By Kevin E. Noonan

DNA has a unique status among biological molecules, being both a chemical and information comprising the nucleotide sequence that encodes the amino acid sequence of a gene’s cognate protein.  In a recent article in Science, Michael Gottesman and his colleagues at the National Institutes of Health have demonstrated that even this straightforward understanding of information flow involved in gene expression may be more complicated than it has appeared.  These results have implications not only for understanding the underlying biology but in protecting genes with patents.

The genetic code is degenerate, meaning that the number of possible triplet codons (64) is greater than the number of amino acids (20) corresponding to these codons.  This fundamental concept in molecular biology was first postulated by Francis Crick and verified by Marshall Nirenberg and others in the ’50’s and ’60’s.  The physical embodiment of the degeneracy of the code is transfer RNA (tRNA), small RNA molecules having a triplet codon at one "end" of the molecule and a position at the other end to which the corresponding amino acid is linked; the specificity resides in the "charging" enzymes that match tRNA to its proper amino acid.  Most amino acids are encoded by more than one triplet and have more than one tRNA charged with that amino acid.  Leucine has the most (with 6) and methionine and tryptophan are the exceptions that have only one cognate tRNA.

It is known that not all tRNAs for a particular amino acid are produced in the same abundance, either in cells of a particular species or between species.  Thus, early genetic engineering efforts for some proteins involved changing codon choice for DNA from one species (human) for expression in other species (bacteria and yeast).  The underlying assumption, however, is that there was no difference between these codons other than tRNA abundance, and that an amino acid sequence could be encoded by any one of a number of nucleotide sequences representing "synonyms" of the native sequence.

Gottesman’s work changes all that.  His team looked at the gene encoding human MDR1, a plasma membrane, ATP-dependent transporter that recognizes and transports from the cell many structurally-different chemical compounds.  This protein, termed P-gp, is responsible for protecting colon, kidney, and other cells from environmental toxins; forms a part of the blood-brain barrier; and makes certain cancer cells resistant to chemotherapeutic drugs.  Gottesman expressed two different embodiments of the MDR1 gene, where a certain portion of the amino acid sequences of each were encoded by a synonymous stretch of nucleotides.  In this portion of the gene, the encoded amino acid sequence was the same but the codons used in the two embodiments were different:  one of them used rare triplet codons and the other did not.  His results were startling:  not only were the proteins produced at different rates (which could be expected if the rare codon-containing species was translated more slowly), but the phenotype of the two proteins differed as well.  For example, the rare codon-encoding embodiment showed different sensitivity to certain inhibitors.  In addition, the different forms showed different susceptibility to proteolytic enzymes and different binding characteristics with conformation-specific antibodies.  Gottesman interpreted these results to mean that differences in translation speed or efficiency could change the way the protein was folded as it was produced at the ribosome, and the differences in protein folding caused (or at least contributed to) the observed phenotypic differences.

It is unknown how common it is to have these kinds of phenotypic differences result from triplet codon synonyms.  P-pg differs from many other proteins in significant ways.  For example, it is relatively nonspecific in substrate recognition, being able to transport many structurally-different compounds across the cell’s plasma membrane against a concentration gradient.  Other researchers have found changes in substrate specificity associated with mutations in the amino acid sequence (see Choi et al., 1988, Cell 53:519-29 and Safa et al., 1990, Proc. Natl. Acad. Sci. U.S.A. 87:7225-29).  The protein is also internally-duplicated, and other members of the MDR gene family are involved in dramatically different phenotypes despite having very similar primary amino acid sequences (see Van der Bliek et al., 1987, EMBO J. 6:3325-31 and Smith et al., 1994, FEBS Lett. 354:263-66).

Gottesman’s findings have implications for how patent protection is obtained for genes.  Heretofore, it has been common to recite a gene claim in the form of "an isolated nucleic acid encoding a protein having an amino acid sequence identified by SEQ ID NO: X."  This claim is typically supported by one cloned or otherwise identified sequence, and encompasses all nucleotide sequences encoding the amino acid sequence.  Said another way, the claim encompasses all nucleotide sequence synonyms of the encoded amino acid sequence.  When these synonymous sequences were believed to be equivalent (i.e., to encode the same protein with the same phenotypic characteristics), there was no reason not to consider elucidation of the nucleotide sequence and translation into the predicted amino acid sequence as encompassing all these synonymous sequences.  If the effect of synonymous sequences detected by Gottesman is widespread, however, it may be more appropriate to permit patenting of claims to specific synonymous sequences, provided that they are associated with a phenotype that differs in some way from the phenotype of the protein encoded by the patented sequence.  This would be the equivalent of patenting a chemical species encompassed within a prior art genus, where the species has unexpected or unappreciated properties that distinguish it from the properties of the generic compounds.  These results also illustrate one other example of how the human genome has evolved to maximize genetic variability and helps explain the general observation from the Human Genome Project, that there are many fewer human genes than were expected to be found.

    By Kevin E. Noonan

Earlier this week, the National Eye Institutes (NEI) of the National Institutes of Health announced that it would conduct a study comparing the effectiveness of two Genentech drugs, Lucentis® and Avastin®, for the treatment of age-related macular degeneration (AMD).  This was not good news for Genentech, because although treating AMD with Avastin® is an off-label use (Avastin® is approved for treating colon and certain lung cancers), the cost of the treatment is between 1% and 3% of the cost of Lucentis® treatment ($20-60 per dose versus $2,000/dose).

Lucentis® has been a great success for Genentech despite its cost, because it is one of only two FDA approved drugs for treating neovascular or "wet" AMD, a condition that if untreated leads inexorably to blindness.  The drug has been so successful that it has the lion’s share of the market over Macugen (sold by EyeTech).  The drug has not been without controversy, however, since it is not a cure for AMD but merely stablizes the wet form, and in many patients prevents the condition from worsening.  This means that the drug must be administered like insulin or other maintenance drugs.  At a cost of $2,000/dose and a once-a-month dosing schedule, the cost to treat the half-million "wet" AMD patients in the U.S. is greater than $10 billion per year.  According to Dr. Edward Chaum, Plough Foundation Professor of Ophthalmology, University of Tennessee, this is more than the entire Medicare budget for all of ophthalmology (cataracts, diabetes, glaucoma, and everything else) combined.

Avastin® is so much cheaper because it is priced at $600/vial for intravenous use, but is injected into the eye at such small doses (0.1cc) that each vial can deliver 30 doses ($20/dose), says Dr. Chaum.  Genentech has opposed the off-label use of Avastin®, and cautioned ophthalmologists that there is a greater risk of stroke using Avastin.  However, the increased stroke risk was seen at the therapeutic dose for cancer patients, which is 400-fold higher than the dose used to treat AMD.  Recently, Genentech also reported that certain doses of Lucentis used for AMD are also associated with an increased risk of stroke.

Not surprisingly, Genentech’s behavior has angered the Retina ophthalmology community, which has generally ignored Genentech’s warnings and has been widely using Avastin® for AMD.  The results of the NEI study will not be available for about a year, but the situation should be clarified sooner than that, in view of anecdotal evidence from such off-label use and the expected disclosure of preliminary results.

Dr. Chaum and others believe that the results of the head-to-head clinical trial will likely validate use of Avastin® as being at least equivalent to Lucentis®.  If confirmed, Genentech will be forced to choose between removing Avastin® from the marketplace (which is unlikely, since it is a $2 billion/ year drug for cancer indications), or acquiescing to providing the drug in the lower dose amounts that are now provided only by compounding pharmacies.

The hard reality is that AMD is a progressive disease whose best treatment is prevention.  It is an indication of the dire consequences of AMD (blindness in old age) that patients and their physicians endure the therapeutic regimen (direct injection into the eye every 4-6 weeks) for drugs that at best only slow disease progression.  Genentech lost an opportunity with Avastin® and Lucentis® to solidify ties with a large and growing patient population and their doctors.  Perhaps their cooperation, if forthcoming, in the face of the expected equivalence results for these two drugs, and the resulting greater affordability for the majority of patients, can repair the damage.

    By Kevin E. Noonan

The news keeps getting worse for Genentech.  On the heels of the announcement that the National Institutes of Health would fund a study comparing the efficacy of Lucentis® and Avastin® (see "Retinal Specialist on the Avastin®/Lucentis® Controversy"), Genentech disclosed the results of an Avastin® study done by its parent company, Roche Holdings Ltd. of Switzerland.  In this study, Avastin® treatment was shown to be just as effective in lung cancer patients using half the recommended dose as it is with the full dose.  Treating with half the recommended dose would drop the cost, and hence Genentech’s revenues, from $8,800/month to $4,400/month, threatening a franchise worth about $3 million yearly (a number expected to grow as the U.S. population continues to age).

The only silver lining is that the Roche study did not present evidence on the effects of the lower dose on long-term survival.  Although patients treated in the study with the lower dose showed not difference in mortality, the length of the study was insufficient to establish no difference over a longer timeframe.  Thus, low dosing comprises a risk for cancer patients, which may inhibit the 75% of oncologists who have not switched to the lower dose from doing so.  In any event, this study continues to threaten Genentech’s newfound dominance as the market leader in U.S. biotechnology companies.

    By Kevin E. Noonan

The U.S. National Institutes of Health (NIH) announced
Wednesday that it had assembled a database of the entire genomic complement of
2,000 human and avian influenza strains that are publicly available from the
National Institute of Allergy and Infectious Diseases (NIAID).  The information in this database represents
flu samples collected worldwide.

The database is the result of the Influenza Genome
Sequencing project, started at NIH in 2004.  NIAID funded a variety of influenza sequencing efforts by university and
private entities, including The Institute for Genomic Research (TIGR);
Wadsworth Center of the New York State Department of Health in Albany, NY; the
Centers for Disease Control and Prevention in Atlanta; St. Jude Children’s
Research Hospital in Memphis, TN; the World Organization for Animal Health /
Food and Agriculture Organization of the United Nations (OIE/FAO) Reference
Laboratory for Newcastle Disease and Avian Influenza in Padova, Italy; The Ohio
State University in Columbus, OH; Children’s Hospital Boston; Baylor College of
Medicine in Houston; and Canterbury Health Laboratories in Christchurch, New
Zealand.

The work is not completed:  sequencing capacities have grown to more than
200 viral genomes per month and is ongoing.  The sequence information is being posted on the GenBank database as it is
acquired.

This continued effort is relevant to developing vaccines for influenza, which is
characterized by high mutation and recombination rates.  This information is particularly useful for developing vaccines to
so-called "seasonal" influenza, which kills between 250,000 and
500,000 people worldwide each year.  The
information also helps determine the appropriate vaccines for the population of
influenza virus extant each season and in different parts of the world.  The database will be even more important
if a pandemic strain arises, such as the Asian "bird flu" of recent
years, since pandemics can cause much more widespread mortality – an estimated
40-60 million people died in the 1918 pandemic.