US 9725502

I'll research this patent across the sources you specified. Let me run several targeted searches.

Let me check specifically for any court proceedings involving this patent, since those searches were inconclusive.

US Patent 9,725,502 — Analyst Summary

Scope note / confidence statement: I retrieved the authoritative patent text (Google Patents, fetched for this record) plus secondary database entries. I was not able to retrieve the actual granted claim set from USPTO PatentCenter, and no CAFC 2026 docket activity was returned for this number. Details below are flagged accordingly. Per your instruction, I have not auto-corrected any identifiers.

1. Bibliographic data

Legal-status discrepancy (fidelity note): The authoritative record supplied states "Expired – Fee Related." However, one Google Patents search snippet returned for this same number showed "Active." These conflict. I have not resolved it and flag it rather than picking one.

This is a continuation in a large IRB family. Related US members include 8,124,093; 8,287,870; 8,298,539; 8,435,524; 8,603,480; 8,765,132; 9,127,049; 9,149,524; 9,217,028; 9,221,897; 9,249,213; 9,365,636; 9,371,372; 9,491,906; 9,527,902; 9,611,316; 9,725,502; 9,796,771; 9,796,772; 9,803,000; 10,040,845; 10,414,817; 10,889,632. Post-grant filings linked from this patent include US 15/637,513, US 15/716,818, US 16/025,296, US 16/527,836 and US 17/011,934 (US 2021/0087257 A1). I list these only to delimit the family; they are not results for 9,725,502 itself.

2. Abstract

The supplied text is the patent's descriptive/definitions portion and does not include the printed abstract. The family abstract (identical title, same inventors/assignee) reads:

"The invention relates to neutralizing antibodies, and antibody fragments thereof, having high potency in neutralizing hCMV, wherein said antibodies and antibody fragments are specific for one, or a combination of two or more, hCMV gene UL products. The invention also relates to immortalized B cells that produce, and to epitopes that bind to, such antibodies and antibody fragments. In addition, the invention relates to the use of the antibodies, antibody fragments, and epitopes in screening methods as well as in the diagnosis, prevention, and therapy of disease."

Caveat: this abstract text was retrieved from a secondary database entry keyed to sibling US 8,603,480, not read off the 9,725,502 front page. Treat as very likely identical but not front-page-confirmed.

3. Plain-language overview of the independent claims

Important caveat: The authoritative text I have is the specification/definitions section, which recites the invention's aspects ("the invention comprises…"). I could not pull the literal granted claim set, so the following paraphrases the independent-claim subject matter as disclosed. Where the searched Google Patents term-index showed claim language ("injection," "infusion," "preservative," "epithelial cell"), I note it as claim-language evidence, but treat all of the below as a specification-grounded reconstruction, not a verbatim claim chart.

Core subject matter groups (each typically a separate independent claim):

1. Anti-UL128 neutralizing antibody. A monoclonal/human monoclonal antibody, or antigen-binding fragment, that binds an epitope in the hCMV UL128 protein and neutralizes hCMV infection (potency recited as IC₉₀ < about 2 µg/mL, with narrower ranges down to ≤ 0.0005 µg/mL).

2. Complex/multi-protein epitope antibodies. Antibodies binding an epitope formed by:

   • gH, gL, UL128 and UL130; or
   • UL128, UL130 and UL131A; or
   • UL130 and UL131A;

   each neutralizing hCMV. The specification explains "formed by" allows the epitope to sit on one subunit while the others are required for binding.

3. Additional target classes: antibodies binding epitopes in gB, in gH, or formed by gM and gN, all with the same potency limitations.

4. CDR-defined antibodies. An antibody comprising at least one CDR having ≥95% sequence identity to a recited set of SEQ ID NOs (e.g., 188–193, 204, 205, 210, 174–177, 149, 178, 65–70, 81–86, 97–102, 129–134, 145–150, 113, 161–164, 1–6, 17–22, 33–38, 49–54, 114–118; and separately 216–221, 232–235, 149, 236, 246–251, 278–283, 296–301, 312, 316–321, 332, 336–341, 352, 360, 361, 262–267), wherein the antibody neutralizes hCMV.

5. Consolidated CDR-set antibodies. Claims reciting a heavy-chain CDR1/CDR2/CDR3 selected from enumerated SEQ ID NO groups, and light-chain CDR1/CDR2/CDR3 from enumerated groups.

6. VH/VL pair antibodies. Antibodies defined by specific heavy-chain/light-chain variable-region SEQ ID NO pairings (e.g., VH 200 + VL 201; VH 200 + VL 213; VH 208 + VL 201/213; VH 212 + VL 201; and analogous pairs such as VH 228/VL 229, VH 242/VL 243, VH 258/VL 259, VH 290/VL 291, VH 294/VL 291).

7. Potency-limited antibodies (functional claims). An antibody that neutralizes infection of endothelial, epithelial, retinal, myeloid, dendritic, fibroblast or mesenchymal stromal cells by a clinical isolate of hCMV where the IC₉₀ is 1.2 µg/mL or less; and a parallel claim at 10 µg/mL or less, expressly excluding MSL-109 and 8F9.

8. Deposit-based claims. Antibodies/antigen-binding fragments produced by the deposited immortalized B-cell clones (8I21, 2C12, 8C15, 4N10, 11B12, 3G16, 4H9, 6B4, 10C6, 6L3, deposited with the Advanced Biotechnology Center (ABC), Genoa, under the Budapest Treaty on Jul. 9, 2008 under accession numbers PD 08005, PD 08007, PD 08006, PD 08009, PD 08011, PD 08012, PD 08013, PD 08004, PD 08014, PD 08010, respectively; and 7H3 deposited Jul. 16, 2008 under PD 08017), plus antibodies having the same amino acid sequences.

9. Nucleic acid claims. A nucleic acid molecule comprising a polynucleotide encoding an hCMV-neutralizing antibody/fragment of the invention.

10. Cell claims. A cell expressing an antibody of the invention.

11. Immunogenic polypeptide claims. An isolated/purified immunogenic polypeptide comprising an epitope that binds an antibody of the invention.

12. Pharmaceutical composition claims. A composition comprising an antibody/fragment, a nucleic acid, or an immunogenic polypeptide of the invention plus a pharmaceutically-acceptable diluent or carrier — the searched claim-term index indicates claim language covering injection and infusion administration and a preservative (consistent with a formulation independent claim).

13. Combination composition claims. A composition comprising a first antibody of the invention and a second antibody that neutralizes hCMV (defined by differing target–epitope pairings across UL128, UL130/UL131A, UL128/UL130/UL131A, gH/gL/UL128/UL130, gB, gH, gM/gN, gL, gO).

14. Method/use claims. Use of the antibody, fragment, nucleic acid, immunogenic polypeptide, or composition (i) in manufacture of a medicament for treating hCMV infection, (ii) in a vaccine, or (iii) in diagnosis of hCMV infection; use for monitoring anti-hCMV vaccine quality by confirming the correct-conformation epitope is present; and epitope-per-se claims for therapy, medicament manufacture, vaccine use, and screening for neutralizing ligands.

Representative disclosed antibodies include 15D8 (and variants 1 and 2), 4N10, 10F7, 10P3, 4I22, 8L13, 2C12, 8C15, 916, 7B13, 8J16, 8I21, 7I13, 7H3 (and variant 1), 6B4, 5F1, 10C6, 4H9 (and variant 1), 2B11, 11B12, 13H11, 3G16 and 6L3. The specification also expressly disclaims the antibody being MSL-109, 8F9, 3E3 or R551A, and disclaims 1F11, 2F4, 5A2 or 6G4 from U.S. Ser. Nos. 11/969,104 and 12/174,568.

4. Litigation / CAFC 2026 docket status

• No CAFC 2026 docket activity was found for US 9,725,502. My searches for "9725502" + CAFC/docket/litigation and for the patent number in court contexts returned zero relevant results.
• The only litigation signal is a hyperlink on the Google Patents page to a Darts-IP record: "First worldwide family litigation filed — family=41550776." That link is a pointer, not a case listing; I could not open it and cannot confirm what case(s), parties, venue, or dates it refers to. Do not treat this as a confirmed litigated patent.
• Explicit uncertainty: I cannot state whether US 9,725,502 has been asserted, IPR'd, or appealed to the CAFC in 2026. Absence of search hits is not evidence of absence, especially for very recent dockets.
• Separately, the "when will biosimilars launch" database associates the family with biologic SYLVANT (siltuximab, BLA 125496, Eusa Pharma) under sibling patent 8,603,480 — but that mapping looks like a database artifact and I would not rely on it without verification.

5. Bottom line

US 9,725,502 (IRB; Lanzavecchia & Macagno) claims a genus of high-potency human anti-hCMV neutralizing antibodies and fragments, defined by target (UL128 alone, or complexes UL130/UL131A, UL128/UL130/UL131A, gH/gL/UL128/UL130, plus gB, gH, gM/gN), by CDR/VH-VL sequences, by IC₉₀ potency thresholds (including ≤1.2 µg/mL and ≤10 µg/mL with MSL-109/8F9 carved out), and by deposited B-cell clones; with downstream claims to nucleic acids, expressing cells, epitope polypeptides, pharmaceutical and multi-antibody compositions, and therapeutic/diagnostic uses. It is a 2008-priority continuation granted 2017-08-08, with anticipated expiration 2029-07-15.

Main gaps I could not close: the verbatim granted claim set (especially exact independent-claim wording and count), the front-page abstract as printed, and any court/PTAB/CAFC proceedings. If you need claim-exact language or docket confirmation, pull the USPTO PatentCenter "Claims" tab for application 15/284,799 and the Darts-IP family 41550776 record directly — I can then map claim-by-claim.

I'll search for litigation involving US patent 9725502.

Litigation Search Report — U.S. Patent No. 9,725,502

Bottom line: I found no verifiable litigation specifically asserting or involving U.S. Patent No. 9,725,502. I am reporting that as a negative result rather than supplying case numbers I could not confirm.

1. What I searched and what I found

Per the operating rule, results for similar-but-different numbers (notably 7,238,550) are not attributable to 9,725,502 and are excluded from the findings below.

2. The one litigation-adjacent data point I can confirm

The Google Patents page for US9725502B2 carries a "Family has litigation" flag, sourced to Darts-IP, stating: "First worldwide family litigation filed." It links to a Darts-IP family record (family ID 41550776):

https://patents.darts-ip.com/?family=41550776

What this tells us: at least one member of the worldwide patent family (the PCT/EP national-phase family claiming priority to 2008‑07‑16, inventors Antonio Lanzavecchia and Annalisa Macagno) has been involved in some patent litigation somewhere in the world, at some time.

What it does not tell us: the flag does not identify the court, the parties, the case number, the filing date, the patent member asserted, or the outcome. Critically, the family covers roughly 137 members (e.g., EP 2487187, ES 2682596‑T3, NZ 591055, PH‑12015500782, plus numerous US grants listed as siblings: 8,124,093; 8,287,870; 8,298,539; 8,435,524; 8,495,282/…; 8,605,480; 8,765,132; 9,127,049; 9,221,897; 9,249,213; 9,371,372; 9,363,636; 9,491,906; 9,527,902; 9,795,771; 9,795,772; 9,803,000; 10,040,845; 10,414,817; 10,889,632). A family-level flag therefore cannot be attributed to the '502 patent itself, and the '502 patent could easily have issued long after the flagged proceeding.

I therefore cannot state, and do not represent, that U.S. 9,725,502 was ever asserted in litigation.

3. Requested fields (plaintiff, defendant, jurisdiction, case number, filing date, status)

I cannot populate this table for U.S. 9,725,502 without fabricating details:

4. Relevant US prosecution/post-grant facts (non-litigation, for context)

These are confirmed from the patent record and are useful when verifying whether any case exists:

• Patent: US 9,725,502 B2, "Human cytomegalovirus neutralizing antibodies and use thereof"
• Application: 15/284,799, filed 2016‑10‑04; granted 2017‑08‑08
• Priority: 2008‑07‑16 (earliest US filing in the chain: 12/174,568, 2008‑07‑16)
• Assignee: Institute for Research in Biomedicine (IRB), Switzerland
• Status on Google Patents: Expired – Fee Related; anticipated expiration listed as 2029‑07‑15
• Post‑grant: No IPR/PGR/CBM proceeding for this patent surfaced in my searches
• Note the family also includes a flagged US 15/637,513 → US 9,803,000 and continuations 15/716,818 → 10,040,845; 16/025,296 → 10,414,817; 16/527,836 → 10,896,32x — worth checking separately, since an assertion may have been brought on a sibling rather than on the '502.

5. Recommended verification steps (to close the gap)

If you need a definitive negative or positive, these sources are authoritative and were not fully reachable in this session:

1. PACER / CourtListener (RECAP) — search the patent number "9,725,502" and "9725502" in the "Patents" / docket-text fields; RECAP's free docket search will return a case number if any pleading names it.
2. Unified Patents litigation portal (https://portal.unifiedpatents.com/litigation/caselist) and Unified's Patent Litigation database — search by patent number.
3. Darts-IP family record 41550776 (subscription) — the source of the Google Patents flag; this is where the "first worldwide family litigation" would be itemized, and it would reveal whether the asserted member was the '502 or a sibling.
4. USPTO Patent Center / Assignment records for 15/284,799 — to confirm no reissue or supplemental examination (which can be litigation-driven).
5. CAFC docket search on "9,725,502" or on the assignee "Institute for Research in Biomedicine."

6. Caveats

• I did not confirm or deny the existence of foreign litigation (the Darts-IP flag is expressly "worldwide"), and this report covers only what I could verify for the US.
• Absence of a hit in the sources I reached is not proof of no litigation — it means I could not confirm any. I will not supply a case number, party, or date that I cannot ground in a cited source.
• A separate caution: the Helix Microinnovations cases returned in searches involve U.S. 7,238,550, a semiconductor packaging patent with entirely different parties. Those results are unrelated to 9,725,502 and are excluded.

I'll start by checking the structured data default and then searching for any PTAB activity the ODP may not have indexed.

Let me probe further for any proceedings in the family and any litigation-linked IPR.

Let me confirm via the Unified Patents portal and a targeted PTAB search.

Proceedings overview

No AIA trial proceedings are on file for US 9,725,502. The canonical USPTO Open Data Portal ("PTAB proceedings on file") block returns zero AIA trials — no IPR, PGR, or CBM — and my independent web searches (Google Patents family/legal-status pages, Unified Patents portal records for the IRB hCMV family, and general queries on the patent number and the "Institute for Research in Biomedicine" / Lanzavecchia-Macagno hCMV antibody family) surfaced no petition, institution decision, Final Written Decision, or Federal Circuit appeal naming this patent. Bottom line defensive posture: there is no PTAB kill shot to point to, and there is no IPR estoppel to inherit — but the patent itself is flagged by the source data with an "Expired - Fee Related" legal status, which (if accurate) is a far cheaper defense than any IPR.

⚠️ Caveat on completeness: The ODP block is the canonical list, and I found no contradicting web evidence; but I could not run a live PTAB E2E / PTAB Decisions docket query or a CourtListener docket pull in this session (search steps exhausted). Treat "zero proceedings" as high-confidence-but-not-certified, and confirm against PTAB E2E before filing anything that depends on it.

No proceedings to enumerate

There are no proceeding numbers to report. Per your constraint, I will not invent any (e.g., I will not manufacture an "IPR2019-XXXXX" for this patent). The list is empty.

What that means in each of your requested buckets:

Strategic summary

Claim status of 9,725,502. Because no AIA trial ever reached a Final Written Decision, no claim of 9,725,502 has been canceled, narrowed through IPR, or judicially construed in a PTAB record. Every claim — including the antibody/antigen-binding-fragment claims drawn to UL128, UL130/UL131A, UL128/UL130/UL131A, gH/gL/UL128/UL130, gB, gH, and gM/gN epitopes, and the deposited-clone/sequence-based claims — is untested at the PTAB. There is no certificate cancelling claims to hand a court, and no claim-level disposition to quote. The full text of the granted patent (as reproduced in the authoritative source) reflects the claims as issued.

The real defensive lever here is the patent's legal status, not estoppel. The Google Patents record for US 9,725,502 lists the current legal status as "Expired - Fee Related" with a listed anticipated expiration of 2029-07-15. "Expired - Fee Related" is the standard designation for lapse due to non-payment of a maintenance fee. That is a potential case-terminating fact for a defendant — an expired patent cannot be infringed going forward, and damages would be limited to pre-expiration accrual (at most). However, flag the internal inconsistency: the same record shows an anticipated expiration of 2029-07-15, which is inconsistent with an early fee lapse. This discrepancy may reflect a delisting/downgrade of a family member, a data artifact, or a corrected status. Verify directly in USPTO Patent Center / the fee-payment history before relying on it. If the fee lapse is real and unrevived, this discussion largely ends there.

Estoppel landscape. With zero IPRs, there is no § 325(e) / § 315(e)(2) estoppel on any party, and no petitioner privies to worry about. Conversely, there is also no estoppel protecting you — the patent owner is not boxed in by any adverse PTAB findings. If you are a defendant today, every prior-art ground remains available to you, including: (i) § 102/§ 103 art on the 2008-07-16 priority date; (ii) the sizable pre-2007 art already of record in the family (e.g., Wang & Shenk, PNAS 102:18153-58 (2005) and J. Virol. 79:10330-38 (2005); Adler et al. (2006); Hahn et al. (2004); Patrone et al. (2005); Gerna et al. (2005/2008); Shimamura et al. (2006); Schoppel et al. (1996); Ohlin et al. (1993)); and (iii) the family's own earlier filings, which generate § 102(b)/§ 103 self-collision and obviousness-type double-prosecution angles. Also note the numerous terminal disclaimers across the family (e.g., US 8,603,480 recites a terminal disclaimer), which can matter for the priority/obviousness-of-variant analysis. Because the family's earliest priority traces to GB 0700133.2, filed 2007-01-04, the 2008-07-16 priority date is itself open to challenge in litigation — a fertile ground no IPR has ever tested.

Pattern signals. There is no repeat petitioner (because there is no petitioner), no defensive aggregator (no Unified Patents / RPX filing surfaced for the IRB hCMV family), and no patent-owner appeal activity at the Federal Circuit that I could confirm. Notably, the family is flagged by Google Patents as having litigation (the Darts-ip "Family has litigation / First worldwide family litigation filed" indicator), i.e., this patent family has been asserted somewhere in the world, yet no party ever filed an AIA trial against 9,725,502 or its siblings in the records I could reach. That combination — litigation exposure without any US post-grant challenge — is the single most interesting signal in this file. Plausible explanations, in rough order of likelihood: (a) the family was litigated outside the US or settled early; (b) the patent expired/lapsed, mooting an IPR; or (c) the asserted member was a different family patent (e.g., US 9,796,771/9,796,772/9,803,000), not 9,725,502. I could not verify which, and you should not assume any of them.

Recommended next steps

1. Do not look for an FWD — there isn't one. If your demand-letter or pre-suit analysis cites a PTAB Final Written Decision on 9,725,502, that citation is wrong. There is no FWD, no institution decision, and no Federal Circuit docket to link. Per your instruction, I am not linking a nonexistent opinion.
2. Verify the "Expired - Fee Related" status first. Pull the maintenance-fee payment history and Patent Center status for 9,725,502. If the 11.5-year fee was missed and not petitioned for revival under 37 C.F.R. § 1.378, the patent has lapsed and infringement exposure is confined to the pre-lapse window. Reconcile this against the listed 2029-07-15 anticipated expiration. This is a five-minute check that can moot the entire invalidity analysis.
3. If the patent is live, treat the field as wide open. With no IPR and no estoppel, you can run a full § 102/§ 103 challenge — in the district court or by filing first at the PTAB. If you file an IPR, note the statutory clock: institution decision within 6 months of the petition's filing date (35 U.S.C. § 314(b)), and a Final Written Decision within 12 months of institution (§ 316(a)(11), extendable up to 6 months for good cause). If you are later sued, a parallel IPR can support a stay motion.
4. If no PTAB activity and no fee lapse — and the patent owner is actively asserting — the absence of any IPR is itself a signal worth interrogating in diligence: check whether prior defendants took licenses, whether the family dispute is foreign-only, and whether the asserted claims are the sequence/deposit claims (which are notoriously hard to invalidate with printed publication art, but correspondingly narrow).

Confidence statement: I have high confidence that the canonical ODP list is empty and high confidence that no IPR/PGR/CBM number for 9,725,502 appears in the searchable public web I could reach. I have medium-to-low confidence in the "Expired - Fee Related" status being a real, unrevived lapse, because of the internal conflict with the 2029-07-15 anticipated expiration figure — verify at the source before acting. I could not verify the litigation flag's parties, forum, or the identity of the asserted family member; I will not guess at them.

I'll research the assignment record for this patent across the specified sources.

Ownership-chain reconstruction — US 9,725,502 B2

Source-access caveat (read first). I queried for the Assignment Center record for patent 9,725,502 / application 15/284,799. I could not retrieve the recorded reel/frame numbers, execution dates, or the correspondent of record for this patent. Google Patents (the authoritative text supplied for this analysis) exposes only two legal-event lines with assignment content, and neither carries reel/frame data. Everything below is therefore grounded in (a) the Google Patents legal-events/reassignment block, (b) the patent text itself, and (c) SEC filings by Vir Biotechnology describing the IRB licence relationship. I have not fabricated any reel/frame numbers, execution dates or attorney names. Where a field is unknown, I say so.

Inventors

Unusual-pattern screen — departure/attrition: not present. Neither inventor left the IRB/Humabs orbit around filing. Lanzavecchia remained IRB Director through 2020 and Humabs CSO; Humabs was acquired by Vir Biotechnology in August 2017 and became Vir's Bellinzona subsidiary, with Lanzavecchia and the IRB antibody platform retained. There is no “all inventors out within 12 months → portfolio fire-sale” precursor here.

Original assignee

Institute for Research in Biomedicine (IRB) — listed on the issued patent as both original and current assignee (“Institute for Research in Biomedicine IRB”). IRB is a non-profit academic research institute in Bellinzona, Switzerland, affiliated with the Università della Svizzera italiana (USI), founded in 2000 and funded by the City of Bellinzona, Canton Ticino, the Swiss Confederation, plus competitive grants and donations (per the Italian-language IRB coverage, e.g. eticinforma.ch and mattinonline.ch).

• Primary line of business: basic biomedical research. IRB does not manufacture or sell antibody products.
• Did it ship a product embodying the claims? No. IRB monetised this family by exclusive licence, not by product sales: an original 2004 exclusive IRB licence plus a May 2008 Humabs–IRB exclusive licence, amended and restated 16 December 2011, covering hCMV antibodies among other assets (Vir Biotechnology 10-K disclosures). Humabs then sub-licensed the CMV antibody programme to MedImmune, LLC (AstraZeneca) on 20 March 2012. IRB reported roughly CHF 12 million in accumulated royalty income from the Vir/Humabs relationship (2022 press coverage).
• Current status: operating (academic institute). Not acquired, not dissolved, not in bankruptcy. No bankruptcy, reorganisation or wind-down of IRB appears in any source reviewed.
• Ownership nuance: the patent’s owner is IRB; the commercial exploiter is Humabs BioMed SA / Vir Biotechnology under licence. Those are two different legal relationships and only one of them is recorded at USPTO.

Assignment timeline

The Assignment Center record for this patent could not be retrieved, so reel/frame, execution date and correspondent are unknown. The single recorded conveyance visible in the available legal-events data is:

• Executed: not retrieved / recorded 2017-06-29 — Reel/frame: not retrieved
  • Conveyance: Assignment (“Assignment of Assignors’ Interest”)
  • Assignor: LANZAVECCHIA, ANTONIO; MACAGNO, ANNALISA (as individuals)
  • Assignee: INSTITUTE FOR RESEARCH IN BIOMEDICINE
  • Correspondent: not retrievable from the sources available to me — I cannot confirm the attorney/firm of record, so the repeat-correspondent signal cannot be evaluated.
  • Context: inventor-to-institution title perfection for the continuation application (15/284,799, filed 2016-10-04). Recorded ~9 months after filing and shortly before the 2017-08-08 grant. This is the routine “confirm the employer owns the continuation” filing, not an arm’s-length transfer of rights to a third party.
  • Name-form note (not a change-of-name record): the assignee is recorded as “INSTITUTE FOR RESEARCH IN BIOMEDICINE” while Google Patents’ assignee field reads “Institute for Research in Biomedicine IRB.” I have not found a separate Change of Name, Merger or Correction conveyance reconciling the forms — flagging rather than resolving.

No other conveyances are evidenced for this patent: no second assignment, no Security Agreement, no Merger, no Release, no recorded Licence, and no downstream assignment to any operating company, holding LLC, financing vehicle or aggregator. On the available record, IRB remains the owner of US 9,725,502.

Licensing is not assignment — and matters here. The commercially important IRB→Humabs and Humabs→MedImmune agreements are licences documented in SEC filings (Vir Biotechnology Forms 10-K, Exhibits 10.28–10.58), not as recorded assignments. Anyone reading only the Assignment Center record would wrongly conclude the CMV portfolio is inert; the exploitation runs through the unrecorded-here licence chain instead.

Timeline diagram

timeline
    title Ownership of US 9725502
    2008 : Priority filing by IRB Bellinzona
         : Humabs takes exclusive IRB licence
    2016 : Continuation filed by IRB
    2017 : Inventors assign interest to IRB
         : Patent granted to IRB
    2026 : IRB still recorded owner

NPE / troll-pattern signals

Verdict

Insufficient data — applying the rubric’s own definition (“no records, or only the original assignment”). The only recorded conveyance is the 2017-06-29 “Assignment of Assignors’ Interest” from Lanzavecchia and Macagno to INSTITUTE FOR RESEARCH IN BIOMEDICINE, and the original assignee remains the named owner. Substantively this is not an NPE chain — an academic non-profit that licenses (Humabs BioMed / Vir Biotechnology; sub-licence to MedImmune) rather than asserts — and not a defensive-aggregator chain either; the verdict is driven by the incompleteness of the assignment record rather than by any adverse evidence. The main reason I cannot promote this to a clean “non-asserting institution” call is the unretrieved reel/frame and correspondent, and the unresolved Darts-IP “first worldwide family litigation filed” pointer flagged in the earlier summary.

Verify directly: search patent number 9,725,502 (or application 15/284,799) at https://assignmentcenter.uspto.gov/ and https://assignment.uspto.gov/patent/index.html to pull the reel/frame, execution date and correspondent of record; then pull the Darts-IP family 41550776 record to confirm what, if anything, was filed.

Contradictions / corrections against the previously generated sections

1. Legal status (carried forward): the earlier summary flagged “Expired – Fee Related” (authoritative text) vs an “Active” search snippet. That conflict is still unresolved here and I do not resolve it — it matters for the ownership question only insofar as an expired patent cannot be asserted.
2. Litigation: my findings are consistent with the earlier §4 — no CAFC/PTAB/litigation record for 9,725,502 was returned. The Google Patents Darts-IP family 41550776 “first worldwide family litigation filed” hyperlink remains a pointer only; I could not confirm parties, venue or dates, and it attaches to the family, not necessarily to this patent.
3. Biologic mapping is still dubious: the earlier-noted association of this family with SYLVANT (siltuximab, BLA 125496, Eusa Pharma) via sibling 8,603,480 is almost certainly a database artifact — siltuximab is an anti-IL-6 antibody, whereas this family is hCMV-neutralising antibodies licensed to Humabs/Vir/MedImmune. Do not rely on it.
4. One genuinely useful family-level observation not in the earlier sections: at least one foreign family member appears indexed under applicant “Humabs LLC” (e.g. MX20151000542A, 2008 priority), while a later Mexican grant in the same family is indexed to “Inst For Res In Biomedicine” (MX343490B). That naming split is worth checking on the PCT/national-phase ownership record, but it is not a US assignment and I have not treated it as one.

I'll search for the specific cited references on US 9,725,502.

US 9,725,502 B2 — Prior-Art Citation Analysis

Analyst note on scope and confidence (read first): I attempted multiple targeted searches to pull the examiner/applicant "Patent Citations" and "Cited By" tables for US 9,725,502 from Google Patents/USPTO. Those tables did not surface in any search result I could retrieve. The Google Patents page text that returned is truncated and includes only the Info/Links/classification/term-index portions — not the citation list. I also hit a hard search-step limit before I could open a secondary source (FreePatentsOnline / PATENTSCOPE / PatentCenter) that reproduces the citation table.

I will therefore not invent citation numbers, dates, or examiner references. What follows is (a) an explicit statement of that gap, and (b) the prior art that is documented inside the authoritative patent text you supplied, mapped to claims with § 102/§ 103 cautions. Everything below is either quoted from the supplied text or explicitly flagged as unresolved.

1. Status of the requested citation table

Do not treat any specific reference below as a confirmed examiner citation unless it is quoted from the patent text. Two searches returned a Google Patents snippet for a different document (WO 1994/016730 A1) in which 9,725,502 appears in a family/citation-adjacent list — that is a cross-reference, not a citation by 9,725,502.

2. Prior art that IS documented within the patent text (best available grounding)

The authoritative text you supplied identifies prior art in three places: (i) the numbered bracketed references in the "Definitions"/background passages, (ii) express claim/specification disclaimers of named prior antibodies, and (iii) cross-cited U.S. patent applications by serial number. These are the only defensible "citations" I can attribute.

2.1 Expressly named/disclaimed prior-art antibodies

The specification states the antibody of the invention "is not MSL-109, 8F9, 3E3 or R551A" and further "is not 1F11, 2F4, 5A2 or 6G4, disclosed in U.S. application Ser. Nos. 11/969,104 and 12/174,568." It also states (in the potency claim aspect) an antibody with IC₉₀ ≤ 10 µg/mL "wherein the antibody is not MSL-109 or 8F9."

§ 102 caveat: None of these named antibodies is relied on as an anticipating reference against the granted claims, because the claims/description affirmatively exclude them. Their role is as the closest known art that the applicant distinguished. Whether any related genus (e.g., a published application disclosing these clones) could anticipate a claim would require the actual claim language, which I could not retrieve.

2.2 Cross-cited U.S. patent applications (document citations)

2.3 Non-patent literature cited in the specification (bracketed key)

The supplied text preserves a numbered reference key but truncates the full bibliographic entries. The visible anchors are:

• [1]–[8] — background on gH, gL, gO complex (fibroblast entry) and the gH/gL/UL128–UL131 complex (endothelial, epithelial, dendritic-cell entry) — i.e., the receptor-complex biology the claims build on.
• [7], [10], [11] — prior antibodies to gH, gB, UL128 and UL130 products with stated neutralizing activity; the text characterizes their potency as "modest" (0.5–20 µg/mL).
• [9] — hyperimmune globulins commercialized for hCMV prophylaxis.
• [12]–[14] — IMGT numbering system references (CDR definitional, not prior art on patentability).
• [15]–[33] — assay, labelling, conjugation, PEG/polymer, liposome and drug-delivery references (enablement-support).
• [34]–[42] — hybridoma/EBV-immortalization methodology and expression-host references (e.g., [38] PER.C6 (Crucell); [39] & [40] HKB-11 (Bayer); [41] & [42] myeloma cells).

I cannot give full citations (journal, volume, page, year) for these bracketed references because the entries themselves are not present in the text you supplied — only the bracketed in-text callouts are. Restating them with invented bibliographic details would violate the no-fabrication rule.

3. Claim-to-prior-art mapping (as far as the text permits)

4. What I could not establish (explicit gaps)

1. The actual examiner "References Cited" list for US 15/284,799 / 9,725,502 — not retrieved.
2. The "Cited By" list — not retrieved.
3. Full bibliographic details of bracketed NPL references [1]–[42] — truncated in the source text.
4. Whether any cited reference has been applied in a rejection (and under which statutory subsection) — no file-history (PAIR/PatentCenter) data retrieved; § 102 vs. § 103 characterization above is analytical inference, not file-history fact.
5. Any Darts-IP family-41550776 litigation record — link only; not opened.

5. Precise sources to close the gap

To get the verbatim citation table you asked for, one of these will reproduce it (I recommend the first):

• USPTO PatentCenter → Application 15/284,799 → "References Cited" / "Patent Citations" tab (authoritative examiner citations + NPL list).
• USPTO Patent Public Search (ppubs) → document US 9725502 B2.
• FreePatentsOnline and EPO PATENTSCOPE → "Cited Documents" view for this number.
• Darts-IP → family 41550776 (for the litigation pointer).

If you paste the References Cited table (or the full claim set) from any of those, I can convert it into a claim-by-claim § 102 anticipation chart with exact dates and statutory subsections. As it stands, I will not fabricate that chart, because the citation data was not retrievable in this session.

I'll ground the prior-art side with targeted searches before analyzing.

Let me ground the patent-art side and the earlier IRB applications the '502 itself cites.

Analytical note on inputs (read first)

Two gaps constrain this analysis, and I will not paper over them:

1. No "Prior Art" section was present in the material supplied for this page. The only prior-art metadata in the fetched record is Google Patents' machine-generated fields — Prior art keywords: antibody; seq; hcmv; antibodies; composition and Prior art date: 2008-07-16. There is no reference list, no citation table, and no examiner-cited art in what I received. Rather than invent a "Prior Art section," I reconstructed the landscape from three grounded sources: (a) the patent's own Background/Definitions admissions in the authoritative text; (b) the reference numerals the patent itself cites (the list of references [2]–[42] is not included in the fetched text — i.e., I can see "[7]", "[9]", "[10, 11]", "[36]", but not what they point to); and (c) independently verified literature retrieved for this analysis, with dates checked against the 2008-07-16 critical date.
2. The verbatim granted claim set is still missing (flagged in the earlier Patent Summary section). The § 103 analysis below is therefore organised by claim group, using the specification's own recitation of invention aspects as the best available proxy for claim scope, and using the claim-language fragments already surfaced (injection, infusion, preservative, epithelial cell).

Contradiction flagged: the patent's own background asserts that antibodies to UL128 and UL130 "show very low potency in neutralizing infection of endothelial cells [7]," while the verified Shenk/Wang art (US 7,704,510, discussed below) reports anti‑pUL128 and anti‑pUL130 antibodies that do neutralize ARPE‑19 and HUVEC but not fibroblasts. The patent's characterisation is about potency, not existence — but this distinction is the entire crux of the obviousness question and I flag it as a disputed characterisation rather than accept it.

Threshold legal question flagged: US 9,725,502 was filed 2016-10-04 but claims priority to 2008-07-16. Whether pre‑AIA § 103(a) or AIA § 103 governs depends on whether the application ever contained a claim with an effective filing date on or after 16 March 2013. I have not verified this from the file wrapper. The analysis below applies pre‑AIA § 103(a) (KSR as construed), which is the framing the specification's own "prior art date 2008-07-16" implies.

Obviousness Analysis — US 9,725,502 B2 under 35 U.S.C. § 103

1. Framework

Graham v. John Deere Co., 383 U.S. 1 (1966) requires: (1) scope and content of the prior art; (2) differences between the prior art and the claims; (3) level of ordinary skill; and (4) secondary considerations. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) supplies the operative test for the combination theories below: a claim is obvious where the prior art elements are known, the combination was "obvious to try," there was a design incentive or market pressure to combine, and the results were predictable. In re Kubin, 561 F.3d 1351 (Fed. Cir. 2009), applied KSR to biotechnology: where the antigen is isolated and a routine method of obtaining antibodies/binding molecules exists, the "obvious to try" doctrine bites.

(Case citations are given from memory and should be verified before use in any filing; I am deliberately not citing antibody-specific Board or Federal Circuit decisions whose holdings I cannot state with confidence.)

2. Level of ordinary skill in the art (PHOSITA)

As of 16 July 2008: a Ph.D. or M.D. with 2–5 years' experience in herpesvirology and antibody engineering, routinely able to (i) clone and co-express HCMV glycoprotein genes in HEK293T cells (the very assay the '502 uses in FIG. 1 and FIG. 3), (ii) immortalise human memory B cells and screen supernatants for neutralisation (the '502's own method, referencing its reference [36]), (iii) run fibroblast (MRC-5), epithelial (ARPE-19) and endothelial (HUVEC) microneutralisation assays, and (iv) sequence and reformat VH/VL into IgG. All four were standard by 2008 — several are disclosed in the patent itself as routine.

3. Scope and content of the prior art (verified)

The specification's own admissions as prior art

The '502's Background and Definitions are applicant admissions usable under § 103:

• "Neutralization by these antibodies was observed at antibody concentrations ranging from 0.5 to 20 µg/ml."
• "Known antibodies to UL128 and UL130 show very low potency in neutralizing infection of endothelial cells."
• "the neutralizing potency of the antibodies isolated so far is modest."

⚠️ The bracketed numeral [7] supporting the UL128/UL130 statement cannot be resolved, because the '502's reference list is absent from the fetched text. In similarly ordered lists in later IRB/Princeton-family documents, position [7] corresponds to Gerna et al. 2008 — but I will not assert that mapping. Reference numerals must be decoded from the printed patent or file wrapper before any of these admissions are relied on.

4. Differences between the prior art and the claims

Reading the claim groups against the table above, the differences reduce to four:

5. Combination theories

Combination A — the "core pentamer" combination (strongest for the target/genus claims)

A1 + A2 + A3 + A4 = Hahn 2004 + Wang & Shenk PNAS 2005 + Ryckman 2008 + Gerna 2008

Claim groups targeted: the anti‑UL128 antibody; the UL130/UL131A, UL128/UL130/UL131A and gH/gL/UL128/UL130 antibodies; and the functional IC₉₀ ≤ 10 µg/mL ("not MSL‑109 or 8F9") claim.

Why a POSHITA would combine them: These four do not merely coexist — they form an express, stepwise teaching:

1. Hahn 2004 and Wang & Shenk 2005 establish the locus as the tropism switch for the very cell types (endothelial, epithelial, dendritic) in which hCMV causes disease.
2. Wang & Shenk PNAS 2005 identifies the physical target — a discrete, isolable gH/gL/pUL128/pUL130 complex incorporated into virions.
3. Ryckman 2008 characterises that complex, supplying the immunogen and the binding assay.
4. Gerna 2008 (three months before the critical date) closes the loop by teaching that human serum neutralising activity against E/E cells is attributable to antibodies to the pUL131A/pUL130/pUL128 products, and that fibroblast assays under-report this activity. That is a near-verbatim instruction to make antibodies against those gene products and screen them on E/E cells.

Reasonable expectation of success: US 7,704,510 had already demonstrated, a year earlier, that anti‑pUL128 and anti‑pUL130 monoclonal antibodies neutralise ARPE‑19 and HUVEC and spare fibroblasts. So the outcome (a neutralising anti‑UL128/UL130 mAb) was not merely predictable — it was reported. Under KSR, the § 103 case for a genus claim covering "an antibody that binds an epitope in UL128 and neutralises hCMV" is strong, particularly where the claim is not tied to any structure.

Where this combination falls short — D3. Combination A supplies no motivation to expect sub-µg/mL IC₉₀. It would, if anything, suggest the opposite: 3E3 at ~20 µg/mL and the 0.5–20 µg/mL art range would lead a POSHITA toward antibody cocktails or glycoengineering rather than toward a single picomolar mAb. This is the heart of the applicant's case, and it is set out in § 7 below.

Combination B — Princeton art + routine mAb generation (strongest for the sequence-independent claims)

B1 + B2 + B3 = US 7,704,510 + Baldanti 2006 + the EBV-immortalisation/plasma-cell method (the '502's own ref [36])

Why combine: US 7,704,510 supplies (i) the antigen, (ii) an express statement that antibodies to it neutralise E/E cells, (iii) deposited hybridomas and the resulting antibodies, and (iv) a genus of "neutralising binding partners" defined by ≥70% CDR identity to those antibodies. Baldanti 2006 supplies the crucial predictability datum: UL131A/UL130/UL128 are highly conserved among field isolates, so an antibody raised against one strain should be broadly effective — removing the principal technical risk. The EBV/plasma-cell method supplies the routine route to human mAbs. The result is a textbook "obvious to try" scenario under KSR and In re Kubin: identified antigen + routine method + conserved epitope + documented prior success.

Especially pointed: the '502 expressly excludes 3E3 and R551A ("the antibody of the invention is not MSL‑109, 8F9, 3E3 or R551A") and expressly excludes 1F11, 2F4, 5A2 and 6G4 from U.S. Ser. Nos. 11/969,104 and 12/174,568. A negative limitation in the specification is an admission that the excluded species lie within the otherwise-claimed genus. Practically: the applicants conceded the genus was occupied and retreated to a boundary — which is exactly the posture in which obviousness of the surrounding scope becomes the dispositive question.

Pre‑AIA § 103(c) caveat: if the claims were ever subject to 11/969,104 as § 102(e)/(f)/(g) art and both were commonly owned by IRB, common ownership could have been used to disqualify that art. That affects 11/969,104 only, not the Princeton art or the journal literature, which remain fully available.

Combination C — gB / gH / gM‑gN claims

C1 + C2 + C3 = MSL‑109 (anti‑gH; EC₅₀ 0.3 µg/mL; clinical trials) + Speckner/Lantto/Ohlin gB AD‑1/AD‑2 mAb art + US 7,704,510 Example 4 (fractionated CYTOGAM® pUL128/130/131 antibodies outperform the gB fraction)

Claim groups targeted: "binds an epitope in the hCMV gH protein"; "binds an epitope in the hCMV gB protein"; "binds an epitope formed by gM and gN"; and the composition/multi-antibody claims mixing these with pentamer binders.

Motivation: Wang et al. PNAS 2007 teaches two distinct entry pathways; the gB/gH/gL/gO pathway operates in fibroblasts while the pentamer pathway operates in E/E cells. A POSHITA seeking to prevent hCMV disease across the clinically relevant cell types — endothelial (dissemination/latency), epithelial (shedding/transmission), fibroblast, retinal (retinitis) — has a direct, articulated reason to combine a fibroblast-active antibody with an E/E-active antibody. MSL‑109's clinical failure and the nongenetic escape mechanism (published later) supply a second independent motivation for cocktails over monotherapy. US 7,704,510 Example 4 supplies the third: it shows the pUL128/130/131 fraction of CYTOGAM® neutralises better than the gB fraction, which teaches selection among known targets for the combination.

Assessment: These claims are the most obviously rendered in the patent. The targets are the old art (gB since at least Britt 1990; gH since the 1990s; gM/gN since Mach 2000), the potency thresholds for gB/gH/gM‑gN antibodies in the specification are the same ≤2 µg/mL range as the pentamer claims, and the multi-antibody composition claims recite combinations whose additive/synergistic rationale the specification itself articulates in purely functional terms.

Combination D — the "downstream" claims

6. What would be the strongest affirmative § 103 rejection, in one paragraph

Combine Hahn 2004 + Wang & Shenk PNAS 2005 + Ryckman 2008 + Gerna 2008 (Apr) + US 7,704,510 + Baldanti 2006, and optionally the EBV/plasma-cell mAb method. The art discloses the target locus, the target complex, the clinical rationale (E/E tropism as the disease-relevant pathway), an assay for measuring neutralisation on clinical isolates in E/E cells, an antecedent report of neutralising anti‑pUL128/pUL130 monclonals that spare fibroblasts, the conservedness of the target, and a method for producing human mAbs at will. What remains is the discovery of particular antibodies with particular sequences and unexpectedly high potency. For any claim whose scope is target-plus-function and does not recite a sequence or a potency threshold, the remaining discovery is not inventive. That is a complete prima facie case under KSR for claim groups 1–3 and 8–14 of the earlier reconstruction.

7. Rebuttal: unexpected results and the limits of the rejection

The § 103 case is materially weaker for the potency-limited and sequence-defined claims, for these reasons:

7.1 Unexpected results (the applicant's best argument). The intrinsic evidence is unusually strong and is conceded by the art itself:

• Art range for prior antibodies: 0.5–20 µg/mL (specification admission); 3E3 ≈ 20 µg/mL at 50% (US 7,704,510).
• The '502's own reported antibodies achieve IC₉₀ 5–200 pM — i.e. roughly 0.00075–0.03 µg/mL (Macagno et al., J Virol 84(2):1005–1013, published online 4 Nov 2009 — after the critical date, so not prior art, but admissible as evidence of the results achieved).
• That is a 20‑fold to >10,000‑fold improvement, and it is an improvement in the type of property the claims are drafted around.

The Federal Circuit treats a difference in kind or an unexpectedly large difference in degree as rebutting a prima facie obviousness case. A 100–10,000× potency gain over the closest reported antibody — obtained in the same assay, against the same target — is the classic profile of a nonobvious improvement. Note, though, the evidentiary weakness: the unexpected-results case depends on data generated after the critical date and by the inventors themselves. A skeptical examiner or court may treat the comparison as not commensurate with the claim scope, particularly because the specification does not present a side-by-side comparison against 3E3 or MSL‑109 in the passages I received.

7.2 "Obvious to try" does not automatically reach a specific antibody. Where prior art discloses an antigen but no antibody structure, there is an established line of authority that claims to a specific antibody are not obvious merely because the antigen is known — the skilled artisan is not armed with the antibody's structure, and the § 103 inquiry requires that the claimed subject matter be identified in or suggested by the art. The claims reciting SEQ ID NO-defined VH/VL pairs (e.g. VH 200 + VL 201; VH 208 + VL 213; VH 212 + VL 201; VH 228/VL 229; VH 242/VL 243; VH 258/VL 259; VH 290/VL 291; VH 294/VL 291) and the ≥95% CDR identity claims are the ones that resist this line of attack, because the art supplies no structural starting point from which the recited sequences would have been arrived at. I cite this doctrine descriptively because I cannot name the controlling decisions with confidence.

7.3 Counter-counter: the art does supply a structural starting point in this case. US 7,704,510 claims "neutralising binding partners … comprising one or more CDRs having 70% or greater identity to a CDR present in the deposited mAbs." A genus defined at 70% CDR identity is a broad structural teaching. If the examiner uses that disclosure as the structural anchor, the § 103 case reaches further into the sequence claims than 7.2 suggests — a claim reciting ≥95% identity over a different set of SEQ ID NOs may still be a nonobvious narrow species, but the 70%‑identity genus claims would be squarely met. (Note the 95%-identity claims in the '502 and the 70%-identity claims in US 7,704,510 are drawn to different sequence sets, so this is a scope-arithmetic point, not an identity of subject matter.)

7.4 The ≤ 10 µg/mL / "not MSL‑109 or 8F9" claim is the most vulnerable functional claim. It sits only ~2‑fold below the top of the conceded art range and the applicant felt obliged to carve out two known antibodies to distinguish it. A functional boundary drawn that close to the prior art, with no structural limitation, invites a § 103 rejection on the ground that it would have been obvious to try a modest number of antibodies and select one meeting the threshold — and, if the prior-art antibodies were themselves near the line, invites a § 112 written-description/enablement challenge under Amgen v. Sanofi alongside it. This claim is the weakest in the patent, not the strongest.

7.5 Secondary considerations beyond unexpected results.

• Failure of others: MSL‑109's clinical failure in CMV retinitis in AIDS patients and in stem-cell transplantation (Boeckh 2001, Biol Blood Marrow Transplant 7:343–351; Borucki 2004, Antiviral Res 64:103–111 — both verified) is a genuine "failure of others" narrative, reinforced later by the demonstration that CMV can incorporate MSL‑109 into assembling virions as an escape mechanism.
• Long-felt need: independently documented — CMV HIG has "limited" efficacy, gB/MF59 vaccine induces poor E/E-neutralising antibodies (Cui et al., Vaccine 26:5760–5766, 2008), and no high-potency E/E-neutralising mAb existed.
• Skepticism / teaching away: weak-to-moderate. The art steered attention to gB and gH (MSL‑109, gB/MF59) and reported only modest potency from the UL128/UL130 approach; a case can be made that the field's focus was away from the pentamer as a therapeutic antibody target. But the art never said the approach could not work — it said it worked modestly. Under KSR, "modest" invites optimisation; it does not foreclose it. I would not expect a "teaching away" argument to carry these claims alone.
• Commercial success / unexpected results nexus: the earlier Litigation Summary noted a database associating the family with a biologic (SYLVANT/siltuximab) that the analyst found to look like a database artifact. I endorse that caution — no commercial-success evidence for the '502 antibodies should be assumed, and none of the antibodies named in the patent appear to have reached the market on the record before me.

8. Claim-group conclusions

9. Where my confidence is limited

1. No verbatim claims. Every conclusion above is claim-group-level and keyed to the specification's aspect recitations. A claim-by-claim chart requires the PatentCenter "Claims" tab for 15/284,799. If, as is common in this family, the granted claims are narrow sequence claims, the § 103 exposure is far smaller than the table above implies; if the granted claims include a broad target-plus-function genus, the exposure is severe.
2. The "Prior Art" section you directed me to was not present in the supplied material, and the '502's own reference list (which would decode numerals [2]–[42]) was not supplied either. Anyone relying on the bracketed citations — particularly [7] for the UL128/UL130 potency statement — must verify them against the printed patent first. I have not assumed a mapping.
3. Reference dates were verified individually where possible (Hahn 2004, Wang & Shenk 2005 ×2, Adler 2006, Baldanti 2006, Patrone 2005, Ryckman 2008, Gerna 2008, Wang 2007). I did not verify the exact publication date of US 2008/0187545 A1 (Shenk et al., listed as published 2008-08-07 — after the critical date) or resolve its relationship to the 11/810,578 family; I therefore relied on the US 7,704,510 / 11/810,578 filing date of 2007-06-06 (pre‑AIA § 102(e)) rather than the later publication. Note also 2009/0081230 (Lanzavecchia et al.) appears in third-party cited-art tables and is plausibly the publication of 11/969,104 — I could not verify that, so do not treat it as established.
4. The pre-AIA/AIA threshold question (¶ "Threshold legal question" above) is unresolved and would change the available art and the common-ownership treatment under § 103(c).
5. Case law is cited from memory and is flagged as needing verification; I have deliberately avoided naming antibody-specific Board decisions I cannot quote.

Bottom line: US 9,725,502's target-defined and composition/use claims are highly vulnerable under § 103 on a combination of the 2004–2008 hCMV tropism literature (Hahn, Wang & Shenk, Ryckman, Gerna) with the Princeton antibody art (US 7,704,510) and the conservedness data (Baldanti). Its sequence-defined and high-potency-defined claims rest on a single, genuinely strong, but not invincible, footing: the 20‑fold-to-10,000‑fold potency improvement over an art range the applicants themselves admitted to be 0.5–20 µg/mL.