United States Patent 5,840,299: Humanized anti–α4 integrin immunoglobulin (mouse 21-6) and US landscape impact

US 5,840,299 claims a family of humanized IgG variable domains (heavy and light) derived from the mouse 21-6 antibody that “specifically binds to alpha-4 integrin,” with a binding affinity lower limit of about 10^7 M^-1 and an upper limit of five-times the binding affinity of the mouse 21-6 antibody. The claims further anchor humanization to defined mouse CDR sequences (mouse 21-6 variable domains referenced as SEQ ID NO: 2 for light and SEQ ID NO: 4 for heavy) and to human framework scaffolds (light kappa framework from SEQ ID NO: 6 “RE1,” heavy framework from SEQ ID NO: 10 “21/28′CL”), while restricting key Kabat framework positions that may be retained from the mouse or substituted into human frameworks.

The practical patent value of 5,840,299 is that it is not “genus-only” around an anti-α4 integrin antibody. It is a detailed variable-domain humanization specification that recites:

the exact CDR source (mouse 21-6 CDRs),
the exact framework source (specific human kappa and heavy frameworks by sequence ID),
position-level humanization constraints at selected Kabat positions,
and a functional binding-affinity band versus mouse 21-6, plus downstream claims on compositions, nucleic acids, and methods (diagnosis, adhesion inhibition, inflammatory disease including multiple sclerosis).

What follows is a claim-by-claim critical analysis of scope, enforceability posture, design-around pressure points, and how the rest of the US competitive landscape typically evaluates such antibody humanization patents.

What exactly does claim 1 lock down?

Core binding target and affinity constraint

Claim 1 defines a humanized immunoglobulin that:

specifically binds α4 integrin, and
has binding affinity “having a lower limit of about 10^7 M^-1 and an upper limit of about five-times the binding affinity of the mouse 21-6 immunoglobulin**.”

This is a functional limitation tied to the comparator antibody (mouse 21-6). That matters legally because it can narrow scope beyond mere binding. If a competing antibody binds α4 integrin with dramatically higher or lower affinity than the “five-times” band, it can fall outside claim coverage.

CDRs are anchored to mouse 21-6 variable domains

Claim 1 requires:

Humanized light chain CDR1/2/3 sequences are taken from the corresponding CDRs of mouse 21-6 light variable domain (SEQ ID NO: 2).
Humanized heavy chain CDR1/2/3 sequences are taken from the corresponding CDRs of mouse 21-6 heavy variable domain (SEQ ID NO: 4).

So the invention keeps the antigen-contacting regions as a direct transplant, then swaps frameworks under position restrictions.

Frameworks are anchored to human templates with limited “mouse residue retention” at specified Kabat positions

For the light chain variable region:

Framework is from a human kappa light variable framework “provided that at least one position selected from L45, L49, L58, L69 is occupied by the same amino acid residue present in the equivalent position of the mouse 21-6 light variable framework.”

For the heavy chain variable region:

Framework is from a human heavy chain variable framework “provided that at least one position selected from H27, H28, H29, H30, H44, H71 is occupied by the same amino acid residue present in the equivalent position of the mouse 21-6 heavy variable framework.”

Two critical scope effects follow:

1) “At least one position” is a low threshold. The claim does not require all listed positions to be retained. A competitor can choose a human framework and retain only one of the specified residues (or satisfy the condition another way if there are sequence identities due to framework similarity).

2) The claim language uses “occupied by the same amino acid residue… in the equivalent position of the mouse.” This makes the framework constraint partly a sequence-identity constraint, not merely “humanization by typical methods.”

Binding to α4 integrin is specified at method level but not in structural form

Claim 1 does not recite explicit residues that determine α4 specificity beyond the CDRs taken from mouse 21-6. That makes CDR identity central and makes affinity band a second pillar.

How do dependent claims constrain the frameworks and shrink design-around space?

Claim 2: Light framework must be RE1 with additional position constraint versus “any other human kappa other than RE1”

Claim 2 narrows the light framework to:

RE1 variable framework sequence (SEQ ID NO: 6),
with an additional requirement that at least one position selected from L104, L105, L107 is occupied by the same amino acid residue present in the equivalent position of a kappa light chain from any human immunoglobulin other than RE1 (SEQ ID NO: 6).

This clause is unusual: it forces alignment of certain positions with a generic “other human kappa” template. The scope likely functions as an additional structural discriminator that prevents trivial substitutions among human kappa frameworks.

Claim 3: Heavy framework must be 21/28′CL (SEQ ID NO: 10)

Claim 3 further narrows heavy framework source to:

21/28′CL variable framework sequence (SEQ ID NO: 10).

Claim 4–6: Minimum numbers of residues retained from mouse and exact position selections

Claim 4 imposes minimum residue retention counts:

Light framework: “at least three amino acids from mouse 21.6… at positions in the first group” and three amino acids from a human kappa from any human immunoglobulin other than RE1 at positions in the third group.
Heavy framework: “at least five amino acids from mouse 21.6… at positions in the second group.”

Claim 5 elevates the precision:

Light framework identical to RE1 except for the “at least three positions” from the first group and “three positions” from the third group.
Heavy framework identical to 21/28′CL except for the “at least five positions” from the second group.

Claim 6 then makes the mapping explicit:

Light retained positions are L45, L58, L69.
Heavy retained positions are H27, H28, H29, H30, H71.

This sequence-position mapping is the strongest design-around barrier in the document. Many antibody humanization efforts can vary frameworks broadly, but if the claim requires specific Kabat positions to match mouse residues and also fixes the framework templates by SEQ ID, it becomes harder to “escape” while maintaining CDR identity and α4 binding.

Claim 7–8: CDR3 details at H98 (phenylalanine optional vs required)

Claim 7 states the humanized CDRs are identical to mouse 21-6’s, except CDR3 of heavy chain may or may not have a phenylalanine at H98. Claim 8 requires phenylalanine at H98.

This creates two tiers of scope:

claim 1 and some intermediates cover both H98 variants,
claim 6/14/16-based descendants can be constrained to the F-containing variant depending on which claim path is used.

Claims 9–16: Enumeration of mature variable sequences (La/Lb/Ha/Hb/Hc and variants)

Claims 9–13 define mature light and heavy variable amino acid sequences by FIG references and SEQ ID numbers:

Light: La (SEQ ID NO: 7) and Lb (SEQ ID NO: 8)
Heavy: Ha (SEQ ID NO: 11), Hb (SEQ ID NO: 12), Hc (SEQ ID NO: 13)

Then claims 14–16 specify heavy sequence identity within the “La” path:

Ha / Hb / Hc in combinations with claim 9 (La).

From an enforcement perspective, these are “hard” sequences. If a competitor uses any alternative heavy variant sequence outside those enumerated, it can avoid literal infringement where those specific claim-dependent branches apply.

Do the later claims broaden beyond the variable sequences?

Claim 17: Antigen-specific binding fragment

Claim 17 covers binding fragments of the humanized immunoglobulin of claim 14 or 16. Such fragments could include Fab, F(ab’)2, scFv variants (depending on definitions in the specification), and any fragment with the claimed binding specificity.

Because claim 17 is dependent on a specific heavy sequence (claim 14 or 16), the fragment scope likely stays tethered to those variable-domain sequences.

Claims 18–20: Constant region effector function switch

Claim 18 adds “constant region domain.” Claim 19 requires constant region that is “incapable of complement fixation and antibody dependent cellular toxicity.” Claim 20 requires constant region with complement fixation or ADCC activity.

This splits effector function coverage. For infringement analysis, constant region differences do not change variable domains, so both pro- and non-effector formats can be covered if the variable domains match.

Claims 21–22: Nucleic acids

Claim 21 and 22 cover nucleic acids encoding heavy or light chain variable domains or binding fragments.

These are typical enforcement tools against gene constructs used in expression systems.

Claim 23: Pharmaceutical compositions

Covers compositions with the humanized immunoglobulin (claim 14 or 16) and a carrier.

Claims 24–29: Methods (diagnosis, adhesion inhibition, inflammatory disease, multiple sclerosis)

Claim 24: detection of α4 integrin in tissue samples via binding complex detection.
Claim 25–26: inhibition of leukocyte adhesion to endothelial cells; includes brain cell endothelial targeting.
Claim 27–29: treating inflammatory disease including multiple sclerosis, including arrests symptoms.

These are downstream therapeutic method claims that generally rise or fall with the underlying antibody infringement.

Where is the patent strongest versus vulnerable?

Strength: structural specificity at variable-domain level

The tight coupling of:

CDR sequences copied from mouse 21-6 (SEQ ID NO: 2 and 4),
human framework scaffolds defined by SEQ ID (SEQ ID NO: 6 and 10),
explicit Kabat positions (L45/L58/L69 and H27/H28/H29/H30/H71),
and optionally required H98 phenylalanine

creates a “precision humanization” profile. If a product matches those sequence constraints, it is difficult to argue it is outside literal scope.

Vulnerability: “at least one position” language appears earlier than the most explicit position list

Even though claims 4–6 become explicit, the earlier broad claim 1 is framed as “at least one position selected from” two groups.

That broad language can cut both ways:

For enforcement, it could expand coverage to variants that keep only one required mouse residue at a specified Kabat position, potentially making it harder to design around.
For validity challenges, it can also raise questions about whether the claim scope captures antibodies that were not enabled or not clearly supported by the disclosure, depending on the support for each combination path (e.g., if only certain variants were exemplified).

Vulnerability: affinity band as a functional limitation may invite measurement disputes

The “about 10^7 M^-1 to five-times binding affinity of mouse 21-6” introduces dependence on how affinity is measured (assay conditions, formats). In litigation, that can become a technical battleground.

It also creates a potential design-around: a competitor could aim for affinity outside the upper bound while retaining similar CDRs.

What competitive design-around strategies are most plausible against this specific claim set?

1) Alter heavy framework substitutions outside the required Kabat-position set while staying close to RE1 and 21/28′CL templates.

Because claim 6 requires specific position identities (L45, L58, L69 in the light; H27, H28, H29, H30, H71 in heavy), omitting any one of those while attempting to preserve binding risks moving outside the more specific descendant claims.

2) Use a different heavy CDR3 H98 status and/or a heavy variable sequence not matching Ha/Hb/Hc.

Claims 8 and 14–16 create sub-scope around H98 phenylalanine and specific heavy sequence enumerations.

3) Reformat constant region only if effector-function coverage is what is targeted.

Constant region differences likely do not avoid infringement for variable domains, but they can affect which dependent claim branches apply (complement fixation/ADCC capable vs incapable).

4) Create a fragment format (Fab/F(ab’)2/scFv) that does not embody the claimed “mature variable sequences” tied to the enumerated La/Lb and Ha/Hb/Hc.

Claim 17 is tethered to claim 14 or 16, so avoiding those exact variable sequences is the cleanest escape from fragment claims.

How does this map onto known α4 integrin therapeutic antibody paradigms?

This patent sits in the broader universe of α4 integrin (CD49d) biology, where therapeutic antibodies were developed to modulate leukocyte trafficking and treat inflammatory diseases including multiple sclerosis.

Within that paradigm, the US 5,840,299 approach is classic for antibody engineering patents:

define CDR donor (mouse 21-6),
humanize frameworks using specific human templates,
preserve key framework residues at selected Kabat positions,
then claim composition and method use.

The claim set’s strongest identifying markers are not “α4 integrin” itself; it is the combination of:

mouse 21-6 CDR identity, plus
human framework sequence IDs, plus
explicit Kabat position constraints, plus
enumerated matured variable sequences.

That combination is what makes this patent more product-specific than a general “humanized anti-CD49d” claim.

Patent landscape implications for US filings and freedom-to-operate

What matters for FTO analysts

For product teams developing anti-α4 integrin antibodies in the US, the relevant infringement question is whether the candidate antibody:

uses the mouse 21-6 CDRs (SEQ ID NO: 2 and 4) in full,
uses the specified human framework templates (light RE1, heavy 21/28′CL) with the specific Kabat-position matches,
and has binding affinity within the stated band.

Even if a candidate binds α4 integrin, mismatch at any of these layers can place it outside literal coverage of the narrower dependent claim set (claims 2–6, 8, 9–16).

Where competing patents often cluster

In this technology space, competitors typically attempt to:

change CDR sequences while retaining α4 binding (escape the “mouse 21-6 CDR identity” requirement),
use different human framework scaffolds or alter the specified Kabat positions (escape the framework-position constraints),
add different binding formats or hinge/constant region architectures (usually not enough to avoid variable-domain-tethered claims),
or pursue different epitope angles that require different CDRs (stronger escape).

Because this patent claims down to explicit mature variable sequences (La/Lb and Ha/Hb/Hc), competitors are incentivized to design new sequences rather than merely “conservative” humanization.

Key Takeaways

US 5,840,299 claims a humanized anti–α4 integrin IgG family built from mouse 21-6 CDRs (light SEQ ID NO: 2; heavy SEQ ID NO: 4) combined with fixed human framework scaffolds (light RE1 SEQ ID NO: 6; heavy 21/28′CL SEQ ID NO: 10).
The broadest anchor is claim 1: α4 integrin binding with an affinity band (about 10^7 M^-1 to about 5x mouse 21-6) plus the CDR identity.
The narrowest enforceability leverage is claim 6 (explicit Kabat positions): light L45/L58/L69 and heavy H27/H28/H29/H30/H71, plus heavy CDR3 H98 phenylalanine (claims 7–8) and enumerated mature sequences (La/Lb; Ha/Hb/Hc).
Downstream claims (binding fragments, nucleic acids, compositions, diagnostic and therapeutic methods including multiple sclerosis) generally track the underlying variable-domain infringement.
The clearest design-around is not “changing constant regions,” but changing at least one of: CDR donor identity, framework template identity, required Kabat-position matches, or the enumerated matured heavy/light variable sequences tied to dependent claims.

FAQs

1) What is the central biological target claimed in US 5,840,299?
Alpha-4 integrin (CD49d), bound by the humanized immunoglobulin.

2) What role do the mouse 21-6 sequences play in claim scope?
They define the CDR sequences: light CDRs come from mouse light variable SEQ ID NO: 2 and heavy CDRs come from mouse heavy variable SEQ ID NO: 4.

3) Are the framework scaffolds generic, or fixed?
Fixed for the key dependent claims: the light framework is RE1 (SEQ ID NO: 6) and the heavy framework is 21/28′CL (SEQ ID NO: 10), with specific Kabat positions required to match mouse residues.

4) How does the affinity limitation change infringement risk?
It adds a functional band requirement in claim 1 (about 10^7 M^-1 to about 5x mouse 21-6), which can exclude antibodies binding α4 integrin with affinity outside that range.

5) Does the patent cover methods for multiple sclerosis?
Yes. It claims treating inflammatory disease including multiple sclerosis via administering the pharmaceutical composition comprising the antibody of the specified dependent claims.

References

[1] United States Patent 5,840,299. “Humanized immunoglobulin specifically binding to alpha-4 integrin.” Claims 1–29.

Title:	Humanized antibodies against leukocyte adhesion molecule VLA-4
Abstract:	The invention provides humanized immunoglobulins that specifically bind to the VLA-4 ligand, and methods of treatment using the same. The methods are particularly useful for treatment of multiple sclerosis.
Inventor(s):	Bendig; Mary M. (London, GB), Leger; Olivier J. (Hertfordshire, GB), Saldanha; Jose (Enfield Middlesex, GB), Jones; S. Tarran (Radlett, GB), Yednock; Ted A. (Fairfax, CA)
Assignee:	Athena Neurosciences, Inc. (South San Francisco, CA)
Application Number:	08/561,521
Patent Claims:	see list of patent claims
Patent landscape, scope, and claims summary:	United States Patent 5,840,299: Humanized anti–α4 integrin immunoglobulin (mouse 21-6) and US landscape impact US 5,840,299 claims a family of humanized IgG variable domains (heavy and light) derived from the mouse 21-6 antibody that “specifically binds to alpha-4 integrin,” with a binding affinity lower limit of about 10^7 M^-1 and an upper limit of five-times the binding affinity of the mouse 21-6 antibody. The claims further anchor humanization to defined mouse CDR sequences (mouse 21-6 variable domains referenced as SEQ ID NO: 2 for light and SEQ ID NO: 4 for heavy) and to human framework scaffolds (light kappa framework from SEQ ID NO: 6 “RE1,” heavy framework from SEQ ID NO: 10 “21/28′CL”), while restricting key Kabat framework positions that may be retained from the mouse or substituted into human frameworks. The practical patent value of 5,840,299 is that it is not “genus-only” around an anti-α4 integrin antibody. It is a detailed variable-domain humanization specification that recites: the exact CDR source (mouse 21-6 CDRs), the exact framework source (specific human kappa and heavy frameworks by sequence ID), position-level humanization constraints at selected Kabat positions, and a functional binding-affinity band versus mouse 21-6, plus downstream claims on compositions, nucleic acids, and methods (diagnosis, adhesion inhibition, inflammatory disease including multiple sclerosis). What follows is a claim-by-claim critical analysis of scope, enforceability posture, design-around pressure points, and how the rest of the US competitive landscape typically evaluates such antibody humanization patents. What exactly does claim 1 lock down? Core binding target and affinity constraint Claim 1 defines a humanized immunoglobulin that: specifically binds α4 integrin, and has binding affinity “having a lower limit of about 10^7 M^-1 and an upper limit of about five-times the binding affinity of the mouse 21-6 immunoglobulin.” This is a functional limitation tied to the comparator antibody (mouse 21-6). That matters legally because it can narrow scope beyond mere binding. If a competing antibody binds α4 integrin with dramatically higher or lower affinity than the “five-times” band, it can fall outside claim coverage. CDRs are anchored to mouse 21-6 variable domains Claim 1 requires: Humanized light chain CDR1/2/3 sequences are taken from the corresponding CDRs of mouse 21-6 light variable domain (SEQ ID NO: 2). Humanized heavy chain CDR1/2/3 sequences are taken from the corresponding CDRs of mouse 21-6 heavy variable domain (SEQ ID NO: 4). So the invention keeps the antigen-contacting regions as a direct transplant, then swaps frameworks under position restrictions. Frameworks are anchored to human templates with limited “mouse residue retention” at specified Kabat positions For the light chain variable region: Framework is from a human kappa light variable framework “provided that at least one position selected from L45, L49, L58, L69 is occupied by the same amino acid residue present in the equivalent position of the mouse 21-6 light variable framework.” For the heavy chain variable region: Framework is from a human heavy chain variable framework “provided that at least one position selected from H27, H28, H29, H30, H44, H71 is occupied by the same amino acid residue present in the equivalent position of the mouse 21-6 heavy variable framework.” Two critical scope effects follow: 1) “At least one position” is a low threshold. The claim does not require all listed positions to be retained. A competitor can choose a human framework and retain only one of the specified residues (or satisfy the condition another way if there are sequence identities due to framework similarity). 2) The claim language uses “occupied by the same amino acid residue… in the equivalent position of the mouse.” This makes the framework constraint partly a sequence-identity constraint, not merely “humanization by typical methods.” Binding to α4 integrin is specified at method level but not in structural form Claim 1 does not recite explicit residues that determine α4 specificity beyond the CDRs taken from mouse 21-6. That makes CDR identity central and makes affinity band a second pillar. How do dependent claims constrain the frameworks and shrink design-around space? Claim 2: Light framework must be RE1 with additional position constraint versus “any other human kappa other than RE1” Claim 2 narrows the light framework to: RE1 variable framework sequence (SEQ ID NO: 6), with an additional requirement that at least one position selected from L104, L105, L107 is occupied by the same amino acid residue present in the equivalent position of a kappa light chain from any human immunoglobulin other than RE1 (SEQ ID NO: 6). This clause is unusual: it forces alignment of certain positions with a generic “other human kappa” template. The scope likely functions as an additional structural discriminator that prevents trivial substitutions among human kappa frameworks. Claim 3: Heavy framework must be 21/28′CL (SEQ ID NO: 10) Claim 3 further narrows heavy framework source to: 21/28′CL variable framework sequence (SEQ ID NO: 10). Claim 4–6: Minimum numbers of residues retained from mouse and exact position selections Claim 4 imposes minimum residue retention counts: Light framework: “at least three amino acids from mouse 21.6… at positions in the first group” and three amino acids from a human kappa from any human immunoglobulin other than RE1 at positions in the third group. Heavy framework: “at least five amino acids from mouse 21.6… at positions in the second group.” Claim 5 elevates the precision: Light framework identical to RE1 except for the “at least three positions” from the first group and “three positions” from the third group. Heavy framework identical to 21/28′CL except for the “at least five positions” from the second group. Claim 6 then makes the mapping explicit: Light retained positions are L45, L58, L69. Heavy retained positions are H27, H28, H29, H30, H71. This sequence-position mapping is the strongest design-around barrier in the document. Many antibody humanization efforts can vary frameworks broadly, but if the claim requires specific Kabat positions to match mouse residues and also fixes the framework templates by SEQ ID, it becomes harder to “escape” while maintaining CDR identity and α4 binding. Claim 7–8: CDR3 details at H98 (phenylalanine optional vs required) Claim 7 states the humanized CDRs are identical to mouse 21-6’s, except CDR3 of heavy chain may or may not have a phenylalanine at H98. Claim 8 requires phenylalanine at H98. This creates two tiers of scope: claim 1 and some intermediates cover both H98 variants, claim 6/14/16-based descendants can be constrained to the F-containing variant depending on which claim path is used. Claims 9–16: Enumeration of mature variable sequences (La/Lb/Ha/Hb/Hc and variants) Claims 9–13 define mature light and heavy variable amino acid sequences by FIG references and SEQ ID numbers: Light: La (SEQ ID NO: 7) and Lb (SEQ ID NO: 8) Heavy: Ha (SEQ ID NO: 11), Hb (SEQ ID NO: 12), Hc (SEQ ID NO: 13) Then claims 14–16 specify heavy sequence identity within the “La” path: Ha / Hb / Hc in combinations with claim 9 (La). From an enforcement perspective, these are “hard” sequences. If a competitor uses any alternative heavy variant sequence outside those enumerated, it can avoid literal infringement where those specific claim-dependent branches apply. Do the later claims broaden beyond the variable sequences? Claim 17: Antigen-specific binding fragment Claim 17 covers binding fragments of the humanized immunoglobulin of claim 14 or 16. Such fragments could include Fab, F(ab’)2, scFv variants (depending on definitions in the specification), and any fragment with the claimed binding specificity. Because claim 17 is dependent on a specific heavy sequence (claim 14 or 16), the fragment scope likely stays tethered to those variable-domain sequences. Claims 18–20: Constant region effector function switch Claim 18 adds “constant region domain.” Claim 19 requires constant region that is “incapable of complement fixation and antibody dependent cellular toxicity.” Claim 20 requires constant region with complement fixation or ADCC activity. This splits effector function coverage. For infringement analysis, constant region differences do not change variable domains, so both pro- and non-effector formats can be covered if the variable domains match. Claims 21–22: Nucleic acids Claim 21 and 22 cover nucleic acids encoding heavy or light chain variable domains or binding fragments. These are typical enforcement tools against gene constructs used in expression systems. Claim 23: Pharmaceutical compositions Covers compositions with the humanized immunoglobulin (claim 14 or 16) and a carrier. Claims 24–29: Methods (diagnosis, adhesion inhibition, inflammatory disease, multiple sclerosis) Claim 24: detection of α4 integrin in tissue samples via binding complex detection. Claim 25–26: inhibition of leukocyte adhesion to endothelial cells; includes brain cell endothelial targeting. Claim 27–29: treating inflammatory disease including multiple sclerosis, including arrests symptoms. These are downstream therapeutic method claims that generally rise or fall with the underlying antibody infringement. Where is the patent strongest versus vulnerable? Strength: structural specificity at variable-domain level The tight coupling of: CDR sequences copied from mouse 21-6 (SEQ ID NO: 2 and 4), human framework scaffolds defined by SEQ ID (SEQ ID NO: 6 and 10), explicit Kabat positions (L45/L58/L69 and H27/H28/H29/H30/H71), and optionally required H98 phenylalanine creates a “precision humanization” profile. If a product matches those sequence constraints, it is difficult to argue it is outside literal scope. Vulnerability: “at least one position” language appears earlier than the most explicit position list Even though claims 4–6 become explicit, the earlier broad claim 1 is framed as “at least one position selected from” two groups. That broad language can cut both ways: For enforcement, it could expand coverage to variants that keep only one required mouse residue at a specified Kabat position, potentially making it harder to design around. For validity challenges, it can also raise questions about whether the claim scope captures antibodies that were not enabled or not clearly supported by the disclosure, depending on the support for each combination path (e.g., if only certain variants were exemplified). Vulnerability: affinity band as a functional limitation may invite measurement disputes The “about 10^7 M^-1 to five-times binding affinity of mouse 21-6” introduces dependence on how affinity is measured (assay conditions, formats). In litigation, that can become a technical battleground. It also creates a potential design-around: a competitor could aim for affinity outside the upper bound while retaining similar CDRs. What competitive design-around strategies are most plausible against this specific claim set? 1) Alter heavy framework substitutions outside the required Kabat-position set while staying close to RE1 and 21/28′CL templates. Because claim 6 requires specific position identities (L45, L58, L69 in the light; H27, H28, H29, H30, H71 in heavy), omitting any one of those while attempting to preserve binding risks moving outside the more specific descendant claims. 2) Use a different heavy CDR3 H98 status and/or a heavy variable sequence not matching Ha/Hb/Hc. Claims 8 and 14–16 create sub-scope around H98 phenylalanine and specific heavy sequence enumerations. 3) Reformat constant region only if effector-function coverage is what is targeted. Constant region differences likely do not avoid infringement for variable domains, but they can affect which dependent claim branches apply (complement fixation/ADCC capable vs incapable). 4) Create a fragment format (Fab/F(ab’)2/scFv) that does not embody the claimed “mature variable sequences” tied to the enumerated La/Lb and Ha/Hb/Hc. Claim 17 is tethered to claim 14 or 16, so avoiding those exact variable sequences is the cleanest escape from fragment claims. How does this map onto known α4 integrin therapeutic antibody paradigms? This patent sits in the broader universe of α4 integrin (CD49d) biology, where therapeutic antibodies were developed to modulate leukocyte trafficking and treat inflammatory diseases including multiple sclerosis. Within that paradigm, the US 5,840,299 approach is classic for antibody engineering patents: define CDR donor (mouse 21-6), humanize frameworks using specific human templates, preserve key framework residues at selected Kabat positions, then claim composition and method use. The claim set’s strongest identifying markers are not “α4 integrin” itself; it is the combination of: mouse 21-6 CDR identity, plus human framework sequence IDs, plus explicit Kabat position constraints, plus enumerated matured variable sequences. That combination is what makes this patent more product-specific than a general “humanized anti-CD49d” claim. Patent landscape implications for US filings and freedom-to-operate What matters for FTO analysts For product teams developing anti-α4 integrin antibodies in the US, the relevant infringement question is whether the candidate antibody: uses the mouse 21-6 CDRs (SEQ ID NO: 2 and 4) in full, uses the specified human framework templates (light RE1, heavy 21/28′CL) with the specific Kabat-position matches, and has binding affinity within the stated band. Even if a candidate binds α4 integrin, mismatch at any of these layers can place it outside literal coverage of the narrower dependent claim set (claims 2–6, 8, 9–16). Where competing patents often cluster In this technology space, competitors typically attempt to: change CDR sequences while retaining α4 binding (escape the “mouse 21-6 CDR identity” requirement), use different human framework scaffolds or alter the specified Kabat positions (escape the framework-position constraints), add different binding formats or hinge/constant region architectures (usually not enough to avoid variable-domain-tethered claims), or pursue different epitope angles that require different CDRs (stronger escape). Because this patent claims down to explicit mature variable sequences (La/Lb and Ha/Hb/Hc), competitors are incentivized to design new sequences rather than merely “conservative” humanization. Key Takeaways US 5,840,299 claims a humanized anti–α4 integrin IgG family built from mouse 21-6 CDRs (light SEQ ID NO: 2; heavy SEQ ID NO: 4) combined with fixed human framework scaffolds (light RE1 SEQ ID NO: 6; heavy 21/28′CL SEQ ID NO: 10). The broadest anchor is claim 1: α4 integrin binding with an affinity band (about 10^7 M^-1 to about 5x mouse 21-6) plus the CDR identity. The narrowest enforceability leverage is claim 6 (explicit Kabat positions): light L45/L58/L69 and heavy H27/H28/H29/H30/H71, plus heavy CDR3 H98 phenylalanine (claims 7–8) and enumerated mature sequences (La/Lb; Ha/Hb/Hc). Downstream claims (binding fragments, nucleic acids, compositions, diagnostic and therapeutic methods including multiple sclerosis) generally track the underlying variable-domain infringement. The clearest design-around is not “changing constant regions,” but changing at least one of: CDR donor identity, framework template identity, required Kabat-position matches, or the enumerated matured heavy/light variable sequences tied to dependent claims. FAQs 1) What is the central biological target claimed in US 5,840,299? Alpha-4 integrin (CD49d), bound by the humanized immunoglobulin. 2) What role do the mouse 21-6 sequences play in claim scope? They define the CDR sequences: light CDRs come from mouse light variable SEQ ID NO: 2 and heavy CDRs come from mouse heavy variable SEQ ID NO: 4. 3) Are the framework scaffolds generic, or fixed? Fixed for the key dependent claims: the light framework is RE1 (SEQ ID NO: 6) and the heavy framework is 21/28′CL (SEQ ID NO: 10), with specific Kabat positions required to match mouse residues. 4) How does the affinity limitation change infringement risk? It adds a functional band requirement in claim 1 (about 10^7 M^-1 to about 5x mouse 21-6), which can exclude antibodies binding α4 integrin with affinity outside that range. 5) Does the patent cover methods for multiple sclerosis? Yes. It claims treating inflammatory disease including multiple sclerosis via administering the pharmaceutical composition comprising the antibody of the specified dependent claims. References [1] United States Patent 5,840,299. “Humanized immunoglobulin specifically binding to alpha-4 integrin.” Claims 1–29. More… ↓ ⤷ Start Trial**

Applicant	Tradename	Biologic Ingredient	Dosage Form	BLA	Approval Date	Patent No.	Expiredate
Genentech, Inc.	AVASTIN	bevacizumab	Injection	125085	February 26, 2004	⤷ Start Trial	2015-11-21
Biogen Inc.	TYSABRI	natalizumab	Injection	125104	November 23, 2004	⤷ Start Trial	2015-11-21
>Applicant	>Tradename	>Biologic Ingredient	>Dosage Form	>BLA	>Approval Date	>Patent No.	>Expiredate

Country	Patent Number	Estimated Expiration
World Intellectual Property Organization (WIPO)	9718838	⤷ Start Trial
World Intellectual Property Organization (WIPO)	9519790	⤷ Start Trial
United States of America	8246958	⤷ Start Trial
United States of America	7435802	⤷ Start Trial
United States of America	2015064177	⤷ Start Trial
>Country	>Patent Number	>Estimated Expiration

Patent: 5,840,299

Summary for Patent: 5,840,299