United States Patent 5,712,155: Critical Claim Validity and US Patent Landscape Analysis

US Patent 5,712,155 claims isolated DNA and engineered expression constructs encoding TNF-binding TNF-R polypeptides defined by (i) specific amino-acid ranges from FIG. 2A/FIG. 3A, (ii) DNA hybridization to a complement under “moderately stringent” conditions, and (iii) sequence identity thresholds (at least 88%) tied to functional TNF binding (with quantified binding thresholds in select dependent claim groupings). The landscape risk is driven by two claim architectures that often overlap with prior art in TNF-binding receptor engineering: DNA/hybridization-defined nucleic acids and receptor/ligand binding variants defined through identity percentages plus functional assays.

This analysis decomposes the claims into enforceable boundaries, tests the claim logic against typical novelty and obviousness attack paths for this class of biologics patents, maps the practical freedom-to-operate (FTO) effects of the claim language, and highlights the likely competitive patent space around TNF-receptor fusion proteins and engineered TNF-binding domains in the US.

What exactly does US 5,712,155 claim, in enforceable terms?

Claim set architecture (what is actually being protected)

The patent’s claim set clusters into three functional buckets:

Bucket	Claim numbers	What the patentee owns
Nucleic acids	1-3, 10-12, 15	Isolated DNA sequences defined by amino-acid ranges and/or hybridization plus identity and functional TNF binding
Expression constructs	4-6, 13, 16	Recombinant expression vectors comprising the claimed DNA
Biological producers	7-9, 14, 17	Host cells transformed/transfected with the claimed vectors

Core structural definition: amino-acid span and identity

Claims 1-3 and 10-12 are built around the same polypeptide backbone definition:

Backbone amino acids: “amino acids 1 to X of FIG. 2A and amino acids 1 to 233 of FIG. 3A, wherein X is an amino acid from 163 to 235.”
Identity threshold: “at least 88% identical” to the polypeptide encoded by the DNA of (a).
Hybridization gate (in (b) elements for claims 1-3): DNA capable of hybridization to the complement under moderately stringent conditions (50° C., 2× SSC).
Function gate:
- Claims 1 and 2: binding to TNF; claims 2 and 3 add quantified binding thresholds.
- Claims 2 and 11: “greater than 0.1 nmoles TNF per nmole TNF-R.”
- Claims 3 and 12: “greater than 0.5 nmoles TNF per nmole TNF-R.”

Dependent claim engineering levers: glycosylation, protease cleavage, cysteine modification

Claims 10-12 expand protection to engineered variants that are “identical” to the base except for modifications selected from:

(i) inactivated N-linked glycosylation sites
(ii) altered KEX2 protease cleavage sites
(iii) conservative amino-acid substitutions
(iv) substitution or deletion of cysteine residues
(v) combinations of modifications (i)-(iv)

These variants are tethered to the TNF binding function:

Claim 11: “greater than 0.1 nmoles TNF per nmole”
Claim 12: “greater than 0.5 nmoles TNF per nmole”

“Full length to 235” fallback coverage

Claim 15 tightens the amino-acid requirement to:

(a) “amino acids 1-235 of FIG. 2A”
(b) hybridization-defined DNA encoding TNF-binding polypeptide with ≥88% identity to the polypeptide encoded by DNA of (a)

This is the broadest type of anchor in the nucleic-acid definition because it reduces ambiguity in the amino-acid upper bound relative to the variable X range in claims 1-3 and 10-12.

How broad is the practical claim scope against realistic TNF-binding biologics design routes?

Broadness from hybridization + identity thresholds

The claim language uses two common breadth-expansion mechanisms:

Hybridization definition
“Capable of hybridization to the complement of the DNA sequence of (a) under moderately stringent conditions (50° C., 2× SSC).”
This can cover nucleic acids that are not strictly identical but share substantial complementarity. In enforcement, this typically shifts fights toward expert-based interpretation of assay conditions and what sequences “hybridize” under those parameters.
Identity at 88%
For proteins up to roughly 235 aa, 88% identity can still tolerate multiple substitutions, including potentially in surface-exposed residues depending on the alignment strategy. That level of identity can leave room for obvious design-around variants that keep binding while drifting sequence identity.

Function limitations reduce some breadth but are assay-dependent

Quantified binding thresholds (claims 2, 3, 11, 12) reduce overbreadth by requiring TNF binding activity above a defined level:

0.1 nmoles TNF per nmole TNF-R
0.5 nmoles TNF per nmole TNF-R

However, function-based limitations can still be vulnerable in validity and claim construction:

They raise issues in prior art searches (were such binding metrics measured similarly?).
They create potential enforceability friction if accused products meet binding via different assay format, oligomerization state, glycosylation, or presentation that affects apparent binding units.

Engineering modifications map to typical expression optimization

The glycosylation and protease cleavage site modifications are the most operationally actionable components of the claims:

N-linked glycosylation site inactivation is a common route to control heterogeneity, stability, or expression yield.
KEX2 cleavage site alteration suggests a yeast or fungal expression processing context (KEX2 is well known as a protease in yeast). This implies the claimed receptor variants are optimized for particular expression systems that process precursor forms.
Cysteine substitution/deletion targets disulfide management, mispairing reduction, or changing redox properties.

These are common levers in biotech prior art, which increases obviousness pressure when combined with a known TNF receptor binder scaffold.

What are the most likely novelty and obviousness attack paths?

The patent’s defenses against invalidity will center on whether the claimed scaffold and the engineered variants were previously disclosed with the same combination of:

the specific amino-acid span constraint (including the FIG.-based region definition),
the 88% identity language,
the specific hybridization condition definition, and
the engineered site modifications tied to quantitative binding performance.

Novelty pressure: TNF-R binders and receptor fragments

The claims cover DNA encoding a polypeptide that binds TNF. The prior art landscape for TNF-binding proteins was active long before this filing period:

purified TNF receptors and receptor fragments (extracellular domain constructs) were broadly investigated as TNF antagonists,
receptor engineering variants were disclosed to tune binding and improve developability (glycosylation, protease sites, cysteine management).

Likely novelty failure scenarios:

A prior art reference discloses the same or highly overlapping receptor-binding domain sequence (or nucleotide sequence) and its encoding nucleic acid.
Another reference discloses engineered variants (glycosylation site inactivation, protease cleavage site changes, cysteine variants) with measured TNF binding.

Because claims 10-12 require “identical … except for modification(s)” selected from a list, if prior art already includes such modifications on a disclosed TNF-R sequence, those claims can be vulnerable.

Obviousness pressure: combining known scaffold with known expression optimizations

Even if the exact sequence is not in a single reference, obviousness often proceeds by:

identifying a known TNF-binding receptor scaffold (or fragment),
proposing routine expression modifications (glycosylation site inactivation, cleavage site tuning, cysteine management) as known techniques,
applying predictable functional outcomes (retained TNF binding).

The key obviousness vulnerability is that claims 10-12 do not require a specific mutation pattern beyond the category list. “Conservative amino acid substitutions” and “combinations” expand the set of allowable variant designs, which makes it easier for an obviousness challenger to argue that a range of variants would be predictable.

Functional thresholds as secondary considerations

Measured TNF binding above 0.1 or 0.5 nmol per nmol may be argued as improved performance. But these thresholds face two common validity risks:

Reproducibility and assay comparability: if prior art used a similar assay but reported different units or different experimental constructs, function-based differentiation can collapse.
Optimization not invention: if the modifications are standard expression and folding optimization steps, challengers argue that any improvement is the predictable result of better expression and presentation rather than a new technical contribution.

How defensible are the hybridization-defined nucleic-acid claims?

Hybridization language is broad but sometimes hard to enforce

The claims’ hybridization definition (“50° C., 2× SSC”) is a standard-style technique used to cover nucleic acid variants. In practice:

It can capture sequences that are not explicitly disclosed as such.
It can be difficult to enforce against an accused nucleic acid unless the test is carefully controlled and the parties agree on what “capable of hybridization” means for the relevant product DNA.

Hybridization + identity may still be attacked for “overbreadth”

The patent combines hybridization constraints with an identity metric (≥88%). Identity does not eliminate hybridization breadth; it just narrows it to a subset that must match the function and identity threshold to the polypeptide in the figure-defined anchor.

In invalidity litigation, challengers often argue:

The identity threshold is low enough to cover many plausible receptor variants.
The hybridization element makes claim coverage non-local and potentially includes sequences that were not the focus of the disclosure.

Where are the likely competitive patent clusters in the US?

Without embedding external document text, the dominant US competitive clusters for TNF antagonists typically include:

TNF receptor extracellular domain constructs
- DNA constructs encoding the TNF-binding extracellular domain (or fragments) of the receptor.
- Claims often focus on amino-acid sequences of the extracellular domain and conserved cysteine patterns.
Sequence variants and glycoengineering
- Mutations to N-linked glycosylation sites to change expression profile.
- Cysteine substitutions/deletions to manage disulfides and aggregation.
Host system processing and protease site tuning
- For yeast or fungal expression, edits to protease processing sites (KEX2-related) are common.
- Patents often tie cleavage site engineering to improved yield and correct processing.
Functional TNF binding potency requirements
- Patents frequently include potency or binding metrics measured by ELISA-like formats, binding kinetics, or cell-based assays.
- The presence of quantified thresholds in 5,712,155 (0.1 and 0.5 nmol TNF per nmol TNF-R) positions it among those clusters where competitors may match the same biological activity metrics.

Competitive implication for FTO: if competing products are produced by engineered TNF-R extracellular domain variants that preserve TNF binding while using similar glycosylation/protease/cysteine strategies, they may fall within claim scope if their sequences also satisfy the ≥88% identity and hybridization criteria. If they use substantially diverged sequences, change the target receptor architecture, or avoid the specific processing-linked engineered scaffold, risk shifts down.

What are the most important design-around vectors (and why they matter)?

1) Sequence divergence beyond the “≥88% identity” threshold

The identity ceiling is a critical boundary.

A design-around can reduce risk if the polypeptide variant falls below the claimed identity threshold relative to the FIG.-anchored polypeptide sequence.

2) Avoiding the claimed modification set language

Claims 10-12 define variants “identical … except for modification(s)” from a closed set (i)-(iv) plus combinations. If a competitor introduces additional changes outside that list (even if conservative) to meet manufacturability or pharmacokinetics goals, they may reduce “except for modification” coverage.

3) Avoiding the hybridization conditions characterization

A competitor could avoid nucleotide-level overlap by using distinct nucleotide sequences even when amino-acid sequences are similar. But because claims ultimately tie nucleic acid to encoded polypeptide and identity thresholds, nucleotide-only design-around has limited impact unless identity drops or function is altered.

4) Using different biological constructs that do not map to the claimed polypeptide span

The FIG.-based span definition (X from 163 to 235; plus claims anchored to 1-235) is a practical mapping constraint. Competitors using different extracellular domain boundaries or different receptor constructs can escape if they are outside the defined spans.

How does the claim structure affect enforcement strategy?

Direct infringement targets nucleic acids and expression constructs

Claims 1-3, 4-9, 10-12, and 13-17 are layered so that:

If a competitor makes the DNA (directly or via a provider), claim coverage can extend to the DNA product itself (isolated DNA sequences).
If they express it, vector and host-cell claims follow.

A typical enforcement sequence:

capture an accused construct as a recombinant expression vector or host cell,
use DNA sequence and translation alignment to test the amino-acid and 88% identity conditions,
use hybridization test interpretation to bind variants,
use TNF binding assay results to test threshold limitations.

Key Takeaways

US 5,712,155 protects TNF-binding receptor (TNF-R) polypeptides through a layered definition: FIG.-anchored amino-acid span, ≥88% identity, and in many claims moderate-stringency hybridization (50° C., 2× SSC) to a complementary sequence, plus in select claims quantified TNF binding thresholds (>0.1 or >0.5 nmol TNF per nmol TNF-R).
The breadth-expansion levers are hybridization and 88% identity, while the narrowing levers are specific engineered modification categories (glycosylation site inactivation, KEX2 cleavage site alteration, conservative substitutions, cysteine substitution/deletion) and functional binding thresholds.
Validity pressure is high in this patent class because TNF receptor binders and receptor engineering strategies (especially glycosylation, protease site, and cysteine management) are common biotech prior art themes. The claims’ “category list” nature in 10-12 supports obviousness combinations unless the patent can anchor on a narrow, non-obvious combination disclosed and measured in the same framework.
The biggest design-around levers are to move the product’s polypeptide identity below 88% to the FIG.-anchored sequence, use constructs with different domain boundaries outside the claimed spans, or introduce changes outside the “except for modification(s)” categories.
Enforceability will likely hinge on aligning the accused sequences to the patent’s FIG.-defined anchors and translating that into a credible mapping for hybridization capability and the stated TNF binding metrics.

FAQs

1) What parts of the claims are most likely to be litigated in claim construction?

The most litigated elements are typically the hybridization conditions (“50° C., 2× SSC”), the meaning of “capable of hybridization,” and how 88% identity is calculated (alignment method and region). The TNF binding threshold units and assay format also commonly become dispute points.

2) Do claims 1-3 and 10-12 rely on the same core polypeptide definition?

Yes. Both groups use FIG.-based amino-acid span constraints and an identity threshold of at least 88%, with 10-12 adding explicit categories of permitted engineered modifications.

3) Which claims include quantitative TNF binding thresholds?

Claims 2, 3, 11, and 12 include quantified binding thresholds: >0.1 and >0.5 nmoles TNF per nmole TNF-R, respectively.

4) How do the “KEX2 protease cleavage site” modifications affect competitor risk?

They suggest the engineered scaffold is tuned for expression processing in systems where KEX2 acts. Competitors using different expression processing routes or leaving cleavage processing unchanged can reduce overlap with the engineered variant claims.

5) Are the host-cell and vector claims dependent on the nucleic-acid claims?

Yes. The vector and host-cell claims are explicitly tethered to earlier claims (e.g., claim 4 depends on claim 1; claim 7 depends on claim 4; and similarly for the 10-12 and 13-17 chain). If the nucleic-acid claims fall, dependent construct claims often fall with them.

References

US Patent 5,712,155. (Document content as provided in prompt: claims 1-17 text).

Applicant	Tradename	Biologic Ingredient	Dosage Form	BLA	Approval Date	Patent No.	Expiredate
Immunex Corporation	ENBREL	etanercept	For Injection	103795	November 02, 1998	⤷ Start Trial	2014-11-29
Immunex Corporation	ENBREL	etanercept	For Injection	103795	May 27, 1999	⤷ Start Trial	2014-11-29
Immunex Corporation	ENBREL	etanercept	Injection	103795	September 27, 2004	⤷ Start Trial	2014-11-29
Immunex Corporation	ENBREL	etanercept	Injection	103795	February 01, 2007	⤷ Start Trial	2014-11-29
Immunex Corporation	ENBREL MINI	etanercept	Injection	103795	September 14, 2017	⤷ Start Trial	2014-11-29
Immunex Corporation	ENBREL	etanercept	Injection	103795		⤷ Start Trial	2014-11-29
>Applicant	>Tradename	>Biologic Ingredient	>Dosage Form	>BLA	>Approval Date	>Patent No.	>Expiredate

Country	Patent Number	Estimated Expiration
South Africa	907072	⤷ Start Trial
World Intellectual Property Organization (WIPO)	9406476	⤷ Start Trial
World Intellectual Property Organization (WIPO)	9319777	⤷ Start Trial
World Intellectual Property Organization (WIPO)	9103553	⤷ Start Trial
United States of America	RE36755	⤷ Start Trial
United States of America	7459528	⤷ Start Trial
>Country	>Patent Number	>Estimated Expiration

Patent: 5,712,155

Summary for Patent: 5,712,155