Mechanisms in Thyroid Eye Disease: The TSH Receptor Interacts Directly With the IGF-1 Receptor

Abstract

The pathogenesis of thyroid eye disease (TED) has been suggested as due to signal enhancement in orbital fibroblasts as a result of autoantibody-induced, synergistic interaction between the TSH receptor (TSHR) and the IGF-1 receptor (IGF-1R). This interaction has been explained by a "receptor cross-talk," mediated via β-arrestin binding. Here we have examined if this interaction can be mediated via direct receptor contact using modeling and experimental approaches. First, we docked a model of the leucine-rich domain of the TSHR ectodomain (ECD) to an available cryo-electron microscopy-based structure of the active-state IGF-1R dimer and demonstrated the stability of the complex using molecular dynamics simulations. We then extended the complex with the full-length TSHR and the transmembrane helices of the IGF1R and a 3000 ns simulation also showed stability of this complex. We then performed coimmunoprecipitation studies with anti-TSHR and anti-IGF-1R antibodies using cells expressing the IGF-1R and the full-length TSHR and also cells that expressed the IGF-1R and only the TSHR-ECD and, therefore, unable to bind β-arrestin. These studies showed a 360 kD complex protein in the immunoprecipitation, which was present in both the full-length TSHR and the TSHR-ECD-only expressing cells, evidencing a direct interaction of receptors via their ectodomains in the absence of arrestin. Colocalized staining of TSHR and IGF-1R in the TSHR-ECD cells further supported this direct interaction. These data showed that the TSHR and IGF-1R can interact directly and in the absence of β-arrestin binding. Understanding these interactions is important in the pathogenesis of TED and its therapeutic intervention.

Keywords: insulin-like growth factor 1 receptor; molecular dynamics; thyroid eye disease; thyroid stimulating hormone receptor; transmembrane domain.

Figure 1.

Orientation of models of IGF-1R and TSHR with respect to the membrane: shown on the left (A) is our published model structure of the full-length TSHR (9) consisting of 3 domains: LRD (gray), LR (red), and TMD (blue). Shown on the right (B) is a partial structure of the active human IGF-1R dimer's extracellular domain (PBDid 6vwj) linked to the model of the single-helix transmembrane domain; the 2 monomers are blue and red, respectively.

Figure 2.

Representative pose of the IGF-1R-TSHR complex. The ClusPro-generated pose of the IGF-1R-TSHR complex is shown both in cartoon and space-filling representations. Red and blue: the 2 IGF-1R monomers, respectively, gray: the LRD of the TSHR.

Figure 3.

History of contacts between the ectodomains. The contact history of IGF-1R-TSHR-LRD residue pairs that were in contact at least 20% of the time. Each line represents a residue pair; lines are broken when that residue pair is not in contact. The respective residues of each receptor and the percentage of time under which the residues from TSHR LRD and that from the IGF-1R remain together are shown on the left.

Figure 4.

History of contacts between the transmembrane helices of IGF-1R. The contact history of the residue pairs on different transmembrane helices of the IGF-1R that were in contact at least 20% of the time with each other suggesting movement and proximity of the TMDs of IGF1R.

Figure 5.

Complete model of the IGF-1R-TSHR complex: the full-length structure of the TSHR represented here in grey, which is in contact via its LRD with the ectodomain of 1 of the monomeric structures of IGF-1R shown in blue. The TMD loops of the IGF-1R were built using the ModLoop server.

Figure 6.

Representative images from a long MD simulation. Cartoon representations of a randomly selected conformation of the IGF-1R-TSHR complex from the extended MD simulation (grey: TSHR; red and blue: the 2 IGF-1R monomers, respectively). The TMDs of each receptor are shown as transparent spheres. These representations show conformations changes at the ectodomain and with the TMDs and indicate that IGF-1R TMD is in close contact with the TSHR TMD.

Figure 7.

Space-filling model of IGF-1R and TSHR ECD complex after 3000 ns of MD simulation. Space-filling representation of the overlay of the representative structure of the largest blocks of IGF-1R (monomers of IGF-1R: blue and red) with TSHR ECD shown as transparent grey spheres.

Figure 8.

Hydrogen bonding between the IGF-1R and the TSHR. (A) The LRD (cartoon representation) with residues, shown as spheres, forming hydrogen bonds with the IGF-1R more than 50% of the time; blue: 50%-60%, green: 60%-70%, yellow: 70%-80%, red: 80%-100%. (B) Histories of the IGF-1R-TSHR residue pairs being hydrogen bonded at least 20% of the time.

Figure 9.

Secondary-structure elements of the TSHR LR in the IGF-1R-TSHR complex. DSSP plot of the LR showing the different secondary-structure elements formed during the simulation.

Figure 10.

Superimposed models of M22 and TSH to IGF-1R-TSHR complex. (A) M22 antibody (Fab and Fc) LRD (21) superimposed on the TSHR LRD of the 3000 ns IGF-1R-TSHR complex. (IGF-1R: red and blue, TSHR: gray, M22: transparent spheres, colored by atom names.) (B) TSH LRD superimposed to the LRD of the IGF-1R-TSHR complex at 3000 ns (IGF-1R: red and blue, TSHR: gray, TSH: transparent yellow).

Figure 11.

Coimmunoprecipitation showing the interaction of TSHR ECD with IGF-1R. The blot on the left shows the 360KD complex, which consists of the TSHR protein interacting with IGF-1R as seen in the CHOTSHR (JPO9) lane and CHOTSHR- ECD (G13 cells) lane but not in only the CHO lane. For this experiment, the lysate was immunoprecipitated with anti-TSHR and blot probed with anti IGF-1R antibody (boxed area). The graph on the right shows the signal-to-noise ratio from these bands as a graph as a single read of the band intensity calculated by compass software in the WES protein simple machine.

Figure 12.

Colocalization of TSHR ECD with IGF-1R in CHOTSHR ECD (G13) cells. (A) CHO TSHR ECD (G13 cells) fixed and permeabilized costained with anti-TSHR (MC1 1;200) antibody and probed with anti-mouse Alexa 488 (1:500); (B) IGF-1R was stained using a rabbit IGF-1R (1:500) and detected with Alexa 594. (C) Overlaid image. Controls are CHO cells stained in a similar manner as shown in the inset under each panel.