BMP2-Induced Inflammation Can Be Suppressed by the Osteoinductive Growth Factor NELL-1

Antibodies and reagents

Human recombinant NELL-1 was purified by Aragen Bioscience (Purity >98%). Human BMP2 (INFUSE Bone Graft) was purchased from Medtronic. Primary antibodies used in this study were anti-TNFα (AF-510-NA; R&D) and anti-IL6 (AF506; R&D). NIH3T3 fibroblasts, Human endothelial cells (HUVEC), C2C12 myoblasts, and J774A.1 macrophages were purchased from American Type Culture Collection (ATCC). NIH3T3, C2C12, and J774A.1 cells were cultured in DMEM+10% FBS. HUVEC were grown in EBM2 MV Bulletkit (Lonza). All the cells were sustained in an incubator that maintained an environment with a temperature of 37°C and 5% CO₂.

Surgical procedure and study groups

All animal procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals of the University of California and approved by the Chancellor's Animal Research Committee. Surgical procedure was conducted as previously described.²⁶ A concentration of 4 mg/mL of BMP2 was used in this study as that concentration is proven to consistently induce inflammation.²⁶ The ratio of volume of reagent to sponge area was kept constant between all test groups and was based on the ratio recommended in the manufacturer's manual (INFUSE Bone Graft, Medtronic). Briefly, 74 male Lewis rats aged 12–14 weeks were randomized into four experimental groups according to sacrifice timepoints of 3, 7, or 14 days postoperative (Table 1). Helistat absorbable collagen sponges (ACS; thickness 1 mm, porosity 88.2%; Integra, Plainsboro, NJ) were cut into 1×2 cm strips. Sponges were saturated by dropwise application of 0.2 mL of either: (1) phosphate buffered saline (PBS), (2) BMP2 (total dose: 0.8 mg), (3) NELL-1 (total dose: 0.8 mg), or (4) NELL-1+BMP2 combined (total dose of BMP2: 0.8 mg; total dose of NELL-1: 0.8 mg), and allowed to soak for 15 min before implantation. Human BMP2 and NELL-1 were dissolved in PBS.

Animals were anesthetized by isoflurane inhalation. After anesthesia, the hind limbs were shaved, aseptically scrubbed and a transverse incision was made at the level of the midpoint of the femur. The vastus lateralis and biceps femoris were elevated through an anterolateral approach with care to keep the periosteum intact along the surface of the bone. ACS with or without growth factors was placed over the femoral surface, adjacent to the muscle and fascia. Each layer was closed with 4–0 Vicryl absorbable suture. After surgery, all animals were housed in separate cages and allowed to eat and drink ad libitum. Weight bearing was started immediately postoperatively, with daily monitoring. Buprenorphine was administered for 2 days as an analgesic, and trimethoprim/sulfamethoxazole was administered for 10 days as an antibiotic. Animals were sacrificed at 3, 7, and 14 days postoperative.

Serum TNFα and IL6 measurements

Blood (∼1 mL) was extracted at 3, 7, and 14 days postoperative. After collection of whole blood, it remained at room temperature for 20 min, allowing it to clot. Clots were removed by centrifuging at 2000 g for 10 min in a refrigerated centrifuge. Serum TNFα and IL6 was measured using rat TNFα single analyte ELISArray kit (SER06411A; SABiosciences) and rat IL6 single analyte ELISArray kit (SER06481A; SABiosciences) as per manufacturers' instructions. The serum TNFα and IL6 concentrations were calculated based on standard calibration curves.

Histological and immunohistochemical analyses

Hematoxylin and Eosin (H&E) and immunohistochemical (IHC) stains were performed as previously described.⁴³ Briefly, tissue from the biceps femoris, vastus lateralis, and the femur and its overlying periosteum adjacent to and inclusive of the collagen sponge were harvested after sacrifice, fixed in formalin, and decalcified using Cal-Ex solution (Fisher Scientific) for 7 days. All staining was performed on frozen sections. Primary antibodies were used at a dilution of 1:100. Appropriate biotinylated secondary antibodies (Dako) were used at a dilution of 1:200. ABC complex (Vectastain Elite ABC Kit; Vector Laboratories, Inc., catalog no. PK-6100) incubation was followed by development with AEC chromagen (Dako; catalog no. K346911-2). Photomicrographs were acquired using Olympus BX51 (100× and 1000× magnification lens, UPLanFL; Olympus). The intensity of staining was analyzed using commercial software Image-Pro Plus 6 and quantified by the mean optical density of staining signal x per percent area positively stained ×100.⁴⁶ The software analyzed and averaged the results from six random fields per sample. For cell counting, eight fields (1000× magnification) from each sample of each group were randomly selected to evaluate for the presence of polymorphonuclear and mononuclear cells based on the morphology. Results were reported as the mean number of inflammatory cells per high power field.

Real-time PCR

To determine the local expression of TNFα and IL6 in rats, 1.5×2×1 cm tissue (4 rats/group) around the implanted sponges was dissected and homogenized in Trizol (Invitrogen). Total RNA extraction and real-time PCR were performed as previously described.⁴⁷ Real-time PCR was performed using the 7300 real-time PCR System instrument (Applied Biosystems). The reactions were incubated in 96-well optical plates at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s, and at 60° for 10 s. The threshold cycle (Ct) data was determinate using default threshold settings. The Ct was defined as the fractional cycle number at which the fluorescence passes the fixed threshold. Mix primer sequences of mouse TNFα were sense: ATGAGCACAGAAAGCATGATC and anti-sense: TACAGGCTTGTCACTCGAATT; those of mouse IL6 were sense: ATGGATGCTACCAAACTGGAT and anti-sense: TGAAGGACTCTGGCTTTGTCT; those of rat TNFα were sense: TACTGAACTTCGGGGTGATCG and anti-sense: CCTTGTCCCTTGAAGAGAACC; those of rat IL6 were sense: AAATCTGCTCTGGTCTTCTGG and anti-sense: TTAGATACCCATCGACAGG; and primers for human TNFα were sense: TGAGGCCAAGCCCTGGTAT and anti-sense: CGAGATAGTCGGGCCGATT.

The relative expression level of the transcripts was compared to GAPDH. The primer sequences of the mouse GAPDH control were sense: CTCCCACTCTTCCACCTTCG and anti-sense: CCACCACCCTGTTGCTGTAG. The primer sequences of rat GADPH control were sense GCCTCGTCTCATAGACAAGATGG and anti-sense: TGCTCCTGCTACTTTAGACTCCG. The primer sequences of the human GAPDH control were sense: TGCACCACCAACTGCTTAGC and anti-sense: GGCATGGACTGTGGTCATGAG.

Transfection and luciferase assays

NIH3T3 and J774A.1 cells were plated at a density of 5×10⁴ cells/well on 24-well plate. Transient transfection was performed using lipofectamine 2000 (Invitrogen) according to the instructions provided by the manufacturer. Cells were transfected using 4 μg of NFκB luciferase reporter vector and 0.5 μg of renilla luciferase reporter vector as a transfection control. Thus, the total amount of DNA for each transfection was kept constant. Luciferase activity was assessed using Dual-luciferase Reporter Assay System (Promega).

Measurement of ROS

The total ROS was detected by Total ROS Detection Kit (Enzo) and performed as per manufacturer's descriptions. Briefly, NIH3T3 cells were seeded at 2×10⁵ cells/well on six-well plates in DMEM+10% FBS and starved overnight with DMEM+0.1% FBS. Cells were treated with BMP2 and NELL-1 at the indicated dose for 4 h before staining for ROS. Flow cytometry was used to detect the intensity of ROS and percentage of positive stained cells. 10,000 events were collected per sample. Samples not stained for ROS were used to quantify the average percentages of positively stained cells, serving as our negative control. DIVA version 6.1.3 was the software used to collect and analyze the data.

Statistical analyses

Means and standard deviations were calculated from numerical data. To allow multiple comparisons between groups, statistical analyses were performed using appropriate ANOVA followed by post-hoc Tukey's range test to directly test two different treatment groups.

NELL-1 inhibits BMP2-induced local inflammation

The femoral onlay procedure was performed using a collagen sponge implant (Fig. 1). The procedure was relatively atraumatic, and did not result in perceptible swelling upon gross examination of soft tissues at site of implantation. This was in agreement with our previous observations.⁴⁸ Next, we examined the histology of implant sites treated with control carrier, BMP2 alone, NELL-1 alone, or a combination of NELL-1+BMP2 (see Table 1 for a detailed breakdown of treatment groups). Histological analysis was performed at 3, 7, and 14 days postimplantation to assess for local inflammation.

H&E staining showed that control treatment (sponge with PBS) resulted in minimal inflammation with rare scattered inflammatory cells (Fig. 2A–F). In striking contrast, diffuse inflammation was observed with BMP2 treatment. At 3 days postimplantation, a dense polymorphonuclear infiltrate was observed in sponge and surrounding edematous soft tissue (Fig. 2G, H). At 7 days postimplantation, BMP2 treated samples showed a dense, mixed mononuclear and polymorphonuclear infiltrate in a background of exudative fibrous soft tissue (Fig. 2I, J). A persistent mononuclear infiltrate existed 14 days postimplanation with predominant macrophages and lymphocytes with scattered neutrophils was observed, on a background of newly formed osteoid (Fig. 2K, L).

Similar to the control treatment, treatment with NELL-1 alone resulted in minimal inflammation (Fig. 2M–R). After observing the group treated with NELL-1 14 days postoperative, it was concluded that NELL-1 did not induce ectopic bone formation (Fig. 2Q, R). This is consistent with the findings of previous studies,^43,48 where the direct delivery of adenoviral NELL-1 to muscle did not cause ectopic bone formation, unlike the effects of BMP2. Additionally, in vivo ectopic bone formation was quantified using microCT (Supplementary Fig. S1; Supplementary Data are available online at www.liebertpub.com/tea). Analysis of bone mineral density showed no significant difference across all groups (Supplementary Fig. S1A). Analysis of bone volume showed that treatment groups, including BMP2 significantly induced ectopic bone formation compared to control, while revealing no significant difference between BMP2 only and NELL-1+BMP2 groups (Supplementary Fig. S1B). Next, the inflammatory response from combination treatment of NELL-1+BMP2 was analyzed (Fig. 2S–X). At 3 days postimplantation, a striking reduction of inflammation was apparent in comparison to BMP2 treatment, with few neutrophils observed at high powered fields (Fig. 2S, T in comparison to Fig. 2G, H). This reduction in inflammation was observed throughout the 14 day study period. Abundant bone matrix was present at 14 days, and with only rare inflammatory cells present (Fig. 2W, X).

The qualitative observations of inflammation were next quantified, by counting the number of the inflammatory cells per high power field (Fig. 3). First, the presence of a polymorphonuclear infiltrate was analyzed and showed a significant increase in BMP2-treated samples at both 3 and 7 days postoperative (Fig. 3, left). This increase in acute inflammation was completely reversed by NELL-1. Next, the presence of a mononuclear inflammatory infiltrate was analyzed, and showed a statistically significant increase in BMP2-treated samples at 7 and 14 days (Fig. 3, right). This chronic inflammation was either completely or partially reversed by NELL-1 combination treatment. Thus and in summary, combination treatment with NELL-1+BMP2 was observed to mitigate the proinflammatory effects of BMP2 in the femoral onlay model.

NELL-1 inhibits BMP2-induced local inflammatory cytokine production

Next, we investigated the local expression of the proinflammatory cytokines TNFα and IL6 with BMP2 and/or NELL-1 treatment. IHC staining for TNFα confirmed our findings from H&E staining (Fig. 4). TNFα protein expression was dramatically induced at day 3 by BMP2 in comparison to control and staining was observed predominantly in areas of intense neutrophil infiltrate (Fig. 4A–F). In contrast, NELL-1 alone and NELL-1+BMP2 treatment groups both stained predominately negative for TNFα with focal areas of weak positivity (Fig. 4G–L). Semiquantification of TNFα immunostaining confirmed a significant induction by BMP2, which was reversed by NELL-1 protein addition (Fig. 4M). Local TNFα mRNA levels was also evaluated by microdissection and real-time PCR at 7 days (Fig. 6A). Similar to what has been seen from the IHC staining; BMP2 by itself significantly induced TNFα expression, which was inhibited by the addition of NELL-1.

We next performed IHC staining for IL6, which revealed a similar pattern (Fig. 5). IL6 protein expression was dramatically induced by BMP2 in comparison to control, and remained persistently elevated throughout the 14 day study period (Fig. 5A–F). In contrast, NELL-1 alone treatment exhibited nearly entire negative staining for IL6 (Fig. 5G–I). Moreover, NELL-1 reversed BMP2 induction of IL6 expression, with only sporadic single cell staining throughout each sample (Fig. 5J–L). Again, semiquantification of immunostaining confirmed a significant induction by BMP2, which was reversed by NELL-1 protein addition (Fig. 5M). Similar results were seen by real-time PCR at 7 days postsurgery for local IL6 mRNA levels (Fig. 6B). In summary, BMP2 reliably induces an acute inflammatory response with an intensely TNFα+, IL6+ infiltrate in the femoral onlay model. In contrast, NELL-1 does not induce local inflammation in the same model and at similar dosages. Moreover, treatment with NELL-1 significantly attenuates the local TNFα, IL6 inflammatory infiltrate induced by BMP2.

NELL-1 inhibits BMP2-mediated increases in serum TNFα and IL6

To further investigate the proinflammatory effects of BMP2 and the extent to which this is reversed by NELL-1, we evaluated serum levels of TNFα and IL6 at 3,7, and 14 days postoperative by cytokine specific ELISA (Fig. 7). As before, BMP2 and/or NELL-1 were implanted overlying the femoral shaft and evaluation for systemic inflammation was performed at 3, 7, and 14 days postoperative. A significant, transient increase in TNFα and IL6 serum levels was observed with BMP2 treatment at 7 days postoperative, which normalized by 14 days. In contrast, NELL-1 alone did not induce an increase in TNFα or IL6 serum levels at any time point. Excitingly, NELL-1 significantly reduced BMP2-mediated increases of both TNFα and IL6 serum peak levels at 7 days (^#p<0.05). Thus and in summary, in our femoral onlay model, treatment with BMP2 protein induced expression of inflammatory markers both locally and systemically. In contrast, NELL-1 protein significantly attenuated BMP2 induced inflammatory markers, again at both local and systemic levels.

NELL-1 inhibits BMP2 induced inflammatory response in vitro in a ROS-dependent mechanism

Finally, we sought to understand the mechanisms, by which NELL-1 suppresses BMP2-induced inflammation by performing a series of in vitro studies. NIH3T3 cells, an established cell type for the study of BMP2-induced inflammation,²⁹ were studied after being treated with NELL-1±BMP2. In striking similarity to our in vivo findings, treatment with BMP2 alone significantly increased mRNA expression of inflammatory cytokines TNFα and IL6, in a dose-dependent manner (Fig. 8A, B). The inflammation induced by BMP2, however, was significantly suppressed in the combined treatment with NELL-1 (observed across all BMP2 test doses) (Fig. 8A, B). Additionally, we evaluated TNFα expression in HUVEC (Fig. 8C), J774A.1 macrophages (Fig. 8D), and C2C12 mouse myoblast cells (Fig. 8E). Results confirmed both that BMP2 treatment increases transcription of the proinflammatory cytokine TNFα, and that NELL-1 can significantly suppress this expression.

Considering NF-κB plays a key role in regulating immune and inflammatory responses, as well as being downstream of TNFα,⁴⁹ we next performed luciferase reporter assays using NF-κB promoter construct. Interestingly, NELL-1 suppressed BMP2-induced NF-κB transcriptional factor activity in NIH3T3 (Fig. 8F) and J774A.1 (Fig. 8G) cells. We then investigated ROS, which are crucial intracellular signaling molecules that trigger various transduction pathways, including inflammation.⁵⁰ We assessed ROS generation in NIH3T3 cells treated with NELL-1±BMP2. As documented in previous studies in other cell types,^29,30 BMP2 increased ROS generation in a dose-dependent manner. Interestingly, NELL-1 significantly reduced ROS generation induced by BMP2 (Fig. 8H–K). In summary, NELL-1 attenuates BMP2 induced inflammation, associated with decreased transcription of TNFα and IL6, decreased NF-κB activity, and reduced generation of ROS.