HIGH AFFINITY ANTICALINS DIRECTED AGAINST HUMAN CD98hc

20220153790 · 2022-05-19

    Inventors

    Cpc classification

    International classification

    Abstract

    The present invention relates to a cluster of differentiation 98 heavy chain (CD98hc)-specific binding protein, wherein the CD98hc-specific binding protein is a lipocalin 2 (Lcn2)-derived binding protein and binds to CD98hc with a K.sub.D of 200 nM or lower.

    Claims

    1. A cluster of differentiation 98 heavy chain (CD98hc)-specific binding protein, wherein the CD98hc-specific binding protein is a lipocalin 2 (Lcn2)-derived binding protein and binds to CD98hc with a K.sub.D of 200 nM or lower.

    2. The CD98hc-specific binding protein of claim 1, wherein the CD98hc-specific binding protein specifically binds to an epitope of CD98hc comprising one or more amino acids from amino acid positions 128 to 137 of SEQ ID NO: 1 and/or one or more amino acids from amino acid positions 374 to 404 of SEQ ID NO: 1.

    3. The CD98hc-specific binding protein of claim 1, wherein the CD98hc-specific binding protein (a) comprises or consists of an amino acid sequence as represented in formula I: TABLE-US-00011 QDSTSD(X1)(X2)PAPPLSKVPLQQNFQDNQF(Q/H)GKWY(X3)VG (X4)AG(X5)(X6)(X7)(X8)(X9)E(X10)(X11)(X12)(X13) (X14)(X15)M(X16)ATIYELKEDKS(Y/F)NVT(X17)V(X18) (X19)(X20)(X21)(K/T)KC(X22)(Y/N)(X23)(X24)(X25) (T/S)(X26)VPG(C/S)QPGE(F/Y)(T/N)(X27)G(N/K)I(X28) S(Y/R/G/A)P(X29)(X30)(X31)S(X32)L(X33)RVVSTNYNQ (H/Y)A(M/L)VF(X34)K(X35)(V/E)(X36)(X37)N(X38)E (X39)(X40)(X41)I(X42)L(X43)GRTKELTSELKE(N/I/Y)FIR FSKSLGLPE(X44)(X45)IVFPVPIDQCIDG, wherein (X.sub.1) is L; (X.sub.2) is I; (X.sub.3) is V; (X.sub.4) is R; (X.sub.5) is N; (X.sub.6) is L or T, preferably T; (X.sub.7) is G; (X.sub.8) is L; (X.sub.9) is R; (X.sub.10) is D; (X.sub.11) is K; (X.sub.12) is D; (X.sub.13) is P; (X.sub.14) is A or G, preferably G; (X.sub.15) is K; (X.sub.16) is F; (X.sub.17) is Y; (X.sub.18) is W; (X.sub.19) is F or S, preferably S; (X.sub.20) is D or G, preferably G; (X.sub.21) is L or Q, preferably Q; (X.sub.22) is K or M, preferably M; (X.sub.23) is S; (X.sub.24) is I (X.sub.25) is H, G or V, preferably V; (X.sub.26) is F; (X.sub.27) is L; (X.sub.28) is K; (X.sub.29) is G; (X.sub.30) is H, R or Q, preferably R; (X.sub.31) is T; (X.sub.32) is W; (X.sub.33) is V; (X.sub.34) is F; (X.sub.35) is W or S, preferably S; (X.sub.36) is G or T, preferably T; (X.sub.37) is Q; (X.sub.38) is R; (X.sub.39) is N, G or E, preferably G; (X.sub.40) is F; (X.sub.41) is A; (X.sub.42) is T; (X.sub.43) is Y; (X.sub.44) is N; and (X.sub.45) is H; (b) an amino acid sequence which is at least 80% identical to the amino acid sequence of (a), provided that amino acid positions (X.sub.1) to (X.sub.45) are maintained as defined in item (a); (c) an amino acid sequence which is at least 80% identical to the amino acid sequence of (a), provided that amino acid positions (X.sub.1) to (X.sub.45) are maintained as defined in item (a) or are changed by one or more conservative amino acid substitutions; or (d) an amino acid sequence which is at least 80% identical to the amino acid sequence of (a), provided that no more than ten amino acid positions of (X.sub.1) to (X.sub.45) as defined in item (a) are changed by amino acid substitutions, wherein the amino acid substitutions are preferably conservative amino acid substitutions.

    4. The CD98hc-specific binding protein of claim 1 or 2, wherein the CD98hc-specific binding protein (a) comprises or consists of the amino acid sequence of TABLE-US-00012 (SEQ ID NO: 14) QDSTSDLIPAPPLSKVPLQQNFQDNQFHGKWYVVGIAGNSMLREDKDPF KMTATIYELKEDKSYNVTRVRFDDKKCLYRILTFVPGSQPGEFTLGNIK SRPGVVTSWLVRVVSTNYNQHAMVFFKRVHQNRETFWITLYGRTKELTS ELKENFIRFSKSLGLPENHIVFPVPIDQCIDG  (b) an amino acid sequence which is at least 80% identical to the amino acid sequence of (a), provided that the underlined amino acids are maintained as defined in item (a); (c) an amino acid sequence which is at least 80% identical to the amino acid sequence of (a), provided that the underlined amino acids are maintained as defined in item (a) or are changed by one or more conservative amino acid substitutions; or (d) an amino acid sequence which is at least 80% identical to the amino acid sequence of (a), provided that no more than ten of the underlined amino acids as defined in item (a) are changed by amino acid substitutions, wherein the amino acid substitutions are preferably conservative amino acid substitutions.

    5. The CD98hc-specific binding protein of claim 1 or 2, wherein the CD98hc-specific binding protein comprises or consists of (a) the amino acid sequence of any one of SEQ ID NOs 2 to 14; or (b) an amino acid sequence which is at least 80% identical to the amino acid sequence of (a).

    6. A nucleic acid molecule encoding the CD98hc-specific binding protein of claim 1.

    7. A vector comprising the nucleic acid molecule of claim 6.

    8. A host cell transformed with the vector of claim 7.

    9. A method for the production of a CD98hc-specific binding protein comprising culturing the host cell of claim 8 under suitable conditions and isolating the CD98hc-specific binding protein produced.

    10. A protein conjugate or fusion protein comprising the CD98hc-specific binding protein of claim 1.

    11. The protein conjugate or fusion protein of claim 10, wherein the CD98hc-specific binding protein is conjugated to or is part of a fusion protein wherein the fusion partner is (a) a fluorescent dye or a fluorescent protein, (b) a radionuclide, (c) a toxic compound, (d) a photosensitizer, (e) an enzyme or truncated version thereof, (f) a membrane protein or functional fragment thereof retaining the enzyme function, (g) a contrast agent, (h) a cytokine, (i) a chemokine, (j) a pro-coagulant factor, (k) an acetylcholineesterase inhibitor, (l) an inhibitor of Aβ aggregation, (m) a nucleic acid molecule, or (n) a nanoparticle.

    12. The protein conjugate or fusion protein of claim 10, wherein the CD98hc-specific binding protein is conjugated to a binding protein, wherein the binding protein is preferably selected from an antibody or immunoglobulin, an antibody fragment and an antibody mimetic, wherein the antibody mimetic is preferably selected from the group consisting of an Anticalin, Affibody, Adnectin, DARPin, Avimer, Nanofitin, Affilin, β-Wrapin, ADAPT, Monobody, Rasln, FingR, Pronectin, Centyrin, Affimer, Adhiron, Affitin, αRep, Repebody, i-body, Fynomer and a Kunitz domain protein.

    13. A pharmaceutical composition or a diagnostic composition comprising at least one of (i) the CD98hc-specific binding protein of claim 1; (ii) a nucleic acid molecule encoding the CD98hc-specific binding protein of (i); (iii) a vector comprising the nucleic acid molecule of (ii); (iv) a host cell transformed with the vector of (iii), and/or (v) a protein conjugate or fusion protein comprising the CD98hc-specific binding protein of (i).

    14. The CD98hc-specific binding protein of claim 1 for use in therapy and/or diagnosis, preferably for use in treating or diagnosing a disease in the brain and/or the spinal cord.

    15. The CD98hc-specific binding protein of claim 1 for use in the therapy and/or diagnosis of a tumor or a neurological disease, wherein the tumor or neurological disease is preferably located in the brain and/or the spinal cord.

    16. The nucleic acid molecule of claim 6, a vector comprising the nucleic acid molecule, or a host cell transformed with the vector for use in therapy and/or diagnosis, preferably for use in treating or diagnosing a disease in the brain and/or the spinal cord.

    17. The nucleic acid molecule of claim 6, a vector comprising the nucleic acid molecule, or a host cell transformed with the vector for use in the therapy and/or diagnosis of a tumor or a neurological disease, wherein the tumor or neurological disease is preferably located in the brain and/or the spinal cord.

    18. The protein conjugate or fusion protein of claim 10 for use in therapy and/or diagnosis, preferably for use in treating or diagnosing a disease in the brain and/or the spinal cord.

    19. The protein conjugate or fusion protein of claim 10 for use in the therapy and/or diagnosis of a tumor or a neurological disease, wherein the tumor or neurological disease is preferably located in the brain and/or the spinal cord.

    Description

    [0254] The figures show.

    [0255] FIG. 1. Phage Display selection of lipocalin variants with affinity towards the ectodomain of hCD98hc. (A) Graphical depiction of the CD98 heavy chain (CD98hc) covalently linked to a CD98 light chain, illustrating the two main biochemical functions fulfilled by CD98hc. The structure of CD98hc ectodomain used for phage display selection of Anticalins is shown as a molecular surface model. (B) Analytical size exclusion chromatography of the CD98hc-specific lipocalin variants P1E4, P3A12 and P3D11 selected via phage display and wtLcn2, showing a monomeric elution behavior with just minor aggregate formation (eluting at Vo). (C) SPR real-time binding analysis of the picomolar affinity variant P3D11 for hCD98hcED. The deduced kinetic constants are listed in Table 1.

    [0256] FIG. 2. X-ray structure of Anticalin P3D11 in complex with the human CD98hc ectodomain. (A) Cartoon representation of the P3D11.circle-solid.hCD98hcED complex (P3D11 dark gray, hCD98hcED light gray). The epitope loops L1 and L2 of hCD98hc are highlighted in dark gray and black (both in bold), respectively. (B) Dissection of the contact interface. Both complex components are shown as molecular surface and rotated by 90° in opposite directions to visualize their interface. P3D11 interface residues are colored light grey, in order to indicate contacts with loops L1 and L2 of hCD98hc (see panel A). Contacting residues of hCD98hc with P3D11 are colored dark gray. Hydrogen bond donors and acceptors are indicated in black for both molecules. Water molecules that mediate hydrogen bonds are depicted as dark spheres. (C) Surface representation of P3D11.circle-solid.hCD98hcED complex in context of its four complex N-glycans (spheres), rotated by 110° with regard to panel A.

    [0257] FIG. 3. Affinity and stability engineering of the lipocalin variant P3D11 via bacterial cell surface display. (A) FACS analysis of E. coli cells presenting wtLcn2 (left), starting variant P3D11 (middle) or the error-prone library (right) incubated with 100 nM (round 1, upper panel) or 1 nM (round 7, lower panel) biotinylated BAP-huCD98hcED, respectively. (B) Single clone FACS analysis of wtLcn2, P3D11 and an identified variant depicted D11.1 after seven selection cycles using 1 nM BAP-huCD98hcED (FIG. 11). (C) Thermal denaturation of P3D11 and D11vs using CD spectroscopy.

    [0258] FIG. 4. Cytofluorometry and immunofluorescence microscopy of CD98hc expressing human cancer cell lines using the PASylated lipocalin variant D11vs. (A) Flow cytometry analysis of three B-cell lymphoma (SU-DHL-4, Raji and Ramos) and two prostate carcinoma (DU-145 and PC-3) cell lines with Cy5.5 labeled PASylated D11vs (gray60 histogram). For competition of D11vs binding, a 10-fold molar excess of either unlabeled D11vs-PAS200 or soluble hCD98hcEDg was used (gray80 and gray50 histograms, respectively). The respective cell line without the addition of a labeled binding protein is shown in gray70. Immunocytochemical detection of hCD98hc using Cy5.5 labeled PASylated D11vs is shown for (B) Ramos, (C) PC-3 and (D) Caco-2 cells (middle panel). As controls for specific recognition of hCD98hcED by D11vs, cells were stained with wtLcn2 (left panel) or binding was competed using a 10-fold molar excess of unlabeled D11vs-PAS200 (right panel). Cell nuclei were stained with DAPI.

    [0259] FIG. 5. In vivo .sup.89Zr PET/CT imaging study in a prostate carcinoma xenograft model. (A) Mice (♀) bearing PC-3 xenografts were i.v. injected with 2.85±0.15 MBq .sup.89Zr.circle-solid.D11vs-PAS200-DFO followed by PET/CT imaging 24, 48 and 72 h post injection. Signals were detected in the xenograft tumor (arrowheads), liver (I), kidneys (k), bladder (bl) and the joints (*). (B) PET images were analyzed by threshold-based image segmentation (lower threshold at 50% of the hottest voxel), activity trajectories are depicted. Error bars indicate standard deviation and an unpaired student's t-test was performed between the blocked and non-blocked group (* p<0.05). Epitope blocking was achieved by i.v. injection of a 100-fold molar excess of the cold PASylated Anticalin 2 h before .sup.89Zr-tracer injection. (C) Explanted tumors were split in transversal direction, from one half consecutive cryosections of 10 μm were prepared followed by autoradiography and HE staining. (D) Second half of the tumor was embedded in paraffin and sections were stained for HE, CD31 (blood vessels) and CD98hc using appropriate antibodies.

    [0260] FIG. 6. In vivo and ex vivo PET/CT imaging and biodistribution analysis. (A) Mice (♂) bearing PC-3 tumors were injected with 3.96±0.12 MBq .sup.89Zr.circle-solid.D11vs-PAS200 and (B) mice in the blocking group additionally received 250-fold molar excess of unlabeled D11vs-PAS200 2 h before the radiotracer injection. For each mouse in each cohort the in vivo PET/CT (left panel), ex vivo PET/CT (middle panel) and a photograph of the tumors (right panel) are depicted. (C) For biodistribution analysis mice organs were explanted and subsequently weight and radioactivity were determined. Error bars show standard deviation and an unpaired student's t-test was performed between the blocked and unblocked group (** p<0.005).

    [0261] FIG. 7. Biochemical characterization of the monobiotinylated human and murine CD98hc ectodomain soluble produced in E. coli or HEK cells. (A) Schematic illustration of the covalently linked CD98hc/CD98lc heterodimer and plasmid constructs used for the soluble expression of unglycosylated and glycosylated m/hCD98hc ectodomain. The extracellular domain of CD98hc soluble expressed and used for Anticalin selection is framed by a rectangle (B) Apparent mass determination of the glycosylated and unglycosylated m/hCD98hcED using analytical SEC (Table 2) verifying a considerable mass increase due to glycosylation. (C) Differences in electrophoretic mobility of unglycosylated and glycosylated m/hCD98hcED during SDS-PAGE, confirming glycosylation after production in eukaryotic HEK cells, while absent after production in E. coli. (D) Enzymatic processing of N-linked sugars from m/hCD98hcEDg by Peptide-N-Glycosidase F (PNGase F) and subsequent comparison with the fully glycosylated ectodomain via SDS-PAGE.

    [0262] FIG. 8. Sequence analysis, expression and functional characterization of the selected lipocalin variants P1E4, P3A12 and P3D11. (A) Amino acid sequence alignment of the selected lipocalin variants compared to wtLcn2. The central randomized gene cassette flanked by a pair of BstXI sites is underlined, β-stands and structurally hypervariable loops are labeled with letters A-H and numbers 1-4, respectively. (B) Coomassie-stained SDS-PAGE of the recombinant lipocalin variants after production in E. coli. The increased electrophoretic mobility under non-reducing conditions confirms formation of the single structural disulfide bridge in the Lcn2 scaffold. (C) Biomolecular interaction analysis between the immobilized hCD98hc ectodomain produced in E. coli (ARU=225) and the selected lipocalin variants P3A12 and P1E4 via SPR measurement. (D) Competitive binding analysis between the lipocalin variants P3D11 and P3A12 by SPR measurement. Binding sites on hCD98hcEDg were saturated with an injection of 100 nM P3D11 followed by injection of 100 nM P3A12, which did not cause an additive RU signal (black). Injection of 100 nM P3A12 without prior blocking of the hCD98hcED epitope with P3D11 evokes an expected binding signal for P3A12 (grey). Two consecutive buffer injections are shown as a negative control (light grey).

    [0263] FIG. 9. Crystallization of the P3D11.circle-solid.hCD98hcED complex. (A) SEC purification of the P3D11.circle-solid.hCD98hcED complex. (B) Analytical SDS-PAGE of the purified P3D11.circle-solid.hCD98hcED complex with and without the addition of 2-mercaptoethanol. (C) Diffraction quality crystals obtained by vapor diffusion in hanging drops at 20° C. in 18% (w/v) PEG3350 and 100 mM Na-malonate pH 4.75. (D) Diffraction pattern obtained for the large crystal in (B).

    [0264] FIG. 10. Conformational differences of the individual complex components. (A) Conformational changes of hCD98hcED loop L1 (dark gray) and L2 (black) upon complex formation with P3D11. (B) Structural superposition of P3D11 with wild-type Lcn2 (PDB code 1L6M), Anticalin N7E (PDB code 5N47) and Anticalin PRS #003 (PDB code 3BX7), using the 58 conserved Cα positions that are conserved among different lipocalins. The structurally variable loops are highlighted for P3D11, wild-type Lcn2, N7E and PRS #003, respectively.

    [0265] FIG. 11. Engineering of lipocalin variant P3D11 via error prone PCR and bacterial cell surface display. (A) Schematic depiction of surface displayed Lcn2 variants, which are presented by fusion to the engineered β-domain of auto-transporter EspP. (B) Single clone FACS analysis after six selection cycles of BCSD in comparison to Lcn2 and the starting variant P3D11 using 1 nM biotinylated hCD98hcED. Additionally, mean intensity of PE fluorescence is shown for the unique lipocalin variants identified during stability and affinity engineering of P3D11. (C) Amino acid sequence alignment of the identified lipocalin variants compared to Lcn2 and the starting variant P3D11. The lipocalin variant D11vs was generated based on the sequence information provided by the selected variants. (D) Thermal stability testing of lipocalin variant P3D11 and its engineered version D11vs incubated at 37° C. for the indicated time points followed by SDS-PAGE analysis. (E) SPR real-time binding analysis of the stability and affinity improved variant D11vs for hCD98hcED. The deduced kinetic constants are listed in Table 1.

    [0266] FIG. 12. Structural context of side chain substitutions that increase stability and/or affinity of P3D11. Residues that differ from wild-type Lcn2 are labeled with asterisks. (A) Hydrophobic cavity (surface illustration) lined by residues Tyr68, Gly81, Phe83 and Leu94 (upper panel). Substitution of Gly in position 81 of the Anticalin (translucent surface) with Val fills the cavity (lower panel). (B) In addition, Val81 increases the contact interface with hCD98hc (solid surface). (C) Substitution of Phe71 by Ser disrupts the aromatic stacking between Phe71 and Gln174, but permits hydrogen bond formation of Ser71 with Lys50 and Gln174.

    [0267] FIG. 13. In vivo .sup.89Zr PET/CT imaging study in a B-cell Non-Hodgkin's lymphoma xenograft model. Mouse (♀) bearing Ramos xenograft was i.v. injected with 2.9 MBq .sup.89Zr.circle-solid.D11vs-PAS200-DFO followed by PET/CT imaging 24, 48 and 72 h post injection. Signals were detected in the xenograft tumor (arrowheads), liver (I), kidneys (k), bladder (bl) and the joints (*).

    [0268] FIG. 14. Immunohistochemical characterization of PC-3 and Ramos tumor sections. Tissue sections of PC-3 (A) and Ramos (B) tumors stained with hematoxylin and eosin (H&E), CD31 (blood vessels) and CD98hc using appropriate antibodies. Order of magnification is indicated.

    [0269] FIG. 15. .sup.18F-fludeoxyglucose (FDG) and .sup.18F-fluoroethyl-L-tyrosine (FET) PET/CT imaging. Mice (♀) bearing Ramos/PC-3 xenografts were i.v. injected with (A) 12.3/12.7 MBq .sup.18F-FDG or (B) 11.8/12.1 MBq .sup.18F-FET followed by PET/CT imaging 45 min post injection. Signals were detected in the xenograft tumor (arrowheads), brain (br), heart (h), liver (I), kidneys (k), intestine (i) and the bladder (bl).

    [0270] FIG. 16. Contact area of P3D11 with CD98hc. (A) Residues of the lipocalin variant P3D11 that interact with hCD98hcED according to PISA analysis. Residues with contact area ≥Å.sup.2 are highlighted in gray, whereas randomized positions are indicated with asterisks. (B) Summary of contact area and contact types of the lipocalin variant P3D11.

    [0271] FIG. 17. Molecular design and characterization of PASylated D11vs. (A) Schematic depiction of Anticalin D11vs with a C-terminally attached PAS200 polypeptide, including an engineered Cys residue at the end. (B) Non-reducing SDS-PAGE of the PASylated D11vs variants used for in vitro and in vivo studies. For the D11vs-PAS200-Cy5.5 conjugate (rightmost lane), fluorescence was detected on a fluorescence scanner. (C and D) ESI-MS measurement (deconvoluted, raw data are shown as inset) for D11vs-PAS200-Cys after coupling to Sulfo-Cyanine5.5 (C) or Dfo (D) via maleimide chemistry, confirming a homogeneous site-specific labelling (Table 7).

    [0272]

    TABLE-US-00004 TABLE 7 Protein mass determination of PASylated D11vs variants via ESI MS Mass Theoretical Measured difference to Protein mass [Da] mass [Da] uncoupled [Da] D11vs-PAS200 37933.5 37933.6 (Δ0.1) — D11vs-PAS200-Cys 38036.7 n.d. — D11vs-PAS200-DFO 38748.5 39749.4 (Δ0.9) 712.7 DFO: 711.8 D11vs-PAS200-Cy5.5 39060.9 39062.1 (Δ1.2) 1025.4 S-Cy5.5: 1024.3

    [0273] The examples illustrate the invention.

    EXAMPLE 1—MATERIAL AND METHODS

    [0274] Phage display selection, identification and production of hCD98hcED-specific lipocalin variants Selection of CD98hc-specific lipocalin variants was performed via filamentous phagemid display from a combinatorial library based on Lcn2 randomized at 20 positions within the structurally variable loops comprising 1×10.sup.10 variants [50]. Biotinylated recombinant hCD98hcED (cf. Supplementary Information) was immobilized on streptavidin- or NeutrAvidin-coated paramagnetic beads (Sigma-Aldrich, Munich, Germany and Thermo Fisher Scientific, Waltham, Mass., respectively), incubated with the phagemid library (starting titer 1×10.sup.12), and bound phagemids were eluted under denaturing conditions using 4 M urea. After five consecutive cycles, pooled phasmid DNA from the enriched library was prepared and subcloned on pNGAL98 in order to perform high-throughput ELISA screening from microcultures expressing the soluble lipocalin variants, as previously described [50]. Briefly, periplasmic extract was prepared and applied to 96-well MaxiSorp plates (Thermo Fisher Scientific) coated with 250 nM of hCD98hcED. After 1 h incubation bound lipocalin variants were detected by means of their C-terminal Strep-tag II using an ExtrAvidin/alkaline phosphatase (AP) conjugate (Sigma-Aldrich). Signals were developed using 0.5 mg/ml p-nitrophenyl phosphate in AP buffer (0.1 M NaCl, 5 mM MgCl2, 0.1 M Tris/HCl, pH 8.8) and absorbance measured at 405 nm with an Infinite 200 PRO microplate reader (Tecan, Mannerdorf, Switzerland). For clones showing significant binding towards hCD98hcED the expression cassette on the plasmid DNA was sequenced, and the soluble lipocalin variant was produced in a preparative scale as soluble, secretory protein in E. coli JM83 with a C-terminal His.sub.6-tag using the plasmid pNGAL118 [50]. After periplasmic protein extraction the recombinant proteins were purified by immobilized metal ion affinity chromatography (IMAC) using a Ni(II)-charged HisTrap HP column (GE Healthcare, Munich, Germany) and subsequently subjected to size-exclusion chromatography (SEC) in PBS (4 mM KH.sub.2PO.sub.4, 160 mM Na.sub.2HPO.sub.4, 115 mM NaCl pH 7.4) on a 24 ml Superdex 75 10/300 GL column (GE Healthcare). Furthermore, plasmid construct design and production of recombinant D11vs-PAS200, D11vs-PAS200-Cys and Lcn2-PAS200-Cys was performed as previously described [51] and purified as described above.

    [0275] Protein Crystallization and Structure Determination

    [0276] For protein crystallization a variant of hCD98hcED with N-terminal Strep-tag II was produced in E. coli stain BL21 as described previously (Deuschle et al. submitted) (FIG. 12A). After incubation with the purified lipocalin variant P3D11 in a 1:1 molar ratio for 1 h at 4° C., the P3D11.circle-solid.hCD98hcED complex was isolated via SEC on a Superdex 200 10/300 GL column (GE Healthcare) and directly eluted in crystallization buffer comprising 10 mM Hepes/NaOH pH 7.5, 100 mM NaCl and 0.02% (w/v) NaN.sub.3. The complex was concentrated to 15.7 mg/ml using a 30 kDa MWCO Amicon centrifugational filter (Merck Millipore, Burlington, Mass.) and subjected to crystallization by vapor diffusion in hanging drops at 20° C. Diffraction quality crystals were obtained by mixing 1 μl of the P3D11.circle-solid.hCD98hcED solution with 1 μl of reservoir solution containing 18% (w/v) PEG3350 and 100 mM Na-malonate pH 4.75. Suitable crystals were transferred into cryoprotectant consisting of 19% (w/v) PEG3350, 100 mM Na-malonate pH 5.0 and 20% (v/v) ethylene glycol prior to flash cooling in liquid nitrogen. X-ray diffraction data were collected at the Helmholtz-Zentrum Berlin, BESSY beamline 14.2 [52] and reduced with the XDS package [53] (Table 4).

    TABLE-US-00005 TABLE 4 Data collection and refinement statistics Data collection Space group C2 Unit cell parameters a = 202.81 Å, b = 46.05 Å, c = 137.09 Å, α = γ = 90°, β = 106.64° Wavelength [Å] 0.9184 Resolution Å 30.0 − 1.80 (1.90 − 1.80) Completeness [%] 99.8 (99.9) Unique reflections 113253 (16844) Multiplicity 6.8 (7.0) Mean I/σ (I) 23.9 (2.3) R.sub.meas [%] 4.9 (93.4) Wilson B-factor [Å] 37.4 Refinement Resolution [Å] 30.0 − 1.80 (1.85 − 1.80) Reflections (working) 111018 (8149) Reflections (test).sup.b 2234 (187) R.sub.cryst [%] 18.4 (60.8) R.sub.free [%] 22.6 (67.0) Protein molecules per au 4 Number of atoms: protein/solvent.sup.c 9356/909 B-values of atoms: protein/solvent  33.3/39.2 [Å.sup.2] Ramachandran plot.sup.d: favored/outliers 96.8/0.2 [%] RMSD bonds [Å]/angles [°]  0.02/1.76 .sup.aValues in parentheses refer to the highest resolution shell. .sup.bTest set corresponds to 2% of all reflections .sup.cSolvent refers to waters, ions as well as ordered buffer or cryoprotectant molecules .sup.dRamachandran statistics were calculated with MolProbity

    [0277] The crystal structure was solved by molecular replacement with Phaser [54] using coordinates of the hCD98hcED (PDB entry 2DH2) [55] and of the Anticalin N7A (PDB entry 4GH7) [50] as search models. Manual rebuilding and refinement were done with Coot [56] and Refmac5 [57], respectively (Table 4). Translation, libration and screw (TLS) groups were determined with TLSMD [58]. The asymmetric unit of space group C2 contained two P3D11.circle-solid.hCD98hcED complexes, of which the complex comprising chain pairs A and B, with overall lower B-factors, was used for analysis.

    [0278] Biomolecular Interaction Analysis Via Surface Plasmon Resonance (SPR) Spectroscopy

    [0279] Real time surface plasmon resonance (SPR) spectroscopy was performed on a BIAcore 2000 system (BIAcore, Uppsala, Sweden) at 25° C. using HBS-T (20 mM Hepes/NaOH pH 7.5, 150 mM NaCl, 0.005% v/v Tween20) as running buffer. The purified biotinylated m/hCD98hcED (3 μg/ml protein solution in HBS-T produced in HEK cells or E. coli) was immobilized (ARU-225) via streptavidin on a CAP sensorchip of the Biotin CAPture kit (GE Healthcare). SPR single cycle kinetic experiments were performed using five consecutive injections of a 1:2 dilution series of the purified lipocalin variant at a flow rate of 25 μl/min, with 288 s contact time and 3500 s dissociation time after the fifth injection. For multi cycle kinetic experiments, eight samples from a consecutive 1:2 dilution series of the purified lipocalin variant were measured. Rate constants of association and dissociation were calculated from reference-corrected sensorgrams by fitting to a global 1:1 Langmuir binding model using BIAevaluation software (BIAcore). The equilibrium dissociation constants (K.sub.D value) were calculated as the quotient k.sub.off/k.sub.on. To test competitive binding of the lipocalin variants P3D11, P3A12 and P1E4 to glycosylated hCD98hcED, 100 nM of P3D11 was first injected and followed by a subsequent injection of either 100 nM P3A12 or 100 nM P1E4. HBS-T injections served as negative control.

    [0280] Error-Prone Library Construction and Bacterial Cell Surface Display Selection

    [0281] Stability and affinity engineering of the lipocalin variant P3D11 was accomplished by means of directed evolution and bacterial surface display (BSD) as previously published [51, 59, 60]. Briefly, error-prone PCR was executed using the GeneMorph II random mutagenesis kit (Agilent, Santa Clara, Calif.) with 10 μg of the central coding region for the lipocalin variant P3D11 used as template. After digest of the PCR product with BstXI, the randomized DNA fragment was ligated with the backbone of pNGAL146 and used for electrotransformation of E. coli JK321, yielding 5×10.sup.9 transformants plated on LB agar medium supplemented with 100 μg/ml ampicillin. The bacterial lawn was scraped from the plate(s), resuspended in 50 ml LB/Amp medium, followed by inoculation to an initial OD.sub.550 of 0.15 at 37° C. Gene expression was induced at OD.sub.550=0.5 with 10 ng/ml anhydrotetracycline (aTc) for 2.5 h. Approximately 2×10.sup.8 cells were sedimented and resuspended in PBS with biotinylated hCD98hcED at different concentrations (cycles 1 and 2: 100 nM; cycles 3 and 4: 10 nM; cycles 5 and 6: 1 nM) for 1 h at 4° C. After one washing step with PBS, the bacteria were incubated with 25 μg/ml streptavidin/phycoerythrin (PE) conjugate (Biolegend, San Diego, Calif.) and 3 μM dye-labeled Fab A3C5-DY634 [61] in PBS for 30 min followed by a last washing step in PBS. Bacteria were sorted on a FACSAria Ilu instrument (BD Bioscience, Heidelberg, Germany) using a 488 nm LASER diode with a 585/42 band pass filter or a 633 nm HeNe LASER in combination with a 660/20 band pass filter for excitation/detection of PE and DY634 fluorescence, respectively. Sorted bacteria were plated on LB/Amp agar, incubated at 37° C. for 12 h and then subjected to a new sorting cycle. After BSD cycle 6, single clone analysis was performed by cytofluorimetric analysis of individual colonies. Data were processed using FlowJo v10 software (FlowJo, Ashland, Oreg.).

    [0282] Flow Cytofluorimetry

    [0283] The human cancer cell lines Ramos, Raji, SU-DHL-4, PC3 and DU-145 were cultivated in Roswell Park Memorial Institute (RPMI) 1640 medium containing 2.0 g/L NaHCO.sub.3 and stable L-glutamine (Biochrom, Berlin, Germany), supplemented with 10% (v/v) fetal bovine serum (FBS) (PAA Laboratories, Pasching, Austria) at 37° C. under humidified 5% CO.sub.2 atmosphere. Adherent cells (PC3, DU-145 and Caco-2) were washed with PBS (w/o Ca.sup.2+ and Mg.sup.2+) (Biochrom) and detached using PBS (w/o Ca.sup.2+ and Mg.sup.2+) supplemented with 0.5 mM EDTA for 10 min at 37° C. Cells growing in suspension (Ramos, Raji and SU-DHL-4) were directly washed with PBS (w/o Ca.sup.2+ and Mg.sup.2+). For each measurement, 250,000 viable cells were resuspended in 250 μl FACS buffer (PBS+10% (v/v) FBS) containing 1 μM D11vs-PAS200-Cy5.5 (cf. Supplementary information) and incubated for 1 h at 4° C. Competition experiments were performed by adding a 10-fold molar concentration of either unlabeled D11vs-PAS200 or the soluble glycosylated hCD98hcED. After incubation, cells were washed three times in PBS and finally resuspended in 250 μl FACS buffer. Flow cytofluorometric experiments were performed on a FACSAria Ilu instrument using an excitation wavelength of 650 nm and an emission band-path filter of 780/60 nm.

    [0284] Immunofluorescence Microscopy of CD98hc Expressing Cells

    [0285] PC-3 and Caco-2 cells were cultivated at 37° C. on poly-D-lysine (PDL) coated Lab-Tek II chamber slides (Thermo Fisher Scientific, Waltham, Mass.) in RPMI 1640 medium containing 2.0 g/L NaHCO.sub.3 and stable L-glutamine, supplemented with 10 (v/v) % FBS, until 60-70 confluence was reached. After three times washing with PBS, the cells were incubated with 1 μM D11vs-PAS200-Cy5.5 for 1 h at 37° C., followed by another three washing steps. Ramos cells, on the other hand, were grown in suspension and stained in the same manner using a reaction tube, followed by transfer to a PDL coated Lab-Tek II chamber slide. For control experiments, the D11vs-PAS200-Cy5.5 conjugate was premixed with a 10-fold molar concentration of unlabeled D11vs-PAS200 or the equivalent conjugate of wtLcn2 (Lcn2-PAS200-Cy5.5) was applied. Cells were fixed and counterstained by 5 min incubation with ice-cold methanol containing 1 μg/ml 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich). Digital fluorescence images were recorded on an Axiovert 40 CFL microscope equipped with an AxioCam MRm camera (Carl Zeiss Microscopy, Jena, Germany) using 365/12 nm (DAPI) and 716/40 nm (Cy5.5) band-pass filters with identical exposure times.

    [0286] Circular Dichroism (CD) Spectroscopy and Thermal Denaturation

    [0287] CD spectra and thermal unfolding of purified Lcn2 variants were recorded using a Jasco J-810 spectropolarimeter (Jasco, Pfungstadt, Germany) controlled by Spectra manager software (ver. 1.53.05) equipped with a PT-423S Peltier element. Proteins were dialyzed against 20 mM KP.sub.i pH 7.5 and 50 mM K.sub.2SO.sub.4 and applied at a concentration of 1 μM. In order to identify the wavelength with maximum change in CD signal for thermal denaturation studies, spectra were measured in a 1 mm path length quartz cuvette (Hellma, Müllheim, Germany) from 190-250 nM at 20° C. and 90° C. Thermal unfolding of the lipocalin variants was measured at a wavelength of 214 nm by heating from 20° C. to 90° C. at a rate of 60° C./h. Data were fitted to an equation for a one-step unfolding transition, and the melting temperature (T.sub.m), enthalpy of unfolding (ΔH.sub.m) and Gibb's free energy of unfolding (ΔG.sup.0) at ambient temperature was calculated as previously described [62].

    [0288] In Vivo Xenograft Studies

    [0289] Animal experiments were conducted with permission from the District Government of Upper Bavaria (application no.: 55.2-1-54-2532-216-15). CB17-SCID mice at an age of 6 weeks (♂) and (♀) were purchased from Charles River Laboratories (Wilmington, Mass.) and housed under specific pathogen free (SPF) conditions. Mice were injected subcutaneously with 100 μl RPMI 1640 medium (without supplements) containing 6×10.sup.6 PC-3 cells above the right shoulder. On day 12 p.i. the tumor had reached a diameter of ˜0.5 cm and the mice were used for in vivo studies.

    [0290] .sup.89Zr Positron Emission Tomography

    [0291] For .sup.89Zr-labeling of Dfo-conjugated [63] D11vs-PAS200 (for protein production and coupling cf. Supplementary information), 250 μg of the protein dialyzed against 250 mM Na-acetate pH 5.5 was incubated with 200 μl 0.5 M Hepes/NaOH pH 7.0, 50 μl 0.5 M gentistic acid and 111 MBq .sup.89Zr in 1 M oxalic acid (neutralized with 3 M Na.sub.2CO.sub.3) (Perkin Elmer, Waltham, Mass.). After 1 h incubation at 37° C., radiolabeling was confirmed by radio thin layer chromatography (radio-TLC) on strips using Na-citrate pH 5.0 as mobile phase. The radio-labeled protein tracer was separated by gel filtration using a PD-10 column (GE Healthcare) equilibrated with 0.9% NaCl solution (B. Braun, Melsungen, Germany). Radiochemical purity of the protein tracer was confirmed by radio-TLC prior to its use for injections.

    [0292] For the biodistribution study, mice were injected with the protein tracer via the tail vein at a dose of 3.96±0.12 MBq. For blocking experiments, a 100-fold (FIG. 5) or 250-fold (FIG. 6) molar amount of D11vs-PAS200 was injected 2 h prior to the tracer injection. Static PET/CT imaging was performed at indicated time points under isoflurane anesthesia in an Inveon PET/CT small-animal scanner (Siemens Medical Solutions, Erlangen, Germany) with 20 min PET and 5 min CT acquisition. Data were analyzed with the Inveon Research Workplace software (Siemens Medical Solutions, Erlangen, Germany) and reconstructed using the 0.8 mm high resolution OSEM-3D algorithm. Quantification of regions of interest (ROI) was performed by threshold-based image segmentation with a lower threshold at 50% of the hottest voxel for in vivo PET and a threshold of 10 kBq/ml for ex vivo PET images.

    [0293] Biodistribution Analysis and Autoradiography

    [0294] Tumor, blood and organs were dissected from the sacrificed animals at the end of the experiment in order to determine weight and quantify radioactivity using a 2480 Wizard.sup.2 automatic gamma counter (PerkinElmer). A 1% fraction of the injected protein tracer dose was used as reference. Uptake values were calculated as % ID/g and corrected for radioactive decay from the time point of injection.

    [0295] After radioactivity measurement, the tumor was transversally bisected and used for immunohistochemistry and autoradiography analysis. For autoradiography the tumor was embedded in Tissue-Tek O.C.T. (Sakura Finetek, Alphen aan den Rijn, The Netherlands) in an embedding mold and frozen. Subsequently, 10 μm sections were prepared using a CM1950 cryostat (Leica Biosystems, Nußloch, Germany) and the sections were mounted on a superfrost microscopic slide (Thermo Fisher Scientific). The slide was exposed, together with a standard dilution row in silicon isolators (Grace Bio-Labs, Bend, Oreg.), to a storage phosphor screen BAS-IP super resolution film (GE Healthcare) for one week, followed by quantification using a CR 25 BIO autoradiography scanner (Durr Medical, Bietigheim-Bissingen, Germany) and data analysis with AIDA software ver. 4.24.036 (Raytest, Straubenhardt, Germany). Chosen pictures show the sections with the highest exposure level to allow comparison of blocked and unblocked tumor sections.

    [0296] Immunohistochemistry

    [0297] Tumor tissue was fixed for 48 h in 4% paraformaldehyde in PBS and stored in PBS at 4° C. until radioactivity decayed. Tissue was dehydrated under standard conditions (ASP300S; Leica Biosystems) and embedded in paraffin. Serial 2-μm-thin sections prepared with a rotary microtome (HM355S; Thermo Fisher Scientific) were collected and subjected to histological and immunohistochemical analysis. Hematoxylin-eosin staining was performed on deparaffinized sections.

    [0298] Immunohistochemistry was done under standard conditions using appropriate antibodies (AT-2; Leica Biosystems). Representative images were collected using Aperio Imagescope software (version 12.3; Leica Biosystems).

    [0299] Statistics

    [0300] For mean comparison between two groups of animals, the Student's t test for unpaired data was used. P values 0.05 were considered statistically significant (* P≤0.05; ** P≤0.005). Experimental data were analyzed using Prism 6 (GraphPad Software, San Diego, Calif.) and standard deviations (S.D.) are indicated.

    [0301] Soluble Production and Purification of the Monobiotinylated Human and Murine CD98hc Ectodomains

    [0302] The human and murine CD98hcED (UniProt ID P10852-1, residues Glu105-Ala526 and UniProt ID P08195-2, residues Glu111-Ala529, respectively) were produced in E. coli strain BL21 using pASK-IBA5(+)-BAP-m/hCD98hcED and in human embryonic kidney (HEK) 293E cells (MEXi expression system, IBA Lifesciences, Goettingen, Germany) using the expression plasmid pDSG-BM-His.sub.6-BAP-m/hCD98hcEDg-Igk-BirA-StrepDEL (StrepDEL=Fusion of the Strep-tag II amino acid sequence WSHPQFEK and the endoplasmic reticulum retention signal sequence KDEL).

    [0303] For expression in E. coli, BL21 was co-transformed with pBirAcm encoding biotin ligase (Avidity LLC, Aurora, Colo.), and heterologous gene expression was induced with 0.5 mg/L anhydrotetracycline (aTc) and 1 mM isopropyl-f3-D-thiogalactopyranoside (IPTG) for 12 h at 26° C. in 2 L 2xYT medium. Cells were harvested by centrifugation, resuspended in anion-exchange chromatography (AEX) buffer (20 mM Tris/HCl pH 8.0, 1 mM EDTA) and disrupted with a PandaPLUS 2000 homogenizer (GEA Niro Soavi, Parma, Italy). The recombinant ectodomain was purified from the whole cell extract by AEX using a MacroCap Q column (GE Healthcare, Munich, Germany) equilibrated with AEX buffer using a linear concentration gradient of 0-500 mM NaCl. Appropriate fractions were pooled, dialyzed against 100 mM Tris/HCl pH 8.0, 50 mM NaCl, 1 mM EDTA, and applied to a 5 ml column carrying a streptavidin mutant which allows binding of biotinylated proteins and competitive elution via an excess of free biotin (unpublished). After elution using 5 mM biotin in the same buffer, preparative size-exclusion chromatography (SEC) was performed on a 24 ml Superdex 200 10/300 GL column (GE Healthcare) using PBS (4 mM KH.sub.2PO.sub.4, 16 mM Na.sub.2HPO.sub.4, 115 mM NaCl, pH 7.4) as a running buffer.

    [0304] Expression procedure of the m/hCD98hcEDg in MEXi cells was essentially performed as recommended by the manufacturer. Briefly, cells were cultivated in MEXi culture medium supplemented with 50 mg/I G-418 and 8 mM L-alanyl-L-glutamine at 37° C. under humidified 5% CO.sub.2 atmosphere. 250 ml containing 5×10.sup.6 cells/mi were transfected using 5 μg plasmid DNA/1×10.sup.6 cells, and mixed at a 1:3 mass ratio with polyethylenimine in MEXi transfection medium. After 4 h incubation at 37° C., 500 ml of MEXi cultivation medium was added, and secretory expression was continued for 4 days. After sedimentation of the HEK cells by centrifugation, the supernatant was dialyzed against an immobilized metal ion affinity chromatography (IMAC) buffer (20 mM Tris/HCl pH 8.0, 500 mM NaCl), and IMAC was performed on a Ni(II)-charged HisTrap HP column (GE Healthcare). Elution fractions obtained after applying a linear concentration gradient of 0-300 mM imidazole/HCl in IMAC buffer were pooled and further purified via streptavidin affinity chromatography and preparative SEC as described above.

    [0305] Analytical SEC and SDS-PAGE

    [0306] Analytical SEC was performed using a 24 ml Superdex 200 10/300 GL column (GE Healthcare) in PBS pH 7.4 at a flow rate of 0.5 ml/min. For apparent molecular weight determination, the column was calibrated with the following standard proteins (Sigma-Aldrich, Munich, Germany): thyroglobulin (669 kDa), alcohol dehydrogenase (150 kDa), bovine serum albumin (66 kDa), carbonic anhydrase (29 kDa), cytochrome c (12.4 kDa) and aprotinin (6.5 kDa). The void volume of the column was determined using Blue dextran (Sigma-Aldrich). Based on the elution volumes, the partition coefficients K.sub.av were calculated and used to interpolate the apparent molecular sizes of the analyzed proteins.

    [0307] SDS-PAGE was performed using a high molarity Tris buffer system with or without the addition of 2-mercaptoethanol in the sample buffer, followed by staining with Coomassie brilliant blue or direct detection of Cy5.5 fluorescence (for corresponding protein conjugates) using an Ettan DIGE fluorescence scanner (GE Healthcare) with an excitation wavelength of 635/30 nm and a 680/30 nm emission band-path filter

    [0308] Enzymatic Cleavage of N-Linked Sugars from m/hCD98hcEDg Using PNGase F

    [0309] N-linked glycans were enzymatically removed from m/hCD98hcEDg produced in HEK cells using Peptide-N-Glycosidase F (PNGase F) (New England Biolabs, Ipswich, Mass.). 5 μg protein were first denatured by adding 0.5% sodium dodecyl sulfate (SDS) and 40 mM dithiothreitol (DTT) for 10 min at 100° C. Subsequently, 50 mM Na.sub.3PO.sub.4 pH 7.5, 1% (v/v) NP-40 and 250 units of PNGase F were added, incubated for 1 h at 37° C. and subjected to SDS-PAGE for comparison with the untreated ectodomain (Fig. S7D).

    [0310] Site-Specific Thiol Labelling of PASylated Lipocalin Variants Via Maleimide Chemistry

    [0311] For site-specific labelling at the engineered C-terminal Cys side chain (FIG. 17), maleimide-functionalized deferoxamine (Dfo; Macrocyclics, Plano, Tex.) or Sulfo-Cyanine5.5 (Cy5.5, Lumiprobe, Hannover, Germany) was used. To ensure presence of the free thiol for homogeneous C-terminal labelling without cleaving the intramolecular disulfide bridge of the lipocalin variant, the purified protein was incubated with a 20-fold molar concentration of DTT for 1 h at 20° C. in PBS pH 7.4. Then, the buffer was exchanged against 50 mM NaH.sub.2PO.sub.4 pH 5.5, 100 mM NaCl, 1 mM EDTA by gel filtration on a PD-10 column (GE Healthcare). After adjusting the pH to 7.4 using an appropriate volume of Na.sub.3PO.sub.4 and quantification of the protein concentration, a 5-fold molar concentration of the respective coupling reagent was added and incubated for 12 h at 4° C. Residual reagents were removed using a 24 ml Superdex 200 10/300 GL column equilibrated in PBS (pH 7.4). Successful 1:1 coupling with Dfo or Cy5.5 was verified using ESI-MS and the target affinity of each conjugate was verified.

    [0312] ESI Mass Spectrometry

    [0313] Mass spectra of proteins were measured on a maXis mass spectrometer with an electrospray ionization (ESI) source (Bruker Daltonics, Bremen, Germany) in the positive ion mode. To measure the intact protein mass (under denaturing conditions), the purified protein was dialyzed against 10 mM ammonium acetate pH 6.6 followed by the addition of 50% (v/v) methanol and 0.1% (v/v) acetic acid and application to the mass spectrometer via a syringe pump operated at 180 μL/h. The following conditions for the ion-transfer were used: 3400 V capillary voltage, 500 V endplate offset, 4 L/min dry gas at 200° C. temperature, 0.3 bar nebulizer pressure and 3 eV collision energy. Raw spectra were collected and deconvoluted with the Bruker Compass Data Analysis Software using the MaxEnt algorithm.

    EXAMPLE 2—RESULTS

    [0314] Phage Display Selection Divulged a hCD98hcED-Specific Lcn2 Variant with Pico-Molar Affinity

    [0315] The soluble, monomeric extracellular domain of human CD98hc (UniProt ID P08195-2, residues Glu105-Ala526) was produced fully glycosylated in eukaryotic human embryonic kidney (HEK) cells (CD98hcEDg) and unglycosylated in E. coli (CD98hcED) (FIG. 7). The four occupied N-glycosylation sites of CD98hcEDg (Asn264, 280, 323 and 405) account for 32.4 of its apparent mass as determined by analytical SEC and reduce its electrophoretic mobility (FIG. 7 and Table 2).

    TABLE-US-00006 TABLE 2 Size determination of the glycosylated and unglycosylated m/hCD98hcED Theoretical Apparent Apparent mass mass mass Protein [kDa] (SEC)[kDa] increase [%].sup.a hCD98hcED 48.4 49.4 — mCD98hcED 49.3 51.6 — hCD98hcEDg 49.3 73.1 32.4 mCD98hcEDg 50.2 79.8 37.1 .sup.aApparent mass increase caused by the N-linked glycosylation of h/mCD98hcED after production in HEK cells compared to production in E. coli

    [0316] This extensive glycosylation of hCD98hcED could impede in vitro phage display selection of cognate Anticalins due to poor sterical accessibility of potential surface epitopes, as previously seen in the selection of Anticalins against the prostate-specific membrane antigen [65]. To avoid the necessity of glycan removal, which either requires harsh reaction conditions or results in incomplete deglycosylation, here the unglycosylated ectodomain produced in E. coli was used as the initial molecular target for selection. N-terminal fusion of the biotin acceptor peptide (BAP) resulted in the intracellular enzymatic attachment of a single biotin group upon co-expression of the BirA ligase (verified via western blot analysis, data not shown) for both CD98hcEDg and CD98hcED, regardless of the expression system used [66]. This modification allowed immobilization of the ectodomain to beads or surfaces in an orientation that resembles the in vivo situation for the type II membrane protein, accessible for binding of lipocalin variants during phage display selection, ELISA screening and in vitro binding studies.

    [0317] Anticalin candidates were selected from a Lcn2 random library with a combinatorial complexity of 1×10.sup.10, which also had served for the development of other hapten- and protein-specific Anticalins in the past [50]. After six selection cycles, three independent CD98hc-specific lipocalin variants were identified by ELISA screening, then expressed as soluble proteins in E. coli at the shake flask scale and purified to homogeneity, resulting in monomeric proteins with apparent molecular size similar to wild type (wt) Lcn2 (FIG. 1B, FIGS. 8A and B). To determine the rate constants of association and dissociation of the selected Anticalin candidates, SPR real-time analyses were performed using both the glycosylated and unglycosylated immobilized CD98hcED. The three lipocalin variants showed equilibrium dissociation constants in the single-digit to sub-nanomolar range (Table 1, FIG. 1C and FIG. 8C).

    TABLE-US-00007 TABLE 1 Affinities for hCD98hcED and thermal stabilities of engineered lipocalin variants Affinity Stability Protein K.sub.D [nM] k.sub.on [M.sup.−1 × s.sup.−1] k.sub.off [s.sup.−1] T½ [min] T.sub.m [° C.] ΔH.sub.m [kJ/mol] ΔG.sub.U [kJ/mol] P1E4 4.5.sup.a  2.5 × 10.sup.6 1.1 × 10.sup.−2 1.1 68.8 812 104.0 P3A12 2.8.sup.b  6.2 × 10.sup.4 1.7 × 10.sup.−4 67.7 64.1 635 73.6 P3D11 0.15.sup.b 2.2 × 10.sup.5 3.2 × 10.sup.−5 359.4 55.2 368 33.8 D11vs 0.05.sup.b 4.5 × 10.sup.5 2.6 × 10.sup.−5 442.3 63.5 646 73.9 .sup.a= multiple cycle kinetics .sup.b= single cycle kinetics

    [0318] Notably, the lipocalin variant P3D11 revealed a very low K.sub.D value of 150 pM for hCD98hcED and a long complex half-life of 6 h, hence suitable for tumor targeting in vivo (FIG. 1C). Interestingly, a series of competitive SPR experiments indicated that all three Anticalins recognize overlapping epitopes on hCD98hcED (FIG. 8D, example shown for the variants P3D11 and P3A12), despite considerable sequence deviation in particular between variants P1E4 and P3D11 (cf. FIG. 8A).

    [0319] Protruding Loops Constitute the Major Epitope for hCD98hcED Recognition by the Engineered Lipocalin

    [0320] To identify the epitope region targeted by these Anticalin candidates and to understand the structural mechanisms of the tight molecular interaction of P3D11 with the ectodomain, we crystallized P3D11 in complex with hCD98hcED (carrying a N-terminal Strep-tag II) after isolation of the 1:1 complex via SEC (FIG. 9). The hCD98hcED.circle-solid.P3D11 complex crystallized in space group C2 with two complexes per asymmetric unit, and its X-ray structure was refined to a resolution of 1.8 Å (Table 4). hCD98hcED resembles a glycoside hydrolase fold, comprising a central (α/β).sub.8-TIM barrel core domain (residues 114-438) and a C-terminal β-sandwich domain (residues 439-529), as previously described [55]. Both N- and C-termini point toward the cytoplasmic membrane, hence mainly the C-terminal membrane distal side of the TIM barrel domain is accessible to bind molecules. Indeed, the engineered lipocalin binds the membrane-distal part of hCD98hcED with its four variable loops #1-#4 protruding from the β-barrel (FIG. 2). The β-barrel axis of P3D11 is tilted by ˜45° with regard to the TIM barrel axis of hCD98hcED. The contact interface between both molecules involves a total buried surface area (BSA) of 1473 Å.sup.2 (1424 Å.sup.2 on the side of the lipocalin and 1521 Å.sup.2 on the one of the membrane protein), 15 hydrogen bonds and 3 salt bridges (Table 5), resulting in the largest interface observed for Anticalin.circle-solid.protein complexes thus far [67].

    TABLE-US-00008 TABLE 5 Summary of the contact area and contacts provided by the lipocalin variant P3D11 BSA [Å.sup.2] BSA mutated [Å.sup.2] HB/SB Loop #1 486 156 5/2  Loop #2 175 169 2/— Loop #3 90 1.0 1/— Loop #4 97 77 1/— β-barrel 576 364 6/1  SUM 1424 767 15/3 

    [0321] In part the interaction is driven by electrostatics as P3D11 is positively charged, favoring interaction with the predominantly negatively charged epitope on hCD98hcED, which is reflected by the calculated pl values of 8.8 and 5.2 for P3D11 and hCD98hcED, respectively. The epitope of hCD98hcED essentially consists of two loops, L1 (residues 128-137) and L2 (374-404), which follow to the first and the eighth β-strand of the TIM barrel, respectively.

    [0322] Almost 75% of the contact interface is contributed by the interaction with L2, which penetrates deep into the β-barrel of the engineered lipocalin, whereas L1 provides approximately 20% of the total BSA (Table 6)

    TABLE-US-00009 TABLE 6 Summary of the contact area and contacts provided by hCD98hcED BSA [Å.sup.2] HB/SB hCD98hcED 1521 15/3  Loop L1 288 6/1 Loop L2 1111 6/1

    [0323] Further to the direct protein.circle-solid.protein contacts, 16 water molecules mediate hydrogen bonds at the complex interface. Based on the distribution of hydrogen bond donors, acceptors as well as water molecules, the P3D11 interface with the hCD98hcED can be dissected into two distinct areas, a rather polar region around the lipocalin loops #1 and #2 including the cavity as well as a rather apolar region around loop #3 and #4 (FIG. 2B).

    [0324] Interestingly, binding of P3D11 causes an induced fit of hCD98hcED. Superposition of the hCD98hcED Cα positions in complex with P3D11 and for hCD98hcED crystallized alone (PDB entry 2DH2) [55], excluding the epitope loops L1 and L2, revealed an RMSD value of 0.97 Å (for 375 equivalent Cα positions), while in this context loops 1 and 2 deviate by 1.6 and 1.8 Å, respectively (FIG. 10A). Likewise, P3D11 shows conformational differences among its loops #1-4 compared with wtLcn2 (PDB code 1L6M); yet, these may be caused by the sequence changes (FIG. 10B).

    [0325] Although the selection of P3D11 was performed with the unglycosylated hCD98hcED, its affinity for the glycosylated hCD98hcEDg was indistinguishable (Table 3).

    TABLE-US-00010 TABLE 3 Affinities of selected and engineered lipocalin variants for hCD98hcEDg determined via SPR analysis Protein K.sub.D [nM] k.sub.on M.sup.−1 × s.sup.−1] k.sub.off [s.sup.−1] T.sub.1/2 [min] P1E4 162 5.6 × 10.sup.5 9.0 × 10.sup.−2 <1 P3A12 1.9 2.2 × 10.sup.4 4.0 × 10.sup.−4 29 P3D11 0.17 3.0 × 10.sup.5 5.0 × 10.sup.−5 230 D11vs 0.05 4.6 × 10.sup.5 2.2 × 10.sup.−5 523

    [0326] Modeling of the glycan moieties suggests that the Anticalin binds the protruding epitope in close proximity to two of the oligosaccharides but is not sterically hindered by them (FIG. 2C). Interestingly, P3D11 does not show any affinity for the murine antigen (mCD98hcED, data not shown), which can be explained by significant differences both in the amino acid sequences and in the glycosylation pattern between the CD98hc orthologs. In particular, loop L2 of mCD98hcED is shorter by one residue, carries a N-glycosylation site and shows considerable sequence changes (Deuschle et al. submitted). Notably, this observation is in line with the known properties of the clinical-stage mAb IGN523, which also recognizes loop L2 (residues 374-401) of hCD98hcED with picomolar affinity but shows no detectable affinity for mCD98hc [20, 68].

    [0327] Directed Evolution Results in an Anticalin with Improved Affinity and Stability

    [0328] Despite its tight binding activity towards hCD98hcED, the lipocalin variant P3D11 suffered from low thermal stability, a caveat for further preclinical as well as clinical development (Table 1). To tackle this issue, stability engineering was performed via directed evolution using bacterial surface display (BSD) starting from an error-prone library with moderate amino acid mutation rate (with approximately two amino acid exchanges on average) (FIG. 11A).

    [0329] To select variants of P3D11 with enhanced thermal stability, expression of the membrane-anchored protein was induced at 37° C. and only bacteria showing both strong hCD98hcED binding and high cell surface display level were gated in the FACS experiment (FIG. 3A). After six enrichment cycles, the lipocalin library exhibited a strong average binding signal after incubation with 1 nM hCD98hcED, whereas almost no binding was detectable for the initial clone P3D11 if tested under the same conditions. Single clone analysis with 1 nM hCD98hcED verified strong binding of several of the new P3D11 variants (D11.1 shown as an example, FIG. 3B and FIG. 11B) whereas just very modest binding was detected for P3D11, and no signal for wtLcn2 as expected. Considering that P3D11 exhibits a K.sub.D value of 150 pM for hCD98hcED the increased binding signal is likely caused by the more efficient bacterial surface display of lipocalin variants with improved protein folding and/or stability. Subsequent sequence analysis revealed that replacement of Gly81 and Phe71 by Val and Ser, respectively, in the P3D11 sequence appear to be crucial in this regard (FIG. 11C). Therefore, the P3D11 variant D11vs, which carries both mutations, was generated and characterized with respect to its thermal stability and affinity.

    [0330] CD thermal unfolding studies at physiological pH demonstrated a significant rise by 8.6° C. in the melting temperature (T.sub.m) for D11vs (T.sub.m=63.5° C.) compared to P3D11 (T.sub.m=55.2° C.) (FIG. 3C and Table 1). Importantly, the cooperativity of the unfolding transition was much steeper, corresponding to a higher enthalpy of denaturation (ΔH.sub.m) and extrapolated free energy of denaturation (ΔG.sub.U) at standard conditions (25° C.). This is in line with a lower tendency of D11vs to aggregate during incubation at 37° C. compared to P3D11 (FIG. 11D). Additionally, D11vs showed even higher affinity towards hCD98hcED, with a K.sub.D value of 50 pM, due to both faster association and slower dissociation rate constants (FIG. 11E and Table 1).

    [0331] The enhanced stability and affinity of D11vs can be explained on the basis of the crystal structure solved for the hCD98hcED.circle-solid.P3D11 complex: residue Gly81 was introduced during the initial selection campaign, where it replaced Arg81 present in wtLcn2, thus creating space for tight target binding (FIG. 11C). However, due to the lack of a β-carbon, glycine causes a higher backbone conformational flexibility, which destabilizes the β-barrel [69]. Accordingly, substitution by valine leads to enhanced protein stability while filling a hydrophobic cavity in this region lined by the lipocalin residues Tyr68, Gly81, Phe83 and Leu94 (FIG. 12A), of which Tyr68 also was introduced during the initial selection. Moreover, Val81 increases the contact area with hCD98hcED, which likely contributes to the enhanced affinity (FIG. 12B). On the other hand, replacement of Phe71 by Ser enables the formation of stabilizing hydrogen bonds to Asn174 and Lys50 within the Anticalin (FIG. 12C) Furthermore, the smaller Ser side chain may enable loop #2 of the engineered lipocalin to bend slightly away from the β-barrel axis, thus creating some space for better target binding.

    [0332] D11vs Binds Human CD98hc on Different Tumor Cell Lines

    [0333] For in vivo studies in mice, a modified version of D11vs with moderately prolonged plasma half-life was designed using PASylation technology [70]. To this end, a structurally disordered polypeptide comprising in total 200 proline, alanine and serine residues (PAS200) was genetically fused to the C-terminus of the engineered lipocalin, thus increasing its hydrodynamic molecular volume which leads to a retarded kidney filtration (FIG. 17A) [49]. In the context of in vivo imaging applications, previous investigations have shown that fine-tuning of the pharmacokinetics in this manner can boost protein tracer accumulation and tumor contrast [46]. Furthermore, D11vs-PAS200 was equipped with an engineered Cys residue at the C-terminus of the PAS-tag, which enabled homogenous site-specific conjugation of chemical groups for in vitro and in vivo imaging as well as drug delivery (FIG. 17).

    [0334] The ability of D11vs-PAS200 to bind hCD98hc in a native cellular environment was investigated by cytofluorometry and immunofluorescence microscopy using human cancer cell lines of different origins exhibiting strong expression of this tumor marker. Flow cytometric analysis of the B-cell lymphoma cell lines SU-DHL-4, Raji and Ramos, the prostate carcinoma cells DU-145 and PC-3 and the colorectal adenocarcinoma cell line Caco-2 with Cy5.5-labeled D11vs-PAS200 revealed strong binding of the membrane-associated target protein (FIG. 4A). Competition experiments with excess unlabeled D11vs-PAS200 as well as soluble glycosylated hCD98hcED resulted in signal suppression, hence confirmed high binding specificity of the Anticalin towards native hCD98hc. Immunofluorescence microscopy experiments with D11vs-PAS200-Cy5.5 revealed pronounced membrane and also cytoplasmic staining of Ramos, PC-3 and Caco-2 cancer cells (FIG. 4B-C). Cell staining was effectively blocked when adding a 10-fold molar concentration of the unlabeled lipocalin variant, again confirming its specificity. As expected, no fluorescence staining for any of the tested cell lines was detected when using the Cy5.5-labeled PASylated wtLcn2.

    [0335] D11vs Shows Strong and Specific Accumulation in Two Human Xenograft Models

    [0336] The applicability of PASylated D11vs for in vivo targeting of hCD98hc was evaluated by a PET/CT imaging study with the .sup.89Zr-labeled protein in two murine xenograft tumor models. For this purpose, the human cancer cell lines Ramos (FIG. 13) and PC-3 (FIG. 5) were chosen, considering that CD98hc was described as a promising tumor marker for lymphoma [24] as well as other hematological cancers [17] and is known to be overexpressed in metastatic prostate cancer [23, 71]. In a first study, five female mice each were injected with ˜3 MBq of the radiolabeled protein tracer, and PET/CT imaging was performed at time points t=24, 48 and 72 h p.i. (FIGS. 5A and 13A-C). To assess target-specificity of the injected Anticalin, two of these five mice received a 100-fold molar excess of the unlabeled PASylated Anticalin 2 h before .sup.89Zr-tracer injection. After 24 h, the best imaging contrast and a high accumulation of 3.0±0.3% ID/g was observed for the PC-3 tumors, with decreasing tumor to background contrast after 48 and 72 h. Threshold-based segmentation of the PET images (FIG. 5B) showed a significant (P=0.024) difference in radiotracer accumulation between tracer-injected mice (N=3) and those mice (N=2) which were additionally injected with unlabeled Anticalin. Apart from the strong uptake by the tumors, elimination-related accumulation of radioactivity in liver, kidney and bladder was observed. Furthermore, signals in the joints were detected, which can be explained by accumulation of liberated zirconium caused by the limited complex stability of the linear chelator Dfo [72]. Notably, signals detected in joints and kidneys showed no significant difference in mice with or without blocked CD98hc epitopes, which indicated a tracer-independent uptake into these tissues (FIG. 5B).

    [0337] Furthermore, the .sup.89Zr-labeled D11vs-PAS200 tracer allowed good visualization of the Ramos xenografts 24 h p.i., yet at a lower level compared to the PC-3 tumors. Of note, the Ramos xenografts were less well localized and fast growing, which led to blurred imaging signals. Consequently, the PC-3 xenografts were used for further experiments. After in vivo PET imaging, explanted PC-3 tumors were used for autoradiography of tissue cross sections. In this analysis, a homogeneous distribution of the radiotracer within the PC-3 tumor tissue was detected (FIG. 5C), which was strongly reduced in tumors collected from the blocked mice. Concomitant immunohistochemical analysis of vessel distribution (via staining of CD31) as well as hCD98hc abundancy, using cognate antibodies, revealed a well-vascularized tumor with high expression of this tumor marker for both PC-3 and Ramos xenografts (FIG. 5D and Fig. S10). For comparison, one mouse from each cohort was imaged with the small molecule tracers .sup.18F-fluorodeoxyglucose (FDG) and .sup.18F-fluoroethyl-L-tyrosine (FET) in PET/CT (FIG. 15). Interestingly, the segmentation-derived values for uptake of FET, which is mainly mediated by the CD98lc Lat-1, was higher in Ramos (15.9±2.9% ID/g) than in PC-3 (9.7±1.4% ID/g), which is in contrast to the observed accumulation of the .sup.89Zr-Anticalin tracer.

    [0338] In order to confirm these imaging results and to obtain quantitative data from ex vivo biodistribution experiments at the optimal time point 24 h p.i., a second study using male CB17-SCID mice baring PC-3 tumors was conducted (FIG. 6). Two cohorts with each five mice were injected with either 3.96±0.12 MBq .sup.89Zr-labeled D11vs-PAS or the protein tracer together with a 250-fold molar amount of the unlabeled PASylated Anticalin (t=−2 h). After 24 h, PET/CT scans were performed, tumors were explanted, subjected to an ex vivo PET scan and, finally, a biodistribution analysis was conducted. PET/CT images of the mice without prior blocking of hCD98hc showed pronounced accumulation of radioactivity in the tumor while the tumors were merely visible in the blocked mice, in line with the ex vivo PET/CT ROI segmentation analysis (P=0.027). Finally, biodistribution study revealed high accumulation of radioactivity (8.6±1.1% ID/g) in the tumor (vs. 5.4±1.1% ID/g in the blocked cohort, P=0.0032), with good tumor-to-blood and tumor-to-muscle ratios of 3.5 and 11.8, respectively. Radioactivity in the excretion-related organs matched the observations above with a particularly pronounced accumulation in the spleen. This can be explained by formation of insoluble Zr-phosphate, which gets quickly absorbed [73]. In summary, the hCD98hc-specific PASylated Anticalin tracer showed high and specific tumor uptake, which may in part be explained by the intracellular accumulation of the residualizing .sup.89Zr radiometal after targeting the highly expressed and rapidly internalizing tumor marker.

    EXAMPLE 3—DISCUSSION

    [0339] The pathophysiological significance of CD98hc expression in several solid and blood cancers is well established today. Even though also expressed weakly in healthy cells and tissues such as activated lymphocytes, proximal tubules, placenta or testis, strong overexpression of this membrane protein has been shown to correlate with tumor progression and aggressiveness, thus allowing the assessment of disease prognosis, treatment response as well as overall survival in oncologic diseases. Hence, CD98hc constitutes a highly attractive biomedical target for diagnostic as well as therapeutic purposes in oncology. However, no CD98hc-specific reagents for in vivo imaging and only few drug candidates addressing this target are available to date.

    [0340] To meet this need, an Anticalin was developed with exquisite specificity and high (picomolar) affinity towards the extracellular domain of hCD98hc. The initial Anticalin candidates were selected from a random library based on human lipocalin 2 (Lcn2), an abundant siderophore-binding plasma protein. This previously designed library, which carries an optimally distributed set of randomized amino acid positions within the binding site of the protein scaffold, was successfully applied in previous studies to generate Anticalins against proteinaceous biomedical targets (ED-B, Hsp70, VEGFR-3 or PSMA) [50, 60, 71, 72], peptides (monomeric Aβ) [73] and hapten-type ligands (Colchicin, Y.sup.III-DTPA, Petrobactin) [50, 60, 71, 72]. The lipocalin variants selected towards hCD98hcED from this naïve Lcn2 library show K.sub.D values already in the pico- to single-digit nanomolar range combined with a stable monomeric behavior, which once again validates the utility of this library.

    [0341] Based on its association and dissociation rate constants, with a remarkable complex dissociation half-life of ˜6 h, P3D11 was chosen as the lead candidate for further characterization and optimization. X-ray structural analysis of its complex with the CD98hc ectodomain revealed deep penetration of loop L2 within the membrane-distal part of CD98hc into the cup-shaped β-barrel of the Anticalin, with a buried total paratope surface area of 1424 Å.sup.2. Competition analysis for binding of glycosylated hCD98hc showed that all the selected Anticalins, including P3D11, recognize the same epitope region. P3D11 and P3A12 share a similar set of amino acid exchanges compared with wtLcn2, with 8 identical residues among the 20 randomized positions. Highly similar sequence stretches are seen for loop #1, loop #2 as well as their neighboring β-strands A/B and C/D (FIG. 8A), which also strongly contribute to the protein interface in the hCD98hcED.circle-solid.P3D11 complex (Table 5).

    [0342] While N-glycosylation has no influence on target affinity of P3D11 (or its improved version D11vs) and P3A12, binding of P1E4 is significantly diminished (˜100-fold). This behavior can be explained by minor variations in the mutual orientations between hCD98hcED and the different Anticalins, which could lead to varying sensitivity towards sterically demanding carbohydrate side chains in the neighborhood despite targeting of a similar epitope region (FIG. 2C). Indeed, such a kind of structural paratope plasticity was previously observed for a set of Anticalins directed against the extra-domain B of oncofetal fibronectin [64]. Of note, the clinical-stage mAb IGN523 recognizes the same loop L2 (residues 374-401) of hCD98hcED, also with picomolar affinity [20, 66]. Treatment of patients in a phase I clinical trial did not lead to severe adverse effects, suggesting that this epitope of CD98hc can be targeted safely.

    [0343] Despite remarkable antigen affinity, which would be suitable for in vivo tumor targeting, especially the Anticalin candidate P3D11 suffered from lower thermal stability and showed pronounced aggregation after prolonged incubation at 37° C. To overcome this obstacle, directed evolution via moderate random mutagenesis was employed in combination with high throughput sorting using a recently developed system for the bacterial surface display of Anticalins [59]. By applying high temperature (37° C.) during protein expression selection of variants showing improved protein folding and stability was forced. With just two additional amino acid substitutions—Phe71Ser and Gly81Val (located in loop #2 and β-strand D, respectively)—the variant D11vs exhibited both increased affinity (approximately three-fold) as well as significantly increased melting temperature (by ˜8° C.; cf. Table 1 and FIG. 3C), thus nicely illustrating the power of directed evolution for stability engineering of proteins. (FIG. 11C).

    [0344] Small animal PET/CT imaging and biodistribution analysis with the plasma half-life optimized and .sup.89Zr-labeled D11vs radiotracer showed high accumulation in PC-3 xenograft tumors with good imaging contrast. Blocking of CD98hc binding by addition of unlabeled D11 vs-PAS200 reduced tracer accumulation by ˜40%. This is in line with in vitro cytofluorometry and immunofluorescence microscopy, where highly specific binding of D11vs on several CD98hc-positive human cancer cell lines, but not on cells with the blocked CD98hc epitope, was evident. Tracer uptake in kidney, liver, spleen and joints was as expected due to the known in vivo release of .sup.89Zr(IV) from the radiochelator [69, 74]. Superior .sup.89Zr chelating reagents with higher in vivo stability, as for example fusarinin C, a siderophore-based cyclic chelator derived from Aspergillus fumigatus [75, 76], may be an option for future experiments.

    [0345] In conclusion, human CD98hcED-specific Anticalin was selected and engineered with high affinity and successfully applied it for CD98hc-targeted PET diagnostic imaging. This novel radiotracer showed high tumor uptake in prostate carcinoma as well as B-cell lymphoma xenograft models in mice with excellent tumor visualization. To our knowledge, this is the first in vivo imaging study of CD98hc overexpression on tumor cells described in the literature. Considering the clinical importance of elevated CD98hc expression in several human cancers, this Anticalin constitutes a promising new tool for preclinical and, potentially, clinical applications in oncology.

    REFERENCES

    [0346] 1. Cantor J M, Ginsberg M H. CD98 at the crossroads of adaptive immunity and cancer. Journal of cell science. 2012; 125: 1373-82. [0347] 2. Fotiadis D, Kanai Y, Palacin M. The SLC3 and SLC7 families of amino acid transporters Molecular aspects of medicine. 2013; 34: 139-58. [0348] 3. Kanai Y, Segawa H, Miyamoto K, Uchino H, Takeda E, Endou H. Expression cloning and characterization of a transporter for large neutral amino acids activated by the heavy chain of 4F2 antigen (CD98) The Journal of biological chemistry. 1998; 273: 23629-32. [0349] 4. Pineda M, Fernandez E, Torrents D, Estevez R, Lopez C, Camps M, et al. Identification of a membrane protein, LAT-2, that Co-expresses with 4F2 heavy chain, an L-type amino acid transport activity with broad specificity for small and large zwitterionic amino acids The Journal of biological chemistry. 1999; 274: 19738-44. [0350] 5. Torrents D, Estevez R, Pineda M, Fernandez E, Lloberas J, Shi Y B, et al. Identification and characterization of a membrane protein (y+L amino acid transporter-1) that associates with 4F2hc to encode the amino acid transport activity y+L. A candidate gene for lysinuric protein intolerance The Journal of biological chemistry. 1998; 273: 32437-45. [0351] 6. Broer A, Wagner C A, Lang F, Broer S. The heterodimeric amino acid transporter 4F2hc/y+LAT2 mediates arginine efflux in exchange with glutamine. The Biochemical journal. 2000; 349 Pt 3: 787-95. [0352] 7. Fukasawa Y, Segawa H, Kim J Y, Chairoungdua A, Kim D K, Matsuo H, et al. Identification and characterization of a Na(+)-independent neutral amino acid transporter that associates with the 4F2 heavy chain and exhibits substrate selectivity for small neutral D- and L-amino acids. The Journal of biological chemistry. 2000; 275: 9690-8. [0353] 8. Sato H, Tamba M, Ishii T, Bannai S. Cloning and expression of a plasma membrane cystine/glutamate exchange transporter composed of two distinct proteins The Journal of biological chemistry. 1999; 274: 11455-8. [0354] 9. Nakamura E, Sato M, Yang H, Miyagawa F, Harasaki M, Tomita K, et al. 4F2 (CD98) heavy chain is associated covalently with an amino acid transporter and controls intracellular trafficking and membrane topology of 4F2 heterodimer The Journal of biological chemistry. 1999; 274: 3009-16. [0355] 10. Nicklin P, Bergman P, Zhang B, Triantafellow E, Wang H, Nyfeler B, et al. Bidirectional transport of amino acids regulates mTOR and autophagy. Cell. 2009; 136: 521-34. [0356] 11. Koppula P, Zhang Y, Zhuang L, Gan B. Amino acid transporter SLC7A11/xCT at the crossroads of regulating redox homeostasis and nutrient dependency of cancer. Cancer communications. 2018; 38: 12. [0357] 12. Yoshida G J. Metabolic reprogramming: The emerging concept and associated therapeutic strategies. Journal of experimental & clinical cancer research: CR. 2015; 34: 111. [0358] 13. Cantor J M, Ginsberg M H, Rose D M. Integrin-associated proteins as potential therapeutic targets. Immunological reviews. 2008; 223: 236-51. [0359] 14. Miyamoto Y J, Mitchell J S, McIntyre B W. Physical association and functional interaction between β1 integrin and CD98 on human T lymphocytes. Molecular immunology. 2003; 39: 739-51. [0360] 15. Zent R, Fenczik C A, Calderwood D A, Liu S, Dellos M, Ginsberg M H. Class- and splice variant-specific association of CD98 with integrin β cytoplasmic domains. The Journal of biological chemistry. 2000; 275: 5059-64. [0361] 16. Henderson N C, Collis E A, Mackinnon A C, Simpson K J, Haslett C, Zent R, et al. CD98hc (SLC3A2) interaction with β1 integrins is required for transformation. The Journal of biological chemistry. 2004; 279: 54731-41. [0362] 17. Bajaj J, Konuma T, Lytle N K, Kwon H Y, Ablack J N, Cantor J M, et al. CD98-mediated adhesive signaling enables the establishment and propagation of acute myelogenous leukemia. Cancer cell. 2016; 30: 792-805. [0363] 18. Ye Y, Wang M, Wang B, Yang X-M, Chen Z-N. CD98, a potential diagnostic cancer-related biomarker, and its prognostic impact in colorectal cancer patients. Int J Clin Exp Pathol. 2017; 10: 5418-29. [0364] 19. Kaira K, Oriuchi N, Imai H, Shimizu K, Yanagitani N, Sunaga N, et al. CD98 expression is associated with poor prognosis in resected non-small-cell lung cancer with lymph node metastases. Ann Surg Oncol. 2009; 16: 3473-81. [0365] 20. Hayes G M, Chinn L, Cantor J M, Cairns B, Levashova Z, Tran H, et al. Antitumor activity of an anti-CD98 antibody. International journal of cancer. 2015; 137: 710-20. [0366] 21. El Ansari R, Craze M L, Diez-Rodriguez M, Nolan C C, Ellis I O, Rakha E A, et al. The multifunctional solute carrier 3A2 (SLC3A2) confers a poor prognosis in the highly proliferative breast cancer subtypes. Brit J Cancer. 2018; 118: 1115-22. [0367] 22. Furuya M, Horiguchi J, Nakajima H, Kanai Y, Oyama T. Correlation of L-type amino acid transporter 1 and CD98 expression with triple negative breast cancer prognosis. Cancer Sci. 2012; 103: 382-9. [0368] 23. Wang Q, Tiffen J, Bailey C G, Lehman M L, Ritchie W, Fazli L, et al. Targeting amino acid transport in metastatic castration-resistant prostate cancer: Effects on cell cycle, cell growth, and tumor development. Journal of the National Cancer Institute. 2013; 105: 1463-73. [0369] 24. Salter D M, Krajewski A S, Sheehan T, Turner G, Cuthbert R J, McLean A. Prognostic significance of activation and differentiation antigen expression in B-cell non-Hodgkin's lymphoma The Journal of pathology. 1989; 159: 211-20. [0370] 25. Digomann D, Kurth I, Tyutyunnykova A, Chen O, Lock S, Gorodetska I, et al. The CD98 heavy chain is a marker and regulator of head and neck squamous cell carcinoma radiosensitivity. Clinical cancer research: an official journal of the American Association for Cancer Research. 2019. [0371] 26. Yang H, Zou W, Li Y, Chen B, Xin X. Bridge linkage role played by CD98hc of anti-tumor drug resistance and cancer metastasis on cisplatin-resistant ovarian cancer cells. Cancer biology & therapy. 2007; 6: 942-7. [0372] 27. Kaira K, Ohde Y, Endo M, Nakagawa K, Okumura T, Takahashi T, et al. Expression of 4F2hc (CD98) in pulmonary neuroendocrine tumors Oncology reports. 2011; 26: 931-7. [0373] 28. Rietbergen M M, Martens-de Kemp S R, Bloemena E, Witte B I, Brink A, Baatenburg de Jong R J, et al. Cancer stem cell enrichment marker CD98: A prognostic factor for survival in patients with human papillomavirus-positive oropharyngeal cancer. European journal of cancer (Oxford, England: 1990). 2014; 50: 765-73. [0374] 29. Prager G W, Poettler M, Schmidinger M, Mazal P R, Susani M, Zielinski C C, et al. CD98hc (SLC3A2), a novel marker in renal cell cancer. European journal of clinical investigation. 2009; 39: 304-10. [0375] 30. Isoda A, Kaira K, Iwashina M, Oriuchi N, Tominaga H, Nagamori S, et al. Expression of L-type amino acid transporter 1 (LAT1) as a prognostic and therapeutic indicator in multiple myeloma. Cancer Sci. 2014; 105: 1496-502. [0376] 31. Kaira K, Oriuchi N, Imai H, Shimizu K, Yanagitani N, Sunaga N, et al. L-type amino acid transporter 1 and CD98 expression in primary and metastatic sites of human neoplasms. Cancer Sci. 2008; 99: 2380-6. [0377] 32. Toyoda M, Kaira K, Shino M, Sakakura K, Takahashi K, Takayasu Y, et al. CD98 as a novel prognostic indicator for patients with stage III/1V hypopharyngeal squamous cell carcinoma Head & neck. 2015; 37: 1569-74. [0378] 33. Hayashi K, Anzai N. Novel therapeutic approaches targeting L-type amino acid transporters for cancer treatment. World J Gastrointest Oncol. 2017; 9: 21-9. [0379] 34. Salisbury T B, Arthur S. The regulation and function of the L-type amino acid transporter 1 (LAT1) in cancer. Int J Mol Sci. 2018; 19: 2373. [0380] 35. Barollo S, Bertazza L, Watutantrige-Fernando S, Censi S, Cavedon E, Galuppini F, et al. Overexpression of L-type amino acid transporter 1 (LAT1) and 2 (LAT2): Novel markers of neuroendocrine tumors. PloS one. 2016; 11: e0156044-e. [0381] 36. Sugano K, Maeda K, Ohtani H, Nagahara H, Shibutani M, Hirakawa K. Expression of xCT as a predictor of disease recurrence in patients with colorectal cancer. Anticancer Res. 2015; 35: 677-82. [0382] 37. Ji X, Qian J, Rahman S M J, Siska P J, Zou Y, Harris B K, et al. xCT (SLC7A11)-mediated metabolic reprogramming promotes non-small cell lung cancer progression Oncogene. 2018; 37: 5007-19. [0383] 38. Kaira K, Oriuchi N, Imai H, Shimizu K, Yanagitani N, Sunaga N, et al. Prognostic significance of L-type amino acid transporter 1 (LAT1) and 4F2 heavy chain (CD98) expression in early stage squamous cell carcinoma of the lung. Cancer Sci. 2009; 100: 249-54. [0384] 39. Nakanishi K, Ogata S, Matsuo H, Kanai Y, Endou H, Hiroi S, et al. Expression of LAT1 predicts risk of progression of transitional cell carcinoma of the upper urinary tract. Virchows Arch A. 2007; 451: 681-90. [0385] 40. Bixby D, Wieduwilt M J, Akard L P, Khoury H J, Becker P S, Van Der Horst E H, et al. A Phase I Study of IGN523, a Novel Anti-CD98 Monoclonal Antibody in Patients with Relapsed or Refractory Acute Myeloid Leukemia (AML). Blood. 2015; 126: 3809. [0386] 41. Wang Q, Hoist J. L-type amino acid transport and cancer: Targeting the mTORC1 pathway to inhibit neoplasia. Am J Cancer Res. 2015; 5: 1281-94. [0387] 42. Klasner B D, Krause B J, Beer A J, Drzezga A. PET imaging of gliomas using novel tracers: A sleeping beauty waiting to be kissed. Expert review of anticancer therapy. 2010; 10: 609-13. [0388] 43. Ikotun O F, Marquez B V, Huang C, Masuko K, Daiji M, Masuko T, et al. Imaging the L-type amino acid transporter-1 (LAT1) with Zr-89 immunoPET. PloS one. 2013; 8: e77476-e. [0389] 44. Hutterer M, Nowosielski M, Putzer D, Jansen N L, Seiz M, Schocke M, et al. [18F]-fluoro-ethyl-L-tyrosine PET: A valuable diagnostic tool in neuro-oncology, but not all that glitters is glioma. Neuro-oncology. 2013; 15: 341-51. [0390] 45. Singh N, Ecker G F. Insights into the structure, function, and ligand discovery of the large neutral amino acid transporter 1, LAT1. Int J Mol Sci. 2018; 19. [0391] 46. Mendler C T, Friedrich L, Laitinen I, Schlapschy M, Schwaiger M, Wester H J, et al. High contrast tumor imaging with radio-labeled antibody Fab fragments tailored for optimized pharmacokinetics via PASylation. mAbs. 2015; 7: 96-109. [0392] 47. Richter A, Eggenstein E, Skerra A. Anticalins: Exploiting a non-Ig scaffold with hypervariable loops for the engineering of binding proteins. FEBS letters. 2014; 588: 213-8. [0393] 48. Rothe C, Skerra A. Anticalin® proteins as therapeutic agents in human diseases. BioDrugs. 2018; 32: 233-43. [0394] 49. Schlapschy M, Binder U, Borger C, Theobald I, Wachinger K, Kisling S, et al. PASylation:

    [0395] A biological alternative to PEGylation for extending the plasma half-life of pharmaceutically active proteins. Protein engineering, design & selection: PEDS. 2013; 26: 489-501. [0396] 50. Gebauer M, Schiefner A, Matschiner G, Skerra A. Combinatorial design of an Anticalin directed against the extra-domain b for the specific targeting of oncofetal fibronectin Journal of molecular biology. 2013; 425: 780-802. [0397] 51. Friedrich L, Kornberger P, Mendler C T, Multhoff G, Schwaiger M, Skerra A. Selection of an Anticalin® against the membrane form of Hsp70 via bacterial surface display and its theranostic application in tumour models. Biological chemistry. 2018; 399: 235-52. [0398] 52. Mueller U, Darowski N, Fuchs M R, Forster R, Hellmig M, Paithankar K S, et al. Facilities for macromolecular crystallography at the Helmholtz-Zentrum Berlin. J Synchrotron Radiat. 2012; 19: 442-9. [0399] 53. Kabsch W. XDS. Acta Crystallographica Section D Biological Crystallography. 2010; 66: 125-32. [0400] 54. McCoy A J, Grosse-Kunstleve R W, Adams P D, Winn M D, Storoni L C, Read R J. Phaser crystallographic software. J Appl Crystallogr. 2007; 40: 658-74. [0401] 55. Fort J, de la Ballina L R, Burghardt H E, Ferrer-Costa C, Turnay J, Ferrer-Orta C, et al. The structure of human 4F2hc ectodomain provides a model for homodimerization and electrostatic interaction with plasma membrane. The Journal of biological chemistry. 2007; 282: 31444-52. [0402] 56. Emsley P, Lohkamp B, Scott W G, Cowtan K. Features and development of Coot. Acta Crystallographica Section D Biological Crystallography. 2010; 66: 486-501. [0403] 57. Murshudov G N, Skubak P, Lebedev A A, Pannu N S, Steiner R A, Nicholls R A, et al. REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallographica Section D Biological Crystallography. 2011; 67: 355-67. [0404] 58. Painter J, Merritt E A. TLSMD web server for the generation of multi-group TLS models. J Appl Crystallogr. 2006; 39: 109-11. [0405] 59. Barkovskiy M, Ilyukhina E, Dauner M, Eichinger A, Skerra A. An engineered lipocalin that tightly complexes the plant poison colchicine for use as antidote as well as bioanalytical applications. Biological chemistry. 2018. [0406] 60. Binder U, Matschiner G, Theobald I, Skerra A. High-throughput sorting of an Anticalin library via EspP-mediated functional display on the Escherichia coli cell surface. Journal of molecular biology. 2010; 400: 783-802. [0407] 61. Costa J, Grabenhorst E, Nimtz M, Conradt H S. Stable expression of the Golgi form and secretory variants of human fucosyltransferase III from BHK-21 cells. The Journal of biological chemistry. 1997; 272: 11613-21. [0408] 62. Schlehuber S, Skerra A. Tuning ligand affinity, specificity, and folding stability of an engineered lipocalin variant—a so-called ‘anticalin’—using a molecular random approach. Biophys Chem. 2002; 96: 213-28. [0409] 63. Vosjan M J, Perk L R, Visser G W, Budde M, Jurek P, Kiefer G E, et al. Conjugation and radiolabeling of monoclonal antibodies with zirconium-89 for PET imaging using the bifunctional chelate p-isothiocyanatobenzyl-desferrioxamine. Nat Protoc. 2010; 5: 739-43. [0410] 64. Wollscheid B, Bausch-Fluck D, Henderson C, O'Brien R, Bibel M, Schiess R, et al. Mass-spectrometric identification and relative quantification of N-linked cell surface glycoproteins Nature biotechnology. 2009; 27: 378-86. [0411] 65. Barinka C, Ptacek J, Richter A, Novakova Z, Morath V, Skerra A. Selection and characterization of Anticalins targeting human prostate-specific membrane antigen (PSMA) Protein engineering, design & selection: PEDS. 2016; 29: 105-15. [0412] 66. Beckett D, Kovaleva E, Schatz P J. A minimal peptide substrate in biotin holoenzyme synthetase-catalyzed biotinylation Protein science: a publication of the Protein Society. 1999; 8: 921-9. [0413] 67. Schiefner A, Gebauer M, Richter A, Skerra A. Anticalins reveal high plasticity in the mode of complex formation with a common tumor antigen. Structure (London, England: 1993). 2018; 26: 649-56.e3. [0414] 68. Lippincott J, Van der Horst E T H, Zachwieja J, Tran H. Anti-CD98 antibodies and methods of use thereof. In: Organization WIP, editor.; 2013. [0415] 69. Matthews B W, Nicholson H, Becktel W J. Enhanced protein thermostability from site-directed mutations that decrease the entropy of unfolding. Proc Natl Acad Sci USA. 1987; 84: 6663-7. [0416] 70. Binder U, Skerra A. PASylation®: A versatile technology to extend drug delivery. Curr Opin Colloid Interface Sci. 2017; 31: 10-7. [0417] 71. Papetti M, Herman I M. Controlling tumor-derived and vascular endothelial cell growth: Role of the 4Ff2 cell surface antigen. Am J Pathol. 2001; 159: 165-78. [0418] 72. Fischer G, Seibold U, Schirrmacher R, Wangler B, Wangler C. (89)Zr, a radiometal nuclide with high potential for molecular imaging with PET: Chemistry, applications and remaining challenges. Molecules (Basel, Switzerland). 2013; 18: 6469-90. [0419] 73. Abou D S, Ku T, Smith-Jones P M. In vivo biodistribution and accumulation of 89Zr in mice Nuclear medicine and biology. 2011; 38: 675-81. [0420] 74. Richter A, Skerra A. Anticalins directed against vascular endothelial growth factor receptor 3 (VEGFR-3) with picomolar affinities show potential for medical therapy and in vivo imaging. Biological chemistry. 2017; 398: 39-55. [0421] 75. Rauth S, Hinz D, Borger M, Uhrig M, Mayhaus M, Riemenschneider M, et al. High-affinity Anticalins with aggregation-blocking activity directed against the Alzheimer beta-amyloid peptide. The Biochemical journal. 2016; 473: 1563-78. [0422] 76. Mendler C T, Gehring T, Wester H J, Schwaiger M, Skerra A. (89)Zr-labeled versus (124)I-labeled αHER2 Fab with optimized plasma half-life for high-contrast tumor imaging in vivo. Journal of nuclear medicine: official publication, Society of Nuclear Medicine. 2015; 56: 1112-8. [0423] 77. Summer D, Garousi J, Oroujeni M, Mitran B, Andersson K G, Vorobyeva A, et al. Cyclic versus noncyclic chelating scaffold for (89)Zr-labeled ZEGFR:2377 Affibody bioconjugates targeting epidermal growth factor receptor overexpression. Molecular pharmaceutics. 2018; 15: 175-85. [0424] 78. Summer D, Rangger C, Klingler M, Laverman P, Franssen G M, Lechner B E, et al. Exploiting the concept of multivalency with (68)Ga- and (89)Zr-labeled Fusarinine C-Minigastrin Bioconjugates for targeting CCK2R expression. Contrast media & molecular imaging. 2018; 2018: 3171794.