» 

Adeno-Associated Virus — An Introduction

Adeno-Associated Virus (AAV) was discovered by Dr. Bob Atchison (1) of Pittsburgh University and Dr. Wallace Rowe (2) of NIH as contaminants of adenovirus preparations in 1965. Upon closer examination, the researchers realized that replication of the AAV viral particles was only possible in the presence of Adenovirus, leading to its name and classification in the Dependovirus genus of the family Paroviridae (3).

Physical Properties

Adeno-Associated Virus serotype 2 (AAV2) is the most extensively studied and serves as a "prototype" for the majority of the AAV family. Most of the following information on the AAV genome and structure is therefore derived from knowledge and studies of AAV2 (4).

The AAV genome is a single-stranded DNA molecules about 4.7 kb in length with two 145 nucleotide terminal inverted repeats (ITR) (Figure 1) (5). The ITR has the ability to form Watson-Crick base pairing with itself, which takes the form of a T-shaped hairpin structure containing cis elements required for replication and packaging (6). The AAV genome itself only contains two genes, REP and CAP. Together, they encode for three structural proteins (VP1, VP2 and VP3) and four non-structural proteins which are necessary for replication (Rep78, Rep68, Rep52 and Rep40). The Rep proteins all exhibit helicase activity, which are necessary for integration into host cell chromosomes, as well as being essential for capsid formation initiation (4). Three viral promoters have been identified in the viral genome as p5 and p19, which direct synthesis of the viral regulatory proteins, and p40, which promotes the transcription of the CAP gene(7).

Figure 1 – Adeno-Associated Virus Genome map. The positions of the three promoters as well as the seven protein coding regions of the AAV have been highlighted.

The virus is in the form of an icosahedral non-enveloped particle with an encapsulated ssDNA genome. At a size of roughly 22 nm in diameter, AAV is one of the smallest encapsulated viruses discovered (Figure 2). Its capsid is comprised of 60 subunits made up of the three capsid proteins VP1, VP2 and VP3 in a 1:1:10 ratio (4). The major difference between the three proteins lies in their N-terminus residues. The VP1 protein contains phospholipidase that is necessary for infectivity while the VP2 and VP3 proteins form the structure of the envelope coat (8).

Figure 2 – Images from http://goo.gl/YRqjvB and http://goo.gl/y8nZNG.

Because it does not encode a polymerase, AAV therefore relies on cellular polymerase activity to replicate its own genome (9). AAV enters a human cell via a clathrin coated vesicle activated through capsid interaction with receptors on the outer cellular surface. It is still unknown how the virus actually reaches the cell nucleus but many different methods and pathways have been proposed (10). Once within the nucleus, the relatively slow shedding of the AAV viral coat begins (11) (12). The presence of a helper virus such as adenovirus is required for wild-type AAV to initiate replication and replicate at a high capacity. Without adenovirus or another similar helper virus to co-infect the host, limited expression of Rep68/Rep78 occurs due to Ying Yang 1 (YY1) repression of the p5 promoter. p5 promoter repression, as a result, inhibits AAV genome replication and gene expression, and initiates AAV chromosome integration (13). AAV specifically integrates into a mapped site, known as the AAVS1, on human chromosome 19. Wild type AAV has been shown to be the only eukaryotic virus to site-specifically integration into the human genome (Figure 3) (14).



Figure 3 – The AAV infection cycle. AAV enters the cell through receptor mediated endocytosis, and is transported to the nucleus via clathrin coated vesicle. Once within the nucleus the virion capsid is shed and the viral genome released, where it can either integrate into the host cell genome (in the absence of a co-infected helper virus) or proceed to lytic cycle (in presence of helper virus).


cartWe offer recombinant AAV with different serotype options, see if your gene of interest is in stock here.


Life Cycle and Pathogenicity

For AAV to re-enter the lytic phase it requires co-infection with a helper virus that will activate AAV genome replication. These viruses include adenovirus and even larger species such as Herpes Simplex Virus. Specific adenovirus genes such as E1a, E1b55k, E2a, E4orf6 and associated viral protein have been identified to provide known helper functions for AAV (15) (16). The functions of these genes vary: E1a activates the adenovirus promoters of E1b55k, E2a, E4orf6, and associated viral protein and bind to the YY1 repressor to relieve AAV p5 promoter repression. This consequently allows p5 to promote the expression of large amounts of Rep68/Rep78. E2a stimulates processivity of AAV replication in vitro, and the E1b55k and E4orf6 proteins work to promote AAV replication and second strand DNA synthesis (17) (18) (19). The viral associated proteins stimulate AAV expression through the blocking of phosphorylation of EIF2alpha translation factor, which when phosphorylated blocks AAV gene expression. (20). AAV replication can also be naturally stimulated through cellular stress caused by UV radiation and hydroxyurea, although the exact method of why this occurs remains unknown (21).

Adeno-Associated virus has been found to be present in nearly 80% of the human population and that 60% of the global population possess neutralizing antibodies to AAV serotypes 1, 2, 3, and 5 at the age of 10 (4). It is believed that since AAV cannot replicate without a helper virus in tandem, it has evolved to be almost undetectable in its human host. To date, no known illness is associated with AAV. Its low immunogenicity has been demonstrated in experimental settings where even though 96% of subjects tested positive for AAV, no detrimental effects or lytic infection was observed until they were exposed to adenovirus, and even then, the results were nearly negligible (4).


cartRequire high-titer AAV production? Check out our custom AAV cloning, packaging purification service.


Modern Day System: Recombinant AAV

Modern Adeno-Associated virus adaptations, that enhance and improve its function as a viral vector, have brought AAV to the forefront in gene therapy and regenerative medicine. In 1984, recombinant AAV (rAAV) was used for the first time by replacing the viral genome with a transgene which was then transfected into cells that were infected with adenovirus (3). The two viruses were harvested together as a pool, but the adenovirus was then killed by heat inactivation. This early day method proved that recombinant AAV was possible. Modern day recombinant AAV (rAAV) vectors are generated by replacing all of the viral genome between the ITRs with a transcriptional cassette of less than approximately 5 kb in length (Figure 4). The resulting construct is then co-expressed with two other plasmids: 1) a plasmid that provides the Rep and Cap genes in trans (separate from the ITR/Transgene cassette) and 2) a plasmid that harbours the adenoviral helper genes. HEK293 cells are used as the packaging cell line since they provides the E1a protein in trans as well. By modifying the Rep and Cap genes, scientists can control the serotype to guide the recombinant AAV infection towards certain tissues. This 3-plasmid co-transfection system liberates the need for adenovirus in the production, which greatly simplifies the purification process. Furthermore, since the recombinant AAV lacks the REP gene and its cis-active intercistronic expression element (IEE), both of which are required for site specific integration, recombinant AAV persists as extra-chromosomal element in the cell (3) (22).

Figure 4 – Recombinant AAV can be produced by removing the AAV genes between the ITR regions and replacing them with the transgene of choice. rAAV particles can be packaged when the Rep/Cap plasmid, Ad-Helper plasmid and the transgene plasmid are co-transfected into HEK293 cells.


cartOur recombinant AAVs are packaged using the 3-plasmid co-transfection system, learn more about the different promoters and reporter gene options we offer here.


Tropism and Serotypes

Because of its exhibition of natural tropism toward certain cell or tissue types, Adeno-Associated virus has garnered considerable attention. The AAV serotypes were discovered from humans and other animals: AAV2, AAV3, AAV5, and AAV6 were each discovered from human cells while the other serotypes AAV1, AAV4, AAV7, AAV8, AAV9, AAV10, and AAV11 were found in non-primate species (24) (25). Differences between cell surface receptors of the AAV serotype determine its tropism. In addition to cell surface receptors, tropism is also believed to be affected by cellular uptake, intracellular processing, nuclear delivery of the vector genome, uncoating, and second strand DNA synthesis (23). After the discovery that the capsid surface proteins determines the tropism, scientists have genetically modified the viral capsid, and generated mosaic vectors to create chimeric virions by swapping domain’s or amino-acids between serotypes. (26) (27). This has allowed researchers to specifically target cells with certain serotypes to effectively transduce and express genes in a localized area.

AAV2 is the most widely studied AAV serotype and therefore a majority of all information regarding AAV structure, genetic makeup and lifecycle has been discovered while studying this particular serotype. AAV2 can naturally infect a wide range of tissues (28), but it has preference for the following tissue types: skeletal muscles (29), neurons (30), vascular smooth muscle cells (31) and hepatocytes (32). The primary receptor for AAV2 is the heparan sulfate protoglycan (HSPG), while the fibroblast growth factor receptor 1 (FGFR-1) and avBs integrin may play a role as co-receptors and are utilized for endocytosis. HSPG mainly functions as the primary receptor, which initiates the endocytosis of the AAV virion. (33) (34) (35). Having a great potential in leading current gene therapy, AAV Serotype 2 has been used in cancer treatments to kill a diverse group of cancer cells without showing any effect on healthy cells (36).

To date, a total of 11 serotypes of AAV have been described and each with its own unique traits and tropisms. AAV1 and AAV5 have been studied and found to be very similar in which cells they transduce, although they do differ. AAV5 in particular is very important in that it transduces astrocytes very effectively, while AAV1 is effective in retinal transduction as well as transducing heart and lung tissue (37). AAV6 has been found to be highly effective at transduction of airway epithelial cells. In addition, AAV6 exhibits lower immunogenicity compared to AAV2, which may provide a significant advantage in treatment of chronic disorders such as cystic fibrosis (38) (39) (27). Both AAV7 and AAV8 are known for transduction of hepatocytes. AAV8 has a 10 to 100 fold increased transduction rate in liver cells as compared to AAV7. AAV7 on the other hand also transduces murine skeletal muscle cells effectively - to the same extent as AAV1. AAV4 has tropism for heart and lung cells, being mainly localized in the chest cavity of the host after infection. Another AAV serotype that targets heart tissue is AAV9 (40). The Table 1 below summarizes the natural tropisms and recommended applications of each AAV serotype:


Table 1 — Table of AAV Serotypes and their respective Tropisms.

AAV Serotype CNS/Retina Heart Lung Liver Skeletal Muscle
AAV1 X X X X
AAV2 X X X
AAV3 X X X
AAV4 X X
AAV5 X X
AAV6 X X X X
AAV7 X X X
AAV8 X X X
AAV9 X X X X X

cartUnsure about which serotype to use? Pilot your serotype selection studies with our AAV serotype Starter kit here.


Advantages and Disadvantages of Using AAV as a Viral Vector

Adeno-Associated Virus is a very promising viral vector, and an important candidate for highly efficient gene editing for a variety of reasons. The advantage to using adeno-associated virus over other viral vectors is its ability to infect both dividing and quiescent cells, allowing genetic material to be delivered to a highly diverse range of cell types. Wild Type AAV can integrate into host cells site specifically, which makes it highly predictable as they feature a reliable insertion pattern. Although modern day recombinant AAV has been modified to not insert into the host genome, it can persist in cells in an extra-chromosomal form for a long period of time. AAV brings long term expression in non-dividing cells, causes almost negligible pathogenecity and exerts a very mild immune response, all of which are favourable for use in gene therapy.

Although AAV does present many advantages, the use of the virus is not without some disadvantages. AAV’s major drawback is its very small genome size compared to other existing viral vectors (Table 2). For instance, compared to adenovirus or lentivirus, AAV is 5 to 10X smaller in genome size. The smaller genome size restricts the size of the gene that can be inserted to less than 4.5 kb in length.


Table 2 - Comparison between Lentivirus, Adenovirus and AAV viral vectors.

Features Lentivirus Adenovirus AAV
Packaging Capacity 5Kb 8-9Kb 3.4Kb
Efficiency *** **** ***
Cell Type Most Dividing/Non-Dividing Cells Most Dividing/Non-Dividing Cells and High Transduction Rate Towards Primary Cells All Cell Types
Integrating Yes No 90% Not, 10% May Integrate
Immune Response *** ***** **

cartIn addition to AAV, we also offer recombinant adenovirus and lentivirus, check out our collections!


Clinical Trials

Viral-vectors based on Adeno-Associated Virus have been used in over 117 clinical trials worldwide to date, including a number of promising trials for Cystic Fibrosis, Hemophilia, congestive heart failure, lipoprotein lipase deficiency, and Parkinson's disease (41). Two currently ongoing clinical trials as well as an approved gene therapy breakthrough treatment are described below.


Heart Failure

The FDA has approved Mydicar (Celladon Corporation), a “breakthrough therapy” designation for reducing hospitalizations for heart failure in neutralizing antibody (NAb)-negative patients with New York Heart Association class III or IV chronic heart failure. Mydicar is being developed as a first-in-class therapy for patients with chronic heart failure due to systolic dysfunction. It uses genetic enzyme-replacement therapy to correct deficiencies in the enzyme SERCA2a that result in inadequate pumping of the heart. Mydicar transfers the SERCA2a gene directly into cardiac muscle cells using a nonpathogenic recombinant adeno-associated virus. In the phase 2a CUPID 1 trial, a single intracoronary infusion of high-dose Mydicar in patients with advanced heart failure due to systolic dysfunction reduced heart failure-related hospitalizations and improved patients’ symptoms, quality of life, and key markers of cardiac function that were predictive of survival (42).


Cystic Fibrosis

AAV has been used in trials for cystic fibrosis (CF) treatment since it is a monogenic disease. In cystic fibrosis, the CF transmembrane regulator (CFTR) is inactivated by a rare autosomal mutation, which directly leads to problems with transmembrane electric potential and accumulation of fluid within the lungs. This has consequences such as increased risk of pulmonary infection, loss of regular ciliary activity, and other problems including the loss of pancreatic functions. Thirteen protocols have so far been approved for phase I and phase II clinical trials with AAV vectors. The trials involve AAV vectors with an inserted copy of human CFTR cDNA and delivery to the lungs via a bronchoscope or aerosol. The AAV administration caused a very minor immune response and the trial’s outcome was measured through improvement in pulmonary function. Although changes were not statistically significant in the first trial, in patients who had the vector instilled in the maxillary sinus, there was an in increase in interleukin-10 levels which has a role in anti-inflammation. This has been seen as very promising, and further trials are ongoing. The major challenge to this treatment is the uptake of the AAV dosage in the lungs, which is currently poor due to the method of delivery. In addition, those cells are shed quite rapidly, which poses another barrier (43).


Hemophilia

AAV gene therapy is also being applied to hemophilia. Hemophilia used to be a fatal illness but is currently controlled with modern medicine. This disease prevents blood clots from forming due to a deficiency in factors VIII (Hemophilia A) or IX (Hemophilia B). AAV was chosen as a potential candidate for gene therapy treatment since it can encode all of the IX gene as well as its regulatory region. Early studies showed that the vector could treat mice and dogs successfully. After initial trials it was discovered that although no toxicity was detected with the use of AAV, the IX gene function could not be detected either after injection into human muscles. The vector was then targeted to the liver, and administered to the patients in increasing doses for sufficient IX factor expression. Unfortunately, toxic immunogenic response caused by AAV capsid proteins accompanied the IX factor expression. As a result, researchers are currently in need of a new AAV capsid that elicits low immunogenic response to move forward (44).


cartHave your own idea? Ask us know how we can help you along the way by emailing technical@abmgood.com.


References
  1. Adenovirus associated defective virus particles. Atchinson. s.l. : Science., 1965., Vol. 49.
  2. Studies of small DNA viruses found various adenovirus preparations: physical, biological and immunological characteristics. Rowe, W.P. et al. 1966, Proc. Natl. Acad. Sci. USA, Vol. 55, pp. 1467-1472.
  3. Adeno-associated Virus and the Development of Adeno-associated Virus Vectors: A Historical Perspective. Blacklow, N. R. 2004, Molecular Therapy, Vol. 10, pp. 981-989.
  4. Adeno-Associated Virus Biology. Weitzman, M.D., Linden, R.M. 2011, Methods in Molecular Biology, Vol. 807.
  5. Biology of Adeno-Associated Virus. Berns, K. I., Giraud, C. 1996, Curr Top Microbiol Immunol, Vol. 218, pp. 1-23.
  6. Self-complementarity of terminal sequences within plus or minus strands of adenovirus-associated virus DNA. Koczot FJ, Carter BJ, Garon CF, Rose JA. 1973, Proc Natl Acad Sci USA, Vol. 70, pp. 215 - 219.
  7. Processing of adeno-associated virus RNA. Qiu, J., Pintel, D. Front Biosci, Vol. 13, pp. 3101-3115.
  8. The VP1 capsid protien of adeno-associated virus type 2 is carrying a phospholipase A2 domain required for virus infectivity. Girod, A., Wobus, C. E., Zadori, Z., Ried, M., Leike, K., Tijssen, P., Kleinschmidt, J. A., and Hallek, M. J Gen Virol, Vol. 83, pp. 973-978.
  9. Parvovirus replication. KI, Berns. 1990, Microbiol Rev , Vol. 54, pp. 316-329.
  10. Infectious entry pathway of adeno-associated virus and adeno-associated virus vectors. Bartlett JS, Wilcher R, Samulski RJ. 2000, J Virol, Vol. 74, pp. 2777-2785.
  11. Intracellular trafficking of adeno-associated viral vectors. Ding, W., Zhang, L., Yan, Z., Engelhardt, J.F. 2005, Gene Ther, Vol. 12, pp. 873-880.
  12. Adeno-associated virus mediated gene transfer,. Srivastava, A. J Cell Biochem, Vol. 105, pp. 17-24.
  13. The adeno-associated virus (AAV) Rep protein acts as both a repressor and an activator to regulate AAV transcription during a productive infection,. Pereira, D. J., McCarty, D. M., Muzyczka, N. J Virol, Vol. 71, pp. 1079-1088.
  14. The cryptic life style of adeno-associated virus. Berns, K. I., Linden, R. M. 1995, Bioessays, Vol. 17, pp. 237-245.
  15. Transcriptional repression by YY1, a human GLI-Kruppel-related protein, and relief of repression by adenovirus E1A protein. Shi, Y., Seto, E., Chang, L. S., and Shenk, T. 1991, Cell, Vol. 67, pp. 377-388.
  16. Adeno-associated virus P5 promoter contains an adenovirus E1A-inducible element and a binding site for the major late transcription factor. Chang, L. S., Shi, Y., and Shenk, T. 1989, J Virol, Vol. 63, pp. 3479-3488.
  17. Role of the adenovirus DNA-binding protein in in vitro adeno-associated virus DNA replication. Ward, P., Dean, F. B., O’Donnell, M. E., and Berns, K. I. 1998, J Virol, Vol. 72, pp. 420-427.
  18. Adenovirus E1B 55-Mr polypeptide facilitates timely cytoplasmic accumulation of adeno-associated virus mRNAs. Samulski, R. J., and Shenk, T. 1988, J Virol, Vol. 62, pp. 206-210.
  19. Transduction with recombinant adeno-associated virus for gene therapy is limited by leaidng stran synthesis. Fisher, K. J., Gao, G. P., Weitzman, M. D., DeMatteo, R., Burda, J. F., Wilson, J. M. 1996, J Virol, Vol. 70, pp. 520-532.
  20. Adeno-associated viruses can induce phosphorylation of eIF2alpha via PKR activation, which can be overcome by helper adenovirus type 5 virus associated RNA. Nayak, R., and Pintel, D. J. 2007, J Virol, Vol. 81, pp. 11908–11916.
  21. DNA amplification of adeno-associated virus as a response to cellular genotoxic stress. Yalkinoglu, A. O., Heilbronn, R., Burkle, A., Schlehofer, J. R., zur Hausen, H. 1988, Cancer Res, Vol. 48, pp. 3123-3129.
  22. Gene Therapy Using Adeno-Associated Virus Vectors. Berns K.I., Daya S.,. 4, October 2008, Clin Microbiol Rev., Vol. 21, pp. 583-593.
  23. Adeno-associated Virus Serotypes: Vector Toolkit for Human Gene Therapy. Zhijian Wu, Aravind Asokan, R. Jude Samulski. 2006, Vol. 14, pp. 316-327.
  24. Clades of Adeno-associated viruses are widely disseminated in human tissues. Gao, G., Vandenberghe, L. H., Alvira, M. R. Lu, Y., Calcedo, R., Zhou, X., and Wilson, J. M. 2004, Vol. 78, pp. 6381-6388.
  25. Tailoring the AAV vector capsid for gene therapy. Vandenberghe, L. H., Wilson, J. M., Gao, G. 2009, Gene Ther, Vol. 16, pp. 311-319.
  26. Generation and characterization of chimeric, recombinant AAV vectors. Hauck, B., Chen, L., and Xiao, W. 2003, Mol Ther, Vol. 7, pp. 419-425.
  27. Cross-dressing the virion: the transcapsidation of adeno-associated virus serotypes functionally defines subgroups. Rabinowitz, J. E., Bowles, D. E., Faust, S. M., Ledford, J. G., Cunningham, S. E., Samulski, R. J. 2004, J Virol, Vol. 78, pp. 4421–4432.
  28. Insertional mutagenesis of AAV2 capsid and the production of recombinant virus. Rabinowitz J.E., Xiao W, Samulski R.J. 2, December 1999, Virology, Vol. 265, pp. 274-285.
  29. AAV-mediated factor IX gene transfer to skeletal muscle in patients with severe hemophilia B. Manno CS, Chew AJ, Hutchison S, et al. 8, 2003, Blood , Vol. 101, pp. 2963-2972.
  30. Selective and rapid uptake of adeno-associated virus type 2 in brain. Bartlett JS, Samulski RJ, McCown TJ. 8, 1998, Human Gene Therapy, Vol. 9, pp. 1181-1186.
  31. Adeno-associated virus vector transduction of vascular smooth muscle cells in vivo. Richter M, Iwata A, Nyhuis J, et al. 3, 2000, Physiological Genomics, Vol. 2, pp. 117-127.
  32. Persistent expression of human clotting factor IX from mouse liver after intravenous injection of adeno-associated virus vectors. Koeberl DD, Alexander IE, Halbert CL, Russell DW, Miller AD. 4, 1997, Proceedings of the National Academy of Sciences of the United States of America, Vol. 94, pp. 1426-1431.
  33. Human fibroblast growth factor receptor 1 is a co-receptor for infection by adeno-associated virus 2. Qing K, Mah C, Hansen J, Zhou S, Dwarki V, Srivastava A. 1, 1999, Nature Medecine, Vol. 5, pp. 71-77.
  34. Membrane-Associated Heparan Sulfae Proteoglycan Is a Receptor for Adeno-Associated Virus Type 2 Virions. Summerford C, Samulski RJ. 2, 1998, Journal of Virology, Vol. 72, pp. 1438-1445.
  35. AlphaVbeta5 integrin: a co-receptor for adeno-associated virus type 2 infection. Summerford C, Bartlett JS, Samulski RJ. 1, January 1999, Nature Medecine, Vol. 5, pp. 78-82.
  36. Harmless Virus Kills Some Cancers. State., Penn. June 23, 2005, Science Daily.
  37. Efficient Transduction of Vascular Endothelial Cells with Recombinant Adeno-Associated Virus Serotype 1 and 5 Vectors". Chen S, Kapturczak M, Loiler SA, et al. (February 2005). 2, 2005, Human Gene Therapy , Vol. 16, pp. 235-247.
  38. Novel adeno-associated viruses from rhesus monkeys as vectors for human gene therapy. Gao GP, Alvira MR, Wang L, Calcedo R, Johnston J, Wilson JM. 18, 2002, Proceedings of the National Academy of Sciences of the United States of America, Vol. 99, pp. 11854-11859.
  39. Adeno-Associated Virus Type 6 (AAV6) Vectors Mediate Efficient Transduction of Airway Epithelial Cells in Mouse Lungs Compared to That of AAV2 Vectors. Halbert CL, Allen JM, Miller AD. 14, 2001, Journal of Virology, Vol. 75, pp. 6615-6624.
  40. Analysis of AAV Serotypes 1–9 Mediated Gene Expression and Tropism in Mice After Systemic Injection. Carmela Zincarelli, Stephen Soltys, Giuseppe Rengo, and Joseph E Rabinowitz. 6, 2008, Vol. 16, pp. 1073-1080.
  41. Vectors used in gene therapy clinical trials. The Journal of Gene Medicine Online Library. [Online] 2014. http://www.abedia.com/wiley/vectors.php.
  42. New Drugs/Drug News/New Medical Devices. Pharmacy and Therapeutics. 2014;39(6):392-423.
  43. A phase I study of aerosolized administration of tgAAVCF to cystic fibrosis subjects with mild lung disease. Aitken ML, Moss RB, Waltz DA, Dovey ME, Tonelli MR, McNamara SC, Gibson RL, Ramsey BW, Carter BJ, Reynolds TC. 15, October 10, 2001, Human Gene Therapy, Vol. 12, pp. 1907-1916.
  44. Adenovirus-Associated Virus Vector–Mediated Gene Transfer in Hemophilia B. Amit C. Nathwani, M.B., et al. December 10, 2011, The New England Journal of Medecine, Vol. 365, pp. 2357-2365.