Background
Rheumatoid arthritis (RA) is a chronic and debilitating autoimmune disease that causes joint damage, decreased quality of life and cardiovascular complications, among other comorbidities. With 5–50 per 100,000 new cases annually, RA occurs in 0.5–1 % of adults in industrialized countries, being more frequent in women and elderly people [
1].
Interleukin-7 (IL-7) is an anti-apoptotic cytokine, essential for T cell proliferation, development and homeostasis. It is also involved in B cell development. The IL-7 receptor (IL-7R) comprises two heterodimeric subunits, IL-7Rα and the common γ-chain (γc) respectively. IL-7Rα is composed of a 195 amino acid intracellular domain, a 25 amino acid transmembrane domain and an extracellular region comprising 219 amino acids. Four invariant cysteine residues located at the N-terminus of the extracellular domain are involved in intrachain disulfide bond formation. A Trp-Ser-X-Trp-Ser (WSXWS) motif is located close to the transmembrane domain of the extracellular region of IL-7Rα, which also contains a fibronectin (FN) type III-like domain. The intracellular domain of IL-7Rα is responsible of signal transduction via the recruitment of signal transducing molecules, such as the Janus kinase 1 (Jak1), the signal transducer and activator of transcription 5 (STAT5) and the src family of tyrosine kinases, and is involved in the IL-7-dependent activation of phosphatidylinositol 3-kinase (PI3K) [
2,
3]. IL-7 binding to its receptor triggers several signaling cascades, i.e., Jak/STAT, PI3K, Ras and mitogen-activated protein kinase (MAPK)/extracellular signal-related kinase (ERK), being essential for lymphocyte survival, homeostasis and differentiation [
2‐
4].
IL-7 and IL-7R are over-expressed in the RA synovium. IL-7 plays a crucial role in T cell activation and osteoclastogenesis by upregulating T cell-derived cytokines, including the receptor activator of nuclear factor-κB ligand (RANKL) [
5]. In RA synovitis, not only T cells, but also synovial macrophages and fibroblasts over-express IL-7Rα, thereby making IL-7Rα the transcript most differentially expressed between RA and other inflammatory joint conditions such as osteoarthritis, systemic lupus erythematosus, psoriatic arthritis and gout [
6‐
8]. Synovial fibroblasts also produce a high quantity of soluble IL-7Rα subsequent to their stimulation by cytokines such as tumor necrosis factor-α (TNFα), IL-1 and IL-17 [
9‐
11]. According to several observations, soluble IL-7Rα stabilizes IL-7 and amplifies its T cell stimulatory effects [
12]. The increased expression and high serum concentrations of soluble IL-7Rα is associated with poor response to anti-TNFα therapy in patients with RA [
13,
14]. The pathogenic role of the IL-7/IL-7R axis in RA is further illustrated in a mouse model of the disease (collagen-induced arthritis, CIA), in which both IL-7 or IL-7R blockade using monoclonal antibodies results in significant improvements in disease activity [
15‐
17]. Of note, recent observations also highlight the potential involvement of IL-7/IL-7R in other autoimmune disorders, such as systemic lupus erythematosus and Sjögren’s syndrome [
18,
19].
As it is currently an incurable disease, diagnosis and treatment of RA before its progression towards a debilitating stage is imperative for patients. Magnetic resonance imaging (MRI) is reported to be the best clinical imaging technique for the diagnosis of RA, allowing observation of the characteristic inflammation and lesions that are not adequately displayed using conventional radiography. Molecular imaging of specific pathological processes in synovitis would increase the likelihood of early diagnosis, disease staging and monitoring. Any of the molecular actors involved in chronic inflammation, cell proliferation and apoptosis represents putative targets for functionalized imaging probes, thus optimizing the diagnostic capability of clinical imaging techniques [
20‐
23].
Our work is integrated within this clinical and scientific context, by trying to develop molecular tools to image IL-7Rα in vivo as a diagnostic biomarker, and as a marker of response to RA therapy. To achieve this goal, a randomized cyclic heptapeptide phage display library was screened against IL-7Rα, leading to the identification of two peptides specific to this biomarker, one of them being a putative blocking agent for IL-7 (called P725). The peptides were characterized in terms of affinity, biologic activity and diagnostic potential by MRI. The peptide dedicated to MRI applications (called P258) was coupled to ultra-small superparamagnetic particles of iron oxide (USPIO, an MRI contrast agent producing a negative contrast) and its pharmacokinetics, biodistribution and diagnostic potential were evaluated in mice. USPIO present a particular interest for molecular imaging due to their excellent MRI efficacy, biocompatibility and biodegradability. The blood half-life of this material is significantly prolonged by coating it with hydrophilic polymers, such as poly(ethylene glycol) (PEG), which reduce its opsonization and the clearance by the reticuloendothelial system, thus improving the targeting of specific biomarkers [
24,
25]. The vectorizing peptide P258 and PEG used as a coating material were covalently coupled to the carboxyl groups exposed at the surface of USPIO.
Discussion
The clinical imaging methods currently used for RA diagnosis and monitoring (i.e., radiography, computed tomography, MRI and ultrasound) provide information on bone erosions and joint space narrowing [
20,
21]; however, no information is obtained on the cellular and molecular mechanisms of the disease that precede the development of destructive lesions, which sometimes trigger serious debility in less than 2 years in 10 % of cases. Most imaging probes used for RA diagnosis and monitoring are unspecific, although several classic or novel targeted imaging agents have been assessed, such as
18F-fluorodeoxyglucose (
18F-FDG), [
11C]Choline, (R)-[
11C]PK11195 and other translocator protein (TSPO)-targeted radiotracers, [
67Ga]Citrate, [
99mTc]- and [
111In] human immunoglobulin G (HIG), [
99mTc]- and [
111In]anti-E-selectin, [
99mTc]- and [
111In]Octreotide, [
99mTc]Anti-TNF-α and [
99mTc]Annexin V, etc. [
21‐
23]. Among them, only
18F-FDG, TSPO-targeted radiotracers (e.g.,
11CPBR28,
18F-GE-180), the α
vβ
3-specific imaging probe
68Ga-BNOTA-PRGD2 and the PET probe
18F-FHBG employed for the imaging of reporter genes, are currently undergoing clinical trials according to the US National Institutes of Health [
32]. Examples of new imaging tracers include
99mTc-labeled derivative of octreotide peptide employed to image somatostatin receptor expressed by T lymphocytes [
33], and several radiotracers used to monitor therapy response in human or experimental arthritis, such as
99mTc-NTP 15–5 targeted to proteoglycans in the RA joints [
34],
111In-RGD2 (α
vβ
3-targeted),
111In-anti-fibroblast activation protein antibody and
111In-antimurine macrophage antibody [
35]. Therefore, there is active research to meet the increasing demand of imaging methods and probes able to provide precocious and reliable information on the clinical outcome, pathophysiological process, the disease severity and location, and the disease response to novel molecular therapies.
The role of IL-7R and IL-7 in the pathogenesis of RA is well-documented. Both molecules are expressed in RA synovial tissue and blockade of the IL-7/IL-7R axis in CIA results in significant clinical improvement. In addition, IL-7Rα was recently identified as a diagnostic marker in early RA, and a marker of severity and poor response to therapy in early and established disease. Our attempts to develop in vivo IL-7Rα imaging tracers are based on these observations, and fit in a novel approach to the taxonomy of inflammatory joint disorders, in which diagnostic and therapeutic decisions are based on the identification of specific molecular pathways, rather than broad clinical diagnostic categories. Thus, we identified several IL-7Rα-specific heptapeptides, which are potential vectors for RA-dedicated imaging probes. Some of them are putative therapeutic agents that work by blocking IL-7 ligation to IL-7R. To the best of our knowledge, no other IL-7Rα-targeted small molecule has been discovered and exploited in the framework of RA diagnosis and treatment.
During the screening of the randomized cyclic heptapeptide phage display library, it was observed that most clones had high affinity against FN type III-like domain of IL-7Rα (A
131-I
231), suggesting its involvement in ligand binding. Indeed, specialized literature has shown that 5 of the 12 amino acids (S
51, F
99, L
100, L
101, I
102, K
104, D
122, H
154, K
157, Y
159, V
160, H
211) are involved in IL-7 ligation [
36]. As a consequence, excessive preselection against FN has led to a decline in phage pool affinity for IL-7Rα during the third and particularly the fourth round of panning. The peptide sequence of the 12 phage clones selected from the second and the third rounds of panning presented a high frequency of basic (His, Lys) and alcohol (Ser) amino acids, but also a Pro repetition. The first three amino acids are potentially involved in IL-7Rα binding based on the proposed model [
36], whereas Pro can induce 20° distortions in the axis of an alpha helix, which suggests that peptides can present a certain 3D conformation. The three best peptide clones (C-PHPQRPA-C, C-KIMKSMP-C and C-ASACPPH-C) characterized by the highest specific affinity against IL-7Rα have shown interesting homologies with molecules involved in signal transduction, cell adhesion, extracellular matrix, cytoskeletal organization, cell migration, embryogenesis and inflammation, proving that their selection was not accidental.
Before synthesis, peptide C-PHPQRPA-C presented the highest binding specificity, whereas peptide C-ASACPPH-C was the weakest candidate. However, after synthesis, peptide C-PHPQRPA-C (encoded as P722) had lost its affinity against IL-7Rα, displaying equivalent binding to FN. Peptide C-KIMKSMP-C (encoded as P725) had better affinity for FN than for IL-7Rα, but its μM K
d indicated stronger binding than that of P722. To our surprise, the highest specific binding to IL-7Rα was observed in the case of peptide C-ASACPPH-C (encoded as P726), which thus became the most promising candidate for diagnostic applications. Aiming to eventually improve its affinity constant, the peptide was synthesized in a linear version (encoded as P258), checking in this way the necessity of Cys-constraint. This chemical strategy has indeed enhanced its affinity in the order of nanomolar and improved its specificity against IL-7Rα. The linear peptides present a more flexible spatial conformation, with functional groups optimally exposed for a chemical interaction with the targeted biomarker [
37].
One can conclude that peptide P258 presents a chemical interaction with IL-7Rα, the conformational compatibility (imposed by Cys-constraint) not being a prerequisite. Peptide P258 was thus subsequently used to functionalize a MRI contrast agent such as USPIO (USPIO-P258). This reversal of the affinity parameters of the three candidate peptides after their synthesis is explained by the monovalent exposure to the target, after being presented in a pentavalent display on the phage entity. It is very well-known that pentavalent presentation is responsible for an avidity effect that may facilitate the binding of certain peptides to their target.
In our experimental conditions, IL-7 was characterized by a K
d (1.7 × 10
−7 M) close to that of P258, being inferior to the value reported previously (i.e., 2 × 10
−10 M against the high-affinity IL-7R and 10
−8 M against the low-affinity IL-7R) [
38]. The truncated recombinant IL-7Rα protein (241 amino acids instead of 459 amino acids), removed from its cellular environment, may be responsible for the lower affinity constant observed in our experimental conditions. In addition, two antibodies were used to detect IL-7, which means that additional rinsing steps may contribute to extensive protein removal and lower apparent affinity. However, an interesting observation was that P725 displayed competitive abilities against IL-7, which highlighted its role as a potential therapeutic agent by blocking IL-7 binding to its receptor.
With regard to imaging applications of our IL-7Rα-targeted peptide, USPIO-P258 had good ability to distinguish stimulated from non-stimulated Jurkat cells and its binding to IL-7Rα co-localized with anti-IL-7Rα antibody, confirming its specificity. The blood clearance of USPIO-P258 is much faster than that of USPIO-PEG, and its elimination is likely to mainly occur via a renal pathway. In addition, USPIO-P258 does not seem to accumulate in the main organs, which are simply transited via the blood stream. We have previously observed [
24,
25] that peptide grafting to USPIO is responsible for enhanced blood clearance, an enlarged VD
ss and increased urinary excretion, probably due to diminished PEG grafting that is partly replaced by peptides on the surface of nanoparticles. Moreover, these biodistribution and pharmacokinetic properties are amplified by peptides with a hydrophilic character.
USPIO-P258 produced a significant negative contrast in the paws of CIA mice, which was not equivalent to any of the control mice. The negative contrast corresponded to the tarsal, metatarsal and phalangeal joints, and persisted for about 2 hours post injection. The contrast agent accumulation in the hind limbs of CIA mice was confirmed by relaxometry and histochemical analysis at the end of the imaging session. The contrast observed at long image acquisition times suggests the specific binding to the targeted receptor, as most of the contrast agent has been cleared from the blood. On histological examination, USPIO-P258 co-localized with IL-7Rα expression in the paws of CIA mice, attesting for its specific accumulation at this level. A non-specific accumulation of USPIO-PEG was also observed in the diseased paws, probably as a result of its phagocytosis by the local macrophages, but its capture was less important as compared to USPIO-P258. Taken together, these results confirm the imaging ability of USPIO-P258 as an IL-7Rα imaging marker, the diagnostic faculty being furthermore confirmed by the high positive correlation between its accumulation in the diseased paws, the IL-7Rα expression and the disease severity.
One of the most important signaling pathways that are activated by IL-7 ligation to its receptor is Jak/STAT, which triggers the phosphorylation of cytoplasmic tyrosine kinases associated with IL-7Rα and the common γ-chain, respectively. Once activated, Jak1 can phosphorylate the Tyr
449 residue of IL-7Rα, which recruits the transcription factor STAT5 (a heterodimer comprising STAT5a and STAT5b) that is itself tyrosine-phosphorylated by Jak. Phosphorylated STAT5 can dimerize and translocate to the nucleus, where it regulates the transcription of several genes involved in T cell survival and proliferation [
2‐
4]. The inhibition of IL-7 engagement with its receptor by a specific competitor may have potential therapeutic effects in various inflammatory conditions such as RA. The competitive character of P725 was thus confirmed by immunofluorescence, using ADC-stimulated Jurkat cells as a model. P725 inhibited IL-7-induced STAT5 activation by 82 %, which was similar to that produced by anti-IL-7Rα antibody (86 %).
It has been shown that IL-7Rα is rapidly inactivated by lysosome and proteasome-dependent degradation subsequent to its endocytosis triggered by IL-7 activation. In fact, IL-7 signal transduction requires clathrin-dependent endocytosis of IL-7Rα followed by its degradation by proteolysis [
30]. In order to check whether STAT5 inactivation induced by P725 treatment is associated with a diminished IL-7Rα endocytosis, we have indirectly measured the lysosome content of Jurkat cells in the same experimental conditions as for STAT5. Our results confirm the increased lysosome content of Jurkat cells stimulated by IL-7, and suggest that P725 ligation to IL-7Rα blocks IL-7 binding and inhibits the endocytosis and lysosome-dependent degradation of IL-7Rα.
Acknowledgements
This work was supported by the project KeyMarker “Pôles de compétitivité BioWin” of the Walloon Region, Belgium (01/11/2006-31/03/2010). The Walloon Region (program First spin-off), FNRS (Fond National de la Recherche Scientifique), and ARC Programs of the French Community of Belgium are also gratefully acknowledged. The authors thank the Center for Microscopy and Molecular Imaging (CMMI, supported by the European Regional Development Fund and the Federation Wallonia Brussels).