US 6593123

I'll search for information on this specific patent number, including any 2026 litigation activity.

Let me run one more targeted check for any litigation or CAFC activity specifically involving this patent.

Let me retrieve the actual claim set of US 6,593,123 B1 (the truncated text cuts off before the claims).

US Patent 6,593,123 B1 — Summary

⚠️ Search-scope note up front: I found no evidence of any CAFC (Federal Circuit) 2026 docket, appeal, IPR, or district-court action involving U.S. Patent 6,593,123 specifically. My searches of 2026 Federal Circuit summaries, PTAB/PTACTS petition records, and litigation coverage returned nothing naming this patent number. The AAV gene-therapy patent litigation that is active in 2024–2026 (Genzyme/Avigen successor v. Novartis over Zolgensma; Genzyme v. Sarepta over Elevidys) asserts different patents — e.g., U.S. 9,051,542; 7,704,721; 7,125,717; 7,846,729; 8,093,054; 7,785,888; 6,596,535; 12,013,326; 12,031,894. I cannot confirm any adjudication of '123. Treat the litigation statement as "nothing found," not as "confirmed inactive."

Bibliographic data (from the patent front page / Google Patents)

Identifier caveat (no auto-correction applied): the front-page inventor listing reads "Quang Qu," but the recorded assignment entries on Google Patents name the assignor as both "QU, QUANG" (2000-08-07 record) and "QU, GUANG" (2001-12-10 record). I am reporting both spellings as they literally appear rather than reconciling them. Also note both "Microfluidics International Corp." and "Pall Membrane Technology Corporation" and "Pall Corporation" appear as the cited vendor names — reported as written.

Family (related filings): PCT/US2001/018443 → WO 2002/012455 A1; AU 2001269757 A1; and U.S. 10/440,704 → US 2003/0207439 A1 (a continuation of 09/633,834, published with a different claim set — do not conflate the two documents' claims).

Abstract (verbatim, as printed)

"Methods are provided for large-scale purification of recombinant AAV (rAAV) virions that were produced in the absence of infectious adenovirus. Preferably, the rAAV is produced in a host cell line via triple-transfection with an accessory function vector, an AAV vector, and an AAV helper vector. The methods include preparing a lysate from the host cell line and passing that lysate over various combinations of ion exchange chromatography media and/or affinity chromatography media. The affinity chromatography medium is an AAV receptor or an antibody with binding affinity for AAV, e.g., heparin sulfate. A variety of cation exchange and anion exchange media are contemplated by the present invention. In certain embodiments, optional purification steps may be included, such as filtering the lysate through one or more filters, or treating the lysate with a nuclease."

Plain-language overview of the independent claims

The patent has three independent claims — claims 1, 12, and 20.

Claim 1 — Anion-exchange flow-through, then heparin/affinity capture, with a sarcosine wash.
A method of purifying rAAV from contaminants that strings together the production and the purification:

• (a)–(c) introduce an AAV vector into a host cell; introduce a first nucleic acid supplying AAV helper functions and a second nucleic acid supplying accessory functions, where the accessory nucleic acid and the host cell together lack a gene needed to make adenovirus (i.e., helper-virus-free production); then culture the cell to make rAAV;
• (d) prepare a lysate;
• (e)–(f) run the lysate over an anion exchange medium under conditions where the contaminants bind but rAAV flows through (flow-through mode), and collect that flow-through;
• (g)–(i) pass the flow-through over an affinity medium, wash with an anionic detergent comprising sarcosine, then elute purified rAAV.
  The sarcosine wash (as opposed to, e.g., urea or deoxycholate) is the distinguishing purification step — the specification reports it removed the most contaminating protein.

Claim 12 — Cation-exchange capture, then affinity capture, plus sarcosine wash.
Same purification goal, but starting from a simpler premise ("providing a host cell comprising rAAV virions," with no recitation of how it was made):

• (b)–(e) make a lysate, load it on a cation exchange medium so the rAAV binds (bind-elute mode rather than flow-through), wash, and collect the eluate;
• (f)–(i) pass the eluate over an affinity medium, wash with sarcosine-containing anionic detergent, and elute purified rAAV.

Claim 20 — Anion-exchange flow-through, then affinity capture (truncated in the available text).
Also a "purifying rAAV from contaminants" method, with steps including "(f) passing said eluate over an anion exchange chromatography medium, thereby binding said contaminants to said anion exchange chromatography medium" and "(g) passing the eluate obtained in step (f) over an affinity chromatography medium…".
Uncertainty: the full text of claim 20 (steps (a)–(e) and the closing steps) and claim 21 were cut off in the sources I could retrieve; I do not have authoritative wording for the complete claim 20 or for dependent claim 21, and I am not going to reconstruct them from the continuation's claim set, which differs.

Dependent claims — quick map

• Claim 2 / 3: order of the AEX vs. affinity steps (AEX-first or affinity-first).
• Claim 4 / 5 (and 18/19): affinity medium is an AAV-binding molecule — an AAV receptor or an anti-AAV antibody; preferably heparin sulfate.
• Claim 6: lysate made by microfluidization.
• Claim 7: further comprising filtering the lysate.
• Claim 8: host cells cultured substantially serum-free.
• Claim 9: treating the lysate with a nuclease (specification: BENZONASE™).
• Claim 10: host cell line cultured in a bioreactor.
• Claim 11: host cell line transfected with an accessory function vector, an AAV vector, and an AAV helper vector (triple transfection).
• Claims 13–17: step-order permutations for the claim 12 cation-exchange embodiment.
• Claim 14: cation exchange medium = sulfo-, phospho-, carboxy-, or carboxy-methyl-based resin.

Practical context

• Why it mattered: the patent's stated problem is that prior rAAV purification processes were built to remove infectious helper adenovirus and therefore required heat inactivation (≈50% titer loss). Because rAAV here is made without infectious helper virus, the process can be a shorter, scalable, column-only train (roller bottles / bioreactor → microfluidization → filtration → chromatography), with the specification reporting >80% overall rAAV recovery by Q-PCR across the exemplified AEX + heparin process.
• Cited prior art of record includes WO 96/27677, WO 97/08298, WO 99/11764, WO 99/61643, WO 00/22152, WO 00/23116, plus Ferrari et al. (Nature Medicine 1997), Xiao et al. (J. Virol. 1998), Qu et al. (ACS abstract 2000), Anderson et al. (2000), Grimm et al. (1998), and Clark et al. (1999).
• Downstream relevance: '123 is frequently cited as background art and expressly incorporated by reference in later AAV manufacturing/production patent families (e.g., EP 1,951,888; EP 2,781,596; EP 2,529,020; and recent WO/US applications on ion-exchange AAV purification) — so its main modern footprint is as cited literature, not as an asserted or adjudicated patent.

Bottom line on your two questions: (1) The bibliographic and claims summary above is well-grounded in the patent text itself. (2) As for "USPTO database and CAFC 2026 dockets for 6,593,123" — I found no 2026 CAFC docket or pending litigation on this patent, and given its recorded expiration on 2020-08-07, any such proceeding would be unusual. I cannot rule out a filing that isn't indexed in the sources I could reach, so please verify against PACER/CAFC docket search directly if that detail is load-bearing.

I'll run targeted searches for litigation specifically naming US 6,593,123.

Let me run additional targeted checks, including the assignee names and litigation-portal sources.

Litigation search results for U.S. Patent No. 6,593,123

Bottom line

I found no litigation — district court, CAFC, ITC, or PTAB — in which U.S. Patent No. 6,593,123 ('123) is named as an asserted, challenged, or adjudicated patent.

I searched for the patent number in both 6,593,123 and 6593123 formats, alone and in combination with the assignees (Avigen, Genzyme), the technology (AAV purification, rAAV), and patent-litigation sources (Unified Patents litigation materials, CourtListener/RECAP, PTACTS/USPTO petition records, CAFC opinions, Justia dockets, and litigation-prone securities filings). Every hit for the number was a false positive — a different patent with a coincidentally similar digit string ('593, '054, '542, '888, '326, 6,535,815, etc.) — or a citation of '123 as prior art/incorporated disclosure in a later patent document (e.g., EP 2,781,596; EP 2,529,020). No docket names '123.

Returning zero results here is the honest answer. Per the operating rule, I am not padding this with the adjacent AAV cases that look related but involve different patents. I also cannot rule out a filing that isn't indexed in the sources I could reach.

The adjacent AAV litigation — confirmed NOT to involve '123

These are the cases a search engine will surface for "Avigen/Genzyme AAV patent litigation." For each, the asserted patents are identified and '123 is not among them:

Note also the counterclaim excerpt from the D. Del. AAV case (Dkt. 18 in 1:22-cv-00780-style docket) that attacks the '054, '542, and '888 patents under §§ 102/103/112 — again, not '123.

Why no litigation is found

The record is consistent with '123 never being asserted:

• Timing. The patent issued July 15, 2003, was filed August 7, 2000, and reached anticipated expiration on 2020-08-07 (Google Patents legal status: Expired – Lifetime). A method patent of this vintage — and one whose owner (Avigen) exited gene therapy via the December 2005 asset sale to Genzyme — would have had to be asserted during roughly 2003–2019 to generate recoverable damages. I found no such suit.
• Commercial role. '123 describes manufacturing/purification process technology (flow-through AEX + heparin affinity; Mustang S cation exchange; sarcosine wash). Zolgensma and Elevidys are made by third-party CDMOs under processes that, if anything, post-date or differ from the '123 examples. Even in the active Genzyme cases, the asserted patents are vector/capsid and analytics patents, not the '123 purification patent.
• Post-expiration posture. With expiry in 2020, an IPR or DJ action now would be unusual (and would face mootness/standing questions). My PTAB/PTACTS searches returned nothing naming '123.

Caveats and how to verify (load-bearing detail)

• Search scope limit: my tooling reached litigation-aggregator pages, CourtListener/Justia excerpts, PTACTS petition PDFs, and CAFC filings, but it is not equivalent to a complete PACER Docket Report or a full CAFC docket query. Treat "none found" as "no evidence," not "confirmed never asserted."
• To close this out authoritatively, run:
  1. PACER — party search for "Avigen" and "Genzyme Corporation" as plaintiff across all districts 2003–2020, then grep complaints for 6,593,123;
  2. CAFC docket search (cafc.uscourts.gov) for the patent number;
  3. PTAB/PTACTS proceeding search on 6,593,123;
  4. Unified Patents litigation portal and Docket Navigator / Lex Machina patent-number search (the LLM-accessible free layers of these do not expose full dockets).

Cross-reference note

This is consistent with the previously generated sections — which also found no CAFC 2026 docket, appeal, IPR, or district-court action involving 6,593,123. No contradiction between the two passes: both report "nothing found." The predecessor analysis additionally flagged (and I confirm here) that '123's modern footprint is as cited literature — it is cited in EP 2,781,596 (at [0073]) and EP 2,529,020 (at [0047]) as background/incorporated material — not as an asserted or adjudicated patent.

I'll start by verifying the structured data against public sources before writing anything up.

Nothing relevant yet. Let me run a few more targeted checks covering alternative proceeding-number formats and the parties most likely to have filed.

Let me do two final verification passes — one on the third-party patent portals, and one against the earlier Avigen/Genzyme IPR cases to confirm none involved this patent number.

Proceedings overview

Total AIA trial proceedings on file for U.S. Patent 6,593,123: ZERO (0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denied). The structured USPTO Open Data Portal block supplied in this prompt returns no AIA trial proceedings for this patent, and my independent searches of PTAB decision/petition records and litigation coverage surfaced no IPR, PGR, or CBM naming 6,593,123. Because there are no proceedings, there is no per-proceeding section to write — inventing proceeding numbers would be fabrication, and I will not do it.

Bottom-line defensive posture: neither of the framing options in the task prompt applies. The patent has not "survived two IPRs and is hardened," and no claims have been canceled. The accurate statement is: the '123 claims are UNTESTED at the PTAB — all 21 claims stand as issued and unamended, with no FWD, no certificate, and no estoppel anywhere in the chain. For a defendant, that means (a) no claim is pre-canceled for you to leverage, and (b) conversely, no petitioner has ever run the prior art that is on the '123 face at the Board, so the PTAB record is a blank slate rather than a hostile one.

---

Per-proceeding detail

None to report. Every field below would be populated only if a proceeding existed; I am stating the null result rather than filling the template with placeholders:

Verification performed

• Structured source (canonical): the "PTAB proceedings on file" block for this patent returns an empty AIA-trial set from the USPTO Open Data Portal. Treated as authoritative; no deviation found.
• Web verification: searches on the patent number, on the title ("Large-scale recombinant adeno-associated virus (rAAV) production and purification"), and on the assignee chain (Avigen → Genzyme) returned no petition, institution decision, or FWD. Hits for 6,593,123 were patent-family, citation, and prior-art-listing pages — e.g., Google Patents, Justia, and third-party patent portals listing '123 only as cited art — not as a challenged patent.
• Negative-control check: the AAV IPRs that do exist in this technology space target different patent numbers. The Novartis v. Genzyme dispute I surfaced involves U.S. 9,051,542 (the '542 patent) — a different Avigen-origin patent with a different specification and claim set. I could not confirm a proceeding number for it from the retrieved material, so I am not asserting one, and in any event it is not a proceeding on '123. Do not conflate the two.

Why the absence is credible here (not just an indexing gap)

1. PGR is time-barred. PGR requires a petition within 9 months of grant (35 U.S.C. § 321(c)). '123 issued 2003-07-15. A PGR was never legally available after ~2004-04-15.
2. CBM is unavailable by subject matter and by sunset. This is a biopharmaceutical manufacturing/purification patent, not a "covered business method" patent under AIA § 18(d). Separately, CBM review sunset for new petitions on 2020-09-16.
3. Only IPR was ever a realistic vehicle — and the patent's 20-year term ran from 2000-08-07, with the record showing anticipated expiration on 2020-08-07 and legal status "Expired – Lifetime."
4. Expiration is not itself a bar — this is the important practitioner nuance. The Board does institute IPRs against expired patents where past damages are in play (the Supreme Court cert petition in Uniloc v. Sega was litigated on precisely that Article III/Seventh Amendment question). So the absence of IPRs is not explained away by expiration; it more likely reflects that '123 was never asserted in a campaign that drew a counterattack, or that any accused party chose district-court invalidity instead.

---

Strategic summary

Claim status. All of claims 1–21 are UNTESTED. None canceled, none disclaimed, none amended, no statutory disclaimer of record appears in the sources I retrieved, and no reissue or reexamination certificate altering them was found. Practically, this means the claim set described in the prior section — independent claims 1 (AEX flow-through → heparin/affinity with sarcosine wash), 12 (CEX bind-elute → affinity with sarcosine wash), and 20 (AEX → affinity), plus dependents 2–11 and 13–21 — remains exactly as issued. There is no PTAB narrowing to plan around.

Estoppel landscape. There is none. Section 315(e)(2) estoppel only attaches to a petitioner that has obtained an FWD; with no petitioner and no FWD, no party (and no privy) is estopped from raising any § 102/§ 103 ground in a civil action. This cuts both ways: a defendant gets a clean IPR slate, but also cannot borrow the benefit of someone else's invalidity work product. Separately, because no proceeding exists, the § 315(b) one-year bar has not been triggered by anyone's service of complaint — the running of that clock for a future defendant depends solely on when that defendant is served.

Pattern signals. No repeat petitioner, no defensive aggregator (I found no Unified Patents or similar challenge on '123; third-party portals list it only as prior art). No patent-owner PTAB appeal activity is possible, since there is no Board decision to appeal — which is consistent with the prior section's finding of no CAFC 2026 docket on '123. The two sections agree; no contradiction to flag on that point. The one item worth flagging explicitly: the task prompt's illustrative examples ("survived two IPRs and is hardened" / "claims 1-5 have been canceled") describe scenarios that do not exist for this patent — do not let those phrases migrate into any client-facing memo as if factual.

---

Recommended next steps

1. If you are a defendant: do not assume the patent is dead or unassailable. '123's modern footprint is as cited and incorporated-by-reference background art in later AAV manufacturing families (the prior section identifies EP 1,951,888; EP 2,781,596; EP 2,529,020, and recent WO/US ion-exchange AAV purification filings). That is a pointer — the references those later patents cite against themselves are a ready-made starting kit for an obviousness theory, but none of that art has ever been tested at the PTAB against '123.
2. Check the real assertion risk first. With a recorded expiration of 2020-08-07, pre-suit damages recovery under 35 U.S.C. § 286 is limited to the 6 years before filing — for a complaint filed now (2026), that window opened around 2020-09, largely after the patent expired, so the recoverable past-damages tail is short and may be entirely foreclosed depending on the exact complaint date and the patent's terminal-disclaimer/term-adjustment history. Pull the full prosecution history and any PTA/terminal disclaimer from Patent Center before spending on validity work; economic exposure may moot it.
3. If you do file an IPR: note the current institution climate is materially stricter than pre-2025. Discretionary denial factors have been expanded and centralized, and "settled expectations" based on patent age has been a driver of denials — a patent in force for ~20 years before expiration is the paradigm case for that factor. Any petition on '123 would need to be written to clear the discretionary gate, not just the merits.
4. Verify directly, since my conclusion is a negative finding. Absence of evidence in search results is not proof of non-existence. Confirm against primary sources before relying on this:
   • PTAB E2E / PTAB Decisions: https://ptab.uspto.gov/ and https://developer.uspto.gov/ptab-api/ (or the PTAB "Patent Trial and Appeal Board" document search at https://www.uspto.gov/patents/ptab)
   • USPTO Patent Center for the file history, any disclaimer, and term adjustments: https://patentcenter.uspto.gov/
   • CourtListener for any Federal Circuit docket involving '123: https://www.courtlistener.com/?q=%226593123%22
   • Google Patents legal-events page for the patent: https://patents.google.com/patent/[US6593123B1](/patent/US6593123B1)/en
5. One item I could not affirmatively verify and therefore flag rather than assert: whether any ex parte reexamination (a non-AIA, ex parte proceeding that would not appear in the ODP AIA-trial block) was ever filed against '123. My searches surfaced none, but reexaminations are indexed separately and I did not have a source that affirmatively rules them out. This is outside the "AIA trial proceeding" scope of this section, but if the file history shows a reexam certificate, it would change the claim-status analysis in step 1 and must be checked in Patent Center.

I'll research the assignment record for this patent before constructing the analysis.

Assignment & Ownership Analysis — US 6,593,123 B1

⚠️ Data-access limitation stated up front (do not read this as a finding of no records): I could not load the USPTO Assignment Center (or its assignment-api.uspto.gov endpoint) in this session — only web-search retrieval was available, and the Assignment Center pages/API are not surfaced by search. I therefore could not retrieve the reel/frame numbers, execution dates, or the correspondent-of-record for the recorded assignments. Rather than invent reel/frame values (explicitly prohibited), I mark those fields [not retrieved] below and reconstruct the chain from Google Patents Legal Events plus primary SEC/contract documents, which are internally consistent. Verify reel/frame and correspondent directly at https://assignmentcenter.uspto.gov/ by searching patent number 6593123.

Contradiction flag vs. the prior summary: none on substance, but note one addition — the prior summary listed only the 2008-01-18 Genzyme recordation. The underlying transaction was executed 2005-12-19, not 2008. The 2008 date is a recordation date, ~25 months after execution. That distinction matters for any pre-litigation-transfer timing analysis.

---

Inventors

Pattern assessment — no adverse tell. The tell to look for is all inventors decamping within 12 months of filing (a leading indicator of a portfolio fire-sale). That is not present here: both inventors were Avigen staff at filing (2000-08-07) and the patent issued to Avigen in 2003. Avigen's gene-therapy wind-down came in 2005, five years later, and was a corporate strategic pivot — not an inventor-driven abandonment.

Identifier caveat carried forward (no auto-correction applied): the front page and the 2000-08-07 record spell the second inventor "QU, QUANG"; the 2001-12-10 record spells him "QU, GUANG." Both are reported literally. This looks like a data-entry variance across two recordations of the same inventor's interest, not two different people.

---

Original assignee

Avigen, Inc., Alameda, California (Nasdaq: AVGN).

• Entity on the issued patent: Avigen, Inc. (assignment records show Avigen as assignee from 2000; current Google Patents assignee field reads "Genzyme Corp" post-2008).
• Primary line of business: gene-therapy platform company — recombinant AAV (rAAV) vector production, process development and clinical AAV programs. It operated an AAV manufacturing facility at 1201 Harbor Bay Parkway, Alameda (Assignment Agreement §1.26), i.e., it was a genuine R&D/production operating company, not a holding vehicle.
• Did it ship a product embodying the claims? No commercial product. Its most advanced AAV program was AV201, a Phase 1/2 Parkinson's disease trial (initiated Dec 2004 at UCSF/LBNL). The '123 claims are process/manufacturing claims that Avigen practiced internally to make clinical-grade vector.
• Current status: Avigen no longer exists as an independent company. It sold its non-pain AAV assets to Genzyme Corporation by Assignment Agreement executed 2005-12-19 ($12M upfront plus milestones/royalties on all products developed under the portfolio). After a 2008 restructuring, an early-2009 proxy fight, and a stated intent to liquidate if no sale occurred, Avigen merged into MediciNova, Inc. (announced 2009-08-20; the SDCA complaint records the merger as occurring 2009-12-18). MediciNova assumed Avigen's rights under the Genzyme agreement, i.e., the contingent milestone/royalty stream — not the patents themselves, which had already passed to Genzyme.

---

Assignment timeline

Two primary-source anchors for link 3 (grounding the substance where the USPTO fields are missing):

• Assignment Agreement, Avigen → Genzyme, effective Dec 2005, including §2.8(c) recordation mechanics and definitions of "Gene Therapy Listed Patents" — https://contracts.justia.com/companies/avigen-inc-50338/contract/[840881](/patent/840881)/
• Joint press release, 2005-12-21: "Genzyme has made an up-front cash payment of $12 million to Avigen, and will make potential milestone and royalty payments…" (republished within the same Justia contract page).

No post-2008 assignment is recorded. The chain terminates at Genzyme Corporation — consistent with Google Patents' "Current Assignee: Genzyme Corp" and the Expired – Lifetime legal status (anticipated expiration 2020-08-07).

---

Timeline diagram

timeline
    title Ownership of US 6593123
    2000 : Application filed 7 Aug 2000
         : Inventors assign to Avigen Inc
    2001 : Second inventor assignment recorded
         : Same parties and same assignee
    2003 : Patent issued 15 Jul 2003
    2005 : Avigen sells AAV estate to Genzyme
         : Executed 19 Dec 2005
    2008 : Avigen to Genzyme recorded at USPTO
    2009 : Avigen merges into MediciNova
    2020 : Patent expires 7 Aug 2020

---

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The only substantive transfer is Avigen → Genzyme, both operating companies. Avigen ran a GMP AAV manufacturing facility; Genzyme is a large-cap biologics manufacturer (Sanofi subsidiary since 2011). Neither is a "IP Holdings/Licensing/Ventures" shell, and neither assignee address is a registered-agent service address.

2. Known asserter in the chain — not present. Neither Avigen, Inc. nor Genzyme Corporation appears on the Acacia / Marathon / IV / IPNav / Wi-LAN / Mosaid-Conversant / Vringo / Pendrell / Round Rock / Spangenberg NPE rosters. Genzyme is a product-side defendant-and-plaintiff in gene-therapy patent litigation, the inverse of the NPE profile.

3. Repeat correspondent across the chain — unclear / not retrievable. This is the single most diagnostic field for this assignment, and it is precisely the field I could not pull from Assignment Center. I am not going to infer a correspondent from firm reputations. Retrieve it and re-check: if the same attorney/firm appears on both the 2000/2001 inventor assignments and the 2008 Avigen→Genzyme recordation, that is normal house-counsel continuity (weak signal); the strong version of this signal requires the same correspondent appearing across unrelated assignee entities, which does not arise here because there are only two assignee entities in the chain.

4. Cascading transfers — not present. Two links, ~5 years apart (2000/2001 → 2005 exec / 2008 recordation). No chained-LLC sequence inside any 24-month window.

5. Pre-litigation transfer — not present for '123. The Genzyme transfer (exec. 2005-12-19) predates by ~14 years the Genzyme/Sanofi AAV assertions (the first amended complaint asserting U.S. 9,051,542 against Novartis was filed 2022-02-23 in D. Del.). No infringement suit naming '123 was found in the earlier analysis. So the "arranged-to-enable-assertion" tell does not apply to this patent.

6. Bankruptcy fire-sale — not present for this patent. Avigen's 2005 AAV divestiture was a strategic pivot ("strategic move from a gene therapy company to… small molecule therapeutics"), not a Chapter 7/11 sale. Avigen's distress (Nov 2008 restructuring; 2009 liquidation threat; MediciNova merger) post-dates the transfer and involved only Avigen's residual small-molecule assets and the Genzyme royalty stream — never '123.

7. Privateering — not present. Genzyme asserts the Avigen portfolio on its own behalf as the record owner and as the party practicing/prosecuting the technology, not as a proxy asserting another operating company's patent against that company's competitors.

8. Defensive aggregator — not present. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN.

Cross-reference note (relevant, but does not change the verdict): MediciNova holds a contingent economic interest (milestones/royalties) in products covered by the same patent estate, pursuant to the Avigen→Genzyme agreement (SDCA 3:14-cv-02513; MediciNova, Inc. v. Genzyme Corp., claim-construction on "a stock of recombinant adeno-associated virus" of U.S. 6,376,237). That is a royalty-monetization structure, not a patent-assertion structure — MediciNova does not own '123 and does not sue on it.

---

Verdict

Operating-company assertion.

Justification: the entire recorded chain runs inventors → Avigen, Inc. (2000/2001) → Genzyme Corporation (executed 2005-12-19; recorded 2008-01-18) — two operating companies and zero licensing-only entities, with no cascading LLCs, no shell assignees, no defensive aggregator, and no pre-suit transfer. The NPE tells in the rubric are uniformly absent; the closest thing to an assertion posture is Genzyme/Sanofi (an operating manufacturer) later suing actual competitors over other patents from the same Avigen estate (e.g., U.S. 9,051,542 against Novartis/Zolgensma; the Genzyme v. Sarepta line) — and no assertion of '123 itself was identified, which is why this is an operating-company call made at the portfolio level rather than a patent-specific one. Caveat, stated plainly: with no reel/frame and no correspondent retrieved, signals 3 and any finer recordation-anomaly check remain unverified; if the reel/frame pull shows a correspondent or assignee entity not accounted for here, that would be the one thing that could move this verdict.

Verify at: https://assignmentcenter.uspto.gov/ (search patent no. 6593123) — mirrored legal events also at https://patents.google.com/patent/[US6593123B1](/patent/US6593123B1)/en.

I'll research the prior-art references cited on the face of US 6,593,123 and verify their bibliographic details before analyzing.

Let me research the specific cited prior-art references individually to verify their bibliographic details.

Let me verify the remaining cited references (WO publications and U.S. patents).

Prior-Art Analysis — US 6,593,123 B1

⚠️ Scope & method note (read first)

1. What I verified vs. what I inherited. I confirmed the target patent directly: US 6,593,123 B1, application 09/633,834, filed 2000-08-07, issued 2003-07-15, "Large-scale recombinant adeno-associated virus (rAAV) production and purification." The full claim 1 text was independently retrieved (patents.justia.com) and matches the earlier section verbatim, including the sarcosine-containing anionic detergent wash (step (h)) and the helper-virus-free production clause ("said second nucleic acid and said host cell collectively lack a gene necessary for adenovirus production").

2. The citation list itself is inherited, not re-pulled. In this session I could not load a page containing the patent's own front-page "References Cited" / "Patent Citations" block. The reference list below therefore comes from the earlier generated sections of this analysis (which listed WO 96/27677; WO 97/08298; WO 99/11764; WO 99/61643; WO 00/22152; WO 00/23116; U.S. 5,139,941; U.S. 5,173,414; U.S. 6,004,797; U.S. 6,001,650; U.S. 5,225,347; and NPL: Ferrari 1997; Xiao 1998; Qu 2000; Anderson 2000; Grimm 1998; Clark 1999). If the issued patent's actual citation list differs from this, this analysis must be re-run. Google Patents' own "Prior art keywords" for the patent — raav, aav, chromatography medium, lysate, exchange chromatography — corroborate that the examiner's core art was AAV purification/chromatography, which is consistent with the list.

3. Legal caution on "anticipates." Anticipation under § 102 requires a single reference disclosing every limitation, arranged as claimed, enabled. A reference merely cited on the face of the patent is not thereby anticipatory art — many of these are § 102 background citations, and the substantive rejections the examiner considered were almost certainly § 103 combinations. I flag confidence on each item; where I could not verify a title/date this session I say so rather than guess.

---

A. Critical-date framework (pre-AIA § 102 applies)

Consequences for the cited set:

• § 102(b) bars (published >1 yr before filing): Ferrari 1997; Grimm 1998; Xiao 1998; Clark 1999 (Apr); WO 96/27677; WO 97/08298; and those 1999 WO publications actually printed before Aug 1999.
• § 102(a)-only (within the year): Anderson et al. 2000; WO 00/22152; WO 00/23116; any 1999 WO published after Aug 1999.
• Likely NOT prior art at all (post-filing): Gao et al., Hum. Gene Ther. 11:2079-2091 (2000) — if that is the "Qu et al. 2000" item, its late-2000 issue date places it after 2000-08-07, so it would not be § 102(a)/(b) art (journal articles are not § 102(e) art). Flagged for confirmation of the exact venue/date.

---

B. Reference-by-reference

U.S. patent documents

1. U.S. 6,001,650 — Avigen, Inc.; "High-efficiency wild-type-free AAV helper functions"; priority 1995-08-03, granted 1999-12-14. (VERIFIED via Google Patents family listing: "US6001650A, 1995-08-03, Avigen, Inc., High-efficiency wild-type-free AAV helper functions.")

• Description: AAV helper constructs (rep/cap) engineered to avoid generating wild-type AAV by homologous recombination — the same concept the '123 specification invokes for its "modified AAV p5 promoter lacking an intact TATA box."
• § 102 mapping: Bears on the helper-function construct recited in independent claim 1 step (b) and on claim 11 / claim 17-type "helper vector" dependents. Does not disclose lysing or any chromatographic step → cannot anticipate claims 1, 12, or 20.

2. U.S. 6,004,797 — described in the '123 specification as disclosing "the production of rAAV without infectious helper virus" (adenoviral VA RNA / E4 ORF6 / E2a 72 kD accessory functions on a vector rather than by infection).

• Description: Replication-incompetent accessory-function vector system; expressly lacks the adenovirus fiber gene (i.e., "collectively lack a gene necessary for adenovirus production").
• § 102 mapping: This is the closest cited U.S. document to the production half of claim 1, steps (a)–(c), and to claim 11 (triple transfection) and claim 10 (bioreactor). Still cannot anticipate any independent claim because it discloses no lysis and no chromatography. (Publication/grant date not independently re-verified this session.)

3. U.S. 5,139,941 — the '123 specification cites it twice for AAV rep-region-containing vectors (ATCC 53222–53226) and cap-region-containing vectors.

• Description: Foundational AAV expression/vector constructs.
• § 102 mapping: Background art for AAV vector + helper-construct elements. No production, lysis, or purification teaching → not anticipatory of claims 1, 12, 20. (Title/inventor/grant date not re-verified this session — reporting the number and the patent body's characterization only.)

4. U.S. 5,173,414 — cited as an AAV vector-construction reference (recombinant genome construction).

• § 102 mapping: Background; AAV vector construction. Not anticipatory of the purification claims. (Title/date not re-verified.)

5. U.S. 5,225,347 — cited in the '123 specification as "U.S. Pat. No. 5,225,347 to Goldberg et al." in connection with ribozymes.

• § 102 mapping: Relates to a heterologous payload (ribozyme), not to production or purification. No bearing on claims 1/12/20; at most background for the "heterologous nucleotide sequence" discussion. (Attributes taken from the patent text; not re-verified.)

PCT / WO publications

6. WO 00/22152 — cited and expressly incorporated by reference in the '123 specification for cell-culture media and transfection (100–500 mL media, 0–20% FCS, PEI/Superfect lipofection).

• § 102 mapping: Cell-culture/transfection process-parameter art. Could bear on dependents reciting serum-free culture (claim 8) or media/transfection, but discloses none of the chromatographic steps → not anticipatory of claims 1/12/20.

7. WO 00/23116 — cited art in the AAV production/purification space.

• § 102 mapping: As a WO "00" publication (early 2000) it is at most § 102(a) art. (I could not independently verify its title/subject this session — flagged as unverified; treat its substantive mapping as provisional.)

8. WO 96/27677 and 9. WO 97/08298 — 1996/1997 PCT publications; § 102(b) art by date.

• (I could not independently verify their titles/subjects in this session. They are cited for the AAV vector/production background. Substantive mapping provisional — do not treat as anticipatory without pulling the documents.)

10. WO 99/11764 and 11. WO 99/61643 — 1999 PCT publications.

• § 102(b)/(a) art depending on exact print date around the 1999-08-07 line. Both sit in the AAV production/purification art cluster. (Titles/subjects not independently verified this session — provisional.)

Non-patent literature

12. Clark, K.R., Liu, X., McGrath, J.P., Johnson, P.R. (1999) — "Highly purified recombinant adeno-associated virus vectors are biologically active and free of detectable helper and wild-type viruses," Human Gene Therapy 10(6):1031–1039, published 1999-04-10 (PMID 10223736; DOI 10.1089/10430349950018427). (VERIFIED.)

• Description: Single-step heparin affinity chromatography purification of AAV-2 from crude lysate, >70% recovery, product free of detectable adenovirus and cellular contaminants; plus real-time-PCR vector quantification.
• § 102(b) art. This is the single most damaging reference to the affinity step and to any claim attempt to monopolize "purify rAAV by heparin affinity chromatography." But: it discloses no anion-exchange flow-through, no cation-exchange step, and no sarcosine wash → cannot anticipate claims 1, 12, or 20 (it lacks the (e)–(f)/(h) limitations). Strong § 103 partner.

13. Grimm, D., Kern, A., Rittner, K., Kleinschmidt, J.A. (1998) — "Novel tools for production and purification of recombinant adenoassociated virus vectors," Human Gene Therapy 9(18):2745–2760, published 1998-12-10 (PMID 9874273; DOI 10.1089/hum.1998.9.18-2745). (VERIFIED.)

• Description: The pDG packaging/helper plasmid supplies all AAV and adenovirus functions → helper-virus-free rAAV stocks; and an A20 monoclonal anti-AAV-2-capsid antibody affinity purification procedure.
• § 102(b) art. Discloses (i) helper-virus-free production (claim 1 steps (a)–(c) concept; claim 11) and (ii) an anti-AAV-antibody affinity medium (the alternative recited in claims 4/5). Cannot anticipate claims 1/12/20 (no ion-exchange, no sarcosine). Excellent § 103 partner, especially for the "affinity medium = anti-AAV antibody" dependents.

14. Xiao, X., Li, J., Samulski, R.J. (1998) — "Production of high-titer recombinant adeno-associated virus vectors in the absence of helper adenovirus," J. Virology 72(3):2224–2232 (March 1998).

• Description: rAAV production without helper adenovirus using adenoviral genes (E2a, E4, VA RNA) supplied on a plasmid rather than by infection — i.e., the classic "accessory-function vector" method leading to the helper-virus-free platform.
• § 102(b) art. Directly relevant to claim 1 steps (a)–(c) (helper-function + accessory-function nucleic acids; host cell lacking the adenovirus-production gene) and to claim 11. Cannot anticipate an independent claim on its own (no lysis/chromatography). (Bibliographic detail stated from field knowledge; the retrieved search evidence corroborates the "absent helper adenovirus" subject matter.)

15. Anderson, R., Macdonald, I., Corbett, T., Whiteway, A., Prentice, H.G. (2000) — "A method for the preparation of highly purified adeno-associated virus using affinity column chromatography, protease digestion and solvent extraction," Journal of Virological Methods 85:23–34 (DOI 10.1016/S0166-0934(99)00150-0). (VERIFIED via citing reference list.)

• Description: Purification of AAV by affinity column chromatography combined with protease digestion and solvent extraction.
• § 102(a)-era art (published 2000, within one year of filing). Relevant to the affinity step of claims 1/12/20, and its protease digestion teaching is arguably the closest environment to the nuclease-treatment dependent (claim 9). Not a single-reference anticipation of claims 1/12/20.

16. Ferrari, F.K., Samulski, T., Shenk, T., Samulski, R.J. (1997) — "Second-strand synthesis is a rate-limiting step for efficient transduction by recombinant adeno-associated virus vectors," Nature Medicine (1997).

• Description: Biology of rAAV transduction (second-strand DNA synthesis), not production or purification.
• § 102(b) art by date, but of low substantive relevance — it discloses nothing about lysate preparation, ion exchange, heparin affinity, or the sarcosine wash. No anticipation of any claim in the claim set; appears to be a general AAV-background citation.
• (Publication volume/page and exact authors taken from field knowledge; not independently re-verified this session.)

17. Qu et al. (2000) — "ACS abstract." Cited in the earlier section as an ACS-meeting abstract (2000) by (presumably) inventor Quang Qu / "Qu, G." on AAV column purification.

• § 102 status is date-critical and currently unresolved. If the abstract predates 2000-08-07 (e.g., a spring 2000 ACS national meeting), it is § 102(a) art; if later, it is not prior art. Note the risk of conflation with Gao, G., Qu, G., et al., Hum. Gene Ther. 11:2079–2091 (2000) ("Purification of recombinant adeno-associated virus vectors by column chromatography and its performance in vivo"), which on its late-2000 issue date would be post-filing and non-citing art. These must not be merged — the same "Qu" surname appears in both, and one is an abstract while the other is a full paper. (Flagged as unverified.)

---

C. Which claims are potentially anticipated under § 102

Bottom line: no cited reference anticipates any independent claim (1, 12, or 20).

The reason is structural, not rhetorical:

The load-bearing point: the sarcosine-wash limitation in claims 1(h) and 12(g) is not disclosed by any verified cited reference. A single reference that omits one required element cannot anticipate. That limitation alone defeats § 102 against claims 1 and 12 on the cited art.

Where the cited art is genuinely relevant is § 103, in clean two-cluster combinations:

• Cluster A — helper-virus-free production: US 6,004,797 + US 6,001,650 + Xiao 1998 + Grimm 1998 → maps to claim 1 steps (a)–(c), claim 11 (triple transfection), claim 10 (bioreactor), claim 8 (serum-free media via WO 00/22152).
• Cluster B — chromatography purification: Clark 1999 (heparin affinity) + Anderson 2000 (affinity + protease) + Grimm 1998 (anti-AAV antibody affinity) + whichever of WO 99/11764 / WO 99/61643 / WO 96/27677 / WO 97/08298 / WO 00/23116 discloses ion-exchange AAV purification → maps to claims 1(e)–(g), 12(c)–(f), 20, claims 4/5 (AAV receptor / anti-AAV antibody / heparin), claim 6 (microfluidization), claim 7 (filtration), claim 9 (nuclease).

The sarcosine wash (claims 1/12) and the specific AEX-then-affinity ordered train (claims 1/20) are the elements a § 103 challenge would have to supply from outside this cited set (or from the patent's own admissions about prior "many-step" processes). This is consistent with the earlier sections' conclusion that the references are § 103/background art, not true anticipatory art.

Dependents — same analysis, not independent: claims 2–11 and 13–21 each incorporate all limitations of their independent claim. Even where a cited reference discloses the added feature (e.g., Grimm's A20 antibody for claims 4/5; Clark's heparin for claims 4/5; Anderson's protease for claim 9), the dependent claim still cannot be anticipated absent the independent claim's full limitations — which no cited reference meets.

---

D. Verification gaps I am flagging rather than papering over

1. The patent's actual "References Cited" block was not retrievable this session — the WO citation list was inherited. Pull the front page from Google Patents US6593123B1 or Patent Center and diff it against the list above.
2. WO 96/27677, WO 97/08298, WO 99/11764, WO 99/61643, WO 00/23116 — subjects/titles not independently verified. My substantive mapping for these is provisional; do not treat as anticipatory art until each document is pulled and its exact teaching confirmed.
3. The "Qu et al. 2000 ACS abstract" — venue/date/identity unresolved, and it must not be conflated with Gao, Qu et al., Hum. Gene Ther. 11:2079–2091 (2000), which is likely post-filing and therefore not prior art.
4. U.S. patent grant dates/titles (5,139,941; 5,173,414; 5,225,347; 6,004,797) — reported as the '123 specification characterizes them; not re-verified this session.
5. Claim 20 and claim 21 full text remain truncated in the available sources (per the earlier section). The claim-20 mapping above assumes the earlier section's partial recitation (f) AEX step → (g) affinity step; confirm against the issued claims.

Net answer to the task: For US 6,593,123 specifically (not a look-alike number), the most relevant cited prior art is the purification cluster — Clark 1999 (heparin affinity), Anderson 2000 (affinity + protease), Grimm 1998 (antibody affinity) — and the helper-virus-free production cluster — Xiao 1998, Grimm 1998, and U.S. 6,004,797 / 6,001,650. None constitutes a § 102 single-reference anticipation of claims 1, 12, or 20, chiefly because the sarcosine-wash and ordered AEX-plus-affinity limitations appear in no cited reference; they are properly § 103 combination art.

I'll verify the prior-art references of record before mapping them onto the claims.

Good — I now have the verified claim 1 text and several reference identities. Let me ground the remaining references.

Obviousness Analysis — U.S. Patent No. 6,593,123 B1 under 35 U.S.C. § 103

Builds on / does not repeat: the bibliographic data, claim summary, litigation and PTAB null-results, and assignment chain already generated. This section assumes those and adds only the § 103 analysis.

Governing framework. '123 was filed 2000-08-07 and has no earlier priority claim on its face, so the critical date is 2000-08-07 and pre-AIA § 103(a) applies (not AIA § 103). The analysis therefore uses the Graham v. John Deere factors — scope/content of the prior art, differences from the claims, PHOSITA level, secondary considerations — with KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) supplying the permissible rationales (TSM is one rationale among several; "obvious to try" over a finite set of identified solutions is another).

---

1. Two threshold corrections that shape everything below

(a) The "Prior art keywords" on the Google Patents page are not a record of examiner-cited art. The terms raav, aav, chromatography medium, lysate, exchange chromatography are algorithmic page metadata. They are not evidence of what was cited, considered, or overcome during prosecution. Anyone using that block as a proxy for the art of record is mis-reading the page. The actual art of record is the front-page reference list (below).

(b) A large share of the art asserted against later Genzyme/Avigen patents is date-barred against '123. The IPR materials I surfaced (e.g., the 2025 petition on U.S. 7,704,721 discussing Auricchio, Konz, and Wu) are irrelevant to '123 as § 102 art: Konz (WO 2003/…, published 2003-11-27) and Auricchio (2001) both post-date 2000-08-07. They cannot be § 102(a) or § 102(b) art against '123, and would be usable only in a § 103 combination if a POSA would have had them — which fails as a matter of the critical date. Do not import that art into an '123 obviousness theory.

---

2. The prior art of record — verification status and what it supplies

I state plainly which items I could independently verify this session and which I am taking only from '123's own face and background. Content I could not verify is flagged and no teaching is attributed to it.

A note on which reference is "closest prior art." For claims 1 and 20, WO 97/08298 (or the combination its own ISR shows the art already made with WO 99/11764) is the closest art on purification; Xiao/Ferrari/'797 are the closest art on production.

---

3. The patentee's own admissions — the strongest § 103 foundation

Before reaching the references, note that '123's Background contains applicant admissions that do most of the work. Statements in a specification characterizing the prior art are usable as admissions (MPEP 2129; In re Font). '123 admits:

1. rAAV had been purified with a cation-exchange column alone: "A method of purifying rAAV using only cationic column-chromatography has been described. However, the method fails to remove enough contaminating DNA and protein to be suitable for commercial use." → Claim 12's cation-exchange step is admitted art.
2. rAAV had been purified by anion or cation exchange in combination with affinity purification: "A second chromatographic technique designed to purify rAAV… involves several steps, including chromatography over anion or cation exchange columns as well as purification through tangential flow filtration or affinity purification." → Claim 1's AEX→affinity architecture and claim 20's AEX→affinity architecture are admitted art.
3. CsCl was the clinical standard but not scalable; the art's stated deficiency was that these processes were built to remove infectious helper virus and therefore needed heat inactivation (≈50% titer loss).
4. The order of the columns is immaterial — by the patentee's own words: "It should be noted that the order of the chromatographic media is not considered to be important to the ultimate purification of the rAAV particles." → fatal to claims 2, 3 and 13–17 (order-of-steps claims) absent evidence of a critical, unexpected order-dependent result.
5. Every claim-6-to-10 unit operation is described in '123 as conventional: freeze/thaw, sonication and osmotic shock are "suitable cell lysis protocols"; "several clarification techniques are known"; buffers "are well known in the art"; "the selection of growth medium varies."

---

4. Claim 1 — element-by-element obviousness map

Claim 1 (text verified verbatim this session from the granted patent) has nine steps. Element by element:

Conclusion for claim 1. Seven of nine steps are admitted or squarely disclosed. The claim survives § 103 only if (f) (AEX flow-through mode) and/or (h) (sarcosine wash) are non-obvious. Note that (f) is a mode/order limitation that the applicant's own specification calls immaterial as to order, and that (h) was reached in '123 by screening exactly three candidate washes (6 M urea, 0.5 % deoxycholate, 15 mM sarcosine) — see § 9.

---

5. Claim 12 — obviousness map

Claim 12 assumes a host cell comprising rAAV virions (no production recitation) and requires: lysate → cation-exchange bind-elute (rAAV binds, wash, collect eluate) → affinity → sarcosine anionic-detergent wash → elute.

• Cation-exchange capture of rAAV is admitted art (Background: cationic-column-only purification "has been described"). Dependent claim 14's recited resins (sulfo-, phospho-, carboxy-, carboxy-methyl-based) are a generic recitation of the four standard CEX chemistries, and '123 itself designates a commercial sulfo membrane (MUSTANG S™) — i.e., the resin selection is a design choice within a known genus.
• Appending a heparin affinity step to a CEX step is taught/obvious over WO 97/08298, WO 99/11764 (combination of ion-exchange chromatographies with an AEX/affinity follow-on), and Clark 1999, with the scale-up motivation supplied by WO 97/08298's express criticism of density-gradient purification.
• GAP #2 (sarcosine) is the only substantive difference for claim 12. My honest assessment: claim 12 is narrower in number of unconventional limitations than claim 1, and therefore more exposed to § 103.

Dependent claims 13–17 (order permutations of the claim-12 steps). Anticipated obviousness problem: the specification says column order "is not considered to be important." A mere order change with no different result is not patentable subject matter for a process claim (In re Vaidyanathan; MPEP 2144.04 — "changing the order of steps" where no new or unexpected result). Expect these to fall first.

---

6. Claim 20 — flagged as incomplete

'123's independent claim 20 is an AEX/affinity process, but the full text of steps (a)–(e) and the closing steps was truncated in every source I could retrieve, and the continuation US 10/440,704 (US 2003/0207439 A1) has a different claim set that must not be substituted. What is visible ("(f) passing said eluate over an anion exchange chromatography medium, thereby binding said contaminants…"; "(g) passing the eluate obtained in step (f) over an affinity chromatography medium…") indicates a multi-column train with AEX→affinity as the last two stages — consistent with '123's field-of-invention statement that a cation-exchange step is optionally included. I do not attribute a complete claim scope to claim 20, and no definitive § 103 conclusion is offered for it. Retrieve claim 20 verbatim before relying on anything here.

---

7. Dependent claims — each is weak in a different way

---

8. The specific combinations and the stated motivation to combine

Under KSR, each combination below is stated with its rationale. All are premised on a PHOSITA being an experienced AAV gene-therapy process-development scientist (M.S./Ph.D. + several years in viral-vector manufacturing) as of August 2000.

Combination A — primary attack on claims 1, 11, 20 (and 2–10, 18, 19)

Xiao et al. (1998) + U.S. 6,004,797 + WO 97/08298 + WO 99/11764.

• What each contributes: Xiao and '797 supply the helper-virus-free production steps (a)–(c) and the triple-transfection step (claim 11). WO 97/08298 supplies the recognition that density-gradient purification cannot be scaled and that ion-exchange and affinity chromatography are the scalable substitutes for AAV. WO 99/11764 supplies the combination of ion-exchange chromatography steps performed on rAAV from cell lysates/supernatants in a helper-free preparation.
• Motivation, stated: (1) Known problem — WO 97/08298 expressly frames density-gradient centrifugation as incompatible with large-scale manufacture, and '123's own background concedes CsCl is unsuitable for commercial scale; (2) Design incentive flowing from Xiao/Ferrari — once rAAV is made without infectious adenovirus, the entire rationale for the prior art's heat-inactivation and helper-removal steps disappears, so a POSA is affirmatively motivated to substitute the shorter, scalable column train; (3) Orthogonal-mode logic — ion exchange (charge) and affinity/heparin (receptor-ligand) resolve different contaminants, so serially combining them is the textbook bioprocess approach and was already being taught for AAV; (4) Reasonable expectation of success — WO 97/08298 reports viable recovery of AAV from column chromatography, and WO 99/11764 reports high-titer helper-free preparations from such steps.
• Not in Combination A: the AEX flow-through mode and the sarcosine wash.

Combination B — attack on claims 1, 20 and 3–5

WO 97/08298 (O'Riordan/Genzyme) + the receptor literature cited on the '123 face (Summerford & Samulski 1998; Summerford 1999; Qing 1999) + the '123 Background's admitted AEX-and-affinity art.

• Motivation: WO 97/08298 teaches affinity chromatography as a mode; the applicant's own cited references identify heparin sulfate proteoglycan (and integrin/FGFR) as an AAV receptor. Selecting heparin — the commercially available, chemically stable analogue of that receptor — as the affinity ligand is the predictable use of a known binding pair for its known purpose, i.e., the essence of affinity chromatography. This is a § 103 attack built substantially out of the applicant's own prior-art citations, which makes it rhetorically and legally robust.

Combination C — attack on claim 12 (and 13–17)

The '123 Background's admitted "cationic column chromatography only" rAAV purification + WO 97/08298 / WO 99/11764 (affinity/ion-exchange combination) + [sarcosine-closing reference, § 9].

• Motivation: the admitted art already used a CEX column to capture rAAV but "fails to remove enough contaminating DNA and protein to be suitable for commercial use." That is an express statement of a problem in the art that the art itself identified — the strongest possible motivation to add a second (affinity) dimension. Adding a heparin affinity capture after a CEX capture is a combination of known elements with predictable results (KSR).

Combination D — the unit-operation dependent claims (6–10)

'123's own admissive text + WO 00/22152 (culture/media/serum) + WO 99/61643 (nuclease). Each element is either admitted conventional ('123 says so in terms) or supplied by an incorporated/cited reference. Where a claim element is a routine unit operation, a single reference plus Official Notice ("it is well known that…") is sufficient; the claims are written at a level of generality (filtration, bioreactor, nuclease) that invites this.

---

9. The weak links — and how they get closed (or don't)

GAP #2 — the sarcosine wash — is the only limitation I cannot close on the record in front of me, and it is present in both independent claims 1 and 12. Honest framing:

1. The claim is narrow but the selection set is finite. '123 tested exactly three wash additives — 6 M urea, 0.5 % deoxycholate, 15 mM sarcosine — and picked the best of three by silver-stained SDS-PAGE (FIG. 2(a)). Under KSR, "a finite number of identified, predictable solutions" supports obviousness: anionic detergents were the recognized class for stripping contaminating protein from affinity matrices, and sarcosine/sarkosyl was a known member of that class. The patent's own data frame the choice as an optimization result, not a discovery.
2. A closing reference must be located. A § 103 case on claims 1/12 needs art teaching or suggesting an anionic-detergent (ideally sarcosine/sarkosyl) wash of an AAV-binding column. I did not verify such a teaching in any of the six references of record this session. Free-text "detergent wash" art is common in the affinity-chromatography and virus-purification literature; this is the specific gap a searcher should close.
3. Self-collision caveat. '123's inventors are the named inventors on the WO 99/61643 family. If that document (published 1999-12, within one year of the 2000-08-07 filing) contains the sarcosine-wash example, using it against '123 raises two problems: (i) it is the inventors' own work, so the "by others" precondition of pre-AIA § 102(a)'s knowledge/use prong and its use as § 103 art is contestable; and (ii) it is inside the § 102(b) grace period. The same caution applies to Qu et al. (ACS abstract 2000) — the inventors' own presentation within the grace year, which cannot be § 102(b) art and is not § 102(a) art "by others." Both should be treated as research leads, not asserted art.

GAP #1 — AEX flow-through. '123 presents the flow-through behavior as its own observation ("rAAV flows through the column while much of the contaminating protein remains bound… at 7.0–8.5 pH and under 200 mM NaCl"). No verified reference teaches it. Note, however, that this is a mode-of-operation limitation, and the specification elsewhere treats the AEX/affinity order as immaterial to the outcome — an internal inconsistency that a challenger would exploit to argue the mode is an obvious, result-neutral choice.

Antedating risk on the two April-2000 PCTs. WO 00/22152 and WO 00/23116 published only months before the 2000-08-07 filing. As § 102(a) art they are subject to being sworn behind under 37 CFR 1.131 with proof of prior invention (highly available here, since the application was filed in the same year as the work). Any theory built on those two documents invites an antedating fight and should carry that caveat.

§ 103(c) consideration. To the extent any WO 00/22152 / WO 99/61643 / Qu et al. material is Avigen's own commonly-owned work qualifying as § 102(e)/(f)/(g) art, pre-AIA § 103(c) disqualifies it from use in a § 103 combination. Verify the common-ownership chain (the assignment section already traced Avigen's ownership at filing).

---

10. What the patentee would say in rebuttal

Fairness requires stating the patentee's side:

1. Long-felt, recognized need with a concrete failure of the art. '123's background documents a genuine, articulated problem: prior processes existed to remove infectious helper virus and therefore required heat inactivation at a cost of ~50 % of functional titer; CsCl was not commercial-scale; CEX-only "fails to remove enough contaminating DNA and protein." That is a classic Graham secondary-considerations setup.
2. Objective evidence of performance. The specification reports up to 100 % recovery through clarification, >1 000-fold purification on MUSTANG S™ with >80 % recovery, >90 % recovery across AEX, and >80 % overall rAAV recovery by Q-PCR across the AEX + heparin process (FIG. 14). If framed as unexpectedly high relative to the closest art, this is the patentee's best § 103 rebuttal — but the nexus must be to the claimed features (the flow-through mode and the sarcosine wash), not merely to the overall train.
3. Teaching away? The art taught bind-elute chromatography for viruses and taught that heat inactivation was necessary because helper virus was present. '123 would argue the art pointed toward bind-elute, not flow-through, and toward decontaminating a helper-virus-containing preparation. This is the most substantive rebuttal, but it is weakened by the applicant's own concession that column order is immaterial.
4. Commercial success / nexus. Avigen used the process to make clinical-grade vector (the AV201 program), and the '123 process was later cited and incorporated as background by subsequent AAV manufacturing families (EP 1,951,888; EP 2,781,596; EP 2,529,020). Whether that rises to nexus-qualified commercial success is unproven on this record.

---

11. Bottom-line assessment

Where I land: on the record available to me, the strongest defensible § 103 position is a two-part attack — Combination C on claim 12 and Combination A/B on claim 1 — with the sarcosine wash as the single load-bearing limitation. I could not, this session, close that limitation with a verified reference. That is a gap in my search coverage, not a finding that the art is absent. Until it is closed, I would not represent that claims 1 or 12 are invalid as obvious; I would represent that the dependent claims and the production steps are highly vulnerable and that the independent claims' validity narrows to one claimed wash reagent.

---

12. Cross-reference against the previously generated sections

• No contradiction found. The claim-1 text quoted here (verified verbatim from the granted patent this session) matches the claim summary already generated, including the sarcosine wash at step (h) and the microfluidization/serum-free/nuclease/bioreactor/triple-transfection dependents.
• One addition, not a conflict: the earlier PTAB section reported no proceedings on '123. The only AAV IPR activity surfaced in this pass is on U.S. 7,704,721 (petition materials citing Auricchio, Konz, Wu). That is a different patent, and its art post-dates '123's critical date — it is of no use against '123. Consistent with the earlier "cited literature, not asserted patent" characterization.
• One item re-flagged: the earlier section noted the Qu et al. (ACS abstract 2000) reference. For § 103 purposes it is the inventors' own work and is not available as § 102(b) art — a point worth carrying into any opinion so nobody reaches for it as invalidating art.
• Continuation caution carried forward: do not back-fill claim 20 from US 2003/0207439 A1. Its claim set differs.

13. Verification checklist before this analysis is used externally

1. Retrieve claim 20 and claim 21 verbatim from the granted patent PDF (patentimages / Patent Center) — my claim-20 analysis is suspended.
2. Pull the full texts of WO 97/08298, WO 99/11764, WO 96/27677, WO 00/22152, WO 00/23116, Grimm 1998, Clark 1999, Anderson 2000, and the Qu et al. abstract, and confirm their publication dates against the 2000-08-07 critical date.
3. Search specifically for sarcosine / N-lauroylsarcosine (sarkosyl) washes of AAV- or virus-binding chromatography media in 1995–2000 literature — this is the decisive missing reference.
4. Check WO 99/61643's examples for a sarcosine wash; if present, run the § 102(a) "by others" and § 103(c) common-ownership analysis before using it.
5. Confirm the absence of an earlier priority document (provisional) for 09/633,834 in Patent Center — this sets the critical date.
6. Confirm no ex parte reexamination certificate alters the claims, per the flag already raised in the PTAB section.

Sources for verified statements: US 6,593,123 granted text and front page (https://patents.google.com/patent/[US6593123B1](/patent/US6593123B1)/en; https://patentimages.storage.googleapis.com/1a/e3/cc/37c2eed4c99505/US6593123.pdf; https://patents.justia.com/patent/6593123); US 6,143,548 / WO 97/08298 (https://patentimages.storage.googleapis.com/2e/f2/d4/16d51a856c3467/US6143548.pdf); WO 99/11764 international search report (https://patentimages.storage.googleapis.com/27/10/ef/b09eb1a59f04cc/WO1999011764A3.pdf); WO 99/61643 abstract as reproduced in PTAB records (https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1558103](/patent/1558103)). Reference contents I could not retrieve are flagged inline and no teaching is attributed to them.