Anti-CD98 antibody and its applications

Abstract

The present disclosure provides antibodies and antigen-binding fragments thereof that bind to murine CD98 (mCD98) or hCD98 (hCD98). Specifically, the antibodies of the present disclosure bind to their targets in a pH-dependent manner. Also provided are nucleotides encoding the antibodies or fragments thereof, compositions or combinations that include the antibodies or fragments thereof, and methods of treating diseases, such as cancer, by using the antibodies or fragments thereof.

Description

[Technical field]

[0001] The present disclosure provides monoclonal antibodies, particularly monoclonal antibodies that specifically bind to human or mouse CD98. Also provided are nucleic acid molecules encoding anti-CD98 antibodies, as well as expression vectors and host cells comprising said nucleic acid molecules. Also disclosed are pharmaceutical compositions, conjugates and multispecific antibodies comprising anti-CD98 antibodies. Also provided are methods for treating various diseases, particularly tumors, such as tumors expressing CD98. [Background technology]

[0002] The type II transmembrane protein CD98 (also called CD98 heavy chain (CD98hc) or 4F2hc, encoded by the SLC3A2 gene) interacts with multiple light chains to form different heterodimeric amino acid transporters (HATs). While the light chains act in the amino acid transport process, CD98 is involved in the transport activity by stabilizing the light chain structure and contributing to the localization of the light chain to the cell membrane. CD98 has been shown to be involved in tumor formation, tumor progression and metastasis by promoting amino acid transport activity, promote cell survival, enhance integrin signaling, and increase cell spreading, migration, survival and proliferation. In the past decade, CD98 has become an attractive target for developing cancer therapies, as its expression is upregulated in various types of solid tumors and hematological malignancies and is associated with poor clinical outcomes. The CD98 protein is widely expressed in normal tissues, including the brain, spleen, kidney, small intestine, testis, hematopoietic system, etc.

[0003] According to Hayes et al. (Int J Cancer 137(2015), 710-720), the anti-human CD98 (hCD98) antibody IGN523 has shown potent antitumor activity in xenograft tumor models of leukemia, lymphoma, and lung cancer.

[0004] CN105385694B discloses a monoclonal antibody that binds to CD98 and teaches its application as a carrier for delivering anti-tumor or anti-inflammatory drugs. The disclosure of CN105385694B only provides in vitro data showing that anti-CD98 antibody can bind to lung cancer cells in cell lysate, but no in vivo data is shown. The in vivo performance of the antibody, such as tumor-specific binding activity or pharmacokinetic properties, is unknown.

[0005] WO2007114496A discloses several types of anti-CD98 antibodies, some of which have inhibitory effects on leucine uptake in bladder cancer cell lines and show antitumor effects in mice bearing the mouse CT26 colon cancer cell line expressing hCD98 / hLAT1-EGFP. One of the clones, C2IgG1, shows antitumor effects in mice implanted with the human Burkitt's lymphoma cell line Ramos.

[0006] Additionally, other patent application disclosures including WO2017214456A1, WO2017214458A2, WO2017214462A2 by AbbVie, and WO2015146132A1 by Daiichi Sankyo Co Ltd. provide conjugates comprising an anti-CD98 antibody and a drug such as a Bcl-xL inhibitor.

[0007] However, none of the above publications discuss the problems that may result from the binding of anti-CD98 antibodies to CD98 in normal tissues. This problem does not exist in animal models carrying cancer cell lines genetically engineered to express hCD98 or in xenograft models, because the normal tissues in the animal models express mouse CD98, which is not the target of the antibody. On target / off tumor binding of anti-CD98 antibodies may destroy the function of CD98 present in normal tissues, while weakening the antitumor effect of the antibody on the target tumor cells.

[0008] In fact, the widespread expression of CD98 in normal tissues poses several major challenges to clinical application. Targeted disruption of the CD98 gene leads to embryonic death (Tsumura et al., (2003), Biochemical and biophysical research communications 308, pp. 847-851). Conditional knockout of CD98 impairs the proliferation and regeneration capacity of hematopoietic stem and progenitor cells (Bajaj et al., (2016), Cancer Cell 30, pp. 792-805). Therefore, the problem of targeting side effects and "antigen sink" has become a major focus for developing antibody therapy targeting CD98.

[0009] Currently, there is an unmet need for anti-CD98 antibodies as antitumor therapies, and there is a need to develop anti-CD98 antibodies that have minimal effect on the normal physiological functions of CD98 and reduce the side effects of the "antigen sink" phenomenon. Summary of the Invention

[0010] In a syngeneic tumor model established by xenograft and CD98-humanized mice, the inventors successfully identified an anti-CD98 antibody (S1-F4) that elicits broad-spectrum Fc-dependent antitumor activity independent of interfering with the physiological function of CD98. The antitumor activity of S1-F4 was observed through the activation of FcγRs, macrophages, dendritic cells and CD8 + It requires innate and adaptive immune components including T cells, and to overcome the problem of "antigen sink" due to CD98 antibody binding in normal tissues, we solved the structure of the S1-F4 / CD98 complex and generated a series of pH-dependent binding variants, which increased tumor-specific binding and significantly improved pharmacokinetic characteristics, thereby promoting overall antitumor activity.

[0011] In a first aspect, the present disclosure provides an antibody, or antigen-binding fragment thereof, that binds to human or mouse CD98, comprising: A heavy chain CDR1 (HCDR1) comprising or consisting of an amino acid sequence selected from SEQ ID NO: 10, 20, 30, 40, 50, 60, and 70; A heavy chain CDR2 (HCDR2) comprising or consisting of an amino acid sequence selected from SEQ ID NOs: 11, 21, 31, 41, 51, 61, and 71; A heavy chain CDR3 (HCDR3) comprising or consisting of an amino acid sequence selected from SEQ ID NO: 12, 22, 32, 42, 52, 62, and 72; A light chain CDR1 (LCDR1) comprising or consisting of an amino acid sequence selected from SEQ ID NO: 15, 25, 35, 45, 55, 65, and 75; A light chain CDR2 (LCDR2) comprising or consisting of an amino acid sequence selected from SEQ ID NOs: 16, 26, 36, 46, 56, 66, and 76; and The present invention relates to an isolated antibody or antigen-binding fragment thereof, comprising a light chain CDR3 (LCDR3) comprising or consisting of an amino acid sequence selected from SEQ ID NOs: 17, 27, 37, 47, 57, 67, and 77.

[0012] In one embodiment, the antibody or fragment thereof comprises any of the following: (a) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 10, 11 and 12, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 15, 16 and 17, (b) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 20, 21 and 22, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 25, 26 and 27, respectively. (c) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 30, 31 and 32, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 35, 36 and 37, respectively. (d) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 40, 41 and 42, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 45, 46 and 47, respectively. (e) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 50, 51 and 52, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 55, 56 and 57, respectively. (f) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 60, 61 and 62, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 65, 66 and 67, respectively. (g) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 70, 71 and 72, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 75, 76 and 77, respectively. An antibody or a fragment thereof comprising any one of the above (a) to (f) specifically binds to hCD98, and an antibody or a fragment thereof comprising (g) specifically binds to mouse CD98.

[0013] In one embodiment, the antibody or fragment thereof comprises: A heavy chain variable region (VH) comprising an amino acid sequence having at least 80%, at least 85%, at least 90%, at least 95%, or at least 98% identity to an amino acid sequence selected from SEQ ID NOs: 13, 23, 33, 43, 53, 63, and 73; and / or and a light chain variable region (VL) comprising an amino acid sequence having at least 80%, at least 85%, at least 90%, at least 95% or at least 98% homology to an amino acid sequence selected from SEQ ID NOs: 18, 28, 38, 48, 58, 68 and 78.

[0014] In another embodiment, the antibody or fragment thereof comprises: VH comprising an amino acid sequence selected from SEQ ID NO: 13, 23, 33, 43, 53, 63 and 73, and / or VL comprising an amino acid sequence selected from SEQ ID NO:18, 28, 38, 48, 58, 68 and 78.

[0015] In a specific embodiment, the antibody or fragment thereof comprises any of the following: (a) to (g): (a) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 13, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 18; (b) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 23, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 28; (c) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 33, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 38; (d) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 43, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 48; (e) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 53, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 58; (f) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 63, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 68; (g) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 73, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 78; An antibody or a fragment thereof comprising any one of the above (a) to (f) specifically binds to hCD98, and an antibody or a fragment thereof comprising (g) specifically binds to mouse CD98.

[0016] In one embodiment, the anti-CD98 antibody or fragment thereof has a pH-dependent binding to hCD98, and the binding activity at acidic pH is higher than that at neutral pH. For example, the binding activity at acidic pH is at least 2-fold, 3-fold, 5-fold, 10-fold, 15-fold, 20-fold, 30-fold, 40-fold, 50-fold, 60-fold, 70-fold, 80-fold, 90-fold, or 100-fold higher than that at neutral pH. The binding activity can be measured by any method known in the art, such as ELISA, FACS, or surface plasmon resonance (e.g., Biacore®). For example, when the binding activity to hCD98 is measured by ELISA, the EC50 of the antibody or fragment thereof at acidic pH is at least 2-fold, 3-fold, 5-fold, 10-fold, 15-fold, 20-fold, 30-fold, 40-fold, 50-fold, 60-fold, 70-fold, 80-fold, 90-fold, or 100-fold lower than that at neutral pH. An acidic pH is 5.8, 5.9, 6.0, 6.1, 6.2, 6.3, 6.4, 6.5, 6.6, 6.7 or 6.8, in particular 6.5. A neutral pH is 7.0, 7.1, 7.2, 7.3, 7.4, 7.5 or 7.6, in particular 7.4. In one embodiment, the antibody or fragment thereof has a desired binding activity at an acidic pH, but no or significantly reduced binding activity at a neutral pH.

[0017] In one embodiment, the anti-CD98 antibody or fragment thereof exhibits preferential binding to hCD98 present on or within tumor cells compared to hCD98 not present on or within tumor cells. Specifically, such binding preference is achieved by the antibody or fragment thereof binding to hCD98 in a pH-dependent manner. For example, the binding activity of the anti-CD98 antibody or fragment thereof to hCD98 is significantly greater at the pH of the tumor microenvironment of a subject (e.g., a human) than at the normal physiological pH of the subject (e.g., a human). For example, when the binding activity to hCD98 is measured by ELISA, the EC50 of the antibody or fragment thereof at the pH of the tumor microenvironment is at least 2-fold, 3-fold, 5-fold, 10-fold, 15-fold, 20-fold, 30-fold, 40-fold, 50-fold, 60-fold, 70-fold, 80-fold, 90-fold, or 100-fold lower than the EC50 at normal physiological pH. In one embodiment, the tumor microenvironment pH may be a pH of about 6.0 to 7.0, e.g., pH 7.0, 6.9, 6.8, 6.7, 6.6, 6.5, 6.4, 6.3, 6.2, 6.1, 6.0 or lower, e.g., pH 6.5. In another embodiment, normal physiological pH may be normal blood or serum pH, e.g., pH 7.4. In one embodiment, the antibody or fragment thereof has the requisite binding activity in the tumor microenvironment and no or significantly reduced binding activity at normal physiological pH.

[0018] In one embodiment, the present application relates to an antibody or fragment thereof that binds to the same epitope on hCD98 as any of the above antibodies or antigen-binding fragments thereof. Specifically, the antibody or fragment thereof binds to an epitope at amino acid residues 135, 376-384, and 391-399 of the amino acid sequence of hCD98 (SEQ ID NO:9). More specifically, the antibody or fragment thereof directly binds to an epitope in which amino acid residues H135, E384, E392, and D397 are directly involved in antibody binding.

[0019] In one embodiment, the anti-CD98 antibody comprises a heavy chain constant region of the IgG1, IgG2, IgG3, IgG4 subclass or a variant thereof, and a light chain constant region of the kappa or lambda type or a variant thereof. In one preferred embodiment, the anti-CD98 antibody comprises an IgG1 heavy chain constant region.

[0020] In one embodiment, the antigen-binding fragment is a single chain variable region, Fab, Fab', F(ab')2, scFv, dsFv, or ds-scFv.

[0021] In one embodiment, the antibody may be a multispecific antibody, such as a bispecific antibody, a trispecific antibody, a diabody, or a minibody.

[0022] In one preferred embodiment, the antibody or fragment thereof comprises HCDR1, HCDR2, HCDR3 having the amino acid sequences of SEQ ID NOs: 10, 11 and 12, and LCDR1, LCDR2, LCDR3 having the amino acid sequences of SEQ ID NOs: 15, 16 and 17, respectively, and one or more amino acids in any of the six CDRs are mutated to aspartic acid (D) or glutamic acid (E) or mutated to histidine (H), and the antibody or fragment thereof exhibits pH-dependent binding to hCD98. Preferably, the amino acid mutated to D or E is in a position that closely interacts with a histidine within or near the binding epitope on hCD98, or the amino acid mutated to H is in a position that closely interacts with an aspartic acid or glutamic acid within or near the binding epitope on hCD98.

[0023] In one embodiment, the antibody or fragment thereof is used in antibody-based therapy.

[0024] In a second aspect, the disclosure provides an isolated polynucleotide encoding an antibody or fragment thereof of the first aspect. In one embodiment, the isolated polynucleotide comprises a nucleotide sequence having at least 80% identity, at least 85% identity, at least 90% identity, at least 95% identity, at least 98% identity, or 100% identity to a nucleotide sequence selected from SEQ ID NOs:14, 19, 24, 29, 34, 39, 44, 49, 54, 59, 64, 69, 74, or 79.

[0025] In a third aspect, the present disclosure relates to an expression vector comprising the isolated polynucleotide of the second aspect.

[0026] In a fourth aspect, the present disclosure relates to a host cell comprising the isolated polynucleotide of the second aspect or the expression vector of the third aspect.

[0027] In a fifth aspect, the disclosure relates to a composition, e.g. a pharmaceutical composition, comprising an anti-CD98 antibody or fragment thereof of the first aspect and a pharma- ceutically acceptable carrier. Preferably, the pharmaceutical composition comprises a therapeutically effective amount of the anti-CD98 antibody or fragment thereof.

[0028] In a sixth aspect, the disclosure provides a method for reducing tumors, inhibiting tumor cell proliferation, treating cancer, or preventing recurrence of cancer in a subject in need thereof comprising administering to the subject a therapeutically effective amount of an anti-CD98 antibody or fragment thereof of the first aspect, or a composition of the fifth aspect.

[0029] In a seventh aspect, the disclosure provides a method for treating an autoimmune disease in a subject in need thereof comprising administering to the subject a therapeutically effective amount of an anti-CD98 antibody or fragment thereof of the first aspect, or a composition of the fifth aspect.

[0030] In an eighth aspect, the disclosure provides the use of an anti-CD98 antibody or fragment thereof according to the first aspect in the manufacture of a medicament. In one embodiment, the medicament is used for reducing tumors, inhibiting tumor cell proliferation, treating cancer, preventing cancer recurrence. In one embodiment, the medicament is used for treating an autoimmune disease.

[0031] In one embodiment of the sixth or eighth aspect, the tumor or cancer expresses CD98, in particular hCD98. In one embodiment of the sixth or eighth aspect, the pH of the microenvironment in the tumor or cancer is lower than the normal physiological pH of the subject. For example, the microenvironment in the tumor or cancer has an acidic pH, such as pH 7.0, 6.9, 6.8, 6.7, 6.6, 6.5, 6.4, 6.3, 6.2, 6.1, 6.0 or lower, such as about pH 6.5, and the normal physiological pH of the subject is the pH of normal blood or serum, such as pH 7.4. More specifically, said cancer is selected from lymphoma, acute myeloid leukemia, acute promyelocytic leukemia, hepatocellular carcinoma, pancreatic cancer, pancreatic epithelioid carcinoma, breast cancer, colorectal adenocarcinoma, skin epidermoid carcinoma, melanoma, fibrosarcoma, non-small cell lung cancer, gastric cancer, acute myeloid leukemia, glioma, tongue cancer, hypopharyngeal squamous cell carcinoma, cholangiocarcinoma, osteomalacia, osteosarcoma, renal cancer, and neuroblastoma.

[0032] In one embodiment of the seventh or eighth aspect, the autoimmune disease is caused by an abnormal proliferation of immune cells, such as T cells or B cells, hi one embodiment, the autoimmune disease is selected from multiple sclerosis, type I diabetes, or rheumatoid arthritis.

[0033] Recent studies have identified CD98hc as a potent receptor-mediated cytoplasmic transport (RMT) target that can facilitate and enhance brain delivery of therapeutic antibodies (Y. Joy Yu Zuchero et al., Neuron 89, pp. 70-82, January 6, 2016). Antibodies with poor blood-brain barrier (BBB) ​​penetration can be paired with anti-CD98 antibodies to form bispecific antibodies to improve their transport across the blood-brain barrier and brain accumulation.

[0034] Thus, in a ninth aspect, the present disclosure relates to a fusion protein, such as a bispecific antibody, comprising a first antigen-binding fragment that is an antigen-binding fragment that specifically binds to hCD98 of the first aspect, and a second antigen-binding fragment that specifically binds to a second antigen distinct from that bound by the antigen-binding fragment of the first aspect. In one embodiment, the bispecific antibody is a therapeutic antibody or a diagnostic antibody. In one embodiment, the bispecific antibody exhibits higher brain accumulation following systemic administration than a monospecific antibody that specifically binds to the second antigen.

[0035] In a tenth aspect, the present application relates to a combination, conjugate, or composition for treating a tumor comprising (a) an anti-hCD98 antibody of the first aspect, and (b) a second anti-tumor reagent. In one embodiment, the second anti-tumor reagent may be a reagent that modulates an immune checkpoint protein, including, but not limited to, PD-1, PD-L1, CTLA-4, 4-1BB, 4-1BBL, CD28, CD40, CD40L, CD47, OX40, OX40L, TIM-3, TIGIT, NKG2A, B7-H3, B7-H4, VISTA, LAG3, 2B4. In one embodiment, the reagent that modulates an immune checkpoint protein is an antibody that specifically binds to an immune checkpoint protein.

[0036] The present inventors have demonstrated that the anti-hCD98 antibody of the present application exerts its antitumor activity by targeting macrophages and CD8 + Thus, in a preferred embodiment of the tenth aspect, the second anti-tumor agent may be an agent that enhances the phagocytic function of macrophages, or may be an agent that enhances the phagocytic function of CD8 T cells. + The second anti-tumor agent may be a reagent that enhances the action of T cells. In one embodiment, the second anti-tumor agent is a reagent that enhances the phagocytic function of macrophages by targeting a phagocytosis inhibitor (e.g., CD47). In one specific embodiment, the reagent that enhances the phagocytic function of macrophages is an anti-CD47 antibody. In another embodiment, the second anti-tumor agent is a reagent that enhances the phagocytic function of macrophages by targeting a phagocytosis inhibitor (e.g., CD47). +Enhance the action of T cells, e.g., by blocking or reversing negative regulation of cell-mediated immune responses, e.g., by targeting inhibitory receptors, such as CD8 + and enhances the action of T cells. In one specific embodiment, the second anti-tumor reagent is an antibody that specifically binds to PD-1, CTLA-4, PD-L1, or 4-1BB. In another embodiment, the anti-CD98 antibody of the present application cannot be used in combination with a reagent that depletes macrophages (e.g., an anti-CSF1R antibody). [Brief description of the drawings]

[0037] [Figure 1] Figure 1 shows the antitumor activity of HN2-G9 and IGN523 as a control in the Raji tumor model. NOD SCID mice bearing Raji tumors are randomly divided into three groups with similar mean tumor volumes. Antitumor activity of 10 mg / kg of antibody is measured. Tumor volumes are shown as mean ± SEM. (ns: not significant, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, 2-way ANOVA). Similar statistical methods and illustrations apply to all lined scattergrams for tumor volumes described below. [Diagram 2] Figure 2 shows the binding specificity and avidity of IGN523, HN2-G9 and S1-F4 as measured by FACS. Wild-type (WT) CHO or CHO-hCD98 cells were incubated with the indicated concentrations of each antibody and then labeled with a FITC-conjugated anti-human IgG secondary antibody. Binding was determined from the FITC fluorescence intensity. [Diagram 3] FIG. 3 shows the analysis of the binding kinetics between anti-hCD98 antibodies and hCD98 ECD by SPR. [Figure 4] Figure 4 shows the antitumor activity of S1-F4, and rituximab and IGN523 as controls in the Raji tumor model. NOD SCID mice bearing Raji tumors are randomly divided into three groups with similar mean tumor volumes. Antitumor activity of 10 mg / kg of antibody is measured. [Diagram 5]FIG. 5 is a table showing human-derived cell lines for constructing xenograft tumor models. [Figure 6] Figures 6 and 7 show the antitumor activity of S1-F4 in HepG2 (Figure 6) and Ramos (Figure 7) tumor models. CB-17 SCID mice bearing HepG2 tumors or NOD SCID mice bearing Ramos tumors were randomized into 4-5 groups with similar mean tumor volumes and treated with S1-F4 at the indicated doses. [Figure 7] Figures 6 and 7 show the antitumor activity of S1-F4 in HepG2 (Figure 6) and Ramos (Figure 7) tumor models. CB-17 SCID mice bearing HepG2 tumors or NOD SCID mice bearing Ramos tumors were randomized into 4-5 groups with similar mean tumor volumes and treated with S1-F4 at the indicated doses. [Figure 8] Figure 8 shows the antitumor activity of S1-F4 in HL60, HCT-8, BxPC-3 and PANC-1 tumor models. CB-17 SCID mice bearing HL60, BxPC-3 or HCT-8 tumors or NOD SCID mice bearing PANC-1 tumors were randomly divided into two groups with similar tumor volumes: a control group or a S1-F4 (15 mg / kg) treatment group. The time points of antibody treatment are indicated by arrows. [Figure 9] Figure 9 shows the antitumor activity of S1-F4 in the MDA-MB-231-LN tumor model. CB-17 SCID mice bearing MDA-MB-231-LN tumors were randomly divided into two groups with similar mean tumor volumes: vehicle and S1-F4 (15 mg / kg). The schematic diagram shows the timeline of tumor inoculation and the S1-F4 treatment scheme. Bioluminescence imaging analysis was performed on days 8, 18, 29, and 41 after tumor inoculation. [Figure 10]Figure 10 shows the antitumor activity of S1-F4 in HCT 116, A549, A-431, Hep3B and HT-29 tumor models. CB-17 SCID mice bearing HCT 116, A549 or A-431 tumors or NOD SCID mice bearing Hep3B or HT-29 tumors were randomly divided into two groups with similar mean tumor volumes, vehicle or S1-F4 (15 mg / kg). Antibody treatment time points are indicated by arrows. Tumor volumes are shown as mean ± SEM. (ns: not significant, two-way ANOVA). [Figure 11A-E] 11A-F show the construction strategy of CD98 humanized mice and analysis of CD98 expression. (A) Schematic showing the strategy of inserting hCD98 ECD cassette into mouse Slc3a2 locus. The positions of sgRNA-CD98 and indicated PCR primers in the targeted and wild-type alleles are shown. (B-D) Gel electrophoresis of PCR reaction products from genomic DNA templates of F0 generation mice. Primer pairs used (see Table 2) are Y1-F / Y1-R (B), WT1-F / WT1-R (C), and IF-1 / Y1-R (D). (E) Kinetic analysis of binding of anti-mouse CD98 antibody BC8 to mouse CD98 (mCD98) ECD. (F) Binding specificity and affinity of BC8 are measured by FACS. WT HEK293T and HEK293T-mCD98 cells were incubated with the indicated concentrations of BC8 and then labeled with a FITC-conjugated anti-human IgG secondary antibody, and binding was determined by FITC fluorescence intensity. [Figure 11F]11A-F show the construction strategy of CD98 humanized mice and analysis of CD98 expression. (A) Schematic showing the strategy of inserting hCD98 ECD cassette into mouse Slc3a2 locus. The positions of sgRNA-CD98 and indicated PCR primers in the targeted and wild-type alleles are shown. (B-D) Gel electrophoresis of PCR reaction products from genomic DNA templates of F0 generation mice. Primer pairs used (see Table 2) are Y1-F / Y1-R (B), WT1-F / WT1-R (C), and IF-1 / Y1-R (D). (E) Kinetic analysis of binding of anti-mouse CD98 antibody BC8 to mouse CD98 (mCD98) ECD. (F) Binding specificity and affinity of BC8 are measured by FACS. WT HEK293T and HEK293T-mCD98 cells were incubated with the indicated concentrations of BC8 and then labeled with a FITC-conjugated anti-human IgG secondary antibody, and binding was determined by FITC fluorescence intensity. [Figure 12] Figure 12 shows the analysis of CD98 expression in mouse kidneys by immunofluorescence staining. Kidney samples from wild-type C57BL / 6 mice and CD98-humanized mice are incubated with BC8 (mCD98-specific antibody, 15 μg / ml) or IGN523 (hCD98-specific antibody, 15 μg / ml) and then labeled with FITC-conjugated anti-human IgG secondary antibody. Magnification: ×1.3. [Figure 13] Figure 13 shows the expression of CD98 on mouse leukocytes by FACS analysis. Leukocytes from C57BL / 6 and CD98-humanized mice were incubated with BC8 or S1-F4 (10 μg / ml) and then labeled with FITC-conjugated anti-human IgG secondary antibody. [Figure 14] Figure 14 shows the expression of hCD98 and mCD98 in mouse cancer cells by FACS analysis. Cells were incubated with S1-F4 or BC8 (15 μg / ml) and then labeled with a FITC-conjugated anti-human IgG secondary antibody. [Figure 15] 15A-B show the antitumor activity of S1-F4 in CD98 humanized mice (50 mg / kg) (A) and in C57BL / 6 mice bearing syngeneic tumors (15 mg / kg) (B). [Figure 16] Figure 16A-B shows the ADCC effector function induced by S1-F4 and its Fc variants. ADCC activity is measured by LDH release. Data are shown as mean ± SEM. (A) NK92-MIhCD16 cells are used as effector cells and Raji or HepG2 cells are used as target cells. E:T ratio is 7:1. Antibody is measured at the indicated concentrations. (ns: not significant, ****p<0.0001, two-way ANOVA). (B) Raji cells are used as target cells. Left: mBMDM are used as effector cells. Control IgG or S1-F4 is measured at the indicated concentrations. E:T ratio is 1:1. (****p<0.0001, two-way ANOVA). Right: hPBMC are used as effector cells. Vehicle or S1-F4 (10 μg / ml) is used. E:T ratio is 10:1. (**p<0.01, two-tailed unpaired Student's t-test). [Figure 17A] 17A-B show the ADCP effector function induced by S1-F4 and its Fc variants. (A) Raji cells (target cells) are labeled with CFSE fluorescent dye and then mixed with mouse bone marrow derived macrophages (mBMDM) labeled with anti-F4 / 80-Alexa Fluor 633 antibody. The E:T ratio is 1:2. ADCP activity is monitored by fluorescence microscopy and the phagocytic index is measured as the number of CFSE-positive target cells per 100 effector cells. The scale is 100 μm. (B) hPBMCs are used as effector cells. Raji cells (target cells) are labeled with CFSE fluorescent dye and then mixed with hPBMCs labeled with Deep-Red. Control IgG or S1-F4 (10 μg / ml). The E:T ratio is 1.5:1. ADCP activity is monitored by fluorescence microscopy. Representative micrographs are shown. The scale is 100 μm. [Figure 17B]17A-B show the ADCP effector function induced by S1-F4 and its Fc variants. (A) Raji cells (target cells) are labeled with CFSE fluorescent dye and then mixed with mouse bone marrow derived macrophages (mBMDM) labeled with anti-F4 / 80-Alexa Fluor 633 antibody. The E:T ratio is 1:2. ADCP activity is monitored by fluorescence microscopy and the phagocytic index is measured as the number of CFSE-positive target cells per 100 effector cells. The scale is 100 μm. (B) hPBMCs are used as effector cells. Raji cells (target cells) are labeled with CFSE fluorescent dye and then mixed with hPBMCs labeled with Deep-Red. Control IgG or S1-F4 (10 μg / ml). The E:T ratio is 1.5:1. ADCP activity is monitored by fluorescence microscopy. Representative micrographs are shown. The scale is 100 μm. [Figure 18] Figure 18A-B shows the CDC effector function (A) and binding to human C1q (B) of S1-F4 and its variants. (A) Raji cells are incubated with the indicated antibodies in the presence of 10% human complement serum. CDC activity is measured by LDH release assay. Antibodies are measured at 10 μg / ml. (B) Binding of S1-F4, Fc variants of S1-F4, and rituximab to human C1q is measured by ELISA. [Figure 19] FIG. 19 shows the effect of S1-F4 on HPG transport in HepG2 and Raji cells. HepG2 or Raji cells were incubated with control IgG (15 μg / ml), S1-F4 (15 μg / ml) or BCH (10 mM), respectively, and then cultured in a medium containing HPG (70 μM). Then, HPG in the cell lysate was biotinylated, and biotinylated HPG was detected using streptavidin-HRP by ELISA. HPG uptake activity is shown as the percentage of the HPG uptake of the measurement sample relative to the HPG uptake of the control IgG-treated group. Data are shown as mean ± SEM. (ns: no significant difference, **p<0.01, ***p<0.001, two-tailed unpaired Student's t-test) [Figure 20]FIG. 20 shows the effect of S1-F4 on the proliferation of Raji, Ramos, HepG2 and HCT-8 cells. Cell proliferation is analyzed using WST-8 Cell Counting Kit-8. Cells are incubated with control IgG or S1-F4 for 72 hours. The percentage of cell proliferation is shown as the percentage of the measured sample group relative to the detection value of the untreated group. Raji or Ramos cells are incubated with control IgG or S1-F4 at the indicated concentrations. (****p<0.0001, two-way ANOVA). HepG2 or HCT-8 cells are incubated with control IgG or S1-F4 (15 μg / ml). Data are shown as mean ± SEM. (ns: not significant, two-tailed unpaired Student's t-test). [Figure 21] FIG. 21 shows the antitumor activity of S1-F4 and its Fc variants in CD98 humanized mice bearing EL4-hCD98 or MC38-hCD98 tumors (S1-F4 or S1-F4DANA, 50 mg / kg), NOD SCID mice bearing Raji tumors, or CB-17 SCID mice bearing HepG2 tumors (control IgG, S1-F4, S1-F4KA, or S1-F4DANA, 15 mg / kg). [Figure 22] Figure 22 shows that S1-F4 antibody treatment induced antitumor immunity upon subsequent tumor challenge. C57BL / 6 mice bearing EL4-hCD98 tumors that achieved complete response (CR) after S1-F4 treatment were rechallenged with EL4-hCD98, EL4, B16F10-hCD98 or B16F10 tumor cells, respectively, 77 days after the first tumor cell inoculation. Naive mice of similar age served as control mice. After (re)inoculation with tumor cells, tumor size was assessed. Tumor volumes are shown as mean ± SEM. The percentage of tumor-free mice is shown. [Figure 23]Figures 23A-B show the pharmacokinetics and biodistribution of S1-F4 in monkeys and mice. S1-F4 concentration in serum of cynomolgus monkeys and mice is plotted against time. S1-F4 treatment times are indicated by arrows. Serum concentrations of S1-F4 were measured by ELISA and are shown as mean ± SEM. A single dose (iv) of S1-F4 to monkeys is 20 mg / kg (A). A single dose (ip) of S1-F4 to C57BL / 6 or CD98 humanized mice is 15 mg / kg (B). [Figure 24A] Figure 24A-B shows typical dissection images (A) of CD98 humanized mice sacrificed 53 hours after injection of S1-F4-Cy7 (15 mg / kg) or vehicle. The concentration of S1-F4 in the kidney (B) is analyzed by immunofluorescence staining. C57BL / 6 mice and CD98 humanized mice are treated with vehicle or S1-F4 (50 mg / kg), and mouse kidney samples are taken 2 days later and incubated with FITC-conjugated anti-human IgG secondary antibody. The concentration of S1-F4 is then measured by the fluorescence intensity of FITC. The magnification is X1.3. [Figure 24B] Figure 24A-B shows typical dissection images (A) of CD98 humanized mice sacrificed 53 hours after injection of S1-F4-Cy7 (15 mg / kg) or vehicle. The concentration of S1-F4 in the kidney (B) is analyzed by immunofluorescence staining. C57BL / 6 mice and CD98 humanized mice are treated with vehicle or S1-F4 (50 mg / kg), and mouse kidney samples are taken 2 days later and incubated with FITC-conjugated anti-human IgG secondary antibody. The concentration of S1-F4 is then measured by the fluorescence intensity of FITC. The magnification is X1.3. [Diagram 25] FIG. 25 shows a striped schematic of S1-F4 scFv and hCD98 ECD in orthogonal view, and a detailed view of the S1-F4-CD98 interface. [Figure 26] FIG. 26 is a close-up of the interface between S1-F4 scFv and hCD98 ECD, showing the positions of E384, D391, E392, D397 of CD98, and HCDR2, HCDR3, LCDR1 of S1-F4. [Figure 27]FIG. 27 shows the binding of S1-F4, S1-F4 Y97E and H15L54 to hCD98 ECD at pH 6.5 or pH 7.4 as analyzed by ELISA. [Figure 28] FIG. 28 shows the binding of S1-F4, H15L1, H15L35 and H15L54 to A-431 at pH 6.5 and pH 7.4 as analyzed by FACS. [Figure 29] Figures 29 and 30 show that H15L54 preferentially binds to CD98 in tumors. CD98 humanized mice bearing EL4-hCD98 tumors are treated with S1-F4-Cy7 or H15L54-Cy7 (15 mg / kg). Imaging analysis is performed on organs at 46, 90 and 160 hours after injection. (A) Representative images. (B) Quantified tissue biodistribution. Mean luminescence intensity is shown as mean ± SEM. [Diagram 30] Figures 29 and 30 show that H15L54 preferentially binds to CD98 in tumors. CD98 humanized mice bearing EL4-hCD98 tumors are treated with S1-F4-Cy7 or H15L54-Cy7 (15 mg / kg). Imaging analysis is performed on organs at 46, 90 and 160 hours after injection. (A) Representative images. (B) Quantified tissue biodistribution. Mean luminescence intensity is shown as mean ± SEM. [Diagram 31] Figure 31 shows the enrichment of the indicated antibodies in kidney cells of CD98-humanized mice. Mice are treated with 15 mg / kg of the indicated antibodies. After 2 days, mouse kidney samples are taken and incubated with FITC-conjugated anti-human IgG secondary antibody. The enrichment of S1-F4 is then measured from the fluorescence intensity of FITC. Images are converted to IHC images by Inform software. The magnification is X20. [Diagram 32] Figure 32 is a graph showing the time course of antibody concentrations in serum of CD98 humanized mice. Mice were treated with the indicated antibodies (15 mg / kg, ip) on day 0. Serum concentrations of S1-F4 or H15L54 were measured by ELISA and are shown as mean ± SEM. [Diagram 33]Figures 33A-D show the antitumor activity of H15L54 or S1-F4 in CD98-humanized mice bearing EL4-hCD98 tumors (A), MC38-hCD98 tumors (B), B16F10-hCD98 tumors (C), and CB-17 SCID mice bearing A-431 tumors (D). S1-F4 or H15L54 (15 mg / kg) (A and C), H15L54 (20 mg / kg) (B), S1-F4 or H15L54 (15 mg / kg) (D). [Diagram 34] Figure 34 shows the antitumor activity of Ab8332 (8332) or S1-F4 in CD98-humanized mice bearing EL4-hCD98 tumors. Arrows indicate administration. Dose: 15 mg / kg. [Diagram 35] FIG. 35 shows the binding of S1-F4 and Ab8332 to hCD98 ECD-His6-Avi-Biotin protein at pH 6.5 and pH 7.4 as measured by ELISA. [Diagram 36] FIG. 36 shows the binding of different antibodies to hCD98 ECD-His6-Avi-Biotin protein at pH 6.5 and pH 7.4 as measured by ELISA. [Figure 37] FIG. 37 shows the distribution of S1-F4 and Ab8332 in hCD98ECD mice bearing MC38-hCD98 tumors as detected by IVIS. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0038] Unless expressly defined elsewhere herein, all technical and scientific terms used herein have the meaning commonly understood by one of ordinary skill in the art to which this invention belongs.

[0039] Materials and Methods Details of the experimental model and subjects Cynomolgus monkey experiments were performed at JOINN Laboratories (Beijing) according to an approved IACUC scheme.

[0040] Mouse experiments were performed at Beijing Institute for Life Sciences in accordance with the China National Laboratory Animal Breeding and Care Guidelines and with an approved IACUC scheme. C57BL / 6, CB-17 SCID and NOD SCID mice were purchased from Charles River. CD11c-DTR mice were purchased from Jackson Laboratory. CD98 humanized mice were bred and maintained in the Animal Care Facility of Beijing Institute for Life Sciences.

[0041] CHO or CHO-derived cell lines, HEK293T or HEK293T-derived cell lines, Raji, Ramos, HepG2, Hep3B, MDA-MB-231-LN, A549, PANC-1, EL4-hCD98, and B16F10-hCD98 cells were cultured in Dulbecco's Modification of Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum (FBS). Raji, Ramos, HL60, BxPC-3, HCT-8, HT-29, HCT 116, A-431, MC38-hCD98, and MCA205-hCD98 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum. These cells were cultured in a humidified incubator at 37°C with a 5% CO2 atmosphere. FreeStyle 293F cells were from Life Technologies and cultured according to the manufacturer's protocol.

[0042] Protein expression and purification FreeStyle 293F cells were transiently transfected to produce His6-Avi tagged fusion proteins CD98 ECD or FcγRs and purified by affinity chromatography. Full-length IgG antibodies included GC33 (Ishiguro et al., 2008) and IGN523 (Hayes et al., 2015), whose VH and VL coding sequences were subcloned into human IgG1 H chain (HC) and L chain (LC) expression vectors, respectively. Two types of IgG expression plasmids (HC plasmid + LC plasmid) were co-transfected into 293F cells at a 1:1 ratio. 3-6 days after transfection, cell culture supernatants were collected and IgG1 was purified by Protein A affinity chromatography.

[0043] Panning of antibody library and screening of anti-CD98 antibodies To generate anti-hCD98 antibodies (e.g., HN2-G9) or anti-mouse CD98 antibodies (e.g., BC8), the ECD of human or mouse CD98 was fused to a His6-Avi tag and biotinylated by BirA ligase. These two proteins were used as antigens in panning experiments with a human non-immune antibody library (Li et al., 2017). Phage-scFvs were screened by two rounds of panning that specifically bound to CD98 ECD. Approximately 400 monoclonals were randomly selected and screened for monoclonals that bound to human or mouse CD98 by ELISA. Selected clones were used to generate purified phage-scFv particles or converted to full-length human IgG1 format for further characterization.

[0044] Construction and selection of VH chain replacement sub-libraries The VL gene of HN2-G9 was cloned from non-immune VH genes (~1 × 10 10 The size of the constructed strand displacement sublibrary was ∼1 × 10 8Similar to the above antibody library selection using captured CD98 ECD-His6Avi, three rounds of library selection were performed, with the difference that HN2-G9 IgG1 was introduced as a competing antibody in the third panning round. Approximately 400 monoclonals were randomly selected and screened for monoclonals that bind to hCD98 by ELISA. The selected clones were either generated as purified phage-scFv particles or converted to full-length human IgG1 format for further characterization.

[0045] Screening of pH-dependent anti-hCD98 antibodies from a phage-displayed human non-immune Fab library To screen for pH-dependent antibodies, the constructed phage-Fab library and hCD98 ECD-His6-Avi-Biotin protein were incubated with pH 6.5 solution, washed with pH 6.5 wash buffer, and then eluted with pH 7.4 buffer. Then, monoclonals were randomly selected and screened by ELISA for monoclonals that bind to hCD98 at pH 6.5 and pH 7.4. Clones that bind to hCD98 at pH 6.5 significantly better than at pH 7.4 were converted to full-length human IgG1 format for further characterization.

[0046] Binding measurement by ELISA For antibody-antigen binding analysis by ELISA, biotinylated protein antigens were captured by streptavidin (Sigma-Aldrich)-coated 96-well plates (Nunc, MaxiSorp®), followed by serial dilutions of antibodies and detection by adding HRP-labeled goat anti-human IgG Fc polyclonal antibody (Thermo Fisher Scientific).

[0047] For antibody-human C1q binding analysis by ELISA, serially diluted antibodies were coated onto 96-well plates (Nunc, MaxiSorp®), followed by addition of 3% human complement serum (Sigma-Aldrich) and detection with HRP-labeled sheep polyclonal anti-human C1q (Abcam).

[0048] Binding kinetics analysis by surface plasmon resonance (SPR) Kinetic analysis of the binding of anti-CD98 antibodies to the ECD of CD98 and the ECD of FcγRs was performed on a Biacore T200 instrument (Biacore, GE Healthcare). Anti-hFc antibodies or Protein A / G (Thermo Fisher Scientific) were covalently coupled to the surface of a CM5 sensor chip using an amine coupling kit (GE Healthcare). An optimal concentration of antibody was captured on the chip, followed by two-fold serially diluted concentrations of analyte (CD98 or FcγRs). Binding kinetics were evaluated using a 1:1 Langmuir binding model. ka, kd and KD values ​​were calculated using the Biacore T200 evaluation software.

[0049] Flow cytometry analysis of cell lines To examine the expression of CD98 in tumor cell lines, tumor cells were labeled with anti-CD98 antibody and then with goat anti-human IgG Fc FITC polyclonal antibody (Thermo Fisher Scientific). FITC was used to detect CD98 expression in tumor cell lines.

[0050] To test the antibodies that bound to CD98 variants, we constructed hCD98-GL expression plasmids containing different alanine mutations. Then, we transfected CHO cells with these plasmids. After 2 days, the transfected CHO cells were labeled with IGN523 (Hayes et al., 2015), S1-F4 or anti-GL antibody (GC33) (Ishiguro et al., (2008) Cancer Res 68, 9832-9838), and then with goat anti-human IgG Fc FITC polyclonal antibody (Thermo Fisher Scientific).

[0051] Tissue processing and flow cytometric analysis of immune cells Tumor volume 500mm 3 If the tumor size exceeded 100%, the xenograft model mice were treated with antibodies. Three days after treatment, tumors were harvested from the mice and FACS analysis was performed on single cell suspensions. Briefly, tumors were isolated and treated with Red Blood Cell Lysis Buffer. The cell suspension was passed through a 40 μm cell filter to obtain a single cell suspension. The cells were then labeled with various antibodies (Key Resource Table).

[0052] Construction of human or mouse CD98 expressing cell lines First, expression plasmids were constructed by inserting DNA encoding human or mouse CD98. Then, the expression plasmids were transfected into HEK293T, CHO, EL4, MC38, MCA205, or B16F10 cells. After transfection, cells expressing hCD98 were selected by FACS and cultured in medium containing G418 to generate cell lines stably expressing hCD98.

[0053] Cell proliferation assay Cell proliferation was analyzed using WST-8 Cell Counting Kit-8 (Dojindo Molecular Technologies). Cells (10,000-25,000 cells / well) suspended in RPMI 1640 medium containing 1% FBS were seeded into a 96-well plate and incubated for 72 hours. Then, CCK-8 solution (10 μl) was added to each well and incubated at 37°C for 1-4 hours. The absorbance at 450 nm was measured using a microplate reader. The cell proliferation percentage was expressed as the percentage of the total number of cells in the measured sample group relative to the total number of cells in the untreated group.

[0054] Amino acid intake measurement Cells were incubated with 15 μg / ml of antibody in growth medium for 24 hours. Then, cells were equilibrated with DMEM without Met at 37°C for 30 minutes. HPG (Thermo Fisher Scientific) was added to the cells at a final concentration of 70 μM. At the same time as HPG addition, BCH (2-amino-2-norbornane-carboxylic acid) (Sigma-Aldrich) was added at a final concentration of 10 μM as a positive control. After 2-3 hours, cells were washed with PBS and lysed with lysis buffer (50 mM Tris-HCl, pH 8.0 with 1% SDS) containing complete protease inhibitors (Roche). HPG in the cell lysate was biotinylated according to the manufacturer's instructions (Thermo Fisher Scientific). The cell lysate was then transferred to a 96-well plate (Nunc, MaxiSorp®) and incubated overnight at 4°C. The amount of HGP in the cell lysate was then detected using streptavidin-HRP (Thermo Fisher Scientific) by ELISA.

[0055] ADCC, ADCP and CDC measurements To perform ADCC measurements, target cells (10,000-20,000 cells / well) were seeded into wells of a U-bottom 96-well cell culture plate and incubated with different concentrations of antibodies for some time. Then, effector cells (E:T=4:1-10:1) were added to the wells containing target cells and antibodies and incubated at 37°C for 4-8 hours in RPMI 1640 medium supplemented with 5% FBS. ADCC activity was measured by lactate dehydrogenase (LDH) release according to the instructions of the Cyto Tox96® Non-Radioactive Cytotoxicity Assay Kit (Promega). Cytotoxicity rate was calculated according to the manufacturer's instructions.

[0056] To perform ADCP measurements, mouse bone marrow-derived macrophages (mBMDMs) and human peripheral blood mononuclear cells (hPBMCs) were used as effector cells. Mouse bone marrow cells were harvested from the tibia and femur of C57BL / 6 mice and induced with GM-CSF in L929 supernatant for 3 days to generate mBMDMs. Raji cells were labeled with CFSE (Thermo Fisher Scientific) and used as target cells. mBMDMs were labeled with anti-mouse F4 / 80-Alex Fluor647 (Thermo Fisher Scientific) and then incubated with target cells. CFSE-labeled target cells were incubated with different antibodies for 15 min at room temperature, then added to the labeled mBMDMs in DMEM medium supplemented with 10% heat-inactivated FBS at an E:T ratio of 1:2 and incubated for 2 h at 37 °C. The phagocytosis of CFSE-labeled target cells by anti-mouse F4 / 80 antibody-labeled macrophages was recorded using a Nikon A1R confocal microscope. hPBMCs were induced to differentiate with 20 ng / mL macrophage colony-stimulating factor (M-CSF) (Pepro Tech) for 9 days and used as effector macrophages. ADCP measurement using hPBMCs was the same as above, except that hPBMCs were labeled with Deep-Red dye (Thermo Fisher Scientific).

[0057] To perform CDC measurements, target cells were seeded at 400,000 cells / well in U-bottom 96-well plates and incubated with various antibodies in the presence of 5% human serum (Sigma-Aldrich). After 2 h of incubation, LDH release in the supernatant of each well was analyzed using the Cyto Tox96® non-radioactive cytotoxicity assay kit (Promega).

[0058] Pharmacokinetics and safety evaluation of S1-F4 in monkeys Healthy cynomolgus monkeys (3-4 years old) weighing approximately 3 kg were intravenously injected (iv) with S1-F4 on days 0 and 15. Blood samples were collected at different time points to measure serum antibody concentrations by hCD98 binding ELISA. Antibody serum concentrations were then calculated and plotted with Graphpad Prism 6. PK data were evaluated with WinNonlin software. Body weight, temperature, blood biochemistry (alanine aminotransaminase (ALT), aspartate aminotransferase (AST) and creatinine), red blood cell content (RBC) and white blood cell content (WBC) were evaluated at the designated time points at JOINN Laboratories (Beijing).

[0059] Pharmacokinetic analysis in mice C57BL / 6 or CD98 humanized mice aged 6-8 weeks were used in the study. After a single intraperitoneal (ip) injection of the antibody to be analyzed, blood was collected at different time points. Antibody concentrations in serum were measured using a human IgG ELISA quantification kit (Bethyl Laboratories). Antibody serum concentrations were then calculated and plotted using Graphpad Prism 6. PK data were evaluated using WinNonlin software.

[0060] Structural characterization of the S1-F4 scFv and hCD98 ECD complex For X-ray crystallographic analysis, S1-F4-scFv-His6 and hCD98 ECD recombinant proteins were used. The amino acid sequence of hCD98 ECD corresponds to the Glu111-Ala529 residues of hCD98. S1-F4-scFv-His6 was expressed in FreeStyle 293F cells, and hCD98 ECD was expressed in E. coli. S1-F4-scFv-His6 was purified by immobilized metal ion affinity chromatography (IMAC) using Ni-NTA agarose microbeads (QIAGEN), and hCD98 ECD was purified first with the same microbeads and then with a HiTrap Q HP anion exchange chromatography column (GE Healthcare). S1-F4-scFv-His6 and hCD98 ECD were then mixed to form a complex, which was then purified by size exclusion chromatography with a Superdex S200 10 / 300 GL column (GE Healthcare). 1 μL of protein (dissolved in 10 mM Tris-HCl pH 8.0 and 150 mM NaCl at a concentration of 10 mg / mL) was mixed with 1 μL of a stock solution containing 0.1 M trisodium citrate dihydrate (pH 5.0, 9% PEG20000, 5% PEG400, 9% glycerol), and the purified S1-F4-scFv-His6 / hCD98 ECD complex was concentrated and crystallized at 20°C by hanging drop vapor diffusion.

[0061] Lamellar crystals appeared after 7 days. X-ray diffraction data were collected at beamline BL19U1 of Shanghai Synchrotron Radiation Light Source and processed by XDS. Starting from the hCD98 ECD (PDB 2DH2) structure, the structure was determined at 2.8 Å resolution by molecular replacement with Phaser. The initial model from molecular replacement was further refined with Phenix and manually rebuilt with Coot. The final model includes 235 residues in S1-F4 scFv and residues 115-526 in hCD98 ECD. MolProbity analysis showed that 96.55% of the residues were located in the favored region and 3.41% of the residues were located in the allowed region.

[0062] In vivo tumor assay For the mouse tumor models, 6-8 week old mice were inoculated subcutaneously with various tumor cells (in 100 μL of DPBS or medium). MDA-MB-231-LN cells were inoculated into the mammary fat pad of female mice, while the remaining cells were all inoculated on the right side. For the xenograft tumor models, mice were randomly divided into groups (n=2-6 / group) based on similar mean tumor volume and intraperitoneally injected with various antibodies. For the syngeneic tumor models, tumor cells (1×10 5 After inoculation with 1000 cells / mL of DPBS, mice were randomly divided into groups (n=3-10 / group) and intraperitoneally injected with various antibodies. Tumor volumes were measured with electronic calipers and calculated using the modified ellipsoidal formula 1 / 2×(length×width). 2 Tumor volumes were calculated using the following formula: (a) 100% 100% 15% 10% 15% 20% 25% 30% 35% 40% 40% 50% 60% 70% 80% 90% 100% 15% 20% 35% 40% 50% 60% 70% 80% 90% 100% 1

[0063] TIFF0007676594000001.tif129154

[0064] CD4 + or CD8 +To eliminate T cells, tumor-bearing CD98-humanized mice were injected with 15 mg / kg anti-CD4 antibody (clone GK1.5, BioXCell) or 10 mg / kg anti-CD8α antibody (clone 2.43, BioXCell) every 3-5 days, 1 day before S1-F4 treatment. To eliminate NK cells, mice were injected with 2.5 mg / kg anti-Asialo-GM1 polyclonal antibody (Poly21460, Biolegend) every 6 days, 1 day before S1-F4 treatment. To eliminate neutrophils, mice were injected with 20 mg / kg anti-Ly6G antibody (clone 1A8, BioXCell) every 3 days, 1 day before S1-F4 treatment. To eliminate macrophages, mice were injected with 25 mg / kg of anti-CSF1R antibody (clone AFS98, BioXCell) every 3 days 1-2 days before S1-F4 treatment. To eliminate dendritic cells, mice were injected with 4 μg / kg of diphtheria toxin (Sigma-Aldrich) every 2 days 1 day before S1-F4 treatment. The efficacy of the above immune cell depletion method was demonstrated using tumor-naive mice.

[0065] To perform tumor rechallenge studies, 1 × 10 5 Tumor cells were inoculated into the flanks of C57BL / 6 mice that showed complete response (CR) to EL4-hCD98 tumors after S1-F4 treatment and age-matched naïve C57BL / 6 mice (EL4 and EL4-hCD98 inoculated on the left side, B16F10 and B16F10-hCD98 inoculated on the right side). Tumors were measured as described above.

[0066] Immunofluorescence labeling assay To detect antibody enrichment in mouse kidneys, mice were sacrificed 2-3 days after injection of various antibodies. Mouse kidneys were then harvested and frozen sectioned. Kidney sections were labeled with goat anti-human IgG Fc FITC (Thermo Fisher Scientific) polyclonal antibody. CD98 expression on the kidney cell surface was detected by FITC using a Vectra Polaris instrument (PerkinElmer).

[0067] To examine CD98 expression in tumors, tumors were harvested and frozen sections were prepared. Tumor sections were labeled with BC8 or IGN523 (10 μg / ml) followed by goat anti-human IgG Alexa Fluor 633 polyclonal antibody (Thermo Fisher Scientific). CD98 expression on the tumor cell surface was detected by Alexa Fluor 633 using a Nikon A1R confocal microscope.

[0068] To examine CD98 expression in mouse tissues, various mouse tissues were collected and frozen sections were prepared. The sections were labeled with BC8 or IGN523 (10 μg / ml) and then with goat anti-human IgG Fc FITC polyclonal antibody (Thermo Fisher Scientific). FITC was used to detect CD98 expression in mouse tissues.

[0069] Optical imaging in tumor-bearing mice Tumor-bearing mice were optically imaged using an IVIS spectrometer (PerkinElmer) and analyzed using Live Image 4.4 software (PerkinElmer). All images in one experiment were acquired using identical illumination settings, and fluorescence emission spectra were normalized according to routine procedures in bioluminescence imaging.

[0070] To perform fluorescence imaging of antibody-Cy7 distribution in CD98 humanized mice, tumor-bearing CD98 humanized mice (n=2-3 / group) were intraperitoneally injected with vehicle or Cy7-labeled antibodies (S1-F4-Cy7 (Cy7 / antibody=0.845) or H15L54-Cy7 (Cy7 / antibody=1.076)). Mice were dissected and fluorescence images were obtained at different time points. Fluorescence images were acquired by an IVIS spectrometer equipped with 745 nm excitation and 800 nm emission filters.

[0071] Bioluminescence imaging of the MDA-MB-231-LN tumor model was performed by injecting 150 mg / kg D-fluorescein (PerkinElmer) into mice bearing MDA-MB-231-LN tumors (peritoneally). Tumor bioluminescence was measured 10 min after D-fluorescein injection. Imaging was performed every 4-5 days until the last day of survival for all mice in all groups.

[0072] statistical analysis In this specification, specific comparisons are made using GraphPad Prism 6 (GraphPad), and p values ​​are shown in the relevant figures. Two-way ANOVA or two-tailed unpaired Student's t-test is adopted. Antitumor activity is analyzed by two-way ANOVA. p<0.05 is considered statistically significant (ns: no significant difference, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001). Working Example

[0073] Example 1. Generation of anti-hCD98 antibody S1-F4 A recombinant protein containing the hCD98 extracellular domain (hCD98 ECD) was generated and used to select anti-hCD98 antibodies from a non-immune phage display single-chain Fv (scFv) human antibody library (Li et al., (2017) Elife6). Antibodies binding to hCD98-expressing CHO cells (CHO-hCD98) were analyzed by FACS to screen for the best performing antibodies. From the seven candidates, HN2-G9 was identified and its antitumor activity was evaluated using xenograft mouse tumor models (Raji, Burkitt lymphoma). HN2-G9 showed antitumor activity comparable to the prior art anti-CD98 antibody IGN523 (Hayes et al., 2015) (Figure 1). However, surface plasmon resonance (SPR) analysis showed that HN2-G9 had a relatively low binding affinity for CD98 (1.29 μM) compared to IGN523 (Figure 3).

[0074] To improve the binding affinity of HN2-G9 to CD98, we genetically engineered HN2-G9 using variable heavy chain (VH) chain replacement (Li et al., 2017). Both FACS and SPR analyses identified an antibody (S1-F4) with significantly higher binding affinity (58.2 nM) than HN2-G9 (Figures 2-3).

[0075] We then measured the antitumor activity of S1-F4 against Raji xenograft tumors, and found that it exerted a significantly stronger antitumor effect than IGN523 and was similar to that of rituximab, an anti-CD20 antibody approved by the FDA for treating B-cell lymphoma (Figure 4).

[0076] Example 2. Broad-spectrum antitumor activity of S1-F4 in xenograft tumor models Because CD98 is highly expressed in many tumor types, the antitumor effects of S1-F4 were evaluated in various xenograft tumor models.

[0077] Prior to the in vivo study, CD98 expression was detected in 13 human tumor cell lines (including Raji) derived from different tissues (Figure 5). All tested cell lines expressed high levels of CD98. Therefore, these cell lines were used to establish xenograft tumor models in immunodeficient CB-17 SCID or NOD SCID mice to evaluate the therapeutic effect of S1-F4.

[0078] S1-F4 treatment demonstrated broad-spectrum antitumor activity in 8 of 13 xenograft models (Figure 4, Figure 6-Figure 10). Of particular note, low-dose (1 mg / kg) S1-F4 treatment led to complete tumor regression in the HepG2 (hepatocellular carcinoma) tumor model (Figure 6).

[0079] Example 3. Generation of a mouse model expressing hCD98 (CD98 humanized mice) S1-F4 was found to bind to human and monkey CD98, but not to mouse CD98 (mCD98) (Table 1). Considering that CD98 is widely expressed in many normal human tissues, the results obtained from the above mouse model with human xenografts may not accurately reflect the potential efficacy and / or safety of S1-F4 in humans.

[0080] Table 1. Binding affinity of S1-F4 to CD98 from different sources. [Table 1]

[0081] To solve this problem, we replaced the mCD98 ECD with its human counterpart in a C57BL / 6 genetic background (using C57BL / 6 mice as recipients) by CRISPR / Cas9 to generate a CD98-humanized mouse model (Figure 11A-D). We used an sgRNA (sgRNA-CD98:ccgcgctgccgtgagctgcctgt, SEQ ID NO:1) targeting a sequence within the first intron to integrate the hCD98 ECD (amino acids 107-529) sequence into the Slc3a2 locus in C57BL / 6 mouse germ cells and ablate the mouse Slc3a2 gene. As a template for homology-directed repair (HDR), we constructed a targeting vector containing a 1272-bp sequence encoding amino acids 107-529 of hCD98, a polyA element, an 829-bp upstream fragment and a 945-bp downstream fragment extending from the sgRNA-CD98 target site (Figure 11A). Cas9 protein, sgRNA-CD98 and targeting vector were pronuclear microinjected into C57BL / 6 fertilized eggs to obtain two live mice (F0). Correct insertion of hCD98 ECD in the two F0 mice was verified by multiple pairs of primers (Figure 11A-D, Table 2). Heterozygous F0 mice were crossed with wild-type C57BL / 6 mice to obtain F1 generation mice. Heterozygous F1 animals were then crossed to generate homozygous hCD98 knock-in mice (named CD98 humanized mice).

[0082] Table 2. PCR primers used to identify CD98-humanized mice. [Table 2]

[0083] The generated CD98 humanized mouse model retains the physiological function of mCD98, so that the expression of CD98 can be detected in normal tissues by S1-F4 and the immune system is intact. By immunofluorescence labeling assay, we detected the expression spectrum of hCD98 in different tissues of the established CD98 humanized mice, and confirmed that the expression pattern of hCD98 in tissues was similar to that of mCD98 in wild-type mice (Figures 12-13).

[0084] Example 4. Antitumor activity of S1-F4 against aggressive and refractory tumors Using CD98-humanized mice, we evaluated the antitumor activity of S1-F4 in four aggressive and refractory mouse tumor models. The models were established using mouse tumor cell lines stably expressing hCD98 (Figure 14).

[0085] S1-F4 treatment showed significant antitumor activity against three of the four tumor models, EL4-hCD98, MC38-hCD98, and MCA205-hCD98, but no significant effect against B16F10-hCD98 tumors (Figure 15A). These results indicate that S1-F4 exerts broad-spectrum antitumor activity against a variety of malignancies in immunodeficient xenograft and immune-activated syngeneic mouse models.

[0086] Furthermore, when the effect of S1-F4 treatment on EL4-hCD98 and B16F10-hCD98 tumors was assayed in wild-type C57BL / 6 mice (i.e., mice that do not express any hCD98 in non-tumor tissues), S1-F4 treatment induced a significant inhibition of tumor growth in both tumor models (Figure 15B). The antitumor effect of S1-F4 on B16F10-hCD98 in wild-type mice but not in CD98-humanized mice indicates that a lack of "antigen sink" in wild-type mice may contribute to the antitumor effect of S1-F4 in such highly aggressive tumor models. "Antigen sink" refers to antigen-mediated clearance of antibodies targeting cell membrane antigens. These results indicate that the antitumor effect of anti-CD98 antibodies observed in wild-type mice bearing human xenografts does not necessarily reasonably predict its antitumor effect in humans.

[0087] Example 5. Dependence of antitumor activity of S1-F4 on Fc-FcγR interaction This example investigated whether the anti-tumor activity of anti-hCD98 antibodies depends on Fc-mediated immune responses, such as complement-dependent cytotoxicity (CDC), antibody-dependent cell-mediated cytotoxicity (ADCC) or phagocytosis (ADCP).

[0088] To analyze the Fc-related functions of S1-F4, KA and S1-F4 DANA We generated two S1-F4 variants (Figure 3). KA S1-F4 carries the K322A (KA) mutation to abolish Fc and C1q binding and to abolish CDC function. DANA S1-F4 carries the D265A / N297A (DANA) mutation to abolish Fc binding to FcγRs and to abolish ADCC and ADCP functions. The ability of S1-F4 and its two variants to mediate ADCC, ADCP, and CDC was measured. In vitro assays showed that S1-F4 and S1-F4 KAinduced ADCC (Fig. 16A, B) and ADCP killing effects (Fig. 17A, B) against Raji and / or HepG2 cells, whereas S1-F4 DANA Neither S1-F4 nor the mutants tested were able to bind human C1q or induce CDC killing effect on Raji cells (Fig. 18A and B).

[0089] These results indicated that S1-F4 could bind to FcγRs and induce ADCC and ADCP killing effects, but could not bind to C1q or induce CDC killing effects.

[0090] Example 6. Effect of S1-F4 on CD98 function This example investigated whether S1-F4 interferes with the biochemical or cell biological functions of CD98 in tumor cell lines previously found to be highly sensitive to S1-F4 treatment.

[0091] Amino acid uptake assays were performed to analyze the effect of S1-F4 on amino acid transport. The methionine mimetic molecule L-homopropargylglycine (HPG) was used as a traceable substrate for the LAT1 (L-type / large neutral amino acid transporter 1) / CD98 heterodimer HAT complex, and the LAT1 / CD98 complex inhibitor BCH (2-amino-2-norbornane-carboxylic acid) was used as a positive control to disrupt transport function.

[0092] As shown in Figure 19, when the HPG uptake efficiency of cells treated with control IgG, S1-F4, and BCH was compared, only BCH impaired the transport activity of the HAT complex. These results indicated that S1-F4 did not affect amino acid transport.

[0093] Because CD98 enhances integrin signaling and increases cell proliferation, the inhibitory effect of S1-F4 on the proliferation of HepG2, HCT-8, Raji, and Ramos cells (four cell lines highly sensitive to S1-F4) was measured in a cell proliferation assay.

[0094] As shown in Figure 20, S1-F4 dose-dependently inhibited the proliferation of Raji and Ramos cells, but no S1-F4-mediated effect on the proliferation of HepG2 and HCT-8 cells was observed. These results indicated that inhibition of cell proliferation is not necessary for the antitumor activity of S1-F4.

[0095] We also investigated the mechanism of action (MOA) of S1-F4 by comparing the antitumor effects of S1-F4 and its two Fc variants in xenograft and syngeneic tumor models. For the xenograft tumor models, HepG2 and Raji tumors were selected as typical resistant and sensitive cells to cell growth inhibition by S1-F4, respectively. For the syngeneic tumor models, EL4-hCD98 and MC38-hCD98 cells were selected to establish tumors in CD98-humanized mice.

[0096] S1-F4 DANA showed no therapeutic effect in any of the four models (Figure 21), indicating that the antitumor activity of S1-F4 is dependent on Fc-FcγR interactions. KA The inhibitory effects of S1-F4 on tumor growth were similar (Figure 21, lower left panel), demonstrating that CDC function is not required for the antitumor activity of S1-F4. In summary, these results demonstrated that the antitumor activity of S1-F4 depends on ADCC and ADCP mediated by Fc-FcγR interaction.

[0097] Example 7: Involvement of immune cells in the antitumor activity of S1-F4 To determine which immune cell subpopulation is involved in the antitumor activity of S1-F4, we performed experiments using a HepG2 tumor model and depleted macrophages, NK cells, or neutrophils using subpopulation-specific antibodies before and during S1-F4 treatment, respectively. We found that the antitumor activity of S1-F4 was lost by the clearance of macrophages, but the clearance of neutrophils and NK cells did not affect the antitumor activity of S1-F4. In the HepG2 xenograft model, the antitumor activity of S1-F4 is presumed to be dependent on macrophages.

[0098] Two xenograft models (A549 and A-431) were resistant to S1-F4 treatment, and the effect of S1-F4 treatment on macrophage infiltration was detected. S1-F4 treatment significantly increased the proportion of macrophages in HepG2 tumors by more than two-fold. In contrast, S1-F4 treatment did not increase the number of macrophages infiltrated in S1-F4-resistant A-431 tumors or HCT 116 tumors. These results indicated that the lack of macrophage infiltration in tumors may be one of the reasons why tumor cells are resistant to S1-F4 treatment.

[0099] In addition to the immunodeficient xenograft tumor model, the effect of macrophage, NK cell or neutrophil depletion on S1-F4 antitumor efficacy was also evaluated in the EL4-hCD98 syngeneic tumor model established using CD98 humanized mice. Consistent with the results obtained in the HepG2 tumor model, only macrophage clearance affected the antitumor efficacy of S1-F4, whereas neither neutrophil nor NK clearance affected the antitumor activity of S1-F4. Thus, in both xenograft and syngeneic tumor models, macrophages are crucial for the antitumor action of S1-F4.

[0100] CD98-humanized mice have an intact immune system and therefore provide a suitable model system to study the effect of T cell immune responses in S1-F4 antitumor efficacy. To determine whether S1-F4 can recruit T cells to attack tumor cells, we utilized the EL4-hCD98 tumor model established in CD98-humanized mice to investigate the role of CD4 + T cells or CD8 + We evaluated the effect of T cell clearance on the antitumor activity of S1-F4. + Depletion of T cells did not alter the antitumor activity of S1-F4, which is consistent with the CD4 + These results suggest that CD8 T cells are not involved in the antitumor activity of S1-F4. + Clearance of T cells significantly attenuated the inhibitory effect of S1-F4 on EL4-hCD98 tumor growth. + T cells are involved in the antitumor activity of S1-F4, and macrophages alone are insufficient to induce long-term antitumor effects.

[0101] The present inventors further demonstrated that dendritic cells (DCs) express CD8 + We investigated whether DCs play a role in priming APCs of T cells. Specifically, CD11c-DTR (diphtheria toxin receptor) (Itgax-DTR / EGFP) mice bearing EL4-hCD98 tumors were treated with diphtheria toxin (DT) before (and during) S1-F4 treatment to remove DCs. Clearance of DC cells significantly affected the inhibitory effect of S1-F4 on EL4-hCD98 tumor growth. Thus, DCs are crucial for the complete antitumor response induced by S1-F4.

[0102] Finally, we found that C57BL / 6 mice that showed complete response (CR) to EL4-hCD98 tumors by S1-F4 treatment were protected when subsequently challenged with the same EL4-hCD98 tumor cells, or EL4, B16F10, or B16F10-hCD98 tumor cells, and showed significant tumor growth delay or complete tumor rejection, indicating that S1-F4 has an effect in preventing tumor recurrence (Figure 22). As expected, CR mice could not survive the challenge with B16F10 tumor cells (Figure 22), suggesting that B16F10 and EL4-hCD98 tumor cells are unlikely to have a common antigen, supporting the ability of S1-F4 treatment to induce antigen-specific antitumor immune memory.

[0103] Considering these results comprehensively, the present inventors have concluded that the MOA of S1-F4 is as follows: 1) S1-F4 binds to FcγRs on macrophages via its Fc, and induces ADCC and ADCP to attack tumor cells expressing CD98; 2) tumor cell death in step 1) produces tumor cell-associated antigens, which are then delivered to DCs expressing CD8. + and 3) cross-presentation to CD8 T cells. + The step involves T cells being activated and attacking tumor cells. In this process, macrophages play a role in initiating the antitumor response, but they also express cytotoxic CD8 + T cells are induced after the initiation of antitumor responses, particularly through enhanced DC-mediated cross-presentation, to express cytotoxic CD8 + Thus, S1-F4 treatment may bridge the innate and adaptive T cell immune systems to attack tumor cells and induce long-term immunological memory.

[0104] Example 8. Generation of pH-dependent anti-hCD98 antibodies We evaluated the safety and pharmacokinetic (PK) properties of S1-F4 in cynomolgus monkeys. Dual S1-F4 treatment (20 mg / kg on day 0 and 10 mg / kg on day 15) did not cause significant side effects, indicating that S1-F4 has the necessary safety profile at therapeutically effective doses.

[0105] However, in monkeys, S1-F4 serum concentrations declined rapidly over time (Figure 23A), and the half-life was very short (t 1 / 2 In addition, serum S1-F4 concentrations in C57BL / 6 and CD98-humanized mice after single-agent administration were also measured, and the serum S1-F4 concentration in CD98-humanized mice was found to have decreased rapidly (t 1 / 2 ~69.6hr), whereas in C57BL / 6 mice it declined slowly (t 1 / 2 ~341.5 hr) (Figure 23B).

[0106] After treatment with fluorescently labeled S1-F4 (S1-F4-Cy7), strong fluorescent signals were observed in the kidneys of CD98-humanized mice (Figure 24A), consistent with the strong expression of mouse CD98 in the kidneys. Immunofluorescence labeling also showed that after S1-F4 treatment, S1-F4 was abundantly enriched in the kidneys of CD98-humanized mice, but not in C57BL / 6 mice (Figure 24B). The rapid clearance of S1-F4 in cynomolgus monkeys and CD98-humanized mice is speculated to be due to the "antigen sink" phenomenon (S1-F4 binds to CD98 protein present in normal tissues).

[0107] To overcome the influence of the antigen sink effect and seek to improve the PK characteristics, we creatively engineered S1-F4 into a pH-dependent antibody that binds strongly to CD98 under acidic conditions (e.g., tumor microenvironment) (pH 6.5-6.9) but weakly (or not at all) under neutral conditions (pH 7.2-7.5) found in most normal tissues. A suitable pH-dependent antibody can preferentially bind to antigens in solid tumors while maintaining S1-F4 antitumor activity.

[0108] To support the rational engineering of S1-F4, the crystal structure of the complex between hCD98 ECD and S1-F4 was analyzed to obtain precise information on the binding interface (Figure 25). The crystal structure of the S1-F4 scFv-hCD98 ECD complex was analyzed at 2.8 Å resolution and showed that S1-F4 recognizes a conformational epitope on the hCD98 ECD, including residues 376-384 and 391-399 from two cyclic regions (Figure 24). The structure revealed that seven hCD98 ECD residues (L378, P379, V387, L389, F395, I398, and V402) form a hydrophobic core, stabilizing the conformation of the two loops and promoting the formation of the conformational epitope (Figure 25).

[0109] Evaluation of the epitope bound to the VH and variable light chain (VL) of S1-F4 on the specific structure showed that residues 391-399 of hCD98 ECD bound to LCDR1, LCDR3, and HCDR3 of S1-F4. The side chain hydroxyl groups of residues Y27d and Y32 of S1-F4 LCDR1 formed hydrogen bonds with the main chain amide groups of CD98 ECD residues E392 and P396, respectively. Additionally, the aromatic side chains of Y32 of S1-F4 LCDR1 and Y92 of LCDR3 sandwiched the P399 residue of CD98 ECD through non-polar interactions (Figure 25). The side chain of D397 of CD98 ECD formed hydrogen bonds with the main chain of R100d in the HCDR3 region of S1-F4. Additionally, S1-F4 VH mediated the interaction of the antibody with residues 376-384 of CD98 ECD. In this loop, a kink formed by P379 and G380 fits tightly into a hydrophobic groove containing S1-F4 VH residues Y33, I50, and I100f, which are located in HCDR1, HCDR2, and HCDR3, respectively. In the vicinity of this epitope, the side chain of H135 of CD98 ECD and Y97 of S1-F4 HCDR3 are hydrogen bonded (Figure 25).

[0110] FACS binding assays showed that alanine substitutions of two amino acids located at the binding interface (L389A and P399A) significantly weakened the binding of S1-F4 to hCD98, supporting that the conformational epitope certainly mediates the binding of the S1-F4 scFv-hCD98 ECD complex. The L378A, F395A, and D397A mutations also weakened binding, supporting that these residues in the hydrophobic core affect the stability of the binding affinity of S1-F4 scFv-hCD98 ECD. None of these alanine substitution mutations affected the binding of IGN523 to hCD98, indicating that S1-F4 and IGN523 bind to different epitopes.

[0111] Previous studies have shown that protonation of ionizable residues such as histidine (H) can promote pH-dependent binding (Chaparro-Riggers et al., 2012; Johnston et al., 2019; Sarkar et al., 2002). Therefore, we hypothesized that increasing the degree of interaction between H and acidic amino acids (aspartic acid (D) or glutamic acid (E)) may promote pH-dependent binding of S1-F4 to CD98. The structure shows that four acidic residues (E384, D391, E392, D397) are located within the S1-F4 epitope on hCD98 ECD (Figure 26), and H135 is located near the epitope and makes close contact with Y97 of S1-F4 HCDR3 (Figure 25), thereby providing a molecular basis for rational genetic engineering of pH-dependent S1-F4 variant antibodies.

[0112] First, a Y97E mutation was introduced into S1-F4 HCDR3 to generate S1-F4 Y97E. The binding activity of S1-F4 to CD98 ECD was comparable at pH 6.5 (EC50=0.059nM) and pH 7.4 (EC50=0.046nM) as measured by ELISA. In contrast, the binding of S1-F4 Y97E to CD98 was significantly stronger at pH 6.5 (EC50=1.686nM) than at pH 7.4 (EC50=4.527nM) (Figure 27), verifying that changing residues in close contact with acidic residues in the epitope to acidic residues is an effective strategy to obtain antibody variants with pH-dependent antigen binding.

[0113] To obtain anti-CD98 antibodies with improved binding activity and pH-dependence, we also screened a phage display sub-library derived from S1-F4 VHY97E reflecting mutations at four acidic residue positions facing CD98: D397 facing HCDR3, E384 facing HCDR2, and D391 and E392 facing LCDR1 (FIG. 26). Another pH-dependent anti-hCD98 antibody, H15L54, was identified.

[0114] H15L54 had a binding activity at pH 6.5 (EC50=0.247nM) approximately 7-fold higher than that of S1-F4 Y97E (EC50=1.686nM), and its pH dependence of binding (EC50 ratio, pH7.4 / 6.5=38.0) was approximately 14-fold higher than that of S1-F4 Y97E (EC50 ratio, pH7.4 / 6.5=2.7) (Figure 27). FACS showed that S1-F4 did not bind to A-431 cells at pH6.5 or pH7.4, but H15L54 bound to A-431 cells at pH6.5 significantly stronger than that at pH7.4 (Figure 28).

[0115] The model structure artificially constructed with PyMOL showed that after S1-F4 was genetically engineered, four pairs of HD / E interactions were formed between H15L54 and CD98, and the VH H54, VHE97, and VH H100 of H15L54 were involved. d and VL H27f We showed that H15L54 interacts with E384, H135, D397, and E392 of the CD98 ECD, respectively. These four pairs of HD / E interactions may contribute to the highly selective binding between H15L54 and CD98 at low pH.

[0116] Example 9. Preferential binding of pH-dependent anti-hCD98 antibodies to CD98 in tumors and improved anti-tumor activity The distribution of the antibody was examined by treating CD98-humanized mice bearing EL4-hCD98 tumors with fluorescently labeled H15L54-Cy7 antibody. S1-F4-Cy7 accumulated in the kidney, whereas H15L54-Cy7 accumulated mainly in the tumor (Figures 29-30). Using immunofluorescence labeling, we detected the enrichment of S1-F4 and H15L54 antibodies in the kidney of CD98-humanized mice. Consistent with previous findings, mice treated with S1-F4 had a strong FITC signal in the kidney, whereas mice treated with H15L54 or control IgG had only a very weak FITC signal in the kidney (Figure 31). Comparison of the PK properties of S1-F4 and H15L54 in CD98-humanized mice revealed that the serum concentration (t 1 / 2 ~217.1hr) is S1-F4(t 1 / 2 The results showed that the CD98 binding time was decreased much slower than that in normal tissues at physiological pH in vivo (~69.6 hr) (Figure 32). These results supported that H15L54 preferentially binds to CD98 in tumors rather than in normal tissues at physiological pH in vivo due to its pH-dependent binding properties.

[0117] The antitumor activity of H15L54 was evaluated in various tumor models. In EL4-hCD98 and MC38-hCD98 tumor models established in CD98-humanized mice, both of which are sensitive to S1-F4 treatment, H15L54 exhibited significant antitumor activity (Figure 33A-B). H15L54 also exhibited significant antitumor activity in tumor models resistant to S1-F4 treatment, including B16F10-hCD98 and A-431 xenograft tumor models established in CD98-humanized mice (Figure 33C-D). Thus, the pH-dependent anti-hCD98 antibody H15L54, by preferentially binding to the antigen at the relatively low pH condition of the tumor microenvironment, avoids the potential "antigen sink" problem of S1-F4 and provides the necessary PK properties and improved antitumor activity.

[0118] Example 10. pH-dependent anti-CD98 antibody In addition to H15L54, pH-dependent antibodies Ab8332, H15L1, and H15L35 were also identified in a similar manner, and their binding activity, pH dependence, and antibody distribution were examined in a mouse model.

[0119] When administered at the times indicated (arrows) and at a dose of 15 mg / kg (Figure 34), Ab8332 demonstrated tumor growth inhibitory activity comparable to S1-F4 in an established EL4-hCD98 tumor model in CD98-humanized mice.

[0120] Binding of S1-F4 and Ab8332 to hCD98 ECD-His6-Avi-Biotin protein at pH 6.5 and pH 7.4 was detected by ELISA, confirming the pH-dependent binding property of Ab8332. As shown in Figure 35, the EC50 ratio (pH 7.4 / 6.5) of Ab8332 was high at 50.7, but there was no significant difference with S1-F4. The EC50 ratios (pH 7.4 / 6.5) of H15L1 and H15L35 were 11.8 and 16.0, respectively, as measured by ELISA (Figure 36). The binding of H15L1 and H15L35 to the A-431 tumor model detected by FACS also supported the pH-dependent binding (Figure 28).

[0121] Similar to H15L54, fluorescence imaging of Ab8332-Cy7 in hCD98ECD mice bearing MC38-hCD98 tumors at days 2 and 6 post-injection showed that it was primarily concentrated in the tumors and not in the kidneys (Figure 37).

[0122] Sequence Listing The amino acid and nucleotide sequences of the S1-F4, H15L54 and Ab8332 variable regions can be found in the sequence listing with the SEQ ID NOs shown in the table below. In this application, the CDRs are determined based on the Kabat numbering scheme.

[0123] Table 3: S1-F4, H15L1, H15L35, H15L54 and Sequence information of 8332 anti-CD98 antibody [Table 3] The present disclosure includes the following embodiments. [1] An antibody or antigen-binding fragment thereof that binds to human or mouse CD98, A heavy chain CDR1 (HCDR1) comprising or consisting of an amino acid sequence selected from SEQ ID NO: 10, 20, 30, 40, 50, 60, and 70; A heavy chain CDR2 (HCDR2) comprising or consisting of an amino acid sequence selected from SEQ ID NOs: 11, 21, 31, 41, 51, 61, and 71; A heavy chain CDR3 (HCDR3) comprising or consisting of an amino acid sequence selected from SEQ ID NO: 12, 22, 32, 42, 52, 62, and 72; A light chain CDR1 (LCDR1) comprising or consisting of an amino acid sequence selected from SEQ ID NO: 15, 25, 35, 45, 55, 65, and 75; A light chain CDR2 (LCDR2) comprising or consisting of an amino acid sequence selected from SEQ ID NOs: 16, 26, 36, 46, 56, 66, and 76; and An isolated antibody or antigen-binding fragment thereof, comprising a light chain CDR3 (LCDR3) comprising or consisting of an amino acid sequence selected from SEQ ID NO:17, 27, 37, 47, 57, 67, and 77. [2] Any of the following (a) to (g) is included: An antibody or a fragment thereof comprising any one of (a) to (f) specifically binds to hCD98; The antibody or fragment thereof of embodiment 1, which specifically binds to mouse CD98, comprising (g) (a) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 10, 11 and 12, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 15, 16 and 17, (b) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 20, 21 and 22, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 25, 26 and 27, respectively. (c) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 30, 31 and 32, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 35, 36 and 37, respectively. (d) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 40, 41 and 42, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 45, 46 and 47, respectively. (e) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 50, 51 and 52, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 55, 56 and 57, respectively. (f) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 60, 61 and 62, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 65, 66 and 67, respectively. (g) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs:70, 71 and 72, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs:75, 76 and 77, respectively. [3] A heavy chain variable region (VH) comprising an amino acid sequence having at least 80% homology to an amino acid sequence selected from SEQ ID NO: 13, 23, 33, 43, 53, 63, and 73, and / or The antibody or fragment thereof according to embodiment 1 or 2, comprising a light chain variable region (VL) comprising an amino acid sequence having at least 80% homology to an amino acid sequence selected from SEQ ID NOs: 18, 28, 38, 48, 58, 68 and 78. [4] VH comprising an amino acid sequence selected from SEQ ID NO: 13, 23, 33, 43, 53, 63 and 73, and / or The antibody or fragment thereof of embodiment 3, comprising a VL comprising an amino acid sequence selected from SEQ ID NO:18, 28, 38, 48, 58, 68 and 78. [5] Any of the following (a) to (g) is included: An antibody or a fragment thereof comprising any one of (a) to (f) specifically binds to hCD98; The antibody or fragment thereof of embodiment 3, which specifically binds to mouse CD98, comprising (g) (a) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 13, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 18; (b) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 23, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 28; (c) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 33, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 38; (d) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 43, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 48; (e) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 53, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 58; (f) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 63, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 68; (g) a VH comprising or consisting of the amino acid sequence of SEQ ID NO:73, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO:78. [6] The antibody or fragment thereof according to any one of embodiments 1 to 5, wherein the antibody or fragment thereof has a pH-dependent binding to hCD98, and the binding activity at an acidic pH is higher than the binding activity at a neutral pH. [7] The antibody or fragment thereof of embodiment 6, wherein the EC50 of the antibody or fragment thereof at acidic pH is at least 2-fold, 3-fold, 5-fold, 10-fold, 15-fold, 20-fold, 30-fold, 40-fold, 50-fold lower than the EC50 at neutral pH when measuring binding of the antibody to hCD98, when the binding activity is measured by ELISA. [8] The antibody or fragment thereof according to embodiment 6 or 7, wherein the acidic pH is 5.8, 5.9, 6.0, 6.1, 6.2, 6.3, 6.4, 6.5, 6.6, 6.7 or 6.8, and the neutral pH is 7.0, 7.1, 7.2, 7.3, 7.4, 7.5 or 7.6, preferably, the acidic pH is 6.5 and the neutral pH is 7.4. [9] An isolated polynucleotide encoding the antibody or fragment thereof described in any one of embodiments 1 to 8.

[10] An expression vector comprising the isolated polynucleotide described in embodiment 9.

[11] A host cell comprising the expression vector described in embodiment 10.

[12] A composition comprising the antibody or fragment thereof described in any one of embodiments 1 to 8 and a pharma- ceutically acceptable carrier.

[13] A method for reducing tumors, inhibiting tumor cell proliferation, treating cancer, or preventing recurrence of cancer in a subject in need thereof, comprising administering to the subject a therapeutically effective amount of the antibody or fragment thereof described in any of embodiments 1 to 8, or the composition described in embodiment 12.

[14] The method of embodiment 13, wherein the tumor or cancer is a tumor or cancer that expresses CD98.

[15] The method of embodiment 13, wherein the tumor or cancer has a microenvironment that exhibits an acidic pH, e.g., pH 7.0, 6.9, 6.8, 6.7, 6.6, 6.5, 6.4, 6.3, 6.2, 6.1, 6.0 or lower, and normal physiological pH in the subject is about pH 7.4.

[16] The method of embodiment 13, wherein the tumor or cancer is selected from lymphoma, acute promyelocytic leukemia, hepatocellular carcinoma, pancreatic cancer, pancreatic epithelioid carcinoma, breast cancer, colorectal adenocarcinoma, cutaneous epidermoid carcinoma, melanoma, fibrosarcoma, non-small cell lung cancer, gastric cancer, acute myeloid leukemia, glioma, tongue cancer, hypopharyngeal squamous cell carcinoma, cholangiocarcinoma, osteomalacia, osteosarcoma, renal carcinoma, and neuroblastoma.

Claims

1. 1. An isolated antibody or antigen-binding fragment thereof that binds to human or mouse CD98, Any of the following (a) to (g): (a) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 10, 11 and 12, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 15, 16 and 17; (b) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 20, 21 and 22, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 25, 26 and 27; (c) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 30, 31 and 32, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 35, 36 and 37; (d) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 40, 41 and 42, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 45, 46 and 47; (e) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 50, 51 and 52, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 55, 56 and 57; (f) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 60, 61 and 62, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 65, 66 and 67, respectively; or (g) HCDR1, HCDR2 and HCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 70, 71 and 72, respectively, and LCDR1, LCDR2, LCDR3 comprising or consisting of the amino acid sequences of SEQ ID NOs: 75, 76 and 77; an antibody or fragment thereof comprising The antibody or fragment thereof comprising any one of (a) to (f) specifically binds to hCD98, The antibody or fragment thereof comprising (g) specifically binds to mouse CD98. , an isolated antibody or antigen-binding fragment thereof.

2. A heavy chain variable region (VH) comprising an amino acid sequence having at least 90% sequence identity to an amino acid sequence selected from SEQ ID NOs: 13, 23, 33, 43, 53, 63, and 73, and / or The antibody or fragment thereof according to claim 1, comprising a light chain variable region (VL) comprising an amino acid sequence having at least 90% sequence identity to an amino acid sequence selected from SEQ ID NOs: 18, 28, 38, 48, 58, 68 and 78.

3. Any of the following (a) to (g): (a) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 13, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 18; (b) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 23, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 28; (c) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 33, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 38; (d) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 43, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 48; (e) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 53, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 58; (f) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 63, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 68; or (g) a VH comprising or consisting of the amino acid sequence of SEQ ID NO: 73, and a VL comprising or consisting of the amino acid sequence of SEQ ID NO: 78; An antibody or fragment thereof comprising: The antibody or fragment thereof comprising any one of (a) to (f) specifically binds to hCD98, The antibody or fragment thereof comprising (g) specifically binds to mouse CD98. The antibody or fragment thereof described in claim 2.

4. The antibody or fragment thereof has a pH-dependent binding to hCD98, with the binding activity being higher at an acidic pH than at a neutral pH, and / or When the binding activity is measured by ELISA, the EC50 of the antibody or fragment thereof at an acidic pH is at least 2-fold, 3-fold, 5-fold, 10-fold, 15-fold, 20-fold, 30-fold, 40-fold, or 50-fold lower than the EC50 at a neutral pH when measuring the binding of the antibody to hCD98; The antibody or fragment thereof according to any one of claims 1 to 3, wherein the acidic pH is preferably 5.8, 5.9, 6.0, 6.1, 6.2, 6.3, 6.4, 6.5, 6.6, 6.7 or 6.8, and the neutral pH is 7.0, 7.1, 7.2, 7.3, 7.4, 7.5 or 7.6, and more preferably the acidic pH is 6.5 and the neutral pH is 7.

4.

5. An isolated polynucleotide encoding the antibody or fragment thereof according to any one of claims 1 to 4.

6. An expression vector comprising the isolated polynucleotide of claim 5.

7. A host cell comprising the expression vector of claim 6.

8. A composition comprising the antibody or fragment thereof according to any one of claims 1 to 4 and a pharma- ceutically acceptable carrier.

9. The antibody or fragment thereof according to any one of claims 1 to 4, or the composition according to claim 8, for use in reducing a tumor, inhibiting the proliferation of tumor cells, treating cancer, or preventing the recurrence of cancer in a subject in need thereof.

10. the tumor or cancer is a tumor or cancer that expresses CD98, 10. The antibody or fragment thereof or composition for use according to claim 9, wherein the tumor or cancer has a microenvironment that exhibits an acidic pH, such as pH 7.0, 6.9, 6.8, 6.7, 6.6, 6.5, 6.4, 6.3, 6.2, 6.1, 6.0 or lower, and wherein normal physiological pH in the subject is pH 7.

4.

11. The antibody or fragment thereof, or composition for use according to claim 9, wherein the tumor or cancer is selected from lymphoma, acute promyelocytic leukemia, hepatocellular carcinoma, pancreatic cancer, pancreatic epithelioid carcinoma, breast cancer, colorectal adenocarcinoma, skin epidermoid carcinoma, melanoma, fibrosarcoma, non-small cell lung cancer, gastric cancer, acute myeloid leukemia, glioma, tongue cancer, hypopharyngeal squamous cell carcinoma, cholangiocarcinoma, osteomalacia, osteosarcoma, renal carcinoma, and neuroblastoma.

Images