**3.2 Which TLRs are important in RA?**

216 Rheumatoid Arthritis – Treatment

Protein in RA blood monocytes, tissue macrophages yes (Iwahashi et al.,

Protein in RA synovial fibroblasts yes (Kim et al., 2007;

**TLR7** Protein in RA synovium > OA or healthy joints yes (Roelofs et al., 2005;

F= function = the ability of the TLR to respond to its cognate ligand in each cell/tissue type

nd (Tamaki et al., 2011)

yes (Sacre et al., 2007)

yes (Seibl et al., 2003)

nd (Radstake et al., 2004)

2007)

yes (Kim et al., 2007)

yes (Brentano et al.,

yes (Ospelt et al., 2008)

yes (Sacre et al., 2007)

yes (Radstake et al.,

2008)

2005)

2004; Huang et al.,

2005; Roelofs et al.,

(Tamaki et al., 2011) (Huang et al., 2007)

2004) (Ospelt et al.,

Wu et al., 2010)

nd (Tamaki et al., 2011)

nd (Tamaki et al., 2011)

nd (Tamaki et al., 2011)

Roelofs et al., 2009)

(Tamaki et al., 2011).

**TLR EXPRESSION F REFERENCE** 

**TLR1** Protein in DCs> macrophages > fibroblasts from RA

or bone, around small vessels and in areas of

Protein in RA > OA or healthy joints in synovial lining, sublining and perivascular regions

Protein in fibroblasts from RA > OA joints > healthy

Protein in fibroblasts from early RA > OA or healthy

macrophages but not T cells or fibroblasts from RA

Protein in synovial tissue from RA > OA > healthy

**TLR5** Protein in DCs> macrophages > fibroblasts from RA

**TLR6** Protein in DCs> macrophages > fibroblasts from RA

**TLR9** Protein in DCs> macrophages > fibroblasts from RA

Table 3. Distribution of TLR expression in the RA joint

**TLR3** mRNA and protein in RA > OA or healthy synovium, in fibroblasts of the synovial lining and sublining,

**TLR4** mRNA in RA synovial tissue, protein in DCs and

joints, in early and longstanding RA

and in the perivascular areas

macrophages but not T cells or fibroblasts from RA

mRNA in RA > OA or non arthritic joints, at synovial lining, sites of attachment and invasion into cartilage

infiltrating lymphocytes (fibroblasts not macrophages

**TLR2** mRNA in RA synovial tissue, protein in DCs and

joint

joint

or T cells)

skin

synovium

joint

joint

joint

joint

nd = not determined

Evidence of a role in RA for both cell surface TLRs and endosomal TLRs in human disease is accumulating. In particular, over expression of dominant negative Mal, an adaptor protein required exclusively by TLR2 and 4, has been shown to inhibit cytokine and protease synthesis in RA synovial cells, supporting the contribution of these two family members to the synthesis of pro-inflammatory mediators in the RA joint (Sacre et al., 2007). Blockade of the function of TLR2 and 4 using neutralizing antibodies has also been reported, and while commercially available antibodies to either TLR2 or TLR4 had no effect on cytokine production in isolated RA synovial cells at 10 g/ml (Sacre et al., 2007), 1 g/ml of an anti-TLR2 antibody (OPN301) inhibited spontaneous cytokine release in RA tissue explants as effectively as anti-TNF antibodies (Nic An Ultaigh et al., 2011). Inhibition of TLR4 by the naturally occurring antagonist LPS isolated from *Bartonella Quintana*, also partially inhibited cytokine release in RA synovial biopsies (Abdollahi-Roodsaz et al., 2008). Stimulation of TLRs 2, and 4 has also been shown to induce cytokine synthesis in cell cultures isolated from RA synovia (Sacre et al., 2008). While the same workers found that TLRs 7 and 9 were not responsive to their respective ligands in RA cultures, stimulation of TLRs 3 and 8 did increase cytokine production.

The contribution of endosomal TLRs to cytokine synthesis in RA is also supported by other studies; chloroquine, which prevents intracellular TLR function by inhibiting endosomal acidification, reduces cytokine release in synovial cells (Sacre et al., 2008). The selective serotonin reuptake inhibitors, antidepressant drugs fluoxetine and citalopram and the antidepressant small molecule mianserin are also efficacious in inhibiting synovial cell cytokine release (Sacre et al., 2010). These drugs also inhibit TLR3, -7, -8 and -9 activity, by mechanisms which are yet unknown. More specifically, the small molecule imiquimod, which targets TLR8, also inhibited the production of TNF from human RA synovial membranes (Sacre et al., 2008). There have also been anecdotal reports of improved symptoms in RA patients taking anti-depressants (Krishnadas et al., 2011). Taken together these studies suggest a significant role for TLR2 and 4 as well as the endosomal TLRs 3 and 8 in human disease.

#### **3.3 The role of TLRs in animal models of RA**

In addition to studies in human tissue, the contribution of TLRs to inflammation and joint destruction has been examined in rodent models of arthritis. Mice with targeted deletions in TLRs have demonstrated that specific family members are important in driving disease pathogenesis *in vivo*.

Targeting DAMP Activation of

in these subjects.

2005; Sacre et al., 2007).

are summarized in Table 4.

disease chronicity.

mediate this activation is not clear.

Toll-Like Receptors: Novel Pathways to Treat Rheumatoid Arthritis? 219

TLR4 activation, has improved symptoms in all RA patients tested, causing disease remission in 3 out of 23 in a small clinical trial (Vanags et al., 2006). Cbio et al are also examining the recombinant analogue of HSP10, XToll®, in a phase II clinical trial for RA. The DNA-based TLR7/9 antagonist, IMO-3100, developed by Idera, has also shown promising results *in vivo* for several autoimmune disease models. Phase I clinical trials of IMO-3100 in healthy subjects are underway and it appears to be well tolerated with no major adverse effects; in addition to reducing the release of cytokines such as TNF and IL-1

Taken together, data from human, animal and pharmaceutical studies suggests a significant role for TLRs 2 and 4, in addition to the endosomal TLRs in synovial inflammation, in RA. Intriguingly however, the identity of the factor or factors that

Infection has long been purported to be a key underlying factor in RA pathogenesis. However, whilst pathogenic stimuli may trigger inflammation in RA, a causative infectious agent for RA has not been found and there is little evidence to suggest that PAMPs generate sustained joint inflammation (Schumacher et al., 1999; Chen et al., 2003). In contrast, data implicating DAMPs in RA pathogenesis have emerged from a number of independent studies in which factors derived from the serum, synovial fluid or synovial cells of RA patients can activate TLR mediated signalling pathways (Brentano et al., 2005; Roelofs et al.,

DAMPs are endogenous molecules that are immunologically silent in healthy tissues but become active upon tissue injury. They include intracellular molecules released from necrotic cells or secreted from activated cells, extracellular matrix molecule fragments created by tissue damage or proteolysis and extracellular matrix molecules that are specifically expressed upon tissue injury. In normal circumstances they act as danger signals that alert the organism to tissue damage and initiate the process of tissue repair. In addition to this physiological role however, there is evidence which indicates that DAMPS also contribute to the pathogenesis of many inflammatory and autoimmune diseases

High levels of some DAMPs are detected in the destructive milieu of the RA joint (Table 2) (reviewed in (Piccinini et al., 2010), where they are hypothesized to drive chronic inflammation by invoking a perpetual destructive cycle where inflammation leads to the creation of new stimulators of inflammation (Roelofs et al., 2008). A number of approaches have been taken to examine the effect of DAMP administration, deletion or blockade in animal models of arthritis and data supporting the role of specific molecules in such models

In particular, the administration of the fibronectin EDA domain (FNEDA), fibrinogen, HMGB-1 and tenascin-C intra-articularly to mice provokes pathological inflammation *in vivo,* (Pullerits et al., 2003; Gondokaryono et al., 2007; Midwood et al., 2009). Moreover, targeted deletion of tenascin-C protects mice from experimental disease; synovial inflammation is induced but is transient and little tissue destruction occurs in contrast to wild type mice (Midwood et al., 2009) suggesting that tenascin-C plays a crucial role in

**4. Which TLR activators drive chronic inflammation in RA?** 

characterized by aberrant TLR activation including RA.

Many experimental models of joint disease utilize TLR ligands to initiate or sustain disease induction, making interpretation of the contribution of each TLR to disease induction or progression difficult (Joosten et al., 2003; Frasnelli et al., 2005; Lee et al., 2005). However, in a serum transfer model where induction of arthritis occurs independently of TLR administration, disease was not sustained in TLR4 null mice (Choe et al., 2003). Likewise, the severity of spontaneous, IL-17 driven arthritis in mice lacking IL-1RA is significantly reduced when crossed with TLR4 null mice, concomitant with blunted expression of IL-17 suggesting a key role for TLR4. In this model TLR2 null mice showed increased disease severity whereas TLR9 knockout had no effect on disease (Abdollahi-Roodsaz et al., 2008).

Whilst most data point towards a role for TLR4 in disease progression in mice *in vivo*, recent data have also suggested a role for the endosomal TLRs. Fluoxetine and citalopram reduce disease progression in murine collagen induced arthritis (Sacre et al., 2010). TLR3 was found to be the most significantly up-regulated TLR during pristine induced arthritis in rats, where it appeared in the spleen early after disease initiation. Stimulation of TLR3 with polyI:C also exacerbated disease severity and silencing TLR3 expression reduced disease severity in these animals (Meng et al., 2010).

Despite evidence from human studies highlighting the contribution of TLR8 in synovial inflammation, the lack of activation of murine TLR8 by its cognate ligand suggests this PRR is not biologically active in mice (Heil et al., 2004). Nor is TLR10 present in mice (Hasan et al., 2005). This makes investigation of the *in vivo* function of these TLRs challenging. Combined with the fact that TLR signalling and gene activation is speciesspecific, notably most recently highlighted by examination of the differences in human and murine TLR4-mediated nickel recognition that confers contact hypersensitivity specifically to man (Schmidt et al., 2010), extrapolation of data between species should be undertaken cautiously.

#### **3.4 Targeting TLRs as a therapy in RA**

Studies examining the blockade of TLRs in mouse models have confirmed the importance of TLRs in disease and have provided evidence that TLR antagonism may be a viable means to reduce inflammation in RA. Treatment with LPS from *Bartonella quintana* and ST2 protein expressed by mast cells and T helper cell type 2 (Th2) inhibits TLR4-mediated signalling in experimental models of arthritis, resulting in disease amelioration (Leung et al., 2004; Abdollahi-Roodsaz et al., 2007). Further evidence for a role for the endosomal TLRs in CIA has also come from studies with short DNA oligonucleotides (ODNs). These act as Immunoregulatory Sequences (IRS) and inhibit endosomal TLR activity (Barrat et al., 2005; Ranjith-Kumar et al., 2008; Lenert, 2010). Prophylactic administration of ODNs in CIA and CpG-induced arthritis has been shown to abrogate disease progression (Zeuner et al., 2002; Dong et al., 2004).

In the light of the wealth of evidence implicating TLRs in both animal models of RA and in human disease considerable commercial, pharmaceutical activity has also been focused on designing TLR inhibitors for use in treating RA. TLR antagonists in preclinical development for RA include NI-0101, a TLR4 specific antibody developed by NovImmune, OPN305, a TLR2 specific antibody developed by Opsona, VTX-763, a small molecule inhibitor targeting TLR8 developed by VentiRx Pharmaceuticals and DV-1179, a DNA based TLR7/9 antagonist, developed by Dynavax. There are also several compounds currently in trial. For example, Heat shock protein 10 (HSP10) (chaperonin 10) which can inhibit LPS mediated

Many experimental models of joint disease utilize TLR ligands to initiate or sustain disease induction, making interpretation of the contribution of each TLR to disease induction or progression difficult (Joosten et al., 2003; Frasnelli et al., 2005; Lee et al., 2005). However, in a serum transfer model where induction of arthritis occurs independently of TLR administration, disease was not sustained in TLR4 null mice (Choe et al., 2003). Likewise, the severity of spontaneous, IL-17 driven arthritis in mice lacking IL-1RA is significantly reduced when crossed with TLR4 null mice, concomitant with blunted expression of IL-17 suggesting a key role for TLR4. In this model TLR2 null mice showed increased disease severity whereas TLR9 knockout had no effect on disease (Abdollahi-Roodsaz et al., 2008). Whilst most data point towards a role for TLR4 in disease progression in mice *in vivo*, recent data have also suggested a role for the endosomal TLRs. Fluoxetine and citalopram reduce disease progression in murine collagen induced arthritis (Sacre et al., 2010). TLR3 was found to be the most significantly up-regulated TLR during pristine induced arthritis in rats, where it appeared in the spleen early after disease initiation. Stimulation of TLR3 with polyI:C also exacerbated disease severity and silencing TLR3 expression reduced disease

Despite evidence from human studies highlighting the contribution of TLR8 in synovial inflammation, the lack of activation of murine TLR8 by its cognate ligand suggests this PRR is not biologically active in mice (Heil et al., 2004). Nor is TLR10 present in mice (Hasan et al., 2005). This makes investigation of the *in vivo* function of these TLRs challenging. Combined with the fact that TLR signalling and gene activation is speciesspecific, notably most recently highlighted by examination of the differences in human and murine TLR4-mediated nickel recognition that confers contact hypersensitivity specifically to man (Schmidt et al., 2010), extrapolation of data between species should be

Studies examining the blockade of TLRs in mouse models have confirmed the importance of TLRs in disease and have provided evidence that TLR antagonism may be a viable means to reduce inflammation in RA. Treatment with LPS from *Bartonella quintana* and ST2 protein expressed by mast cells and T helper cell type 2 (Th2) inhibits TLR4-mediated signalling in experimental models of arthritis, resulting in disease amelioration (Leung et al., 2004; Abdollahi-Roodsaz et al., 2007). Further evidence for a role for the endosomal TLRs in CIA has also come from studies with short DNA oligonucleotides (ODNs). These act as Immunoregulatory Sequences (IRS) and inhibit endosomal TLR activity (Barrat et al., 2005; Ranjith-Kumar et al., 2008; Lenert, 2010). Prophylactic administration of ODNs in CIA and CpG-induced arthritis has been shown to abrogate disease progression (Zeuner et al., 2002;

In the light of the wealth of evidence implicating TLRs in both animal models of RA and in human disease considerable commercial, pharmaceutical activity has also been focused on designing TLR inhibitors for use in treating RA. TLR antagonists in preclinical development for RA include NI-0101, a TLR4 specific antibody developed by NovImmune, OPN305, a TLR2 specific antibody developed by Opsona, VTX-763, a small molecule inhibitor targeting TLR8 developed by VentiRx Pharmaceuticals and DV-1179, a DNA based TLR7/9 antagonist, developed by Dynavax. There are also several compounds currently in trial. For example, Heat shock protein 10 (HSP10) (chaperonin 10) which can inhibit LPS mediated

severity in these animals (Meng et al., 2010).

**3.4 Targeting TLRs as a therapy in RA** 

undertaken cautiously.

Dong et al., 2004).

TLR4 activation, has improved symptoms in all RA patients tested, causing disease remission in 3 out of 23 in a small clinical trial (Vanags et al., 2006). Cbio et al are also examining the recombinant analogue of HSP10, XToll®, in a phase II clinical trial for RA. The DNA-based TLR7/9 antagonist, IMO-3100, developed by Idera, has also shown promising results *in vivo* for several autoimmune disease models. Phase I clinical trials of IMO-3100 in healthy subjects are underway and it appears to be well tolerated with no major adverse effects; in addition to reducing the release of cytokines such as TNF and IL-1 in these subjects.

Taken together, data from human, animal and pharmaceutical studies suggests a significant role for TLRs 2 and 4, in addition to the endosomal TLRs in synovial inflammation, in RA. Intriguingly however, the identity of the factor or factors that mediate this activation is not clear.
