- Email: [email protected]
Amyloid micronetworks for cartilage repair
S158
Abstracts / Osteoarthritis and Cartilage 25 (2017) S76eS444
after injury, nor contribute to the resulting fibrocartilage patch. Furthermore, many published studies have shown that MSCs rarely directly contribute to formation of new tissues after injury, but rather they direct other cell types to promote repair. Therefore, in this study we have examined the role of endogenous Prrx1þ cells after FTCD injury in C57 mice. The MSC/progenitor marker, Paired Related Homeobox 1 (Prrx1), was employed in this study, as Prrx1þ cells can give rise to bone, fat and cartilage (cartilaginous callus) in vivo and these cells are present in the bone marrow, periosteum, fat pad and synovium within the joint. Methods: Prrx1CreERT2-GFPR26RTdTomato; C57BL/6 (C57) 8 week old mice were used for all experiments. FTCD (~0.4mm) were induced in the trochlear groove (Fig. 1 A-D). Fluorescence imaging was used to identify the location of undifferentiated Prrx1þ MSCs (GFPþ;TdTomato-) and their differentiated progeny (GFP-;TdTomatoþ) before and after cartilage injury. Immunohistochemistry was used on serial sections to identify actively proliferating cells (Ki-67þ). Data was collected 1 day or 1, 2 and 4 weeks after FTCD injury. Results: Neither Prrx1þ cells (GFPþ) or their differentiated progeny (GFP-TdTomatoþ) were observed inside AC defects in C57 mice at any time point (Fig. 1E-H).
However, Prrx1þ cells and their progeny were observed directly adjacent to AC defects at 1 week post-injury (Fig. 1 F), and were also observed in the subchondral bone (SCB), patella/tendon, periosteum, and blood vessels within the bone marrow at all time points after injury (Fig. 1 E-H). While Prrxþ cells and their progeny were observed adjacent to the defect, no GFP or TdTomato staining was observed within the resulting fibrocartilage or remodeled SCB at any time point postinjury. Actively proliferating cells (Ki-67þ) were observed at all time points in the AC injury site (Fig 1I-L). Specifically, Ki-67þ cells were located adjacent to, and also appeared to fill AC defects 1 week postinjury (Fig. 1J), but were also found within the underlying SCB as early as 1 day post-injury (Fig. 1 I-L). Interestingly, in a number of instances, Ki67þ cells were observed in ‘chains’ between the SCB and AC injury, and were also found on the surface of the AC in and around the defect area (Fig. 1 I-K), however, Ki-67þ cells did not co-localize with GFP or TdTomato staining. Conclusions: These results support the hypothesis that endogenous MSCs (Prrx1þ) do not directly contribute to the formation of new tissues (e.g. fibrocartilage or SCB) after FTCD in C57 mice. These results can neither confirm nor refute that Prrx1þ cells in the injury area are directing repair through trophic effects, but is interesting that Prrx1þ cells are observed at the margins of the defect area, but not within the defect itself. Furthermore, the resultant fibrocartilage and underlying remodeled SCB are derived from a replicating cell population likely originating from the bone marrow. Additionally, since Ki-67þ staining rarely co-localized with GFPþ/TdTomatoþ staining at the time points examined post-injury, this suggests that Prrx1þ cells do not proliferate once they have reached the defect area, but instead must proliferate and then migrate to the defect margins. Since we observed GFPþ/TdTomatocells in proximity to each other after 1 week after injury, this suggests that new Prrx1þ cells were produced from another cell type and able to migrate within this time period in response to injury. Future studies utilizing other progenitor/MSC lineage tracking mice such as GDF-5 or leptin-receptor may help elucidate the composition of the fibrocartilage patch observed during the healing response, however, from this study it is clear that Prrx1þ cells generate neither the fibrocartilage nor bone observed after FTCD in mice.
239 COMPARISON OF UNDIFFERENTIATED VS CHONDROGENIC PREDIFFERENTIATED HUMAN UMBILICAL CORD BLOOD-DERIVED MESENCHYMAL STEM CELLS FOR CARTILAGE REPAIR IN A RAT MODEL Y.-B. Park y, C.-W. Ha z, J.-A. Kim z, J. Rhim z, W.-J. Han z, S. Choi z, K. Lee z, H. Park z, H.-J. Park z. y Chung-Ang Univ. Hosp., Chung-Ang Univ. Coll. of Med., Seoul, Republic of Korea; z Samsung Med. Ctr., Stem Cell & Regenerative Med. Res. Ctr., Sungkyunkwan Univ. Sch. of Med., Seoul, Republic of Korea Purpose: Human umbilical cord blood-derived mesenchymal stem cells (hUCB-MSCs) have gained great interest as a promising cell source for regenerative medicine due to non-invasive collection, readily availability, high expansion capacity and low immunogenicity, but limited in vivo studies have reported cartilage repair. In addition, there is a lack of studies investigating the effects of chondrogenic predifferentiation of hUCB-MSCs on cartilage repair. The purpose of this study was to compare the effectiveness of transplanting chondrogenically predifferentiated hUCB-MSCs and undifferentiated hUCB-MSCs for treatment full-thickness articular cartilage defects. Methods: Full-thickness osteochondral defects (3 mm in diameter and 3 mm depth) were created in the trochlear groove of the femur of the rat. In experimental group, a composite of chondrogenically predifferentiated hUCB-MSCs (0.5 107 cells/mL) and 4% HA hydrogel was transplanted into the full-thickness defect in the right knee, whereas undifferentiated hUCB-MSCs (0.5 107 cells/mL) and 4% HA hydrogel was transplanted into the left knee. In control group, 4% HA hydrogel without MSCs was transplanted into the full-thickness defect in the right knee, whereas the full-thickness defect in the left knee was left untreated. Animals were sacrificed at 8 and 16 weeks post-transplantation. The cartilage repair was evaluated grossly and histologically, which included the International Cartilage Repair Society macroscopic score, the modified O’Driscoll score for histologic grading. Results: Transplantation of undifferentiated hUCB-MSCs resulted in overall superior cartilage repair compared to chondrogenically predifferentiated hUCB-MSCs, HA alone or no treatment. The articular surface of the defect site in the undifferentiated hUCB-MSCs was relatively smooth, which had similar coloration with the surrounding normal cartilage, compared to other groups. In addition, cellular architecture and collagen arrangement at 16 weeks were similar to those of surrounding normal articular cartilage tissue in the undifferentiated hUCB-MSCs. The histological scores also revealed that cartilage repair in undifferentiated hUCB-MSCs were good compared to chondrogenically predifferentiated hUCB-MSCs, HA alone or no treatment. Conclusions: This study demonstrated that undifferentiated hUCBMSCs and a 4% HA hydrogel composite resulted in favorable cartilage repair grossly and histologically compared to chondrogenically predifferentiated hUCB-MSCs in a rat model. These findings suggest that chondrogenic predifferentiation of hUCB-MSCs prior to transplantation do not enhance cartilage repair in osteochondral defects. 240 AMYLOID MICRONETWORKS FOR CARTILAGE REPAIR M. van Dalen, M. Karperien, M. Claessens, J. Post. Univ. of Twente, Enschede, Netherlands Purpose: To improve current cartilage repair strategies, we studied the potential of amyloid micronetworks (AMN). These networks mimic the local environment of chondrocytes, the extracellular matrix (ECM), that provides native cues to the cells. AMN are composed of self-assembled polypeptide fibrils that are popular structural components in nature. The nanoscale fibrils consist of proteins coupled by inter-protein bsheets, have a Young’s modulus comparable to collagen, and can selfassemble into AMN that resemble hydrogel particles of several tens of micrometres in diameter. Therefore, we hypothesized that AMN improve cartilage repair. Methods: Self-assembly of the similarly sized, but differently charged proteins a-synuclein (aS, -), b-lactoglobulin (bLG, -), and lysozyme (LZ, þ) into AMN was induced by incubation at defined conditions. The formation of AMN was confirmed spectroscopically. Viability and phenotype of bovine chondrocytes were investigated after 3 days of culture in the presence of AMN. Their effect on cell viability was quantified using Calcein AM and flow-cytometry. Metabolic activity
Abstracts / Osteoarthritis and Cartilage 25 (2017) S76eS444
was measured by a MTT assay. Changes in relative expression levels of cartilage marker genes were studied with qPCR. ECM formation was evaluated after culturing bovine chondrocytes with AMN for 5 weeks in 3D. Sulphated glycosaminoglycan content was visualized histologically with Safranin O. Electron microscopy was used to detect collagen bundles and qPCR was used to determine changes in cartilage maker gene expression. Results: aS, bLG, and LZ self-assembled into b-sheet-rich fibrils of several nanometres in diameter. These fibrils formed networks with diameters >10 mm and all structures stained positively for the amyloid binding dye Thioflavin T, confirming successful self-assembly into AMN. Flow-cytometry did not reveal differences in Calcein-positive cell populations due to the presence of AMN. Metabolic activity was significantly lowered if AMN were present during culture. We observed significantly increased expression levels of the ECM-related genes ACAN (proteoglycan) and COL2A1 (collagen) in the presence of LZ AMN, suggesting a positive effect of the LZ AMN on chondrocyte phenotype (Figure 1). Interestingly, aS and bLG AMN increased expression levels of collagenases MMP1 and MMP3, suggesting a possible adverse effect on tissue repair.
Figure 1. LZ AMN increase ACAN and COL2A1 expression after 3 days (mean ± SE). Cartilage ECM formation after 5 weeks correlated with the results after 3 days. LZ AMN induced increased glycosaminoglycan production, and collagen bundles with characteristic D-banding were observed (Figure 2). In contrast, aS and bLG AMN did not support ECM formation. Low amounts of glycosaminoglycan were observed and collagen production appeared hampered. ACAN and COL2A1 expression remained upregulated after 5 weeks in the presence of LZ AMN, but MMP1 and MMP3 expression returned to control levels.
Figure 2. Electron microscopy image of collagen bundles (arrows) among LZ AMN after 5 weeks.
S159
Conclusions: We show that AMN can be used for cartilage repair strategies. LZ AMN improve chondrocyte phenotype preservation during 2D culture and increase ECM formation. This effect was induced in chondrocytes without additional stimulation from differentiation factors. Surprisingly, this effect is protein-specific. Despite the similar secondary structure of the proteins in the fibrils and comparable morphology of the AMN used, the presence of aS or bLG AMN did not have a beneficial effect on either phenotype or ECM production. This effect might be caused by the net charge of the proteins; LZ is the only tested protein with a positive net charge. Based on these findings, improved cartilage repair can be achieved by choosing or designing the self-assembling polypeptide for AMN formation. 241 THE EFFECTS OF POLYGLYCEROL SULFATE-BASED HYDROGELS WITH TUNABLE MECHANICAL INTEGRITY ON CARTILAGE REGENERATION IN OSTEOARTHRITIS S. Hemmati-Sadeghi y, z, T. Dehne y, P. Dey z, B. von Lospichl x, R. Haag z, e, Berlin, Germany; z Free M. Sittinger y, J. Ringe y, M. Gradzielski x. y Charit Univ. of Berlin, Berlin, Germany; x Technical Univ. of Berlin, Berlin, Germany Purpose: The current therapeutic approaches do not halt OA progression or reverse joint damage that results in substantial ECM degradation and inflammation. Furthermore, hyaluronic acid (HA), a standard visco-supplement, injected for pain management in OA patients, has a rapid clearance. Our group has synthesized nondegradable hydrogels from a heparin-analogous polymer dendritic polyglycerol sulfate (dPGS), which can be tuned with respect to its rheological properties and has anti-inflammatory effects. In this study, we further characterized this hydrogel in our OA in-vitro model to test its effect on OA regeneration. Importantly, several concentrations from 3.6 to 4.8 wt% of dPGS and, as standard visco-supplement, blends of commercially available HAs were investigated to find out a suitable concentration for intra-articular injections which mimic HA in terms of viscoelastic and mechanical properties. Methods: To test the dPGS potential for OA-treatment, porcine chondrocytes were isolated and maintained in 96-multiwell format to establish micromass cultures (pellets). Recombinant porcine tumor necrosis factor alpha (TNF-a) was used to induce OA-like changes. To test the effects of hydrogel and later on compare it with hyaluronic acid, these reagents were parallelly added with TNF,-a but separately in concentrations of 2.5 and 0.25 wt% respectively. To document ECM formation, cartilage-typical-sulfated glycosaminoglycans (GAG) were stained with Safranin O and cartilage-specific type II collagen was detected immunohistochemically. The rheological measurements were performed with a temperature-controlled Bohlin Gemini 200 HR nano rheometer. From oscillatory measurement data, storage (G') and loss (G'') modulus were deduced as function of the oscillating frequency u. Results: Histomorphometrical analysis of Safranin O stainings demonstrated an increasing decline of proteoglycans dependent on time up to 70% loss at day 14 in TNF-a treated control (w/o) and HA groups compared to the same but non-stimulated groups. Significantly higher levels of GAG were detected in pellets treated with dPGS than in controls and HA. There were no significant difference between stimulated and non-stimulated dPGS groups (Figure 1). Immunostainings of all samples demonstrated the clear presence of cartilage-specific collagen type II in the ECM. Our current rheological findings show that an overall polymer concentration of 4.0 wt% dPGS has viscoelastic properties that are comparable to hyaluronic acid in the medically relevant frequency range of ~1 Hz. Of course, the dPGS is a permanently cross-linked gel, which hyaluronic acid forms highly viscous systems, but without a yield stress. Conclusions: Our results show that dPGS prevents TNF- a induced GAG loss and therefore can play a role in cartilage regeneration. To further investigate the mechanism behind this protective effect of dPGS, these samples are currently under analysis by Microarrays. Furthermore, in our rheological experiments, we found that the 4.0 wt% dPGS had comparable viscoelastic properties to HA. These findings suggest that dPGS hydrogels have similar mechanical properties, but might have an advantage in control of inflammation and of being much less easily disappear from its injection place.
Abstracts / Osteoarthritis and Cartilage 25 (2017) S76eS444
after injury, nor contribute to the resulting fibrocartilage patch. Furthermore, many published studies have shown that MSCs rarely directly contribute to formation of new tissues after injury, but rather they direct other cell types to promote repair. Therefore, in this study we have examined the role of endogenous Prrx1þ cells after FTCD injury in C57 mice. The MSC/progenitor marker, Paired Related Homeobox 1 (Prrx1), was employed in this study, as Prrx1þ cells can give rise to bone, fat and cartilage (cartilaginous callus) in vivo and these cells are present in the bone marrow, periosteum, fat pad and synovium within the joint. Methods: Prrx1CreERT2-GFPR26RTdTomato; C57BL/6 (C57) 8 week old mice were used for all experiments. FTCD (~0.4mm) were induced in the trochlear groove (Fig. 1 A-D). Fluorescence imaging was used to identify the location of undifferentiated Prrx1þ MSCs (GFPþ;TdTomato-) and their differentiated progeny (GFP-;TdTomatoþ) before and after cartilage injury. Immunohistochemistry was used on serial sections to identify actively proliferating cells (Ki-67þ). Data was collected 1 day or 1, 2 and 4 weeks after FTCD injury. Results: Neither Prrx1þ cells (GFPþ) or their differentiated progeny (GFP-TdTomatoþ) were observed inside AC defects in C57 mice at any time point (Fig. 1E-H).
However, Prrx1þ cells and their progeny were observed directly adjacent to AC defects at 1 week post-injury (Fig. 1 F), and were also observed in the subchondral bone (SCB), patella/tendon, periosteum, and blood vessels within the bone marrow at all time points after injury (Fig. 1 E-H). While Prrxþ cells and their progeny were observed adjacent to the defect, no GFP or TdTomato staining was observed within the resulting fibrocartilage or remodeled SCB at any time point postinjury. Actively proliferating cells (Ki-67þ) were observed at all time points in the AC injury site (Fig 1I-L). Specifically, Ki-67þ cells were located adjacent to, and also appeared to fill AC defects 1 week postinjury (Fig. 1J), but were also found within the underlying SCB as early as 1 day post-injury (Fig. 1 I-L). Interestingly, in a number of instances, Ki67þ cells were observed in ‘chains’ between the SCB and AC injury, and were also found on the surface of the AC in and around the defect area (Fig. 1 I-K), however, Ki-67þ cells did not co-localize with GFP or TdTomato staining. Conclusions: These results support the hypothesis that endogenous MSCs (Prrx1þ) do not directly contribute to the formation of new tissues (e.g. fibrocartilage or SCB) after FTCD in C57 mice. These results can neither confirm nor refute that Prrx1þ cells in the injury area are directing repair through trophic effects, but is interesting that Prrx1þ cells are observed at the margins of the defect area, but not within the defect itself. Furthermore, the resultant fibrocartilage and underlying remodeled SCB are derived from a replicating cell population likely originating from the bone marrow. Additionally, since Ki-67þ staining rarely co-localized with GFPþ/TdTomatoþ staining at the time points examined post-injury, this suggests that Prrx1þ cells do not proliferate once they have reached the defect area, but instead must proliferate and then migrate to the defect margins. Since we observed GFPþ/TdTomatocells in proximity to each other after 1 week after injury, this suggests that new Prrx1þ cells were produced from another cell type and able to migrate within this time period in response to injury. Future studies utilizing other progenitor/MSC lineage tracking mice such as GDF-5 or leptin-receptor may help elucidate the composition of the fibrocartilage patch observed during the healing response, however, from this study it is clear that Prrx1þ cells generate neither the fibrocartilage nor bone observed after FTCD in mice.
239 COMPARISON OF UNDIFFERENTIATED VS CHONDROGENIC PREDIFFERENTIATED HUMAN UMBILICAL CORD BLOOD-DERIVED MESENCHYMAL STEM CELLS FOR CARTILAGE REPAIR IN A RAT MODEL Y.-B. Park y, C.-W. Ha z, J.-A. Kim z, J. Rhim z, W.-J. Han z, S. Choi z, K. Lee z, H. Park z, H.-J. Park z. y Chung-Ang Univ. Hosp., Chung-Ang Univ. Coll. of Med., Seoul, Republic of Korea; z Samsung Med. Ctr., Stem Cell & Regenerative Med. Res. Ctr., Sungkyunkwan Univ. Sch. of Med., Seoul, Republic of Korea Purpose: Human umbilical cord blood-derived mesenchymal stem cells (hUCB-MSCs) have gained great interest as a promising cell source for regenerative medicine due to non-invasive collection, readily availability, high expansion capacity and low immunogenicity, but limited in vivo studies have reported cartilage repair. In addition, there is a lack of studies investigating the effects of chondrogenic predifferentiation of hUCB-MSCs on cartilage repair. The purpose of this study was to compare the effectiveness of transplanting chondrogenically predifferentiated hUCB-MSCs and undifferentiated hUCB-MSCs for treatment full-thickness articular cartilage defects. Methods: Full-thickness osteochondral defects (3 mm in diameter and 3 mm depth) were created in the trochlear groove of the femur of the rat. In experimental group, a composite of chondrogenically predifferentiated hUCB-MSCs (0.5 107 cells/mL) and 4% HA hydrogel was transplanted into the full-thickness defect in the right knee, whereas undifferentiated hUCB-MSCs (0.5 107 cells/mL) and 4% HA hydrogel was transplanted into the left knee. In control group, 4% HA hydrogel without MSCs was transplanted into the full-thickness defect in the right knee, whereas the full-thickness defect in the left knee was left untreated. Animals were sacrificed at 8 and 16 weeks post-transplantation. The cartilage repair was evaluated grossly and histologically, which included the International Cartilage Repair Society macroscopic score, the modified O’Driscoll score for histologic grading. Results: Transplantation of undifferentiated hUCB-MSCs resulted in overall superior cartilage repair compared to chondrogenically predifferentiated hUCB-MSCs, HA alone or no treatment. The articular surface of the defect site in the undifferentiated hUCB-MSCs was relatively smooth, which had similar coloration with the surrounding normal cartilage, compared to other groups. In addition, cellular architecture and collagen arrangement at 16 weeks were similar to those of surrounding normal articular cartilage tissue in the undifferentiated hUCB-MSCs. The histological scores also revealed that cartilage repair in undifferentiated hUCB-MSCs were good compared to chondrogenically predifferentiated hUCB-MSCs, HA alone or no treatment. Conclusions: This study demonstrated that undifferentiated hUCBMSCs and a 4% HA hydrogel composite resulted in favorable cartilage repair grossly and histologically compared to chondrogenically predifferentiated hUCB-MSCs in a rat model. These findings suggest that chondrogenic predifferentiation of hUCB-MSCs prior to transplantation do not enhance cartilage repair in osteochondral defects. 240 AMYLOID MICRONETWORKS FOR CARTILAGE REPAIR M. van Dalen, M. Karperien, M. Claessens, J. Post. Univ. of Twente, Enschede, Netherlands Purpose: To improve current cartilage repair strategies, we studied the potential of amyloid micronetworks (AMN). These networks mimic the local environment of chondrocytes, the extracellular matrix (ECM), that provides native cues to the cells. AMN are composed of self-assembled polypeptide fibrils that are popular structural components in nature. The nanoscale fibrils consist of proteins coupled by inter-protein bsheets, have a Young’s modulus comparable to collagen, and can selfassemble into AMN that resemble hydrogel particles of several tens of micrometres in diameter. Therefore, we hypothesized that AMN improve cartilage repair. Methods: Self-assembly of the similarly sized, but differently charged proteins a-synuclein (aS, -), b-lactoglobulin (bLG, -), and lysozyme (LZ, þ) into AMN was induced by incubation at defined conditions. The formation of AMN was confirmed spectroscopically. Viability and phenotype of bovine chondrocytes were investigated after 3 days of culture in the presence of AMN. Their effect on cell viability was quantified using Calcein AM and flow-cytometry. Metabolic activity
Abstracts / Osteoarthritis and Cartilage 25 (2017) S76eS444
was measured by a MTT assay. Changes in relative expression levels of cartilage marker genes were studied with qPCR. ECM formation was evaluated after culturing bovine chondrocytes with AMN for 5 weeks in 3D. Sulphated glycosaminoglycan content was visualized histologically with Safranin O. Electron microscopy was used to detect collagen bundles and qPCR was used to determine changes in cartilage maker gene expression. Results: aS, bLG, and LZ self-assembled into b-sheet-rich fibrils of several nanometres in diameter. These fibrils formed networks with diameters >10 mm and all structures stained positively for the amyloid binding dye Thioflavin T, confirming successful self-assembly into AMN. Flow-cytometry did not reveal differences in Calcein-positive cell populations due to the presence of AMN. Metabolic activity was significantly lowered if AMN were present during culture. We observed significantly increased expression levels of the ECM-related genes ACAN (proteoglycan) and COL2A1 (collagen) in the presence of LZ AMN, suggesting a positive effect of the LZ AMN on chondrocyte phenotype (Figure 1). Interestingly, aS and bLG AMN increased expression levels of collagenases MMP1 and MMP3, suggesting a possible adverse effect on tissue repair.
Figure 1. LZ AMN increase ACAN and COL2A1 expression after 3 days (mean ± SE). Cartilage ECM formation after 5 weeks correlated with the results after 3 days. LZ AMN induced increased glycosaminoglycan production, and collagen bundles with characteristic D-banding were observed (Figure 2). In contrast, aS and bLG AMN did not support ECM formation. Low amounts of glycosaminoglycan were observed and collagen production appeared hampered. ACAN and COL2A1 expression remained upregulated after 5 weeks in the presence of LZ AMN, but MMP1 and MMP3 expression returned to control levels.
Figure 2. Electron microscopy image of collagen bundles (arrows) among LZ AMN after 5 weeks.
S159
Conclusions: We show that AMN can be used for cartilage repair strategies. LZ AMN improve chondrocyte phenotype preservation during 2D culture and increase ECM formation. This effect was induced in chondrocytes without additional stimulation from differentiation factors. Surprisingly, this effect is protein-specific. Despite the similar secondary structure of the proteins in the fibrils and comparable morphology of the AMN used, the presence of aS or bLG AMN did not have a beneficial effect on either phenotype or ECM production. This effect might be caused by the net charge of the proteins; LZ is the only tested protein with a positive net charge. Based on these findings, improved cartilage repair can be achieved by choosing or designing the self-assembling polypeptide for AMN formation. 241 THE EFFECTS OF POLYGLYCEROL SULFATE-BASED HYDROGELS WITH TUNABLE MECHANICAL INTEGRITY ON CARTILAGE REGENERATION IN OSTEOARTHRITIS S. Hemmati-Sadeghi y, z, T. Dehne y, P. Dey z, B. von Lospichl x, R. Haag z, e, Berlin, Germany; z Free M. Sittinger y, J. Ringe y, M. Gradzielski x. y Charit Univ. of Berlin, Berlin, Germany; x Technical Univ. of Berlin, Berlin, Germany Purpose: The current therapeutic approaches do not halt OA progression or reverse joint damage that results in substantial ECM degradation and inflammation. Furthermore, hyaluronic acid (HA), a standard visco-supplement, injected for pain management in OA patients, has a rapid clearance. Our group has synthesized nondegradable hydrogels from a heparin-analogous polymer dendritic polyglycerol sulfate (dPGS), which can be tuned with respect to its rheological properties and has anti-inflammatory effects. In this study, we further characterized this hydrogel in our OA in-vitro model to test its effect on OA regeneration. Importantly, several concentrations from 3.6 to 4.8 wt% of dPGS and, as standard visco-supplement, blends of commercially available HAs were investigated to find out a suitable concentration for intra-articular injections which mimic HA in terms of viscoelastic and mechanical properties. Methods: To test the dPGS potential for OA-treatment, porcine chondrocytes were isolated and maintained in 96-multiwell format to establish micromass cultures (pellets). Recombinant porcine tumor necrosis factor alpha (TNF-a) was used to induce OA-like changes. To test the effects of hydrogel and later on compare it with hyaluronic acid, these reagents were parallelly added with TNF,-a but separately in concentrations of 2.5 and 0.25 wt% respectively. To document ECM formation, cartilage-typical-sulfated glycosaminoglycans (GAG) were stained with Safranin O and cartilage-specific type II collagen was detected immunohistochemically. The rheological measurements were performed with a temperature-controlled Bohlin Gemini 200 HR nano rheometer. From oscillatory measurement data, storage (G') and loss (G'') modulus were deduced as function of the oscillating frequency u. Results: Histomorphometrical analysis of Safranin O stainings demonstrated an increasing decline of proteoglycans dependent on time up to 70% loss at day 14 in TNF-a treated control (w/o) and HA groups compared to the same but non-stimulated groups. Significantly higher levels of GAG were detected in pellets treated with dPGS than in controls and HA. There were no significant difference between stimulated and non-stimulated dPGS groups (Figure 1). Immunostainings of all samples demonstrated the clear presence of cartilage-specific collagen type II in the ECM. Our current rheological findings show that an overall polymer concentration of 4.0 wt% dPGS has viscoelastic properties that are comparable to hyaluronic acid in the medically relevant frequency range of ~1 Hz. Of course, the dPGS is a permanently cross-linked gel, which hyaluronic acid forms highly viscous systems, but without a yield stress. Conclusions: Our results show that dPGS prevents TNF- a induced GAG loss and therefore can play a role in cartilage regeneration. To further investigate the mechanism behind this protective effect of dPGS, these samples are currently under analysis by Microarrays. Furthermore, in our rheological experiments, we found that the 4.0 wt% dPGS had comparable viscoelastic properties to HA. These findings suggest that dPGS hydrogels have similar mechanical properties, but might have an advantage in control of inflammation and of being much less easily disappear from its injection place.