Serum albumin binding immunoglobulin single variable domains

Abstract

The present invention relates to amino acid sequences that can bind to serum albumin. In particular, the present invention relates to immunoglobulin single variable domains, and in particular heavy-chain immunoglobulin single variable domains, that can bind to serum albumin. The invention also relates to proteins, polypeptides and other constructs, compounds, molecules or chemical entities that comprise at least one of the immunoglobulin single variable domains binding to serum albumin that are described herein.

Claims

1. An immunoglobulin single variable domain (ISVD) capable of binding to serum albumin that comprises: a CDR1 (according to Abm) that is the amino acid sequence GLTFSSYAMG (SEQ ID NO: 12); and a CDR2 (according to Abm) that is the amino acid sequence SISRGGGYTY (SEQ ID NO: 13); and a CDR3 (according to Abm) that is the amino acid sequence ARYWATGSEYEFDY (SEQ ID NO: 14), optionally wherein the serum albumin is human serum albumin.

2. The ISVD according to claim 1, that can bind to human serum albumin with an affinity better than 100 nM or better than 50 nM, as determined by surface plasmon resonance.

3. The ISVD according to claim 1, that is a heavy-chain immunoglobulin single variable domain.

4. The ISVD according to claim 1, that is a VHH, a humanized VHH or a camelized VH, optionally wherein the camelized VH is a camelized human VH.

5. The ISVD according to claim 1, wherein the ISVD has a serum half-life in human (expressed as t1/2 beta) that is more than 6 hours, 12 hours, 24 hours, or 72 hours.

6. The ISVD according to claim 1, that has: a degree of sequence identity with the sequence of SEQ ID NO: 15 of at least 85%, at least 90%, or at least 95%; or that has: no more than 7, no more than 5, no more than 3, no more than 2, or no more than 1 amino acid differences with the sequence of SEQ ID NO: 15; wherein the CDRs are not taken into account for determining the degree of sequence identity or the amino acid differences.

7. The ISVD according to claim 1, that is a VHH and that contains, compared to the sequence of SEQ ID NO:15, one or more humanizing substitutions.

8. An immunoglobulin single variable (ISVD) domain capable of binding to serum albumin and that is chosen from SEQ ID NO's: 15 to 77, optionally wherein the serum albumin is human serum albumin.

9. A polypeptide that comprises at least one ISVD according to claim 1.

10. The polypeptide according to claim 9, that comprises at least one therapeutic moiety or entity.

11. The polypeptide according to claim 9, that is a fusion protein.

12. The polypeptide according to claim 9, that has a serum half-life in human (expressed as t1/2 beta) that is more than 6 hours, more than 12 hours, more than 24 hours, or more than 72 hours.

13. The polypeptide according to claim 9, wherein: the polypeptide further comprises a C-terminal ISVD, wherein the C-terminal ISVD polypeptide comprises a C-terminal extension (X)n, wherein each X is an amino acid residue that is independently chosen from naturally occurring amino acid residues, and wherein n is 1-10; and/or the polypeptide further comprises a N-terminal ISVD, wherein the N-terminal ISVD polypeptide comprises a D at position 1.

14. A pharmaceutical composition comprising a polypeptide according to claim 9.

15. A nucleic acid that encodes an ISVD according to claim 1, optionally wherein the nucleic acid is a genetic construct.

16. An expression vector that comprises the nucleic acid of claim 15.

17. A host cell that comprises an expression vector according to claim 16.

18. A method for preparing a polypeptide comprising cultivating or maintaining a host cell according to claim 17 under conditions such that said host cell produces or expresses the polypeptide, and optionally further comprising isolating the polypeptide.

19. The polypeptide according to claim 10, that has a serum half-life in human (expressed as t1/2 beta) that is more than 6 hours, more than 12 hours, more than 24 hours, or more than 72 hours.

20. The ISVD according to claim 1, wherein the ISVD cross-blocks the binding of the amino acid sequence of SEQ ID NO: 15 to human serum albumin.

Description

(1) The invention will now be further described by means of the following non-limiting preferred aspects, examples and figures, in which:

(2) FIG. 1 is a table listing some of the amino acid positions that will be specifically referred to herein and their numbering according to some alternative numbering systems (such as Aho and IMGT);

(3) FIG. 2 lists the amino acid sequences referred to herein;

(4) FIGS. 3A and 3B show an alignment of the sequence of SEQ ID NO:15 (invention) with the prior art sequences of SEQ ID NOs: 1 and 2 and the reference sequences of SEQ ID NO: 79 and 80 (which are based on SEQ ID NO:1 and SEQ ID NO:2, respectively);

(5) FIG. 4A shows an alignment of SEQ ID NOs: 15 to 56 and FIG. 4B shows an alignment of SEQ ID NOs: 15 and 57 to 77

(6) FIG. 5 is a graph showing competitive binding for the serum albumin binder of SEQ ID NO:1 (reference), the serum albumin binder of SEQ ID NO:15 (invention) and an irrelevant Nanobody (cAblys3-Flag3His6), as generated in Example 2;

(7) FIG. 6 is a graph showing binding of human serum albumin to FcRn in the presence of the serum albumin binder of SEQ ID NO:15. Human FcRn-human β2 microglobulin heterodimer was immobilized on CMS chip. Binding of 1 μM HSA in absence or presence of 2 μM Nanobody in 50 mM NaPO4+150 mM NaCl+0.05% Tween-20 pH 6.0 was monitored on a Biacore T100 instrument;

(8) FIG. 7 shows the data collection and processing statistics used in Example 7 in determining/calculating the crystal structure for the crystal structure of human serum albumin and the amino acid sequence of SEQ ID NO:15;

(9) FIG. 8 lists the refinement statistics used for used in Example 7 in determining/calculating the crystal structure for the crystal structure of human serum albumin and the amino acid sequence of SEQ ID NO:15;

(10) FIG. 9 is a graph showing serum concentrations for the constructs of SEQ ID NO:82 (invention), SEQ ID NO:89 (reference) and SEQ ID NO:90 (reference), respectively, as determined in Example 5. The symbols in the graph represent individual data points and the lines represent the mean concentration.

(11) FIG. 10 shows the most important residues on human serum albumin and on SEQ ID NO: 15, respectively, that based on the crystal structure data generated in Example 7 are assumed to be involved in the binding interaction between human serum albumin and SEQ ID NO: 15. FIG. 10 also shows the main binding interactions between these respective amino acid residues.

(12) FIG. 11 gives an alignment of serum albumin from different species of mammal. In the sequence of human serum albumin, the stretches of amino acids that are assumed to be part of the putative epitope of SEQ ID NO:15 (see also FIG. 10) have been highlighted.

(13) The entire contents of all of the references (including literature references, issued patents, published patent applications, and co pending patent applications) cited throughout this application are hereby expressly incorporated by reference, in particular for the teaching that is referenced hereinabove.

EXPERIMENTAL PART

Example 1: Affinity for Serum Albumin

(14) The affinity of the serum albumin binder of SEQ ID NO:15 for human (Sigma-Aldrich A3782), cynomolgus monkey (generated in-house), mouse (Albumin Bioscience 2601), rat (Sigma-Aldrich A4538), rabbit (Sigma-Aldrich A0764), guinea pig (Gentaur GPSA62), pig (Sigma-Aldrich A4414), sheep (Sigma-Aldrich A3264) and bovine (Sigma-Aldrich A3059) serum albumin (SA) was measured via Surface Plasmon Resonance (SPR) on a ProteOn XPR36 (BioRad) instrument. Serum albumin was immobilized via amine coupling on GLC ProteOn chip using ProteOn Amine Coupling Kit (BioRad). Different concentrations (300 nM, 100 nM, 33.3 nM, 11.1 nM, 3.7 nM and 1.23 nM) of the serum albumin binder of SEQ ID NO:15 were injected in HBS-P+pH 7.4 buffer (GE Healthcare) at 45 μL/min for 120 s, followed by dissociation for 900 s. There was no or very low binding observed for the serum albumin binder of SEQ ID NO:15 on rabbit, pig, sheep and bovine SA. The affinity of the serum albumin binder of SEQ ID NO:15 for human, cynomolgus monkey, rat, mouse and guinea pig SA was higher compared to the respective affinities of serum albumin binder of SEQ ID NO:1 (reference) as determined in a separate experiment. The results are shown in Table 1.

(15) TABLE-US-00004 TABLE 1 Kinetic parameters for binding of the serum albumin binder of SEQ ID NO: 15 on SA from different species. SEQ ID NO: 15 SEQ ID NO: 1 (reference) SA ka (s−.sup.1M−.sup.1) kd (s−.sup.1) KD (M) ka (s−.sup.1M−.sup.1) Kd (s−.sup.1) KD (M) human 8.1E+05 1.5E−04 1.9E−10 4.9E+05 1.6E−03 3.3E−09 cyno 7.7E+05 9.4E−05 1.2E−10 4.6E+05 1.4E−03 3.1E−09 rat 1.0E+06 2.2E−02 2.1E−08 3.9E+05 2.6E−01 6.7E−07 mouse 1.2E+06 2.4E−03 1.9E−09 6.6E+05 3.0E−02 3.9E−08 guinea 1.4E+06 2.1E−03 1.5E−09 9.4E+05 1.9E−02 2.0E−08 pig

(16) The long half-life of albumin in blood is mainly driven by two characteristics: (i) the large size (65 kDa) of albumin limits its glomerular filtration and (ii) albumin binds to FcRn at low pH (pH 6), which protects albumin from degradation in the lysosomes after passive endocytosis in endothelial and epithelial cells, by recycling from early endosome back to the extracellular environment. For albumin-binding Nanobodies to result in long serum half-life through albumin binding and subsequent recycling, these should stay bound to albumin in the pH range from 5.0 to 7.4. The dissociation rate of the serum albumin binder of SEQ ID NO:15 from HSA at pH 5, pH 6 and pH 7.4 was measured on a ProteOn instrument as described above, including serum albumin binder of SEQ ID NO:1 as a reference. The serum albumin binder of SEQ ID NO:15 and serum albumin binder of SEQ ID NO:1 (reference) were injected at 500 nM and 300 nM respectively in HBS-P+pH 7.4 buffer. Dissociation buffers were 50 mM NaOAc/HOAc+150 mM NaCl+0.05% Tween-20 pH 5.0, 50 mM NaOAc/HOAc+150 mM NaCl+0.05% Tween-20 pH 6.0 and HBS-P+pH 7.4 respectively. Dissociation was analysed for 2700 s. As can be seen from the data shown in Table 2, the dissociation rates for the serum albumin binder of SEQ ID NO:15 do not differ significantly across the pH range from 5.0 to 7.4.

(17) TABLE-US-00005 TABLE 2 Dissociation rate of T023500010 from HSA at different pH. kd (s.sub.−1) SEQ ID NO: 15 SEQ ID NO: 1 (reference) pH 7.4 1.1E−04 1.3E−03 pH 6.0 6.9E−05 9.2E−04 pH 5.0 5.8E−05 1.1E−03 SEQ ID NO: 15 or SEQ ID NO: 1 were injected on immobilized human serum albumin. Dissociation was monitored at pH 5.0, 6.0 and 7.4 on a ProteOn instrument.

Example 2: Epitope

(18) Epitope binning was analysed in a competition ELISA. Human serum albumin was coated at 125 ng/ml in PBS at 4° C. over night. After blocking with PBS+1% casein, 1.5 nM serum albumin binder of [SEQ ID NO:1-cMycHis6] and a concentration series of competitors ([His6Flag3-SEQ ID NO:15], [His6Flag3-SEQ ID NO:1] as positive control or hen egg lysozyme binding single domain antibody cAblys3-Flag3His6 as negative control) were added. Bound [SEQ ID NO:1-cMycHis6] was detected with goat anti-cMyc (Abcam ab19234) and HRP-labelled rabbit anti-goat (Genway 18-511-244226) antibodies.

(19) The results are shown in FIG. 5. It was found that the serum albumin binder of SEQ ID NO:15 and serum albumin binder of SEQ ID NO:1 (reference) do not bind identical epitopes on human serum albumin.

Example 3: Interference with Interaction Between SA and FcRn

(20) For the serum albumin binder of SEQ ID NO:15 to result in long half-life via albumin binding and subsequent recycling, it should not interfere with the binding of albumin to FcRn. This was analysed in SPR on a Biacore T100 (GE Healthcare) instrument. Human FcRn-human β2 microglobulin heterodimer (Sino Biological CT009-H08H) was immobilized on CMS chip via standard amine coupling (Biacore amine coupling kit). A mixture of 1 μM HSA and 2 μM Nanobody ([His6Flag3-SEQ ID NO:15], [His6Flag3-SEQ ID NO:1] (reference) or cAblys3-Flag3His6) in 50 mM NaPO4+150 mM NaCl+0.05% Tween-20 pH 6.0 was injected at 10 μl/min for 120 s, followed by dissociation for 600 s. Binding curves were qualitatively compared with binding curve of 1 μM HSA in absence of Nanobody. As can be seen from FIG. 6, the serum albumin binder of SEQ ID NO:15 did not interfere with the binding of HSA to FcRn.

Example 4: Physical Stability

(21) The stability of the serum albumin binder of SEQ ID NO:15 was assessed, using the serum albumin binders of SEQ ID NOs:1 and 2 as reference. Melting temperature (Tm) was determined in Differential Scanning calorimetry (DSC). In addition, the physical stability was analysed by measuring the following parameters before and after storage at 40° C. in D-PBS at 5 mg/ml: Turbidity (OP.sub.500nm), percentage high molecular weight variants (SE-HPLC), content (OD280) and chemical variants (RP-HPLC).

(22) The results are shown in Table 3. For all constructs, storage at 40° C. resulted in an increase in pre-peak in SE-HPLC (high molecular weight variants), which was clearly lower for the serum albumin binder of SEQ ID NO:15 and that of SEQ ID NO:2 compared to serum albumin binder of SEQ ID NO:1.

(23) TABLE-US-00006 TABLE 3 Summary data physical stability the serum albumin binder of SEQ ID NO: 15. Stability 40° C. 2 weeks Δ pre-peak Turbidity at Protein Tm SE-HPLC (%) T = 2 w (OD.sub.500 nm) loss (° C.) SEQ ID NO: 15 0.3 0.01 no 72.4 SEQ ID NO: 1 11.8 0.1 no 66.0 (reference) SEQ ID NO: 2 0.6 0.01 no 72.0 (reference) Tm was measured in DSC. The serum albumin binders were stored for 2 weeks at 40° C. at a concentration of 5 mg/mL in PBS. Turbidity, SE-HPLC profile and protein content (RP-UHPLC) were measured before and after storage.

Example 5: PK Profile of the Serum Albumin Binder of SEQ ID NO:15 in Rat

(24) The pharmacokinetics of a representative compound of the invention (SEQ ID NO:82) comprising the serum albumin binder of SEQ ID NO:15 after single i.v. dose were studied in Sprague Dawley rats and compared to similar constructs (SEQ ID NOs: 89 and 90) comprising the reference serum albumin binders of SEQ ID NOs: 79 and 80, respectively. Alignments of SEQ ID NO:15 with the reference sequences of SEQ ID NOs: 1, 2, 79 and 80 are given in FIGS. 3A and 3B. The constructs of SEQ ID NOs: 82, 89 and 90 comprise the relevant serum albumin binder linked via a 35GS linker to a representative Nanobody (the anti-HER2 Nanobody of SEQ ID NO:81) as well as an E1D mutation and a C-terminal alanine, and were produced in Pichia pastoris.

(25) Radio-iodination of the bivalent Nanobodies was conducted using N-succinimidyl 3-.sup.125I-iodobenzoate (.sup.125I-SIB) in borate buffer 0.2 M pH 8.3. There was no indication that labelling of the Nanobodies interfered with binding to HSA. Rats were dosed with 20 μg of the .sup.125I-labelled construct (3 rats per group) at a specific activity of 3.5-4 mCi/mg Nanobody construct. Blood samples were taken at 5 min, 1 h, 4 h, 8 h, 24 h, 48 h, 96 h, 168 h, 240 h, 336 h and 504 h post dosing. Radioactivity was measured in each blood sample and converted to a protein concentration based on the specific activity of the labelled Nanobody construct. The decay of the radioactive label over time was taken into account in the calculations. The measured concentrations of the constructs in blood over time are shown in FIG. 9. PK parameters were calculated by non-compartmental analysis: the relevant data are listed in Table 4). In line with the expected higher affinity for rat SA compared to the references of SEQ ID NOs: 79 and 89, a higher exposure and reduced clearance was observed for the serum albumin binder of SEQ ID NO:15.

(26) TABLE-US-00007 TABLE 4 Results from PK profiling in rats. KD on PK parameters rat SA V.sub.ss Cl MRT T.sub.1/2eff Construct (M) (mL) (mL/h) (h) (h) SEQ ID NO: 82  2.0E−08 33 0.75 44.1 30.6 (invention) SEQ ID NO: 89 >1.0E−07 45.6 4.22 11.4 7.9 (reference) SEQ ID NO: 90 >1.0E−07 38.3 1.81 21.8 15 (reference) KD on rat SA was determined in SPR. The constructs of SEQ ID NOs: 82 (invention) and 89 and 90 (reference) were injected at different concentrations on immobilized rat SA on a ProteOn instrument. Binding and dissociation were analysed at pH 7.4. PK parameters were calculated from non-compartmental analysis of blood concentration over time after single i.v. injection of 20 μg of the Nanobodies. V.sub.ss: volume of distribution at steady state; Cl: clearance; MRT: Mean Residence Time; T.sub.1/2eff: effective half-life calculated from MRT (T.sub.1/2eff = ln2*MRT).

Example 6: In Vivo Safety of the Serum Albumin Binder of SEQ ID NO:15 in Rat

(27) Albumin is a carrier protein for many natural ligands, such as bilirubin, lipids, ions, sugars, metabolites. To use the serum albumin binder of SEQ ID NO:15 for half-life extension of therapeutic compounds, it should not displace binding of natural ligands. This was assessed in a safety study in Crl:CD(SD) rats. Animals were injected i.v. with 100 mg/kg of the construct of SEQ ID NO: 82 or vehicle (D-PBS) on day 1, 4 and 7. Blood was collected on day 4, 7 and 12 and clinical parameters were measured. The compound of the invention was well tolerated and did not result in any adverse clinical observation, food consumption, body weight, or clinical chemistry changes.

Example 7: Crystal Structure

(28) Crystals of human serum albumin in complex with the serum albumin binder of SEQ ID NO:15 were flash-frozen and measured at a temperature of 100 K. Diffraction data for the co-crystallized complex were collected at the SWISS LIGHT SOURCE (Villigen, Switzerland). Data collection and processing statistics are summarized FIGS. 7 and 8.

(29) The resulting electron density maps reveal that the crystals contain one HSA: the serum albumin binder of SEQ ID NO:15 complex in the asymmetric unit and show the unambiguous binding mode for the serum albumin binder of SEQ ID NO:15, binding to domain II of HSA. A structural model was constructed and refined to a final resolution of 2.80 Å. The model comprises residues Glu1 to Ser123 of the serum albumin binder of SEQ ID NO:15 and Lys4 to Leu583 of HSA. The main residues that were found to be involved in the interaction of the serum albumin binder of SEQ ID NO:15 with HSA are listed in FIG. 10 (see also FIG. 11).