FXII Contact Activation Products Have an Inhibitory Effect on αFXIIa

Abstract

It is accepted that the contact activation complex of the intrinsic pathway of blood coagulation cascade produces active enzymes that lead to plasma coagulation following biomaterial contact. In this study, FXII was activated through contact with hydrophilic glass beads and hydrophobic octadecyltrichlorosilane (OTS)-modified glass beads from neat buffer solutions. These FXII contact activation products generated from material interaction were found to suppress the procoagulant activity of exogenous αFXIIa, and this inhibition was dependent on surface wettability and the concentration of exogenous αFXIIa. Higher relative inhibition rates were generally observed at low concetrations of αFXIIa (1-2 μg/ml) while both hydrophobic and hydrophilic materials showed similar inihibtion levels (~39%) at high concentrations of αFXIIa (20 μg/ml). The presence of prekallikrein in the activtion system increased the amount of FXIIa produced during FXII contact activation, and also suppressed the apparent levels of inhibitors on hydrophilic surfaces, while having no effect on apparent levels of inhibitors on hydrophobic surface. The combination of FXII contact activation products and activator surfaces was found to dramatically increase inhibition of αFXIIa activity compared to the activation products alone, regardless of activator surface wettability and the presence of prekallikrein. This finding of inhibitors in the suite of proteins generated by contact activation provides additional knowledge into the complex series of interactions that occur when plasma comes into contact with material surfaces.

1. Introduction

With the increasing demand for blood contacting medical devices in cardiovascular healthcare, biomaterials exhibiting good hemocompatibility have been an essential enabling technology. However, even after decades of focused research, thrombosis remains a significant barrier to development and implementation of advanced blood-contacting medical devices^1,2. Clearly, an improved understanding of the molecular basis of blood coagulation is essential to the prospective engineering of advanced cardiovascular biomaterials. Blood hemocompatibility is a multicomponent process, and the formation of blood clots and thrombus arises from both the protein mediated plasma coagulation pathways and blood thrombosis arising from both proteins and platelets. The intrinsic pathway of the plasma coagulation cascade, trigged by blood contact with artificial materials, is an important aspect of the hemocompatibility of biomaterials^3-7.

The initiation of the intrinsic pathway is thought to be the activation of blood zymogen Hageman factor (FXII) into an active enzyme form FXIIa through a surface-contact activation complex, and in fact this process is easily observed during in-vitro experiments^8-11. Blocking FXIIa activity should reduce induced blood coagulation leading to thrombosis, and FXII has been an increasing important therapeutic target in thromboembolic disease^12,13 and development of new anticoagulation strategies using direct FXII inhibitors ¹⁴. Negatively charged artificial or biological surfaces, such as kaolin, celite, glass, dextran sulfate, misfolded protein aggregates, collagen, nucleic acids, and platelets, are generally regarded as specific initiators of the coagulation cascade^15-18. Binding of FXII to anionic surface is modeled as inducing a conformational change in proteins, resulting in protein cleavage to autoactivate FXII to FXIIa ¹⁹. FXIIa can in turn cleave prekallikrein bound to the surface with high molecular weight kininogen (HMWK) and facilitate more FXIIa formation through reciprocal activation. FXIIa can also hydrolyze FXII by autohydrolysis. Ultimately, FXIIa activates factor XI (FXI) bound at the surface as complex with HMWK to generate FXIa, leading to propagation of subsequent coagulation cascade reactions^8,20,21. There is evidence to show that all three reactions do not occur to the same extent. Chatterjee et al. found that the reciprocal activation was the principal contributor to FXIIa generation (~75%) while autoactivation produced ~25% of the total FXIIa²². Autohydrolysis is found a facile reaction in neat buffer solutions of FXIIa but this is not a significant contributor to FXIIa in plasma²³. Similar observations reported by other investigators^24,25 suggested that autohydrolysis of a zymogen by its activated enzymatic form can be negligible when compared with cleavage of by other enzymes.

Surface-contact autoactivation of FXII is dependent on the surface properties of materials and protein composition in the fluid. Early experimental results showed that plasma coagulation activation was more efficient by contact with anionic^26,27 or hydrophilic surfaces^28,29, leading to a general conclusion that contact activation of FXII was specific to anionic hydrophilic surfaces. However, other studies have demonstrated that hydrophobic and hydrophilic surfaces have nearly equal autoactivation properties in neat buffer solutions of FXII³⁰. Contact activation of FXII in neat-buffer solution was found a strong function of activator surface energy, which exhibited a parabolic profile when scaled as a function of surface energy, with nearly equal activation observed at both extremes of activator water wettability and falling sharply through a broad minimum for materials with water contact angles in the range of 55°< θ <75° ³¹. Contact activation of blood plasma is also demonstrably dependent on plasma-volume-to-activator-surface-area ratio. Rapid decreases in coagulation time were observed on a variety of activators with increasing activator surface areas ^32-34. However, contact activation of FXII appears far less dependent on surface area in buffer, protein cocktail, serum and FXI deficient plasma, suggesting the activator surface-area dependence observed in contact activation of plasma coagulation does not solely arise at the FXII activation step of the intrinsic pathway³³.

Despite these observations about contact activation of FXII and plasma coagulation, including surface energy dependence, surface area dependence and protein composition dependence, an interesting phenomenon is that coagulation time of plasma or FXIIa yield in buffer solutions are always limited and a plateau is generally reached where coagulation can occur no faster, regardless of how much surface area or FXIIa is present in the system. In fact, the apparent FXIIa yield from FXII contact activation is small, generally less than 10% of the original FXII content³⁵. This leads to the concept of some sort of autoinhibition in FXII contact activation, where the FXIIa generated somehow inhibits production of additional FXIIa³⁶. In contrast to FXII contact activation in blood coagulation cascade, far less is known about this potential autoinhibition function.

The biochemistry of contact autoactivation of FXII is complex. Cleavage of FXII results in generation of mixture of fragments with procoagulant activity such as αFXII_a (HFa or factor XII_HMW) and βFXII_a (HFf or factor XII_LMW)^24,37-39. Other minor cleavage products of intermediate molecular weight between αFXIIa and βFXIIa may be present during FXII autoactivation but their procoagulant activity and the related distribution of these fragments are unknown. Golas et al.³⁵ construed autoactivation as a biochemical reaction and assigned the term FXII_act to represent a suite of protein fragments arising from autoactivation of FXII in buffer solution, and which includes all protein fragments produced by contact activation of FXII, such as procoagulant fragments (protease activity that induces coagulation of plasma) and amidolytic fragments (cleavage of amino acid bonds in s-2302 chromogen but unable to produce blood clotting). It was hypothesized in these studies that there may be as yet unknown fragments produced that may be responsible for suppressing further activation of FXII.

This work investigates the auto-inhibitory phenomena of FXII activation following contact with anionic/hydrophilic and hydrophobic activators in neat buffer so that the effects of other plasma proteins or potential inhibitors of FXII contact activation can be removed during study. The procoagulant activity of products of FXII activation was assessed and used to indicate the yield of FXIIa. The results showed that the products of FXII arising from contact activation with either hydrophilic or hydrophobic surfaces inhibited the procoagulant activity of exogenous αFXIIa, and the inhibition level is moderated by both the presence of prekallikrein and surface wettability of the materials. The findings confirmed our previous hypothesis about the presence of an autoinhibitor within FXII contact activation products and provide further insight into the mechanisms of FXII autoactivation that will be useful in the surface engineering of materials with improved hemocompatibility.

2. Materials and Methods

2.1. Plasma and coagulation proteins

Citrated human platelet-poor plasma (PPP) was prepared by pooling 5 units of salvaged human plasma (outdated within 5 days of expiration) from the blood bank at the Pennsylvania State University Milton S. Hershey Medical Center. PPP preparation and storage were followed as the previous description with minor modifications⁴⁰. Briefly, the pooled PPP was centrifuged two times for 20 min at 1500 g and 25 °C. The plasma supernatant was then stored in 12 ml aliquots in 15-ml polypropylene tubes (VWR) at −20 °C and thawed for ~20 min in a 37 °C water bath prior to use.

Human FXII, αFXIIa, and prekallikrein were used as received from Enzyme Research Laboratories (South Bend, IN). The purities of human FXII and prekallikrein were determined by the vendor using SDS-PAGE gels, and were >95% based on the certificate of analysis. The complete activation of αFXIIa was observed on SDS-PAGE. The activity of both FXII and αFXIIa was specified by the vendor in mg/ml and traditional units of plasma-equivalent-unit-per-ml (PEU/ml). Phosphate buffered saline (PBS) (150 mM NaCl, pH 7.4) purchased from Sigma Chemicals was prepared using water from a Millipore Simplicity 185 System incorporating dual UV filters (185 and 254 nm) to remove carbon contamination. PBS was used to prepare FXII and αFXIIa solutions.

2.2. Preparation and characterization of material activators

Soda lime glass beads with diameter of 425-600 μm (Cat.# G9268, Sigma Aldrich) were used in clean or silanized forms as the contact activators. The nominal specific area of glass beads was 5×10⁻³ m²/g and 100 mg of beads (equivalent to 500 mm²) was used to activate FXII in this work. The glass was prepared by first cleaning in aquaregia solution (HNO₃ : HCl = 1:3) for 2 h, rinsing with Millipore water and then cleaning in piranha solution (30%H₂O₂:concentrated H₂SO₄ = 1:4) overnight, followed by another thorough rinse with Millipore water. Clean glass beads were dried in the oven overnight. Glass surfaces treated in this manner were found to be fully water wettable and designated as hydrophilic surfaces for this work. Octadecyltrichlorosilane (OTS, Gelest Inc., Morrisville, PA) was used to create glass particles with hydrophobic surface character. Cleaned glass was incubated in 5% OTS in chloroform for 1.5 h with occasional shaking to facilitate uniform chemical treatment for all particles. Silanized samples were then rinsed 3× with chloroform before drying overnight in a vacuum oven at 110 °C. The OTS-treated glass particles were nonwettable and designated as hydrophobic surfaces. To characterize the surface wettability of glass particles, witness sample glass coverslips were treated using identical methods.

The water wettability of glass coverslips was determined by sessile drop measurements of the advancing water contact angle (θ) using a Krüss contact angle goniometer. The cleaned glass coverslips showed a fully wettable surface with water contact angle <10°, while the OTS treated glass coverslips show a very hydrophobic surface with contact angle ~110°.

2.3. Plasma coagulation time assay

An in vitro assay was used to measure the plasma coagulation activity in terms of coagulation time (CT), defined as the time from activation of the coagulation cascade to the appearance of visible clot. The assay has been described in previous publications^22,41. Briefly, 0.5 ml of plasma (PPP) was recalcified with 0.1 ml of 0.1 M CaCl₂ and mixed with a known dose of procoagulant (material activators or αFXIIa) in a 1.5 mL polystyrene semi-micro cuvette (BRAND+CO KG, Germany). The volume was adjusted by adding 0.01M PBS to obtain a 1 ml solution with a 1:1 dilution of plasma in buffer. The order in which reagents and solid materials were added was varied depending on the assay type, with recalcification of plasma always occurring last to ensure a common zero-time CT. The cuvettes were capped with parafilm and rotated at 8 rpm on a hematology mixer, and the corresponding CT was recorded.

2.4. Contact activation of FXII in buffer solution and inhibition of activity of exogenous αFXIIa

Contact-activation of FXII was performed in a polypropylene micro-centrifuge tube (1.5 ml, VWR) containing 100 mg of either hydrophilic (clean glass) or hydrophobic (OTS-coated) glass beads. Each tube contained 1200-1500 μl of FXII at 30μg/mL (physiological concentration) and was capped. The tube was incubated at room temperature on a hematology mixer with rotation at 8 rpm. After the beads and FXII interacted for the desired time, the tube was centrifuged at 1000 rpm for 30 s to obtain the supernatant following FXII contact activation. The material obtained following FXII reaction with particle activators will be referred to as contact activation products throughout the manuscript. For purposes of this work, we will refer to the procoagulant enzyme created during autoactivation as FXIIa activation products, distinguishing it from the exogenous αFXIIa added during subsequent steps. In the case of experiments to test the effect of prekallikrein on FXII contact activation and for inhibition of FXIIa activity, prekallikrein (final concentration 20 μg/ml) was added to FXII solutions (30 μg/ml) with bead activators at the beginning of the contact activation process.

To test for suppression of αFXIIa activity, a 150 μl aliquot of the FXII contact activation products was incubated with exogenous αFXIIa at 1, 2, 5, 10, and 20 μg/ml for 10 min at room temperature. After this reaction, the mixture of material-activated FXII activation products and the purified exogenous αFXIIa was used to assess the CT using the plasma assay as described above.

2.5. Properties of FXII contact activation products

The hydrophilic and hydrophobic bead activators (100 mg) were incubated with FXII solution (30 μg/ml) in PBS buffer for up to 2 h. At desired time points, 50 μl of solution (n=3 for each sample) was removed for the plasma coagulation time assay. To test for inhibitory activity of FXII contact activation products, concentrated exogenous αFXIIa was added into the FXII contact activation products after 30 minutes of incubation, to yield a mixture with αFXIIa concentration of 1 μg/ml, and after 70 minutes of incubation addition (40 minutes after the first addition), additional αFXIIa was added to yield a final αFXIIa concentration of 2 μg/ml. These 3 phases are designated as Periods I, II and III, respectively. Plasma coagulation assays from Period I illustrate the procoagulant activity of the contact activation products alone, while samples of the mixture of contact activation products and exogenous αFXIIa from Periods II and III were used in the plasma coagulation assay to determine the apparent activity of the mixture of contact activation products and exogenous αFXIIa. Purified αFXIIa at both 1 μg/ml and 2 μg/ml (final concentration in solution) were used in the plasma coagulation assay alone to determine an expected baseline coagulation time for these enzyme activators.

2.6. Suppression of αFXIIa by FXII contact activation products

Similar as in section 2.4, FXII (30μg/ml) in buffer solution was reacted with hydrophilic or hydrophobic bead activators for 1 h but in this case, prekallikrein (20 μg/ml) was included in some samples. 150 μl of bead-free coagulation activation products was removed and incubated with exogenous αFXIIa (final concentration of 5 μg/ml) for 10 min, after which the procoagulant activity of the coagulation activation products and αFXIIa was measured by the CT assay. Simultaneously, αFXIIa (final concentration of 5 μg/ml) was also added into the remaining FXII solution and beads mixture, and incubated for an additional 60 min. The procoagulant activity of the solution containing both the contact activation products and the exogenous αFXIIa was measured using CT assay.

2.7. Statistical analyses

Data are presented as mean values ± standard deviation from at least 3 independent measurements. Statistical analyses were performed by ANOVA using InStat software (GraphPad software). The differences were considered statistically significant for p<0.05.

3. Results and Discussion

3.1. FXIIa assay by titration of exogenous FXIIa in PPP

Titration of exogenous αFXIIa into PPP was used to generate a standard curve to measure the procoagulant activity of material-activated FXII. Figure 1 shows the decreasing CT values arising from increasing amounts of exogenous FXIIa addition into PPP. Over this wide range of FXIIa concentrations (10⁻⁴ – 10 PEU/ml), the trends between CT and [FXIIa] generally fit the Michaelis-Menten enzyme kinetics-based model for material-induced blood coagulation, which was developed by Guo et al.⁴⁰, and described as $CT=\frac{a[FXIIa]+b}{[FXIIa]+C}$. The parameters $b$ and $c$ account for background activation during the assay, including contact activation due to the polystyrene vial, trace amounts of active enzyme, remnant platelets and platelet-derived microparticles³⁴. With this model, Guo et al. demonstrated that the primary mechanism of activation of coagulation involves contact autoactivation of FXII and kallikrein-mediated reciprocal activation, but the αFXIIa -induced self-amplification of FXII is insignificant in plasma⁴⁰. It should be noted that the coagulation time response to exogenous αFXIIa will vary with different batches of human plasma, resulting in different values of these parameters but showing the same trends. The values of parameters $a,b,c$ shown in Fig. 1a are the best fit of model for this batch of plasma.

Figure 1.

FXIIa titration in PPP showing experimental coagulation time (CT) vs. concentration of exogenous αFXIIa, (a) curve fitting based on model of $CT=\frac{a[FXIIa]+b}{[FXIIa]+C}$ , where a =4.86±1.14, b=5.30±0.95, and c=0.15 ±0.029, R²=0.987, and (b) the linear fitting to the logarithm scale of FXIIa concentration, $CT=-4.505ln[FXIIa]+10.765$, R² = 0.975 (n=3)

Although the above mathematical model described the properties of blood plasma coagulation, using this model to determine the production of FXIIa in FXII contact activation is complicated. In experiments, we found that the coagulation time responses to exogenous αFXIIa often varied across batches of plasma and storage time of plasma. It becomes necessary to calibrate the plasma by titration of αFXIIa frequently, and errors are often observed at extremely high and low concentrations of exogenous αFXIIa. However, as seen in Figure 1b, the mid-range exogenous αFXIIa concentrations (10⁻³ – 5 PEU/ml) show the relationship between CT and FXIIa fits a simple linear regression when FXIIa concentration is set as a logarithmic scale, yielding a regression coefficient R² >0.95 (Fig. 1b). This method has been used to characterize the blood plasma coagulation properties in previous publications^31-33,35. In this work, we used a logarithmic linear calibration curve to determine the production of FXIIa and measure the procoagulant activity after FXII contact activation. To avoid the errors induced by plasma and plasma storage time, αFXIIa titration of PPP was always carried out with each experiment.

3.2. Suppression of αFXIIa activity by FXII contact activation

Figure 2 illustrates the plasma coagulation times for FXII contact activation products following contact with hydrophilic and hydrophobic surfaces. The first 30 minutes is designated Period I (0-30 min, no αFXIIa addition), during Period II (30-70 min) exogenous αFXIIa was added to yield a total concentration of 1 μg/ml, and during Period III (70-110 min) additional exogenous αFXIIa was added to yield a final concentration of 2 μg/ml. The coagulation time of the plasma control with no activators was 40.5 ±1.8 min. Adding FXII contact activation products from glass bead activators led to a drop in the plasma coagulation time. After 5 min of FXII activation by hydrophilic glass activators led to a coagulation time of 29.1 ±0.4 min, while 5 min of activation with hydrophobic activators led to a CT of 30.6 ±0.5 min. The CT values remained essentially constant over the rest of Period I, suggesting that there is a rapid activation of FXII following surface contact with both hydrophobic and hydrophilic surfaces. This is consistent with the result of Golas et al, who found that the FXII activation in buffer solution was effectively instantaneous within the minimum elapsed-time resolution of experiment (about 2 min)³¹. No additional drop in the CT suggests that the process of subsequent FXII contact activation is either terminated or somehow suppressed, both of which would inhibit further production of activated FXII, and which was termed as autoinhibition in FXII activation model³⁶.

Figure 2.

Coagulation time of FXII activation products from hydrophilic and hydrophobic surfaces at different reaction time. Exogenous αFXIIa was added with concentration of 1 μg/ml and 2 μg/ml at periods II and III, respectively. The gray lines indicate the CT corresponding to pure αFXIIa. (n=3)

By looking at the data in periods II and III following addition of 1 μg/ml or 2 μg/ml of αFXIIa, respectively, we observe what appears to be suppression of the expected αFXIIa activity. At 30 min, exogenous purified αFXIIa (1 μg/ml) was added to the FXII reaction solution and incubated for 40 min. Aliquots were taken during this time period and used as activators for the CT assay. A small drop in CT was measured during Period II compared to the CT in Period I, with average changes of CT of 1.17 min, with a 95% confidence interval of (0.48 min, 1.86 min) for the activation products produced by hydrophilic glass surfaces. For the hydrophobic activators, the change in CT following addition of exogenous αFXIIa was1.18 min with a 95% confidence interval of (−0.21, 2.57) (Table 1). Note in Figure 2 that 1 μg/ml αFXIIa in the CT assay resulted in a CT of 22.84 ±0.55 min (dark line in Figure 2), nearly 5 min less than the CT seen for that same concentration of αFXIIa and the contact activation products together. A Period II drop in CT compared to Period I was expected to the addition of both the exogenous αFXIIa in solution and the contact activation products seen in Period I. However, this was clearly not the case as the CT in Period II was significantly higher than the CT of 1 μg/ml exogenous αFXIIa alone, demonstrating that the material-induced FXII contact activation products actually extended the coagulation time and suppressed the procoagulant activity of αFXIIa. Similar effects are seen for the hydrophobic activators, with an even greater increase in the CT seen in Period II.

Table 1.

Statistical analysis of CT (min) of plasma following FXII (30 μg/ml) contact activation with hydrophilic and hydrophobic surfaces. Exogenous αFXIIa was added into FXII solution at period II (30-70 min, 1μg/ml) and period III (70-110 min, 2 μg/ml).

Surface	Period I	Period II (compared to period I)			Period III (compared to period II)
	CT (min)	CT (min)	ΔCT (min)	95% confidence interval of ΔCT	CT (min)	ΔCT (min)	95% confidence interval of ΔCT
Hydrophilic clean glass	28.35±0.71	27.19±0.66	1.17	(0.48, 1.86)	27.44±0.63	0.91	(0.24, 1.59)
Hydrophobic OTS glass	31.70±1.68	30.52±0.87	1.18	(−0.21, 2.57)	30.71±2.05	0.99	(−0.90, 2.87)

These results suggest that the FXII-contact activation products exhibit an inhibitory effect on αFXIIa for both hydrophilic and hydrophobic activators. Note that these incubations are all done in the absence of activating particles so this change in activity cannot be attributed to loss of active enzyme to the activating materials. In Period III, we repeated the sequence of events but in this case, adding an additional dose of αFXIIa for a final concentration of 2 μg/ml. Again, significant inhibition of αFXIIa activity by FXII activation products was observed for both hydrophilic and hydrophobic activators. The average CT after this 2^nd dose of αFXIIa was 27.44±0.63 min and 30.71±2.05 min for activation products from hydrophilic surface and hydrophobic surfaces, respectively. There is no difference in Period III CT values compared to CT values in period II (p>0.05) despite the doubling of exogenous αFXIIa. Comparing to the value of CT corresponding for 2 μg/ml αFXIIa activation at 20.78±0.13 min, the results again demonstrate that incubation of αFXIIa with the FXII contact activation products diminishes the levels of activity and strongly suggest that the products of FXII contact activation inhibit the procoagulant activity of the exogenous αFXIIa.

3.3. Suppression of exogenous αFXIIa activities by FXII contact activation products is dose and surface dependent

Figure 3 compares the procoagulant activity after combining the products of FXII contact activation with different concentrations of exogenous αFXIIa, as well as the activity of pure αFXIIa (without any FXII contact activation products). FXII in PBS buffer at physiological concentration of 30 μg/ml was incubated with either hydrophilic glass beads or hydrophobic OTS-glass beads for 1 h and produced FXII activation products measured as 0.060 ±0.011 PEU/ml and 0.021±0.005 PEU/ml, respectively (the grey bars in Figure 3). Exogenous αFXIIa showed rapidly increasing activity that varied with the amount of αFXIIa added (hatched bars). When αFXIIa at these same 5 concentrations (αFXIIa = 1, 2, 5, 10 and 20 μg/ml) was mixed with the contact activation products, the apparent procoagulant activity of the mixture increased slightly above that of the activation products alone, but was still far less than the activity of the exogenous αFXIIa alone, although with greater increases in apparent activity seen with increasing αFXIIa concentration, as expected (black bars). This loss of activity in the exogenous αFXIIa once again demonstrates that the process of mixing material activation products with αFXIIa leads to a loss in activity, and this loss appeared much more significant at low concentrations of αFXIIa. For example, the activity of the mixture of hydrophilic glass surface activation products with 1 μg/ml αFXIIa was 0.061 ± 0.017 PEU/ml, almost identical to the original activation product level of 0.060 ±0.011 PEU/ml. In essence, the FXII activation products from glass completely removed all the activation potential of exogenous αFXIIa. For glass activation products plus 2 μg/ml exogenous αFXIIa, there was a larger increase in activity, and the effects of the exogenous continued to increase as the concentrations went up, but the activities were always lower than the exogenous αFXIIa alone. These results strongly suggest that the activity of exogenous αFXIIa is inhibited by the products of FXII contact activation, and Table 2 lists the reductions in the activity of αFXIIa (i.e., inhibition capacity) of FXII contact activation solutions. Results show that the relative inhibition rates decrease with increasing αFXIIa levels. The products of FXII contact activation on hydrophilic glass surfaces show higher efficiency in inhibiting activity of exogenous FXIIa at low dose than did those from hydrophobic OTS-glass surfaces, but with increasing levels of exogenous αFXIIa, the inhibition of glass-activated products and OTS-activated products come to the same value of approximately 39%.

Figure 3.

Apparent FXIIa activity of supernatant products of FXII contact activation products with exogenous αFXIIa. (a) FXII contact activation on hydrophilic glass surface and (b) FXII contact activation on hydrophobic OTS glass surface. (n=3

Table 2.

αFXIIa reduction amounts (Δ[FXIIa]) and activity inhibition rate of αFXIIa by FXII contact activation products. (Δ[FXIIa] = [FXIIa]_dose – ([FXIIa]_apparent - [FXIIa]_material), Inhibition rate = $\frac{\Delta [FXIIa]}{[FXIIa]dose}\times 100\%$ , where [FXIIa]_apparent is the activity of the mixed solution with αFXIIa, and [FXIIa]_material is the activity of FXII contact activated solution before addition of αFXIIa.)

αFXIIa added		Δ[FXIIa] (PEU/ml)		Inhibition rate (%)
(μg/ml)	[FXIIa]dose (PEU/mL)	Hydrophilic glass	Hydrophobic OTS glass	Hydrophilic glass	Hydrophobic OTS glass
1	0.079	0.079	0.041	99.9	70.1
2	0.133	0.12	0.079	87.7	48.9
5	0.370	0.15	0.19	41.4	53.3
10	0.747	0.31	0.39	41.2	47.7
20	1.437	0.56	0.55	38.8	38.6

This inhibition of αFXIIa activity by the products from FXII contact activation solution strongly suggests that inhibitory products are generated during contact activation. It is known that production of FXII through contact activation generates a mixture of fragments, including αFXIIa and βFXIIa, which have different procoagulant activity⁴², and also other unknown minor cleavage fragments. In a study of the amidolytic and procoagulant activity of protein fragments produced by contact activation of the FXII in buffer solution, Golas et al.³⁵ found both autoactivation and autoinhibition reactions present in system, and which were dependent on activator surface chemistry/energy. They proposed the term FXII_act to represent the suite of protein fragments generated during contact activation, including procoagulant molecules and enzymes that are amidolytic (cleave chromophores that measure FXII activity) but that are not procoagulant. They also proposed that an unknown number of protein fragments could be produced by contact activation, some of which may be responsible for autoinhibition of the autoactivation reactions. Golas et al proposed three possible reasons for the suppression of FXII autoactivation in buffer solution including denaturation of FXII, autoinhibition by FXIIa, and inhibition by an amidolytic enzyme³⁵. The findings in this work suggest that these unknown protein fragments may also contain molecules that inhibit the activity of exogenous αFXIIa, acting as a self-inhibitor of FXIIa activity within the intrinsic pathway of blood coagulation cascade.

The capability of these inhibitors to lower the activity of FXIIa appears dependent on the product sources and concentration of αFXIIa added. At lower concentrations of exogenous αFXIIa (1 μg/ml), the activity of exogenous αFXIIa is almost completely inhibited by the activation products from hydrophilic glass surfaces. With increasing exogenous αFXIIa concentration, relative inhibition rates decreased, but the amount of activity inhibited continued to increase by contact activation products from both hydrophilic and hydrophobic surfaces (Table 2). The hydrophilic and hydrophobic surfaces appear to have a greater effect on the relative inhibition rates at low doses of αFXIIa, but tend towards similar levels at high doses of αFXIIa. This suggests that the inhibitors generated from FXII activation on hydrophilic surface and hydrophobic surfaces may have similar capabilities towards inhibition of αFXIIa activity.

3.4. Inhibition of procoagulant activity of αFXIIa by FXII contact activation products generated in the presence of prekallikrein

Figure 4 illustrates the procoagulant activity of activation products of FXII contact activation generated in the presence of prekallikrein and then once again dosed with exogenous αFXIIa. The apparent activities of products of FXII contact activation in the presence of prekallikrein increased to 0.14±0.03 PEU/ml and 0.067±0.016 PEU/ml for samples activated on hydrophilic clean glass surfaces and hydrophobic OTS-glass surfaces, respectively, about 2 and 3 times higher than those samples without prekallikrein (Figure 3) and consistent with the role of reciprocal activation in contact activation²². For the FXII contact activation products from hydrophilic glass surfaces, the procoagulant activity of mixed activation products and exogenous αFXIIa is higher than the activity of corresponding pure αFXIIa for concentrations of αFXIIa ≤ 5 μg/mL, while the activity of the mixed solution became lower than the pure αFXIIa when the concentrations of αFXIIa were greater than 5 μg/mL (Fig. 4a). In the case of FXII contact activation on OTS hydrophobic surfaces, the activity of mixed solution was lower than the activity of pure FXIIa at the corresponding doses, similar to what was seen in Figure 3 where prekallikrein was not used, suggesting that the activity of exogenous αFXIIa is suppressed even by the FXII contact activation products produced by hydrophobic surfaces in the presence of prekallikrein. The decrease in activity and the relative inhibition rates are tabulated in Table 3. In the presence of prekallikrein during FXII contact activation, the inhibition amounts and rates seen with products of contact activation on hydrophilic glass surfaces are smaller than those for FXII contact activation without prekallikrein (refer to Table 2), while inhibition amounts and rates are similar for the products of FXII activated on hydrophobic OTS-glass surfaces whether in the presence or absence of prekallikrein. This suggests that reciprocal activation with cleavage of prekallikrein on hydrophilic surfaces either suppresses the generation of inhibitors or that reciprocal activation produces an increase in FXIIa that substitutes for the loss of the exogenous αFXIIa, but that this does not occur with hydrophobic surfaces. Recall that surface wettability affects protein adsorption and that hydrophobic surfaces can show an apparent loss in prekallikrein activity⁴¹ . It is believed that the protein-adsorption competition on hydrophobic surfaces may lead to the similar generation of inhibitors during FXII contact activation, but with an increasing amount of FXII activation product being generated in the presence of prekallikrein by reactions that occur in bulk, while for hydrophobic surfaces those activated proteins are unable to move into the bulk.

Figure 4.

Apparent procoagulant activity of contact activation products generated in the presence of prekallikrein and with exogenous αFXIIa. (a) FXII contact activation on hydrophilic glass surface and (b) FXII contact activation on hydrophobic OTS glass surface. (n=3)

Table 3.

Effect of prekallikrein on reduction amount (Δ[FXIIa]) and activity inhibition rate of αFXIIa due to FXII contact activation products in the presence of prekallikrein (PK=20 μg/ml). (Δ[FXIIa] = [FXIIa]_dose – ([FXIIa]_apparent - [FXIIa]_material). Inhibition rate = $\frac{\Delta [FXIIa]}{[FXIIa]dose}\times 100\%$ , where [FXIIa]_apparent is the activity of mixed solution dosed with αFXIIa, and [FXIIa]_material is the activity of FXII contact activated solution before dosed with αFXIIa.)

αFXIIa dosage (μg/ml)	Δ[FXIIa] (PEU/ml)		Inhibition rate (%)
	Hydrophilic glass	Hydrophobic OTS glass	Hydrophilic glass	Hydrophobic OTS glass
1	0.032	0.049	49.4	96.6
2	0.020	0.087	12.3	55.1
5	0.042	0.15	11.8	42.8
10	0.30	0.38	36.5	49.2
20	0.41	0.55	29.2	38.1

3.5. αFXIIa activity suppression of FXIIa by activator surfaces

These results have shown the inhibition of procoagulant activity of αFXIIa by activation products of FXII contact activation in sections 3.3 and 3.4, where the activation products were separated from activating beads and then incubated with exogenous αFXIIa. This section addresses the activity suppression of αFXIIa by coagulation activation products in the presence of activator materials as this is a more realistic representation of the real condition where coagulation occurs near to a biomaterial surface. To test this, exogenous αFXIIa at a concentration of 5 μg/ml was added to FXII contact activation products after1 hour incubation with activator beads. The mixture of FXII activation products, exogenous αFXIIa and activator beads was then incubated together for an additional 60 min. Surprisingly, the apparent activity of the mixture of exogenous αFXIIa and FXII contact activation products generated in the presence of the activating surfaces was only about half of the activity of αFXIIa incubated with activation products alone (no bead activators, Table 4), suggesting that the activity of αFXIIa was further suppressed by the activator beads. The inhibition rates of activation products may vary between the contact activation surfaces and the protein composition in solution (e.g., the presence of prekallikrein), however, the inhibition rates by the mixture of activation products and activator surfaces appear similar, approximately 85%, regardless of activator wettability and protein composition (Table 4).

Table 4.

Apparent procoagulant activity of αFXIIa inhibited by products of FXII contact activation on hydrophilic glass and hydrophobic OTS-glass surfaces. (αFXIIa= 5 μg/ml)

		Without Prekallikrein		With Prekallikrein
	Samples	Hydrophilic glass	Hydrophobic OTS-glass	Hydrophilic glass	Hydrophobic OTS-glass
Apparent activity (PEU/ml)	Material alone (w/o αFXIIa)	0.06±0.01	0.02±0.01	0.14±0.03	0.07±0.02
	αFXIIa + activation products	0.28±0.09	0.19±0.02	0.45±0.11	0.27±0.09
	αFXIIa + activation products + activating particles	0.12±0.02	0.07±0.01	0.20±0.04	0.13±0.03
Relative inhibition rate (%)	Activation products	41.4	53.3	11.8	42.7
	Activation products + activating particles	83.2	86.3	83.8	87.3

The mixture of supernatants and activator surfaces exhibits higher inhibition rates on αFXIIa activity compared to the activation products alone. Zhuo’s mathematical model for FXII contact activation linking FXII and FXIIa adsorption/desorption processes on procoagulant particle surfaces³⁶ proposed that all biochemical reactions occurred within an interphase region that surrounded a procoagulant particle immersed in FXII solutions. Adsorption/desorption of proteins forms the “FXII partition” and “FXIIa partition” at the interphase and FXII activation competed with an autoinhibition reaction, resulted in self-limiting production of FXIIa in plasma and buffer. Because surface activated FXIIa is really a suite of proteins (FXII_act), it is likely that the “FXIIa partition” in this model can be extended to other protein fragments, including inhibitors. Partitioning concentrates proteins (e.g., inhibitors) within the interphase around the particles and the concentration of inhibitors in bulk solution is therefore limited. The partition is dependent on surface properties such as surface wettability, resulting in different inhibition rates for different material activators.

The activation of FXII by material surfaces is far more complicated than the simple reactions described as $FXII\frac{contact}{surface}\to FXIIa$, and the products of reactions show both procoagulant (leading to blood coagulation) and amidolytic (cleave chromophores but do not lead to coagulation) activities³⁵. In this study, we show that this suite of activation products also includes molecules that can inhibit αFXIIa activity. In 2015, Golas et al.⁴³ showed electrophoresis bands for FXII activated by 3 different surface chemistries. In these gel images, 99% of all FXII was seen as a single band at molecular weight of 80 kDa, corresponding to the molecular weight of fully intact FXII. These activation products were termed conformers: full chain FXII molecules whose conformation has been perturbed by surface (or near surface) contact in a manner that changes the function of the protein but does not affect the primary structure or length. It is these conformers that constitute the FXII_act products mentioned here, and that are substantially different from the αFXIIa that is prepared by interactions with dextran sulfate.

Results in this work demonstrate that the products of material activation of FXII reduce the activity of exogenous αFXIIa and suggest that the contact activation process produces unknown protein fragments different from either the procoagulant or amidolytic enzymes previously seen. Although it still remains to identify the molecular characteristics of these different products, the observation of the presence of αFXIIa inhibitors within the suite of activation products produced after material contact adds to our understanding of the complex processes that occur in the intrinsic pathway of the blood plasma coagulation cascade. It should be noted that this work was carried out in a model system with purified coagulation factors. The FXII contact activation will be more complicated in the complex system of blood, and other factors from the blood components may affect both the autoactivation and autoinhibition processes. However, it’s also important to note that after more than 5 decades since the first descriptions of the blood coagulation cascade^44,45, we still do not fully understand even the first step in this highly complicated process, and this in-vitro testing described here provides hypotheses for molecules to search for during in-vivo activation, and provides further information useful in the prospective surface engineering and design of cardiovascular biomaterials with improved hemocompatibility.

4. Conclusions

This work demonstrates that contact activation of FXII by both hydrophilic and hydrophobic surfaces results in suppression of αFXIIa activity. Taken with prior work, this study extends our knowledge of the contact activation process and illustrates that contact activation of FXII in buffer solution appears to produce a suite of protein fragments that exhibit both procoagulant activity and amidolytic activity, and that also reduces the activity of exogenous αFXIIa. The degree of suppression appears dependent on the level of exogenous αFXIIa, suggesting that there is a finite amount of activator produced. At high levels of exogenous αFXIIa, the inhibitors produced from hydrophilic clean glass and hydrophobic OTS-glass surfaces exhibit similar relative inhibition rates.

The presence of prekallikrein in the system increases the apparent yield of FXIIa following FXII contact activation for both hydrophilic and hydrophobic surfaces, but the suppression levels of inhibitors produced depends on surface characteristics. Prekallikrein suppressed the apparent yield of inhibitors on hydrophilic surface, but had no effect on inhibitors in the case of the hydrophobic surface. The combination of FXII contact activation products and activator surfaces dramatically inhibits the activity of exogenous αFXIIa, significantly higher than that by the activation products alone, regardless of activator surface wettability and the presence of prekallikrein. Together, these results demonstrate that the contact activation process is far more complex than the view that surface contact activation produces αFXIIa that either amplifies coagulation through reciprocal activation or propagates the cascade leading to blood coagulation and fibrin formation.