requires a derivatization step, it must be ensured that the resulting derivatization products of the analytes are stable until the sample is measured. Usually, this stability test refers to the sample processing time, sample handling and waiting time in the autosampler. However, there are cases, e.g. instrument failures or need for intermediate storage, advance preparation of large sample numbers, prevention of the need to repeat derivatization for re-measurements etc., when the extracted sample must be stored in the freezer before measurement. As a consequence, it is essential that the derivatized analytes are sufficiently stable until the sample is measured. While a stable isotope standard for each analyte in the sample should ensure accuracy if analyte and stable isotope standard degrade to the same extent, significant degradation rates might lower the concentrations of low abundant metabolites to levels outside the validated calibration range or even below LOQ or LOD.
The human vitamin D 3 free serum was purchased from Sigma-Aldrich (VD-DDC Mass Spect Gold serum).

Preparation of standard solutions and sample preparation of serum samples
Stock solutions of the investigated vitamin D 3 compounds were prepared in methanol at 1 mg/mL and stored at − 20 °C. These solutions were used to spike serum samples with the investigated compounds. Human vitamin D 3 free serum was used and divided into two groups after spiking: group A included 3β-25(OH)D 3 at 20 ng/mL, vitamin D 3 at 30 ng/mL and 1,25(OH) 2 D 3 at 10 ng/mL; group B included 3α-25(OH) D 3 at 5 ng/mL and 24,25(OH) 2 D 3 at 10 ng/mL. The reason for the separation into two groups was the lack of adequate chromatographic separation of 25(OH)D 3 epimers and dihydroxy species. The concentrations of the metabolites were chosen sufficiently high to be able to be measure even without derivatization.
The sample preparation protocol was previously described by Ding et al. [22]. Briefly, protein precipitation was performed in 100 μL of serum using 250 μL of acetonitrile followed by 1 min vortexing and 15 min of centrifugation. A Concentrator plus/Vacufuge® plus (Eppendorf, Hamburg, Germany) was used to evaporate the supernatant to dryness after its transfer to a new vial and two-step liquid-liquid extraction (LLE): firstly, 100 μL of water and 200 μL ethyl acetate were added to the dry residue followed by 30 s vortexing and 5 min centrifugation. After removing the upper organic phase to a fresh vial, the aqueous phase was re-extracted by adding 200 μL of ethyl acetate. From the two combined organic phases, 380 μL were transferred to a new vial and evaporated to dryness. The dry residue was used for chemical derivatization, the details of which are described in the following section.
For every derivatization reagent that was examined, three spiked serum samples from group A and three spiked serum samples from group B were derivatized. Every sample was measured in triplicate. Moreover, a blank sample that underwent the identical sample preparation and derivatization steps was prepared for all reagents and was measured in triplicate.
The coefficient of variation (RSD) in the stability testing was between ± 5%, ± 8% and ± 7% for all metabolite derivatization products at t = 0, after 1 month and after 3 months, respectively. The calculated RSD values exhibited very good reproducibilities, ensuring that loss of observed analyte signals could be confidently assigned to sample degradation.

Derivatization procedures
The individual steps of the derivatization reactions are compared in Table 1 for the single incubation reactions and in Table 2 for the double incubation reactions. The FMP-TS, INC and PTAD-Ac derivatization reactions were developed by our group to further enhance the detection sensitivity.

Design of the stability experiments
To evaluate the storage stability of the vitamin D 3 metabolites' derivatization products, three spiked serum samples were prepared independently for the chosen analytes' concentration level (see above). Every sample was measured in triplicate. Moreover, we also examined the stability of the non-derivatized metabolites. We limited the extent of the study to one concentration level, as otherwise it would have been too time-consuming. We chose the concentration level to be as close as possible to typical concentrations in serum samples, but at the same time still quantifiable as underivatized compounds. The long-term stability testing was carried

Liquid chromatography-tandem mass spectrometry
For, LC-MS/MS, 5 μL of each sample solution were injected into a 1290 Infinity II LC system (Agilent, CA, USA). A Phenomenex (Torrance, CA, USA) Kinetex 2.6 µm C-18 100 Å column (100 × 2.1 mm). The mobile phases consisted of solvent (A) water (+ 0.1% formic acid) and solvent (B) methanol (+ 0.1% formic acid). A gradient elution program was applied for 15 min and solvent (B) increased linearly from 50 to 100%. Subsequently, it was held constant for 2 min at 100% (B) before returning to the initial conditions. Re-equilibration was performed for 3 min. The flow rate of the mobile phase was constant at 0.4 mL/min, and the temperature in the column oven was set at 30 °C. The same elution program was used for all the tested reagents. The slow gradient was chosen to allow optimum separation of the epimers, isomers and stereoisomeric derivatization products. Importantly, this HPLC method was only used for the purpose of the stability study. It is not a routine method optimized for routine, high-throughput analyses of vitamin D metabolites.
The UHPLC system was coupled to a Sciex (Concord, ON, Canada) QTRAP 6500 + triple quadrupole-linear ion trap mass spectrometer equipped with a Turbo-V electrospray ion source in positive ion mode. Data acquisition was performed in MRM mode. The parameters of the ion source were as follows: curtain gas, 35 psi; IonSpray voltage, 5500 V; nebulizer gas, 30 psi; heating gas, 30 psi; collision gas was set to medium. MRM transitions and collision-induced dissociation conditions were optimized for each vitamin D metabolite derivatization product individually. The optimized values for declustering potential (DP), entrance potential (EP), collision energy (CE) and collision cell exit potential (CXP) are summarized in Table S1 (Supplementary Material). The dwell time was chosen to provide 12-20 data points across the chromatographic peaks. Analyst (Sciex) version 1.7 and MultiQuant (Sciex) version 3.0.3 software were used for data analysis.

Results and discussion
In the following section, we present the results of the longterm stability experiments of the different chemical derivatization products of vitamin D metabolites. Three time points were tested (t = 0, after 1 and 3 months) while the derivatized sample extracts were kept at − 20 °C. We chose this storage temperature because − 20 °C freezers are more commonly available in analytical laboratories than − 80 °C freezers. The mean measured signal intensities were compared to the mean signal intensities of the freshly prepared samples at t = 0. (Table 3).
Initially, we investigated the stability of the non-derivatized metabolites in serum after sample preparation. The extracted non-derivatized metabolites were not particularly stable, even after 1 month of storage at − 20 °C. The extracts exhibited 48-90% of the values seen for t = 0 (Table 3), and only vitamin D 3 was sufficiently stable after 1 month of storage maintaining 90% of the observed value at t = 0.
Wan et al. mentioned that the 1,25(OH) 2 D 3 -PyrNO derivatization product was stable for 1 week when kept at − 20 and − 80 °C with values observed ˃90% of those at t = 0 [23]. Importantly, these observations were for standard solutions, not for serum extracts as investigated in this study. Our extracted 1,25(OH) 2 D 3 -PyrNO derivatization product was only 68% of the signal seen for the freshly prepared sample after 1 month of storage at − 20 °C, which may be due to the longer storage time as well as interactions with coextracted sample matrix compounds. In addition, Helmeczi et al. investigated the stability of the extracted 25(OH) D-PyrNO derivatization product after 2 freeze-thaw cycles when stored at − 80 °C [26]. The derivatization product was stable and no degradation was observed. The exact storage time of the extracts at − 80 °C during the two freeze-thaw cycles was not mentioned in the study [26].
Noteworthy is the difference between the observed stabilities of products of PTAD and PTAD-Ac. The results clearly show that the additional acetylation improves the stabilities from 28-46% to 75-82% after 1 month of storage and from 15-25 to 56-65% after 3 months of storage.
In summary, the most stable extracted metabolite products were formed by Amplifex, which after 1 month of storage were within 80-89% of the t = 0 value and after 3 months were within 65-86% of t = 0. The 24,25(OH) 2 D 3 -Amplifex products were the most stable in our study.
Adding a stable isotope internal standard for each metabolite would likely compensate for accuracy issues originating from analyte degradation, assuming that the degradation rates are the same for all species, which is a reasonable assumption. However, vitamin D metabolites present in the sample at very low concentration levels, e.g. those seen for 1,25(OH) 2 D 3 or 3α-25(OH)D 3 , which undergo very high degradation (˃30%) at the same time (e.g. PTAD or PyrNO derivatization products), can potentially diminish to very low concentration levels in the stored extracts, which are then below the validated calibration range or even below LOQ or LOD. The assay then is not fit for purpose anymore.
The EMA and FDA guidelines prescribe that analyte stabilities in the samples should be within ± 15% of the values at t = 0 after the chosen storage conditions and time period [9,10]. Considering this criterion in our comparison study, it allows us to identify the optimum derivatization reagent for each vitamin D 3 metabolite, if they were to be measured individually (Fig. S1 A-E in the Supplementary Material provide a complete breakdown of the data). For 1,25(OH) 2 D 3 , Amplifex and INC the most stable products were provided

Conclusions
Stability studies should investigate the preferred storage conditions over storage periods that are equal or exceed those applied to real samples. Usually, the only criterion for the stability that guidelines specify is that the value after a certain period of time or under certain conditions of storage should be within ± 15% of the value at the start (t = 0) [9,10]. However, a laboratory can set more strict criteria, such as ± 10% or even ± 5% to ensure the validity of their analytical method [7]. Taking the above into consideration, the most stable chemical derivatives of the investigated vitamin D 3 metabolites were formed by the Amplifex reagent, which exhibited good stability within the ± 15% range during the first 1 month of storage; for some analytes even after 3 months. Similarly good stability was observed for some vitamin D metabolite derivatization products of DMEQ-TAD, FMP-TS, INC and PTAD-Ac after 1 month (± 15%). Very poor stability, even after the first month, was seen for the PTAD and PyrNO derivatization products. None of them complied with the given stability criterion.
In comparison to the stability of native, protein-bound vitamin D metabolites in blood serum or plasma, however, these values are somewhat poor, showing that derivatized metabolites are generally much less stable and should not be stored for extended periods of time at − 20 °C. Storing fully derivatized sample extracts at − 20 °C should only be used as an emergency measure in case of instrument failure or if extracts are to be used within a short period of time, e.g. within 1 week. Hollis reported no detectable degradation for 25(OH)D and 1,25(OH) 2 D in pooled human samples stored for over 10 years at − 20 °C [28]. Serum samples stored for 6, 24 or 40 years are reported to have been re-analysed with no significant effect on 25(OH)D levels [20,21]. Even under extreme storage conditions, serum 25(OH)D 2 and 25(OH)D 3 showed very small to no degradation unless they were subjected to prolonged exposure to direct sunlight [29].
In summary, extracts from serum samples containing vitamin D 3 metabolites can be remeasured after 1 or even 3 months storage at − 20 °C, if the proper derivatization reagent is chosen and a small degree of degradation is accepted. An appropriate stable isotope internal standard will then compensate for the degradation in most cases. Storage of derivatized sample extracts is important as it allows remeasurement of selected samples or entire batches if, for example, the instrument breaks down or samples are shipped to another laboratory. It also permits automated, parallel sample processing, including the derivatization step, of large sample numbers, which requires safe storage of samples prior to instrumental analysis.