Generation of molecular-targeting peptides for selective inhibition of the interaction between CTLA-4 and B7 in the dog: a new immune checkpoint inhibitor for cancer therapy

概要

Introduction
　Two signal pathways activate T cells: the antigen-specific signal and the co-stimulatory signal. Most effective co-stimulatory signal is generated by binding of CD28 on T cell with B7 on dendritic cell (DC), a typical antigen-presenting cell. Following the activation, cytotoxic T lymphocyte antigen-4 (CTLA-4) is expressed on T cell, which takes B7 from CD28, and generates suppressive signals. This suppression of T cells by the CTLA-4 signal plays an important role to prevent an excess of immune responses as an immune checkpoint. However, in tumor immunity, the suppression of T cell response by CTLA-4 is a critical mechanism underlying tumor growth. Therefore, inhibition of the binding between CTLA-4 and B7 elicits enhancement of the T cell activity against tumors by restoring the binding between CD28 and B7, and progresses tumor immune therapy.
　Treatment with monoclonal antibody specific for human CTLA-4 has emerged as an effective cancer therapy in humans. But therapeutic antibodies show some limitations such as high molecular weight with complex structure which make them difficult to synthesize, high immunogenicity, which requires a change of the framework from mouse immunoglobulin to that of recipient animal, low in vivo stability, and difficulty to improve affinity to targets. Those biophysical properties have prompted the extensive investigation of alternative binders. A targeting peptide, which has low molecular weight, low immunogenicity, high in vivo stability, and the high possibility to improve affinity is a promising alternative tool for the therapy.
　In this study, the author developed the molecular-targeting peptides against human (h)-CTLA- 4 from the library of peptide with helix-loop-helix (HLH) structure, and increased the affinity to CTLA-4. Then the biological activity of the h-CTLA-4-targeting peptides was confirmed in assays using human immune cells. With the effective results in human, the biological activity of the h-CTLA-4-targeting peptides finally evaluated in canine immune system.

Chapter 1: Generation of CTLA-4-targeting peptides
　To develop molecular-targeting peptides, as same in the case of antibody, the purified target protein is necessary. However, the cloning and the purification of recombinant canine CTLA-4 are difficult and take a long time. On the other hand, recombinant h-CTLA-4 is commercially available and clinically used for immune suppression. Therefore, the author at first examined the cross-reactivity between h-CTLA-4 and canine CTLA-4. The h-CTLA-4 specifically binds to canine DCs expressing canine B7 in flow cytometry (FCM), and inhibits the interaction between canine DCs and canine T cells expressing canine CD28 in allogeneic mixed lymphocyte reaction (MLR). These results suggested that the h-CTLA-4 has ability to bind canine B7 and biological activity to inhibit the binding between canine CD28 and canine CTLA-4. Therefore, the h-CTLA- 4 was used as the target to develop the canine CTLA-4-targeting peptides.
　For development of the CTLA-4-targeting peptides, yeast-displaying HLH peptides which bound to h-CTLA-4 were detected in FCM using fluorescence-labeled h-CTLA-4-Ig, and were isolated using fluorescence activated cell sorter (FACS). After sorting, three different h-CTLA-4- targeting peptides were obtained. Of the three, Y-2 showed the highest binding activity to h- CTLA-4, which was significantly decreased in the presence of B7-1. These results suggested that Y-2 and h-B7-1 share a same binding site on h-CTLA-4. Moreover, Y-2 showed no binding activities to other unrelated proteins such as h-IgG-Fc, h-TNF- , h-EGF, suggesting that Y-2 will not show undesirable effects in vivo. However, the dissociation constant (KD) of h-CTLA-4 binding to the Y-2-expressing yeasts was 1530 nM, and significantly higher than that to h-B7-1 (278 nM), indicating that the affinity of Y-2 to h-CTLA-4 was significantly low.
　Therefore, as the second step, the author improved the binding affinity of Y-2 by randomly changing a few amino acids of Y-2. This random mutation was carried out by error-prone PCR based on the DNA sequence of Y-2. Consequently, three variants were generated after the PCR. The variants, named as ERY2-1, ERY2-4, and ERY2-6, showed significantly higher binding activities to h-CTLA-4 than that of Y-2. The KD of h-CTLA-4 binding to the ERY2-1-, ERY2-4- or ERY2-6-expressing yeasts was 20.75 nM, 18.5 nM or 20.7 nM respectively, about 1/75 of that to the Y-2-expressing yeasts. Also, the binding activities of the variants significantly decreased in the presence of human B7-1. Moreover, the variants showed no binding activities to other unrelated proteins, h-IgG-Fc, h-TNF-α, h-EGF. Thus, these results indicated that the variant peptides generated by random mutation of Y-2 significantly improved the binding affinities to CTLA-4 and that, as same as Y-2, variants had the same binding site of B7-1 on h-CTLA-4.
　As the third step, the variant peptides were synthesized and investigated property of binding to CTLA-4. The KD of the synthesized ERY2-1, ERY2-4, or ERY2-6 binding to h-CTLA-4 was 277.7 nM, 196.8 nM, and 571.7 nM, respectively. These values were comparable to that of h-B7- 1 (278 nM). Since the variant peptides, especially ERY2-4 possessed a comparable affinity with B7-1 in binding to h-CTLA-4, and thus would be capable to inhibit h-CTLA-4 and h-B7-1 interaction. Due to the structural similarity between CD28 and CTLA-4, the CTLA-4-targeting peptides might bind to CD28. Because the binding between CD28 and B7 is essential for T cell activation, interruption of the binding thus induces immune suppression. However, it was confirmed by ELISA that ERY2-4 had significantly lower affinity to CD28 than h-B7-1. These results suggest that the ERY2-4 selectively binds to CTLA-4 for the inhibition of the CTLA-4/B7 binding without disrupting the CD28/B7 binding.
　Taken results together, it is indicated that the CTLA-4-targeting peptide which is capable of inhibiting CTLA-4/B7-1 interaction was successfully generated.

Chapter 2: Biological activity of CTLA-4-targeting peptides in human immune response
　To exert the therapeutic effect, CTLA-4-targeting peptides should functionaly block the interaction between CTLA-4 on DC and B7 on T cells. As the targeting peptide, ERY2-4 was generated against h-CTLA-4, the biological activity of ERY2-4 on human immune cells were at first evaluated. In the FCM assay, by addition of ERY2-4, the binding of h-CTLA-4 to DCs expressing B7 was significantly decreased. These results suggest that the CTLA-4-targeting peptides blocked the interaction between h-CTLA-4 and h-B7. Furthermore, to evaluate the functional blockade of human CTLA-4/B7-1 interaction, a mixed lymphocyte reaction (MLR) was conducted using human lymphocytes and human DCs. Allogeneic lymphocytes in the ERY2-4-treated cultures showed 2-fold enhanced proliferation compared to non-treated cultures. Taking these results together with results described in the Chapter 1, it is suggested that the CTLA-4-targeting peptides increased T cell responses by inhibiting the interaction between CTLA-4 expressed on T cells and B7 on DCs.

Chapter 3: Biological activity of CTLA-4-targeting peptides in canine immune response
　As the final step of this study, the author investigated the effects of the CTLA-4-targeting peptides on the canine immune system, for estimating a possibility for clinical use as new immune checkpoint inhibitor in tumor therapy. In the FCM assay, the binding of h-CTLA-4 to canine DCs was diminished by the h-CTLA-4-targeting peptide, ERY2-4 as same as the results using human DCs. Furthermore, in MLR using canine lymphocytes and canine DC, allogeneic lymphocytes in the ERY2-4-treated cultures showed 1.6-fold enhanced proliferation compared to non-treated cultures. It was demonstrated in the study of Chapter 1 that the h-CTLA-4 has ability to bind canine B7 and biological activity to inhibit the binding between canine B7 and canine CTLA-4, and that ERY2-4 has a significantly high affinity and specificity to h-CTLA-4 in Chapter 2. Taking these results together, it is strongly suggested that the h-CTLA-4-targeting peptide enhanced canine T cell responses due to the blockade of the interaction between CTLA-4 expressed on canine T cells and B7 on canine DCs. These results also indicates the possibility of the h-CTLA-4-targeting peptides for the application of canine tumor therapy, in which the targeting peptides prevent the suppression of T cell activity by the CTLA-4/B7 interaction, providing the affinity of the h-CTLA-4-targeting peptides to canine CTLA-4 will be further increased in future study.

Conclusion
　The CTLA-4-targeting peptides was generated and the biological activity in human and in canine immune responses were evaluated. Based on the results, the following conclusions are drawn;
　1. The CTLA-4-targeting peptide, Y-2 showed specific binding activity to h-CTLA-4Ig, but with a lower binding affinity compared to B7-1.
　2. Production of Y-2 variants by random mutation significantly increased the binding affinities and specificities to h-CTLA-4. Among Y-2 variants, ERY2-4 showed the strongest binding affinity to h-CTLA-4, which is comparable to that of B7-1. Further, it showed no binding activity to CD28.
　3. The h-CTLA-4-targeting peptide, ERY2-4 inhibited the interaction between h-CTLA-4 and B7 expressed on human DCs. Further, it enhanced human lymphocytes proliferation in the response to allogeneic human DCs.
　4. ERY2-4 showed inhibitory activity against the interaction between h-CTLA-4 and B7 expressed on canine DCs, and enhanced canine lymphocytes proliferation in the response to allogeneic canine DCs.

　Overall, the CTLA-4-targeting peptide was biologically active in both humans and dogs: it functionally blocked the immune checkpoint molecule, CTLA-4 to enhance the immune responses.