A, proteins involved in HA matrix stabilization. Ternary and quaternary structures are schematically shown, with the sizes of all proteins and their subunits approximately to scale and known interactions indicated by arrows. Subunits that were used separately in addition to the complete protein are enclosed within dashed boxes. B, schematic illustration of the platform for solid-phase binding assays. A gold support was modified with a protein-repellent oligoethylene glycol monolayer functionalized with biotin, followed by the formation of a dense monolayer of well oriented SAv. HA, Link_TSG6, or PTX3 was grafted to the SAv layer via biotin tags. HA chains were site-specifically functionalized with biotin at their reducing end and could therefore be immobilized at controlled orientation. Proteins were functionalized through primary amines and therefore might be immobilized in a variety of orientations. The thickness of the oligoethylene glycol monolayer and the dimensions of SAv, Link_TSG6, and the PTX3 octamer are drawn approximately to scale; the thickness of the HA brush and the mean distance between HA anchor points are reduced by 10–20-fold for illustrative purposes. C, interaction of PTX3, IαI, and a mixture of IαI/PTX3 with HA film. A control shows that 1 μm IαI, 0.3 μm PTX3, and a mixture of these proteins do not bind to HA in the absence of TSG-6. The start and duration of the incubation with different samples are indicated (arrows). After each incubation step, the solution phase was replaced by buffer. QCM-D did not show any significant interaction between the HA film (58 kDa) and IαI, PTX3, or a PTX3/IαI mixture. The employed HA films are easily permeated by the proteins (cf. B). The control measurements therefore also confirm that our streptavidin-coated surfaces are resistant to nonspecific binding of IαI or PTX3, alone or in a mixture. D, interaction of surface-bound Link_TSG6 with octamer-forming wild type PTX3 and dimer-forming N_PTX3_MUT. Interactions were measured by QCM-D. Biotinylated Link_TSG6 (b-Link_TSG6) but not Link_TSG6 without biotin was immobilized to a streptavidin monolayer, and binding of PTX3 constructs was monitored. The bulk PTX3 concentration refers to the molar concentration of PTX3 monomers. E, interaction of surface-bound PTX3 with Link_TSG6 and rhTSG-6. Interactions were measured by SE. b-PTX3 was immobilized, and binding of Link_TSG6 and rhTSG-6 was monitored. F, interaction of surface-bound PTX3 with IαI. Interactions were measured by QCM-D. b-PTX3 was immobilized, and binding of IαI was monitored. The curves shown in C–F are representative of sets of measurements performed at least in duplicate.

PTX3 does not incorporate into HA films via full-length rhTSG-6, but it does via Link_TSG6. The HA film (837 kDa) was first loaded with Link_TSG6 (A) and rhTSG-6 (B). After equilibrium was established, free TSG-6 protein remained in solution. The addition of PTX3 revealed fast adsorption to the Link_TSG6-loaded HA film but not to the rhTSG-6-loaded HA film. The gray solid line in B is a linear fit revealing an initial binding rate of 44 ng/cm²/min. The addition of 0.08 μm anti-PTX3 antibody (MNB4) in the rinsing phase in B did not affect the unbinding curve. The SE curves shown are representative of sets of measurements performed at least in duplicate.

PTX3 differentially modulates the interactions of Link_TSG6 and full length TSG-6 with HA. A, PTX3 modulates Link_TSG6 interaction with HA. Binding isotherms obtained from titration of Link_TSG-6 into an HA film, monitored by SE, in the presence or absence of 0.3 μm PTX3 in the solution phase. A fit (solid line) of the titration data for Link_TSG6 (filled triangles) with the Hill equation gave a Hill exponent close to 1.0, confirming a simple non-cooperative interaction between HA and Link_TSG6. In the presence of PTX3, the interaction is more complex. The sigmoidal shape of the data (filled circles) indicates cooperative binding. A fit with the Hill equation (dashed line) indeed provides a Hill exponent of 3.6 (i.e. much larger than 1.0). B, competition of PTX3 and HA for TSG-6 binding. Binding isotherms obtained from titration of rhTSG-6 into an HA film, alone (filled circles) and in the presence of 0.3 μm (filled triangles) or 1 μm (empty triangles) PTX3 in the solution phase. At a bulk rhTSG-6 concentration of 0.45 μm, an ∼3-fold decrease in the surface density of adsorbed rhTSG-6 can be detected in the presence of 0.3 μm PTX3. At 1 μm PTX3, the decrease in binding becomes even more pronounced.

PTX3 does not bind to HA films that had previously been exposed to a mixture of IαI/TSG-6. A, binding assays by SE. 1 μm IαI and 0.3 μm TSG-6 were sequentially added to the HA film without premixing. In this case, the film contains an additional fraction of non-covalently but stably bound protein (). The gray solid line is a linear fit revealing an initial binding rate of 13 ng/cm²/min. Incubation with 0.3 μm PTX3 does not affect the surface density of the film. The lack of a significant response upon incubation with 0.08 μm anti-PTX3 antibody (MNB4) confirms the absence of PTX3 binding. The curve shown is representative of a set of measurements performed in duplicate. B, HA films are permeable to PTX3. b-PTX3 was added to SAv-covered surfaces without any further functionalization (solid line) or in the presence of HA (837 kDa) films with a surface density of 35 ± 5 ng/cm². HA films were presented pure (dashed line) or following exposure to 0.3 μm rhTSG-6 (solid line with open circles) or to a mixture of 1 μm IαI and 0.3 μm TSG-6 (premixed for 1 min before the addition to the HA film; solid line with filled squares). Only initial binding is shown. Binding with similar rates is observed for all surfaces. Because PTX3 alone did not show binding on any of these surfaces, the binding of b-PTX3 must occur via the biotin moiety to SAv, indicating that all HA films are permeable to PTX3.

PTX3 incorporates into HA films when presented in a ternary mixture with TSG-6 and IαI. A, binding assay by SE. HA films were first exposed to 0.3 μm PTX3. After 2 min of incubation, 1 μm IαI was added, and after another 2.5 min, 0.3 μm rhTSG-6 was added. Binding ensued after the addition of rhTSG-6; the gray solid line is a linear fit revealing an initial binding rate of 6 ng/cm²/min. The protein mixtures were incubated with the HA films for 2 h. Subsequent binding of anti-PTX3 antibody, incubated at 0.08 μm, indicates successful incorporation of PTX3. The curve shown is representative of a set of measurements performed in duplicate. B, binding assay by QCM-D. Δf (blue lines) and ΔD (orange lines) are shown. Mixtures of PTX3, rhTSG-6, and IαI were exposed to HA films at final protein concentrations of 1.0, 0.6, and 0.2 μm, respectively. In one case (circles), PTX3 was first mixed with TSG-6 for 2 h and then with IαI for another 1 h (all at room temperature) before exposure to HA. In the other case (triangles), TSG-6 and IαI were mixed first (for 2 h), and then PTX3 was added (for 1 h) before exposure to HA. Clear QCM-D responses upon the subsequent addition of anti-PTX3 antibody, incubated at 0.08 μm, indicated successful incorporation of PTX3.

PTX3 in a mixture with IαI incorporates into TSG-6-loaded HA films. A, the HA film was first loaded with 0.3 μm TSG-6, excess protein was removed from the solution phase, and a mixture of 0.3 μm PTX3, 1 μm IαI was added (premixed for 1 min). The addition of the PTX3/IαI mixture first enhanced desorption until a mass fraction of 35% remained, and thereafter incorporation of material into the HA film started. The addition of anti-PTX3, after the removal of excess protein in the solution phase, confirmed PTX3 incorporation into the film. Only a small fraction of about 12% could be eluted with 2 m GdnHCl, whereas most material was eluted with 8 m GdnHCl. A fraction of 10% remained bound in 8 m GdnHCl but could be largely digested by hyaluronidase. The increase in the desorption rate upon the addition of IαI/PTX3 can be appreciated from the linear fits to the data shortly before (green dashed line; 3.7 ng/cm²/min) and after (red dashed line; 27 ng/cm²/min) protein addition. The adsorption process setting in 10 min after incubation with IαI/PTX3 was fitted by an exponential (gray dashed line). The fit revealed a maximal surface density of 660 ng/cm² and a half-time of about 90 min. The curve shown is representative of a set of measurements performed in duplicate. B, equivalent measurement with IαI instead of a mixture of IαI and PTX3, with linear fits to the data shortly before (green dashed line; 2.9 ng/cm²/min) and after (orange dashed line; 31 ng/cm²/min) IαI addition, reproduced from B in Ref. . In this case, the adsorption process, commencing 10 min after incubation with IαI, could be fitted with a straight line (gray dashed line).

PTX3 inhibits the catalytic activity of TSG-6 in HC transfer. In a solution-phase assay, 0.27 μm TSG-6, 1.8 μm IαI, and 20 μm b-HA₁₀ were co-incubated with or without 1.8 μm PTX3 for various times and subsequently analyzed by Western blots with streptavidin-conjugated Alexa 488, which recognizes biotin in b-HA₁₀·HC complexes. A, Western blot with co-incubation times indicated. B, densitometric analysis of Western blots. Error bars, S.E. from three blots. Data are representative of two independent experiments.

Western blot analysis of protein material incorporated into HA films. A–C, HA films were incubated with proteins in binary (IαI/TSG-6) and ternary (IαI/TSG-6/PTX3) mixtures, as shown in A and , respectively. Western blots were made from fractions obtained by stepwise elution with 2 and 8 m GdnHCl and by digestion with hyaluronidase (HA-ase). Collected material was analyzed by Western blots with anti-PTX3 (A), anti-IαI (B), and anti-TSG-6 (C) antibodies. The control reaction mix of TSG-6, HA₁₄, and IαI is expected to contain a total amount of 100 ng of TSG-6 and 25 ng of IαI, the control lane for PTX3 is expected to contain 500 pg of PTX3, and the detection limits are estimated to be around 5 ng for TSG-6, 0.5 ng for IαI, and 50 pg for PTX3. D, direct comparison of IαI and TSG-6 proteins running in 4–12% NuPAGE BisTris gels (used in B and C) and 10% Tris/Tricine gels (used by Rugg et al. ()), stained with Coomassie Blue. The same standards were used for both gel types (as indicated), and their apparent molecular masses were assigned following the manufacturer's indications for NuPAGE BisTris 4–12 with MES (as in B and C) and Tris-glycine gels (as in Ref. ), respectively. IαI and TSG-6 were mixed (1.8 and 2.7 μm, respectively) and co-incubated at 4 °C using the standard conditions (as described in Ref. ). Immediately after mixing (0 h), bands for intact IαI and TSG-6 are dominant; after 2 h of co-incubation, additional bands for HC·bikunin/HC·TSG-6 appear. In the 10% Tris/Tricine gel, IαI, HC·bikunin, the HC·TSG-6 doublet, and TSG-6 run at apparent molecular masses of ∼220, ∼130, ∼120, and ∼38 kDa, respectively, consistent with Rugg et al. (), where these had been identified by Edman degradation and mass spectrometry. In the 4–12% NuPAGE BisTris gels, the apparent molecular masses for IαI and TSG-6 are ∼170 and ∼34 kDa, respectively; the HC·bikunin and HC·TSG-6 species run together with an apparent molecular mass of ∼98 kDa. HMW IαI, a high molecular weight form of IαI with three or four HCs attached, which is a minor species within the IαI preparation purified from serum that also forms as a by-product of HC·TSG-6 complex formation (see Ref. ).

Effect of PTX3 on HA film morphology. A, binding assays by QCM-D. Δf and ΔD at three selected overtones (n = 5, 7, and 9) are shown. HA films were first exposed to the protein mixtures 1.0 μm IαI, 0.3 μm TSG-6; 0.3 μm PTX3, 1.0 μm IαI, 0.3 μm TSG-6; or 0.3 μm N_PTX3_MUT, 1.0 μm IαI, 0.3 μm TSG-6, with all proteins premixed (in the indicated order) shortly before the addition to the HA films. After removal of excess protein from bulk solution, the incorporation of PTX3 into the HA matrices was confirmed by the addition of 0.1 μm anti-PTX3 (MNB4). Films exposed to the binary IαI/TSG-6 mixture did not show any significant response to anti-PTX3. A decrease in Δf (blue curves) accompanied by an increase in ΔD (orange curves) for both ternary mixtures evidences anti-PTX3 binding, confirming that both PTX3 and N_PTX3_MUT can be incorporated into HA films. B, parametric plots reveal morphological differences. Plots of ΔD versus Δf are shown, for the incubation of HA films with binary and ternary protein mixtures. Data were taken from A, and ΔD and Δf were offset to zero shortly before the start of protein incubation. This parametric plot provides a qualitative fingerprint of how the mechanical properties of the HA matrix evolve upon protein incorporation. For the binary mixture and the ternary mixture containing N_PTX3_MUT, all overtones produce a roughly linear response. The slope is comparable between overtones and also between the two protein mixtures. The latter indicates that N_PTX3_MUT, although incorporated into the film, does not affect the film morphology drastically as compared with the binary protein mixture. The response for the ternary protein mixture containing wild type PTX3 is distinct, with a large spread between the three overtones and non-linear shapes for n = 7 and 9. This indicates that incorporation of PTX3 does affect the film morphology appreciably.

Effect of proteins on HA film thickness and cross-linking. A, sketch of the experimental setup. HA films (837 kDa) were formed separately on a planar surface and on colloidal probes, and the probes were then added to the planar surface. The probe-surface distance (interaction range) and the in-plane thermal motion of the probe were measured by RICM, for pure HA films and in the presence of protein mixtures. B, interaction range. Proteins were incubated for 2 h at the indicated concentrations. Proteins were added to the HA film in sequential order: first IαI, either alone or in a mixture with a PTX3 construct (octamer-forming wild type (PTX3) or dimer-forming mutant (N_PTX3_MUT); premixed for 1 min) and thereafter TSG-6. Error bars, S.E. of 10 measurements with different colloidal probes on the same sample; data are representative of at least two sets of experiments on different samples. C, representative traces of the thermally driven random in-plane movement of a bead's center over a period of 62.5 s for pure HA films on the planar surface and on the colloidal probe, alone and in the presence of protein mixtures. The scales in the x and y direction span a range of 800 nm. D, lateral MSD as a function of lag time τ extracted from the curves shown in C. Error bars correspond to statistical uncertainties for a given measurement. E, MSD for τ = 0.625 s. Error bars, S.E. of three measurements with different colloidal probes on the same sample.

Interaction of PTX3 with TSG-6 and HA. Sketches illustrate proposed mechanisms for Link_TSG6 (A) and rhTSG-6 (B).