A Novel 3D-printed Centrifugal Ultrafiltration Method Reveals in vivo Glycation of Human Serum Albumin Decreases its Binding Affinity for Zinc

Abstract

Plasma proteins are covalently modified in vivo by the high-glucose conditions in the bloodstreams of people with diabetes, resulting in changes to both structure and function. Human Serum Albumin (HSA) functions as a carrier-protein in the bloodstream, binding various ligands and tightly regulating their bioavailability. HSA is known to react with glucose via the Maillard reaction, causing adverse effects on its ability to bind and deliver certain ligands, such as metals. Here, the binding between in vivo glycated HSA and zinc (Zn²⁺) was determined using a novel centrifugal ultrafiltration method that was developed using a 3D-printed device. This method is rapid (90 minutes), capable of high-throughput measurements (24 samples), low-cost (<$1.00 USD per device) and requires lower sample volumes (200 μL) compared to other binding techniques. This device was used to determine an equilibrium dissociation constant between Zn²⁺ and a commercially obtained normal HSA (nHSA) with a glycation level of 11.5% (K_d = 2.1 (± 0.5) × 10⁻⁷ M). A glycated fraction of the nHSA sample was enriched (gHSA, 65.5%) and isolated using boronate-affinity chromatography, and found to have a 2.3-fold decrease in Zn²⁺ binding-affinity (K_d = 4.8 (± 0.8) × 10⁻⁷ M) when compared to the nHSA sample. The level of glycation of HSA in control plasma (13.0% ± 0.8, n=3 donors) and plasma from people with diabetes (26.9% ± 6.6, n=5 donors) was assessed using mass spectrometry. Furthermore, HSA was isolated from plasma obtained in-house from a person with type 1 diabetes and found to have a glycation level of 24.1% and K_d = 3.3 (± 0.5) × 10⁻⁷ M for Zn²⁺, revealing a 1.5-fold decrease in binding affinity compared to nHSA. These findings suggest that increased levels of glycated HSA result in reduced binding to Zn²⁺, which may have implications in complications associated with diabetes.

Graphical Abstract

Introduction

Diabetes Mellitus (or diabetes) is a growing epidemic, affecting 34.2 million people in the United States alone.¹ Diabetes is the seventh leading cause of death, and is the number one cause of kidney failure, lower limb amputations and adult blindness in the United States.² Diabetes is a pathology characterized by abnormally high glucose levels, due to destruction of the insulin-producing pancreatic β-cells (as in type 1 diabetes, or T1D) or through a process whereby cells appear to have developed an insulin resistance (type 2 diabetes, or T2D).³ In either case, insulin is given exogenously to facilitate glucose control. Despite this therapy, high concentrations of glucose in the bloodstream persist, leading to glycation of proteins and subsequent complications associated with diabetes that are often attributed to poor microvascular blood flow.^4,5 Due to high glucose concentrations found in diabetes, hemoglobin and other plasma proteins become glycated.^6,7 In fact, diagnosis of diabetes sometimes involves measuring elevated hemoglobin A1C (HbA1C) levels in the bloodstream, or glycated hemoglobin.⁶

Human serum albumin (HSA) is the most prominent plasma protein in the blood, comprising 60% of the total protein content.⁷ Often referred to as a “carrier-protein,” HSA is well-known for its extensive ability to bind and transport various hormones, drugs, fatty acids, and metal ions throughout the body.^7,8 Under hyperglycemic conditions, such as those observed in diabetes, HSA interacts covalently with glucose molecules via the Maillard reaction. This process, known as glycation, has been reported to alter HSA’s ability to bind various ligands.⁸ Altered ligand binding abilities of HSA have been attributed to drug toxicity, cardiovascular disease, formation of reactive oxygen species, oxidative stress, and damage to DNA.^9–14

HSA contains a high-affinity binding pocket for zinc (Zn²⁺), an essential transition-metal in the bloodstream and a necessary micronutrient for such physiological processes as wound healing, blood clotting, thyroid function, and immune system function.^15–19 Zn²⁺ is essential for the function of more than 300 enzymes in vivo, contributing to its indispensable role in regulating homeostasis.¹⁹ In blood plasma, approximately 75–80% of Zn²⁺ is bound to HSA,²⁰ thereby tightly regulating the availability of Zn²⁺ to other proteins and cells. HSA also facilitates the uptake of Zn²⁺ into endothelial and red blood cells.²¹ Alterations in HSA-Zn²⁺ binding are reported to lead to increased levels of free, or unbound, Zn²⁺ in the bloodstream, which is believed to prompt diabetic complications associated with alterations in immune function, thrombosis, poor blood flow, hyperzincuria, and neuronal death.^17,22

Recently, it was reported that in vitro glycated-HSA has a decreased ability to bind Zn²⁺.²³ This in vitro glycation method involved incubating HSA in a high-glucose solution in the laboratory for days to weeks followed by the use of MALDI-TOF mass spectrometry to confirm glycation patterns of HSA vary according to glucose concentrations.²⁴ It is also possible that in vitro glycated HSA has a different structure than in vivo glycated HSA, leading to differences in functional characteristics (e.g. ability to bind ligands). Furthermore, like the in vitro studies showing glycation patterns varying as a function of glucose concentration, there may be differences in functional characteristics of HSA between individuals according to their blood glucose levels.⁶ Collectively, measuring the binding interaction between Zn²⁺ and patient-derived HSA could provide useful information about the pathogenesis of complications associated with forms of diabetes.

Here, we investigate the binding characteristics between Zn²⁺ and in vivo glycated HSA. A custom ultrafiltration device was developed using 3D-printing technologies to accurately determine the affinity between Zn²⁺ and various HSA samples. Ultrafiltration is an approach that uses pressure to drive low molecular weight analytes through a size exclusion membrane, while retaining high molecular weight compounds.²⁵ As depicted in Fig 1a, this method enables separation of a small aliquot containing unbound ligand from the bulk sample containing the bound ligand-receptor complex, thus allowing calculation of an equilibrium binding constant. Contingent upon 2–10% of the initial volume in the ultrafiltrate, equilibrium should not be affected.²⁵ Although commercial ultrafiltration devices are available, we propose that through the customizable nature of 3D-printing, these devices are well suited for measuring a wide variety of ligand-protein interactions with differing molecular weights, such as metal-protein binding experiments.

Figure 1. Characterization of nHSA and T1D HSA Through Mass Spectrometry.

Electrospray ionization time-of-flight mass spectrometry was used to analyze samples from control HSA (nHSA) in (a) and T1D HSA in (b). Depicted, HSA is combined with glucose (G = +162 Da), cysteine (C = +119 Da), acetonitrile (A = +41 Da) and sulfinic acid (SO₂H = +31 Da). The nHSA sample is approximately 11.5% glycated and the T1D sample is approximately 24.1% glycated.

Methods

Plasma collection

The use of human whole blood was approved by the Institutional Review Board (IRB) at Michigan State University, along with the written consent of individuals upon collection. All blood samples collected in-house were obtained via venipuncture into heparinized tubes. Whole blood was centrifuged at 500g for 10 minutes and plasma was collected into a 15 mL tube. The plasma was subsequently stored at −20°C until the day of use.

Isolation and Characterization of Human Serum Albumin from Plasma

HSA was isolated from the plasma of an individual with T1D using a modified method similar to that described in the literature.²⁶ Diluted plasma was placed in a tube containing magnetic beads coated with an antibody specific for human serum albumin (PureProteome™ Millipore, Burlington, MA) and incubated on an orbital shaker. The beads were separated from the rest of the solution using a magnet, and the HSA that was bound to the antibody was removed by the addition of glycine buffer (0.1 M glycine, pH 3.0). This solution was then placed into an Amicon Ultra-15 ultrafiltration centrifugal filter unit (10 kDa MWCO, Millipore, Burlington, MA), where it was washed seven times with H₂O (distilled and deionized, 18 MΩ, DDW) to ensure all glycine was removed; it was determined through mathematical computation that these number of washes decreased the glycine concentration to sub-nM at maximum, thus ensuring its concentration was less than Zn²⁺. Purified HSA was subsequently lyophilized and stored at −20°C. As seen in Figure 1, the purity of the protein and level of glycation was characterized using electrospray ionization time of flight mass spectrometry, as previously described.²⁷

Separation of enriched glycated HSA using Boronate Affinity Chromatography

Albumin from human serum (lyophilized, fatty acid free, ≥99% by agarose gel electrophoresis) was purchased from Sigma-Aldrich (St. Louis, MO) and is referred to as nHSA (normal HSA) throughout the manuscript. An affinity chromatography technique was used to enrich and isolate a glycated fraction of nHSA. This chromatography technique utilized a glycoprotein enrichment resin (Takarta, Mountain View, CA) containing a boronate stationary phase. The enriched glycated HSA (gHSA) eluate was lyophilized and stored at −20°C until further experimentation. Detailed instructions for this protocol are in the Supplemental Information.

Printing Centrifuge-Enabled Ultrafiltration Device

A J750 PolyJet 3D-printer (Stratasys, Eden Prairie, MN) was used to produce all ultrafiltration devices (Figure 2a). The printer was programmed to print without a bed of sacrificial support-material by accessing the Stratasys Parameter Manager software and changing values under the title of carpet height, carpet protector Z, and improve support thick of pedestal to 0 mm, as previously described.²⁸ The device was fabricated from three stereolithography (.STL) files designed in computer-aided design (CAD) software (Autodesk Inventor Professional, San Rafael, CA). A stacked-printing approach was used to print the device in stacked layers of different materials (Figure 2b) without any support-material, with membranes incorporated directly into the printing-process.^28,29 Size exclusion dialysis membranes (12 kDa, SpectraPor, New Brunswick, NJ) and polycarbonate membrane filters (0.1 mm pore size, 76 mm sheets, Sterlitech Corporations, Kent, WA) were placed between two sheets of wax paper and cut out into circles using a 5 cm diameter steel-hole punch.

Figure 2. Fabrication and Characterization of 3D-Printed Ultrafiltration Devices.

In (a), the principles of ultrafiltration binding experiments are illustrated. A stock solution containing protein and ligand are loaded into the device. The bottom of the device contains a size-exclusion membrane with pores to allow small ligands through, while retaining the receptor or receptor complex. Centrifugal pressure causes an aliquot of unbound ligand to pass through the pores, which represents the concentration of unbound ligand above the size-exclusion membrane. In (b), a schematic is provided displaying the individual layers of the 3D-printed device. A photograph of the device in and out of a centrifugation tube is depicted in (c). The characterization of the device is shown in (d) where the effect of different centrifuge speeds and times were examined by measuring the mass of the ultrafiltrate passing through the membrane. The preferred settings determined were 15,000g for 90 minutes. (n=3, error=s.d.)

The printing materials for this device were carefully chosen based on specific characteristics. For example, VeroClear is a rigid material that allows for structural integrity, inhibiting distortion of the device during centrifugation. This material was used in the device to ensure the device would uphold its structure during centrifugation. Tango+ was incorporated to seal the membranes. This rubber-like material is mildly adhesive and allowed a liquid-tight seal to prevent leaking of sample around the membrane. The devices were designed to fit securely within 1.7 mL centrifuge tubes (Posi Click, Denville Scientific, Metuchen, NJ). The entire fabrication process for a batch of 20 devices can be completed in under 90 minutes. Further details on fabrication can be found in Supplemental Information.

Sample Preparation of HSA with ⁶⁵Zn²⁺

Safety procedures for handling radioactive material followed approved policies and procedure from Michigan State University’s Environmental Health and Safety Department. A stock of 90.98 μM ⁶⁵Zn²⁺ (ZnCl₂,PerkinElmer, Waltham MA) was used in the preparation of all samples. An HSA stock solution was prepared by adding 8–10 mg of lyophilized HSA and dissolving in 1 mL of ultrafiltration buffer (10 mM Tris, 150 mM NaCl, pH 7.4, sterile filtered through 0.22 μm filters (Sigma-Aldrich)). Preceding the preparation of samples, the concentration of HSA in the stock solution was confirmed through the bicinchoninic acid (BCA) assay (ThermoFisher Scientific). Samples were prepared in triplicate totaling 615 μL, containing 15.0 μM HSA and increasing concentrations of radioactive ⁶⁵Zn²⁺, ranging from 2 to 18 μM in ultrafiltration buffer. 3D-Printed ultrafiltration devices were placed in centrifugation tubes and 200 μL of each sample were transferred to their respective devices and centrifuged at 15,000g for 90 minutes. Following centrifugation, the ultrafiltrate was transferred via pipette into another centrifugation tube for quantitation of unbound Zn²⁺.

Zinc Binding Analysis

In all studies reported, the Zn²⁺ was determined as a radioactive ⁶⁵Zn²⁺ species with gamma emission detection. The fast separation using ultrafiltration results in small volume of effluent pushed through the membrane pores. As such, typical methods to determine metal concentration in samples, such as flame atomic absorption spectrophotometry or ICP emission spectrometry, may require extensive dilutions in order to aspirate sample into the atomization/ionization source (the flame or plasma). Here, we employed the ⁶⁵Zn²⁺ radiolabel for measurement by gamma emission due to its excellent limits of detection, but perhaps more so because the effluent driven through the pores could be directly placed into the gamma readout instrumentation without any sample removal from the collection container, preparation, or dilution. External standards were prepared in 10.0 μL aliquots containing concentrations of ⁶⁵Zn²⁺ ranging from 0.125–8.0 μM. Sample and standard tubes were placed in a 2480 WIZARD automatic gamma counter to measure ⁶⁵Zn²⁺ samples for 5 minutes each (1116 keV). The concentration of free ligand was determined by comparing the counts per minute (cpm) of each sample to an external standard calibration curve. The bound ⁶⁵Zn²⁺ was determined by subtracting the concentration of free ⁶⁵Zn²⁺ from the concentration of total ⁶⁵Zn²⁺ added, as previously reported.^27,30 Full binding curves were analyzed using non-liner regression software (SigmaPlot 13.0) to determine the K_d and B_max values.

Results

Fabricating Ultrafiltration Devices

Novel additive manufacturing methods enabled integration of membranes and material changes along the z-axis without addition of sacrificial support material. Normally, PolyJet 3D-printers deposit a sacrificial support material beneath each model (carpet-bed), in void spaces (channels or holes), and whenever there is a change in material in the z-axis direction. This support material is waxy and can contaminate the device or clog the membrane pores, leading to undesirable conditions and possibly altered experimental results. Therefore, the printing parameters were altered to print the device without the addition of sacrificial support material. In addition, a “stacked-printing” method was employed, allowing membranes to be integrated directly into the device, as shown in Figure 2b.²⁹ This printing method is thoroughly described in the Supplemental Information.

The effects of centrifugation-time and g-force on volume of ultrafiltrate pushed through the membrane pores of the device were investigated. Printed ultrafiltration devices were placed into 1.7 mL centrifugation tubes, shown in Figure 2c, and exactly 200 μL of H₂O (distilled and deionized, 18 MΩ, DDW) were added to the top portion of each device. The devices were centrifuged for a predetermined amount of time, and the volume pushed through the membranes was determined by massing the volume of the ultrafiltrate. The preferred centrifuge settings were contingent upon collecting 10–13 μL of ultrafiltrate. This volume is low enough so that equilibrium is not disturbed, yet large enough to allow accurate quantitation of the analyte-ligand in the ultrafiltrate.²⁵ The preferred settings were 15,000g for 90 minutes as shown in Figure 2d.

Zn²⁺ binding to normal and enriched-glycated HSA

The device was characterized by measuring the interaction between commercially obtained normal HSA (nHSA) and Zn²⁺ and creating a binding curve from this data, which is displayed in Figure 3a. The resultant binding curve resulted in binding affinity (K_d = 2.1 (± 0.5) × 10⁻⁷ M) and B_max (17.3 (± 1.2) μM, n = 1.2 ± 0.1) values of nHSA that are similar to literature values.³¹ Mass spectrometry was used to determine an extent of glycation of 11.5%, which is similar to values reported in the literature for this particular HSA.³² Based on the K_d measurement and glycation percentage, it was concluded that the device could be used in subsequent ligand-binding experiments to measure the association of HSA-Zn²⁺.

Figure 3. Zn²⁺ Binding to nHSA and enriched gHSA.

The nHSA binding data is shown in (a), representing a binding curve fit to the data using non-linear regression software (SigmaPlot 13.0), which enabled the calculation of an equilibrium dissociation binding constant (K_d = 2.1 (± 0.5) × 10⁻⁷ M) and stoichiometry (B_max = (17.3 (± 1.2) μM, n = 1.2 ± 0.1) (n=3, error=s.d.). In (b), the Zn²⁺ binding affinity of nHSA is compared to enriched gHSA, showing a 2.3x higher K_d, revealing a significant decrease in binding affinity for the glycated species. (n=5–6, error=s.d. *p<0.05)

To determine if glycation affects the binding of HSA to Zn²⁺, a boronate affinity column was used to isolate an enriched-glycated HSA (gHSA) fraction from the nHSA sample in Figures 3a. That is, the 11.5% glycation is an “average” glycation of the entire sample; there is a wide array of HSA molecules with different glycation percentages. Here, we separated a fraction that was highly glycated. Mass spectrometry results confirmed that enriched gHSA was approximately 65.5% glycated. The enriched gHSA was further subjected to Zn²⁺ binding analysis (full binding curve in Supplemental Information) and the binding stoichiometry of enriched gHSA was found to be statistically equal to the nHSA sample, with a B_max value of 18.0 (± 1.1) μM (n = 1.2 ± 0.1). However, as depicted in Figure 3b, the enriched gHSA showed a 2.3-fold decrease in ligand binding ability (K_d = 4.8 (± 0.8) × 10⁻⁷ M) in comparison with nHSA, confirming that glycation alone decreases the affinity for HSA to bind Zn²⁺. These results indicate that as the percent glycation of HSA increases, the affinity for HSA to bind Zn²⁺ decreases.

Differences in in vivo glycation per HSA Samples

The data in Figure 3b shows the functional difference between two extreme glycation patterns of HSA (11.5% vs 65.5%), although, in vivo, the typical level of glycation is between these values. To better understand typical levels of glycation in vivo, HSA was first separated from control plasma or plasma obtained from individuals with diabetes. Using negative immunomagnetic separation, the HSA in each plasma sample was isolated and the percent glycation characterized by electrospray ionization time-of-flight mass spectrometry. Figure 4 shows the average percent glycation of control samples and samples obtained from people with diabetes. Our analysis shows that HSA derived from people with diabetes is significantly (p<0.05) more glycated (26.9% ± 6.6) than control samples (13.0% ± 0.8). The percent glycation found in the HSA samples agrees with values found in the literature referring to in vivo glycated HSA.³²

Figure 4. Variations in Percent Glycation of HSA Isolated from Plasma.

Percent glycation was determined through electrospray ionization time-of-flight mass spectrometry. Non-diabetic control samples contained 13.0% ± 0.8 (n=3) glycated HSA, whereas samples from individuals with diabetes contained 26.9% ± 6.6 (n=5) glycated HSA. (p<0.05, error=s.d.)

Zn²⁺ binding to isolated T1D HSA

The binding interaction between glycated HSA and Zn²⁺ was determined using HSA separated from plasma of a whole blood donor with T1D. Samples were subjected to the same characterization and binding conditions as previously stated. Mass spectrometry results indicate that the T1D HSA sample was 24.1% glycated, a concentration ranging in between the nHSA and enriched gHSA fractions in Figure 3b. A saturation binding curve was created as seen in Figure 5. As observed previously in both the nHSA and enriched gHSA, the curve saturated at 14.5 (± 0.6) μM (n = 1.03 ± 0.04), further indicating 1:1 binding. In addition, the affinity for T1D HSA to bind Zn²⁺ showed approximately a 1.5-fold decrease (K_d = 3.3 (± 0.5) 10⁻⁷ M) when comparing the results to the nHSA sample. The T1D HSA also shows an approximate 1.5-fold increase in binding affinity when compared to the enriched gHSA sample. Interestingly, the binding affinity between T1D HSA and Zn²⁺ falls between the nHSA and the enriched gHSA values, confirming that T1D HSA has a decreased affinity for Zn²⁺ in vivo.

Figure 5. Saturation Binding Curves between T1D HSA and Zn²⁺.

The ultrafiltration binding results are graphed and the data was analyzed using non-linear regression software (SigmaPlot 13.0). The results indicate a binding affinity of 3.3 (± 0.5) 10–7 M and a binding curve that saturated at 14.5 (± 0.6) μM (n=3–4, error=s.d.).

Discussion

Since the discovery of glycated HSA several decades ago, its altered ligand-binding abilities have been measured and are associated with many diseased states and health complications.^5,7,8,32 Zinc (Zn²⁺) is an important transition-metal in the body and modifications in its bioavailability are known to exist in diabetes and Alzheimer’s disease. However, the binding of in vivo glycated HSA to Zn²⁺ has never been quantitatively determined.

While various methods exist to measure the binding of proteins to smaller ligands (equilibrium dialysis, isothermal titration calorimetry, spectroscopic assays, etc.), ultrafiltration provides a rapid, accurate, cost-effective, and reproducible method for measuring binding interactions.^25,33–35 Ultrafiltration devices are commercially available, although the selection of pore sizes and materials is limited. Here, novel 3D-printing technologies enabled the fabrication of ultrafiltration devices, customized to specifically measure the association between HSA and Zn²⁺. Using additive manufacturing, any sized membrane and material could potentially be integrated into the 3D-printed device, thus enabling measurement of protein-ligand interactions with differing molecular weights and more complex associations involving metals, proteins, peptides, and other biological molecules. Specifically, the association between such transition metals as copper, iron, cobalt, etc., and proteins can be measured through customizing the ultrafiltration devices with different pore sizes to isolate the free metal. Furthermore, it is anticipated that this device could also be employed to measure interactions between 2 or more binding species that, perhaps, do not involve metals. For example, if carefully selecting pore size to match an appropriate antigen, this device could be used to measure binding characteristics of antigen-antibody interactions. The concentration of any free species can then be quantified using various direct methodologies such as ICP-MS, graphite furnace atomic absorption, or, if an appropriate label was included, radiochemical methods. Furthermore, a key feature of the 3D-printed devices is their rapid fabrication (<90 minutes), sometimes immediately preceding the actual measurement portion of the analysis. Finally, the small-scale configuration of the device enabled 24 devices to be centrifuged, simultaneously. Thus, this technique enables the user to construct binding curves containing multiple samples at each concentration. The devices have additional advantages with respect to cost (<20% of commercial devices), time (90 minutes), volumes required (200 μL), and the ability to measure a binding constant between commercially available nHSA and Zn²⁺ (Figure 3a) statistically equal to literature values.³¹

Subsequently, the devices described here were used to determine binding constants between Zn²⁺ and HSA under glycated conditions. Specifically, the enriched glycated (gHSA) fraction of nHSA was subjected to binding analysis, as shown in Figure 3b. Due to glycation being the only experimental factor altered in comparison with the nHSA sample, the data suggests that glycation alone is the major determinant altering the interactions between HSA and Zn²⁺. Recent studies have shown that in vitro glycated HSA may not bind Zn²⁺ at all (or has a dramatically decreased affinity),²³ although the process of modifying HSA with glucose in vitro may be different from that which occurs in vivo. Differences in ligand binding characteristics corresponding to the pattern of glycation have been reported,²⁴ therefore in vitro glycated HSA could have a vastly different structure and binding capacity than glycated HSA found in vivo.

To better understand the physiological relevance of our binding studies, HSA was isolated directly from the plasma of individuals with and without diabetes. The microheterogeneity of each sample was assessed by mass spectrometry. The HSA derived from non-diabetic individuals was composed of approximately 13.0% glycated HSA, whereas HSA derived from individuals with diabetes was composed of approximately 26.9% glycated HSA. Importantly, levels of glycated HSA found in the plasma of individuals with diabetes vary greatly, depending on glycemic control, further demonstrating the physiologic relevance of studying in vivo glycated HSA.

Isolating HSA from the plasma of a donor with T1D allowed for a physiologically relevant sample while determining the effect of glycation on Zn²⁺ binding, shown in Figure 5. Importantly, the percent glycation of the T1D sample (24.1% glycated) was statistically higher than the non-diabetic samples in Figure 4, and the equilibrium binding constant was statistically higher (weaker) than the nHSA sample (11.5% glycated) in Figure 3a. These results strongly suggest that glycation interferes with Zn²⁺ binding in vivo and that the extent of glycation of HSA is directly correlated to its binding affinity for Zn²⁺. This alteration in Zn²⁺ equilibrium may have pathophysiological consequences.

Atherothrombosis is the most common cause of death in patients with diabetes, accounting for approximately 80% of all deaths among people with diabetes.³⁶ Recent reports show that increased levels of free fatty acids (FFA) interfere with the binding between HSA and Zn²⁺, leading to thrombosis complications in individuals with diabetes.^37,38 This proposed mechanism states that increased levels of unbound-Zn²⁺ in blood-plasma reacts with histidine-rich glycoprotein (HRG), causing heparin-neutralization and downstream obstructive blood clots. Our data suggest this effect may be compounded by increased levels of glycated-HSA, leading to even more unbound-Zn²⁺ in plasma of individuals with poorly controlled diabetes. This alteration in homeostasis increases the concentration of free Zn²⁺ for other molecules in the bloodstream to bind, and thus potentially causing activation or suppression of other responses, such as those found in atherothrombosis.

Additionally, the uptake of Zn²⁺ by certain cell types (erythrocytes and endothelial cells) requires binding to HSA.^21,39,40 Decreased HSA-binding may inhibit Zn²⁺ uptake by these cells in individuals with diabetes, which could result in cell dysfunction and microvascular complications. Further studies should examine the interaction between these cells and Zn²⁺ in the presence of in vivo glycated HSA, as well as the downstream effects on cell metabolism and function.

Interestingly, Zn²⁺ homeostasis is also known to be altered in the cerebrospinal fluid of patients with Alzheimer’s disease (AD), and is hypothesized to have pathological consequences.^41,42 Several studies have reported that unbound-Zn²⁺ reacts with Aβ peptides to form senile plaques, one of the hallmark features of AD.^37,43,44 Additionally, increased unbound-Zn²⁺ has been reported to act as neurotoxin in AD.^45,46 Researchers have recently referred to AD as Type 3 diabetes, as glucose metabolism in the brain is impaired, leading to increased levels of glycated proteins in the cerebrospinal fluid.^47,48 Increased levels of glycated-HSA in the CSF may result in increased levels of unbound-Zn²⁺, causing detrimental plaque formation. Collectively, the data shown here may help explain the pathophysiological role of glycated HSA in diseases associated with Zn²⁺ dyshomeostasis, such as diabetes and Alzheimer’s Disease.

Significance to metallomics.

Expanding upon current research investigating zinc (Zn²⁺) binding to in vitro glycated human serum albumin (HSA), this research utilizes 3D-printed technologies to measure the binding characteristics between in vivo glycated HSA and Zn²⁺. This work provides a platform for accurately measuring protein-metal interactions using a customizable 3D-printed ultrafiltration device, while utilizing proteins isolated from whole human blood. Furthermore, in vivo glycated HSA shows an approximate >1.5-fold decrease in Zn²⁺ binding affinity. This improves our understanding of Zn²⁺ equilibrium in vivo and may help explain diseases associated with Zn²⁺ dyshomeostasis, such as diabetes and Alzheimer’s Disease.