Concept of the construction of recombinant fusion proteins derived from IL-6 cytokines is not new. It is based on the strategy of linking the cytokines and their cognate soluble α receptor. The α receptors posses extracellular N-terminal domain and one transmembrane domain (with exception of CNTFR, which is linked to the membrane by a lipid anchor). Structurally, following predicted regions can be distinguished: (i) signal peptide, (ii) Ig-like region-D1 domain, (iii) cytokine binding homology domain (CHD) consisting of two fibronectin-type-III-like domains (FNIII) termed D2 and D3., (iv) receptor pre-membrane region, (v) transmembrane, and (vi) cytoplasmic regions. The designed fusion proteins contained different structural parts of its α receptor. Hyper-IL-6 consisted of D2 and D3 domain of IL-6R α chain (AA residues 113-323) connected to IL-6 (AA 29-212) via artificial polypeptide linker [14
]. Hyper-IL-6 is a fully active fusion protein, which mediates response at 10-1000-fold lower doses as compared to the combination of soluble IL-6 and sIL-6R molecules [14
]. In analogy, another superagonist was designed named IL-11/R-FP [13
]. IL-11/R-FP was created by covalently linking D2 and D3 domains of IL-11Rα (position L/109-G/318) with IL-11 (position P/29-L/199) using a 21 amino acid linker. It demonstrated 50 fold higher activity in vitro
than the combination of IL-11 and sIL-11Rα. However, such construct was also composed of interleukin and a truncated segment of the alpha receptor (as Hyper IL-6), so it lacked naturally existing parts of the used receptor. Moreover, the used artificial linker is not a naturally occurring sequence, thus it might demonstrate potential immunogenicity when used for treatment of human patient. Hyper-CNTF was another example of a superagonist composed of sCNTFR (AA 1-346) and CNTF (AA 1-186) [15
]. It differed from the previous IL-6 type fusion proteins (Hyper IL-6 and IL-11/R-FP), since it included the Ig domain of sCNTFR. Although the D3 domain of the interleukin alpha receptor is mainly involved in ligand binding [20
] and its D2-D3 part is sufficient to induce biological activity in vitro
], the role of the Ig-like domain and the membrane proximal region can not be disregarded. It was shown that the Ig-domain of the IL-6R is important for intracellular transport of IL-6R through the secretory pathway [21
]. It can also be involved in interdomain stabilization or induction of conformational change of the ligand. Moreover, for fusion of CNTFR and CNTF the C-terminal end of CNTFR and N-terminal end of CNTF was used, which were linked by one additional glycine residue [15
]. A synthetic polypeptide linker was not applied in order to minimize potential immunogenicity. The mentioned terminal ends are presumably flexible and sufficient to allow access of cytokine to its receptor alpha binding sites. Indeed, Hyper-CNTF was biologically active [15
]. In terms of construction, our Hyper-IL-11 (H11) resembles mostly Hyper-CNTF. It is composed of full length of soluble IL-11 receptor and IL-11, and both elements are connected by natural existing parts of receptor and interleukin. It differs from Hyper-CNTF, since the additional glycine residue was not incorporated in the fusion and also the flexible termini of sIL-11Rα and IL-11 generated a longer linker. For construction of Hyper-IL-6, a 29 AA linker was used, while for IL-11/R-FP 31 AA and Hyper-CNTF 31 AA were applied [13
]. With our hyper-molecule, a 67 AA linker was generated, since the pre-membrane region of IL-11Rα (51 amino acids) together with 16 N-terminal residues of IL-11 is not helical and presumably flexible. It should enable the proper positioning of the fusion molecule. In general, the construction of our hyper-cytokine has two major advantages: (i) its components are as close as possible to the natural forms of both proteins and (ii) omitting of the artificial linker should prevent possible immunogenicity and other side effects due to the non-natural recombinant agent.
In the case of Hyper-IL-6, the high potency of the designer cytokine is based on the fact that the affinity of IL-6 to the IL-6R is in the range of 1 nM whereas the affinity of the complex of IL-6 and IL-6R is 6R is around 10-15 pM, which is about 100 times higher. If the affinities between IL-11 and the two receptor subunits are in the same range, this would explain the high biologic activity of the H11 protein.
Although sIL-11Rα was not found in human body fluids so far, its existence has been postulated due to identification of a transcript potentially encoding sIL-11Rα [22
]. The recombinant sIL-11Rα in vitro
acts as IL-11 agonist [10
]. In addition, the sIL-11Rα not only potentiated effects of the IL-11 on cells that are normally responsive to IL-11, but also in the presence of IL-11 it mediated a signal transduction in cells expressing gp130 molecules only [10
]. However, the concentration of IL-11 required to mediate a biological response using sIL-11Rα had to be 10- 20-fold higher than using membrane receptor [11
]. Expression of IL-11Rα is limited to certain cell types, while gp130 is present on all cells of the body. Thus, the use of the sIL-11Rα significantly widens the range of IL-11 bioactivity. Furthermore, sIL-11Rα can act as an IL-11 antagonist when tested on cells expressing the membrane IL-11Rα and gp130 [11
]. The antagonism probably resulted from the competition between soluble and cellular IL-11Rα for IL-11 and/or was depended on the number of gp130 molecules [11
]. Our Hyper-IL-11 acts as agonist when tested in HepG2-assay, in B9 and in Ba/F3gp130 bio-assays. Since sIL-11Rα presumably is already occupied by IL-11, it does not compete with endogenous IL-11 or membrane bound IL-11Rα, which explains its agonistic activity. The previously described molecule IL-11/R-FP also acted as agonist when tested in HepG2 and Ba/F3-gp130 assays [13
]. When Ba/F3-gp130 cells were stimulated with IL-11/R-FP, a concentration 130-140 pM was sufficient for half-maximum proliferation. Our H11 induced the half-maximum proliferation at the dose of approximately 800-900 pM. However, IL-11/R-FP was non purified and its calculated concentration could be misleading. According to Pflanz et al. the concentration of IL-11/R-FP in the yeast supernatant was calculated from the band intensities of IL-11/R-FP and serial dilutions of IL-11 preparation of a known concentration measured by a Lumi-Imager [13
]. Since the analyzed material was derived from not purified yeast supernatant, presumably the concentration was calculated after immunoblotting (it was not defined). Our experience indicated that antibodies anti IL-11 or IL-11Rα not necessarily recognize fusion molecule with the same affinity. When several ELISA systems were used to measure H11 concentration, different results were obtained (data not shown). The antigenic determinant at a fusion molecule may be hidden or partially hidden when compared with the epitope of parental components (IL-11 or IL-11Rα), what can explain the observed differences. In order to compare correctly the activity of both fusion proteins (H11 and IL-11/R-FP) the experiment should be conducted simultaneously using purified proteins, which concentration was calculated with the same method.
After binding to their cognate α receptor, cytokines IL-6 and IL-11 engage a homodimer of gp130 in order to elicit signal transduction. Since, both designer cytokines, H11 and Hyper-IL-6 compromise of cytokine and its cognate α receptor, they need the same receptor subunits for signal transduction. That is why H11 and Hyper-IL-6 may be functionally equivalent. According to our preliminary results H11 retains in part an IL-11Rα-specific activity that is directed by H11 interaction with gp130. In terms of hematopietic differentiation H11 increased expression of early erythroid antigen. IL-11 is well known to be involved in the expansion of megakaryiocyte progenitors, and megakariocytes derive from common megakaryocyte/erythrocyte progenitor [23
]. However, Hyper IL-6 did not exert the same effect. Contrary, it stimulated myelopoiesis, what is in agreement with previous data [14
]. Although, IL-6 and IL-11 share the common receptor subunit (gp130), they bind to the different sites on it [24
], what can influence on the specific response mediated by gp130 signaling. The proliferation of TF1 cells induced by IL-11 but not by IL-6 could be inhibited by a specific antibody anti gp130 (B-P4). On the other hand, using SK-N-MC cells, the same antibody could not block the phosphorylation of gp130 induced by IL-11 and by IL-6 [24
]. Moreover, PHA-activated T-cells responded to IL-6 by mobilization of a heterodimer STAT1/STAT3 and homodimer STAT1, while IL-11/sIL-11Rα by a STAT3 homodimer and STAT3/STAT1 heterodimer [10
]. These results indicated that although structurally cytokines share the same signal transducing unit, its activation may elicit different effects, like different global phosphorylation level, activation of different STAT molecules what may influence the final biological activity. Such difference may be critical to activation of distinct regulation processes depending on the cell origin. However, whether this kind of differences are responsible for diverse action of H11 and Hyper-IL-6 needs to be evaluated.
Since IL-11 is a pleiotropic cytokine of hematopoietic and anti-inflammatory properties, H11 may be potentially beneficial in the treatment of various diseases. IL-11 is the only agent approved by USA Food and Drug Administration (FDA) to prevent severe thrombocytopenia and reduce the need for platelet transfusion following myelosuppressive chemotherapy for non-myeloid malignancies. Accordingly, studies of H11 in similar models are required. Even if in vivo
studies demonstrate unacceptable toxicity it does not exclude utility of H11 for ex vivo
expansion of megakaryocyte precursor cells. Our preliminary in vitro
studies have indicated great potential of H11 for megakaryocyte expansion (data not shown). Moreover, recent studies have shown that IL-11 regulated the autoimmune demyelination disease-multiple sclerosis (MS) [26
]. Due to its dual action: immunomodulatory and neuroprotective/regenerative, IL-11 may display a unique potential to treat MS. Loss of IL-11Rα expression was associated with severe symptoms of autoimmune encephalomyelitis (EAE) in the mouse model of MS. IL-11 treatment of mice with EAE gave partial, albeit statistically significant, therapeutic benefit. Another example of potential H11 utility is to treat infertility. Mice lacking IL-11Rα are infertile due to defective uterine response to implantation [27
]. Accordingly, H11 which can omit IL-11Rα and directly achieve gp130 may prove to be beneficial in pregnancy.