Patent: 10,112,947

Summary for Patent: 10,112,947

Title:	Substituted 6-aminopurines for targeting HSP90
Abstract:	Described herein are 6-aminopurine compounds comprising formula (III) that may selectively bind to Hsp90, methods of using the compounds, and kits including the compounds. Formula (III) may link to detection moieties such as fluorophores that may allow for selective detection of Hsp90 in a sample. ##STR00001##
Inventor(s):	Haystead; Timothy (Chapel Hill, NC), Hughes; Philip Floyd (Chapel Hill, NC)
Assignee:	Duke University (Durham, NC)
Application Number:	15/653,338
Patent Claims:	see list of patent claims

Executive summary: US Patent 10,112,947 claims a broad class of heat shock protein 90 (HSP90)-targeted conjugates in which a defined HSP90 binding scaffold (A) is connected through a linker (L) to a second payload (B) that is either (i) a detection moiety or (ii) an anti-cancer agent or (iii) an additional HSP90 binding component (B). Claim 1 is extremely expansive on substituent, linker length, payload type, and payload class, which makes the estate potent against “same structure, different payload” and “same payload, different linker length/substituent within ranges” competitors, but also increases invalidity exposure for overbreadth if the specification does not support the full combinatorial space with enabling examples and clear structure-activity correlation. Dependent claims narrow to specific payload categories (detection, fluorophores, named anti-cancer agents) and specific linker parameters (m, p, n). A licensing strategy and litigation posture against a given competitor should focus first on whether their conjugate matches the core “HSP90 binding component (formula II or III) + defined linker architecture (m=2/3; n=4-20; p=2/3; poly(alkylene glycol)-like repeating unit) + payload attachment points (X1/X2/L/X1 moieties) + payload class (detection vs anti-cancer vs second HSP90 binder)” rather than the nominal payload identity.

What patents protect HSP90 binding conjugates with detection moieties or anti-cancer agents like US 10,112,947?

What is claimed in US 10,112,947 (core claim 1 architecture)

US 10,112,947 claim 1 recites a compound of formula (I) in which key claim elements are three modules:

1. Targeting module A (HSP90 binding component)
   A is defined as a heat shock protein 90 binding component of formula (III), with the claim covering substantial chemical variability:

   • R is alkylenyl or heteroalkylenyl
   • Y¹ is independently CH or N
   • Z¹ forms a heterocyclic ring
   • Z² is H or halo
   • Z³ is CH₂, S, O, or NH
   • Z⁴ is H or halo
   • A attaches at a defined “point of attachment” to the formula (I) skeleton

2. Linker module L (defined divalent linker scaffold)
   L is a divalent linker:
   –(CH₂)ᵐ–(OCH₂CH₂)ⁿ–O–(CH₂)ᵖ–
   with:

   • m = 2 or 3
   • n = 4 through 20 (broad)
   • p = 2 or 3

3. Payload module B (detection moiety, anti-cancer agent, or second HSP90 binding component) Claim 1 allows B to be:

   • Detection moiety (fluorophore or radioactive compound), with fluorophores limited to fluorescein, rhodamine, coumarin, cyanine, BODIPY; radioactive is any radioisotope.
   • Anti-cancer agent, but the claim is functionally broad because it lists a very long set of agent classes and specific examples.
   • A second HSP90 binding component (formula II or III), meaning these are also “bis-HSP90 binder” constructs.

4. Attachment chemistry in claim 1 Claim 1 defines:

   • X¹ is NH, O, S, C(O), or S(O)₂
   • X² is NR, O, S, C(O), or S(O)₂
     These heteroatom and sulfone/sulfoxide options expand how B or A is coupled to the core scaffold.

Practical consequence: broad coverage by modular design

The claim’s breadth is driven by three freedoms:

• A variability (multiple ring and substituent options still within “formula (III)”).
• L variability (n from 4 to 20 and m/p each 2 or 3).
• B variability (any of many payload classes and named agents, plus multiple detection chemistries).

This modular structure means US 10,112,947 is most powerful when competitors preserve:

• the same HSP90 binding scaffold family for A (formula II or III depending on payload mode),
• the same linker architecture (the exact glycol-ether-like repeating unit),
• the same heteroatom attachment logic (X¹/X² classes),
• and at least one payload class enumerated within claim 1.

---

Which dependent claims narrow detection moieties, fluorophores, and payload attachment points?

Detection sub-claims (Claims 3-6 and associated narrowing)

• Claim 3: B is the detection moiety.
• Claim 4: B is detection moiety comprising the fluorophore.
• Claim 5: specifies B has a further formula (not reproduced here) and identifies an attachment to the “–X²–L–X¹–A” moiety.
• Claim 6: fluorophore is fluorescein.

Infringement relevance: For imaging agents, the critical factual question is whether the competitor’s fluorophore is one of the enumerated families and whether fluorescein is used (Claim 6). If they use a different dye family (even another fluorescein derivative), Claim 6 may not read, but Claim 4/1 may still depending on whether the structure maps to the fluorophore definitions.

Payload attachment: where design-around typically occurs

Claim 5’s explicit attachment language suggests that the patent expects a specific positional chemistry at the B end that ties into the “X²-L-X¹-A” chain. Design-arounds often try to:

• change X² and/or X¹ heteroatom identity (switch amide vs carbamate vs thioether vs sulfonamide-like coupling),
• alter the linker connectivity pattern (swap where the glycol-ether chain is attached),
• or change the linker into a non-matching scaffold (even if the overall PEG-like length is similar).

Because claim 1 defines the linker formula precisely, “PEG chain length tuning” alone will not avoid infringement if the competitor keeps the same connector endpoints and repeating unit architecture.

---

What anti-cancer agents are explicitly covered in US 10,112,947 and how broad is that list?

Claim 7 and Claim 8: named anti-cancer agents

• Claim 7: B is the anti-cancer agent.
• Claim 8: anti-cancer agent is one of the named list including: methotrexate, topotecan, irinotecan, etoposide, teniposide, lamellarin D, SN-38, camptothecin, belotecan, rubitecan, thalidomide, verteporfin (among others; the excerpt includes these explicitly).

Coverage implication: If a competitor uses one of these exact drug identities as the payload B, it is a cleaner infringement case versus payloads that only fall into a broader “class” concept but are not named.

Claim 1’s “anti-cancer agent” definition is extremely expansive

The body of claim 1 (as provided) includes large enumerations across:

• chemotherapy (alkylators, antimetabolites, platinum agents, topoisomerase inhibitors),
• targeted biologics (EGFR/HER2 antibodies),
• kinase inhibitors (BTK, MEK, PI3K, JAK, etc.),
• photodynamic agents (porphyrins/verteporfin),
• and other named agents.

Litigation consequence: This breadth can produce two opposite outcomes in court:

• Pros: A competitor using a canonical enumerated agent (for example topotecan/SN-38/irinotecan) faces stronger “read-on” arguments.
• Cons: Overbreadth and enablement arguments become more plausible if the specification does not provide adequate guidance across the entire enumerated payload class and linkage modes.

---

How strong is US 10,112,947 for asserting against competitors: claim strength versus design-around space?

What makes the claim strong

1. The linker formula is specific.
   L is not “any PEG” but a defined divalent structure:
   –(CH₂)ᵐ–(OCH₂CH₂)ⁿ–O–(CH₂)ᵖ–
   with constrained endpoint lengths and a wide n range. This specificity can be leveraged in infringement mapping.

2. X¹ and X² are defined heteroatom classes.
   Allowing multiple heteroatom identities increases read-on coverage while still leaving fewer degrees of freedom than completely open linker chemistry.

3. Payload type is functionally broad but still enumerated.
   For detection and anti-cancer modes, the claim text gives concrete dye families and agent categories/examples.

What creates vulnerability

1. High combinatorial coverage.
   With modular variation in A (multiple heterocycle options), L (n=4-20), and B (detection + vast anticancer payload set + second HSP90 binders), the effective covered set can be very large.

2. Potential enablement and best-mode pressure (structural support).
   If the specification lacks guidance and examples that track the breadth of payload permutations, the claim can be attacked as not enabling the full range.

3. Ambiguity risk from formula-defined moieties.
   Where infringement depends on whether a competitor’s A scaffold falls “within formula (III) or (II),” claim construction becomes decisive. If parties dispute whether particular ring substitutions match the defined variable definitions, the infringement case can become brittle.

---

When does US 10,112,947 lose exclusivity under US patent term rules and PTA?

No exclusivity or expiration timeline can be produced from the claim text alone. A complete term analysis requires the patent’s issue date, filing date, and any patent term adjustment (PTA) and terminal disclaimers. With only the claim text and without those bibliographic facts, any expiration date would be fabricated.

What is the Orange Book status of US 10,112,947 and does it block generic entry?

No Orange Book status analysis can be performed from the claim text. Orange Book listings are tied to specific FDA-approved drug products and NDA/ANDA/BLA status, which are not provided here.

Is this a small-molecule conjugate or a biologic conjugate risk for biosimilars?

The claim, as provided, is framed as “a compound” with defined chemical formulas and optional fluorophores/radioisotopes and cytotoxic payloads. That presentation aligns with small-molecule conjugates rather than proteins. However, biosimilar timelines and biosimilarity frameworks depend on whether the underlying therapeutic is an FDA-approved biologic product and whether the claimed subject matter overlaps with a BLA reference product. No such product context is provided, so no biosimilar risk assessment can be stated.

---

What claim elements would drive claim construction: formulas (I)-(III) and attachment points?

A-to-core attachment and “point of attachment”

Claim 1 repeatedly references “point of attachment” within formula (I) and formula (II)/(III). For infringement, courts typically focus on whether:

• the competitor’s A scaffold connects at the same position relative to the variable ring system,
• the competitor’s B module is coupled through the same defined X²-L-X¹ sequence, and
• the competitor’s linker matches the divalent architecture exactly.

How the variable definitions change the infringement map

• Y¹ (CH vs N) and Z³ (CH₂ vs S vs O vs NH) alter aromatic/heteroatom content in the binding scaffold.
• Z²/Z⁴ (H or halo) capture substitution patterns.
• R (alkylenyl or heteroalkylenyl) changes the bridging element connecting substructures inside A.

If a competitor uses a related HSP90-binding scaffold that differs materially in the heteroatom presence (for example, converting CH positions to N or vice versa), it may fall outside the defined formula envelope even if it is functionally similar.

---

How could a competitor design around US 10,112,947 while keeping HSP90 targeting?

Using claim 1’s modular constraints, typical design-around pathways include:

1. Change linker architecture away from –(CH₂)ᵐ–(OCH₂CH₂)ⁿ–O–(CH₂)ᵖ–

   • Replace glycol-ether repeating unit with different PEG-like or charged linkers.
   • Keep similar hydrophilicity but alter connectivity to avoid matching the exact linker formula.

2. Alter the attachment heteroatom classes X¹/X²

   • Switch from allowed heteroatom connectivities (NH/O/S/C(O)/S(O)₂) to a disallowed linkage type.

3. Swap payload modality

   • Use a non-enumerated imaging dye family not captured as fluorescein/rhodamine/coumarin/cyanine/BODIPY.
   • Use an anti-cancer agent that is outside the provided categories and not one of the explicitly named examples (depending on how claim interpretation handles “class” versus “named”).

4. Replace HSP90 binding scaffold identity

   • Use a different HSP90 binding chemotype that does not map to formula (II) or (III) variable definitions.

---

What patent estate questions matter next: continuation coverage, claim scope overlap, and likely litigation themes?

With only US 10,112,947 claim text, estate analysis cannot be completed reliably. A real “landscape” assessment needs at least:

• other family members and continuations,
• prosecution history for claim construction signals,
• and whether other patents cover upstream intermediates or downstream conjugates.

Because none of that data is supplied here, only the internal structure of US 10,112,947 can be analyzed.

---

Key Takeaways

• US 10,112,947 claim 1 is a modular HSP90-conjugate patent: A (HSP90 binder) + L (defined glycol-ether linker with m=2/3, n=4-20, p=2/3) + B (detection dye/radioisotope, anti-cancer agent, or second HSP90 binder).
• Dependent claims narrow payload/detection: fluorescein in Claim 6 and defined detection attachment in Claim 5.
• Anti-cancer coverage includes many enumerated agents and Claim 8 lists named examples (e.g., methotrexate, topotecan, irinotecan, etoposide, SN-38, camptothecin, belotecan/rubitecan, thalidomide, verteporfin).
• The patent’s enforcement strength is driven by linker formula specificity and the defined attachment heteroatom classes, while its enforceability risk grows with the breadth of combinatorial payload and scaffold variability.

---

FAQs

1. What linker changes avoid infringement of US 10,112,947?
   Any modification that takes the structure outside the claimed divalent linker formula –(CH₂)ᵐ–(OCH₂CH₂)ⁿ–O–(CH₂)ᵖ– with the defined m, n, p constraints.

2. Does US 10,112,947 cover PEGylated HSP90 conjugates with different PEG repeat units?
   Coverage depends on whether the repeat-unit structure and endpoints match the claimed linker architecture; “PEG-like” function alone does not control.

3. Are fluorescein-conjugates automatically covered?
   Claim 6 explicitly covers fluorescein when Claim 4/5’s detection framework and attachment structure are satisfied.

4. Can a different cytotoxic payload still infringe if it is not among the listed anti-cancer agents?
   Claim 1 includes an “anti-cancer agent” definition tied to extensive categories and examples; infringement depends on whether the competitor’s payload falls within that defined scope rather than only Claim 8’s named subset.

5. How important are the X¹ and X² heteroatom types for infringement?
   They are central to the coupling architecture in claim 1; changing linkage type can break read-on even if A, L, and B are otherwise similar.

---

References (APA)

1. US Patent 10,112,947. (Claims as provided in prompt text).

Details for Patent 10,112,947

Applicant	Tradename	Biologic Ingredient	Dosage Form	BLA	Approval Date	Patent No.	Expiredate
Recordati Rare Diseases, Inc.	ELSPAR	asparaginase	For Injection	101063	January 10, 1978	10,112,947	2037-07-18
Servier Pharmaceuticals Llc	ONCASPAR	pegaspargase	Injection	103411	February 01, 1994	10,112,947	2037-07-18