Involvement of the Toll-like receptor 4 pathway and use of TNF-alpha antagonists for treatment of the mucopolysaccharidoses

Abstract

Enzyme replacement therapy is currently available for three of the mucopolysaccharidoses (MPSs) but has limited effects on the skeletal lesions. We investigated the involvement of the Toll-like receptor 4 (TLR4) signaling pathway in the pathogenesis of MPS bone and joint disease, and the use of the anti-TNF-alpha drug, Remicade (Centocor, Inc.), for treatment. TLR4 KO (TLR4(lps-/-)) mice were interbred with MPS VII mice to produce double-KO (DKO) animals. The DKO mice had longer and thinner faces and longer femora as revealed by micro-computed tomography analysis compared with MPS VII mice. Histological analyses also revealed more organized and thinner growth plates. The serum levels of TNF-alpha were normalized in the DKO animals, and the levels of phosphorylated STAT1 and STAT3 in articular chondrocytes were corrected. These findings led us to evaluate the effects of Remicade in MPS VI rats. When initiated at 1 month of age, i.v. treatment prevented the elevation of TNF-alpha, receptor activator of NF-kappaB, and other inflammatory molecules not only in the blood but in articular chondrocytes and fibroblast-like synoviocytes (FLSs). Treatment of 6-month-old animals also reduced the levels of these molecules to normal. The number of apoptotic articular chondrocytes in MPS VI rats was similarly reduced, with less infiltration of synovial tissue into the underlying bone. These studies revealed the important role of TLR4 signaling in MPS bone and joint disease and suggest that targeting TNF-alpha may have positive therapeutic effects.

Fig. 1.

(A) TLR4 signaling pathway can be activated via GAG fragments requiring CD44 and MyD88 or through LPS, which requires MyD88, LBP, and CD14. (B) CD44 adhesion receptor (55 kDa) is a hyaluronan-binding cell surface glycoprotein involved in the initiation of GAG-induced inflammation via TLR4. Elevated CD44 expression was detected in 6-month-old MPS VI rat FLSs and articular chondrocytes, confirming the activation of TLR4 via the GAG pathway in MPS. Protein kinase C-α (PKC-α; 82 kDa) was used as a loading control.

Fig. 2.

Localization of TLR4 in MPS VI lysosomes. (A) Rat FLSs were incubated with TLR4 primary polyclonal anti-rat antibody and LysoTracker Green to visualize lysosomes. Visualization of TLR4 was accomplished using a fluorescent secondary antibody, Cy-3, and nuclei were stained with a bis-benzamide Hoechst dye. Note the elevated levels of TLR4 in MPS VI cells and colocalization with the LysoTracker Green dye. (B) Anti-LC3-I (14 kDa) and LC3-II (16 kDa) Western blots of FLSs and articular chondrocytes from 6-month-old normal and MPS VI rats showing the lack of conversion of the LC3-I isoform to LC3-II in the FLSs. In contrast, in MPS VI articular chondrocytes, LC3-I was completely converted to LC3-II. Protein kinase C-α (Pkc-α) was used as a loading control.

Fig. 3.

(A) MPS VII and TLR4^(lps−/−) mice were crossed to produce a double-mutant (DKO) strain (MPS VII/TLR4^(lps−/−)). Note the dramatic size difference between MPS VII and DKO mice. Shown are representative 5-month-old TLR4^(lps−/−), DKO, and MPS VII male mice. (B) Skull measurements, including width (A), length (B), inner canthal distance (C), and nose widths (D), were taken of age- and gender-matched TLR4^(lps−/−), DKO, and MPS VII male mice (n = 6). TLR4^(lps−/−) mice had skull measurements that were equivalent to normal. A significant increase in skull length and reduction in width were observed in the DKO mice compared with the MPS VII mice. *P < 0.001. (C) Micro-CT analysis was performed on femora from TLR4^(lps−/−), DKO, and MPS VII male mice (n = 6). A representative image is shown, revealing that the bones were significantly longer in DKO vs. MPS VII mice. (D) Representative images of the growth plates from TLR4^(lps−/−), MPS VII, and DKO mice. Note that the growth plates were thinner in DKO vs. MPS VII mice and that the columns were more organized. (Scale bar: 15 μm.) (E) Levels of p-STAT1-Ser727 (90 kDa) and p-STAT3-Tyr705 (90 kDa) were analyzed by Western blotting in normal, MPS VII, TLR4^(lps−/−), and DKO mice. β-Actin was used as a loading control (46 kDa). MPS VII mice had elevated p-STAT1-Ser727 and reduced p-STAT3-Tyr705; these were corrected in the DKO mice. (F) TNF-α serum levels were analyzed by ELISA in 5-month-old TLR4^(lps−/−) (gray bar), MPS VII (white bar), and DKO (black bar) mice (n = 6). Significantly elevated TNF-α levels were evident in the MPS VII mice, which were normalized in the DKO mice. (G) Male 5-month-old TLR4^(lps−/−) (gray bar), MPS VII (white bar), and DKO (black bar) mice were placed on an accelerating rotarod to assess coordination and mobility (n = 6). The MPS VII mice could not complete the task and fell off between 54 (0.38 m) and 75 (0.70 m) sec after initiating the run. In contrast, the DKO mice completed the 120-sec (1.4 m) task, equivalent to the control TLR4^(lps−/−) mice. *P < 0.001.

Fig. 4.

Six-month-old MPS VI rats (n = 7) with advanced bone disease and 1-month-old presymptomatic MPS VI rats (n = 7) were injected i.v. with 3 mg/kg of Remicade every third day for 8 weeks in the older animals and 24 weeks in the younger group. Serum was collected every 2 weeks in the treated animals. (A) Remicade treatment prevented the elevation of circulating TNF-α in the presymptomatic group, although the levels were normalized in the animals with advanced disease. (B) Similar response was seen in serum RANKL levels, indicating that inhibition of TNF-α was having downstream effects. (C) Ceramide and S1P plasma levels were measured in normal (gray bar), untreated 8-month-old MPS VI (white bar), and MPS VI rats that were treated for 8 weeks beginning at 6 months of age (black bar). Note the reduced levels of both bioactive lipids in the MPS VI-treated animals. (D) Posttreatment, FLSs were isolated from the MPS VI rats as well as age-matched normal and untreated MPS VI animals. The expression of two molecules that are activated by TNF-α, cyclooxygenase-2 (COX-2; 73 kDa) and p38 (38-kDa), were examined by Western blot analysis. Note that the levels of both proteins were normalized in the treated MPS VI cells. In addition, the expression of TGF-β (30 kDa) and Sphk1 (54 kDa) were normalized in the treated animals. Protein kinase C-α (82 kDa) was used as a loading control. (E) Apoptotic (TUNEL) staining of articular cartilage was performed after 24-week treatment in MPS VI animals. Two hundred cells were counted per animal. Fifty-five percent TUNEL-positive cells were found in the untreated animals (white bar) vs. 29% in the treated animals (black bar). *P < 0.01. The differences between normal (gray bar) and treated MPS VI rats were not statistically significant.