Deletion of P58(IPK), the Cellular Inhibitor of the Protein Kinases PKR and PERK, Causes Bone Changes and Joint Degeneration in Mice

Abstract

Objective: Protein kinase-like endoplasmic reticulum kinase (PERK) and protein kinase R (PKR) are implicated in endoplasmic reticulum stress-induced arthritis and pro-inflammatory cytokine-mediated cartilage degradation in vitro, respectively. We determined whether knockout of the cellular inhibitor of PERK and PKR, P58(IPK) causes joint degeneration in vivo and whether these molecules are activated in human osteoarthritis (OA).

Materials and methods: Sections of knee joints from P58(IPK)-null and wild-type mice aged 12-13 and 23-25 months were stained with toluidine blue and scored for degeneration using the osteoarthritis research society international (OARSI) system. Bone changes were assessed by radiology and high-resolution micro-computed tomography of hind limbs. Sections from the medial tibial plateaus of two human knees, removed in total knee replacement surgery for OA, were immunolabelled for phosphorylated PERK and PKR and P58(IPK).

Results: Knockout mice exhibited narrower tibiae (p = 0.0031) and smaller epiphyses in tibiae (p = 0.0004) and femora (p = 0.0214). Older knockout mice had reduced total volume inside the femoral periosteal envelope (p = 0.023), reduced tibial (p = 0.03), and femoral (p = 0.0012) bone volumes (BV) and reduced femoral BV fraction (p = 0.025). Compared with wild-types, younger P58(IPK)-null mice had increased OARSI scores in medial femoral condyles (p = 0.035). Thirty four percent of null mice displayed severe joint degeneration with complete articular cartilage loss from the medial compartment and heterotopic chondro-osseous tissue in the medial joint capsule. Phosphorylated PERK and PKR were localized throughout human osteoarthritic tibial plateaus but, in particular, in areas exhibiting the most degeneration. There was limited expression of P58(IPK).

Conclusion: This study is the first to reveal a critical role for P58(IPK) in maintaining joint integrity in vivo, implicating the PKR and PERK stress signaling pathways in bony changes underlying the pathogenesis of joint degeneration.

Keywords: P58IPK; PERK; PKR; articular cartilage; bone; osteoarthritis.

Figure 1

Tibial widths and epiphyseal heights are reduced in P58^IPK-null mice. Radiographs and high-resolution μCT of the left limb from wild-type and null mice were used to measure tibial lengths (A), tibial epiphyseal heights (B), femoral lengths (C), and tibiae widths at the tibia/fibula intersect (D). Tibial bone density was measured close to the inferior tibio-fibular joint of wild-type and null mice (E), [Figure S1 in Supplementary Material]. Significant differences were detected by GLM ANOVA and Tukey’s post hoc test: *p ≤ 0.05; **p < 0.01; ***p < 0.001. For tibial length, width, and bone density measurements: 12–13-month-old mice n = 4 null, n = 3 wild-type and 23–25-month-old mice n = 5 null, n = 4 wild-type. For micro CT measurements: 12–13-month-old mice n = 3 null, n = 3 wild-type and 23–25-month-old mice n = 5 null, n = 4 wild-type.

Figure 2

Knee joint degeneration is increased in P58^IPK-null mice. Examples of toluidine blue stained coronal sections from the MTPs from P58^IPK-null (A) and wild-type (B) mice from both age groups (12–13 months: top panels; 23–25 months: bottom panels) are shown representing typical signs of joint degeneration: cartilage tears and degeneration (red arrows); subchondral bone changes (yellow arrows); and osteophyte formation (green arrows). Scale bars = 500 μm. The OARSI scoring method (Table S1 in Supplementary Material) was used to determine whether null mice had degeneration over and above that commonly observed in C57Bl/6 wild-type mice. Total scores [parameters 1–3; (C)] and bone scores [parameter 2; (D)] from the MFC and total scores [parameters 1–3; (E)] and OA scores [parameter 1; (F)] from the LTP obtained using are shown. Significant differences were detected by GLM ANOVA and Tukey’s post hoc test: *p ≤ 0.05. LTP, lateral tibial plateau; MFC, medial femoral condyle. OARSI scores: 12–13-month-old mice n = 3 null, n = 3 wild-type and 23–25-month-old mice n = 4 null, n = 4 wild-type.

Figure 3

Significant bone loss occurs in the knees of P58^IPK-null mice. MicroCT analysis of a ROI incorporating 25 lines (0.25 mm) starting from the central line of the epiphyses and moving proximally for the femur and distally for the tibia were obtained. Representative images are shown of slices taken from the top, middle and bottom of the 25 lines. (A) images from 12 to 13-month-old mice; (B) images from 23 to 25-month-old mice. In each slice, the lateral compartment is at the top and medial at the bottom.

Figure 4

A subset of P58^IPK-null mice reveal a severe degenerative joint phenotype. Coronal sections, stained with toluidine blue, are depicted from the right knee joints of three of the P58^IPK-null mice aged 12–13 months (A) and 23–25 months (F,K) old with the severe phenotype. Significant bone remodeling and cartilage loss can be seen in the MTP and MFC. Magnified views of the heterotopic chondro-osseous tissue observed in the medial capsule of these sections (box) are depicted (B,G,L) showing areas of bone marrow formation (yellow arrows) along with osteophyte formation (red circles) in two of the mice. Coronal radiographs from the contra-lateral legs of each animal depicted in (A,F,K) are shown (C,H,M) highlighting the ectopic radio-dense areas in the medial collateral ligaments and the joint capsule of two of these mice [(C,M); yellow triangles]. High-resolution μCT images of the contra-lateral leg revealed changes in joint shape and osteophyte formation on the tibiae (D,I,N) and femurs (E,J,O). Scale bar = 500 μm except for (A,F,K) = 1 mm.

Figure 5

Achilles tendon calcification is reduced in P58^IPK-null mice. (A) Radiographs from the left leg of wild-type and null mice show areas of calcification within the Achillles tendon (yellow arrow). A ROI was drawn around the calcified tissue and bone density software used to measure the area of calcification (B) with significant differences detected by GLM ANOVA and Tukey’s post hoc test: **p < 0.01; ***p < 0.001. Achilles calcification area: 12–13-month-old mice n = 5 null, n = 3 wild-type and 23–25-month-old mice n = 5 null, n = 4 wild-type.

Figure 6

Expression of phosphorylated PKR and PERK correlate with the degree of degeneration in human OA tissue. Sections from the MTP of two patients undergoing total knee replacement for OA were stained with toluidine blue to establish the degree of tissue degeneration. The subcellular localization of phosphorylated PKR and PERK were determined by immunohistochemistry. Patient 1 had complete cartilage loss and severe bone changes toward the outer edge of the plateau with the remaining cartilage showing signs of significant fibrillation and proteoglycan loss (A). Five regions (black boxes; i–v) highlight areas of limited damage (joint middle) to severe damage (outer edge). Representative staining for phosphorylated PERK (B) and PKR (C) from these five regions are shown revealing more staining in the cartilage as the damage progresses outwards and significant activation in the remodeling bone in the areas underlying the complete loss of cartilage. Patient 2 had cartilage fibrillations and proteoglycan loss, which were more extensive toward the outer edge but the damage was less extensive that patient 1 (D). Four regions (black boxes; i–iv) highlight areas of limited damage (joint middle) to severe damage (outer edge). Representative staining for phosphorylated PERK (E) and PKR (F) from these four regions are shown revealing more staining in the cartilage as the damage progresses outwards but not the extensive activity in the bone as observed for patient 1. Scale bars = 100 μm.