IFNγCXCL10

TNFα

T-bet

TH1

GCA ARTERY

Activated Macrophages/DCs

T cell recruitment and activation

GM-CSFR pathway

• Giant Cell Arteritis (GCA) is a type of large vessel vasculitis that can cause pain and malaise and more severe complicationsincluding aortic aneurysms1.

• A significant unmet medical need remains in GCA, as current treatment options are limited.

• Relapse increases corticosteroid (CS) exposure and toxicity.

• Tocilizumab blockade of IL-6 signaling significantly reduces relapses; however, ~45% of patients in the GiACTA trial did not achieve a sustained remission at 52 weeks following weekly or bi-weekly tocilizumab dosing2.

• Macrophages/dendritic cells (DCs) and T helper (TH1/TH17) lymphocytes are major contributing cell types in GCA pathogenesis1.

• Granulocyte-macrophage colony stimulating factor (GM-CSF) may contribute to GCA pathogenesis by promoting the actions of key cell types involved3,4.

• GM-CSF produced primarily by CD4+ T helper TH1 and TH17 cells can stimulate conventional DCs and promote differentiation of monocyte-derived DCs.

• GM-CSF may drive DCs to program naïve CD4+ cells to TH1, TH17, TH follicular and THGM5 phenotypes characterized by the expression of IFNγ, IL-17, IL-21, and GM-CSF respectively.

• GM-CSF may promote the proliferation and migration of vascular endothelial cells, thus contributing to angiogenesis.

• GM-CSF may stimulate giant cell formation characteristic of the disease.

• GM-CSF mRNA expression has been reported in GCA lesions6 and in peripheral blood mononuclear cells of symptomatic GCA patients7.

• Mavrilimumab (KPL-301), a human monoclonal antibody targeting the GM-CSF receptor alpha subunit (GM-CSF-Rα), is currently under investigation as a treatment for GCA (Phase 2; NCT03827018).

GM-CSF Pathway Signature Identified in Temporal Artery Biopsies of Patients With Giant Cell ArteritisMaria C. Cid1, Rohan Gandhi2, Marc Corbera-Bellalta1, Nekane Terrades-Garcia1, Sujatha Muralidharan2, John F. Paolini2

1 Vasculitis Research Unit, Department of Autoimmune Diseases, Hospital Clínic, University of Barcelona, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain; 2 Kiniksa Pharmaceuticals, Corp.

BACKGROUND

METHODS

American College of Rheumatology 2019 (ACR) | 8-13 November 2019 | Atlanta, GA

mRNA detection• RNA from fresh GCA and control temporal artery biopsies (TABs) (n=10 each) was analyzed for GM-CSF, GM-CSF-Rα, CD83,

PU.1, IFNγ, and IL-23a(p19) by qRT-PCR using TaqmanTM assays. Data were normalized to housekeeping gene GUSb and expressed as fold change with respect to average of control samples.

RESULTS

OBJECTIVES

Figure 1: Transcriptome signature by in situ hybridization (RNAScope by ACD*)

Immunofluorescence (IF) staining for protein expression • TABs (n=3 each) were fixed in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS), embedded in OCT, frozen at

−80°C, sectioned at 10 microns thickness, and incubated at 4°C overnight with antibodies against GM -CSF and GM-CSFRα followed by secondary antibodies.

• Slides were mounted with DAPI Fluoromount-G medium and examined using a laser scanning confocal Leica TCS SP5 microscope. Images were processed with Image J software.

Immunohistochemical (IHC) staining for protein expression• FFPE TAB samples (n=1-3, sectioned at 5 microns thickness) were used for IHC. Briefly, after antigen retrieval with citrate

buffer (pH6, 20 min), samples were immunostained with antibodies to PU.1 or phosphoJAK2, developed with diaminobenzidine and counterstained with hematoxylin.

• Slides were analyzed using the Leica Microsystems’ Bond-max™ automated immunostainer together with the Bond Polymer Refine Detection System (Leica Microsystems).

Ex vivo culture of temporal arteries• Control or GCA temporal arteries (n=7 each) were sectioned (0.8-1 mm), embedded in Matrigel, and cultured in

supplemented RPMI medium.

• Each artery sample was cultured in the presence of Placebo [50 mM sodium acetate, 70 mM sodium chloride, 4% (w/v) trehalose dihydrate, 0.05% (w/v) polysorbate 80, pH 5.8] or Mavrilimumab (20 mcg/ml) for 5 days prior to RNA isolation and qRT-PCR analysis and normalization as mentioned above. Supernatants from these cultures were used for ELISA analysis of relevant soluble factors.

• Each artery sample was cultured in the presence of Placebo [50 mM sodium acetate, 70 mM sodium chloride, 4% (w/v) trehalose dihydrate, 0.05% (w/v) polysorbate 80, pH 5.8] or mavrilimumab (20 mcg/ml) for 5 days prior to RNA isolation and qRT-PCR analysis and normalization as mentioned above. Supernatants from these cultures were used for ELISA analysis of relevant soluble factors.

Table 1: RS Score

RS score Scoring criteria

0 No staining or <1 dot/10 cells

1 1-3 dots/cell

2 4-9 dots/cell

3 10-15 dots/cell

4 >15 dots/cell

Abstract 2689

GCA

PU.1Control

GM-CSF-Rα

GCA

Inti

ma

Med

iaA

dve

nti

tia

Control

GCA

GM-CSF

Cell nuclei (blue); indicated target (red)

Control

Figure 2B.

Figure 2: GCA arteries exhibit increased expression of GM-CSF and GM-CSFR pathway-associated genes (mRNA and protein) which indicates activation of GM-CSFR pathway in disease relevant tissue regions of GCA

• Fresh GCA and control TABs (n=7 each) were sectioned, embedded in Matrigel®, and cultured in supplemented RPMI medium for 5 days. Supernatants were used for soluble factor ELISA, and cells were used for RNA isolation and expression of relevant genes analyzed by qRT-PCR using TaqmanTM assays. Data were normalized to housekeeping gene GUSb and expressed as fold change with respect to average of control samples.

• RNAScope® (RS) was used to assay mRNA expression of multiple genes in commercially sourced FFPE sections of GCA (n=18) and control (n=5) TABs (Figure 1). RNAScope in situ hybridization was performed on FFPE sections of GCA and control biopsies to detect 15 mRNA transcripts associated with TH1, TH17, and GM-CSFR pathways. Expression score was calculated as RS score (dots/cell, Table 1) multiplied by % cells positive (> 1 dot/cell).

• To examine expression of GM-CSF cytokine and receptor in GCA arteries compared to control arteries at the mRNA and protein level.

• To explore the granulocyte-macrophage colony-stimulating factor receptor (GM-CSFR) pathway associated transcriptomic signature in GCA arteries compared to control arteries.

• To measure the effect of mavrilimumab on genes relevant to GCA pathophysiology in ex vivo GCA artery cultures.

Figure 2A.

Figure 2C.

Figure 2D.

Figure 2E.

Figure 5: Mavrilimumab suppresses the increased expression of genes associated with GM-CSFR pathway, inflammation, and immune cell infiltration in cultured GCA arteries

• Increased expression of GM-CSF, GM-CSF-Rα, and TH1 pathway-associated genes was observed in different sets of GCA arteries by independent mRNA analytical techniques

• Increased GM-CSF, GM-CSF-Rα, and downstream pathway-associated protein levels in GCA biopsies were consistent with increased transcriptome signature

• Expression of genes associated with GM-CSFR pathway and inflammatory cell infiltration was suppressed by mavrilimumab in cultured GCA arteries

• These data implicate the GM-CSF pathway in GCA pathophysiology and support confidence in rationale for targeting the GM-CSF pathway in GCA (Figure 6)

Figure 6: GM-CSF pathway in GCA pathophysiology

REFERENCES:1. Terrades-Garcia & Cid. Rheumatology, 2018; 57(2):51-62; 2. Stone et al. N Engl J Med, 2017; 377:317-328; 3. Lemaire et al. Journal of Leukocyte Biology, 1996; 60(4):509-18; 4. Wicks & Roberts. Nature Reviews. Rheumatology, 2016; 12(1):37-48; 5. Herndler-Brandstetter & Flavell. Cell Research, 2014; 24(12): 1379-80; 6. Weyand et al. Annals of Internal Medicine, 1994, 121(7):484-91; 7. Terrier et al. Arthritis and rheumatism, 2012; 64(6):2001-11; 8. Deng et al. Circulation, 2010; 121(7): 906–915; 9. Zielińska et al. Frontiers in Immunology, 2014; (7):592; 10. Burmester et al. Ann Rheum Dis. 2013 Sep 1;72(9):1445-52.

DISCLOSURES:This research was funded by Kiniksa Pharmaceuticals, Ltd. M. Cid, presenting author, has received research grants from Kiniksa and consulting fees from Janssen and Abbvie; co-authors R. Gandhi, S. Muralidharan, and J.F. Paolini are employees of Kiniksa Pharmaceuticals Corp.

CONCLUSIONS

Figure 3A. Figure 3B.

Figure 3C.

Figure 4: GCA arteries exhibit lower than expected expression of TH17-associated genes at mRNA and protein level, potentially due to empiric CS treatment prior to biopsy

Figure 4A. Figure 4B.

Figure 4C.

Figure 3: GCA arteries exhibit elevated expression of TH1-associated genes including IFNγ at mRNA and protein level, consistent with previous findings

• Expression of GM-CSF and GM-CSFR pathway-associated markers was increased in GCA arteries compared to control arteries (Figure 2)

• Significant increase in GM-CSF, GM-CSFRα, PU.1 (GM-CSFR pathway transcription factor) and CD83 (mature/activated DC marker) observed in bulk mRNA (Figure 2A) was confirmed at mRNA level in a different set of GCA samples and was detected in all layers of GCA artery (Figure 2B, 2C)

• Protein levels for GM-CSF and GM-CSF-Rα (target of mavrilimumab) appeared higher in GCA arteries consistent with mRNA levels (Figure 2D)

• Protein levels of phosphorylated JAK2 and nuclear localized PU.1 also appeared higher in GCA arteries indicating active GM-CSFR signaling pathway (Figure 2E)

• Expression of TH1-associated markers was increased in GCA arteries compared to control arteries, which is consistent with previously published8 data showing increased presence of TH1 cells in GCA (Figure 3)

• Significant increase in IFNγ observed in bulk RNA (Figure 3A) was confirmed at mRNA level in a different set of GCA samples and appears localized to intima and media of GCA artery (Figure 3B)

• Trend for increased expression of IFN-stimulated CXCL10 was observed at protein level in supernatants from cultured GCA arteries (Figure 3C)

Control GCA

Cell nuclei (blue); PU.1 (brown)

100x 100x

PU.1

200x

Control GCA

100x 100x

phospho-JAK2

Cell nuclei (blue); phospho-JAK2 (brown)

Cell nuclei (blue); GM-CSF (green)

Control GCA

40x 40x

GM-CSF

40xControl GCA

40x 40x

GM-CSF-Rα

Cell nuclei (blue); GM-CSF-Rα (green)

GM-CSF

GM-CSFR

JAK/STAT PU.1

GCA ARTERY

Activated Macrophages/DCs

INFLAMMATORY DAMAGE, GIANT CELL FORMATION, VASCULAR REMODELING

Marker

Percent increased mRNA in GCA vs. Control

qRT-PCR RNAscope

GM-CSF 840% 250%

GM-CSF-Rα 200% 460%

PU.1 200% 760%

CD83 190% 460%

Figure 2F.

• Significant increases in TH1 transcription factor Tbx21 (T-bet) and TH1 cytokine TNFα mRNA levels were also observed in intima and media of GCA arteries (Figure 3B)

• Expression of TH17-associated markers was not significantly different in GCA arteries compared to control arteries. This suggests patients may have been treated with CS prior to biopsy, which could cause downregulation of TH17 cells (TH17 cells particularly sensitive to CS8,9) (Figure 4)

• No significant changes were observed in TH17 polarizing cytokine IL-23, TH17 transcription factor RORγ and TH17 cytokine IL-17A mRNA levels in independent sets of GCA samples analyzed by qRT-PCR or RNAscope (Figure 4A, 4B)

• No significant changes were observed in TH17 polarizing cytokine IL-6 at protein level in supernatants from cultured GCA arteries (Figure 4C)

• Mavrilimumab treatment of ex vivo GCA artery cultures suppressed the previously-observed elevated expression of GM-CSFR pathway and inflammatory cell infiltration associated markers in GCA indicating the biological effect of mavrilimumab on cells/molecules relevant to GCA pathophysiology (Figure 5)

• Significant decreases in PU.1 mRNA levels were observed with mavrilimumab treatment indicating inhibition of GM-CSFR pathway (Figure 5A)

Marker

Percent increased in GCA vs. Control

qRT-PCR RNAscope ELISA

Tbx21 (T-bet) 110%

TNFα 165% N.D.

IFNγ 260% 90% N.D.

CXCL10 100%

N.D. Near low detection limit; Purple values p<0.05

Purple values p<0.05

Figure 3D.

Figure 4D.

Marker

Percent increased in GCA vs. Control

qRT-PCR RNAscope ELISA

RORc 0%

IL-17A N.D 20%

IL-23A 20% 55%

IL-6 30%

N.D. Near low detection limit

TH1

GCA ARTERY

Activated Macrophages/DCs

T cell recruitment and activation

GM-CSFR pathway

TH17

CS

IL-17A

IL-23AIL-6

RORγ

T-cell-produced GM-CSF

• No significant changes were observed in GM-CSF and DC activation markers CD83 and HLA-DR mRNA levels with mavrilimumab treatment although there was a downward trend in majority of samples for CD83 and HLA-DR in presence of mavrilimumab which could indicate inhibition of downstream activation in GM-CSFR pathway (Figure 5A)

• Significant decreases in T cell markers including TNFα and CXCL10 mRNA levels were also observed with mavrilimumab treatment although overall IFNγ mRNA levels were not significantly changed (Figure 5B)

• Decrease in CXCL10 protein levels, although not significant, was observed with mavrilimumab, and TNFα and IFNγ protein levels were too low for detection in this artery culture experiment (Figure 5C)

• Significant decrease in IL-6 protein levels was consistent with observed decrease in IL-6 mRNA levels, although not significant, with mavrilimumab treatment (Figure 5B, 5C). Reductions in plasma IL-6 have been previously demonstrated in Phase 2 RA patients treated with mavrilimumab10

Figure 5A. Figure 5C.

Figure 5D.

MAVRILIMUMAB

Figure 5B.

* Images Courtesy of ACD

RNASCOPE

1. PermeabilizeUse RNAscopePretreatment Kit to unmask target RNA and permeabilize cells

2. Hybridize~20 target-specific double Z probes hybridize to target RNA molecules

3. AmplifyAmplify hybridization signal via sequential hybridization of amplifiers and label probes

TH1: Tbx21, IFNγ, TNFα, IL1B, IL12A, IL12BTH17: IL17A, IL23A, RORCGM-CSFR pathway: GM-CSF, GM-CSF-Rα, CD83, JAK2, PU.1

T-cell-produced GM-CSF