Founded in 1987, Alkermes is a pharmaceutical company focused on novel drug delivery systems for existing compounds. Alkermes develops injectable very long-acting formulations based on its ProLease and Medisorb technologies, as well as inhaled formulations based on its AIR pulmonary technology.

Alkermes currently has one approved product, Risperdal Consta, which is used to treat schizophrenia, and an NDA currently under review with the FDA for Vivitrex, which would be used to treat alcohol dependence. Alkermes’ pipeline consists of AIR Insulin, Exenatide LAR, AIR epinephrine, and AIR- hGH, with AIR Insulin furthest in development. Potential blockbusters Exenatide LAR and AIR Insulin are the most likely to drive Alkermes's future growth.

TECHNOLOGIES

Medisorb and ProLease: injectable extended-release
Medisorb and ProLease are Alkermes’s technologies for encapsulating drugs in microspheres made of common medical polymers to provide controlled release over time. The Medisorb technology is for small molecules and the ProLease for peptides and proteins. The ProLease process involves stabilization of the molecules under very low temperatures.

With both Medisorb and ProLease, release of the drug is controlled by diffusion of the drug through the microspheres, as well as biodegradation of the polymer. The timing of release can be adjusted through formulation and fabrication variables, including microsphere particle sizing and choice of polymers and excipients. Alkermes believes these technologies can be used to control release of the drug from a few days to several months.

AIR: pulmonary drug delivery
The AIR pulmonary delivery system is for the delivery of both small and macromolecules (including peptides and proteins) to the lungs. The drugs are formulated into dry powders made up of highly porous, low density particles. The plastic delivery devices are breath activated. Alkermes believes that by varying the ratio and type of excipients used, it can utilize the technology for a range of drugs with both immediate and extended release formulations.

MARKETED DRUGS

Risperdal Consta (Long-acting Risperdal Injection)

Risperdal Consta, approved by the FDA in October 2003 for the treatment of schizophrenia, is the long-acting injectable formulation of Janssen’s popular atypical antipsychotic drug Risperdal. It is given every two weeks, and its primary use is for patients who are non- compliant with daily oral medications. The atypical antipsychotics are second-generation agents that have largely replaced first-generation antipsychotics, which carry a risk of causing certain movement abnormalities. Since Risperdal Consta is the only approved long-acting atypical antipsychotic, sales have grown significantly to $263 million worldwide in the first half of 2005. The drug is commercialized by Janssen, and Alkermes receives manufacturing and royalty revenue (expected to total 10% of product sales this year).

While we expect continued strong short-term growth, there is a potential major competitor on the horizon: paliperidone palmitate (Johnson & Johnson, Elan), the active metabolite of Risperdal. Paliperidone palmitate is also being developed by Janssen and may have fewer side effects than Risperdal. The drug uses Elan’s NanoCrystal technology in a once-a-month formulation. A phase III study comparing paliperidone palmitate injections every 4 weeks with Risperdal Consta every 2 weeks began February 2005, and though the patent for Risperdal Consta expires in 2013, paliperidone palmitate is on track to be launched in 2008. One reason for this timing is that oral paliperidone is being developed as a follow-on to oral Risperdal, whose patent expires in 2008. While some physicians may opt for the biweekly injections of Risperdal Consta, we project that paliperidone palmitate will take major share, significantly reducing Risperdal Consta sales beginning in 2009.

DRUGS IN DEVELOPMENT

Risperdal Consta for Bipolar Disorder

Risperdal Consta is also in phase III study for maintenance treatment of bipolar disorder. The atypical antipsychotics are recommended for bipolar maintenance in patients who fail treatment with standard therapy. Similar to its use in schizophrenia, we expect injectable Risperdal Consta to be used primarily in patients who are non-compliant with oral treatment. We are projecting approval in the first quarter of 2008. However, we also believe paliperidone palmitate will be developed for this indication and will start to take significant share from Risperdal Consta in 2010.

Vivitrex (Long-acting Naltrexone Injection)

Vivitrex is Alkermes’s once-a-month injectable formulation of the generically available oral drug naltrexone, which is used to treat alcohol dependence. An NDA for Vivitrex was submitted April 1, 2005, and a priority review designation was granted May 27. However, on September 20, Alkermes announced that the PDUFA date had been extended to December 30, 2005, because the FDA reclassified a response by Alkermes as a major amendment to the NDA. Details regarding the nature of the issue have not been released.

Naltrexone is an inhibitor of the opioid m, d, and k receptors, and the oral formulation is also used in the treatment of opioid dependence. Though the exact mechanism of action in treating alcohol dependence is not clear, it is thought to reduce craving and the reinforcing properties of alcohol.

In a pivotal trial, Vivitrex resulted in an overall 25% decrease in the rate of heavy drinking days. For the typical drinker, the reduction was 48%. While there was a minimal effect on abstinence for the entire group, for the minority of patients who entered the study with at least 7 days of abstinence, the total abstinence rate was 41% versus 17% for placebo. Vivitrex’s effects were stronger in men than women (the study was not powered to separately test efficacy in these subpopulations, but this could become an issue in the labeling).

A major market issue for Vivitrex is that the oral form of naltrexone is not commonly used, so Vivitrex will have to expand the market. Despite 1.3 million patients receiving treatment for alcohol abuse or dependence, only 16,000 to 17,000 prescriptions are written for oral naltrexone a month. One large database found that patients typically only filled one prescription per year. A recent survey found that major reasons for the low usage were patient refusal to take or comply with the medication (23%) and patient inability to afford the medication (21%). Additionally, consistent with clinical study results, physicians perceived the drug to have only a modest or uncertain clinical effect. Oral naltrexone has been shown to control consumption, but not to consistently lead to abstinence.

Vivitrex will also compete with Campral (Forest and Merck KGaA), approved in July 2004. Unlike oral naltrexone, Campral has been shown in studies to increase rates of abstinence, though the studies have been criticized for using patient interviews rather than daily drink diaries and only enrolling patients who were abstinent upon entry. Campral is more popular in Europe than oral naltrexone, and though it was only recently approved in the US, prescriptions are beginning to surpass those of oral naltrexone. While Campral has the benefit of being an oral drug, it has the disadvantage or requiring three times a day dosing. There is some evidence of a benefit when both Campral and naltrexone are used together, and there is a government sponsored trial (COMBINE) evaluating combination use.

A potential more direct competitor for Vivitrex is a depot formulation of naltrexone for monthly intramuscular injections under development by privately funded Drug Abuse Sciences. Though positive date from one phase III trial has been published, according to company information the drug is still in phase III development. The company appears to be in some disarray, with three different CEOs over the past few years, and we believe Vivitrex will be able to hold its own against this drug.

Though it is difficult to compare Vivitrex clinical results with those of competitors, we believe that the Vivitrex results will be perceived very positively. While we expect partner Cephalon to be able to help expand the market somewhat, we also believe that like all of these drugs, success will be limited due to the lack of strong enough motivation in alcohol dependent patients.

One additional issue regarding Vivitrex use is the danger of liver toxicity seen with high doses of oral naltrexone and resulting in a black box warning for the oral formulation. This is particularly an issue for patients dependent on alcohol, since alcohol is also toxic to the liver. However, for many patients, the increased compliance with Vivitrex can result in improvement in liver function. While liver toxicity is likely to be a labeling issue, we feel that use of Vivitrex will be avoided primarily in patients with more severe liver disease.

AIR Insulin (Human Insulin Inhalation Powder [HIIP])

AIR Insulin is currently in phase III development for both type 1 and type 2 diabetes, and we are projecting approval in the second quarter of 2009. AIR Insulin is an inhaled form of insulin that can replace shorter acting injectable insulins used prior to meals. Alkermes is partnered with Lilly for this drug.

In a phase II study of type 1 diabetics, AIR Insulin was shown to be noninferior to injected insulin in glycemic control, though it led to both lower fasting blood glucose and increased nighttime low blood sugar. The investigator noted that the dosing would need to be adjusted for the phase III trials. However, given that with the exception of Technosphere, the remaining inhaled insulins in late stage development have similar pharmacokinetics, we expect them to have similar efficacy. We believe Technosphere’s more rapid action may offer some advantage in post-meal control, but do not believe the effect will be major.

We project that the inhaled insulins will take a significant share of the insulin market. Because of its small size, its ease of use, and Lilly’s strength in the diabetes market, we project AIR Insulin will be the market leader of the inhaled insulins, beating first-to-market Exubera (Nektar, Pfizer, and Sanofi-Aventis), but splitting share with Technosphere (MannKind) and other options coming on to the market later (Technosphere is on track to be approved around the same time as AIR Insulin). Please see our inhaled insulin Revenue Model Update for the competitive landscape and more details on the inhaled insulin market.

Exenatide LAR (Long-acting Synthetic Exendin-4)

Exenatide LAR is the long-acting version of Byetta (Amylin and Lilly) under development for type 2 diabetes. Currently in phase II study, strong results were recently reported, though the number of patients was limited. In type 2 diabetics not achieving adequate glucose control with or without metformin, once-weekly Exenatide LAR resulted in an A1C reduction of 2% compared to placebo (group baseline was 8.5%) and an average weight loss of 9 lbs compared to placebo.

Exenatide LAR will be a major improvement on Byetta, since Byetta is a twice-a-day injection, which is a strong negative for many patients. A major attraction of these drugs is their ability to induce weight loss, which aids diabetes control. One unanswered issue for Exenatide LAR is how it works in combination with the popular sulfonylureas. Byetta usage increases the rate of hypoglycemia for these agents (28% mild to moderate hypoglycemia in one study), and it will be important to see that Exenatide LAR does not exacerbate this complication further.

Because of its convenient weekly dosing, as well as its weight loss, we are currently projecting Exenatide LAR to take significant share in type 2 diabetic patients failing oral medications, and a smaller share earlier on in treatment. Please see our 2005 ADA Post- Conference Report and Type 2 Diabetes Indication Report for more details on Byetta, Exenatide LAR, and the diabetes market.

AIR- hGH

Currently in phase I development, AIR-hGH is an inhaled formulation of human growth hormone to treat growth hormone deficiency. Growth hormone has been typically administered by injection, though there is also needleless system currently available that is less painful than an injection.

AIR- Epinephrine

AIR Epinephrine is an inhaled dry powder formulation of epinephrine to treat anaphylaxis (a severe allergic reaction). The purpose is to replace injectable epinephrine, which patients allergic to certain common items often keep on hand for emergency self-injection. Three phase I studies have been completed and Alkermes plans to review its plans with the FDA prior to the initiation of additional clinical trials.

PARTNERSHIPS

Janssen
Alkermes collaborated with Janssen Pharmaceutica on the development of Risperdal Consta. Janssen provided development funding, is responsible for regulatory approvals, and with an affiliate, has exclusive worldwide commercialization rights. Alkermes exclusively manufactures Risperdal Consta and receives manufacturing and royalty revenues from Janssen. Alkermes expects manufacturing revenue of 7.5% and royalties of 2.5% for their fiscal year ending March 31, 2006.

Cephalon
On June 23, 2005, Alkermes entered into collaboration agreement with Cephalon for sustained-release forms of naltrexone in the US, including Vivitrex. Cephalon also has an exclusive option until December 20, 2005 to negotiate with Alkermes for rights outside the US. Cephalon paid $160 million to Alkermes, and will also make milestone payments of $110 million on US approval and up to an additional $220 million if sales exceed certain levels. For an initial time period, Alkermes is responsible for the first $120 million of overall product losses, and Cephalon for additional losses. The companies will subsequently divide losses equally. Once the product is profitable, all pretax profits will be shared equally and Cephalon will reimburse Alkermes for product expenses.

Eli Lilly
Alkermes entered into a development and license agreement with Lilly in April 2001 for AIR inhaled formulations of insulin and other compounds relating to diabetes. Alkermes is responsible for development of the compound and delivery devices. Lilly is responsible for clinical trials, obtaining regulatory approvals, and marketing. Lilly has worldwide commercialization rights and is required to pay to Alkermes initial fees, research funding, milestone payments, royalties, and commercial manufacturing fees. Both companies will manufacture the products for sales.

Alkermes entered into a similar agreement with Lilly in February 2000 for the development of an inhaled formulation of human growth hormone.

In December 2002, the collaboration between the two companies expanded after development milestones were reached, and the royalty rate payable to Alkermes was increased.

Amylin
Alkermes entered into an agreement with Amylin in May 2000 for the development of a long- acting Medisorb formulation of Exenatide LAR in the treatment of type 2 diabetes. Amylin has an exclusive, worldwide license to the Medisorb technology for injectable extended- release formulations of exendins and other related compounds that Amylin may develop. Amylin also has a collaboration agreement with Lilly for the development and commercialization of Byetta and Exenatide LAR. Alkermes is entitled to receive research and development funding and milestone payments, as well as royalty payments and manufacturing fees.

Targeted Genetics
In 1999, Alkermes entered into a license agreement with Targeted Genetics, in which Targeted Genetics received exclusive rights to an issued patent related to AAV vector manufacturing (the AAV vector is a viral based method for gene therapy). One product covered by this agreement, tgAAC94, is currently in phase I development for rheumatoid arthritis. tgAAC94 is viral gene therapy that involves tumor necrosis factor “a” (TNFa). Alkermes does not have any clinical development or manufacturing responsibilities and will receive milestone payments and royalties.

PATENTS

Alkermes owns approximately 112 issued U.S. patents, of which none expire prior to 2013. Through licensing agreements with third parties, Alkermes has exclusive rights to approximately 38 issued U.S. patents, of which none expire prior to 2016. The Risperdal Consta patent expires in 2013.

UPCOMING CATALYSTS

A major impending catalyst is the PDUFA date for Vivitrex, which the FDA recently moved to December 30, 2005.

Another major near-term catalyst is the full data from Exenatide LAR’s phase II study. Alkermes and its collaborative partners have stated that this data will be presented in a future scientific forum, which we anticipate could be the American Diabetes Association's 66th Annual Scientific Sessions in Washington, DC on June 9-13, 2006. However, per Alkermes, the data could also be released without a venue by the end of this year.

Finally, Alkermes is expected to initiate a pivotal trial for AIR epinephrine by the end of 2005.

REVENUE POTENTIAL

While Risperdal Consta and Vivitrex are likely to supply moderate revenue for Alkermes in the near term, helping to offset costs, we believe their potential will be limited by various market factors. However, Exenatide LAR and AIR Insulin have the potential to fuel Alkermes revenues to over a billion dollars in 2012.

Risperdal Consta for Schizophrenia
We are projecting continued strong growth until 2009, when we expect paliperidone palmitate to start taking significant share from Risperdal Consta. Potential upside would come if paliperidone palmitate does not demonstrate superiority to Risperdal Consta in clinical trials.

For schizophrenia, we currently project 5 and 10-year revenue of $472.6 million and $130.8 million in the US. The corresponding worldwide revenues are $750.9 million and $155 million.

Risperdal Consta for Bipolar Disorder
As the first long-acting atypical antipsychotic approved for this indication, we are projecting strong peak share for the segment of non-compliant patients after approval in the first quarter of 2008. However, we believe that beginning in 2010, before peak share is reached, Risperdal Consta’s growth will be cut short by paliperidone palmitate’s entrance into this market segment. Projected 5 and 10-year peak US revenues are $237.9 million and $77.5 million. The corresponding worldwide numbers are $358.9 million and $134.8 million.

Vivitrex for Alcohol Dependence
Given the difficulties pharmacologic treatments have had in the alcohol dependence market thus far, we are cautious in our outlook for Vivitrex. While we think it will expand the numbers of patients treated, as well as patient compliance (resulting in more prescription refilled), we believe share will be limited by the lack of motivation of patients for the continual monthly injections.

With the increased convenience and improved compliance, we believe that Alkermes will price Vivitrex at a premium to newly approved Campral, targeting patients who are covered by payors and who are strongly motivated to reduce their alcohol consumption. This strategy is likely, however, to leave out a significant segment of patients who are self-paying or who have to pay a high copay.

Significant upside would come if Cephalon is more successful in growing the market. On the downside, with projected revenue many times more than those currently achieved by oral naltrexone, our estimates for usage could be overly optimistic, especially given the probable high price of the drug.

Our model presumes that Vivitrex will be approved by the end of the first quarter of 2006 and launched in the second quarter. Our projections for 5 and 10-year peak US revenue are $177 million and $161.8 million. The corresponding worldwide revenues are $247.8 million and $385 million.

AIR Insulin for Type 1 and Type 2 Diabetes
Though there is still developmental risk (AIR Insulin still needs to meet phase III efficacy and safety endpoints), we project AIR Insulin to be approved in the second quarter of 2009 and to become the market leader for the inhaled insulins. Potential upside could come if the development of potential competitor Technosphere falters.

Our 5 and 10-year US revenue projections are $282.5 million and $1.1 billion for type 2 diabetes and $63.2 million and $220.7 million for type 1 diabetes. The corresponding worldwide numbers are $428.5 million and $2.1 billion for type 2 and $101.4 million and $429.1 million for type 1.

Exenatide LAR for Type 2 Diabetes
With the lack of strong comparable competition, we expect Exenatide LAR to achieve even greater revenue than AIR Insulin in the type 2 diabetes market. There is still significant development risk: phase II results have only been released in a small number of patients, and it is unclear how the drug will fare with the popular sulfonylureas. We expect that if approved (our projection is the first quarter of 2009), most patients will switch from Byetta, leading to high initial revenues for the long-acting version. Exenatide LAR (and AIR Insulin) will also benefit from the rapidly growing type 2 diabetes market.

Our 5 and 10-year US revenue projections for Exenatide LAR are $2.4 billion and $3.7 billion, with corresponding worldwide revenues of $3.5 billion and $7.8 billion.

VALUATION

We believe that for most of its products in later stage development, Alkermes will receive a minority royalty and manufacturing revenue. That manufacturing revenue will typically be for formulating the drug into a long-term product, with the partner supplying raw materials for the active ingredient. Vivitrex is an exception, since Alkermes must purchase naltrexone to create the long-acting version and splits profits equally with its partner.

We expect Alkermes to become profitable in 2009 with the beginning of Exenatide LAR revenues. The 5 year valuation for Alkermes rests largely on the approval and success of Exenatide LAR and Vivitrex. The 10-year valuation rests largely on Exenatide LAR and to a lesser extent AIR Insulin and Vivitrex.

We currently value Alkermes 5 and 10-year pipeline at $19.35 and $19.39 a share. With Alkermes close of $15.96 on October 6, 2005, we rate its stock as SLIGHTLY UNDERVALUED. A major long-term issue for the company is whether it can repeat its success in partnering potential blockbuster drugs, particularly in the face of increasing competition from novel drug delivery technologies.

Upcoming Catalysts and Valuation

Should Vivitrex gain approval, our 5 and 10-year valuations would increase to $20.26 and $19.67 a share – the increase is small partly because Vivitrex already has a high likelihood of approval.

While more complete data from Exenatide LAR’s phase II multi-dose study is an important upcoming catalyst, it will still be unclear whether further phase II studies will be needed. Amylin has been vague, noting this will depend on FDA discussions and the labeling Amylin will ultimately pursue. Should LAR subsequently advance to phase III with, for example, a 10% above average LOA, our 5 and 10-year valuations would be $24.30 and $22.76.

Alkermes Outstanding Debt

Currently Alkermes has approximately $276 million in outstanding long-term debt consisting primarily of $151.4 million in non recourse Risperdal Consta secured 7% notes and $122.3 million in 2.5% convertible subordinated notes due in 2023. $125 million of the notes were issued in September 2003 with an initial conversion premium of 10%, making the notes convertible into common stock at $13.85/share. Alkermes has the option to auto- convert the notes if the common stock trades above 150% of the conversion price ($20.78) for at least 20 trading days out of a 30 trading day period. If the auto-convert is triggered before September 6, 2006 then Alkermes is responsible for the make-whole payment. The bond can be called on or anytime after September 6, 2006 at 101, September 1, 2007 at 100.5, and on September 1, 2008 at par. The bond can be put back to the Company on September 1, 2008, 2013, and 2023 at par.

Given the current market price of the common stock, it is quite feasible that the auto- conversion will be triggered in the coming year and Alkermes will choose to convert all of the outstanding debt. Another possibility would be that given the current market price of the bond (128.625 as of September 23, 2005 according to TRACE) and the common stock, Alkermes will decide to exercise the call feature on September 6, 2006, essentially forcing investors to choose the more profitable option of converting rather than allowing the bond to be called.