MODIFIED AAV CAPSID POLYPEPTIDES FOR TREATMENT OF MUSCULAR DISEASES

Abstract

Described herein is an adeno-associated virus (AAV) capsid polypeptide bonded to a binding peptide including an amino acid sequence RGDX1X2X3X4, with X1 to X4 being independently selected amino acids, for use in treating and/or preventing a muscular disease and/or in muscle regeneration. Also described are polynucleotides, host cells, adeno-associated virus (AAV) capsids, pharmaceutical compositions, uses, and methods related to the AAV capsid polypeptide.

Claims

1. A method for treating and/or preventing a muscular disease and/or for muscle regeneration, the method comprising: contacting a subject with an adeno-associated virus (AAV) capsid polypeptide bonded to a binding peptide comprising an amino acid sequence RGDX1X2X3X4 (SEQ ID NO: 1), with X1 to X4 being independently selected amino acids.

2. The method of claim 1, wherein said binding peptide is inserted into the amino acid sequence of said AAV capsid polypeptide.

3. The method of claim 1, wherein X1, X2, and X3 are independently selected from L, G, V, and A; and X4 being S, V, A, G, or L.

4. The method of claim 1, wherein said binding peptide comprises the amino acid sequence RGDLGLS (SEQ ID NO:2) or RGDAVGV (SEQ ID NO:3).

5. method of claim 2, wherein the insertion site of the binding peptide corresponds to amino acid 588 or 589 of the AAV9 capsid polypeptide.

6. The method of claim 1, wherein said AAV capsid polypeptide further comprises an exchange corresponding to a P504A and/or a G505A amino acid exchange in AAV9.

7. The method of claim 1, wherein said AAV capsid polypeptide is an AAV9 capsid polypeptide.

8. The method of claim 1, wherein said muscle is a striated muscle.

9. The method of claim 1, wherein said muscular disease is a muscular dystrophy, a cardiomyopathy, a myotonia, a muscular atrophy, a myoclonus dystonia, a mitochondrial myopathy, a rhabdomyolysis, a fibromyalgia, and/or a myofascial pain syndrome.

10. (canceled)

11. The method of claim 1, wherein contacting said subject with said AAV capsid polypeptide comprises contacting said subject with a host cell comprising said AAV capsid polypeptide.

12. The method of claim 1, wherein contacting said subject with said AAV capsid polypeptide comprises contacting said subject with an AAV capsid comprising said AAV capsid polypeptide.

13. The method of claim 1, wherein contacting said subject with said AAV capsid polypeptide comprises contacting said subject with a pharmaceutical composition comprising said AAV capsid polypeptide.

14. (canceled)

15. (canceled)

16. An AAV capsid polypeptide comprising an amino acid sequence RGDX1X2X3X4 (SEQ ID NO: 1), with X1 to X4 being independently selected amino acids, wherein said AAV capsid polypeptide mediates tropism of AAV particles to muscle tissues and/or muscle cells.

17. A polynucleotide encoding the AAV capsid polypeptide of claim 16.

18. The method of claim 1, wherein said binding peptide is covalently bonded to said AAV capsid polypeptide.

19. The method of claim 3, wherein at least one of X2 and X3 is G.

20. The method of claim 4, wherein said binding peptide consists of the amino acid sequence RGDLGLS (SEQ ID NO:2) or RGDAVGV (SEQ ID NO:3).

21. The method of claim 5, wherein the insertion site of the binding peptide corresponds to amino acid 588 of the AAV9 capsid polypeptide.

22. The method of claim 6, wherein said AAV capsid polypeptide comprises amino acid exchanges corresponding to P504A and G505A amino acid exchanges in AAV9.

23. The method of claim 8, wherein said muscle is the heart or a skeletal muscle or diaphragm.

Description

EXAMPLE 1: METHODS

[0069] For this study a self-complementary AAV genome was used carrying a CMV promoter-driven EYFP transgene flanked by a BGH (bovine growth hormone) poly-A signal. A 15 nt-long barcode is positioned in the 3-prime UTR and will therefore be transcribed allowing the tracking of the corresponding AAV on the RNA and DNA level. During AAV production, HEK293T cells are transfected by an Adenoviral helper plasmid delivering essential genes for the replication of the AAV. Additionally, a plasmid for the expression of rep2 (rep gene of AAV2) and the respective cap gene is needed as well as a plasmid carrying the ITR-flanked (inverted terminal repeats), barcoded AAV genome. By selecting capsid/barcode pairs for the production, a packaging of a defined barcoded genome into a synthetic capsid was allowed. Each AAV variant was subsequently purified with an iodixanol gradient and viral titers were determined. Equimolar amounts of all variants were pooled to generate a barcoded AAV library, comprising AAV9_P1 and important benchmark variants (Varadi et al. (2012), Yu et al. (2009), Choudhury et al. (2016), Yang et al. (2009), and Adachi et al. (2014)).

[0070] For the biodistribution evaluation of the library, six female C57BL/6 mice were injected with 1×10.sup.12 vg (vector genomes) per mouse via the tail vein. After one week, abdominal aorta (Aa), thoracic aorta (At), brain (B), blood cells (BlC), colon (C), diaphragm (Di), duodenum (Du), eye (Eye), brown fat (FatB), white fat (FatW), heart (H), inner ear (I), kidney (K), liver (Li), lung (Lu), ovaries (0), pancreas (P), quadriceps femoris (QF), spleen (S) and stomach (St) were harvested. Subsequent DNA and RNA isolation was performed to enable, firstly, cDNA synthesis and, secondly, barcode enrichment by PCR. The following next generation sequencing determined the proportion of each variant in every tissue of the six mice based on the occurrence of the barcode sequence. Furthermore q(uantitative)PCRs for the detection of viral genomes in the individual organs provided all necessary values for the multi-step normalization procedure which allows a comparison between different organs for the same AAV variant. Final results were acquired by using a tailored Python script (modified from Marsic et al., 2015).

EXAMPLE 2: RESULTS

2.1 First Screening with AAV9_P1

[0071] The collected data allowed a tissue specificity analysis of one variant over all tissues as well as an efficiency analysis of all variants within one tissue.

[0072] Regarding organ specificity (FIG. 1 and Table 2), AAV9_P1 demonstrated a strong muscle tropism with 74% of the virus ending up in the analyzed muscle tissues diaphragm, heart and quadriceps femoris. In contrast, when using the parental virus AAV9 wt, only 10% of corresponding transcripts were found in the muscle tissues, whereas 50% could be observed in the major off-target of AAV, the liver. Based on these data, a 13-fold reduction was detected in the liver for AAV9_P1. With AAV6 wt, another well-known AAV for muscle targeting, again only 10% of the total viral transcripts could be found in the muscles versus 84% in the liver. Previously identified muscle-specific candidates, AAVpo1wt and AAVpo1_A1, showed a strong muscle specificity with 51% and 60%, respectively, and minor off-targeting in the liver (3% and 6%, respectively).

[0073] A direct comparison of the efficiency of the capsids in the muscle tissues illustrates the superiority of AAV9_P1 (FIG. 2). Out of all viral transcripts expressed from the 82 variants in the library, 54%, 45% and 13% belong to AAV9_P1 in the diaphragm, quadriceps femoris and heart, respectively. A comparison with the benchmarks revealed major improvements in terms of entering the target tissue and transcribing the barcoded transgene (Table 1), as well as muscle specificity (Table 2).

TABLE-US-00002 TABLE 1 AAV9_P1 efficiency comparison. AAV9_P1 demonstrates major efficiency improvements over other AAV capsids that are currently considered as benchmarks for muscle delivery. Values depicted are fold-changes of AAV9_P1 transduction as compared to the indicated variants. Variant Diaphragm Heart Quadriceps f. AAV9wt 10.6 1.6 7.2 AAVpo1wt 18.0 10.8 18.6 AAVpo1_A1 21.6 9.2 25.5 AAV6wt 78.4 6.8 22.4

TABLE-US-00003 TABLE 2 AAV9_P1 specificity comparison. Analyzed was the muscle specificity of AAV9_P1, AAV9wt, AAVpo1wt, AAVpo1_A1 and AAV6wt. Depicted values are percentages of detected transcripts compared to all analyzed tissues. Variant Muscle tissues [%] Liver [%] AAV9_P1 74.6 3.8 AAV9wt 10.0 50.3 AAVpo1wt 51.7 2.9 AAVpo1_A1 60.3 6.2 AAV6wt 10.3 83.9

2.2 Second Screening with AAV9_P1

[0074] To solidify the enhanced muscle specificity and efficiency of AAV9_P1, another NGS-based screening was carried out following the exact same workflow as described in the first screening but now also including three published AAV variants reported to be superior in muscle tissue, AAVM41 (Yang et al. (2009)), AAVB1 (Choudhury et al. (2016)) and AAV2_MTP (Yu et al. (2009)). Furthermore two additional peptide insertion variants were tested, AAV9_P3 and AAV9_K3 (Varadi et al. (2012)), which carry a similar peptide motif like AAV9_P1, offering the possibility to understand the role of the peptide itself for the improved performance in muscle.

[0075] For generation of the second barcoded AAV library, the first library was expanded by adding 75 additional variants including the benchmark variants. Six female C57BL/6 mice were injected with 1.57×10.sup.12 vg/mouse via the tail vein and kept for one week before harvesting the aorta (A), biceps (Bi), colon (C), diaphragm (Di), duodenum (Du), eye (Eye), brown fat tissue (FatB), white fat tissue (FatW), heart (H), inner ear (I), kidney (K), liver (Li), lung (Lu), ovaries (O), pancreas (P), quadriceps femoris (QF) and stomach (St). From the spleen and the lymph nodes, CD3+, CD19+, CD11b+ and CD11c+ cells were isolated by MACS to study the ability of AAVs to enter immune cells. RNA and DNA was extracted from all tissues, and samples were processed and prepared for the NGS-based analysis. The data revealed clear outliers for mouse 3 and 4 for AAV9_P1; therefore, both were excluded. The following graphs show the mean values of mouse 1, 2, 5 and 6 and the corresponding SD.

[0076] As observed previously in the first screening, AAV9_P1 demonstrated an enhanced muscle tropism (66%) as well as detargeting from the liver (5.5%) (FIG. 3). Regarding the wild-type AAVs, only AAVpo1wt displays inherent muscle specificity. The published benchmark variants for muscle transduction, AAVM41, AAVB1 and AAV2_MTP, are clearly inferior to AAV9_P1 (FIG. 4). Only AAVM41 shows a tendency to preferably transduce the four analyzed muscle tissues. AAVB1 and AAV2_MTP predominantly end up in the liver. Particularly informative is a comparison of AAV9_P1 to AAV9_K3 and AAV9_P3. AAV9_K3, for instance, differs in only two amino acids as compared to the P1 peptide but is inserted at position 589, whereas P1 was integrated at position 588. While these seem to be minor changes, the biodistribution analysis of AAV9_K3 reveals a strong tropism for the liver which even exceeds the one seen for AAV9 wt. The peptide motif of AAV9_P3 varies in four amino acids positions but this time a muscle tropism can be seen though to a weaker extent than with AAV9_P1 (FIG. 3). Interestingly, AAV9LD, a variant with two point mutations as compared to AAV9 wt, is massively detargeted from the liver as originally reported (Adachi et al., 2014). However, at the same time, it is enriched in muscle tissues (FIG. 4).

[0077] An analysis of the proportions of each variant within one tissue gave similar results as in the first screening (FIG. 5). AAV9_P1 is the most efficient vector in this study in the diaphragm, biceps, quadriceps femoris and, to a lesser extent, in the heart. A more detailed overview of the differences is seen in Tables 3 and 4.

TABLE-US-00004 TABLE 1 AAV9_P1 efficiency comparison. Efficiency of AAV9_P1 in comparison to AAV9_P3, AAV9wt, AAV9LD, AAV9_K3, AAVB1, AAVpo1wt, AAVpo1_A1, AAVM41, AAV2_MTP and AAV6wt. Values are fold-differences of AAV9_P1 transduction as compared to the indicated variants. Variant Diaphragm Heart Quadriceps f. Biceps AAV9_P3 6.1 2.8 3.1 3.3 AAV9wt 10.1 1.6 5.5 7.0 AAV9LD 16.1 5.0 8.5 10.5 AAV9_K3 31.3 6.0 21.4 21.0 AAVB1 35.6 4.9 42.3 64.5 AAVpo1wt 36.7 17.5 26.8 30.5 AAVpo1_A1 46.1 14.5 38.9 42.3 AAVM41 109.9 14.8 75.7 80.7 AAV2_MTP 543.2 63.6 1135.4 1409.5 AAV6wt 248.2 16.1 146.5 195.6

TABLE-US-00005 TABLE 2 AAV9_P1 specificity comparison. Muscle specificity of AAV9_P1, AAV9_P3, AAV9wt, AAV9LD, AAV9_K3, AAVB1, AAVpo1wt, AAVpo1_A1, AAVM41, AAV2_MTP and AAV6wt. Depicted values are percentages of detected transcripts as compared to all analyzed tissues. Variant Muscle tissues [%] Liver [%] AAV9_P1 66.6 5.5 AAV9_P3 29.6 20.8 AAV9wt 5.3 52.0 AAV9LD 39.4 5.2 AAV9_K3 7.5 77.3 AAVB1 1.4 86.1 AAVpo1wt 48.4 3.6 AAVpo1_A1 51.5 1.5 AAVM41 31.7 47.3 AAV2_MTP 2.9 88.3 AAV6wt 2.2 38.8

EXAMPLE 3: MUSCLE VALIDATION OF AAV9_P1

[0078] To confirm the muscle specificity of AAV9_P1, a follow-up study was conducted using female C57BL/6 mice which were injected with 1×10.sup.11 vg/mouse via the tail vein. Besides AAV9_P1, the lead candidates from a previous screening, AAVpo1wt and AAVpo1_A1, were used as well as the in muscle tissue highly efficient serotype, AAV9 wt. This time, all AAV variants were injected individually into three mice per group. Potential interfering effects originating from receptor competition in a library context are thereby eliminated, allowing to evaluate AAV9_P1 performance. Mice were kept for one week after injection before diaphragm, heart, liver and quadriceps femoris were collected for qPCR analysis against the viral transcripts, using POLR2A as a housekeeper (FIG. 6).

[0079] The evaluation of EYFP relative quantities for AAV9_P1 showed a superior increase compared to AAV9 wt in diaphragm (55-fold), quadriceps femoris (17-fold) and heart (11-fold) while being 9-fold less abundant in the liver. In addition the peptide insertion variant outcompetes AAVpo1wt and AAVpo1_A1 in all muscle tissues. On the other hand both porcine variants exhibit lower cDNA levels in the liver which is however not enough to outweigh their major disadvantage in terms of efficiency.

EXAMPLE 4: IN VIVO ENRICHMENT OF TARGETING SEQUENCES

[0080] An AAV9-library was constructed encoding capsid polypeptides with the sequence of SEQ ID NO:1 inserted between amino acids 588 and 589, in which positions X.sup.1 to X.sup.4 were completely randomized and in which the sequence was flanked by the amino acids S/RG N-terminally and A C-terminally. Virus particles comprising the library were administered into mice via the tail vein, and re-isolated from skeletal muscle, heart, and diaphragm. After the first round of selection, a sequence space represented by the sequence logo of FIG. 11A was obtained, while after the second round of selection, the sequence space represented by the sequence logo of FIG. 11B was retrieved, showing an enrichment of specific amino acids.

EXAMPLE 5: TISSUE TROPISM AND BIODISTRIBUTION OF AAV VARIANTS

[0081] The tissue tropism of AAV9S1_P1 (having a capsid with the amino acid sequence of SEQ ID NO:30) and AAVS10_P1 (having a capsid with the amino acid sequence of SEQ ID NO:28) was compared to AAV9 wt as described herein above in Example 1 for further organs, using 2.5E11 vg per mouse; results are shown in FIG. 12A-C. Further, biodistribution of AAV9P1, AAV9S1, and AAV9S10 in liver was compared to that of AAV9 as described above in Example 1; results are shown in FIG. 12D.

NON-STANDARD LITERATURE CITED

[0082] Adachi et al. (2014), Nat. Commun., vol. 5, no. June 2011, p. 3075. [0083] Choudhury et al. (2016), Mol. Ther., vol. 24, no. 7, pp. 1247-1257. [0084] Grimm et al. (2006), Nature 441, 537-541. [0085] Grimm et al. (2010), J. Clin. Invest. 120, 3106-3119. [0086] Kienle, PhD Thesis, Ruprecht-Karls-Universität Heidelberg, 2014. [0087] Kunze et al. (2018), Glia, vol. 66, no. 2, pp. 413-427. [0088] Lazaro et al. (2017). bioRxiv doi: 10.1101/101188. [0089] Marsic et al. (2015), Mol. Ther. Methods Clin. Dev., vol. 2, no. September, p. 15041. [0090] Michelfelder et al. (2009), PLOS One 4(4):e5122 [0091] Takahashi & Yamanaka (2006), Cell 126, 663-676. [0092] Varadi et al. (2012), Gene Ther., vol. 19, no. 8, pp. 800-809. [0093] Yang et al. (2009), Proc. Natl. Acad. Sci. U.S.A, vol. 106, no. 10, pp. 3946-51. [0094] Yu et al. (2009), Gene Ther., vol. 16, no. 8, pp. 953-962.