RNA drugs work only if they reach the right tissue, and muscle is notoriously hard to deliver to. US11525137B2 - "Compositions and methods of treating Facioscapulohumeral muscular dystrophy," issued December 13, 2022 to Avidity Biosciences, Inc. - claims a delivery answer: bolt the silencing RNA onto a muscle-targeting antibody.

The load-bearing limitation is the conjugate architecture, and claim 1 spells it out with unusual specificity. It recites an siRNA-antibody conjugate in which the antibody is not just any homing antibody but an anti-transferrin receptor antibody or antigen binding fragment thereof, conjugated to an siRNA molecule comprising a sense strand and an antisense strand. The antisense strand is constrained to 19-to-30 nucleotides and to a sequence selected from an enumerated list of SEQ ID NOs (72, 76, 126, 131, 132, 134-136, 212, 216, 266, 271, 272, and 274-276). The construct must mediate RNA interference against human DUX4 mRNA and downregulate one or more DUX4-regulated genes. So three things have to coincide for infringement: the transferrin-receptor targeting arm, a sequence-defined antisense strand, and the DUX4-directed activity. Neither the antibody nor the oligonucleotide alone is the invention - the coupling, with these specified parts, is.

"Disclosed herein are polynucleic acid molecules, pharmaceutical compositions, and methods for treating Facioscapulohumeral muscular dystrophy."- U.S. Patent No. 11,525,137 source

The choice of the transferrin receptor as the address is itself a disclosed mechanism worth noting. Dependent claims 2 through 5 walk through the logic: the antibody binds a transferrin receptor on the cell surface of a muscle cell (claim 2), the conjugate is delivered to a muscle cell (claim 3), that muscle cell expresses human DUX4 mRNA (claim 4), and expression of that mRNA is reduced by the conjugate (claim 5). The transferrin receptor is abundant on muscle, which is what makes it a usable doorway for a payload that otherwise will not cross into the tissue.

The dependent claims also disclose the chemistry that keeps the oligonucleotide stable and active. Claim 11 recites at least one 2'-modified nucleotide, at least one modified internucleotide linkage, or an inverted abasic moiety; claim 15 enumerates the specific 2' modifications (2'-O-methyl, 2'-O-methoxyethyl, 2'-deoxy-2'-fluoro, locked nucleic acid, and others); claim 16 names phosphorothioate and phosphorodithioate linkages; and claim 18 adds a 5'-terminal vinylphosphonate. Claim 19 fixes a siRNA-to-antibody ratio of 1 to 4 - the drug-to-antibody ratio familiar from conjugate engineering. These are not decorative limitations; they are the disclosed means by which a naked siRNA is converted into something that survives circulation and engages the RNAi machinery.

The CPC profile captures the hybrid nature precisely: C12N 15/113 (antisense/RNAi oligonucleotides), A61K 47/6849 and A61K 47/6807 (antibody-conjugate chemistry), C07K 16/2881 (the targeting antibody), with A61P 21/00 / A61P 21/06 (muscular-disorder classification) for the indication. This cross-class signature is the fingerprint of a conjugate platform - antibody chemistry plus oligonucleotide chemistry in one claim - and it is the structural reason the filing reads across two prior-art heritages at once.

Why does the coupling decide scope? Because the prior art holds plenty of silencing oligonucleotides and plenty of antibodies. The claimed novelty is the specific conjugate - transferrin-receptor arm, sequence-defined antisense strand, DUX4 activity - that delivers the RNA to muscle. A competitor administering a naked oligonucleotide, using a different targeting receptor, or using an antisense strand outside the enumerated SEQ ID NOs may fall outside claim 1 as written. The enumerated-sequence limitation in particular is a double-edged sword for the holder: it makes the claim hard to design around if a rival lands on the same sequence, but it also means a different effective sequence is, on its face, outside the literal claim.

It is worth slowing down on what the activity limitation in claim 1 actually requires, because it is doing real work. The conjugate must "mediate RNA interference against human DUX4 mRNA" and "mediate downregulation of one or more DUX4 regulated genes." Claim 21 names those downstream readouts - MBD3L2, TRIM43, PRAMEF1, ZSCAN4, KHDC1L, and LEUTX - genes whose expression tracks DUX4 activity. That construction matters for infringement analysis: a defendant could in principle argue non-infringement not only on the structural parts but on whether an accused conjugate produces the claimed downregulation, turning a portion of the dispute into an evidentiary question about molecular activity rather than chemical identity alone.

Claim 6 ties the indication to its molecular cause, reciting that the FSHD at issue is "caused by genetic and epigenetic changes in D4Z4 DNA region of long arm of chromosome 4." That is the established genetic basis of the disorder, and its appearance in a dependent claim shows the patentee anchoring the claim family to a defined genomic etiology rather than a symptom. Claim 9 splits the indication into type 1 and type 2, and claim 13 requires the antisense strand to hybridize to at least eight contiguous bases of the DUX4 target - a minimal-complementarity floor that, combined with the SEQ-ID list in claim 1, fences the sequence space tightly.

The formulation and ratio claims round out the picture of a fully specified product. Claim 10 recites parenteral administration; claim 17 requires three or more 2'-modified nucleotides drawn from 2'-O-methyl and 2'-deoxy-2'-fluoro; claim 20 gives the sense-strand SEQ ID list paired to the antisense list of claim 1. Read together, claims 1 and 11-20 describe not a concept but a manufacturable molecule: defined sequences, defined sugar and backbone modifications, a defined drug-to-antibody ratio, and a defined route. For a teardown, that completeness is the point - the claim's strength and its narrowness are the same property.

For the landscape, antibody-oligonucleotide conjugates represent a convergence neighborhood - the place where the antibody estate (C07K) and the RNAi estate (C12N 15/113) overlap. A freedom-to-operate read for any tissue-targeted RNA therapy has to cover both heritages, because the conjugate claims sit squarely on the seam between them, and because the dependent-claim chemistry (2' modifications, phosphorothioate linkages, defined DAR) is shared vocabulary across the whole field. The teardown takeaway: this is a narrow, deeply specified species claim, not a broad genus - its strength is its precision, and its boundary is the list of parts it names.