US 5625126

I'll search for the specific patent number and any related litigation.

Let me get the abstract, claim text, and any litigation involving this patent.

Let me try to retrieve the abstract and explicit claim text, plus check for any court dockets.

Summary — US 5,625,126

Caveat on sourcing first: I was not able to query USPTO PatentCenter/PAIR or the CAFC docket system directly. The results below come from web searches that surfaced Google Patents, uspto.report, FreePatentsOnline, Unified Patents, and citing documents. Where I could not verify something authoritatively, I say so.

Also note a date discrepancy in the prompt itself: you state "Current Date: April 26, 2026," while my operating context gives today as 2026-09-28. This does not affect any finding below, but flagging it since you asked for 2026 docket activity.

Bibliographic data (per Google Patents, the highest-quality source retrieved)

Priority lineage (as listed on the Google Patents page): 5,625,126 claims priority from US 07/810,279 (1991-12-17 → US 5,569,825); US 07/834,539 (1992-02-05 → US 5,633,425); US 07/853,408 (1992-03-18 → US 5,789,650); US 07/990,860 (1992-12-16 → US 5,545,806); US 08/053,131 (1993-04-26 → US 5,661,016); and US 08/096,762 (1993-07-22 → US 5,814,318). Later continuations/divisions from this family led to US 5,770,429; US 6,300,129; US 6,255,458; US 7,501,552; US 8,158,419; and US 8,293,480.

Minor discrepancy to note: the Google Patents page states a priority/prior-art date of 1990-08-29, while several third-party aggregator listings (e.g., Unified Patents, in the context of WO 1991000906 and WO 1992003918) show 1990-08-28 for this same document. I am not auto-correcting either value; they are simply inconsistent across sources.

Abstract

I could not retrieve a verbatim abstract from an authoritative source (the Google Patents page I retrieved does not render the abstract field). The closest thing to an abstract text I found was a machine-translated summary on a third-party academic index, which is too garbled to quote as authoritative. I would rather flag this than fabricate one.

What the specification itself says the invention "relates to" (verbatim from the Definitions section of the patent):

"the invention relates to transgenic non-human animals capable of producing heterologous antibodies, transgenes used to produce such transgenic animals, transgenes capable of functionally rearranging a heterologous D gene in V-D-J recombination, immortalized B-cells capable of producing heterologous antibodies, methods and transgenes for producing heterologous antibodies of multiple isotypes, methods and transgenes for producing heterologous antibodies wherein a variable region sequence comprises somatic mutation as compared to germline rearranged variable region sequences, transgenic nonhuman animals which produce antibodies having a human primary sequence and which bind to human antigens, hybridomas made from B cells of such transgenic animals, and monclonal antibodies expressed by such hybridomas."

Independent claims — plain language

The patent's claims section is short. The retrieved claim listing (via uspto.report grant page for 5625126) shows claims 1–5, of which claim 1 is the sole independent claim; claims 2–5 are dependent. I could not independently confirm from a USPTO-hosted source that the claim set ends at claim 5, so treat the completeness of this listing with mild caution.

Claim 1 (independent) — plain language:
A transgenic mouse whose genome contains a transgene in which the following are operably linked: multiple human V genes, multiple human D genes, multiple human J genes, a human μ constant-region gene, and at least two different non-μ human constant-region genes, together with the associated isotype switch sequences. A key structural limitation: the human μ and human γ switch sequences must sit closer to each other than they do in the naturally occurring human immunoglobulin locus (i.e., the transgene is a compressed "mini-locus"). In the mouse's lymphocytes, the transgene must undergo (a) productive VDJ rearrangement and (b) μ-to-γ isotype switching via recombination between the human μ and human γ switch sequences, with the result that the mouse produces serum containing immunoglobulins of at least three human heavy-chain isotypes in response to antigenic stimulation.

The "at least three human heavy chain isotypes" and the "closer proximity" switch-sequence limitation are the practically important narrowing elements — this is not a claim to a broadly humanized mouse, but to a mouse whose transgene is architecturally rearranged to force switch recombination.

Claim 2 (dependent on 1): The same mouse, wherein the lymphocytes additionally have an integrated human light chain transgene.

Claim 3 (dependent on 2): The same mouse, wherein the serum contains a population of IgG that includes a subpopulation of human γ1 + human κ and a subpopulation of human γ3 + human κ.

Claim 4 (dependent on 1): The same mouse, wherein the human γ switch sequences are contained within an approximately 5.3 kb HindIII fragment.

Claim 5 (dependent on 1): The same mouse, wherein the human γ switch sequences are contained within an approximately 4 kb HindIII fragment.

Claims 4 and 5 are structural/characterizational limitations tying the claimed switch region to specific restriction fragments, presumably reflecting the two switch fragments actually used to build the disclosed constructs.

Litigation / CAFC docket check

I found no evidence of any CAFC 2026 docket activity involving US 5,625,126. My attempt to run a targeted docket query hit a tool limit before returning results, so I want to be explicit: this is a negative finding from limited searching, not an authoritative "no litigation exists" conclusion.

Context that bears on the likelihood:

• The patent's anticipated expiration was 2014-04-29 per Google Patents, and its status is "Expired – Lifetime." A patent expiring in 2014 with a six-year damages lookback would have ceased to support new infringement recoveries around 2020. New 2026 appeals on this number are therefore unlikely, though not categorically impossible (e.g., a long-running case with an unusually delayed appeal, or a collateral challenge).
• Separately, US 5,625,126 is heavily cited as prior art and as background in later antibody-patent specifications and prosecution histories (e.g., as a "Humab mouse" reference cited alongside US 5,545,806, US 5,569,825, US 5,633,425, US 5,661,016, and US 5,770,429). Those citations are references, not litigation.
• If you need a definitive docket answer, the authoritative path is a direct PACER/CAFC CM/ECF search plus USPTO PatentCenter (which also carries assignment records — the Google Patents page shows a 1995-04-03 assignment recorded from Kay and Lonberg to Genpharm International, Inc.).

Uncertainty ledger

1. Verbatim abstract — not retrieved from an authoritative source; not provided here.
2. Completeness of the claim set — claims 1–5 retrieved from a secondary aggregator (uspto.report), not from a USPTO-hosted full-text source. There may be additional claims.
3. Priority date 1990-08-29 vs. 1990-08-28 — sources disagree; both reported as found.
4. No CAFC 2026 docket — negative finding from incomplete searching; not conclusive.

I'll research litigation involving US 5,625,126 across patent litigation databases.

Let me refine with litigation-specific searches.

Key lead found: GenPharm sued Abgenix (Cell Genesys subsidiary) in 1997. Let me dig into whether '126 was asserted and get case details.

Bottom line

I could not find any documented litigation in which US 5,625,126 ("'126") was specifically pleaded as an asserted patent. What the record does show is a well-documented 1994–1997 dispute between GenPharm International (the '126 assignee) and Cell Genesys / Abgenix over transgenic human-antibody mouse technology, which was resolved by a global cross-license that covered "all issued and related patent applications" — a category that would reach '126, which issued 29 April 1997, i.e. after the settlement was signed.

Because your instruction is to return only litigation for the specific number, I am flagging confidence levels explicitly below rather than overstating.

---

Case 1 — Cell Genesys, Inc. v. GenPharm International, Inc. (and Frederick Alt)

Sources: BioWorld ("Cell Genesys, GenPharm International Settle Legal Dispute, Form Partnership"); New Scientist, 1 Feb 1997, "No end in sight in the 'human' mouse war" (https://www.newscientist.com/article/mg15320670-800-no-end-in-sight-in-the-human-mouse-war/).

---

Case 2 — GenPharm International, Inc. v. Abgenix, Inc. (Cell Genesys subsidiary)

This is the case that most plausibly implicates the '126 family.

Sources: BioWorld, "Cell Genesys, GenPharm International Settle Legal Dispute, Form Partnership" (https://www.bioworld.com/articles/[485617](/patent/485617)); BioCentury, "Truce reached in transgenic mouse battle," 31 Mar 1997 (https://www.biocentury.com/article/[250931](/patent/250931)); Mary Ann Liebert, Biotechnology Law Report 1997;16:66, "GenPharm Receives U.S. Patents for Mice Producing Human Antibodies; Sues Cell Genesys Subsidiary"; Cell Genesys Form 10-K for FY1996 (https://www.getfilings.com/o0000929624-97-000349.html) and Abgenix S-1 registration statement (SEC, https://www.sec.gov/Archives/edgar/data/[1052837](/patent/1052837)/000091205702033715/a2083916zposam.htm), which state the cross-license "resolved all related litigation and claims between the parties."

---

Why '126 itself shows no standalone suit

1. Timing. '126 issued 29 April 1997, one month after the GenPharm/Cell Genesys settlement and six months before Medarex acquired GenPharm (Oct 1997, ~$62.2M). The commercial dispute was resolved before '126 could be asserted.
2. Blanket cross-license. The 1997 agreement covered "all issued and related patent applications," so the '126 application — then pending (Ser. No. 08/352,322, filed 7 Dec 1994) — fell within the settled subject matter. Abgenix's later SEC filings confirm "all of our financial obligations in connection with the cross-license were recognized in 1997."
3. No later assertion found. Post-1997, the technology moved to Medarex (HuMAb-Mouse) and Abgenix (XenoMouse), which competed commercially but, per contemporaneous reporting, operated free of IP litigation against each other under the cross-license. I found no record of '126 being asserted against a third party (e.g., a licensee or an infringer).
4. Patent is expired. Google Patents lists anticipated expiration 29 April 2014, legal status "Expired – Lifetime." Any infringement action would now be time-barred regardless.

---

What I could not confirm (stated explicitly rather than guessed)

• Docket numbers for either case — not found.
• Exact jurisdiction/venue for either case — not found.
• Whether '126 was ever named in any complaint, anywhere. I found no such pleading. Absence of evidence is not proof of absence, but I searched patent-litigation portals, the Unified Patents patent page for US-5625126-A (which shows assignee/prosecution data but no linked litigation), USPTO assignment records (assignee did change: GenPharm → Medarex lineage), and general web sources.

Data-quality warning on sources

Several aggregator citation lists render "US 5625126 A" alongside an inventor attribution of "WALDMANN HERMAN [GB]" (e.g., EPO citation lists surfaced in search). That is inconsistent with the Google Patents record, which lists inventors Nils Lonberg and Robert M. Kay and original assignee Genpharm International Inc. Treat those aggregator attributions as citation-parse errors and rely on the Google Patents bibliographic record (https://patents.google.com/patent/US5625126/en) as authoritative for '126's identity.

Recommended verification steps

To close the gaps above, the following would be needed (I could not reach them in this session):

• PACER / CourtListener RECAP party search for "GenPharm International" (1994–2000) to obtain the 1994 and 1997 docket numbers and venues.
• Courthouse News / Lex Machina / Docket Navigator patent-case search on 5625126.
• The Cell Genesys 1996 Form 10-K "Legal Proceedings" section and Abgenix S-1, both of which describe the litigation and settlement in more detail than the press releases.

If you want, I can attempt targeted docket-level lookups on the specific party names above.

I'll verify the structured ODP data against public PTAB records before writing anything up.

The ODP block says no AIA trials. Let me press a bit harder before I accept that, since well-asserted patents in this space have been litigated.

PTAB Proceedings — US 5,625,126 ("Transgenic non-human animals for producing heterologous antibodies")

Assignee of record: GenPharm International, Inc. (now within the Medarex / Bristol-Myers Squibb lineage)
Priority date: 1990-08-29 · Filed (this CIP): 1994-12-07 · Granted: 1997-04-29 · Expired: 2014-04-29
Claims on file: claims 1–5 (claim 1 independent; claims 2–5 dependent)

---

Proceedings overview

There are zero AIA trial proceedings on file against US 5,625,126 — the USPTO Open Data Portal returns no IPR, PGR, or CBM for this patent, and my web searches surfaced no PTAB proceeding, institution decision, FWD, or Federal Circuit appeal naming it; accordingly the breakdown is 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials, and the defensive posture this gives a defendant is not "hardened patent" but something more decisive: the patent expired on 2014-04-29, so the absence of PTAB activity is not a sign of an untested patent surviving challenge — it is a sign of a patent that ran out its term without ever being worth an IPR.

Confidence note. The ODP block is authoritative and my searches are consistent with it, but a search tool limitation cut my verification short before I could open PTAB E2E/PTAB Center directly. Treat "0 proceedings" as very likely correct but not exhaustively verified. I found no proceeding to invent, and I have not listed any.

⚠️ Numbering collision — do not get misled

Several search hits referring to "the '512 Patent" and "claims 1 and 13" of a '512 patent are not this patent. They are the Rovi Guides interactive television program guide patent at issue in Comcast Cable Communications, LLC v. Rovi Guides, Inc., IPR2017-00951 (and related ITC/Supreme Court briefing in No. 19-1173). Claim 13 in that case recites "an interactive television program guide implemented on the system." That is a different patent, a different technology, and a different owner. If you see "'512 patent" in a PTAB or ITC docket, check the number before relying on it.

---

Proceedings

None to enumerate. No AIA trial was ever petitioned or instituted against US 5,625,126. There is therefore no petitioner, no panel, no institution decision, no Final Written Decision, no settlement, and no § 318(a) appeal to the Federal Circuit for this patent.

Why the silence, and what it does and does not tell you

• Statutory window was the tail of the patent's life. IPR became available on 2012-09-16 (AIA § 6(c)(2)(A)). This patent expired 2014-04-29. A petitioner in 2012–2013 would have been spending a full IPR budget (median ~$350K through hearing) to take a live claim off the table with roughly 19 months of remaining term. That economics almost never supports a petition, which is the most plausible explanation for the empty docket.
• CBM was unavailable on this subject matter. Covered business method review was confined to financial-products/services claims. A transgenic-mouse claim does not qualify.
• PGR was unavailable by timing. PGR applies only to patents with an effective filing date on or after 2013-03-16. This patent's chain runs to 1990-08-29 / 1994-12-07.
• The owner was sophisticated and license-driven. Genpharm was acquired by Medarex (1997), which was acquired by Bristol-Myers Squibb (2009). This family (US 5,545,806; 5,569,825; 5,633,425; 5,661,016; 5,770,429; 5,789,650; 5,814,318; 5,877,397; 5,874,299; 6,255,458; plus continuations US 6,300,129, 7,501,552, 8,158,419, 8,293,480) was monetized through licensing and platform use, not through an assertion campaign that would draw IPRs. I found no litigation naming this patent.

Term basis (from the record, not inferred). The Google Patents record lists an "[a]nticipated expiration" of 2014-04-29 and a current legal status of "Expired - Lifetime," which equals 17 years from the 1997-04-29 grant date — the pre-URAA term.

---

Strategic summary

Claim status. Because no IPR, PGR, or CBM ever reached a Final Written Decision, there is no PTAB record canceling or sustaining any claim. At expiry, claims 1–5 all remained on the books — none canceled, none held unpatentable, but none adjudicated either. So the correct framing is: claims 1–5 are untested at the PTAB and now expired, not "sustained" and certainly not "narrowed." If a demand letter cites claims 1–5 of this patent, the problem is not that they were invalidated — it is that they lapsed on 2014-04-29.

Estoppel landscape. There is no § 315(e)(2) estoppel on this patent, because estoppel attaches only after a final written decision. There is no petitioner, no RPI, and no privy. Practically, though, estoppel is moot: with the patent expired, the § 102/§ 103 invalidity question is academic unless someone is trying to reach pre-2014-04-29 conduct. If they are, your defenses are § 286's six-year damages lookback (a suit filed in 2026 could reach, at the very earliest, conduct from 2020 — after expiry), laches/notice under § 287, and the ordinary invalidity defenses under §§ 101/102/103/112 in district court, where the PTAB's narrower § 311(b) art limits do not apply. No art is foreclosed to you by estoppel.

Pattern signals. No repeat petitioner, no unified defensive aggregator, and no patent-owner appellate aggression — there is nothing to pattern-match. The relevant pattern here is generational: the Genpharm/Lonberg minilocus family has been cited as foundational prior art by later antibody-platform patents (Regeneron, Kymab, Ablexis, Merus, XOMA filings routinely cite US 5,625,126), so its significance today is as prior art and licensing history, not as an assertion weapon.

---

Recommended next steps

1. If a demand letter cites US 5,625,126, lead with expiry, not validity. The patent expired 2014-04-29. Quote the record: legal status "Expired - Lifetime," "[a]nticipated expiration 2014-04-29" — US5625126A on Google Patents.
2. Run the § 286 math before responding on the merits. Any recovery is limited to infringement within six years of the complaint. If you are reading this on 2026-09-28, the earliest reachable conduct dates to roughly 2020-09-28 — six years after the patent expired. Ask the sender to identify an accused act with a date, and watch the theory collapse.
3. Do not build a defense around the PTAB. There is no FWD to cite and no canceled claim to point to. If the sender claims PTAB outcomes, that is a factual error you can correct.
4. Verify independently before you file anything. Confirm no proceeding exists by running the patent number through PTAB Center (PTAB E2E successor) and the USPTO Patent Center, and confirm the expiry/reinstatement history in the Patent Center "Transactions" view. My search verification was cut short by a tool limit; the ODP block is the authoritative word and it shows nothing, but a one-minute E2E check closes the loop.
5. Watch the sibling patents, not this one. The live assertion risk in this family sits with later-issued continuations (US 6,300,129; 7,501,552; 8,158,419; 8,293,480) and the Regeneron/Kymab/Ablexis-era patents that cite this one. No PTAB activity here says nothing about the freedom-to-operate picture on those.

Bottom line: you are not facing a patent that survived two IPRs. You are facing an expired 1997 patent (a 1990 priority CIP) that was never challenged at the PTAB and lapsed on 2014-04-29. Any demand letter citing US 5,625,126 should be answered with the expiration date and a § 286 damages challenge, not with an IPR budget.

Sources: USPTO ODP "PTAB proceedings on file" block (canonical, negative); US5625126A, Google Patents (expiry, grant, family, claim text); PTAB E2E/PTAB Center and CourtListener referenced for independent confirmation.

I'll research the assignment record for US 5625126 using the available sources.

Let me dig into the specific assignment records and any litigation history.

US 5625126 — Ownership Reconstruction

Verification status up front: the live USPTO Assignment Center (https://assignmentcenter.uspto.gov/) and its legacy mirror (https://assignment.uspto.gov/patent/index.html) are session-driven search applications that my search tools could not query directly. I was therefore unable to retrieve reel/frame numbers, correspondent-of-record entries, or execution/recording dates for this patent's recorded assignments. I am not fabricating reel/frame identifiers. Everything below is reconstructed from Google Patents legal events, SEC filings (Medarex 10-K/8-K, BMS 8-K/Schedule TO), and published case material, and is flagged as verified vs. unverified. Treat the reel/frame column as a to-do, not a finding.

Inventors

Unusual patterns: none detected. The classic "all inventors exit within 12 months of filing" pre-fire-sale tell is absent — the inventors were still with the organization when the portfolio was acquired by Medarex in October 1997 and are named on later continuations filed in the 1994–1998 window (08/352,322; 08/544,404; 08/758,417). The relevant corporate event here is a strategic acquisition of a going concern, not an inventor exodus.

Assignment note: the inventors did execute an inventor-to-employer assignment to Genpharm International, Inc. Google Patents' legal-events feed records a reassignment entry dated 1995-04-03 naming "KAY, ROBERT M." and "LONBERG, NILS" as assignors and GENPHARM INTERNATIONAL, INC. as assignee (recorded ~4 months after the 1994-12-07 filing of application 08/352,322). Correspondent not retrieved.

Original assignee

Genpharm International, Inc. — named on the issued patent (front-page assignee; Google Patents "current assignee" field also shows Genpharm, which reflects original data, not the present owner of record).

• Business: biotech research and platform licensing. Genpharm was founded 1989 and commercialized the HuMAb-Mouse — a transgenic mouse bearing human immunoglobulin transgenes. Its revenue model was out-licensing to pharma partners (Eisai, Centocor, LeukoSite collaborations came with the company into Medarex).
• Did it ship a product embodying the claims? No. The claims of US 5625126 cover transgenic non-human animals / immunoglobulin transgenes, not a marketed therapeutic. Genpharm's output was the mouse itself plus licences. There is no product-in-commerce nexus that would support an "operating-company assertion" label.
• Status: acquired by Medarex, Inc. (announced May 1997; completed October 1997, consideration up to ~$65M in Medarex stock plus contingent payments keyed to licence fees). Genpharm continued to exist as a Medarex subsidiary — it appears as a contracting party (521 Cottonwood Drive, Milpitas, CA 95035) in a 2007 Medarex/Celldex amendment, jointly described with Medarex as "Medarex." Whether the Genpharm legal entity still exists after the 2009 BMS acquisition is unclear; I could not confirm a dissolution filing.
• Litigation context (not an NPE suit): Cell Genesys/Abgenix sued GenPharm in February 1994 for trade-secret misappropriation; the dispute settled via a 26 March 1997 Release and Settlement Agreement and cross-licence among Cell Genesys, Abgenix, Xenotech L.P., Japan Tobacco and GenPharm (listed as Ex. 10.73 in the Medarex 10-K). Issuance of US 5625126 itself triggered a $7.5M payment from Xenotech to Medarex. This is a licence/cross-licence posture, not an assertion posture.

Assignment timeline

The entries below are the ownership events I could verify. Reel/frame and correspondent fields are marked not retrieved rather than guessed.

• 1994-12-07 (filed) / — Reel not retrieved

  • Conveyance: n/a — application 08/352,322 filed as a continuation in the 1990-08-29 priority chain (US 07/810,279 et al.)
  • Assignor: n/a
  • Assignee: Genpharm International, Inc. (applicant)
  • Correspondent: not retrieved
  • Context: initial filing; no assignment yet of record.

• 1995-04-03 (recorded) — Reel not retrieved

  • Conveyance: Assignment (inventor to employer)
  • Assignor: Robert M. Kay and Nils Lonberg
  • Assignee: Genpharm International, Inc.
  • Correspondent: not retrieved — this is the single most useful field for NPE screening and it was not obtainable with the tools available. Because this is an ordinary inventor-to-employer instrument, a recurrence of this correspondent on later links would be the only thing worth flagging, and I have no such recurrence to report.
  • Context: standard employment assignment; no NPE relevance.

• 1997-10 (corporate closing; USPTO recording not verified) — Reel not retrieved

  • Conveyance: Acquisition / merger of GenPharm International, Inc. into Medarex, Inc. (Medarex issued ~3.5M shares plus contingent consideration, up to ~$65M)
  • Assignor: GenPharm International, Inc. shareholders
  • Assignee: Medarex, Inc. (NASDAQ: MEDX; Princeton, NJ)
  • Correspondent: not retrieved
  • Context: strategic acquisition of an operating antibody-discovery company. Note: because some family members continued to issue in later years under the "Genpharm International, Inc." assignee name (e.g., US 7,041,871; US 8,233,877; US 8,293,480) while others issued under "Medarex, Inc." (US 7,501,552; US 8,158,419), the recorded ownership of this family is split across the two entity names — worth reconciling against Assignment Center rather than assuming a single clean transfer.

• 2009-09-01 (merger effective; USPTO recording not verified) — Reel not retrieved

  • Conveyance: Merger / change of control. BMS's subsidiary Puma Acquisition Corporation merged into Medarex; Medarex survived as a wholly-owned subsidiary of Bristol-Myers Squibb Company (Certificate of Merger filed 2009-09-01 with the NJ Department of the Treasury; ~$2.4B, $16.00/share; Schedule TO filing fee valuation $2,503,857,888)
  • Assignor: Medarex, Inc. shareholders
  • Assignee: Bristol-Myers Squibb Company (indirectly, via surviving subsidiary Medarex, Inc.)
  • Correspondent: not retrieved
  • Context: large-cap pharma acquisition. Because Medarex was the surviving corporation, the patent may still stand of record in the name of Medarex, Inc. (as appears on later-issuing family members) with BMS as ultimate parent, rather than in BMS's own name. Confirm which in Assignment Center.

If Assignment Center returns nothing beyond the 1995-04-03 inventor assignment, the correct reading is not "BMS definitely owns it directly" — it is that the 1997 and 2009 transactions were equity acquisitions of the corporate holder (GenPharm, then Medarex), which frequently go unrecorded at the USPTO because the named owner never changed.

Timeline diagram

timeline
    title Ownership of US 5625126
    1990 : Priority application filed
    1994 : Continuation filed by Genpharm
    1995 : Inventors assign to Genpharm
    1997 : Patent issued on April 29
         : Genpharm acquired by Medarex
    2009 : Medarex acquired by Bristol-Myers Squibb
    2014 : Patent expires April 29

NPE / troll-pattern signals

1. Shell-entity transfer — not present. No "IP / Holdings / Ventures / Licensing" LLC anywhere in the chain. Owners of record across the family are Genpharm International, Inc. (a California biotech), Medarex, Inc. (NJ public company, 707 State Road, Princeton NJ), and Bristol-Myers Squibb Company (Delaware, NYSE: BMY). No single-purpose LLC, no registered-agent service address, no Delaware/Texas shell in the reconstructed chain.

2. Known asserter in the chain — not present. No match to Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or Spangenberg entities. I found no Unified Patents or RPX listing tying US 5625126 to a high-frequency plaintiff. Caveat: I could not directly query the RPX/Unified directories in this session; this is a "nothing found," not a confirmed negative.

3. Repeat correspondent across the chain — unclear / not retrievable. No correspondent-of-record fields were obtainable. This is the most important unresolved field. For an operating-company chain, the expected pattern is the original prosecution firm (Genpharm's counsel, later Medarex/BMS in-house or outside counsel) filing the 1995 inventor assignment and any later merger records — i.e., continuity of the same firm, which would be a neutral-to-exculpatory finding. Any newly appearing attorney showing up on a post-2009 record would need explanation.

4. Cascading transfers — not present. Two corporate events across twelve years (1997, 2009), each a publicly filed, disclosed acquisition of a going concern with press releases and SEC filings. No chained LLCs, no <24-month cluster, no shared registered-agent addresses.

5. Pre-litigation transfer — not present. The only enforcement-adjacent litigation (Cell Genesys/Abgenix v. GenPharm, filed February 1994) preceded the patent's issuance and ran against the original assignee; it ended in a March 1997 settlement and cross-licence. No assignment sits within 6 months before an infringement suit naming US 5625126, and I found no infringement suit naming this patent in the material reviewed.

6. Bankruptcy fire-sale — not present. Neither GenPharm (acquired for up to $65M in stock plus contingent consideration) nor Medarex (acquired for ~$2.4B in cash) went through Chapter 7/11. No Kodak/Nortel-style asset sale.

7. Privateering — not present. No operating company is funding an NPE to assert this patent. Notably, the enforcement energy in this technology space ran the other way: the 1997 cross-licence neutralised the Cell Genesys/Abgenix/Xenotech challenge, and Medarex ran the platform as a collaboration and licensing business (Celldex Research and Commercialization + Assignment and License Agreements, 2004, amended 2007; FibroGen options; Kirin collaboration). Licensing, not privateering.

8. Defensive aggregator — not present (but the inverse of the signal applies). The chain does not terminate at RPX, AST, LOT, Unified, or OIN. However, the patent is effectively neutralised for a different reason: it expired 2014-04-29 (17 years from the 1997-04-29 grant; Google Patents status "Expired – Lifetime," with anticipated expiration recorded 2014-04-29). An expired patent is not an assertable asset.

Verdict

Defensive / non-asserting — i.e., not an NPE asset, and not asserted.

The chain is GenPharm International, Inc. (inventor assignment recorded 1995-04-03, assignors Kay and Lonberg) → Medarex, Inc. (October 1997, up to ~$65M) → Bristol-Myers Squibb Company (merger effective 2009-09-01, ~$2.4B, Medarex surviving as a wholly-owned BMS subsidiary). Every signal on the NPE checklist is either absent or attributable to ordinary corporate M&A, and the patent expired 2014-04-29, so no assertion pathway exists; its actual deployment was cross-licensing and platform partnering (the 1997 Cell Genesys/Abgenix/Xenotech/Japan Tobacco settlement and cross-licence, and the $7.5M Xenotech payment triggered by this patent's grant).

Two qualifications you should resolve before relying on this: (a) I could not obtain reel/frame or correspondent-of-record fields, so I cannot rule out an unrecorded or late-recorded link — verify at USPTO Assignment Center by patent number 5625126; (b) confirm whether the surviving owner of record is Medarex, Inc. or Bristol-Myers Squibb Company, since the family shows both entity names on later-issuing members, which suggests a split-ownership or partial-recording situation rather than one clean chain.

I'll research US Patent 5625126 and its cited prior art references.

Let me gather the full list of cited references from the patent's front page.

Let me retrieve the complete "References Cited" list from the patent's front page.

Analysis of US 5,625,126 B1 and Its Cited Prior Art

1. Patent identification (as literally reported)

Transparency note on sources. I could not directly query the USPTO's own PatentCenter/Patent Full‑Text (PatFT) interface from this session; the reference data below is drawn from the Google Patents record for US5625126A and the EveryPatent mirror (http://www.everypatent.com/comp/pat5625126.html). I have not verified in the paper file whether each item is an examiner citation (PTO‑892) or an applicant Information Disclosure Statement citation, and the front‑page list I could retrieve appears short. Treat the list below as the cited references, not as a confirmed exhaustive list of the examiner's applied art.

---

2. The "References Cited" on the 5,625,126 front page

U.S. Patent Documents

Foreign Patent Documents

"Other References" (non-patent literature) — selected

Alt et al., TIG (Aug. 1985); Berman et al., EMBO J. 7:727‑738 (1988); Berton et al., PNAS 86:2829‑2833 (1989); Bollag et al., Annu. Rev. Genet. 23:199‑225 (1989); Bruggemann et al., Eur. J. Immunol. 21:1323‑1326 (1991); Bruggemann et al., PNAS 86:6709‑6713 (1989); Bucchini et al., Nature 326:409‑411 (1987); Capecchi, TIG 5:70‑76 and Science 244:1288‑1292 (1989); Coffman et al., J. Immunol. 136:949‑954 (1986) and 139:3685‑3690 (1987); Doetschman et al., Nature 330:576‑578 (1987); Durdik et al., PNAS 86:2346‑2350 (1989); Esser & Radbruch, EMBO J. 8:483‑488 (1989); Ferrier et al., EMBO J. 9:117‑125 (1990); Forni, Eur. J. Immunol. 20:983‑989 (1990); Gerstein et al., Cell 63:537‑548 (1990); Goodhardt et al., PNAS (1987).

---

3. § 102 framework applied to this patent

Because the priority date is 1990‑08‑29, an item is only §§ 102(a)/(b) prior art if publicly available before that date (or before the applicant's date of invention). Applying that rule to the cited list:

• Qualify as pre‑priority art: US 4,204,244 (1980); EP 0315062 (May 1989); WO 90/04036 (Apr 19, 1990); and essentially all the 1985‑1990 journal articles (Alt, Berman, Bruggemann 1989, Bucchini, Capecchi, Durdik, Goodhardt, etc.).
• Post‑date the priority date, so NOT §§ 102(a)/(b) art (at best § 102(e), and only if "by another" and filed before the invention): US 5,175,384 (1992); US 5,434,340 (1995); WO 90/12878 (Nov 1990 — marginally later); WO 91/00906 (Jan 1991); WO 91/10741 (Jul 1991); WO 92/03918 (Mar 1992).
• Applicant's own work (not "by another," therefore not § 102 prior art at all): WO 92/03918 and the Krimpenfort Genpharm patents to the extent they share inventorship/obligation of assignment.

I also cannot fully map these references to the claims of 5,625,126, because the claim text was not available in the material I retrieved (the record I obtained truncates before the claims). I therefore cannot state a definitive claim‑by‑claim anticipation position, and any statement below is a potential § 102 theory only.

---

4. Most relevant prior art and potential § 102 theories

---

5. Bottom line

1. The most relevant (potentially anticipatory) pre‑priority references cited on the face of 5,625,126 are the journal articles Bruggemann et al. 1989 (PNAS 86:6709), Bucchini et al. 1987 (Nature 326:409), Goodhardt et al. 1987 (PNAS 84:4229), Durdik et al. 1989 (PNAS 86:2346) and Gerstein et al. 1990 (Cell 63:537), together with the patent publication WO 90/04036.
2. None of them, on their face, discloses every limitation of a claim directed to a transgenic non‑human animal bearing unrearranged human heavy‑chain V‑D‑J and light‑chain V‑J transgenes that undergo functional recombination incorporating a D segment and isotype switching to multiple human C_H isotypes. That is the classic distinction used to defeat § 102 here (each reference supplies only an isolated piece).
3. Several cited items (WO 91/00906, WO 91/10741, WO 92/03918, US 5,175,384, US 5,434,340) post‑date the 1990‑08‑29 priority date and therefore are not §§ 102(a)/(b) art; WO 92/03918 is the applicant's own counterpart and is not "by another" at all.
4. I was unable to retrieve the full claim set of 5,625,126 in this session and could not distinguish examiner‑applied art from applicant IDS citations; consequently the per‑claim § 102 mapping above is a potential theory, not a verified office‑action record. To produce a definitive anticipation chart, one would need the USPTO PatentCenter file wrapper (PTO‑892 "Notice of References Cited" vs. IDS transmittals) and the full claim text at https://patents.google.com/patent/US5625126/en (claims section).

Caveats honored: I have reported every patent number and identifier exactly as retrieved (e.g., 4,204,244; 5,175,384; WO 90/04036; WO 92/03918) without normalization, and I have explicitly flagged items I could not confirm rather than guessing.

I'll research the prior art cited against US 5,625,126 and related references to build the § 103 analysis.

I'll pull the cited prior art listed on the patent page and related prosecution/interference references.

Let me pull the actual cited-art lists and the claim set for US 5,625,126.

Obviousness Analysis — U.S. Patent 5,625,126 under 35 U.S.C. § 103

Patent: US 5,625,126 A ("Transgenic non-human animals for producing heterologous antibodies"); inventors Nils Lonberg and Robert M. Kay; assignee Genpharm International Inc.; application 08/352,322 filed 1994-12-07; prior-art date 1990-08-29; granted 1997-04-29; status Expired – Lifetime.
Source: https://patents.google.com/patent/US5625126/en

---

0. Scope, sources, and caveats

What "the Prior Art section of this page" comprises. On the Google Patents page the prior-art signal is (a) the "Prior art keywords" field (human; transgene; gene; sequences; heavy chain), (b) the non-patent literature cited in the specification and in the parallel filings of the same family, and (c) the "Cited By" material (which is later art and therefore not § 102/§ 103 art against this patent). The specification's own reference list and the family's citation lists are the substantive prior art usable here. The most useful corroborated list I could retrieve is the search report of the European counterpart EP 0 814 159 A3 (GenPharm, "Transgenic non-human animals capable of producing heterologous antibodies"), which cites the same core art:

• Nussenzweig, M.C. et al., J. Exp. Med. 167:1969–1974 (1988) — "Allelic exclusion in transgenic mice carrying mutant human IgM genes"
• Durdik, J. et al., PNAS USA 86:2346–2350 (April 1989) — "Isotype switching by a microinjected mu immunoglobulin heavy chain gene in transgenic mice"
• Buttin, G., Trends in Genetics 3(8):205–206 (Aug. 1987) — "Exogenous Ig gene rearrangement in transgenic mice: a new strategy for human monoclonal antibody production?"
• Nikaido, T. et al., Nature 292:845–848 (1981) — "Switch region of immunoglobulin C-mu gene is composed of simple tandem repetitive sequences"
• Brüggemann, M. et al., PNAS USA 86:6709–6713 (1 Sept. 1989) — "A repertoire of monoclonal antibodies with human heavy chains from transgenic mice"
• Ott, D.E. et al., EMBO J. 6(3):577–584 (1987) — "Immunoglobulin heavy chain switch region recombination within a retroviral vector in murine pre-B cells"
• Taussig, M.T. (cited as "A")
  PDF: https://patentimages.storage.googleapis.com/df/62/66/8d4b94493f8af1/EP2314625A1.pdf (EP 0 814 159 A3 search report reproduced therein; see also https://patents.google.com/patent/[US5545807](/patent/US5545807) for the Babraham family's citations)

Additional art relied on expressly in the specification of US 5,625,126 and in sibling filings:

• Brüggemann et al., Eur. J. Immunol. 21:1323–1326 (1991) — human antibody production from 100 kb of the human IgH locus
• Bucchini, D. et al., Nature 326:409–411 (1987) — unrearranged chicken λ transgene
• Goodhart et al., PNAS USA 84:4229–4232 (1987) — unrearranged rabbit κ gene
• Alt, F. et al., TIG 1:1 (Aug. 1985); Storb et al., Immunol. Rev. 89:85 (1986); Storb, in Immunoglobulin Genes pp. 303–326 (Academic Press, 1989)
• Neuberger et al., Nature 338:350–352 (1989) — isotype exclusion/transgene down-regulation in Ig-λ transgenic mice
• Oettinger, M. et al., Science 248:1517–1523 (1990) — recombination signal sequences are conserved and functionally interchangeable across loci and species
• Mills, F.C. et al., Nucleic Acids Res. 18:7305–7316 (1991); Sideras, P. et al., Intl. Immunol. 1:631–642 (1989) — human IgM/IgG switch-region sequences
• Rothman et al., Intl. Immunol. 2:621–627 (1990); Reid et al., PNAS 86:840–844 (1989); Stavnezer et al., PNAS 85:7704–7708 (1988) — sterile transcripts/switch regulation
• Linton et al., Cell 59:1049–1059 (1989); Kabat et al., Sequences of Proteins of Immunological Interest (1991)
• Ausubel et al., Current Protocols in Molecular Biology (1987) — the specification's own molecular-biology how-to citation
• Patent art visible on the same page's citation graph: US 4,736,866 (Harvard, transgenic non-human mammals), US 4,816,567 / EP 0 125 023 (Genentech, recombinant/chimeric immunoglobulins), EP 0 184 187 (Teijin, mouse-human chimaeric heavy chain)

Caveats I want to be explicit about.

1. The full claim text is not in the supplied page extract (the text truncates in the "B-cells preferabl…" passage). My claim-element mapping below is drawn from the "Summary of the Invention" text that is supplied, from the figure list, and from the well-known claim families of this patent. Where I map to a claim, I say so; where I am inferring, I say so.
2. I cannot verify from retrieved material whether Durdik's microinjected μ gene switched cis (to a downstream C_H on the same transgene) or trans (to the endogenous mouse C_H locus). The title establishes only that isotype switching of a microinjected μ heavy-chain transgene occurred in transgenic mice. I treat it accordingly, and this ambiguity is itself analytically important (see § 4).
3. This is an analytical reconstruction, not the actual prosecution record. I have no verified record of a reexamination or successful § 103 rejection of US 5,625,126 itself. I do have verified evidence that the same core reference (Brüggemann) was used by examiners to reject later, related claims on obviousness grounds — see the Regeneron petition appendix (https://www.supremecourt.gov/DocketPDF/17/17-1616/48313/20180525141921206_Regeneron%20Petition%20Appendix%20FINAL%20to%20file.pdf) and the Merus counterclaim (http://www.ipwatchdog.com/wp-content/uploads/2017/07/Merus-counterclaim.pdf), which quote the Examiner as finding that Brüggemann teaches "that an attractive alternative of the mice would be to replace the mouse Ig locus with the human Ig locus."
4. Critical-date discipline. Because the effective date is 1990-08-29, art published after that date is only available if a particular claim cannot benefit from the 1990-08-29 priority — e.g., Brüggemann Eur. J. Immunol. 21:1323 (1991), WO 91/00906 (published 1991-01-24), Bruggemann & Neuberger (1996), Taylor et al. (1992), Lonberg et al. Nature 368:856 (1994), and the 1991-12-17/1992 co-pending applications. Those cannot be used in the primary case; they matter only as corroboration or as § 102(a)/(e) art for claims that fail to inherit the 1990 date.

---

1. The legal framework applied

Graham v. John Deere Co., 383 U.S. 1 (1966): scope and content of the prior art; differences between prior art and claims; level of ordinary skill; secondary considerations. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007): a combination is obvious if (i) the references teach or suggest the claimed elements, (ii) there is a motivation or reason to combine with a reasonable expectation of success, and the combination is "more than a predictable variation" or use of a known technique to improve a similar device; "obvious to try" applies where there is a finite number of identified, predictable solutions. In biotechnology, the Federal Circuit and the Office nonetheless retain a robust reasonable-expectation-of-success requirement (In re O'Farrell, 853 F.2d 894 (Fed. Cir. 1988); Amgen v. Chugai, 927 F.2d 1200 (Fed. Cir. 1991)) — a point that matters enormously here because the specification itself concedes that prior attempts had produced "variable," "incomplete or minimal rearrangement."

PHOSITA here: an immunologist/molecular biologist with an M.D./Ph.D. or Ph.D. plus 2–4 years, having access to standard transgenic-mouse technology (Hogan et al., Manipulating the Mouse Embryo (1986); US 4,736,866), genomic cloning (Sambrook; Ausubel 1987, cited by the patent itself), and Ig locus sequence information (Kabat; Nikaido 1981 for switch repeats).

---

2. What the claims require (element decomposition)

From the supplied specification text, the claimed subject matter falls into these buckets:

Bucket E is a molecular-biology method claim; Buckets A–D are the commercially significant composition/animal claims.

---

3. Combination 1 — Bucket A (unrearranged human heavy-chain transgene that rearranges and incorporates D)

References: Brüggemann 1989 PNAS 86:6709; Bucchini 1987 Nature 326:409; Goodhart 1987 PNAS 84:4229; Alt 1985 TIG 1:1; Storb 1989.

Motivation. The specification itself frames the problem: human monoclonal antibodies are needed, human self-tolerance blocks immunizing humans against human antigens, and human hybridoma technology had failed. Brüggemann 1989 expressly proposed that "the introduction of human immunoglobulin gene segments in their unrearranged configuration into the germ line of mice might allow the production of a repertoire of human antibodies," usable "for the production of human monoclonal antibodies against human antigens." That is the stated motivation for the whole claimed programme.

Reasonable expectation of success. High for Bucket A's broadest genus: Brüggemann 1989 is a working example of exactly this construct class. Adding more V and D segments to a minilocus is a matter of degree, and the art already taught that Ig gene segments recombine in mice (Bucchini; Goodhart; Alt; Storb).

Where the art is weaker. Brüggemann's own result was incomplete: the specification concedes that "the results of such experiments, however, have been variable, in some cases, producing incomplete or minimal rearrangement of the transgene." Whether the claimed D-incorporation thresholds (>10%, >30%, >50% of transcripts containing recognizable D) are obvious depends on whether the art established that a human D-containing minilocus would reliably incorporate D in the mouse. I did not retrieve data showing that Brüggemann 1989 reported robust human D incorporation in expressed transcripts; if it did not, these numerical limitations are the strongest nonobviousness foothold in Bucket A. (Note the patent's own supporting data appear in its Tables 4, 6, 7 and 8.)

---

4. Combination 2 — Bucket B (isotype switching of a heterologous heavy-chain transgene)

References: Durdik 1989 PNAS 86:2346; Nikaido 1981 Nature 292:845; Ott 1987 EMBO J. 6:577; Buttin 1987 TIG 3:205; later, Mills 1991 / Sideras 1989 (sequences) and Rothman 1990 / Reid 1989 / Stavnezer 1988 (sterile transcripts).

The combination.

1. Brüggemann 1989 supplies a transgenic mouse bearing an unrearranged human heavy-chain minilocus with human C_H — the "donor" scaffold.
2. Nikaido 1981 supplies the DNA sequence/structure of the Cμ switch region (simple tandem repeats), i.e., the sequence an artisan would need to include to enable switching; Mills 1991/Sideras 1989 supply the human Sγ, Sα and Sε sequences. The specification itself cites these as the standard source of switch sequences.
3. Durdik 1989 supplies the operative teaching that a microinjected μ immunoglobulin heavy-chain gene undergoes isotype switching in transgenic mice — i.e., the mouse's switch machinery acts on an exogenous heavy-chain transgene.
4. Ott 1987 supplies the mechanism-level teaching that switch-region recombination occurs on an IgH construct in murine pre-B cells, corroborating that switch recombination is a function of the included switch sequences and the host machinery, not of chromosomal location.
5. Buttin 1987 supplies the express motivation: exogenous Ig gene rearrangement in transgenic mice as a strategy for human monoclonal antibody production.

Obviousness rationale. The claim elements are classic "arrangement of old elements with each performing the same function it was known to perform": unrearranged human V/D/J (Brüggemann) + human Cμ and a downstream human Cγ with their natural switch regions (Nikaido/Mills) + the recognition that the host performs switch recombination on transgene-encoded heavy chains (Durdik/Ott). The motivation is set out in the specification itself: surface IgM is required for B-cell maturation, but "isotypes other than IgM are desired for therapeutic uses"; Fc-mediated effector functions (mast-cell/basophil binding via Fcε, complement binding via Fcγ/Fcα) make isotype variation desirable. "Obvious to try" is available to the examiner because the number of candidate constructions is small and well identified (place human Cγ1, Cγ3 etc. downstream of Sμ, or grafted switch sequences, in the minilocus).

The genuine nonobviousness counterargument (and it is strong). The specification states as a present-tense fact of the art: "although transgenic animals have been generated that incorporate transgenes encoding one or more chains of a heterologous antibody, there have been no reports of heterologous transgenes that undergo successful isotype switching." If accepted as an accurate statement of the 1990 art, it defeats the "reasonable expectation of success" prong for Bucket B, because the art's own failure is evidence that the result was not predictable. Key distinctions a patent owner would press:

• Durdik involved a mouse μ gene, not a human heterologous transgene with human switch sequences. The relevant question for § 103 is not whether switching can happen at all, but whether a PHOSITA would have expected human switch regions (a heterologous element, in a mouse recombinase environment) to recombine at usable frequency and in the correct temporal order. The specification itself flags this as uncertain: "the human switch sequences are not [optimized] … a human transgene construct … may produce a higher frequency of isotype switching events if it incorporates switch sequences similar to those that occur naturally in the mouse heavy chain locus."
• If Durdik's switching was trans-switching to the endogenous mouse locus (which is consistent with a microinjected mouse μ gene lacking a downstream C_H in the transgene), then Durdik supports only the trans-switch/chimeric-antibody embodiments, and does not teach cis-switching within a human transgene — the functional core of Bucket B.
• Nussenzweig 1988 (J. Exp. Med. 167:1969) teaches allelic exclusion with mutant human IgM genes and cuts against an expectation that a human transgene would be expressed, rearranged and switched efficiently alongside an intact endogenous mouse locus — the same teaching-away direction as Neuberger 1989 (Nature 338:350, transgene down-regulation).
• Cf. In re O'Farrell and Amgen v. Chugai: "try and see" is not a reasonable expectation of success where the outcome is unpredictable in fact.

Net assessment for Bucket B. A prima facie § 103 case over Brüggemann + Durdik + Nikaido/Mills + Buttin is articulable and, if the examiner treats Durdik as teaching cis-switching of a transgene-encoded heavy chain, likely to be made. But the case is vulnerable on reasonable expectation of success, and the patent owner has unusually strong objective indicia (see § 7).

---

5. Combination 3 — Bucket C (human HC + human κ transgenes in an endogenous-Ig-disrupted host)

References: Brüggemann 1989 (human HC transgene; the 1996 Brüggemann & Neuberger review expressly teaches "replace the mouse Ig loci with the human Ig locus … it might be also possible to retain and exploit any possible regulatory sequences in the mouse loci that are located distal to the protein coding regions"); Neuberger 1989 (transgene down-regulation/competition); US 4,736,866 (Leder; transgenic non-human mammals); Alt 1985/Storb 1986/Storb 1989 (endogenous locus structure and rearrangement); the specification's own citation of positive-negative selection homologous recombination (the Capecchi/Mansour line of work) for functionally disrupting J_H and J_κ.

Motivation. Two well-documented reasons to knock out the endogenous loci rather than merely add transgenes: (i) allelic exclusion/competition — an intact endogenous locus suppresses transgene rearrangement and expression (Nussenzweig 1988; Neuberger 1989); (ii) the therapeutic requirement of fully human antibodies rather than chimeras mixed with mouse chains. The two technologies to be combined — (a) germline transmission of human Ig miniloci in mice (Brüggemann 1989) and (b) targeted disruption of a chosen endogenous allele by homologous recombination in ES cells (Leder '866 provides transgenic mammals; positive-negative selection was established art by 1990) — are from unrelated fields but their combination is the paradigmatic "known technique to improve a similar device" (KSR).

Obviousness rationale and caveat. Note the timing problem: whether the double-knockout/double-transgenic claims were obvious as of 1990-08-29 depends on the state of ES-cell gene targeting as of that date. The patent's own disclosure frames the disruption work as within the invention, and its cited support is the general positive-negative-selection literature. If the examiner's art shows gene targeting in murine ES cells with germline transmission was routine by August 1990, the combination is obvious; if not, these claims retain a nonobviousness argument on the second Graham factor. I cannot resolve that date question from the retrieved materials, and I flag it rather than assert a conclusion.

Note also that Brüggemann & Neuberger (1996), Taylor et al. (1992), Lonberg et al. (1994) and WO 91/00906 are all post-critical-date and cannot carry the primary rejection for claims entitled to the 1990-08-29 date. They are relevant only if priority fails for a given claim.

---

6. Other claim buckets

Bucket D (B-cells, hybridomas, monoclonal antibodies, screening assays). Once the animal is obvious, the derived products and screening methods follow as a matter of routine practice. The hybridoma technology was standard (the specification itself opens by describing priming an animal, harvesting B cells and generating a hybridoma library). The screening method of contacting a hybridoma supernatant with (i) immobilized predetermined antigen and (ii) an anti-human-constant-region antibody that does not cross-react with murine constant regions is a straightforward ELISA design choice — the specification's own wording ("typically … immobilized by adsorption onto a solid substrate") shows the conventionality. Obvious over Brüggemann 1989 (which produced monoclonal antibodies with human heavy chains from transgenic mice) + routine assay practice.

Bucket E (XhoI/SalI-NotI concatemerization method). XhoI (C^TCGAG) and SalI (G^TCGAC) produce compatible cohesive ends — this was elementary, textbook molecular biology (Sambrook; Ausubel 1987, cited by the patent). Using complementary XhoI/SalI termini flanked by non-cleaved NotI sites to concatemerize large restriction fragments in an ordered fashion is a predictable design choice with a finite set of obvious alternatives; expect a § 103 rejection over Sambrook/Ausubel in view of any reference teaching ordered ligation of large inserts. Whatever nonobviousness resides in this family resides in the specific constructs, not in the cloning strategy.

Bucket F (somatic mutation/secondary repertoire). Linton 1989 (Cell 59:1049) and Kabat 1991 describe the switch to non-IgM isotypes and the accumulation of V-region point mutations in the secondary repertoire as the normal, described biology of a humoral response. A claim whose only additional element is "the variable region comprises somatic mutation relative to germline rearranged variable region sequence" is, standing alone, obvious as the expected consequence of immunizing an animal that makes an antigen-driven immune response — unless the claims tie the mutation to a non-obvious level/frequency or to a specific transgene architecture (the patent's FIGS. 66–68 report ~1.1–2.2% mutation frequencies in transgene V regions, which can be argued as unexpectedly high). This is the sort of quantitative limitation where § 103 outcomes turn on whether the prior art provides any guidance, and I have not retrieved art establishing an expected transgene V-region mutation frequency in mice.

---

7. Secondary considerations (Graham factor 4) — where the patent wins

These are the strongest rebuttals to any of the prima facie cases above, and they are well documented:

1. Failure of others / long-felt need. The specification's own admissions are double-edged: they supply motivation (helpful to the examiner) but also establish that the art had repeatedly tried and failed ("variable … incomplete or minimal rearrangement"; "no reports of heterologous transgenes that undergo successful isotype switching"). Under In re O'Farrell, the failure of others is evidence of nonobviousness.
2. Skepticism toward miniloci. Contemporary commentary in the field (e.g., material retrieved describing the minilocus approach as having "theoretical serious drawbacks, that diversity introduced by including only a small number of V, D, and J genes would be insufficient," with published studies apparently confirming impaired B-cell development in minilocus mice) supports a teaching-away from the claimed minilocus architecture.
3. Commercial success and industry adoption. The claimed technology underpinned the "HuMAb-Mouse" platform — licensed to Medarex/Bristol-Myers Squibb and used to generate approved human therapeutic antibodies. See the bibliometric analysis retrieved identifying US 5,625,126 as the highest-degree node (290 in-degree) in the relevant citation network and linking it to WO 1992003918 A1, US 5,633,425, US 5,789,650, US 5,814,318, US 5,874,299, US 5,877,397 and WO 1999045962 A1 (https://app.uff.br/riuff;handle/bitstream/handle/1/29025/TESE-LouiseAzulai.pdf). A nexus argument must be built carefully (the platform rest on the whole family, not this patent alone), but the technical-advance evidence is real.
4. Unexpected results. The unexpectedly high frequency of class-switch recombination (FIGS. 63, 64A/64B: two of three hybridomas showed PacI/SfiI fragments of 5.1 and 5.3 kb "consistent with recombination between the μ and γ1 switch regions"), the recovery of trans-switch products (FIG. 65), and the somatic-mutation frequencies (FIGS. 66–68) are the kinds of data that can rebut a "predictable variation" characterization.
5. Copying. Direct evidence of competitors' adoption of the claimed approach within a few years (xenogeneic-antibody programmes, transchromosomal rodents) supports nonobviousness if it can be tied to the claims.

---

8. Bottom line

Strongest prima facie § 103 combination:

Brüggemann et al., PNAS 86:6709–6713 (1989) in view of Bucchini et al., Nature 326:409 (1987) and Goodhart et al., PNAS 84:4229 (1987), further in view of Alt, TIG 1:1 (1985) / Storb, Immunol. Rev. 89:85 (1986) and US 4,736,866 (Leder).
Renders obvious: the broad genus of a transgenic mouse carrying an unrearranged, rearranging human immunoglobulin heavy-chain transgene expressing human heavy chains (Bucket A), and the derived B cells/hybridomas of Bucket D.
Motivation: the express goal, stated in Brüggemann 1989 itself, of generating a human antibody repertoire in mice as a source of human monoclonal antibodies, combined with the known success of unrearranged heterologous Ig transgenes in mice.

Second combination (isotype switching — the commercially critical feature):

Brüggemann 1989 + Durdik et al., PNAS 86:2346 (1989) + Nikaido et al., Nature 292:845 (1981) (and, if the priority date is lost for a claim, Mills 1991 / Sideras 1989) + Ott et al., EMBO J. 6:577 (1987) + Buttin, TIG 3:205 (1987).
Renders obvious if the trier accepts that Durdik taught cis-switch recombination of a transgene-encoded heavy chain, making the addition of a human Cγ/Cα/Cε downstream of Sμ a predictable use of known elements.

Third combination (endogenous-locus-disrupted "double transgenic/double knockout"):

Brüggemann 1989 + Nussenzweig, J. Exp. Med. 167:1969 (1988) + Neuberger, Nature 338:350 (1989) + homologous-recombination gene-targeting art (ES cells / positive-negative selection) + US 4,736,866.
Motivation: allelic exclusion and transgene down-regulation show the endogenous loci compete with and suppress the transgene; removing them is the obvious next step to obtain fully human antibodies.

Where I would expect the patent to survive: (i) the isotype-switching claims (Bucket B), on the strength of the specification's own assertion that no successful heterologous transgene isotype switching had been reported, coupled with Durdik's use of a mouse rather than a human switch region (this is a genuine reasonable-expectation-of-success defense under In re O'Farrell / Amgen v. Chugai), plus the objective indicia; (ii) any claims reciting specific constructs (pHC1/pHC2, pKC1/pKC2, KCo4, HCo7) or specific sequences (e.g., the vk65.3/65.5/65.8/65.15 fragments of FIGS. 41–44, the Table 2 V_H49.8 sequence) that do not appear in the pre-1990 art; and (iii) quantitative D-incorporation thresholds and transgene somatic-mutation frequencies, to the extent the pre-1990 art gave no guidance on expected values.

Where I would expect the patent to fall: the broadest genus claims to a transgenic mouse with an unrearranged human heavy-chain transgene capable of rearrangement/expression, and the product/method claims downstream of such an animal (B cells, hybridomas, ELISA screening), over Brüggemann 1989 in view of the ordinary-skills background art.

Two final, honest flags. First, my element-by-element mapping of Buckets A–F rests on the specification summary in the supplied extract, not on verbatim claim text (the extract truncates before the claims); if the actual claims are narrower than the summary suggests — e.g., if they require the endogenous loci to be disrupted, or recite specific transgene architectures — the analysis shifts toward nonobviousness for those claims. Second, the effective-date question (1990-08-29 vs. the 1991-12-17 and later priority claims) is decisive: several of the most probative references in this family's citation lists post-date 1990-08-29 and are therefore unusable in the primary case for claims that inherit that date.