To investigate the effectiveness of composite hydrogels as cartilage implants by means of in vivo tests, we have selected the samples with the compositions listed in Table 1. This choice was based on the data previously obtained in the study of the peculiarities of the synthesis of composite hydrogels and of the variation of their mechanical behavior when varying both the cellulose content and ionic groups’ concentration in synthetic polymer chains (Buyanov et al. 2013, 2019). The previous data on biocompatibility of these materials with cartilage and bone tissues of laboratory animals and their ability to maintain integration with these tissues during their functioning as artificial cartilage tissue substitutes (Bozhkova et al. 2016) were taken into account. The available information concerning the mechanical behavior of natural articular cartilage was also considered.
`In this respect, the previously tested samples of hydrogels with cellulose content in the compositions of about 7 and 18 wt.% were of interest. At the first stage of the work the mechanical characteristics of these materials synthesized for in vivo tests were measured in the single-shot compression mode (Table 1). Varying the cellulose content in the compositions enabled us to obtain samples with different equilibrium water contents and different levels of mechanical characteristics. Two non-ionic hydrogels (samples 1 and 3) with cellulose percentages equal to 7 and 18 wt.%, respectively, were used as starting materials for preparation of ionic hydrogels 2 and 4 by alkaline hydrolysis of part of amide groups in PAAm chains. The degrees of hydrolysis of PAAm amide groups were equal to 10 and 25% for hydrogels 2 and 4, respectively, that is, they contained 10 and 25 mol% of carboxylate groups PAA(Na+) in polymer chains PAAm-PAA(Na+).
The presence of ionic groups in hydrogels leads to an increase in their degrees of swelling for well-known reasons (Flory 1953), but the mechanical characteristics of hydrogels deteriorate. It is seen from the data presented in Table 1 that for sample 2 (which contains 10 mol % carboxylate groups in synthetic chains polyacrylamide-sodium polyacrylate (PAAm-PAA(Na+)), the equilibrium water content increased by 9 wt.% as compared to that in the non-ionic hydrogel (sample 1), and the level of mechanical stiffness decreased only slightly. It should be noted that the method of alkaline hydrolysis of PAAm amide groups used in our work to introduce ionic groups into hydrogels does not affect hydrogel structure.
Another sample of non-ionic hydrogel 3, which contains higher amount of cellulose, has too high level of stiffness and at the same time insufficiently high water content (43 wt. %), which is much lower than the range typical of natural cartilages (60–80%) (Mow et al. 1992). At the same time, it is seen from Table 1 that ionic hydrogel 4 obtained from non-ionic sample 3 contains more water (by 25%, i.e., 68 wt.%) and is obviously more suitable as a cartilage substitute in terms of this parameter.
The introduction of ionic groups into hydrogels was also interesting, since it was expected that the presence of these groups would regulate the process of hydrogel mineralization (which has been observed earlier in the course of in vivo experiments with non-ionic hydrogels) (Buyanov et al. 2016). The non-ionic hydrogel samples implanted in the area of contact with subchondral bone were subjected to intensive mineralization with the formation of significant amounts of calcium phosphate (up to 40 wt.%) inside the hydrogel, mainly in the form of hydroxyapatite. At the same time, no calcium phosphates were observed in the cartilaginous area of the implants (out of contact with subchondral bone) after 45 days (Buyanov et al. 2016). It should be clarified that in our works, we simulate the possibility of treatment of deep local defects of articular cartilage tissue, and the majority of implants are located in the area of subchondral bone.
On the other hand, the currently available data do not allow one to predict the consequences of long-term exposure of implants to human body; thus, it is necessary to be able to regulate the rate of implant mineralization in order to minimize the possibility of the formation of mineral phase in the cartilage area of implants. Otherwise, the mineralization process can eventually lead to the formation of a bone-like composite in this area, which will not be able to perform the functions of a cartilage tissue substitute.
While the process of implant mineralization in the subchondral region has a positive role, promoting integration between the implant and bone tissue, the mineralization of the cartilaginous region should be avoided for the reasons mentioned above. According to our preliminary data obtained during the in vitro study of composite hydrogels mineralization (to be published), the presence of ionic carboxyl groups in these hydrogels reduces the mineralization rate by at least an order of magnitude. In order to investigate the possibility of regulating mineralization rate in vivo, the minimum required number of ionic groups was introduced into hydrogels so that their mechanical characteristics remained at a sufficiently high level.
Mechanical behavior of composite hydrogels
As was mentioned in the Introduction, the level of mechanical characteristics of hydrogels is a key parameter that determines the possibility of their application as artificial cartilage substitutes, especially if they are intended to replace damaged areas of articular cartilages. In this connection, let us now consider mechanical behavior of the hydrogel samples listed in Table 1, which were used in the present work as cartilage implants.
The single compression curves of the hydrogels selected for the experiments are given in Fig. 3. The values of secant moduli of compression obtained in the 10–15% region (Е|10−15%) and the stresses corresponding to compressions of 30 and 50% (σ|30% and σ|50%) for these hydrogels are presented in Table 1.
As has been demonstrated in our previous work (Buyanov et al. 2013), the single compression curves of hydrogel samples provide information about the initial stiffness (modular characteristics) of hydrogel materials. However, during repeated compression cycles, significant changes in these curves are observed due to the profound reorganization of the system of physical entanglements and hydrogen bonds that existed in the material before the onset of deformation. These changes are reversible, i.e., after a certain period of time (typically more than two days) the shape of the curves is restored to the original one. However, when a certain level of load is exceeded, chemical bonds and cross-linking nodes start to break down, which eventually leads to disintegration of the material (Buyanov et al. 2013, 2019).
To investigate viscoelastic behavior of hydrogels with the stiffness levels typical of natural articular cartilages and to assess their real functional strength, we developed the testing method involving repeated cyclic compression. In these experiments, a cylindrical sample was subjected to a series of compression cycles (10–100 cycles per series, depending on the objectives of the experiment) with fixed amplitude. After completion of the series of cycles, the experiment was repeated with increased amplitude. Thus, from series to series the amplitude of compression was increased stepwise (from 30–40% to 60–80%) depending on the level of ultimate breaking strain of a material.
The curves of cyclic compression with the amplitudes of 30, 50, and 70% for non-ionic sample 1 (Table 1) are displayed in Fig. 4.
During the first compression cycles of all series (amplitudes of 30, 50, 70%), a rather wide hysteresis loop is visible. However, already in the second cycle of each series, the hysteresis loop “shrinks” dramatically: the compression curve, which in the first cycle of the series passed noticeably higher than the unloading curve, is located lower and practically overlaps with the unloading curve. At the same time, the compression and unloading branches of the cyclic curves of the second and following cycles coincide with the unloading curve of the first cycle. This means that in the process of compression of the sample in the first cycle (as well as in the case of single compression) serious stress-induced changes in the structure of the studied materials occur. Apparently, the system of hydrogen bonds that stabilize the IPN structure is being rearranged (which includes breaking of a part of these bonds); possibly, simultaneous rearrangement of the physical network (entanglements of macromolecules) also occurs. However, this hydrogel sample endures multiple cyclic compressions with amplitude of 70%, and no signs of material failure are observed. The maximum value of compression stresses slightly decreases from cycle to cycle and is close to the value observed in the first compression cycle (11 MPa).
Note that already for the first cycle of the series with amplitudes of 50 and 70%, the shape of compression curve is determined by the sample history: the compression curve for this cycle coincides with the compression/unloading curves of the second and subsequent cycles of the previous series up to the maximum compression amplitude realized in this previous series (Fig. 4).
Similar effect is clearly seen in the cyclic compression curves of ionic hydrogels (Fig. 5). Thus, with increasing amplitude, the envelope of all cyclic curves coincides well with the curve of single compression of the material; it has been demonstrated in our earlier work (Buyanov et al. 2013) that the maximum values of compression stresses in the cycles of each amplitude become approximately similar to the values observed at single compression of hydrogels. The viscoelastic behavior of the studied hydrogels observed in these experiments has been previously described as a characteristic behavior of elastomers - the Mullins effect (Mullins 1969).
As the compression amplitude increases, a significant reduction in the stiffness level of hydrogel material (a decrease in the slope of the curves) is observed up to high deformation values. In this connection, we proposed to characterize the stiffness of hydrogels by the value of the secant compressive modulus in the deformation range of 10–15%, which was calculated for one of the compression cycles following the first cycle, for example, the fifth cycle. Table 2 shows the values of secant compressive moduli Е|10−15% calculated for the fifth cycle compared to the compressive moduli calculated for the first cycle in the same range of strains. Hereinafter, we present the data obtained during the compression cycles with the amplitude of 50%, since we take this value as the maximum amplitude possible during the real functioning of cartilages in joints (Ker 1999).
Table 2
Comparison of mechanical characteristics of the studied hydrogels measured during single and cyclic compression tests
|
Sample No
|
Sample type
|
Single compression
|
Cyclic compression,
the 5th cycle
|
|
Е|10−15%,
MPa
|
σ|50%,
MPa
|
Е|10−15%,
MPa
|
σ|50%,
MPa
|
|
1
|
PC-PAAm
|
8.14
|
3.99
|
2.28
|
3.92
|
|
2
|
PC-PAAm-PAA(Na+)_10
|
5.92
|
4.50
|
2.27
|
4.03
|
|
3
|
PC-PAAm
|
34.9
|
16.6
|
4.92
|
16.78
|
|
4
|
PC-PAAm
-PAA(Na+)_25
|
22.0
|
8.46
|
4.08
|
9.33
|
It can be seen that the values of compressive moduli calculated for the fifth cycle are indeed several times lower than those obtained during single compression of these materials. At the same time, the values of amplitude compressive stresses in the fifth cycles (σ|50%) differ only slightly from the values obtained during single compression of the samples.
Taking into account the results of mechanical tests and estimation of equilibrium degrees of swelling, hydrogel samples 1, 2, and 4 (Table 2) were selected for further in vivo experiments. In general, the level of mechanical characteristics of the selected hydrogels corresponds to that observed for articular cartilages (when we compare the results obtained by similar methods in the experiments carried out in similar testing regimes).
Since cartilages exhibit viscoelastic properties (Lawless et al. 2017), their mechanical behavior depends on the deformation rate. For example, in (Li et al. 2021), it was found that under conditions of unconfined compression, the stress magnitude at 50% deformation for dog knee cartilage increased from approximately 0.7 to 6 MPa as the compression rate increased from 3 to 300%/min. In our works, the compression rates equal to 1 and 10 mm/min were used, which at 4 mm sample height gives deformation rates of 25 and 250%/min.
For a more accurate comparison of the discussed values for cartilages and hydrogels, it was advantageous to obtain the dependences of mechanical characteristics on compression rate for the samples studied in this work. The corresponding dependence for sample 1 is shown in Fig. 6.
At such variation of the deformation rate, very limited changes in the stiffness of the material were registered. Thus, the stress corresponding to compression of the sample by 50% (ε = 0.5) increases from 3.0 to 5.0 MPa with increasing deformation rate over the entire range used in this experiment (Fig. 6). It may be noted that this change in material stiffness turned out to be less significant than that obtained in the experiments involving natural cartilage tissues (Li et al. 2021). This means that for our hydrogel materials, the elastic component contributes much more significantly to the viscoelastic mechanical behavior of the material than in the case of cartilages studied in (Li et al. 2021).
However, as was shown in a number of studies of the mechanical behavior of cartilage tissues (Gore et al. 1983; Deneweth et al. 2013; Li et al. 2021), the variation of their stiffness is quite wide not only for different joints, but also for cartilage regions of different localization within a single joint. Thus, it can be reasonably stated that the deformation behavior of the hydrogel materials studied in this work closely follows the behavior of cartilage tissues.
The experiments carried out in this work showed that the variation of the deformation rate of hydrogel samples within the range of 1–10 mm/min (25–250%/min) causes only insignificant changes in the characteristics of the material: for example, for sample 1, the compression stress measured at 50% increased from 3.80 to 3.87 MPa. Thus, the variation of rate regimes of hydrogel testing within the specified limits (which in some cases was reasonable for technical reasons) does not lead to any noticeable distortion of the obtained results.
Mechanical properties of cartilage implants removed from the rabbit’s knee 90–120 days after operation
A portion of the hydrogel implants (samples 1, 2, and 4) retrieved after in vivo experiments was tested in the cyclic compression regime with increasing amplitude under the same conditions that were used to test the virgin samples.
Figure 7a shows the curves of cyclic compression of the implant (non-ionic sample 1) removed from the loaded area of rabbit knee joint 120 days after operation. Compression cycles were applied to the sample with successively increasing amplitude (by 20%, from 30 to 70%); 100 compression cycles were performed at each amplitude. It is seen that the viscoelastic behavior of this implant, which was previously observed for the original sample (Fig. 4), is retained. It may be noted that for the cycles with the maximum compression amplitude (70%) there is a certain decrease in the stiffness of the material after its exposure to the joint: the stresses in the compression cycles are somewhat lower than those observed at the same deformations for the original sample (Fig. 4). For the implant subjected to the compression cycles with the amplitude of 70%, there is a sequential (from cycle to cycle) decrease in the stresses, which was not observed for the hydrogel in the initial state. It may be assumed that in the process of cyclic loading with this high deformation amplitude, the sample began to fracture gradually: during the last cycles, the amplitude value of stress is 2.5 times lower compared to that in the first cycle (Fig. 7a). It is most likely that this behavior is associated with a decrease in the elasticity of the material due to its significant mineralization in the area of contact with the subchondral bone, which leads to the early development of destruction processes in the sample. However, as it was mentioned above, the real compression deformations for cartilages do not exceed 50% even at maximum loads, and at cyclic compression with this amplitude the implant behavior is sufficiently stable (Fig. 7a). The amplitudes of cyclic stress in the material (more than 4 MPa) are close to those for cartilages.
Figure 7b shows that the curves of the fifth compression cycle at an amplitude of 50% for the implants removed 90 and 120 days after placement are very close to the curve for the initial hydrogel (with a slight decrease in stiffness). The most significant decrease in the level of compression stresses (from 4.67 to 3.65 MPa) is observed for the sample extracted from the loaded area of cartilage 120 days after implantation. The curves presenting the dependences of the “actual" compressive modulus dσ/dε on deformation are also close to each other (Fig. 7c). The stiffness level of the material also slightly decreases compared to the initial one for the sample extracted from the loaded area of cartilage 120 days after implantation.
The results above testify to the absence of palpable difference in the mechanical behavior of the implants removed from the loaded or non-loaded areas of cartilage, as well as implants removed at different in vivo experimental time frames (90 or 120 days). Under this reason and to minimize the amount of the laboratory animals involved in the tests, in the further study of the mechanical behavior of ionic implants, the duration of these experiments was limited to 90 days.
Figure 8 illustrates the results of the analysis of features of mechanical behavior of ionic type implant hydrogels (samples 2 and 4) extracted 90 days after implantation, which are similar to those discussed above for sample 1 (Fig. 7).
The implant sample prepared of ionic hydrogel 2 withstood the compression cycles at the amplitude up to 70% inclusive, but upon increase in the amplitude up to 75% it was destroyed during the first compression cycle (Fig. 8a, curve 5). The stiffer sample of ionic hydrogel 4 withstood the effects of compression cycles at the amplitude of 50%, but collapsed during compression in the first cycle when the amplitude was increased up to 60%. On the contrary, the initial hydrogel samples 2 and 4 showed higher elasticity and tolerated cyclic compressions up to amplitudes of 75 and 65%, respectively (Fig. 5).
Thus, we observe a certain loss of elasticity, i.e. the aging process occurring during functioning in the joint for both non-ionic and ionic types of implants. These processes are intensively manifested at increasing amplitude of compression, namely at amplitudes above 50%. Most likely, this behavior is related to the decrease in the elasticity of a material due to its mineralization in the area of contact with the subchondral bone. At the same time, as it was mentioned above, the compressive deformations to which cartilage in joints is actually subjected, even under maximum loads, do not exceed 50%. Therefore, a drop in compressive strain at the amplitude of more than 50% is not critical.
A certain decrease in the level of compression stresses of the extracted implants in comparison with the original samples is observed for the fifth compression cycles at the amplitude of 50% (Fig. 8c and 8d for samples 2 and 4, respectively). However, such changes are also tolerable, since the implant retained a sufficiently high level of compression loads.
The dependences of the “actual” compressive modulus dσ/dε on the deformation value for the extracted ionic implants are very close to the corresponding curves for the initial samples (Fig. 8e and 8f). A similar behavior of non-ionic hydrogels is demonstrated in Fig. 7c.
Thus, the data discussed above showed that during the long-term functioning of hydrogels as artificial cartilages in the joints of animals (up to 90–120 days) there was no critical deterioration in the level of mechanical characteristics of these materials.
It should be noted that the high level of mechanical properties inherent to the hydrogel materials elaborated by our team as compared to those of other kinds of hydrogels (see the introductory section) originates from a specific structure of our materials composed of a flexible-chain synthetic polymer (PAAm) and a rigid-chain component, namely cellulose. The latter is characterized by a Kuhn segment length one order higher than that of PAAm (Buyanov et al., 2013). The elasticity modulus of hydrogels is known to increase along with the Kuhn segment length (Hasa et al. 1975). This inference was confirmed experimentally (Hasa and Ilavsky 1975; Buyanov et al. 1992). In principal, the effect above provides large possibilities of hydrogels’ strengthening. But in practice they can be realized only for both optimal ratio of the concentrations of IPN components and optimized protocol of hydrogel synthesis (Buyanov et al., 2013, 2019). It should be noted also that, in equilibrium swollen composite hydrogels, the system of physical entanglements of IPN, formed predominantly by the cellulose macrochains, is situated in the constrained state, because the PAAm network has a substantially greater tendency to swell than the physical network of cellulose does (Buyanov et al., 2013). It is evident that the constrained network of physical entanglements between polymer components of composite hydrogels substantially decreases the mobility of polymer chains, thus contributing to the observed high level of mechanical properties of composite hydrogels.
SEM and EDX studies of chemical composition and morphology of hydrogel implants
Non-ionic hydrogels
It has been shown (Buyanov et al. 2016) that the area of non-ionic cellulose/PAAm hydrogel implants located in the subchondral bone undergoes mineralization with the formation of calcium phosphate clusters in the gel volume. Chemical composition of these clusters is close to that of hydroxyapatite. Thus, it was desirable to obtain more detailed data on the morphology and structure of both non-ionic and ionic implants by SEM and to estimate the calcium phosphate content in the implants by EDX.
The fractured surface of implant 1 that was removed from the rabbit organism after 120 days of the in vivo experiment is displayed in Fig. 9. The main part of the implant has a predominantly homogeneous structure, but in its right part, a non-uniform layer with a thickness of about 100 µm can be seen. Separate parts of this layer were investigated by SEM and EDX. This layer is located on the lateral surface of the cylindrical implant sample, which contacted with the subchondral bone.
The SEM image of one region of this layer taken at higher magnification (×500) is shown in Fig. 10. Both local EDX scanning points and two rather large scanning areas marked by rectangles (spectra No. 3 and No. 6) are shown. Scanning area No 6 records high contents of both calcium and phosphorus: 18.3 and 9.05 wt.%, respectively (Table 3), giving about 27 wt.% calcium phosphate in the composition of the material at the studied area. At the same time, about 5.3 wt.% of nitrogen is found; that is, calcium phosphate in this zone is embedded in the polymer network of the hydrogel. The initial hydrogel contained about 20.1 wt.% nitrogen, which is characteristic of formulations containing predominantly PAAm. Scanning region No 6 also recorded a reduced carbon content (25.1 wt.%), but the content of oxygen (41.9 wt.%), which is incorporated into the calcium phosphates, is much higher than that in the initial hydrogel (24.3 wt.%).
The calcium content inside the large rectangular area (spectrum 3, Fig. 10) is very low (0.15 wt.%), and the phosphorus content is equal to 0. Meanwhile, the contents of other elements (C, N, O) are close to those in the initial implant. Thus, it may be stated that the mineralization process occurred inside the local layer of the hydrogel implant, beyond which signs of mineralization are completely absent. In individual scanning points 1, 2 and 8, which are located outside the boundary of the mineralized area of the implant, the contents of elements are close to those observed for scanning area 3.
Table 3
Chemical composition of the initial implant and its “subchondral” area after the in vivo experiment at the points and scanning areas indicated in Fig. 10
|
Spectrum No
|
Concentrations of elements, wt.%.
|
|
C
|
N
|
O
|
Na
|
P
|
Ca
|
|
Initial implant
|
| |
55.59
|
20.13
|
24.28
|
0
|
0
|
0
|
|
Implant after in vivo experiment
|
|
1
|
62.32
|
16.64
|
20.79
|
0
|
0.00
|
0.25
|
|
2
|
57.80
|
20.38
|
21.57
|
0
|
0.00
|
0.25
|
|
3
|
51.94
|
20.39
|
27.52
|
0
|
0.00
|
0.15
|
|
4
|
26.04
|
5.98
|
30.78
|
0.27
|
12.35
|
24.58
|
|
5
|
23.88
|
1.31
|
37.31
|
0
|
12.88
|
24.62
|
|
6
|
25.10
|
5.28
|
41.96
|
0.29
|
9.05
|
18.32
|
|
7
|
65.33
|
14.61
|
19.95
|
0
|
0.00
|
0.11
|
|
8
|
60.49
|
17.57
|
21.77
|
0
|
0.00
|
0.17
|
A more thorough analysis of the changes in the contents of elements in the implant was carried out in the vicinity of the highlighted area No. 6 (Fig. 10), along the line running through the mineralized and non-mineralized areas of the fractured surface of implant 1 with the scanning points 1–15, located with a step of about 10 µm in the direction going into the depth of the implant (Fig. 11, Table 4).
Figure 11 SEM microphotographs of the fractured surface of the subchondral areas of hydrogel implants (the initial sample 1) extracted from the rabbit body after 120 days of the in vivo experiment (a), the mineralized area of the implant at higher magnification (b), fragment of Fig. 11a marked with an arrow (c). Loaded area of cartilage. Magnifications ×5000 (a), ×50000 (b) and ×20000 (c)
There is a clear boundary between the mineralized area of the implant, which contains significant amounts of calcium phosphate (up to 40 wt.%, see Table 4), and the area almost completely free of this substance. The Ca/P ratio in the first area (1.92–2.09) is close to the characteristic ratio for synthetic hydroxyapatite Ca5(PO4)3OH – 2.16. The presence of carbon and nitrogen in scanning points 2–8 indicates that the mineral phase is embedded into the organic polymeric phase, i.e., into the polymeric network of the hydrogel. Two implant regions differing in morphology with the defined interface between them (Fig. 11a) fit quite well to each other.
Table 4
Chemical composition of the “subchondral” area of the implants in the scanning points indicated in Fig. 11a
|
Spectrum No
|
Concentrations of elements, wt.%.
|
|
|
C
|
N
|
O
|
Na
|
P
|
Ca
|
Ca/P
|
|
1
|
20.03
|
2.26
|
37.90
|
0.32
|
13.05
|
26.44
|
2.03
|
|
2
|
34.76
|
10.82
|
30.89
|
0.14
|
8.01
|
15.38
|
1.92
|
|
3
|
19.73
|
2.27
|
36.19
|
0.32
|
14.05
|
27.44
|
1.95
|
|
4
|
16.54
|
1.99
|
43.40
|
0.28
|
12.25
|
25.54
|
2.09
|
|
5
|
22.61
|
1.66
|
35.65
|
0.39
|
13.23
|
26.46
|
2.00
|
|
6
|
18.59
|
1.83
|
33.09
|
0.35
|
15.11
|
31.03
|
2.05
|
|
7
|
22.75
|
1.90
|
38.64
|
0.35
|
12.21
|
24.15
|
1.98
|
|
8
|
62.41
|
17.68
|
18.99
|
0.00
|
0.22
|
0.70
|
-
|
|
9
|
61.08
|
18.53
|
19.86
|
0.00
|
0.00
|
0.53
|
-
|
|
10
|
61.84
|
16.83
|
20.93
|
0.00
|
0.00
|
0.40
|
-
|
|
11
|
60.77
|
17.85
|
21.09
|
0.00
|
0.00
|
0.29
|
-
|
|
12
|
62.56
|
17.29
|
19.65
|
0.00
|
0.00
|
0.50
|
-
|
|
13
|
64.03
|
15.51
|
20.07
|
0.08
|
0.00
|
0.31
|
-
|
|
14
|
62.47
|
17.01
|
20.15
|
0.00
|
0.00
|
0.37
|
-
|
The integration between the mentioned areas of the implants was not disturbed even during cyclic tests up to high degrees of compression. It should be also noted that chemical composition and structure of the mineralized areas of the implants are close to those of bone tissues, and the formation of this mineralized material probably reflects the process of the integration of the implant with living tissues. This assumption is indirectly confirmed by the presence of certain formations with the size up to 10 µm in both implant zones, which can be either traces of cells or their waste products. One of these formations is pointed by arrow in Fig. 11a, and in Fig. 11c a fragment of this drawing is shown in close-up. It is likely that osteoblasts penetrate into the surface layers of implant hydrogels and start to actively form hydroxyapatite clusters in these layers.
Figure 11b displays the mineralized area of the implant at a magnification of ×50 000. Calcium phosphate crystals with the size of 100 nm and higher can be seen; these crystals form much larger clusters with the size reaching several microns (indicated in the image by arrows).
Along the implant-subchondral bone interface, crystalline formations with a certain ordered structure are observed in some places (Fig. 12). Ring-shaped layers are visible around the scanning point labeled with the number 1. In all six scanning points, phosphorus and calcium are present in concentrations up to 18 and 44 wt.%, respectively (Table 5), but the Ca/P ratio is much higher than 2 (2.34–2.52); in two points this value even reaches 3.5 and 9.0 (points 4 and 6, respectively). Thus, these crystalline regions apparently contain calcium phosphates differing in composition from hydroxyapatite. Judging from the presence of nitrogen in all spectra in the amounts ranging from 2.2 to 8 wt.%, we can conclude (similarly to the results discussed earlier) that in Fig. 12 we observe not purely mineral formations, but most likely polymer-inorganic objects, i.e. calcium phosphate is formed inside the polymer network of the hydrogel implant.
Thus, during 120 days of exposure of the non-ionic implant to the subchondral zone of a cartilage defect, the thickness of its mineralized area reaches about 100 µm (Figs. 9 and 10). The characteristics of the hydrogel exposed to the loaded area of rabbit knee joint cartilage obtained by SEM and EDX do not differ principally from those reported earlier for a similar implant extracted from the non-loaded area of the cartilage after 45 days of in vivo experiment (Buyanov et al. 2016). In both cases, the average calcium phosphate content in the mineralized zone reaches 35–40 wt.%, and the Ca/P ratio is close to 2, which is typical of hydroxyapatite. Moreover, the SEM studies revealed no fundamental differences in the morphology of implants extracted from the loaded (this work) and non-loaded regions of cartilage (Buyanov et al. 2016). In view of this observation, further studies were focused on the implants extracted from the loaded areas of cartilage.
Table 5
Chemical composition of one of crystalline areas of the implant in the scanning areas indicated in Fig. 12.
|
Spectrum No
|
Concentrations of elements, wt.%.
|
|
|
C
|
N
|
O
|
Na
|
P
|
Ca
|
Ca/P
|
|
1
|
20.32
|
3.65
|
14.86
|
0.36
|
17.29
|
43.52
|
2.52
|
|
2
|
22.99
|
2.46
|
26.14
|
0.30
|
14.15
|
33.96
|
2.34
|
|
3
|
16.40
|
2.24
|
18.86
|
0.24
|
17.88
|
44.38
|
2.48
|
|
4
|
23.73
|
3.82
|
17.21
|
0.16
|
12.32
|
42.76
|
3.47
|
|
5
|
22.45
|
7.95
|
32.01
|
0.21
|
10.64
|
26.74
|
2.51
|
|
6
|
32.94
|
3.53
|
22.04
|
0.18
|
4.13
|
37.18
|
9.00
|
Ionic hydrogels
The fractured surfaces of ionic sample 2 (Fig. 13) were obtained and tested in the same manner as those of non-ionic implant 1; elemental composition of this sample was determined by EDX in scanning points 1–16 (Table 6).
In this ionic implant, significant amounts of phosphorus and calcium are detected only in three scanning points (8, 12 and 16); they are not located near the lateral surface of the implant, which on the fractured surface runs along the right part of Fig. 13. In points 1–6 (near the lateral surface of the implant), the contents of these elements are insignificant: 0.8–2.46 wt.% of calcium and only 0.12–0.37 wt.% of phosphorus.
In other scanning points, the contents of calcium and phosphorus are also low in comparison with those observed for non-ionic implant 1. It can be stated that in this ionic implant, the mineralization process, as it was supposed, develops much slower than in the non-ionic implant. This conclusion is confirmed by the data for ionic implant 4 presented below.
To study implant 4 by SEM and EDX, hydrogel samples were excised from the joints together with the surrounding bone and cartilage tissue fragments. In Fig. 14, the implant boundaries are indicated by arrows, and the bone tissue surrounding the implant can be seen in the left and lower parts of the image.
Table 6
Chemical composition of implant 2 at the scanning points indicated in Fig. 13
|
Spectrum No
|
Concentrations of elements, wt.%.
|
|
|
|
C
|
N
|
O
|
Na
|
Mg
|
P
|
S
|
Ca
|
|
Initial implant
|
| |
64.08
|
14.64
|
21.28
|
0.00
|
0.00
|
0.00
|
0.00
|
0.00
|
|
Implant after in vivo experiment
|
|
1
|
64.04
|
14.69
|
20.98
|
0.05
|
0.00
|
0.21
|
0.03
|
0.80
|
|
2
|
64.34
|
14.54
|
20.36
|
0.00
|
0.05
|
0.12
|
0.02
|
0.57
|
|
3
|
63.64
|
15.08
|
19.68
|
0.03
|
0.00
|
0.27
|
0.00
|
1.30
|
|
4
|
65.53
|
25.67
|
16.01
|
0.01
|
0.00
|
0.22
|
0.10
|
1.46
|
|
5
|
61.72
|
17.26
|
18.09
|
0.09
|
0.00
|
0.37
|
0.01
|
2.46
|
|
6
|
63.14
|
14.62
|
20.13
|
0.00
|
0.04
|
0.35
|
0.07
|
1.65
|
|
7
|
60.65
|
15.51
|
20.54
|
0.00
|
0.12
|
0.50
|
0.12
|
2.56
|
|
8
|
52.54
|
10.81
|
28.06
|
0.21
|
0.07
|
2.98
|
0.05
|
5.28
|
|
9
|
55.26
|
20.44
|
22.09
|
0.15
|
0.05
|
0.64
|
0.32
|
1.05
|
|
10
|
65.32
|
13.43
|
16.69
|
0.06
|
0.05
|
1.64
|
0.13
|
2.68
|
|
11
|
48.40
|
14.53
|
30.98
|
0.06
|
0.10
|
0.13
|
0.11
|
5.69
|
|
12
|
57.60
|
11.67
|
15.75
|
0.08
|
0.11
|
4.45
|
0.10
|
10.24
|
|
13
|
56.73
|
22.64
|
19.47
|
0.13
|
0.05
|
0.00
|
0.66
|
0.35
|
|
14
|
61.69
|
16.86
|
20.22
|
0.10
|
0.02
|
0.00
|
0.02
|
1.08
|
|
15
|
65.29
|
14.10
|
20.32
|
0.11
|
0.03
|
0.00
|
0.02
|
0.13
|
|
16
|
40.45
|
14.77
|
35.11
|
0.18
|
0.14
|
3.65
|
0.11
|
5.59
|
A portion of bone tissue separated during the preparation of the fractured surface, and the implant edge without bone tissue can be seen in the upper part of Fig. 14.
At higher magnification, two bone beams are observed at the interface between the hydrogel and bone tissue, which penetrate into the hydrogel (Fig. 15).
The SEM image of another section of the bone-implant interface (Fig. 16) also shows bone beam-like structures marked with arrows.
All these results demonstrate a very strong integration of the implant with the bone tissue, which actually "grows" into the hydrogel during the in vivo experiments (in this case, the duration of experiment was 90 days). Indeed, the bone tissues adhere very tightly to the hydrogel implant (Figs. 14–16) and in some areas the tissue fragments penetrate inside the implant (Figs. 15, 16).
The elemental composition of implant 4 (Fig. 17, Table 7) is close to that of implant 2. Only in scanning points 3 and 6, calcium and phosphorus are present simultaneously in significant amounts: 4.02 and 2.10 wt.% (point 3); 3.36 and 1.78 wt.% (point 6) of calcium and phosphorus, respectively. Moreover, while point 3 is located near the edge of the implant, point 6 is located approximately 300 µm to the left of the edge. That is, in this case there is no regularity in the distribution of calcium and phosphorus depending on the distance from the implant edge, which was also observed for ionic implant 2. The Ca/P ratio is close to 2 only for the two scanning points discussed above (3 and 6); in other cases, this value varies from 1.25 to 8.75. Thus, along with hydroxyapatite, other calcium phosphates may be present in this implant.
Table 7
Chemical composition of the fractured surface of implant 4 at the scanning points indicated in Fig. 17
|
Spectrum No
|
Concentrations of elements, wt.%.
|
|
|
|
C
|
N
|
O
|
Na
|
P
|
S
|
Ca
|
Ca/P
|
|
Initial implant
|
| |
65.73
|
12.39
|
21.17
|
0.71
|
0.00
|
0.00
|
0.00
|
-
|
|
Implant after in vivo experiment
|
|
1
|
64.58
|
11.77
|
22.82
|
0.41
|
0.00
|
0.00
|
0.42
|
-
|
|
2
|
64.61
|
9.56
|
21.19
|
0.00
|
0.44
|
0.35
|
3.85
|
8.75
|
|
3
|
60.08
|
9.63
|
23.82
|
0.35
|
2.10
|
0.00
|
4.02
|
1.91
|
|
4
|
66.35
|
10.44
|
22.67
|
0.20
|
0.00
|
0.15
|
0.19
|
-
|
|
5
|
59.84
|
12.54
|
26.81
|
0.34
|
0.00
|
0.08
|
0.39
|
-
|
|
6
|
63.97
|
8.85
|
21.84
|
0.15
|
1.78
|
0.05
|
3.36
|
1.89
|
|
7
|
63.01
|
10.36
|
26.23
|
0.04
|
0.12
|
0.09
|
0.15
|
1.25
|
|
8
|
65.75
|
10.00
|
22.67
|
0.03
|
0.55
|
0.03
|
0.97
|
1.76
|
|
9
|
63.03
|
13.47
|
22.62
|
0.38
|
0.00
|
0.10
|
0.40
|
-
|
|
10
|
63.35
|
11.77
|
22.05
|
0.39
|
0.00
|
0.00
|
2.44
|
-
|
|
11
|
64.84
|
12.54
|
21.81
|
0.30
|
0.00
|
0.08
|
0.43
|
-
|
It must be also kept in mind that in ionic implants calcium ions can bind to carboxyl groups as counterions, replacing sodium ions. Considering the valence of calcium ions, the formation of bridging ionic bonds between two neighboring carboxyl groups is possible; ionic crosslinking between polymer chains can also occur. The EDX data give the total content of calcium, which can exist either in the form of phosphates or as counterions of the negatively charged carboxyl groups.
In addition, a more thorough analysis of the content of elements in ionic implant 4 was carried out; their average content in rather large areas (highlighted in Fig. 18 by rectangles) was determined (spectra 1–3, Table 8).
As seen from the data presented in Table 8, the calcium content in these highlighted areas is low and varies from 0.48 to 1.8 wt.%, while the phosphorus content is equal to zero. At the same time, phosphorus is found at individual scanning points in amounts up to 0.79 wt.% (spectra 6–8). However, the presence of phosphorus at individual points does not contribute noticeably to the average concentration of this element, which is close to zero.
Thus, a fundamental difference in the degrees of mineralization of ionic and non-ionic types of composite hydrogels is shown. In ionic hydrogels, very low amounts of calcium phosphates are formed in the entire implant volume, whereas in non-ionic PC-PAAm hydrogels the content of calcium phosphates in the area contacted with the subchondral bone reaches 40 wt.%.
Table 8
Chemical composition of the fractured surface of implant 4 at the scanning points indicated in Fig. 18.
|
Spectrum No
|
Concentrations of elements, wt.%.
|
|
|
C
|
N
|
O
|
Na
|
P
|
K
|
Ca
|
|
Initial implant
|
|
| |
65.73
|
12.39
|
21.17
|
0.71
|
0.00
|
0.00
|
0.00
|
|
Implant after in vivo experiment
|
|
|
1
|
66.48
|
12.19
|
20.41
|
0.20
|
0.00
|
0.24
|
0.48
|
|
2
|
65.85
|
11.70
|
20.65
|
0.21
|
0.00
|
0.34
|
1.25
|
|
3
|
67.49
|
10.06
|
20.65
|
0.00
|
0.00
|
0.00
|
1.80
|
|
4
|
62.41
|
10.40
|
23.29
|
0.53
|
0.00
|
0.86
|
2.51
|
|
5
|
62.06
|
10.82
|
23.71
|
0.66
|
0.00
|
0.70
|
2.05
|
|
6
|
65.90
|
10.77
|
20.42
|
0.21
|
0.40
|
0.00
|
2.30
|
|
7
|
65.97
|
10.00
|
21.31
|
0.00
|
0.79
|
0.00
|
1.93
|
|
8
|
63.65
|
9.05
|
22.35
|
0.00
|
0.77
|
0.68
|
3.50
|
|
9
|
64.45
|
13.38
|
21.03
|
0.41
|
0.00
|
0.29
|
0.44
|
|
10
|
63.54
|
9.81
|
22.52
|
0.53
|
0.00
|
1.04
|
2.56
|
|
11
|
63.37
|
12.80
|
20.43
|
0.67
|
0.00
|
0.79
|
1.94
|
|
12
|
65.44
|
12.38
|
21.04
|
0.41
|
0.00
|
0.29
|
0.44
|
Most likely, this effect is due to the existence of the Donnan equilibrium between the gel containing ionic groups and the surrounding solution. For gels, this effect is discussed in detail in the monograph by Flory (Flory 1953). As a result of the action of electrostatic factors, the concentration of ions inside the gel decreases compared to their concentration in the external solution. This decrease is more significant the higher the concentration of ionic groups in the polymer network of the hydrogel. In our case, the concentration of ionic groups in implant 2 is approximately 0.3 mol/l, which is two orders of magnitude higher than the calcium concentration in body fluids, which does not exceed 3 mmol/l (Phadke et al. 2010). At such ratios between the concentrations of charged groups of the polymer network and ions in the external solution (or medium), the ions of solution will not be able to penetrate into the ionic hydrogel implant, i.e. their concentration inside the hydrogel will be close to zero. Assuming that the cells of the organism penetrate into the boundary layers of the hydrogel and facilitate the formation of calcium phosphates inside the hydrogel, the amounts of calcium ions and phosphates in the ionic hydrogel implants will be insufficient for mineralization.