US 8206718

I'll search for authoritative information on this specific patent number, then summarize.

Let me verify the claim set and check for litigation history.

US 8,206,718 B2 — Summary

Note on sourcing: The following is based on the Google Patents full-text record for US8206718B2 (https://patents.google.com/patent/US8206718/en) and its public family members. Two caveats flagged up front: (1) the Google Patents page I retrieved reproduced the front matter, description, and abstract but not the numbered claim set, so my claim-level analysis is derived from the patent's own abstract plus the granted claim set of the EP family member (EP1107989B1), which may not be word-for-word identical to the U.S. claims; and (2) I found no CAFC 2026 docket activity for this patent number, so I have no authoritative litigation information.

Bibliographic data

Abstract

A homodimeric protein with angiogenesis-inhibiting activity, consisting of two identical fusion proteins bound together as a homodimer. Each fusion protein comprises an immunoglobulin Fc region and a first target protein linked to it, where the first target protein has the angiogenesis-inhibiting activity of angiostatin or endostatin and is selected from a plasminogen fragment and a collagen XVIII fragment. The immunoglobulin Fc region comprises a hinge region, a CH2 region, and a CH3 region.

Technology in plain language

The invention addresses a manufacturing problem: angiostatin (a ~38 kDa internal plasminogen fragment) and endostatin (a ~20 kDa C-terminal collagen XVIII fragment) are potent anti-angiogenic proteins, but prior expression systems (E. coli, baculovirus, yeast) gave insoluble aggregates or very low yields. The disclosed solution is to express them as "immunofusins" — fusion proteins in which the angiogenesis inhibitor is fused to an immunoglobulin Fc region (hinge + CH2 + CH3) via the pdCs expression vector in mammalian cells (NS/0 myeloma, 293, etc.). The Fc partner acts as a carrier promoting correct folding, solubility, and secretion, and confers Protein A–purifiable, longer-circulating (huFc-huAngiostatin t½ ≈ 33 hr in mouse vs. 4–6 hr) dimeric molecules. Constructs contemplated include Fc–X, X–Fc, and X–Fc–X arrangements, tandem angiostatin/endostatin hybrids joined by Gly-Ser linkers, and Kringle sub-fragments such as Kringle 1 (including a "Cys-67→Ser" inner-loop variant). The specification also contemplates proteolytic cleavage of the inhibitor from the Fc and direct administration of the fusion protein or encoding nucleic acid.

Independent claims — plain language

Based on the abstract (and consistent with the EP family member's claim 1):

• Independent claim 1 (protein): A homodimeric protein having angiogenesis-inhibiting activity, made of two identical fusion proteins dimerized together. Each fusion protein = (a) an immunoglobulin Fc region comprising a hinge region, a CH2 region and a CH3 region, plus (b) a target protein that is a plasminogen fragment (angiostatin activity) or a collagen XVIII fragment (endostatin activity), linked to the Fc region.

Because I could not retrieve the actual U.S. claim listing, I cannot state with confidence how many independent claims US8206718 contains or the precise wording/scope of any independent method or nucleic-acid claims. The patent's description strongly indicates the disclosure supports independent claims to: nucleic acid molecules encoding such fusion proteins (with or without a signal sequence), expression vectors and host cells, methods of producing the fusion protein by culturing a mammalian host cell, and methods of treating angiogenesis-mediated conditions (e.g., tumors) — but treat the counting and exact wording of those as unverified.

Uncertainty and gaps

• CAFC / litigation: No 2026 Federal Circuit docket or litigation involving patent 8,206,718 was found in my searches. I cannot confirm or deny any enforcement, IPR, or appeal activity.
• Claim text: The full U.S. claim set was not retrieved; my claim overview rests on the abstract and the EP counterpart. Verify against the official USPTO PatentCenter record (US 11/475,627) or the granted US8206718B2 claims before relying on scope.
• Expiration: The 2023-08-12 "adjusted expiration" is as reported by Google Patents and is explicitly an assumption, not a legal conclusion.

I'll search for litigation involving this specific patent number.

Let me search more specifically on litigation databases.

Let me try the litigation-specific portals and docket databases.

I'll start by checking the structured data and then running web searches to see whether any AIA trial proceedings exist for this patent.

Let me run more targeted searches to be thorough, including the patent's family and known litigation involving this patent.

Let me do a final round of checks specifically for litigation-driven IPRs, Unified Patents involvement, and the sibling patent.

PTAB Proceedings Analysis — US 8,206,718 B2

Patent: US 8,206,718 B2 — "Expression and export of angiogenesis inhibitors as immunofusins"
Patent Owner / Current Assignee: Merck Patent GmbH (chain of title: Lexigen Pharmaceuticals Corp. → EMD Lexigen Research Center Corp. → EMD Serono Research Center, Inc. → Merck Patent GmbH)
Inventors: Kin-Ming Lo; Yue Li; Stephen D. Gillies
Application: US 11/475,627, filed 2006-06-27; granted 2012-06-26
Earliest priority: 1998-08-25 (provisional 60/097,883); § 371/continuation chain from App. 09/383,315 (filed 1999-08-25)
Recorded legal status: Expired – Fee Related; adjusted expiration 2023-08-12
Continuation child: US 8,703,908 B2 (App. 13/533,250, filed 2012-06-26)

---

Proceedings overview

Total AIA trial proceedings on file: 0. The structured "PTAB proceedings on file" block states the USPTO Open Data Portal returned no AIA trial proceedings for this patent as of the most recent ingest, and my independent web checks (Google Patents family/prosecution view, P-TACTS petition-document hits, and general searches pairing "8,206,718"/"8206718" with IPR/PGR/CBM and petitioner terms such as Unified Patents) surfaced no IPR, PGR, or CBM directed at US 8,206,718 — so the breakdown is 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denied.

Bottom-line defensive posture: There is no PTAB invalidity record to borrow and no estoppel shield to inherit — but there is also no hardened, IPR-survived claim set to fear: the patent is recorded as expired (adjusted expiration 2023-08-12), so for a defendant facing a demand letter today the dominant issue is not claim validity but whether any enforceable term remains. A troll asserting this patent now is asserting an expired right — verify the expiration/term computation and the maintenance-fee record before engaging on the merits.

---

Proceedings

None. No AIA trial proceeding has been instituted or filed against US 8,206,718 on the record I can verify. I will not invent proceeding numbers, panels, or outcomes. For the avoidance of doubt, the following fields are therefore inapplicable for this patent:

• Type / Filed / Status / Judge panel / Petition grounds / Institution decision / FWD / Settlement / Appeal — N/A; no proceeding exists.
• Defensive value of prior proceedings — none available; a defendant cannot point to a prior FWD canceling any claim of this patent.

If a recently filed petition exists that the ODP ingest has not yet captured, the authoritative places to confirm are the PTAB P-TACTS case-tracking system (https://ptacts.uspto.gov/ptacts/ui/home) and PTAB E2E / PTAB Decisions (https://developer.uspto.gov/ptab-web/#/search/decisions). Because this patent expired in 2023, new petitions are unlikely — but a petition challenging it is not categorically impossible if any pre-expiration infringement claim is still being litigated.

---

Strategic summary

Claim-level status: UNTESTED. No claim of US 8,206,718 has been canceled, confirmed, or construed by the Board. Unlike a patent that has been through IPR, there is no claim chart, no claim-construction ruling, and no estoppel record to work from. The claim set as issued (per the abstract and the Google Patents full text) is directed to a homodimeric protein comprising two identical fusion proteins, each having (i) an immunoglobulin Fc region comprising a hinge, CH2, and CH3 region, and (ii) a first target protein with angiogenesis-inhibiting activity of angiostatin or endostatin, selected from a plasminogen fragment and a collagen XVIII fragment. Because there is no FWD, I cannot and do not represent which specific claim numbers are independent vs. dependent, or which would survive scrutiny.

Estoppel landscape: empty. With zero instituted IPRs, no § 315(e)(2) estoppel has attached to anyone. A defendant today is not restricted by any prior petitioner's grounds — every § 102/§ 103/§ 112 theory remains nominally available. Conversely, because no one has litigated this patent at the Board, there is no "reasonably could have raised" trap and no precedential Board construction of the "immunofusin" limitations to rely on. The corollary risk: you would be the first to develop the invalidity record, with no free ride.

Pattern signals: none. No serial petitioner, no defensive aggregator (e.g., Unified Patents) involvement, no Patent Owner appeal activity, and no Federal Circuit progeny tied to this patent are visible on the record. The closest related activity is the continuation US 8,703,908 B2, which is its own patent with its own exposure; a challenge to the family would most naturally target the still-live-in-interest sibling rather than the expired '718. Note also the Merck Patent GmbH / Lexigen portfolio is large and well-litigated in the immunocytokine space, so future activity on related patents is plausible even though '718 itself shows none.

---

Recommended next steps

1. Lead with the expiration, not invalidity. The record shows "Expired – Fee Related, expires 2023-08-12." For a defendant receiving a demand or complaint citing US 8,206,718 today (2026-09-30), the first move is to confirm the term: the 20-year term from the 1999-08-25 non-provisional filing would have run to 2019-08-25, extended by any PTA to the recorded 2023-08-12 date. If the patent is expired, any post-expiration assertion of the '718 claims is baseless, and pre-expiration damages windows (if any) must be tied to actual notice and actual infringement. I flag the term arithmetic as an assumption to verify against the face of the patent and the maintenance-fee record — I have not independently confirmed the PTA/terminal-disclaimer basis for the 2023-08-12 date.

2. Verify no pending proceeding before relying on "zero." Pull the P-TACTS docket for "8206718" at https://ptacts.uspto.gov/ptacts/ui/home and the PTAB Decisions search at https://developer.uspto.gov/ptab-web/#/search/decisions. A zero count is a signal, not proof — but here the ODP block and my searches agree.

3. If assertion is against the live sibling, shift focus to US 8,703,908 B2. That continuation (App. 13/533,250) is the more likely target for a modern assertion. Any invalidity investigation should be run on the '908 claims, and any IPR should be filed against the '908 — not the expired '718 — subject to the § 315(b) one-year bar running from service of the first complaint alleging infringement of that patent.

4. Preserve invalidity positions without PTAB. Because '718 expired, the Board is unlikely to be the forum. If pre-expiration conduct is in suit, run the invalidity analysis in district court (anticipation/obviousness over the same Lo/Gillies immunofusin art and the U.S. Pat. Nos. 5,541,087 / 5,726,044 referenced in the specification), and consider ex parte reexamination or an IPR on the sibling if a live claim set remains in play.

5. Do not cite nonexistent FWDs. The prompt template assumes FWDs to quote. There are none for this patent. Any internal memo or brief that quotes an "IPR201x-##### FWD" for US 8,206,718 would be fabricated and sanctionable — resist the temptation to import a proceeding from the family.

---

Constraint note: No proceeding number, panel, FWD, settlement term, or appeal disposition is stated above because none exists on the verifiable record. All statements about the patent's term and status are drawn from the Google Patents record (https://patents.google.com/patent/US8206718/en) and the structured ODP block, and should be re-verified against the USPTO Patent Center file wrapper before being relied on in a filing.

I'll start by pulling the assignment record and related ownership data for US 8206718.

Let me dig into the ownership chain history and try to surface reel/frame records.

Let me check for any later assignments, litigation, and the continuation's ownership.

⚠️ Sourcing caveat up front. I was able to retrieve the Google Patents legal-events record for US 8,206,718 (which is populated from USPTO reassignment data), the patent's own cross-reference chain, and third-party documentation of the Lexigen → EMD → Merck corporate history. I was not able to retrieve the USPTO Patent Assignment Center's underlying record fields — specifically reel/frame numbers, execution dates, and the correspondent of record. Those fields are therefore left explicitly blank rather than estimated. Everything below is labeled as either verified or open.

---

Inventors

Pattern notes:

• No unusual departure pattern. The published integration case study records that after Merck KGaA's 1998 majority acquisition, "nobody left the company" and that top scientists were contractually bound. Gillies remained the site's principal and is still a named inventor here. So the classic "founders exit within 12 months → fire sale" tell is not present.
• Timing distortion to be aware of: this patent's priority date is 1998-08-25 (provisional 60/097,883), but the application that issued (11/475,627) was filed 2006-06-27 as a second-generation continuation. Inventor employer-of-record therefore spans two corporate identities (Lexigen → EMD Lexigen), which the ownership chain mirrors.

---

Original assignee

Entity named on the issued patent: Merck Patent GmbH (Darmstadt, Germany) — this is the "Original Assignee" field shown on Google Patents, but that field is populated from the post-2008 owner and does not reflect the filing-time applicant.

Filing-time applicant / true original owner: Lexigen Pharmaceuticals Corp. (Lexigen Research Center, Billerica, MA — an oncology biologics research unit; founded 1992 as Fuji Immuno Pharmaceuticals Corp.; Merck KGaA acquired a 57% majority stake in December 1998 for the immunocytokine/"immunofusin" platform).

• Products embodying the claims: No commercial product is claimed to be on the market under this patent. The immunofusin platform (Fc–IL2, Fc–endostatin, Fc–angiostatin constructs; hu14.18-IL2) reached clinical trials at Lexigen/EMD but I found no approved product reading on these claims.
• Primary line of business: Lexigen — oncology biologics R&D. Merck Patent GmbH — patent-holding subsidiary of Merck KGaA (Darmstadt), a ~€20bn diversified pharma/life-science operating group (note: unrelated to Merck & Co., Rahway, NJ).
• Current status: Merck Patent GmbH is operating; Merck KGaA is solvent and active. The Lexigen/EMD Lexigen/EMD Serono Research Center US entity has been folded into Merck KGaA's group structure (the Billerica, MA site has since been wound down).
• Patent status: Expired – Fee Related, adjusted expiration 2023-08-12 (Google Patents legal-status field). Maintenance fees were not maintained — consistent with a legacy asset of no remaining commercial value, not with an assertion campaign.

---

Assignment timeline

Source: Google Patents legal-events block for US 8,206,718 (mirrors USPTO reassignment records). All three records below carry a recording date of 2008-07-09. The indexed view I retrieved does not expose execution dates, reel/frame numbers, or correspondents; those are marked [open] and must be pulled directly from Assignment Center.

• Executed [open] / recorded 2008-07-09 — Reel [open]/[open]

  • Conveyance: Change of Name (see document for details)
  • Assignor: LEXIGEN PHARMACEUTICALS CORP.
  • Assignee: EMD LEXIGEN RESEARCH CENTER CORP.
  • Correspondent: [open — not exposed in the retrieved index]
  • Context: Internal corporate rebranding — Lexigen adopted the "EMD" umbrella brand after Merck KGaA created EMD as its North American pharma identity. Change of name only; no change in beneficial ownership.

• Executed [open] / recorded 2008-07-09 — Reel [open]/[open]

  • Conveyance: Change of Name (see document for details)
  • Assignor: EMD LEXIGEN RESEARCH CENTER CORP.
  • Assignee: EMD SERONO RESEARCH CENTER, INC.
  • Correspondent: [open — not exposed in the retrieved index]
  • Context: Internal rebranding following Merck KGaA's acquisition of Serono and the creation of the Merck Serono / EMD Serono identity. Change of name only; no change in beneficial ownership.

• Executed [open] / recorded 2008-07-09 — Reel [open]/[open]

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: EMD SERONO RESEARCH CENTER, INC.
  • Assignee: MERCK PATENT GMBH (Darmstadt, DE)
  • Correspondent: [open — not exposed in the retrieved index]
  • Context: Intra-group IP consolidation — US operating subsidiary's patent estate transferred up to the parent group's centralized patent-holding company. This is the classic "assign everything to the IP GmbH" housekeeping move, not an arm's-length sale.

No further recorded assignments after 2008-07-09 appear in the retrieved record. The chain has been static for ~15 years, through issuance (2012-06-26) and to expiration (2023-08-12).

Supporting corporate-event context (not USPTO records, but corroborative): A Geneva cantonal filing (ge.ch/grandconseil, doc IUE01444A) records a 10 December 2007 intra-group contract transferring shares of EMD Pharmaceuticals, Inc. and EMD Lexigen Research Center Corp. — i.e., a group restructuring in late 2007 that lines up with the July 2008 batch recording of these three name-change/assignment documents. That is consistent with a single law-department cleanup filing all three at once.

Continuation note: Google Patents records a priority claim from this patent to US 13/533,250 on 2012-06-26 → US 8,703,908 B2. I could not verify that sibling's assignment record within this session; if you are tracing the Merck KGaA immunofusin portfolio, that record should be checked separately (it likely terminates in the same Merck Patent GmbH chain).

---

Timeline diagram

timeline
    title Ownership of US 8206718
    1998 : Provisional filed by Lexigen
         : Merck KGaA buys 57 percent of Lexigen
    1999 : Parent application filed by Lexigen
    2002 : First continuation filed
    2006 : Second continuation filed
    2007 : Group transfers Lexigen shares internally
    2008 : Lexigen renamed EMD Lexigen
         : Renamed EMD Serono Research Center
         : Assigned to Merck Patent GmbH
    2012 : Patent US 8206718 issues
    2023 : Fees lapse and patent expires

---

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT.
Every link in the chain stays inside the Merck KGaA corporate family: Lexigen Pharmaceuticals Corp. → EMD Lexigen Research Center Corp. → EMD Serono Research Center, Inc. → Merck Patent GmbH. Two of the three conveyances are expressly typed "Change of Name," i.e., the same legal person under a new name, not a transfer to a new acquirer. No "IP / Patents / Licensing / Holdings / Ventures" LLC, no registered-agent service address, no single-member Delaware/Texas LLC appears.

2. Known asserter in the chain — NOT PRESENT.
Current assignee Merck Patent GmbH does not match any entity on the NPE lists supplied (Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, Spangenberg entities). Merck Patent GmbH is a centralized patent-holding company for an operating pharma, which is a structurally different animal from a licensing-only assertion vehicle: its purpose is intra-group title-holding, and it files as assignee across the Merck KGaA estate. No RPX or Unified Patents high-frequency-plaintiff listing surfaced for this patent.

3. Repeat correspondent across the chain — UNCLEAR (data gap, not an acquittal).
The correspondent of record is not exposed in the Google Patents legal-events view I retrieved, so I cannot test for recurrence. What I can say: all three conveyances were recorded on the same day (2008-07-09), which is prima facie evidence that a single filer — one attorney or one in-house IP paralegal — handled the entire batch. That is the shape of the repeat-correspondent signal, but it is a structural inference, not a cited finding. I am flagging this as the one item worth pulling from Assignment Center, since the correspondent field is exactly what disambiguates "in-house docket cleanup" from "outside NPE counsel."

4. Cascading transfers — NOT PRESENT (as an NPE signal).
Three conveyances in a single day superficially resembles a cascade, but the substance fails the test: the assignees do not share a registered-agent address or common outside principals, they are not chained LLCs, and the direction of travel is toward the ultimate parent. Two are name changes of the identical legal entity. This is a corporate reorganisation, not a laundering chain. Contrast: a genuine cascade would show three differently-named LLCs each with a single-purpose address, in sequence, within <24 months.

5. Pre-litigation transfer — NOT PRESENT.
The last recorded transfer is 2008-07-09. The patent issued 2012-06-26 and expired 2023-08-12. I found no infringement suit naming this patent. A transfer dated ~13 years before issuance and ~15 years before expiry cannot be a venue-setting or standing-cleansing pre-litigation transfer.

6. Bankruptcy fire-sale — NOT PRESENT.
No Chapter 7/11 proceeding is associated with Lexigen, EMD Lexigen, EMD Serono Research Center, or Merck Patent GmbH. Merck KGaA acquired Lexigen in 1998 precisely because Lexigen was cash-constrained, but the transaction was a strategic acquisition of a majority stake, not a court-supervised sale of a bankrupt estate. There is no bankruptcy docket to cite.

7. Privateering — NOT PRESENT.
Privateering requires assets to flow outward from an operating company to an NPE that asserts against that operating company's competitors. Here the asset flowed inward and upward, from the US R&D subsidiary to the group parent's IP holding company. No SEC-filing, Patent Progress, or EFF coverage describes Merck KGaA asserting this patent through a proxy. (Note: Merck KGaA is not SEC-registered; it reports to BaFin, so a US 10-K/8-K trail would not exist for the parent anyway — the absence of SEC evidence is weak evidence here.)

8. Defensive aggregator — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. It terminates at Merck Patent GmbH, i.e., it never left the original owner's corporate family. The patent was not neutralised by acquisition; it simply ran out.

---

Verdict

Defensive / non-asserting — with an explicit caveat on the label.

Justification: all three recorded conveyances (recording date 2008-07-09, reel/frame [open]) are internal to the Merck KGaA group — two are typed "Change of Name" between Lexigen Pharmaceuticals Corp., EMD Lexigen Research Center Corp., and EMD Serono Research Center, Inc., and the third is an "Assignment of Assignors' Interest" moving title up to Merck Patent GmbH. That is intra-group housekeeping, not a transfer to an assertion vehicle: zero NPE signals are present, no assignee matches any known NPE list, and no litigation was found in the ~15 years the patent remained in force before lapsing for non-payment on 2023-08-12.

Caveat on the menu: the supplied "Defensive / non-asserting" option is worded for chains terminating at a defensive aggregator (RPX/AST/LOT), which is not literally what happened here. Of the five available buckets, it is the only one that does not falsely attribute NPE assertion activity or an operating-company litigation campaign. The accurate characterisation is: operating-company internal ownership, never asserted, now expired.

Open items to close before relying on this as final:

1. Pull the reel/frame, execution dates, and correspondent of record for all three 2008-07-09 records at USPTO Assignment Center (or the legacy Assignment Search, search by patent number 8206718). The correspondent field is the one substantive gap — everything else in this report is corroborated. If all three share a single correspondent, that is still merely internal-housekeeping evidence, not an NPE finding.
2. Confirm whether US 8,703,908 B2 (continuation via 13/533,250) shares the identical Merck Patent GmbH chain.

Verification link: https://patents.google.com/patent/US8206718/en (legal events) · https://assignmentcenter.uspto.gov/ (reel/frame lookup)

I'll research this patent's prior-art citations. Let me run several targeted searches.

Let me look specifically for the cited-references sections of this patent.

Let me try to retrieve the cited-references list from patent databases directly.

Important scope note up front

I could not retrieve a verified front‑page "References Cited" / "Patent Citations" table for US 8,206,718 B2 from the sources available in this session. The Google Patents and Justia records I retrieved returned the bibliographic data, description and claims, but not the examiner/applicant citation list.

Rather than reconstruct that list from memory (which would risk fabricating citations and dates), I am splitting my answer into two parts:

1. References that are explicitly cited and quoted inside the authoritative full text of US 8,206,718 (reliable — these are in the patent body itself), with § 102 analysis.
2. A clearly-flagged note on what I could not verify, and what the user should pull from the USPTO front page / Global Dossier to complete the examiner-citation analysis.

Also, per the operating rules, all numbers are interpreted literally; I have not "corrected" any identifier. Note that SE 8206718 A (a 1984 Swedish document, cited family member "KEINE/NONE") and US 8,206,718 B2 are entirely different documents and are not to be conflated.

---

1. Patent identification (verified)

Claims potentially at issue (from the patent's own claim set, e.g. Justia 20070009538):

• Claim 1: homodimeric protein = two identical fusion proteins, each comprising an immunoglobulin Fc region (hinge + CH2 + CH3) and a first target protein having angiostatin or endostatin activity, selected from a plasminogen fragment and a collagen XVIII fragment.
• Claim 2: adds a second target protein linked to the first target protein or to the Fc region.

---

2. Prior art explicitly cited in the patent text, with § 102 analysis

2(a) Patent documents cited by the applicant

2(b) Non‑patent literature explicitly cited in the patent text

These are quoted verbatim in the US 8,206,718 description (authoritative text supplied):

2(c) References surfaced from related/family records (verify before relying on them)

• Bachelot et al., Proc. AACR 39:271‑2 (March 1998) — "Retrovirus‑mediated gene transfer of an angiostatin–endostatin fusion protein with enhanced anti‑tumor properties in vivo." This was tagged X (novelty‑destroying) in the search report for EP 1 705 248. Because it pre‑dates the 1998‑08‑25 priority, it is the most dangerous single item against claim 2 (angiostatin + endostatin in one molecule). I flag it because it came from an EP search report, not from the US 8,206,718 front page.
• Bussolino et al., TIBS 22(7):251‑256 (1997) — also tagged X in that same EP search report.
• WO 96/35774 (Children's Medical Center, 1996‑11‑14) — angiostatin; cited as "A,D" in the EP report (cited in the application).
• EP 0 141 484 (Biogen); Forsgren et al., FEBS Lett. 213(2):254 (1987); WO 92/16221 (Synergen) — background (A) citations touching plasminogen/angiostatin.
• RU 2240328 C2 ("Homodimeric fused protein with activity of angiogenesis inhibitor…") — appears to be a family counterpart; its record lists non‑patent citations Lo K.M. et al. 1998, Sim B.K. et al. 1997, and O'Reilly M.S. et al. 1997. I could not confirm that those three are the US examiner citations for 8,206,718 — treat as family‑level context only.
• EP 1 705 248 A1 ("Fusion proteins comprising an angiostatin moiety…") and MX PA/a/2001/001970 ("Expression and export of angiostatin and endostatin as immunofusins") appeared in the RU record's cited‑by table.

---

3. Bottom line on anticipation

• No single reference I can verify anticipates claim 1 as a whole. Claim 1 requires a homodimeric Fc fusion in which the target is specifically an angiostatin‑type (plasminogen fragment) or endostatin‑type (collagen XVIII fragment) moiety. The generic immunofusin patents (US 5,541,087 / 5,726,044) and Lo 1998 teach the Fc–X architecture but not these targets; the angiostatin and endostatin patents (US 5,733,876 / 5,837,682 / 5,885,795 / 5,854,205) and O'Reilly 1994/1997 teach the targets but not the Fc homodimer.
• The realistic § 102/§ 103 posture is a combination: US 5,541,087 or US 5,726,044 or Lo et al. 1998 (Fc‑X immunofusin platform) + O'Reilly 1997 / US 5,854,205 (endostatin) and O'Reilly 1994 / US 5,733,876 (angiostatin), with O'Reilly 1997's express statement that recombinant yields were too low providing the § 103 motivation. That is an obviousness argument, not clean anticipation.
• Claim 2 (angiostatin + endostatin tandem/bifunctional) is the claim most exposed to a pure § 102 attack, and Bachelot et al. 1998 is the reference to check hardest.
• Practically, US 8,206,718 is Expired – Fee Related (adjusted expiration 2023‑08‑12), so the anticipation question is now of historical/Litigation‑context interest rather than an active-infringement question.

---

4. What I could not verify (please confirm against the official record)

I was unable to pull the examiner‑cited "References Cited" table for US 8,206,718 B2. To complete the exactly‑as‑requested deliverable, retrieve from USPTO PatentCenter/Global Dossier (https://patents.google.com/patent/US8206718/en → "USPTO PatentCenter") the front page's U.S. Patent Documents, Foreign Patent Documents and Other Publications lists for application 11/475,627, and I will map each listed reference to claims 1 and 2 under § 102 with exact dates. Until that is done, the tables in § 2 are the references that are verifiable inside the patent text, not the official examiner citation list.

I'll research the key prior-art references cited on the patent page before building the obviousness analysis.

Obviousness Analysis — US 8,206,718 B2 (“Expression and export of angiogenesis inhibitors as immunofusins”)

Prepared: 26 April 2026 · Framework: pre‑AIA 35 U.S.C. § 103(a); Graham v. John Deere factors as refined by KSR Int’l v. Teleflex
Primary source: the fetched Google Patents record for US8206718B2 — https://patents.google.com/patent/US8206718/en (with its “Prior art keywords” field: region, endostatin, angiostatin, protein, immunoglobulin; and “Prior art date: 1998‑08‑25”).

---

1. Preliminary caveats (important for how much weight to give this memo)

1. The full text of the granted claims is not in the fetched record. Google Patents renders the Description, Abstract and family data, but the claim set is not reproduced verbatim in the supplied text. I therefore analyze the claim subject matter as it is reflected in the Abstract and in the specification/related family, and I flag where I am inferring.
   • The Abstract recites: “A homodimeric protein … consists of two identical fusion proteins bound together as a homodimer. Each fusion protein comprises an immunoglobulin Fc region and a first target protein … The first target protein has an angiogenesis inhibiting activity of angiostatin or endostatin, and is selected from the group consisting of a plasminogen fragment and a collagen XVIII fragment. The immunoglobulin Fc region comprises a hinge region, a CH2 region, and a CH3 region.”
   • The family corroborates this: the related continuation US 13/533,250 (published as US 2013/0165634, later US8703908B2) claims “A fusion protein comprising an angiogenesis inhibiting collagen XVIII fragment linked to an immunoglobulin Fc region,” with dependent claims to hinge/CH2/CH3, direct linkage, linker linkage, SEQ ID NO: 4, IgG1 Fc, and “A dimer of the fusion protein.” (https://www.patentsencyclopedia.com/app/20130165634)
   • The Russian family member RU2240328C2 is titled “Homodimeric fused protein with activity of angiogenesis inhibitor (variants)…” — consistent with the Abstract being a fair statement of claim 1 of the US patent.
2. Identifiers are taken literally. Notably, the fetched page contains an internal citation inconsistency the patent itself carries: the Detailed Description says “a useful expression vector is pdCs (Lo et al. (1988) Protein Engineering 11:495…)” while Example 1 and Example 15 say “Lo et al. (1998) Protein Engineering 11:495.” The underlying article is Lo, Sudo, Chen, Li, Lan, Kong, Chen, An & Gillies, High level expression and secretion of Fc‑X fusion proteins in mammalian cells, Protein Engineering 11(6):495–500 (1998) (https://doi.org/10.1093/protein/11.6.495). I flag this rather than silently harmonizing it; for §103 purposes the reference is dated 1998, the same year as the 25 August 1998 provisional.
3. Likewise the page uses “SEQ ID NO: 3” as the template for both endostatin (Example 1) and angiostatin (Example 4). That is a written‑description/§112 issue, not a §103 issue, and I do not rely on it.

Effective date. The application that issued as US 8,206,718 (Ser. No. 11/475,627, filed 27 June 2006) is a continuation chain back to Ser. No. 09/383,315, filed 25 August 1999, which claims benefit of provisional 60/097,883, filed 25 August 1998 (the page’s stated priority/prior‑art date). Pre‑AIA §102/§103 therefore govern, with 25 August 1998 as the operative critical date.

---

2. The prior art on the page, sorted by role

Where a reference is only known to me as cited on the patent page (i.e., no independent URL was retrieved), I say so; I do not attribute content to it beyond what the page itself recites.

---

3. The dispositive question

The claims are, at bottom, the application of a known, general‑purpose Fc‑fusion (“immunofusin”) expression platform to a known, well‑characterized pair of target proteins (angiostatin and endostatin) whose identity, boundaries and therapeutic utility were already public. The §103 question is therefore: was it obvious to select angiostatin/endostatin as the “target protein X” in an Fc‑X cassette? Under KSR, the answer is yes.

---

4. Combination A — Endostatin/collagen XVIII claims

References combined: O’Reilly 1997 + US 5,854,205 (endostatin/collagen XVIII) in view of US 5,541,087 / US 5,726,044 and Lo et al. 1998 (immunofusin/pdCs Fc‑X platform), optionally with Capon 1989.

Element mapping (Abstract‑level claim 1):

Motivation to combine (why a POSITA would have done this as of 8/25/1998):

1. A recognized, unsolved production problem. The patent’s own Background concedes the state of the art: clinical‑grade soluble, active angiostatin/endostatin could not be made in commercial quantities in E. coli, baculovirus, yeast or existing mammalian systems; E. coli gave “insoluble protein aggregates”; mammalian/baculovirus gave “very low levels”; and endostatin from E. coli “precipitated during dialysis.” O’Reilly 1997 itself reports administering non‑refolded E. coli material. Where the art identifies a problem and the prior art supplies a known remedy, the combination is motivated.
2. The remedy was the known purpose of the immunofusin platform. US 5,541,087 and US 5,726,044 are expressly directed to “enhanc[ing] the production of a given target protein,” especially proteins “secreted at low levels.” Lo 1998 demonstrated that the same pdCs/Fc‑X cassette gives high‑level secretion for many different, structurally unrelated proteins (“high levels of expression … for many different types of proteins”). Fc‑X was, in 1998, a generic, target‑agnostic platform.
3. The specific favorable properties needed were already attributed to Fc. The patent itself lists as known purposes of Fc fusion: “proper folding, increased solubility, … protein A binding, increased circulation half‑life.” Capon 1989 independently documents the long plasma half‑life of Fc fusions. Given that endostatin’s dosing burden (20 mg/kg/day, Boehm 1997) is a function of rapid renal clearance of a ~20 kDa protein, fusing to Fc to extend half‑life is an obvious, predictable design choice.
4. Reasonable expectation of success. Success did not depend on unpredictable chemistry: a single XmaI/XhoI or HindIII ligation of an existing endostatin cDNA into an existing vector (“the gene encoding the protein of interest can be ligated via the XmaI site” — Lo 1998). Post‑KSR, routine, predictable engineering with a known platform and a known target is obvious.

Claim 1 of the related continuation (US 2013/0165634) — “a fusion protein comprising an angiogenesis inhibiting collagen XVIII fragment linked to an immunoglobulin Fc region” — is squarely met by O’Reilly 1997/’205 + US 5,541,087 + Lo 1998. Dependents adding hinge/CH2/CH3, direct vs. linker linkage, SEQ ID NO: 4, IgG1 Fc, and dimers are met by the same combination (IgG1 Fc and hinge/CH2/CH3 from Lo 1998; direct fusion via the CH3 XmaI site from Lo 1998/US 5,541,087; Gly‑Ser linker from US 5,258,698; dimerization inherent to hinge).

---

5. Combination B — Angiostatin/plasminogen‑fragment claims (incl. Kringle fragments)

References: O’Reilly 1994 + US 5,733,876 / 5,837,682 / 5,885,795, in view of US 5,541,087 / US 5,726,044 + Lo 1998 (and Capon 1989).

The same mapping holds, substituting a plasminogen fragment for the collagen XVIII fragment. Additional motivations:

• O’Reilly 1994 identifies angiostatin as a 38 kDa internal plasminogen fragment containing Kringle regions, and the patent itself concedes that Kringle 1, Kringle 2, Kringle 3 and combinations “exhibit physiologically similar behavior” — i.e., the sub‑fragment claims are supported by, and obvious in view of, the parent disclosure plus routine domain deletion.
• The specification also concedes the reason Fc works here: “when different parts of the angiostatin are expressed as Fc fusion molecules, high levels of expression are obtained, presumably because the Fc portion acts as a carrier.” A stated, predictable mechanism of action is the antithesis of non‑obviousness.
• The patent’s disclosure that Cys‑67 of the murine K1 inner loop was mutated to Ser (Example 11) is a routine engineering step — avoiding an unpaired cysteine — of the kind KSR classifies as predictable optimization.

---

6. Combination C — Homodimer/dimer and “bivalency” limitations

If the granted claims recite the homodimer expressly (as the Abstract and RU2240328C2 suggest), that element is not a saving distinction:

• Capon 1989 discloses Fc immunoadhesins assembled as homo‑dimers (and tetramers), and US 5,541,087/US 5,726,044 likewise describe multimers.
• The patent itself states that the Fc fragment “itself typically is a dimer of the heavy chain fragments including at least a portion of the hinge region and CH3 domain, and optionally the CH2 domain,” and that “disulfide bridges … linking together the portions of two immunoglobulin Fc regions via an immunoglobulin hinge region … are characteristic of native forms of these molecules.” An inherent property of the claimed structure is not an inventive contribution.
• The patent’s asserted benefit (higher avidity from bivalency) is precisely the mechanism Capon 1989 and the immunoadhesin art already described for Fc‑fused receptor domains.

---

7. Combination D — Tandem/bifunctional constructs (Fc‑X‑X, X‑Fc‑X)

References: O’Reilly 1994 + O’Reilly 1997 (both) + US 5,541,087/US 5,726,044 + US 5,258,698 (linker) + Lo 1998.

• O’Reilly 1997 is explicit that its findings “Together with angiostatin data, … validate a strategy … suggest a theme of fragments of proteins as angiogenesis inhibitors.” That is a teach‑to‑combine the two inhibitors.
• The patent concedes the rationale: “Such functional hybrid proteins would allow the synergistic effect of angiostatin and endostatin to be explored in animal models.”
• The linker limitation is met by US 5,258,698, and the specification itself concedes that “the optimal linker length and amino acid composition may be determined by routine experimentation” — a textbook obviousness admission.

---

8. Combination E — Cleavage‑site and species‑matched (murine/canine) limitations

• Proteolytic cleavage site (e.g., plasmin/trypsin or enterokinase Asp₄‑Lys between Fc and target): expressly disclosed in US 5,726,044 (“interpose a proteolytic cleavage site between the encoded secretion cassette and the encoded target protein”) and LaVallie 1993 (Asp₄‑Lys). Motivation: ability to remove Fc and purify the free inhibitor (a stated object of the immunofusin patents).
• Non‑human Fc/target pairs (muFc‑muEndo, caFc‑caAngio, etc.): motivated by the conceded need for species‑matched surrogate proteins so that a human protein does not “elicit an immune response, and/or exhibit different pharmacokinetics,” and by the conceded inadequacy of inbred mouse/induced‑tumor models relative to spontaneous canine tumors. US 5,726,044 itself exemplifies murine Fcγ2a expression. A POSITA would predictably substitute the species‑matched constant region.

---

9. KSR synthesis — why this is not a “teaching away” case

The KSR factors all point toward obviousness:

• Known problem, known solution, finite alternatives. The field knew (i) the production problem and (ii) the Fc‑fusion remedy; the only open choice was which target to fuse, and the two targets were the most prominent angiogenesis inhibitors of the day.
• Predictable results. Increased solubility, secretion, half‑life and Protein A purifiability from Fc fusion were documented expectations (Capon 1989; US 5,541,087; US 5,726,044), not surprises.
• Design incentive / market pressure. The patent Background itself recites the “urgent need to test endostatin and angiostatin clinically” and the 20 mg/kg/day dosing burden — i.e., pressure to improve manufacture and pharmacokinetics.
• Obviousness does not require the combination to be physically mixed in one reference. KSR expressly blesses combining familiar elements according to known methods to yield predictable results.

---

10. Counterarguments a patent owner would press (and my assessment)

(a) “Teaching away.” One could argue the 1997 art taught away from recombinant mammalian expression because “it has not been possible to produce clinical grade soluble, active angiostatin and endostatin … using E. coli, baculoviral, yeast, and mammalian expression systems.” Assessment: weak. A general statement that unmodified expression failed in several hosts is not a specific teaching that Fc fusion in mammalian cells would fail; indeed the stated failure mode (improper folding of an internal fragment; 26 disulfides in angiostatin; hydrophobicity/precipitation of endostatin) is exactly the problem a hydrophilic, secretion‑directing, disulfide‑stabilizing Fc partner is known to address. O’Reilly 1997’s use of non‑refolded E. coli material is a workaround, not a disclaimer.

(b) Unexpected results / secondary considerations. The strongest non‑obviousness narrative is the degree of success: 30–100 µg/ml secretion, half‑life rising from 4–6 h (native angiostatin; O’Reilly 1996, Nat. Med. 2:689) to ~33 h, and greater Lewis‑lung efficacy than the E. coli material. Assessment: mixed. A POSITA would have predicted a half‑life increase and solubility gain (Capon 1989; US 5,541,087). Whether a 5–8× half‑life extension and the obtained expression titers are “unexpected” is a factual battle over the level of ordinary skill in 1998 — but no unexpected property (e.g., a new activity or a broken structure–activity expectation) is established. Secondary considerations must be attributable to the claimed invention and not to the known advantages of the known platform; here they largely inure to Fc fusion generally.

(c) There is no nexus problem in reverse. Commercial or clinical success of any Fc‑endostatin/angiostatin product would need a nexus to these claims; the fetched record contains no such evidence.

---

11. Bottom line

A prima facie case of obviousness under §103(a) is established, at least by the following combinations:

1. O’Reilly 1997 (Cell 88:277) and/or US 5,854,205 (endostatin = collagen XVIII fragment) + US 5,541,087 / US 5,726,044 (immunofusin Fc‑X cassette) + Lo et al. 1998, Protein Eng. 11:495 (pdCs human Fcγ1 hinge/CH2/CH3 Fc‑X with XmaI junction) — renders the endostatin/collagen XVIII fusion claims (and their hinge/CH2/CH3, direct‑linkage, IgG1, and dimer dependents) obvious.
2. O’Reilly 1994 (Cell 79:315) and/or US 5,733,876 / 5,837,682 / 5,885,795 (angiostatin = plasminogen fragment, Kringle‑containing) + the same immunofusin references — renders the angiostatin/plasminogen‑fragment claims obvious.
3. Addition of Capon 1989 (Nature 337:525) for the disulfide‑linked homodimer/bivalency limitations — or, alternatively, those limitations are inherent in the hinge‑containing Fc cassette as the patent itself acknowledges.
4. Addition of US 5,258,698 and/or O’Reilly 1997↔1994 in combination (+ the patent’s own concession of routine linker optimization) for the Fc‑X‑X / X‑Fc‑X bifunctional “synergy” claims.
5. Addition of US 5,726,044 and LaVallie 1993 for proteolytic‑cleavage‑site limitations, and the species‑matched murine/canine variants as predictable substitutions driven by conceded preclinical need.

Motivation is not merely conceivable but expressly documented: the references identify a known production/folding/half‑life problem; the immunofusin patents identify Fc‑X as the general remedy; and the O’Reilly papers identify the two target proteins and invite combining them.

Residual uncertainty I want to be explicit about: because the verbatim granted claim set of US 8,206,718 was not available in the fetched record, this analysis reaches the claim subject matter through the Abstract, the specification, and the related continuation US 2013/0165634 (US 13/533,250 → US8703908B2) and RU2240328C2. If the granted claims contain limitations beyond those reflected there (e.g., specific SEQ ID NOs., specific host cells, or specific pharmacokinetic ranges), those should be charted separately before a final conclusion is stated. Likewise, several dates above are taken from the patent page itself (e.g., “prior art date 1998‑08‑25”) rather than independently verified publication dates for O’Reilly 1994, US 5,733,876, US 5,837,682, US 5,885,795, US 5,854,205 and US 5,258,698; those should be confirmed in a full §102 verification pass.