Patent: 10,100,106

Summary for Patent: 10,100,106

Title:	Single domain serum albumin binding protein
Abstract:	Disclosed herein are single domain serum albumins binding proteins with improved thermal stability, binding affinities, and robust aggregation profiles. Also described are multispecific binding proteins comprising a single domain serum albumin binding protein according to the instant disclosure. Pharmaceutical compositions comprising the binding proteins disclosed herein and methods of using such formulations are provided.
Inventor(s):	Dubridge; Robert B. (Belmont, CA), Lemon; Bryan D. (Mountain View, CA), Austin; Richard J. (San Francisco, CA), Evnin; Luke (San Francisco, CA), Guenot; Jeanmarie (San Francisco, CA)
Assignee:	HARPOON THERAPEUTICS, INC. (South San Francisco, CA)
Application Number:	15/600,582
Patent Claims:	see list of patent claims

United States Patent 10,100,106: claims, claim scope, and US patent landscape for single-domain serum albumin binding proteins (CDR-engineered)

Executive summary. US Patent 10,100,106 claims a single-domain serum albumin binding protein defined by (i) CDR1/2/3 amino-acid sequence constraints (SEQ ID NOs. 1–3), (ii) preferred compositions built from framework residues to yield proteins ≥80% identical to SEQ ID NO. 10, (iii) narrower embodiments specifying permitted r1/r2/r3 SEQ ID NO combinations, and (iv) additional limitations on cross-species binding affinity ratios, serum elimination half-life, and specific residue substitutions relative to a “wt anti-HSA” scaffold (SEQ ID NO. 10). Downstream risk for competitors hinges on whether their candidates land inside the CDR sequence “allowed-variable” grammar and/or the listed SEQ ID NO embodiments. The patent also reaches therapeutic use and multispecific fusions via dependent claims.

---

What is US Patent 10,100,106 claiming about single-domain anti-HSA binding proteins?

Short answer. It covers engineered single-domain antibody-like proteins (or nanobody/variable single-domain format) that bind serum albumin (HSA and cynomolgus/mouse, per claim language), with binding and pharmacokinetic performance tied to CDR-level design, framework backbones, and in vivo elimination half-life thresholds.

Core claim 1: CDR grammar that drives literal infringement

Claim 1 defines a “single domain serum albumin binding protein” with:

• CDR1 sequence (SEQ ID NO. 1) pattern:
  G F X1 X2 X3 X4 F G M S
  with variable constraints:

  • X1 ∈ {Thr, Arg, Lys, Ser, Pro}
  • X2 ∈ {Phe, Tyr}
  • X3 ∈ {Ser, Arg, Lys}
  • X4 ∈ {Ser, Lys, Arg, Ala}

• CDR2 sequence (SEQ ID NO. 2) pattern:
  S I S G S G X5 X6 T L Y A X7 S X8 K
  with constraints:

  • X5 ∈ {Ser, Arg, Thr, Ala}
  • X6 ∈ {Asp, His, Val, Thr}
  • X7 ∈ {Asp, His, Arg, Ser}
  • X8 ∈ {Val, Leu}

• CDR3 sequence (SEQ ID NO. 3) pattern:
  G G S L X9 X10
  with constraints:

  • X9 ∈ {Ser, Arg, Thr, Lys}
  • X10 ∈ {Arg, Lys, Val, Pro, Asn}

• A final anti-pattern exclusion stating that X1–X10 are not simultaneously (Thr, Phe, Ser, Ser, Ser, Asp, Asp, Val, Ser, Arg) in the respective variables. This is a specific “all-at-once” carve-out rather than a broad functional limitation.

Critical scope implication. Claim 1 is literal-sequence anchored. A competitor’s infringement analysis for claim 1 is essentially:

1. map their CDR1/2/3 to the same positional template; 2) check each position’s permitted amino-acid set; and 3) ensure the exact “simultaneous prohibited combination” is not present.

Claim 2: framework residues + “≥80% identical to SEQ ID NO. 10”

Claim 2 narrows claim 1 into a particular architecture:

• Protein is represented as f1-r1-f2-r2-f3-r3-f4
  • r1 = SEQ ID NO. 1
  • r2 = SEQ ID NO. 2
  • r3 = SEQ ID NO. 3
• Framework residues (f1–f4) are selected so that the full protein is at least 80% identical to SEQ ID NO. 10.

Critical scope implication. This introduces a second axis beyond CDR compliance: overall identity to SEQ ID NO. 10. A competitor could satisfy CDR constraints but fail the ≥80% identity requirement if their framework diverges too far.

Claims 3–14: permitted r1/r2/r3 combinations (combinatorial narrowing)

Claims 3–14 specify alternative choices for r1/r2/r3 as allowed SEQ ID NOs. These are “named embodiments” that can act as a strong infringement beacon: if a product’s CDR sets correspond to one of those enumerated r1/r2/r3 assignments, claim coverage becomes easier to argue.

• Claim 3: r1 is SEQ ID NO. 14 or 15 or 16
• Claim 4: r2 is SEQ ID NO. 17–22
• Claim 5: r3 is SEQ ID NO. 23 or 24
• Claims 6–14 further constrain which combinations occur, e.g.
  • Claim 6: r1 = SEQ ID NO. 14
  • Claim 7: r1 = 15; r2 = 17; r3 = 23
  • Claim 10: r1 = 14; r3 = 23
  • Claim 11: r1 = 15; r2 = 19; r3 = 24
  • Claim 12: r1 = 14; r2 = 20
  • Claim 13: r1 = 15; r2 = 21
  • Claim 14: r1 = 15; r2 = 22; r3 = 24

Critical scope implication. These dependent claims may be used in litigation either as:

• alternative claim paths if independent claim 1/2 are contested, or
• a fallback to specific CD R-configuration subsets.

Claims 15–20: specific full-sequence embodiments (SEQ ID NOs. 4–9, 25–27)

Claim 15 lists proteins by full sequences:

• SEQ ID NO. 4, 5, 6, 7, 8, 9, 25, 26, 27.

Claims 16–20 pull out specific ones:

• Claim 16: SEQ ID NO. 4
• Claim 17: SEQ ID NO. 7
• Claim 18: SEQ ID NO. 9
• Claim 19: SEQ ID NO. 26
• Claim 20: SEQ ID NO. 27

Critical scope implication. These “hard-coded” sequences are usually the easiest to match to a candidate sequence. In practice, product teams often publish sequences, patent filings, or deposition sequences that can map directly to these SEQ IDs.

---

What “cross-species binding” and “half-life” limits tighten infringement risk?

Claim 21–22: mouse weaker than human/cynomolgus; hKd/cKd ratio

Claim 21: binds mouse serum albumin with Kd about 1.5 to 20-fold weaker than binding to human and cynomolgus serum albumin.

Claim 22: binds human and cynomolgus with hKd and cKd such that

• hKd/cKd ranges from about 20:1 to about 1:2.

Critical scope implication. These are quantitative functional constraints. They matter in two ways:

1. They can narrow coverage to a performance profile rather than pure sequence grammar.
2. They raise evidence burdens in litigation if the competitor’s sequence could arguably be within the CDR/frame sets but their measured binding differs.

Claim 23: elimination half-time thresholds

Claim 23 adds pharmacokinetics: the protein has an elimination half-time of at least thresholds:

• ≥12h, ≥20h, ≥25h, ≥30h, ≥35h, ≥40h, ≥45h, ≥50h, or ≥100h.

Critical scope implication. This may capture albumin-binding proteins with extended systemic exposure, but it can also create “escape” for variants that bind albumin yet clear faster (e.g., altered stability or dosing-dependent differences). The claim language suggests a menu of minimum half-life values, meaning coverage attaches if the actual half-life is at or above the stated threshold(s).

---

Which specific residue substitutions relative to “wt anti-HSA” are protected?

Claim 24: residue-level substitution map at defined positions in CDR1/2/3

Claim 24 is a targeted variation claim using a baseline SEQ ID NO. 10 (wt anti-HSA) and substituting residues at positions:

• CDR1 positions 28–31 (one or more substituted) with specified allowed residues for each position:

  • position 28 → {Arg, Lys, Ser, Pro}
  • position 29 → {Tyr}
  • position 30 → {Arg, Lys}
  • position 31 → {Lys, Arg, Ala}

• CDR2 positions 56, 57, 62, 64 substitutions:

  • 56 → {Arg, Thr, Ala}
  • 57 → {His, Val, Thr}
  • 62 → {His, Arg, Glu, Ser}
  • 64 → {Leu}

• CDR3 positions 103 and 104 substitutions:

  • 103 → {Arg, Thr, Lys}
  • 104 → {Lys, Val, Pro, Asn}

Critical scope implication. This claim is often valuable because it is position-number anchored. If a competitor uses the same wt scaffold and performs substitutions within those allowed sets at those numbered positions, claim 24 can be easier to test than the more abstract SEQ ID grammar in claim 1.

---

What therapeutic and platform claims extend coverage beyond the binding protein itself?

Claim 25: methods for proliferative/tumorous/inflammatory/immunologic and other diseases

Claim 25 covers treatment/amelioration of a broad set of indications including:

• proliferative/tumorous
• inflammatory
• immunological/autoimmune
• infectious/viral
• allergic reactions
• parasitic reactions
• graft-versus-host and host-versus-graft disease

by administering the protein of claim 1.

Critical scope implication. The breadth of indications means competitors face method claims if they use the same albumin-binding single-domain in therapeutic regimens for any covered disease class. However, enforceability depends on how their product is actually used and whether the protein itself fits claim 1 (or a dependent claim path).

Claim 26: multispecific binding proteins containing the albumin binder

Claim 26 extends to multispecific binding proteins that include the single-domain albumin binding protein of claim 1.

Critical scope implication. This is a common platform-style coverage hook: if a competitor fuses the albumin binder to another binding arm (e.g., another antibody specificity), claim 26 can reach the assembled multispecific.

---

How strong is the patent estate around US 10,100,106 likely to be, claim-by-claim?

Short answer. Strength is concentrated in (i) explicit CDR sequence constraints, (ii) the ≥80% identity framework requirement, and (iii) quantified binding and half-life limits, with additional enumerated SEQ ID embodiments providing litigation-ready “maps” to specific designs. Weaknesses typically arise from (a) potential prior art showing similar anti-HSA CDR engineering and (b) claim ambiguity risks around identity calculation methods or assay-defined Kd/half-life.

Claim 1: High structural determinacy due to CDR templates and explicit amino-acid sets.
Claim 2: Moderately strong due to added ≥80% identity requirement, but identity percent disputes are a typical litigation flashpoint.
Claims 3–14: Stronger for enforcement because they fence to specific SEQ ID combinations.
Claims 21–23: Provide functional narrowing but require assay comparability; these limits can also create evidentiary friction.
Claim 24: Practical strength if competing sequences are built on the same numbering system and scaffold.

---

What would be the most likely infringement and invalidity angles for competitors?

Infringement map: sequence-first screening

For a candidate anti-HSA single-domain binder:

1. Align CDR1/2/3 to SEQ ID NO templates (or positional numbering used in the patent).
2. Check allowed residues at each X position (claim 1) and ensure the excluded simultaneous combination is not present.
3. Evaluate framework identity versus SEQ ID NO. 10 for the claim-2 path.
4. Check whether the candidate is one of the enumerated full sequences (SEQ ID NOs. 4–9, 25–27) for the easiest claim path.
5. If those match, assess functional constraints (mouse vs human/cyno Kd; half-life) for claim 21–23 and 23 thresholds, and residue substitution constraints for claim 24.

Validity map: potential prior art categories (what matters legally)

Even without enumerating every family member here, the patent’s claim structure suggests prior art will be evaluated in three buckets:

• Existing anti-albumin single-domain scaffolds with similar CDR engineering
• Framework choices that maintain binding but differ in sequence identity
• Pharmacokinetic tuning (albumin binder half-life extension) with established measurement methods

The more a competitor can show that anti-HSA binders with these CDR patterns and/or wt scaffold substitutions were already disclosed with predictable performance, the higher the invalidity pressure.

---

US regulatory and Orange Book status: does it affect this patent?

No direct FDA Orange Book listing can be assumed from the patent number alone, and this asset is a biologic-like binding protein patent rather than a small-molecule drug patent typically listed in the Orange Book. The relevant exclusivity framework is usually tied to the biologic’s license and any data exclusivity plus patent listings in the Biologics License Application context, not Orange Book.

---

Key takeaways

• US 10,100,106 is a sequence-defined patent focused on single-domain serum albumin binding proteins with CDR1/2/3 constraints and a ≥80% identity framework requirement relative to SEQ ID NO. 10.
• Dependent claims enumerate allowed r1/r2/r3 SEQ ID combinations and list specific full-length SEQ IDs, creating straightforward infringement “checkpoints.”
• The patent adds quantitative functional narrowing (cross-species Kd relationships and elimination half-life) and a position-based substitution claim (claim 24) tied to the “wt anti-HSA” scaffold.
• Therapeutic method coverage is broad across disease classes, and the platform is extended to multispecific formats.

---

FAQs

1. How do courts typically interpret “at least eighty percent identical” to a reference SEQ ID in albumin binder claims?
2. If a candidate matches claim 1 CDR sequences but misses framework identity, does claim 2 still apply?
3. What evidentiary standards usually decide whether Kd and half-life thresholds are met for functional-dependent antibody claims?
4. How does claim 24’s position-number substitution map reduce design-around opportunities compared with claim 1’s X-position grammar?
5. In multispecific formats, when does including an albumin-binding domain trigger claim 26 versus falling outside due to formatting or chain composition?

---

References (APA)

1. United States Patent 10,100,106 (claims text provided in prompt).

Details for Patent 10,100,106

Applicant	Tradename	Biologic Ingredient	Dosage Form	BLA	Approval Date	Patent No.	Expiredate
Grifols Therapeutics Llc	ALBUKED, PLASBUMIN-20, PLASBUMIN-25, PLASBUMIN-5	albumin (human)	For Injection	101138	October 21, 1942	10,100,106	2037-05-19
Takeda Pharmaceuticals U.s.a., Inc.	BUMINATE, FLEXBUMIN	albumin (human)	Injection	101452	March 03, 1954	10,100,106	2037-05-19
Csl Behring Ag	ALBURX	albumin (human)	Injection	102366	July 23, 1976	10,100,106	2037-05-19