Mechanisms in Graves Eye Disease: Apoptosis as the End Point of Insulin-Like Growth Factor 1 Receptor Inhibition

Abstract

Background:

Graves' eye disease, also called Graves' orbitopathy (GO), is a potentially debilitating autoimmune disease associated with retro-orbital inflammation and tissue expansion, involving both fibroblasts and adipocytes, resulting in periorbital edema, worsening proptosis, and muscle dysfunction with diplopia and may ultimately threaten sight. Accumulating evidence has indicated that autoantibodies to the thyrotropin receptor (TSHR), which induce the hyperthyroidism of Graves' disease, also help mediate the pathogenesis of the eye disease in susceptible individuals through TSHR expression on retro-orbital cells. Since it has long been known that the effects of insulin-like growth factor 1 (IGF-1) and thyrotropin are additive, recent clinical trials with a human monoclonal IGF-1 receptor blocking antibody (teprotumumab; IGF-1R-B-monoclonal antibody [mAb]) have demonstrated its ability to induce significant reductions in proptosis, diplopia, and clinical activity scores in patients with GO. However, the molecular mechanisms by which such an antibody achieves this result is unclear.

Methods:

We have used Li-Cor In-Cell Western, Western blot, and immunohistochemistry to define levels of different proteins in mouse and human fibroblast cells. Proteomic array was also used to define pathway signaling molecules. Using CCK-8 and BrdU cell proliferation ELISA, we have analyzed proliferative response of these cells to different antibodies.

Results:

We now show that a stimulating TSHR antibody was able to induce phosphorylation of the IGF-1R and initiate both TSHR and IGF-1R signaling in mouse and human fibroblasts. IGF-1R-B-mAb (1H7) inhibited all major IGF-1R signaling cascades and also reduced TSHR signaling. This resulted in the antibody-induced suppression of autophagy as shown by inhibition of multiple autophagy-related proteins (Beclin1, LC3a, LC3b, p62, and ULK1) and the induction of cell death by apoptosis as evidenced by activation of cleaved caspase 3, FADD, and caspase 8. Furthermore, this IGF-1R-blocking mAb suppressed serum-induced perkin and pink mitophagic proteins.

Conclusions:

Our observations clearly indicated that stimulating TSHR antibodies were able to enhance IGF-1R activity and contribute to retro-orbital cellular proliferation and inflammation. In contrast, an IGF-1R-B-mAb was capable of suppressing IGF-1R signaling leading to retro-orbital fibroblast/adipocyte death through the cell-extrinsic pathway of apoptosis. This is likely the major mechanism involved in proptosis reduction in patients with Graves' eye disease treated by IGF-1R inhibition.

Introduction

Graves' thyroid eye illness (Graves' orbitopathy [GO]) is a potentially debilitating autoimmune disease characterized by orbital inflammation and retro-orbital tissue formation, resulting in considerable proptosis and visual abnormalities, and may be sight threatening in its most severe form. Thyrotropin receptor (TSHR) antibodies of the stimulating kind (TSHR-S-Ab) appear to be key mediators of pathophysiology in vulnerable people, according to current findings (1–3).

Teprotumumab, a monoclonal insulin-like growth factor 1 receptor (IGF-1R) antagonist (blocking), has been proven in randomized controlled trials to significantly enhance composite outcomes of proptosis and clinical activity scores when compared with placebo, and to a degree not previously seen with corticosteroids (4,5). Indeed, the proptosis outcome was met in 71.4% of the teprotumumab-treated patients as compared with 20% of the placebo-treated patients (p < 0.001). The level of proptosis reduction with teprotumumab was similar to that seen with decompression surgery.

Despite this impressive clinical response, the pathogenesis of GO remains poorly understood. The current theory is that autoantibodies to the TSHR play a pivotal role in the pathogenesis by their action on TSHR-expressing retro-orbital fibroblasts inducing proliferation and adipogenesis accompanied by a retro-orbital inflammatory response with hyaluronic acid accumulation induced by cytokine and chemokine secretion (6). However, such autoantibodies to the TSHR alone cannot explain GO disease evolution in patients who are TSHR-Ab negative. This suggests that not only must T cells be intimately involved (7), but there may also be another receptor at play in the pathogenesis of GO, such as the IGF1-R (1).

TSH and insulin-like growth factor 1 (IGF-1) have been known for many years to be additive in their action on the thyroid cell, and accumulating evidence has shown it to apply equally to fibroblasts (8). Such synergy has been suggested as secondary to direct “cross-talk” between the TSHR and the IGF1-R (8). Evidence of the existence of this functional complex comprising TSHR and IGF-1R in fibroblasts indicates that, in most patients, GO is likely due to the hyperactivation of TSHR/IGF-1R downstream signaling re-enforcing orbital fibroblast proliferation, cytokine and chemokine release, and production of glycosaminoglycans within the orbit leading to increased adipose accumulation and muscle cell dysfunction (9).

IGF-1 binds to insulin-like growth factor receptors on cell surfaces, autophosphorylating the receptors (IGF-1R), and causing their activation. Activation of these receptors triggers tyrosine kinase activity and phosphorylation of signaling molecules, including insulin receptor substrates (IRS) (10,11). After they are phosphorylated, IRS proteins serve as adaptors for downstream signaling pathways through Akt, which is a key effector. Besides being an important therapeutic target outside the realm of Graves' disease, inhibiting insulin/IGF-1 signaling has been shown to be an effective anticancer influence and to improve lifespan in model organisms, including worms, flies, and mice (12–16).

Activation of retro-orbital fibroblasts has been shown to involve widespread cellular stress mechanisms (17) as we have reported extensively in thyroid cells (18,19). During such stress, one of the major cell survival mechanisms is autophagy or “self-eating.” Cells under stress exhibit a tightly regulated catabolic process in which damaged proteins and organelles are sequestered in double-membrane vesicles that are called autophagosomes. Lysosomal proteolysis degrades the cargo in autophagosomes, generating raw materials for biosynthesis of essential macromolecules. In addition, autophagy upregulation in model organisms increases longevity (20). Autophagosomes are formed in mammalian cells from precursor membrane structures containing autophagic proteins (21).

There are several proposed sources for these precursors, including the plasma membrane, the endoplasmic reticulum, and mitochondria. Apart from inducing stress, IGF-1 ligands directly induce proliferation through modulation of autophagy through Akt inhibition, since Akt inhibits autophagy-inhibiting rapamycin complex 1 (mTORC1) (22). The IGF-1 signaling pathway facilitates cell proliferation by activating the PI3/Akt/S6K and Ras/Raf/Erk cascades (23). The balance between these signaling cascades is crucial to cellular proliferation.

To better understand the therapeutic success of IGF-1R blockade, we have now probed the underlying molecular consequences of IGF-1R blockade in mouse and human fibroblast cells. This has included an evaluation of how the fibroblasts respond to stress and their ability to use cell survival pathways including autophagy and the failure of such mechanisms resulting in cell death. These observations have also uncovered a new mechanism of fibroblast cell death through activation of the cell-extrinsic apoptosis pathway.

Methods

Methods are illustrated in the Supplementary Methods section.

Results

Stimulating TSHR-monoclonal antibodies induce proliferation and activate TSHR and IGF-1R signaling in mouse fibroblasts

To determine the effect of stimulating TSH receptor (S-TSHR) monoclonal antibodies (mAbs) (MS1 and M22) on murine fibroblasts, we first analyzed the proliferative response and the activity of multiple signaling pathways using TSH and a cAMP inhibitor (KH7) as positive controls (Fig. 1A). The cAMP and proliferative responses appeared to be similar to the effect of TSH although MS-1 was less potent at inducing proliferation (Fig. 1B). S-TSHR-mAb M22 activated the major signaling cascades PI3K/Akt/mTOR and Ras/Raf/ERK1/2 (Fig. 1C). This in-depth characterization of 850 signaling molecules by proteomic array uncovered many important findings in the S-TSHR-mAb–treated fibroblasts including the observation that the S-TSHR-mAb appeared to induce IGF-1R–related signaling as shown by the detection of enhanced phosphorylated IGF-1R when using appropriate specific mAbs (Fig. 1D).

FIG. 1.

TSHR antibody signaling in mouse fibroblasts. (A) Using TSHR-stimulating mAbs (human M22 and hamster MS1), the induction of cAMP was dose dependent. KH7, a known cAMP inhibitor, inhibited constitutive cAMP production also dose dependently. TSH was used as a positive control and an isotype-mAb was used as a negative control. All data are expressed as the mean ± SEM of three experiments. *p < 0.05 by paired Student's t-test. (B) The TSHR-stimulating antibodies also induced fibroblast proliferation although M22 was the more effective and with both showing a decline with higher concentrations. All data are expressed as the mean ± SEM of three experiments. *p < 0.05 by paired t test. (C) M22-induced data are shown that illustrate activation of the PI3K/Akt/S6k/4EBP1 signaling cascade and the Ras/Raf/Mek1/2/Erk1/2/RSK1/Elk1 signaling pathways when assessed by proteomic array. A greater than 20% increase in activity was considered significant. Dashed line indicates 20% increase. (D) The same M22 mAb also activated the IGF-1R and EGFR and not just the TSHR as evidenced by induction of phosphorylation above the baseline. More than a 20% increase in activity was considered significant. As discussed, detecting such observations is highly dependent on the antibodies utilized. Multiple phosphorylation sites were involved in receptor activation including IGF-1R (Y1280 and Y/Y1165/1166), IRS1 (Y1179), EGFR (Y1172/T693), PDGFR (Y716), and Src (Y530). EGFR, epidermal growth factor receptor; IGF-1R, insulin-like growth factor 1 receptor; IRS, insulin receptor substrates; mAb, monoclonal antibody; PDGFR, platelet derived growth factor receptor; SEM, standard error of the mean; TSH, thyrotropin; TSHR, thyrotropin receptor.

Importantly, this technique demonstrated that S-TSHR-mAbs activated multiple signaling pathways that covered both TSHR and IGF-1R signaling cascades. In addition to signaling molecules pertaining to cell proliferation, the mAbs activated multiple isotypes of PKC and cell cycle-related proteins (Supplementary Fig. S1) and also activated multiple additional receptors including IRS, epidermal growth factor receptor, platelet derived growth factor receptor, and Src (Fig. 1D), which implied that such antibodies have the ability to induce diverse functions including immune-related signaling proteins such as Nuclear factor kappa B, certain Stat proteins, Syk, and Lyn (Supplementary Fig. S1). In addition to the cellular activation signals, these data indicated that antibody-induced stimulation was able to imprint immune signatures on the fibroblast population.

Signaling in fibroblasts by IGF-1R ligand

We then analyzed the fibroblast signaling response to recombinant human IGF-1 ligand. The proliferation induced by IGF-1 and human insulin is shown in Figure 2. As illustrated, rapamycin (a known inducer of autophagy) and IGF-1R inhibitor (linsitinib) suppressed such proliferation (Fig. 2A). Staurosporine a known inducer of cell death also inhibited cell proliferation. IGF-1 ligand induced autophosphorylation of the IGF-1R and activated PI3K/Akt/mTOR and Mek1/2/Erk1/2 signaling cascades in a dose-dependent manner (Fig. 2B) when measured by In-Cell Western. These findings were further confirmed by Western blot analyses (Fig. 2C).

FIG. 2.

Signaling in fibroblasts by IGF-1 ligand. (A) IGF-1 and insulin-activated fibroblast proliferation in a dose-dependent manner peaking at 0.5 μg/mL. This action was inhibited by rapamycin, linsitinib, and staurosporine. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test. (B) Increasing concentrations of IGF-1 ligand activated the PI3K/Akt/mTOR and Mek1/2/Erk1/2 pathways, and the IGF-1R itself as evidenced by changes in phosphorylation. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test. (C) Western blot data, obtained in parallel, concurred with the In-Cell Western data from which the data in (B) were derived. IGF-1, insulin-like growth factor 1.

Additive effects of S-TSHR-mAb and IGF-1

Studies of fibroblast proliferation and signaling cascades showed that both S-TSHR-mAb and IGF-1 induced proliferation and when applied together gave an additive response (Fig. 3A). The fibroblast proliferation paralleled the activity of the responsive signaling cascades when measuring pP13K, pAkt, pmTOR, pErk1/2, and pIGF-1R (Fig. 3B–E). Data shown in Figure 3F confirmed the earlier observation of an increase in IGF-1R phosphorylation by S-TSHR-mAb M22 that was also additive with IGF-1 ligand.

FIG. 3.

Additive influence of stimulating TSHR-mAb (M22) and IGF-1R ligand (L). (A) The addition of IGF-1 ligand (L) to M22 induced an additive increase in cell division. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test. (B–F) These changes in proliferation were paralleled by activation of multiple signaling molecules detected by In Cell-Western analyses and documented as changes in RFI including IGF-1R phosphorylation. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test. RFI, relative fluorescence intensity.

Blockade of IGF-1R signaling by IGF-1R-blocking mAb

To evaluate the expected effects of an IGF-1R-blocking antibody (1H7) on ligand-induced IGF-1R signaling, a fixed concentration IGF-1 ligand at 0.5 μg/mL was used to treat fibroblasts in the presence of increasing concentrations of the IGF-1R-blocking mAb (Fig. 4). IGF-1R inhibitor linsitinib was used as a control. Analysis of signaling activities including the PI3K/Akt/mTOR, Mek1/2/Erk1/2, and IGF-1R pathways in these treated cells by In-Cell Western showed that the blocking antibody inhibited all signaling cascades tested (Fig. 4A–F). These findings were paralleled by Western blot analyses (Fig. 4G).

FIG. 4.

Inhibition of IGF-1 signaling by IGF-1R-blocking mAb. (A–C) In the presence of IGF-1 both the IGF-1R-blocking mAb 1H7 and the known small molecule IGF-1R inhibitor linsitinib attenuated the signaling activity along the PI3K/Akt/mTOR pathway. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test. (D–F) Similarly, the Mek1/2/Erk1/2/pIGF-1R cascade was reduced when assessed by In Cell-Western. All data are expressed as the mean ± SEM of three to four experiments. *p ≤ 0.05, paired t test. (G) Western blot analyses are shown that paralleled (A–F).

IGF-1R-blocking mAb inhibits TSHR-mAb-induced proliferation

While S-TSHR-mAb M22 increased fibroblast proliferation, we found that this effect was suppressed by the IGF-1R-blocking 1H7 antibody (Fig. 5A). In the presence of increasing concentrations of S-TSHR-mAb M22, the 1H7 antibody inhibition of proliferation began to be overcome. Similarly, 1H7 inhibited the effect of IGF-1 ligand (Fig. 5B). With high concentrations of 1H7, M22 consistently induced less proliferation (Fig. 5C). These findings were well correlated with IGF-1R activity as indicated by the phosphorylation status of the receptor (Fig. 5D).

FIG. 5.

Inhibition of TSHR-mAb induced mouse fibroblast proliferation by IGF-1R-blocking mAb. (A) Stimulating TSHR-mAb M22-induced fibroblast proliferation was markedly reduced by IGF-1R-blocking mAb 1H7. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test. (B) IGF-1 ligand-activated proliferation was also reduced by IGF-1R-blocking mAb 1H7. *p ≤ 0.05, paired t test. (C) The proliferation induced by even increasing concentrations of M22 was able to be blocked by increasing concentrations of 1H7. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test. (D) Phosphorylation studies showed blockade of the TSHR-IGF-1R interplay by 1H7 whether initiated by M22 or IGF-1 ligand. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test.

Inhibition of autophagy by IGF-1R-blocking antibody in mouse fibroblasts

In our fibroblast culture system, serum starvation led to a survival response as evidenced by increased autophagy-related proteins as reported previously (31). When we examined the autophagy response, which involves multiple proteins including beclin 1 (BECN1), LC3A, LC3B, ULK1, and p62, we found that BECN1 and LC3A were easily detectable by immunocytochemistry (Fig. 6A(A)). Starvation could be seen to significantly induce both BECN1 and LC3A, whereas 10% serum, mAbs 1H7, and M22 all suppressed autophagy. This effect was secondary to induction of proliferation by serum and M22, whereas the suppression of the starvation effect by 1H7 was incomplete at 24 hours.

FIG. 6.

(A) Inhibition of autophagy by IGF-1R-blocking mAb. (A) ICC showed the induction of autophagy (green fluorescence with DAPI blue nuclear staining) by serum starvation and by rapamycin as evidenced by the induction of both beclin 1 and LC3A using specific antibodies. The induction of autophagy was relieved by 10% serum and TSHR-stimulating mAb M22 (1μg/mL) secondary to enhanced proliferation. The IGF-1R-blocking antibody 1H7 (1 μg/mL) also reduced autophagy without inducing proliferation and with a different staining pattern. All data are expressed as the mean ± SEM of three experiments. Red scale bar = 100 μM. (B) Western blot analyses also showed dose-dependent inhibition of autophagy proteins by 1H7 and linsitinib and induction by rapamycin. Parallel results confirmed these observations using In-Cell Western analyses (Fig. 6B). (C) Conversion of LC3A to LC3B was not evident by 1H7-blocking antibody or linsitinib IGF-1R inhibitor. (B). Inhibition of autophagy by IGF-1R-blocking mAb. (A) The changes in Beclin 1 and LC3A were shown by quantitative ICC as RFI that quantified the effects among the treated samples shown in Figure 6A illustration (A). All data are expressed as the mean ± SEM of three experiments. (B–F) In addition to the changes in beclin 1 and LC3A, In-Cell Western analyses demonstrated that autophagy proteins LC3A, LC3B, ULK1, and p62 were affected by the different treatments in parallel with the Western blot observations in 6A illustration. ICC, immunocytochemistry.

Linsitinib, as an inhibitor of the IGF-1R, and rapamycin, as an inducer of autophagy, were used as positive controls. The IGF-1R-blocking antibody 1H7 suppressed most of the autophagy survival-related proteins in a dose-dependent manner as seen on Western blot analyses (Fig. 6A (B)). Conversion of LC3A to LC3B was shown by the ratio that indicated no significant changes by the IGF-1R-blocking antibody and IGF-1R inhibitor (Fig. 6C). In-cell Western analysis also indicated similar results (Fig. 6B (A–F)).

Inhibition of autophagy and IGF-1R signaling by IGF-1R-blocking-mAb 1H7 in human fibroblasts

To determine the effect of H7 on human retro-orbital fibroblasts, we starved cells and treated them with H7 mAb in a dose-dependent manner. The IGF-1R inhibitor linsitinib was used as a positive control. Expression of autophagy proteins BECN1, LC3A, LC3B, and ULK 1 were measured quantitatively by In-Cell Western analyses as previously (Fig. 7). The IGF-1R-blocking 1H7 antibody inhibited all of the survival signals and also inhibited the expression of Mek1/2/Erk1/2 and Akt/mTOR signaling molecules and IGF-1R phosphorylation (Supplementary Fig. S2).

FIG. 7.

Inhibition of autophagy and signaling by IGF-1R-blocking mAb in human retro-orbital fibroblasts. (A–I) These illustrations show that the 1H7-blocking mAb inhibited autophagy proteins (beclin 1, LC3A, LC3B, and ULK1) induced by starvation. Data were derived from In-Cell Western analyses. These findings were paralleled by reduced signaling activities of the Mek1/2/Erk1/2 and PI3/Akt/mTOR/pIGF-1R cascades. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test.

These findings collectively were similar to that we also observed with human skin fibroblasts (Supplementary Fig. S3) and indicated that the IGF-1R-blocking mAb 1H7 blocked IGF-1R constitutive activity as shown by reduced phosphorylation (Supplementary Fig. S3F) and all signaling functions and survival signals. In addition, the pathway to autophagy was shown to be the noncannonical as the treatments failed to induce both proteins, caspase 4, and caspase 11. More importantly, suppression of mitophagy proteins, perkin and pink, indicated that 1H7 and linsitinib treatments failed to sustain mitophagy that was induced by serum starvation (Supplementary Fig. S4).

IGF-1R-blocking mAb induces cell-extrinsic apoptosis

Since cell survival signals were inhibited by mAb 1H7, we next evaluated which pathways were involved in fibroblast cell death mechanisms. We examined multiple caspases and death-associated proteins to assess the apoptosis pathways. Cells treated with increasing concentrations of the IGF-1R-blocking antibody 1H7 exhibited a robust increase in apoptosis (Fig. 8). Maximum apoptotic events were observed at 10 μg/mL. In-Cell Western analyses showed activation of multiple caspases including caspase-3 and caspase-8 but not caspase-9.

FIG. 8.

IGF-1R-B-mAb activates the extrinsic-cell apoptosis pathway in fibroblasts. (A–F) IGF-1R-blocking antibody 1H7 induced caspase 3, caspase 8, and FADD while it inhibited caspase 9, cyto-c, and Bax as shown by the RFI from In Cell-Western analyses. All data are expressed as the mean ± SEM of three experiments. *p ≤ 0.05, paired t test. (G) Similar results were also observed in parallel by Western blot assays. The indicated control was an isotype mAb. cyto-c, cytochrome-c.

FADD, a protein involved in the cell extrinsic pathway, was also used by these treated cells. To rule out intrinsic apoptotic mechanisms, Bax and cytochrome-c proteins showed no induction. This confirmed that the 1H7 blocking antibody used the cell-extrinsic apoptotic pathway by activating caspase-3 (cleaved), caspase-8, and FADD but not caspase-9 that is known to be involved in the intrinsic apoptosis pathway.

Modeling the interactions of stimulating TSHR-mAb with IGF-1R-blocking antibodies

Stimulating TSHR antibodies activated two well-characterized major signaling pathways and two IGF-1R signaling pathways through receptor phosphorylation (Fig. 9). The balance between these signaling pathways was dependent largely on the concentrations of IGF-1 ligand and TSHR-Ab that were additive. These stimuli induced robust signaling causing profound fibroblast proliferation and adipogenesis. IGF-1R-blocking antibody attenuated these signaling events. However, IGF-1R blockade by an mAb was able to overcome these signaling events, inhibiting the survival signals of autophagy, and induced fibroblast (and likely adipocyte) cell death by apoptosis. These changes would result in a marked reduction in the retro-orbital cellular mass and explain the profound clinical effects of such an antibody.

FIG. 9.

Modeling IGF-1R and TSHR interactions in GO with and without IGF-1R-blocking antibody. Our data illustrate that stimulation of the TSHR and the IGF-1R by their ligands, or by stimulating TSHR antibodies, induced an additive response causing excessive retro-orbital fibroblast proliferation with the induced cell stress known to release cytokines and chemokines causing inflammation and only minor fibroblast/adipocyte cell death. With inhibition of the IGF-1R activity by an IGF-1R-blocking antibody, there was much reduced signaling, reduced cell proliferation, a failure of autophagy for cell survival, and a large increase in cell death by apoptosis, likely reducing the pathological retro-orbital tissue mass and causing improvement in the eye disease. GO, Graves' orbitopathy.

Discussion

Recent studies have added much to clarify the immunopathology of GO. Clearly, antibodies and T cells reactive to the TSHR are heavily involved along with genetic susceptibility to develop this disorder and even the environment has been implicated (18). Clarity from all this information has been secondary to our understanding of the role of the IGF-1 receptor in this disorder. It has long been known that TSH and IGF-1 are additive in their actions on the thyroid cell with enhanced signal transduction and cell proliferation (32).

Once it became clear that extrathyroidal TSHRs may have a physiological role (32,33), for example, when the binding of TSH to fibroblasts was shown to induce a cAMP signal and proliferation (34), then the observation quickly followed that retro-orbital fibroblasts may actually overexpress the TSHR compared with fibroblasts from other sites (35). This immediately implicated this receptor in the pathophysiology of the eye disorder associated with Graves' thyroid disease. Long held observations that high levels of TSHR antibodies, the fundamental distinction of Graves' disease, were usually associated with GO added to the accumulating evidence for their important role (6).

Interestingly, few patients have undetectable TSHR antibodies presenting the possibility of other significant players such as T cells, known to aid B cells in maturation and antibody production. In these cases, T cells were given a prominent role for causing pathology in the absence of antibodies (36,37). For unclear reasons, it took many years to realize the possibility of another receptor being important and the IGF-1R was clearly a likely candidate. Subsequently, both antibodies to the IGF-1R and cross-talk between the TSHR and the IGF-1R have now been implicated in the tissue expansion and immune reaction behind the eye in Graves' disease (38).

Since the evidence for the presence of IGF-1R autoantibodies appears unconvincing at this time, in this study we have confined ourselves to the role of the IGF-1R itself since it appears to be central to this approach. The lack of specificity of IGF-1 for the thyroid and for the eyes was one major cause of the disinterest in its receptor since IGF-1 can interact with almost all cells, but the concept of its linkage to the TSHR overcame this biological doubt by providing the specificity needed for understanding thyroid eye disease immunopathology (39).

Not only does the TSHR appear to be involved in the pathological fibroblast response in Graves' eye disease and is capable of interacting with the IGF-1 receptor, but also stimulation by stimulating TSHR-Abs resulted in IGF-1R signaling as we have confirmed (Fig. 1E) (40). Furthermore, this response, as expected, showed additive effects (Fig. 3) resulting in increased signal transduction and increased proliferation. Our findings also indicated that S-TSHR-Ab exhibited robust signaling not only to IGF-1R but also to multiple receptors: importantly the EGF, PDGF, and SRC receptors (Fig. 1E).

These multiple receptor activations recruit changes in cells other than just fibroblasts to be involved in the eye disease pathology. Since TSHR activation on fibroblasts is also known to induce their differentiation into adipocytes, then such actions are also partly responsible for the marked retro-orbital adiposity in thyroid eye disease (41). Furthermore, along with such excessive cell activation comes cell stress that we have described in detail (42) and both induced cytokine and chemokine production can initiate a local inflammatory response. To survive such stress, many cells use the phenomenon of autophagy to preserve their proteins (43) (Figs. 6 and 7), thus allowing them to continue to proliferate and worsen the disease.

Cells under stress or starvation undergo the tightly regulated catabolic process known as autophagy, which leads to the formation of autophagosomes from cytosolic constituents such as damaged proteins. After autophagosomes degrade their cargo by lysosomal proteolysis, raw materials are produced and vital macromolecules are synthesized. In order for autophagosomes to be assembled from preautophagic structures, the BECN1, VPS34, VPS15, ATG14, and ULK1 (ATG13) proteins must be recruited and activated.

Evidence suggests that responses to diverse autophagy modulating factors contribute to maintaining the balance between prosurvival autophagy and proapoptotic responses. Various phosphorylation events, signaling pathways, and kinases that are involved in autophagy regulation are essential to cell survival, and failure to maintain such signaling can result in cell death. Our observations clarify that IGF-1R-blocking antibody inhibited the action of the IGF-1R and induced inhibition of autophagy in fibroblasts, leaving them without any chance to survive. In the context of multiple regulatory pathways involved in autophagy, further study should aim to clarify what signals initiate and terminate this mechanism. Although ULK1 has been suggested as the initiator in the autophagy protein complex formation, the precise mechanisms remain unclear (44).

The molecular machinery that causes apoptosis has been identified in recent years, revealing the caspases, a family of intracellular proteases that are directly or indirectly responsible for the chemical and structural alterations that characterize the phenomena of apoptosis. Activators and inhibitors of death proteases have also been revealed as regulators of apoptotic inducer proteins. Signaling mechanisms have also been revealed as inputs into the core of the cell death machinery, providing new ways to relate environmental stimuli to cell destiny decisions that result in cell suicide or homeostasis for cell survival. The present studies clearly demonstrated that IGF-1R-blocking antibody induced cell-extrinsic apoptosis as cells lost their ability to survive through autophagy: their major mechanism for cell survival.

With this logic in mind, the use of an antibody to inhibit either the TSHR or the IGF-1R appears to be excellent therapeutic routes for GO that is a difficult disease to treat in its worse form (45). The successful clinical application of a monoclonal IGF-1R antibody is proof of concept (4,46) and the detailed mechanism of why such therapy should be so effective is revealed here. Antibody inhibition of the IGF-1R clearly prevented the use of autophagy to help preserve the cells that were then forced into an apoptotic state (see Fig. 8) as evidenced by the use of the extrinsic-cell pathway for apoptosis by both human and mouse fibroblasts (47).

Conclusions

Our observations bring us to a number of general conclusions summarized in Figure 9. Stimulation of the TSHR and IGF-1R by their ligands, or by stimulating antibodies, induces an additive response causing excessive retro-orbital fibroblast proliferation, adipogenesis, and inflammation with the induced cell stress causing only minor cell death. With inhibition of the IGF-1R, there is much reduced signal transduction, reduced cell proliferation, a failure of autophagy, and a large increase in cell death by apoptosis reducing the retro-orbital tissue mass. So why does autophagy initiation fail? Simply through the lack of IGF-1R signaling that can only be overcome by high levels of TSHR antibodies making such patients more difficult to treat.

Authors' Contributions

S.A.M. performed the experiments with the help of R.M., interpreted the data and helped write the article. R.L. helped interpret the data and write the article. T.F.D. supervised the studies, interpreted the data, and wrote the article.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported, in part, by DK069713 from the National Institutes of Health, the Segal Family Endowment, and the VA Merit Award Program (to T.F.D.).

Supplementary Material

Supplementary Data

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Table S1