Development of molecularly imprinted Acrylamide-Acrylamido phenylboronic acid copolymer microbeads for selective glycosaminoglycan separation in children urine

Materials

All chemicals were obtained from Sigma-Aldrich (USA) and all the solutions were chosen as MS grade. Steel chromatography columns were obtained from Waters and emptied their filling materials to fill MIPs (4.6×1.6 mm). Mass spectrometry measurements were carried out in Tandem Quadrupole Xevo TQD Tandem MS systems. Urine samples were collected with the approval of Ege University Clinical Researches Ethical Committee (permit 14/05/2015 15-4.1/17) from children of 0–3 year-old healthies to obtain high GAG excretion [20]. Parents of the children were informed about the purpose of the study and permission was obtained by signed forms. Urine samples (24 h urine-all the urine excreted over a 24-h period) were collected and pooled to obtain more natural GAG in urine from children who did not have any kidney diseases, idiopathic hypertension, urinary system infection or type-1 diabetes.

Methods

MS optimization and production of GAG fragments

All fragments were ionized in ESI+. The degraded GAG fragments which were dissolved in 10 mM anhydrous ammonium acetate were infused directly into the mass spectrometer to monitor m/z values of the fragments. Ionization parameters were set by considering the maximum peak energy of the fragments in the mass detector, as follows: 2.7 eV capillary voltage, 350°C desolvation temperature, and 650 L/h desolvation gas flow rate. Individual parameters of the fragments in MRM and their collision energy levels are given in Table 1. MS/MS optimization for the concentration of fragments of each GAG was carried out by using the retention time in the mass detector and measuring the response area of mass chromatograms.

From the MS/MS results, three calibration curves were prepared for each GAG concentrations <1000 ng/mL (Figure 1A and B) and GAG concentrations in samples were later calculated using these calibration curves.

Figure 1:

(A, B) Calibration curves.

These calibration curves were prepared with GAG concentrations between 62.5 ng/mL and 1000 ng/mL. Response areas were proportional to concentrations. As a standard equation of the calibration curve (y=mx+n), y represents response, x represents the concentration of GAGs ([DS], [CS]). (A) MRM response areas of CS’s 138.1 daughter ion of the 394.2 parent ion. (B) MRM response areas of DS’s 258.2 daughter ion of the 448.2 parent ion.

SPE column filling and GAG elution

In the polymer synthesis process, we primarily considered the sulfate, carboxyl and hydroxyl groups of GAGs when choosing acrylamide, a basic-neutral monomer to attract negatively charged groups and AAPBA, which has boronic acid residue to attract sugars via diol bond formations [22]. Because of the high molecular mass of GAGs, a surface imprinted polymerization technique was chosen to prevent in case of GAGs would be embedded inside the polymers. Therefore, precipitation polymerization was chosen to form uniform spheres as the GAG imprinted polymer shapes [23]. The elasticity and uniform structure of the spheres make them suitable for high pressure chromatographic separations. It was also easy to perform the column filling process because of this elasticity and uniformity of the material. Polymers were investigated by FTIR to characterize the polymer characteristics (Supplementary Figure 3). The green curve is the GAG imprinted polymer, ~3500 represents the OH stretching, ~2900 represents aliphatic C-H stretching. As it can be seen in supplementary data, OH bond signals were decreased because the imprinting process has occurred via hydrogen bonds. The pale blue represents, GAG-removed polymer, 3500 OH stretching signal was increased because of the free OH groups. ~1720 cm⁻¹ for sugars bound to the phenylboronic acid residues and C=O ester peaks. Pale blue represents CS and DS, purple curve represents GAG removed polymer, where it can be seen OH stretching, the red is only acrylamide polymer and dark purple is control polymer. After polymerization process the polymer-filled steel columns were tested under pressurized liquid chromatography system until the flow rate was steady. Before applying the extraction solvent, first acetonitrile and then absolute methanol were flowed through the column. Methanol that was collected from the outlet side of the column was examined for GAG content, and MS/MS results showed that GAGs were not removed from the polymer by using methanol. We then infused methanol:acetic acid (4:1) through the MISPE column to remove the individual GAG. GAGs (CS, DS and HS) were eluted and chemically degraded for MS/MS analysis.

Column selectivity, repeatability and limit of detection

Urine samples were collected between 0 and 3 years of age (n=10) for higher level of glycosaminoglycans in urine and concentrated [6]. The samples were pooled to perform same concentration of samples to analyze the methods optimization parameters. These urine samples were applied to the non-imprinted columns to examine whether they contained any fragments with the same m/z ratio(s) as GAG standards, as it is shown in Figure 2, after MeOH:AcCOOH application, the chemically degraded extractions were investigated by MS/MS analysis and no response was obtained for column affinity (Figure 2). Two different imprinted polymers of CS and DS were compared to NIP polymer and retention times were given in Supplementary file (Suppl. 1.). This comparison was carried out by analyzing outlet samples of the prepared column, before extraction process.

Figure 2:

Comparison of relative responses from non-imprinted and molecularly imprinted polymers.

Non-imprinted polymer is used as a control polymer to measure the separation specificity of the imprinted polymers. In the eluent solution of methanol DS, CS and HSA were recovered less than 7%, which indicated that imprinted polymers had specific affinity to GAGs.

After examination of selectivity, 62.5 ng/mL standard samples (n=10) were applied to the columns to examine column performance tests. In a 300 μL/min flow of MeOH:AcCOOH (4:1), GAG eluents were analyzed for MS/MS analysis (Table 2).

The reusability of the imprinted columns was examined and the results showed that reproducibility was good (Figure 3A and B); coefficient of variation (CV) results were 5.0% and 3.9%, respectively, for the CS and DS.

Figure 3:

(A, B) The reusability of molecularly imprinted polymers for every 500 ng/mL GAG application consequently.

(A) CS imprinted column material reusability capacity. (B) DS imprinted column reusability capacity.

The amount of glycosaminoglycans that bound to the polymers (Q%) was defined according to the equation;

Q%=C0−CsCS

where C₀ is the initial concentration of sugar and C_s is the concentration of glycosaminoglycan in the urine after rebinding. Q_max is and Q was calculated as 332 ng/mL for 500 ng/mL DS: 332 ng/mL, CS: 432 ng/mL equilibrium constant. The quantity of binding sites of the MIPs was examined by using Scatchard analysis. Scatchard equation is as follows;

Q[C]=Qmax−QKd

where Q is the amount of template bound to the polymer, Q_max is the apparent maximum number of binding sites, KD is the equilibrium dissociation constant and [C] is the equilibrium constant of template [24]. K_d was calculated as for DS=0.016 and CS=0.0123. In most clinical mass spectrometry-chromatographic techniques, internal standards (IS) are used to correct for the loss of analyte during sample preparation or sample inlet. For this process, a known amount of IS was added to every sample, both calibrators and unknowns to base the calibration on the ratio of response between the analyte and the IS (by calculating response factor.) instead of basing it on the absolute response of the target analyte. However, the requirements of the IS (that it should be very similar to the analyte with a similar retention time and similar derivatization as well as it should be stable, not interfering with the sample components) to be selected is often a challenge. In this study, chemically degraded GAG fragments were added to see matrix effect in spiked urine. In spiked urine samples the matrix effect was found to be <5% and real sample GAG concentrations calculated by using this data (Table 3). This indicates that in our method, it is not necessary to use an IS (which means we can avoid the problem that chemically degraded fragments would have to be examined by NMR for IS synthesis).