US 5635517

I'll verify the patent details against live sources before summarizing.

Let me check for the PTAB proceeding and any 2026 Federal Circuit activity.

Verification note

I searched for the literal string 5,635,517 / US5635517 and did not auto-correct or substitute any similar number. Live sources (Google Patents, DrugPatentWatch, PTAB docket documents) agree with the full patent text supplied to me. What follows is grounded in that text plus the sources cited below.

Bibliographic data — US 5,635,517

Abstract (verbatim): "1-Oxo- and 1,3-dioxo-2-(2,6-dioxopiperidin-3-yl) isoindolines substituted with amino in the benzo ring reduce the levels of TNFα in a mammal. A typical embodiment is 1,3-dioxo-2-(2,6-dioxopiperidin-3-yl)-5-aminoisoindoline."

Independent-claim overview (plain language)

The patent has ten claims; claims 1–9 are method claims (claim 1 independent), and claim 10 is an independent compound claim. Claims 2–6 depend from claim 1; claims 7–9 depend from claim 1.

Claim 1 (independent — method):
Administering an effective amount of an amino-substituted 2-(2,6-dioxopiperidin-3-yl)-1-oxo- or 1,3-dioxoisoindoline to a mammal in order to lower undesirable TNFα levels, where the ring system is defined by X and Y: exactly one of X and Y is C=O and the other is either C=O or CH₂. Dependent claims narrow this to the specific amino positional isomers: the 5-, 4-, 6-, and 7-aminoisoindolin-1-one (claims 3–6, where X = C=O and Y = CH₂) and the 4- and 5-aminoisoindoline-1,3-diones (claims 8–9, where both X and Y are C=O).

Claim 10 (independent — compound):
A compound selected from a four-member group: the 1-oxo-2-(2,6-dioxopiperidin-3-yl) aminoisoindolines substituted at the 5-, 4-, 6-, or 7-position. Notably, this compound claim is the one Celgene's prosecution history identifies as covering four compounds first synthesized by the inventors; it includes 1-oxo-2-(2,6-dioxopiperidin-3-yl)-4-aminoisoindoline, which is the compound subsequently developed and marketed as lenalidomide (REVLIMID®) and is also recited in dependent claim 4.

The specification also describes preparation of the claimed compounds by catalytic hydrogenation of the corresponding nitro intermediates (e.g., Example 1 gives 1,3-dioxo-2-(2,6-dioxopiperidin-3-yl)-5-aminoisoindoline from its 5-nitro precursor, mp 318.5–320.5 °C), plus examples of tablets, chewing tablets, gelatin capsules, and injectable solutions.

Post-grant and litigation posture (relevant, but not all "CAFC 2026")

• Ex parte reexamination (1998–1999): No amendments made; the reexamination certificate states "The patentability of claims 1–10 is confirmed." (Reexam certificate, B1.)
• Patent term extension: Celgene filed a §156 application on the '517 patent tied to REVLIMID® (lenalidomide; 5 mg and 10 mg capsules, originally approved for transfusion-dependent anemia in Low-/Intermediate-1-risk MDS with del(5q)). The extension yields the listed adjusted expiration of 2019-10-04.
• PTAB — IPR2015-01169: Coalition for Affordable Drugs VI LLC v. Celgene Corporation, challenging claims 1–10 under §103. The Board denied institution in its entirety (decision dated Dec. 22, 2015), finding no reasonable likelihood of prevailing on any challenged claim; a related Celgene motion for sanctions was also denied (Sept. 25, 2015).
• District court: The '517 patent was at issue in Celgene Corporation v. Natco Pharma Ltd., C.A. No. 2:10-cv-5197 (D.N.J.).
• 2026 materials: A 2026 antitrust complaint filed in New Jersey (Bristol Myers Squibb / Celgene) references the '517 patent and its 2015 reexamination/IPR history; that is a complaint that discusses the patent, not a validity proceeding concerning it.

Uncertainty flags (stated explicitly rather than guessed):

1. I did not identify any Federal Circuit docket in 2026 captioned on U.S. 5,635,517. The 2026 CAFC/PTAB statistics and appeal documents returned by search are unrelated cases (e.g., Contour IP v. GoPro, Headwater v. Verizon, TrackTime). Any 2026 appeal involving this patent would post-date the pre-issued patent's 2019 expiration, which makes an active 2026 validity appeal unlikely but not impossible to confirm from the sources I retrieved.
2. I could not independently verify the exact adjusted expiration calculation beyond Google Patents' listing of 2019-10-04 and the existence of the §156 application; the PTO extension certificate itself was not in the retrieved text.
3. The issue date vs. reexamination certificate date distinction matters: "US 5,635,517 A" published 1997-06-03; "US 5,635,517 B1" (post-reexam) 1999-06-29. Both use the same patent number, as required by the literal-number rule.

Sources: https://patents.google.com/patent/US5635517/en ; https://www.drugpatentwatch.com/p/patent/5635517 ; https://www.docketalarm.com/cases/PTAB/IPR2015-01169/ ; https://www.jonesday.com/en/practices/experience/2015/11/celgene-successfully-defends-revlimid-compound-patent-against-kyle-bass-ipr-petition

I'll research litigation involving US patent 5635517 (the Celgene amino-substituted isoindoline/TNFα patent) across patent litigation databases.

Good results. Let me dig deeper into the specific district court cases and PTAB proceedings.

Let me gather more specifics on additional D.N.J. cases and outcomes.

Litigation involving U.S. Patent No. 5,635,517

Bottom line: U.S. 5,635,517 ("'517 patent") has been extensively litigated — but almost always as one of many patents-in-suit in Celgene's Hatch‑Waxman (ANDA) campaign over REVLIMID® (lenalidomide), not as the sole or lead patent. I found no standalone, single-patent assertion of the '517 patent. The '517 patent also faced one PTAB challenge. Below is what I can ground in search results, with confidence flags where the record is incomplete.

1. The patent itself (context for the litigation)

• Title: "Method of Reducing TNFα Levels with Amino Substituted 2-(2,6-dioxopiperidin-3-yl)-1-oxo-and 1,3-dioxoisoindolines"; inventors Muller, Stirling, Chen; assignee Celgene Corporation.
• Issued: June 3, 1997; reexamination certificate June 29, 1999; term extended 1,167 days under 35 U.S.C. § 156 on March 27, 2008; listed expiration Oct 4, 2019 (adjusted).
• Orange Book–listed for REVLIMID® (NDA 21‑880) and Pomalyst® (NDA 204‑026).
• Source: Celgene's § 156 extension request (Exhibit 1011, IPR2015‑01169) and the Natco complaint (the '517 counts among the "Patents‑in‑Suit").

2. District court litigation (all in D.N.J. — the Revlimid ANDA cases)

The '517 patent appears in the Patents‑in‑Suit list in Celgene's lenalidomide ANDA suits. The following cases are confirmed in the search results:

Confidence note: Cases 1 and 4 are well grounded in document text (the Natco complaint and the Lotus consent judgment both list the '517 patent by number). Cases 2 and 3 are confirmed only as docket entries retrieved from Docket Alarm's patent:(5635517) search — i.e., the '517 patent number was associated with those dockets — but I did not retrieve the complaints or judgments, so I cannot confirm the '517 patent was actually asserted or how each ended. Treat their detailed status as unverified.

Additional/earlier and later cases likely exist. The Docket Alarm search for patent:(5635517) returned "151 results," far more than the four shown, indicating many more ANDA/related dockets reference the '517 patent. I could not enumerate them within this search session. Celgene also asserted a 16‑patent set (including 5635517) against Lotus/Alvogen, per the Taiwanese IPO source.

3. PTAB post-grant proceeding

Sources: Docket Alarm PTAB docket; GreyB/ipverse case page. This was the direct validity challenge to the '517 patent at the PTAB. The petition was premised on thalidomide prior art (e.g., Piper, Agrawal, Sampaio, D'Amato/WO 94/20085, D'Amato Nature 1987) and the '517 prosecution/reexamination file history.

4. Related proceedings where '517 is involved but NOT the challenged/asserted patent (flagged for precision)

• [Dr. Reddy's Laboratories, Inc.](/litigations/by-plaintiff/Dr.%20Reddy's%20Laboratories%2C%20Inc.) et al. v. Celgene Corp., IPR2018-01507 — the '517 patent appears only as Exhibit 2026 (prior-art/background evidence); it was not the challenged patent. Do not count this as a '517 challenge.
• In re Revlimid and Thalomid Purchaser Antitrust Litigation (D.N.J.) — antitrust claims in which the '517 patent's prosecution, reexamination (claims initially rejected, then confirmed via Celgene's declaration), and Orange Book listings are central background. The '517 patent is referenced as the "'517 Compound Patent."
• Hagens Berman / direct-purchaser antitrust complaints (2023, 2024 corrected amended complaint; and a 2026 BMS/Celgene antitrust complaint in New Jersey) — background allegations regarding the '517 reexamination and Celgene's patent-enforcement conduct.
• EPO proceedings — European counterparts (e.g., EP '973, EP 2 177 517 B1, EP 2 070 920 B1) were litigated/revoked, but these are foreign counterparts, not U.S. 5,635,517.

Caveats and how to verify

1. Verification method the user requested: The authoritative sources are the D.N.J. PACER dockets (2:10‑cv‑05197; 2:17‑cv‑02528; 2:17‑cv‑03159; 2:17‑cv‑06842; 2:18‑cv‑11518) and the PTAB docket for IPR2015‑01169. I was unable to open the Unified Patents litigation portal directly in this session; the Docket Alarm patent:(5635517) search (151 hits) is the best proxy I retrieved, plus CourtListener RECAP documents.
2. I did not confirm every case or the final disposition of cases 2 and 3. Stating otherwise would be fabrication.
3. Interpretation of identifiers: All numbers are reported literally as found (e.g., "5,635,517," "2:10-cv-05197"). Note the Docket Alarm/Google entry also shows a "US5635517B1" second publication (June 29, 1999) tied to the reexamination certificate — the same patent, not a different one.
4. The '517 patent expired Oct 4, 2019, so the litigation is historical; suits resolved via settlements and consent judgments, not final invalidity/infringement merits judgments on the '517 patent itself (the institution decision in IPR2015‑01169 was a denial of institution, not a merits win for either side).

If you want, I can drill into any single docket (e.g., pull the Zydus or Par complaints via CourtListener/PACER) to confirm whether the '517 patent was actually asserted there and how each terminated.

I'll verify the structured data against live sources before drafting.

Web search surfaced a proceeding the ODP block doesn't reflect. Digging in.

PTAB / AIA Trial Activity — US 5,635,517

Data-integrity flag, read first. The structured "PTAB proceedings on file" block supplied for this analysis states that the USPTO ODP API returns no AIA trial proceedings for US 5,635,517. That is incomplete. The public record — including a Board decision hosted on the PTAB's own docket and mirrored into a district-court record — shows at least one AIA petition filed against this patent, IPR2015-01169, which was denied institution on 2015-11-16. The ODP block appears to have missed the pre-2016 / Coalition for Affordable Drugs filings. Everything below is grounded in the decision documents I could retrieve; where I could not verify something (e.g., the petition's ground-by-ground art mapping), I say so rather than guess.

Proceedings overview

One (1) AIA trial proceeding is on file against US 5,635,517: IPR2015-01169, status institution denied (2015-11-16). Zero claims were canceled; zero claims were confirmed patentable in a Final Written Decision; there is no settlement and no appeal (institution denials are unreviewable under § 314(d)).

Bottom-line defensive posture for a defendant: the patent has never been through a PTAB trial, so it is neither "hardened" by a win nor narrowed by a loss — it is simply untested at the Board. Critically, the one petition that was filed never reached a Final Written Decision, which means no § 315(e)(2) estoppel attached to anyone. But the more important fact for a real defendant is timing: the '517 patent expired 2019-10-04 (Orange Book / Google Patents legal status: "Expired - Lifetime"), so the live question is not how to beat it at the PTAB — it is what past damages, if any, remain reachable inside the 35 U.S.C. § 286 six-year window.

IPR2015-01169 — Coalition for Affordable Drugs VI LLC v. Celgene Corporation

• Type: Inter Partes Review (post-AIA, § 311)
• Patent at issue: U.S. Patent No. 5,635,517 (Ex. 1001)
• Filed: 2015-05-07 (Petition, Paper 1)
• Status: Decision Denying Institution of Inter Partes Review, 37 C.F.R. § 42.108 — institution denied (2015-11-16, Paper 22). Plain English: the Board read the petition and the Patent Owner Preliminary Response (Paper 16) and refused to open a trial. No FWD, no claim-level validity ruling, no estoppel.
• Judge panel: Administrative Patent Judges Toni R. Scheiner, Jacqueline Wright Bonilla, and Tina E. Hulse; the opinion was authored by APJ Bonilla.
• Petition grounds: The petition challenged claims 1–10 — i.e., all ten claims of the patent, including the method claims (1–9) and the compound claim (10). The Board's decision frames the analysis under the § 314(a) "reasonable likelihood that the petitioner would prevail with respect to at least 1 of the claims challenged" standard. I could not verify, from the documents I retrieved, the petition's ground-by-ground statutory mapping (§ 102 vs. § 103) or the specific reference combinations — the denial decision itself is short and rests on the overall failure of the showing. I am explicitly not asserting a reference list. (The exhibit record shows thalidomide-analog / anti-inflammatory art on the petitioner's side — e.g., Piper et al., Anti-inflammatory immunosuppressive thalidomide analogs; Hassan; King — and thalidomide-mode-of-action and teratology art on the Patent Owner's side, including Powell. Those exhibit titles alone do not establish the grounds.)
• Institution decision: Denied, 2015-11-16. The panel's reasoning, quoted verbatim: "we determine that Petitioner has not established a reasonable likelihood that it would prevail in showing the unpatentability of any claim challenged in the Petition. Accordingly, we decline to institute an inter partes review." Note this was a merits-based denial under § 314(a), not a discretionary efficiency denial, and the panel expressly noted the co-pending Hatch-Waxman case, Celgene Corp. v. Natco Pharma Ltd., C.A. No. 2:10-cv-05197 (D.N.J.) (consolidated with 2:12-cv-04571), as background.
• Final Written Decision: None. The case terminated at the institution stage, so no claim of the '517 patent was ever adjudicated on the merits by the Board.
• Settlement / termination: Not applicable — the proceeding ended by Board denial, not party settlement. (Contrast the district court Revlimid cases, which did settle: the Natco action terminated by settlement agreement and consent judgment entered 2016-01-04, and the Lotus/Alvogen action settled 2019-03-30 on terms that were publicly described as volume-limited entry beginning March 2022 through 2026-01-31. Those are litigation settlements, not PTAB terminations.)
• Appeal: None available and none taken. A denial of institution is not appealable. See 35 U.S.C. § 314(d); Cuozzo Speed Techs., LLC v. Lee, 579 U.S. 261 (2016).
• Defensive value: Low as precedent, but high as an estoppel fact. The denial is non-precedential and claim-nonspecific — it does not validate claims 1–10 and it binds no one. Because there was no FWD, no petitioner is estopped under § 315(e)(2), and any defendant today may raise the same art CFAD raised, plus everything else, free of estoppel. Do not treat the Jones Day press summary — "Celgene successfully defends REVLIMID® compound patent" — as a merits win on validity; it is a description of a declined petition, and the panel made no patentability determination.

Related AIA proceedings that touch the '517 patent but are not challenges to it

These are frequently confused with '517 IPRs in dockets and diligence memos. None of them can cancel a claim of the '517 patent, because the '517 patent is not the challenged patent in any of them:

Two corroborating citations: (i) IPR2018-01507's Patent Owner Mandatory Notices identify the challenged patent as the '717 patent and list the related DRL petitions on the '120 and '740 patents (https://www.docketalarm.com/cases/PTAB/IPR2018-01507/Inter_Partes_Review_of_U.S._Pat._8404717/08-22-2018-Patent_Owner/Mandatory_Notice-4-Patent_Owners_Mandatory_Notices/); (ii) the Alvogen petition against the '569 patent affirmatively states that "[t]he Orange Book lists 27 patents as covering Revlimid, including US 5,635,517, which issued June 3, 1997 and covers the lenalidomide compound… The '517 patent does not expire until October 4, 2019" — the '517 patent used as an obviousness/commercial-success argument, not as a target.

Strategic summary

Claim-by-claim status of US 5,635,517 as of today:

• Canceled: none. Zero claims have ever been canceled by the Board.
• Sustained by a Final Written Decision: none. No claim has ever been adjudicated or confirmed by the PTAB.
• Untested: all of claims 1–10. That includes the two structurally distinct groups — claims 1–9 (method of reducing TNFα, with 2–6 covering the 1-oxo species and 7–9 the 1,3-dioxo species) and claim 10, the compound claim covering 1-oxo-2-(2,6-dioxopiperidin-3-yl)-5-, 4-, 6- and 7-aminoisoindoline (i.e., lenalidomide and its positional isomers). Claim 10 is the commercially important claim and it has never been challenged in an instituted AIA trial.

Estoppel landscape. This is the cleanest part of the analysis and it runs in the defendant's favor. Section 315(e)(2) estoppel is triggered only by a final written decision. There has never been one for this patent. IPR2015-01169 died at institution; the near-miss petitions (IPR2018-01504/-01507/-01509, IPR2018-01714) are estoppel-relevant only to their own patents — the '120, '717, '740 and '569 patents respectively. Therefore no petitioner, real party in interest, or privy is barred from asserting any ground against the '517 patent, and a defendant now facing assertion can deploy the full prior-art universe — including the thalidomide-analog art listed in the CFAD exhibit record — without an estoppel trap. The only meaningful "cost" the denial imposed was on CFAD, and CFAD is a defunct Kyle Bass vehicle, not a defensive aggregator.

Pattern signals. (1) The same petitioner filed multiple petitions against the same patent owner: Coalition for Affordable Drugs VI filed the '517 petition alongside IPRs on the '501 and '720 patents, and the Board instituted on those while denying only the '517 petition — the pattern is a hedge-fund short-and-challenge campaign, not a competitor's defense. (2) No defensive aggregator (Unified Patents, RPX, etc.) appears anywhere in the '517 chain. (3) The patent owner did not pursue a PTAB appeal on this patent (nothing to appeal); Celgene's aggressive appellate activity at the Federal Circuit concerns other patents in the family — the '501 and '720 appeals were consolidated and decided against Celgene in Celgene Corp. v. Peter, 931 F.3d 1342 (Fed. Cir. 2019) (I flag that the citation is reconstructed from a garbled source and should be re-verified before filing). (4) The '517 patent was a workhorse asserted patent in the Hatch-Waxman wars (Celgene v. Natco 2:10-cv-05197 / 2:12-cv-04571; Celgene v. Dr. Reddy's 2:16-cv-07704; Celgene v. Zydus 2:17-cv-02528; Celgene v. Par 2:17-cv-03159; and the Lotus/Alvogen case) — and every one of those settled rather than being tried to judgment on the '517 patent.

Recommended next steps

1. Lead with expiry, not validity. The patent expired 2019-10-04. The twenty-year date from the 1996-07-24 filing is 2016-07-24; the difference reflects a § 156 patent term extension obtained for REVLIMID regulatory review (the file-history exhibit "Application for Extension of Patent Term under 35 U.S.C. § 156" confirms the request; the Orange Book / Google Patents entries confirm the 2019-10-04 expiration). Practical consequences: no injunctive relief is available, no ongoing royalty, and damages are confined to past infringement within 35 U.S.C. § 286's six-year lookback — i.e., conduct on or before 2019-10-04. For a generic contemplating or already marketing lenalidomide today, the '517 patent is a closed damages question, not a bar to entry.

2. If a demand letter cites claims 1–9, press the method-claim problem. Claims 1–9 are method-of-reducing-TNFα claims. They require an administering step and are susceptible to inducement-only theories under § 271(b), which are materially harder to prove against a generic label. Combined with the expired status, a demand premised on claims 1–9 should be met with a straight § 286 / expiry analysis.

3. Treat claim 10 as the only claim worth a real invalidity workup — but only if you need a declaratory judgment. Claim 10 is the compound claim. Because no estoppel applies, an invalidity theory on claim 10 is unmortgaged by any prior proceeding. Note, however, that an IPR petition is likely moot now: the patent has expired, and the Board generally declines to institute where no live controversy supports amendability of the claims. For a past-damages defense, the vehicle is a district-court invalidity defense, not the PTAB.

4. Do not rely on the "PTAB ruled in Celgene's favor" narrative. A party brief in the Revlimid antitrust litigation (D.N.J. 2:19-cv-07532) asserts that "[l]ike the '517 compound patent, these patents were subject to repeated IPR attacks, yet in all five proceedings the PTAB ruled in Celgene's favor." That is advocacy, and it does not hold for the '517 patent: the single '517 petition was denied institution, which is not a ruling of patentability, and in the companion CFAD proceedings the Board actually instituted and invalidated claims of the '501 and '720 patents. If opposing counsel quotes that line, the correct response is to separate institution denials from FWDs.

5. Pull the primary documents before you file anything. The controlling document is one paper:

• Decision Denying Institution, Coalition for Affordable Drugs VI LLC v. Celgene Corp., IPR2015-01169, Paper 22 (P.T.A.B. Nov. 16, 2015) — https://www.docketalarm.com/cases/PTAB/IPR2015-01169/Inter_Partes_Review_of_U.S._Pat._5635517/docs/11-16-2015-Board/Decision_Denying_Institution-22-Decision___Denying_Institution_of_Inter_Partes_Review___37_CFR_42108.pdf (also filed of record at https://storage.courtlistener.com/recap/gov.uscourts.njd.[399896](/patent/399896)/gov.uscourts.njd.399896.18.4.pdf).
• For anything filed after 2015, search PTAB E2E / PTAB Center directly by patent number rather than relying on the ODP feed that produced the incomplete block in this prompt.

6. Caveat on completeness. My search identified one AIA petition against the '517 patent. I cannot affirmatively exclude a second, non-indexed or recently-filed petition, but the combination of (a) the patent's 2019-10-04 expiry and (b) the absence of any PTAB docket activity in the sources I retrieved makes a material undiscovered proceeding unlikely. If precision matters for a filing, confirm against the Board's Patent Trial and Appeal Case Tracking System by patent number before you represent "no other proceedings exist" to a court.

Net answer to the client question: There is exactly one AIA proceeding on this patent, it was denied institution, nothing was canceled, nothing was affirmed, no estoppel attached, and the patent is expired. The defensive value here is expiry plus the absence of estoppel — not a PTAB win.

I'll research the assignment record for US 5635517 before writing up the analysis.

Good — I found a key reel reference. Let me dig into the assignment records and litigation history.

Let me verify whether any post-issuance assignments exist and check the Celgene/BMS acquisition angle.

Assignment & Ownership Analysis — US 5,635,517

"Method of reducing TNFα levels with amino substituted 2-(2,6-dioxopiperidin-3-yl)-1-oxo- and 1,3-dioxoisoindolines"
Application 08/690,258 · Filed 1996-07-24 · Issued 1997-06-03 (US5635517A; B1 certificate 1999-06-29) · PTE expiry 2019-10-04

Inventors

All three executed an assignment of all right, title and interest to Celgene Corporation, which is the standard indicium that they were Celgene employees under a proprietary-information-and-assignment obligation at filing (the patent's own specification describes the work as internal Celgene chemistry; the assays and compounds are Celgene's).

Unusual-pattern check: not present. There is no evidence of inventors departing the original assignee within 12 months. This is the inverse of the "inventor-flight → portfolio fire-sale" tell: the inventors were the source of the asset, not sellers of it, and they appear to have remained Celgene-affiliated for years (Stirling rose to CSO; Muller is a career Celgene scientist). I have not independently verified exact departure dates for Chen or Muller, so treat the "remained" statement as a reasonable inference rather than a documented fact.

Original assignee

Celgene Corporation (then of Warren/Summit, New Jersey; incorporated in Delaware).

• Primary line of business: specialty biopharmaceutical — the leading immunomodulatory-drug (IMiD) franchise. Celgene was a NASDAQ-listed operating company (ticker CELG).
• Product embodying the claims: yes, and it was a blockbuster. Claim 10 (compound claim) recites 1-oxo-2-(2,6-dioxopiperidin-3-yl)-4-aminoisoindoline, which is lenalidomide — Celgene's REVLIMID® (NDA 021880, first approved 2005-12-27). Claim 4 is the corresponding method claim. Claim 8 recites 1,3-dioxo-2-(2,6-dioxopiperidin-3-yl)-4-aminoisoindoline, which is pomalidomide — later marketed by Celgene as POMALYST® (I have not separately verified POMALYST's Orange Book listing for the '517 patent; the listing I have verified is REVLIMID).
• The '517 patent was the foundational compound patent of the "517 family" — the family from which Celgene's Revlimid patents (US 6,281,230; 6,555,554; 7,119,106; 8,288,415) descend, all sharing the 1996-07-24 priority date. Sources: FDA Orange Book listing (5635517, expiry 2019/10/04, compound patent "Y"); Cigna Group v. Celgene complaint (D.N.J.) ¶¶2–5 describing the "'517 family."
• Term extension: Celgene obtained a 35 U.S.C. § 156 patent term extension based on FDA regulatory review of REVLIMID, moving expiry from 2016-07-24 → 2019-10-04 (PTE application, Ex. 1011 in IPR2015-01169).
• Current status: operating, but no longer independent. Celgene was acquired by Bristol-Myers Squibb in a ~$74 billion cash/stock transaction that closed 2019-11-20; Celgene became a wholly owned BMS subsidiary. Note the timing: the '517 patent expired 2019-10-04, roughly six weeks before the merger closed — so the asset was already off the books by closing.

Assignment timeline

One recorded assignment only. USPTO records show a single assignment for the '517 patent — the original inventor-to-corporation assignment. No security agreements, no releases, no change-of-name records, and no post-issuance transfers were identified.

• 1996-07-24 (executed; date per Google Patents legal events) / recorded 1996 — Reel 08147/0954
  • Conveyance: Assignment (ASSIGNMENT OF ASSIGNORS INTEREST)
  • Assignors: George W. Muller; David I. Stirling; Roger Shen-Chu Chen
  • Assignee: CELGENE CORPORATION
  • Correspondent: Not retrievable. I could not pull the correspondent-of-record field for this reel/frame from the Assignment Center index, and no secondary source reproduces it. Flagging as a data gap rather than a finding. The only name-adjacent datum I recovered is a fax-header string, "SHIRSSENI," on the 2006 PTE filing (Exhibit 1011, IPR2015-01169) — that is a prosecution/PTE-side cover-page entry, not an assignment correspondent, and I decline to characterize it as one.
  • Context: Initial inventor assignment — internal, pre-issuance, standard practice. This is the conveyance Celgene relied on in 2006 when it certified to the USPTO that it "is the owner of record… by virtue of assignment" and attached "assignment recordation at the USPTO (Reel 8147, frame 0954)" as Exhibit A to its § 156 PTE application.

Nothing after 1996. The Celgene → Bristol-Myers Squibb combination (2019-11-20) was a stock merger in which Celgene survived as a wholly owned subsidiary, not an asset assignment — so no reel/frame would be expected for the '517 patent even had it not expired first. That absence is itself a finding: Celgene Corporation (now a BMS subsidiary) remained the owner of record from 1996 until expiry.

Verification: Assignment Center — https://assignmentcenter.uspto.gov/ (search US 5,635,517) · Google Patents legal events — https://patents.google.com/patent/US5635517/en

Timeline diagram

timeline
    title Ownership of US 5635517
    1996 : Inventors assign to Celgene Corporation
         : Application 08690258 filed
    1997 : Patent US 5635517 issued
    2006 : Revlimid approved by FDA
         : Patent term extension requested
    2010 : Celgene sues Natco for infringement
    2015 : Bass IPR petition denied by PTAB
    2019 : Patent term extension expires
         : Celgene acquired by Bristol-Myers Squibb

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. The chain contains one assignee: Celgene Corporation, an operating company, per Reel 08147/0954. No "IP / Holdings / Licensing / Ventures"-suffixed entity, no single-purpose LLC, no registered-agent service address appears anywhere in the record.

2. Known asserter in the chain — NOT PRESENT. Neither current nor prior assignee matches any public NPE list (Acacia, Marathon, IV, IPNav, Wi-LAN/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, DGC, Spangenberg entities). Note a common misread: Coalition for Affordable Drugs VI LLC (hedge-fund manager Kyle Bass / Hayman Capital) appears in this patent's history as the IPR petitioner in Coalition for Affordable Drugs VI LLC v. Celgene Corp., IPR2015-01169 — that entity was attacking the patent, never owned it. Institution was denied in full on 2015-11-16. Do not mistake the petitioner for an owner.

3. Repeat correspondent across the chain — NOT PRESENT / NOT ASSESSABLE. With only one link in the chain there is no recurrence to detect, and I could not retrieve the correspondent field for Reel 08147/0954. A single correspondent entry would not be a finding even if recovered — the signal is recurrence, and recurrence is structurally impossible in a one-transfer chain. Marked as a genuine data limitation, not evidence of cleanliness.

4. Cascading transfers — NOT PRESENT. One transfer in 1996; zero subsequent transfers across 23 years of patent life. There is no <24-month LLC chain, no shared correspondent address across assignees, no common principals.

5. Pre-litigation transfer — NOT PRESENT. The assignment was executed 1996-07-24; the first identified assertion naming the '517 patent is Celgene Corp. v. Natco Pharma Ltd., D.N.J. Civ. No. 2:10-05197, filed 2010-10-08 (following Natco's 2010-08-27 paragraph IV notice). That is a ~14-year gap between recordation and assertion — the opposite of venue/standing engineering. Later suits naming the '517 patent include Celgene v. Zydus, D.N.J. 2:17-cv-02528 (2017-04-12) and Celgene v. Par Pharmaceutical, D.N.J. 2:17-cv-03159 (2017-05-04), plus actions against Dr. Reddy's, Alvogen/Lotus and Sun Pharma.

6. Bankruptcy fire-sale — NOT PRESENT. Celgene never filed Chapter 7 or 11. Its exit was a $74 billion negotiated merger with BMS (announced 2019-01-03, closed 2019-11-20), with a companion FTC-mandated divestiture of OTEZLA® to Amgen for $13.4 billion — a regulatory remedy, not a distressed sale. The '517 patent had already expired (2019-10-04) before closing.

7. Privateering — NOT PRESENT. Celgene/BMS asserted in its own name, against actual generic ANDA filers (Natco, Watson, Arrow, Dr. Reddy's, Alvogen/Lotus, Sun, Par, Zydus) in Hatch-Waxman litigation over its own branded product. There is no operating-company-to-NPE transfer, no assertion on Celgene's behalf by a third party, and no SEC disclosure of such an arrangement. (The 2010s litigation around these patents did attract antitrust scrutiny — see the Cigna Group v. Celgene complaint's reverse-payment and "shielding from invalidation" allegations — but that is a branded-pharma settlement-conduct theory, not an NPE/privateering theory. Different animal; do not conflate.)

8. Defensive aggregator (anti-NPE) — NOT PRESENT. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. Defensive aggregators appeared in this story only as databases (the Unified Patents portal indexes the '517 family as Celgene-assigned), not as owners.

Verdict

Operating-company assertion.

The entire ownership record consists of a single 1996-07-24 inventor-to-corporation assignment, Reel 08147/0954, to Celgene Corporation, and nothing else — no LLC, no asserter, no security interest, no post-issuance transfer across 23 years. Celgene (and, after the 2019-11-20 merger, its parent Bristol-Myers Squibb) shipped the claimed subject matter as a commercial product — claim 10's species is lenalidomide, marketed as REVLIMID®, an FDA-approved blockbuster — and enforced the '517 patent itself, in its own name, against genuine generic competitors beginning with Celgene v. Natco on 2010-10-08 after a ~14-year pre-assertion gap. Every NPE signal is either flatly absent or structurally inapplicable to a one-transfer chain; the sole NPE-adjacent entity in the record, Coalition for Affordable Drugs VI LLC, was the IPR petitioner attacking the patent (IPR2015-01169, institution denied 2015-11-16), not its owner.

Caveat on completeness: the conclusions rest on the USPTO Google Patents legal-events record plus the PTE exhibit citing Reel 8147/0954. I was unable to query the Assignment Center directly during this session to pull the raw reel/frame abstract, the exact recordation date, or the correspondent of record. If you need the correspondent field for a definitive repeat-player check, that must be pulled live at https://assignmentcenter.uspto.gov/ — but given that the chain has exactly one link and one corporate assignee, a correspondent finding could not, on its own, move the verdict off operating-company assertion.

I'll research the specific prior art citations for US5635517. Let me run several targeted searches.

Let me retrieve the specific "References Cited" (prior art citations) listed on the face of US5635517.

Prior-Art Analysis — US Patent 5,635,517

0. Identification of the patent (confirmed by search)

Before reaching the references, I should state plainly what the search confirmed, because the citation set only makes sense against the correct document:

This corresponds to lenalidomide (3-(4-amino-1-oxo-1,3-dihydroisoindol-2-yl)piperidine-2,6-dione), i.e. the Revlimid compound patent. The searched record also shows it was the subject of an IPR petition filed May 2015 by Coalition for Affordable Drugs VI LLC (institution denied 2015-11-16).

1. An important direction-of-citation caveat (methodology note)

The request said "look at each patent citation for 5635517." There are two opposite citation streams, and only one of them is prior art:

• "References Cited" / "Patent Citations" (cited in 5635517) — documents the examiner and applicant listed on the face of the patent as background art, and the references cited in the specification. These are what can be relied on under 35 U.S.C. § 102/§ 103. These are the prior art you asked about.
• "Cited By" (documents citing 5635517) — the 267 forward citations (e.g., US 5,712,291 D'Amato; WO 98/03502; US 5,874,448; US 6,281,230; US 6,395,754; etc.). These were published after the 1996-07-24 priority date and legally cannot be prior art to US 5,635,517. They are relevant only to later, related Celgene filings.

⚠️ Honest limitation: The authoritative full text supplied to me for US 5,635,517 was truncated before the face-page "References Cited" table (it jumps from the claims/family data straight into the "Cited By (267)" list). My web searches did not surface the verbatim front-page U.S./foreign patent-document list from the examiner's "References Cited" block. I therefore cannot reproduce that exact enumerated list, and I will not fabricate it. What follows is grounded in (a) the references that are expressly relied upon inside the patent's own specification — which I can cite verbatim from the supplied text — and (b) confirmed bibliographic facts from the searches. Where I am relying on general knowledge rather than the retrieved documents, I flag it.

2. Prior art expressly cited or acknowledged inside US 5,635,517

2.1 The reference that matters most — Jonsson (1972)

• Full citation: Jonsson, N. Å., Acta Pharmaceutica Suecica, Vol. 9, pp. 521–542 (1972).
• Date: 1972 (more than 24 years before the 1996 filing → prior art under § 102(b)).
• What it discloses — per the patent's own words: "Certain of these compounds, such as 1,3-dioxo-2-(2,6-dioxopiperidin-3-yl)-4-aminoisoindoline and 1,3-dioxo-2-(2,6-dioxopiperidin-3-yl)-5-aminoisoindoline are known. See, e.g., Jonsson, Acta Pharma. Succica, 9, 521-542 (1972)."
• Translation to modern naming: Jonsson discloses 4-aminothalidomide (pomalidomide) and 5-aminothalidomide — i.e., the two 1,3-dioxo amino species.
• Claims it potentially bears on:
  • Claims 7 and 9 (1,3-dioxo genus / 1,3-dioxo-5-amino species) and claim 8 (1,3-dioxo-4-amino species) — Jonsson discloses the compounds recited in these dependent claims.
  • Claim 1 (the genus "one of X and Y is C═O and the other is C═O or CH₂") — Jonsson discloses species falling inside this genus.
• Critical § 102 caveat: Claims 1–9 are method claims ("The method of reducing undesirable levels of TNFα …"). Anticipation of a method claim requires the reference to disclose the method steps, not merely the compound. Jonsson discloses the compounds but (on the record here) not their use to reduce TNFα. So Jonsson is, strictly, § 103 prior art (compound known + the extensive TNFα literature of record making the use obvious) rather than a clean § 102 anticipation of the method claims. If Jonsson did also report a biological/pharmacological activity, its § 102 weight would rise — I could not verify from the retrieved material whether it does.

2.2 The other patent document cited in the description

• WO 92/11383 — cited in the specification as: "Prior efforts directed to the suppression of the effects of TNFα have ranged from the utilization of steroids such as dexamethasone and prednisolone to the use of both polyclonal and monoclonal antibodies {Beutler et al., Science 234, 470-474 (1985); WO 92/11383}."
• Date: published 1992 → § 102(b) art.
• Description: PCT publication relating to anti-TNFα antibody approaches.
• Claim relevance: Background only. It supports the § 103 motivation ("reducing TNFα is a recognized therapeutic goal") but does not disclose the isoindoline/glutarimide compound class and therefore cannot anticipate any of claims 1–10.

2.3 Non-patent background literature cited in the specification

These are all cited in the description and each predates the filing (so each is § 102(b)/§ 103 background art). None of them discloses the claimed compounds; they establish the known therapeutic rationale for lowering TNFα and cAMP. Grouped by topic:

TNFα in disease (background / § 103 motivation):

• Tracey et al., Nature 330, 662-664 (1987) — endotoxemia/toxic shock.
• Hinshaw et al., Circ. Shock 30, 279-292 (1990).
• Dezube et al., Lancet 335(8690), 662 (1990) — cachexia.
• Millar et al., Lancet 2(8665), 712-714 (1989) — ARDS.
• Ferrai-Baliviera et al., Arch. Surg. 124(12), 1400-1405 (1989).
• Bertolini et al., Nature 319, 516-518 (1986) and Johnson et al., Endocrinology 124(3), 1424-1427 (1989) — bone resorption/arthritis.
• Holler et al., Blood 75(4), 1011-1016 (1990) — graft-versus-host.
• Grau et al., N. Engl. J. Med. 320(24), 1586-1591 (1989) — cerebral malaria.
• Pignet et al., Nature 344, 245-247 (1990) and Bissonnette et al., Inflammation 13(3), 329-339 (1989) — silicosis/asbestos fibrosis.
• Baughman et al., J. Lab. Clin. Med. 115(1), 36-42 (1990) — sarcoidosis.
• Vedder et al., PNAS 87, 2643-2646 (1990) — reperfusion injury.
• Sherry et al., J. Cell Biol. 107, 1269-1277 (1988); Munro et al., Am. J. Path. 135(1), 121-132 (1989) — endothelial ICAM/ELAM.

TNFα/HIV and NF-κB biology:

• Duh et al., PNAS 86, 5974-5978 (1989); Poll et al., PNAS 87, 782-785 (1990); Monto et al., Blood 79, 2670 (1990); Clouse et al., J. Immunol. 142, 431-438 (1989); Poll et al., AIDS Res. Hum. Retrovirus, 191-197 (1992); Rosenberg et al., Advances in Immunology 57 (1989); Folks et al., PNAS 86, 2365-2368 (1989); Osborn et al., PNAS 86, 2336-2340 (1989); Wright et al., J. Immunol. 141(1), 99-104 (1988).
• NF-κB: Lenardo et al., Cell 58, 227-29 (1989); Dbaibo et al., J. Biol. Chem. (1993) 17762-66; Bachelerie et al., Nature 350, 709-12 (1991); Boswas et al., J. Acq. Immune Defic. Syndr. 6, 778-786 (1993); Suzuki et al., Biochem. Biophys. Res. Commun. 193, 277-83 (1993), 189, 1709-15 (1992), and Biochem. Mol. Biol. Int. 31(4), 693-700 (1993); Shakhov et al., PNAS USA (1990) 171, 35-47; Staal et al., PNAS USA 87, 9943-47 (1990).

cAMP rationale:

• Lowe and Cheng, Drugs of the Future 17(9), 799-807 (1992).

Anti-TNFα prior art (steroids/antibodies):

• Beutler et al., Science 234, 470-474 (1985).

Synthetic-method reference (enablement, not anticipation):

• Robin et al., Tetrahedron: Asymmetry 6, 1249 (1995) — stereopreferential synthesis route cited for making the (R)/(S) enantiomers.

Caveat: several of these items (the Suzuki, Boswas, Dbaibo and Shakhov citations) are dated 1992–1993 and are cited by the patent drafters with imprecise/incomplete page data; I am reporting them exactly as they appear in the specification, without correction.

3. The "Cited By (267)" list — not prior art (for completeness)

Because the query said "each patent citation for 5635517," I want to pre-empt a common error. The many patent numbers the search surfaced — e.g.:

• US 5,712,291 A — D'Amato, The Children's Medical Center Corp., "Methods and compositions for inhibition of angiogenesis," granted 1998-01-27 (priority 1993-03-01).
• WO 98/03502 A1 — Celgene, published 1998-01-29 (the PCT counterpart of the sibling US 5,798,368).
• US 5,874,448 A — Celgene, 1999 (3-fluoropiperidin-3-yl isoindolines).
• US 6,281,230 / US 6,555,554 / US 7,119,106 — Celgene continuations claiming the same 1996-07-24 priority.
• US 5,798,368 A — "Tetrasubstituted 2-(2,6-dioxopiperidin-3-yl)-1-oxoisoindolines," listed as a co-pending application.

None of these is § 102 prior art to US 5,635,517. They either (i) share the same 1996-07-24 priority date (and so are not "prior"), or (ii) published after that date. They belong to the analysis of the later family members, not this patent's novelty.

One data artifact worth flagging literally: the retrieved Google Patents record lists, among "Related Parent Applications," the entry "US08/701,499 Continuation — US5748087A — Remote personal security alarm system — 1996-08-01." Taken literally that is an alarm-system patent, plainly unrelated to TNFα chemistry and inconsistent with the same application number's other appearance as a dioxopiperidin isoindoline case (US 5,798,368). Per the "do not auto-correct identifiers" rule I have not altered it, but I record it as an apparent database/parsing error rather than a genuine related application.

4. Reference-by-claim § 102 mapping (summary table)

Bottom line on anticipation: On the record I can verify, no single reference anticipates all elements of any of claims 1–9, because every one of claims 1–9 is a method claim and no cited reference discloses the step "administering [these compounds] to reduce TNFα." Claim 10 (a composition-of-matter claim to the four 1-oxo-2-(2,6-dioxopiperidin-3-yl)-aminoisoindolines, i.e., the lenalidomide genus) is the claim least exposed to the cited art, because Jonsson discloses only the 1,3-dioxo (both X and Y = C═O) species, not the 1-oxo (X = C═O, Y = CH₂) species. That structural distinction is precisely what carried claim 10 (and the corresponding Markush genus) to grant.

5. What I could not verify, stated explicitly

1. I could not retrieve the verbatim front-page "References Cited" table of US 5,635,517 (examiner-cited U.S. patent documents, foreign patent documents, and "Other References"). The supplied authoritative text omits it, and no search returned it. If you need the exact enumerated list (with examiner initials and the specific claims rejected), the reliable sources are the USPTO Patent Center / Patent Application file wrapper (IFW) for application 08/690,258, or the printed front page of the patent PDF. I would not want to guess at patent numbers for that table.
2. Whether Jonsson (1972) itself reports any biological/TNFα activity — this determines whether it is § 102 or § 103 art against the method claims. I could not confirm this from the retrieved material.
3. My searches were run against public web/patent-aggregator sources (Google Patents, FreePatentsOnline/JUSTIA indexed snippets, EPO publication server), not a direct authenticated USPTO database query. Where the record showed contradictions (e.g., the unrelated US 5,748,087 entry), I have flagged rather than resolved them.

If you can provide the face page (or the IFW "References Cited" block) for 08/690,258, I can complete the per-reference § 102 analysis exactly as requested, including any examiner-cited U.S. patent documents I was unable to surface here.

I'll ground the prior-art set before analyzing. Let me pull the actual references cited on the face of this patent and in the 2015 IPR petition.

Good — I now have the actual IPR prior-art set. Let me pin down what Piper, Kaplan, and Agrawal disclose.

Let me capture the Board's actual reasoning on the merits and the examiner-cited references from the file history.

Scope, sources, and two flags before the analysis

Sources actually used. The Google Patents excerpt I was given does not contain a populated "Prior Art" tab (only the prior-art keywords — dioxopiperidin, oxo, tnfa, aminoisoindoline, acid — and the "Cited By" list). So I grounded the prior-art set in the record that does exist for this patent:

• the references cited in the '517 specification itself (Jonsson 1972; WO 92/11383; the Beutler/Tracey/Hinshaw line);
• the '517 file history exhibits identified in IPR2015-01169 (Ex. 2006 "Notice of References Cited (1997)", Ex. 2007 "Information Disclosure Citation (1997)");
• the IPR2015-01169 exhibit set (Piper, Kaplan, Agrawal, WO '085, Keith, D'Amato, Koch II, Carswell, Jonsson, Schumacher);
• the Board's decision denying institution (Docket Alarm/PTAB OCR, incl. the copy in gov.uscourts.njd.399896.18.4) and Celgene's Preliminary Response (Paper 16, Aug. 19, 2015).

Flag 1 — date contradiction. The prior section states "Today's date: 2026-10-01"; this task states "Current Date: April 26, 2026." I use neither as a legal fact; nothing in the §103 analysis turns on it, but the inconsistency should be resolved in the final record.

Flag 2 — claim 1 construction contradiction. The prior section glosses claim 1 as "exactly one of X and Y is C═O and the other is C═O or CH₂." That is not what the claim says. Literally: "one of X and Y is C═O and the other of X and Y is C═O or CH₂." Claims 2 (X=C=O, Y=CH₂) and 7 ("each of X and Y is C=O") confirm that X is always C=O in claim 1. The mirror-image genus (X=CH₂, Y=C=O) is outside claim 1. This narrows the obviousness attack surface — a challenger must show the art suggests the N-glutarimidyl-isoindolin-1-one / isoindoline-1,3-dione orientation, not merely "some amino-substituted bicyclic imide."

Flag 3 — denial date. The prior section dates the institution denial to Dec. 22, 2015. The docket materials I retrieved place the Board's decision in the Nov. 16, 2015 docket entry (Jones Day's own client alert is dated Nov. 2015). Treat 2015-12-22 as unverified.

1. Legal framework (this is a pre-AIA patent)

Filing/priority date July 24, 1996 → pre-AIA 35 U.S.C. §103(a). Governing law is Graham v. John Deere as refined by KSR Int'l v. Teleflex (motivation may come from any source, but the record must supply it), the lead-compound line (Takeda v. Alphapharm; Eli Lilly v. Zenith Goldline), and the "obvious to try" limit of In re Cyclobenzaprine — the options must be "finite, small, or easily traversed," and the art must point to the route taken. Pre-AIA means the operative question is what the art taught as of July 24, 1996; anything published later (notably Muller et al., J. Med. Chem. 39:3238 (1996) and Engelhardt 2014, Ex. 1029) is not prior art and can only serve as evidence of unexpected results.

The claims are mixed: claims 1–9 are method-of-use claims (administering an effective amount to reduce undesirable TNFα levels); claim 10 is a compound claim (four 1-oxo aminoisoindolines). That distinction drives everything below — for a method claim, prior disclosure of the compound is not enough; there must be a suggestion of the use.

2. The prior-art corpus, and what each reference actually teaches

3. The combinations, claim by claim

The institution decision confirms the challenger's own architecture — three grounds:

• Ground 1: Piper + Kaplan → claims 1, 7–9
• Ground 2: Piper + Kaplan + Agrawal + WO '085 → claims 2–6, 10
• Ground 3: Piper + Kaplan + Agrawal + Keith → claims 2–6, 10

Motivation story the challenger can articulate (steel-manned). Piper and Kaplan share a field (thalidomide analogs), a disease (ENL), and a parent compound (thalidomide). Both teach that thalidomide treats ENL, and Kaplan teaches that thalidomide's benefit runs through TNFα. Piper and Agrawal teach that amino substitution on the benzo ring is a tolerated variation — 4-aminothalidomide "retained the anticarrageenan activity." A POSA seeking an improved TNFα inhibitor would therefore look to the tolerated amino analogs of the known TNFα-suppressing scaffold. Under KSR, "familiar elements according to known methods" plus a finite, enumerated set of analogs can supply the motivation. For the 1-oxo claims, EM 12 (Kaplan's compound H) supplies the 1-oxoisoindolinone skeleton, and Piper/Agrawal supply the amino-substitution teaching.

Where it breaks down, claim group by claim group:

Claims 8 and 9 — the most exposed claims. This is the only place where the composition itself is conceded prior art (Jonsson 1972, admitted in the specification; and the Board accepted that Piper's AH 13/AH 14 "encompass" claims 8 and 9). The only missing element is the use. Kaplan supplies a TNFα-reducing use for a structurally adjacent compound, and pre-1996 literature (Sampaio/Kaplan, J. Exp. Med. 173:699 (1991)) makes thalidomide's TNFα suppression common knowledge. So the strongest §103 sentence available against claims 8–9 is: "Known compound (Jonsson) + known TNFα-inhibitory activity of the parent (Kaplan/'901) + art-recognized tolerance of amino substitution on the benzo ring (Piper/Agrawal) = obvious to use." I rate claims 8–9 as moderately vulnerable — but the bridge fails for two concrete reasons the Board credited: (i) there was no demonstrated relationship between Piper's carrageenan/PFC assays and TNFα (Piper never mentions TNFα), so the "tolerated variation" teaching doesn't reach the claimed use; and (ii) the very analog data relied on affirm the opposite — AH 13 was active in only one assay and AH 14 was inactive/immunostimulatory.

Claims 2–6 and 10 — the 1-oxo genus and species. These are the patentee's actual novel compounds. No reference discloses any amino-substituted 1-oxoisoindoline or its glutarimide. Obviousness here requires the POSA to (1) select EM 12 as lead compound, (2) discard thalidomide's 1,3-dioxo ring in favor of the 1-oxo ring, (3) add an amino group at the 4-, 5-, 6- or 7-position, and (4) expect TNFα reduction — with zero guidance that the 1-oxo ring retains (let alone improves) TNFα activity. That is a double structural modification with an unguided positional selection. I rate claims 2–6 and 10 as low vulnerability.

Claim 1 (genus). No single reference, and no proper combination, discloses the full class called for (amino-benzo ring; X = C=O; Y = C=O or CH₂; administered for TNFα reduction). It is the genus of which the novel 1-oxo compounds are species, so it inherits their low vulnerability. Low.

4. Why the combinations fail on the merits (the part most analysts skip)

1. No motivation to focus on TNFα at the priority date. Piper and Agrawal (1981) predate the TNFα/thalidomide linkage; their endpoints (carrageenan paw edema, plaque-forming-cell assay) are not TNFα readouts. Kaplan postdates Piper by 14 years and supplies a different compound (EM 12) and a different genus.
2. No reasonable expectation of success. Thalidomide SAR is notoriously non-linear, and the record contains the proof: AH 13 active in one assay, AH 14 in neither, thalidomide/AH 3/AH 20 in both. In re Cyclobenzaprine–type "obvious to try" reasoning fails where the "options" are not a finite, predictable set. Celgene's own characterization — "billions of thalidomide-related compounds" — underscores the size of the search space.
3. The solubility motive backfires. Keith, the reference offered to motivate amination, reports that 4-aminothalidomide is no more soluble than thalidomide/EM 12 — the exact opposite of the asserted motivation.
4. Teaching away. Kaplan's own figures show compound G far more active than EM 12, so the art pointed the POSA away from EM 12 as a lead.
5. Prosecution and reexamination history. The examiner twice addressed substantially this art; the file records the examiner's conclusion that "it would not have been obvious to one skilled in the art to substitute NH2 group to Thalidomide to make the instant compounds for reducing TNF alpha levels" (quoted in Celgene's Preliminary Response, citing Ex. 1013). The 1998–99 ex parte reexamination confirmed claims 1–10 over the D'Amato-family disclosures. The Board then denied institution in its entirety on the ground that the motivation/expectation bridges were missing.

Where I would push back on the prior sections: the earlier summary treats the 2015 denial as essentially dispositive of §103. It is strong evidence (the Board agreed with the petitioner on several disclosure findings, then still denied), but a denial of institution is a reasonable-likelihood ruling on a petition record, not a final judgment of validity. Claims 8–9 in particular remain exposed on a cleaner record (e.g., a district-court record proving the Sampaio/Kaplan 1991 TNFα teaching as of 1996), because their compositions are admitted prior art and the asserted novelty is purely the use.

5. Objective indicia that would defeat the strongest ground

• Unexpected results: amino substitution produced a large, unexpected increase in TNFα inhibition relative to the parent scaffold — reported by the inventors themselves in Muller et al., J. Med. Chem. 39:3238 (1996). Careful: this is an inventors' own post-filing publication, so it is not prior art; it is rebuttal evidence of an unexpected property, and its nexus must be argued compound-by-compound.
• Failure of others / long-felt need: the record (Celgene's Preliminary Response) describes a multi-year program of "repeated failures" to find non-teratogenic thalidomide analogs. If documented in contemporaneous records, this is powerful.
• Commercial success: REVLIMID® (lenalidomide = 1-oxo-2-(2,6-dioxopiperidin-3-yl)-4-aminoisoindoline, claim 4 and within claim 10). Nexus is clean for claims 4 and 10; weak for claims 1 and 7–9, which do not cover the commercial species — a point a challenger should exploit.
• Countervailing weakness for the patentee: the '517 specification contains no TNFα assay data whatsoever. There is no in-patent demonstration that the claimed compounds reduce TNFα, and no comparative data against the closest prior art. Any unexpected-results argument must be built entirely from evidence extrinsic to the patent — which invites the challenger to attack nexus and commensurateness of scope.

6. Bottom line

The most defensible proposition an examiner or challenger can assert today is narrow and specific: claims 8 and 9, over Jonsson 1972 admitted as anticipating the compounds, in view of Kaplan's TNFα teaching and Piper/Agrawal's amino-substitution teaching. Every broader theory fails at the motivation/expectation step for the reasons the Board gave, and the 1-oxo claims (2–6, 10) have no structural antecedent anywhere in the 1996 art. Given the '517 patent's adjusted expiration of 2019-10-04, a live §103 ruling matters today chiefly as (i) an obviousness-type-double-patenting/priority anchor against the later family members (US 6,281,230; 6,316,471; 6,335,349; 6,476,052; 6,555,554; 7,041,680; 7,119,106) and (ii) support for the invalidity/inequitable-conduct theories pleaded in the 2026 New Jersey antitrust matter (N.D.N.J. 2:24-cv-…, Dkt. 1), where the '517 patent's prosecution history — including the 1997 Notice of References Cited and the 1998–99 reexamination — is itself the central evidentiary issue.

Confidence and unknowns: I am confident in the reference identities, dates, and the Board's reasoning (all retrieved from PTAB/party documents today). I am not confident in my inference that EM 12 is the 1-oxoisoindolinone that maps onto claims 2–6/10 — the Board described it only as containing "one C=O on the 5-carbon ring," so I am reading that description literally rather than asserting an IUPAC name. Likewise, I could not verify from primary sources which specific claim elements WO '085 was mapped to in Ground 2; that remains an open item.