Aastrom: Mice Scared Me Away

| About: Vericel Corporation (VCEL)
This article is now exclusive for PRO subscribers.

Summary

Aastrom's price has surged due to a positive article from Jason Napodano.

Images suggest potential technical flaws in preclinical mouse experiments.

Keep caution on the menu.

Over the past few trading days, sleepy microcap Aastrom Biosciences (ASTM) has exploded from $3.88 3/19/14 to a peak of $6.24 after a positive article from respected analyst Jason Napodano. Seeking Alpha's sophisticated readership is more than capable of analyzing Aastrom's balance sheet. Instead of crunching the numbers, in this article I take a step back and look at the big picture behind Aastrom's science. For every junior biotech, the main question is: will the science pan out? In this case, Aastrom's mice make me nervous.

Aastrom's only product is ixmyelocel-T, a mixture of amplified hematopoietic and mesenchymal cells derived from the patient's own bone marrow aspirates. The process is outlined in Aastrom' patent, found here. In a nutshell, bone marrow aspirates are freed of residual red blood cells through lysis and centrifugation on a density gradient. Subsequently, cells are cultured in media supplemented with human, horse or bovine serum, L-glutamine, and steroids. After multiplying for about 2 weeks in cell culture, the cells are washed and harvested for use.

Thus far, Aastrom has explored two applications for ixmyelocel-T: critical limb ischemia and dilated cardiomyopathy. In this article, we look at the preclinical mouse experiments forming the basis for clinical trials in ischemic cardiomyopathy.

Ischemic cardiomyopathy

Ischemic cardiomyopathy is the heart dysfunction associated with chronically inadequate blood supply. Ischemic cardiomyopathy is usually caused by atherosclerotic disease of the coronary arteries or the cardiac microvasculature. When cardiac muscle cells are deprived of oxygen and nutrients due to inadequate blood supply, they either die (infarct) or exhibit decreased contractility (cardiomyocyte stunning and/or hibernation). Over time, the ischemic heart increases in size, becomes stiffer, and demonstrates reduced ejection fraction. A host of hormones and paracrine factors released in the setting of ischemia and volume overload result in irreversible structural and biochemical changes in the heart affected by ischemic cardiomyopathy. Patients experience varying degrees of shortness of breath, chest pain, leg swelling, pulmonary edema, and fatigue, often complicated by periodic (and often life-threatening) heart failure exacerbations. For the most part, patients are more symptomatic as ejection fraction decreases. Patient symptoms can be quantified using the New York Heart Association scale, a very well-accepted metric.

Aastrom's strategy has been to directly inject ixmyelocel-T into the heart muscle, either through open heart surgery (very invasive) or through catheter-directed means (minimally invasive). If either strategy is going to be a commercial success, the catheter-directed injection is much more likely because of the reduced risk compared to open chest surgery on patients with advanced heart failure.

Aastrom presented mouse data purporting to demonstrate that ixmyelocel-T has a significant protective effect in a model of ischemic cardiomyopathy. The experiment consisted of causing an infarct by ligating the left anterior descending artery, the major artery providing blood to the anterior aspect of the heart. Two weeks later, the animals received a sham injection or an injection of ixmyelocel-T around the edges of the infarct. After an additional 4 weeks, the animals were sacrificed and the hearts sectioned and stained.

Aastrom claims that this data shows that ixmyelocel-T decreases infarct size. Instead, my interpretation of the images available to the public is that the experiment was technically flawed. Let's review the images (page 15 of source document) and understand what they show.

The hearts were sectioned in the short-axis plane. Normal heart muscle stains red with the trichrome stain whereas collagen fibers and scar tissue stain blue. The left ventricle can be identified as the larger, rounded cavity with thicker muscle whereas the right ventricle is smaller, slit-like, and has thinner muscle. The interventricular septum separates the right and left ventricles. Opposite the septum is the lateral wall of the left ventricle. The anterior and inferior walls of the ventricle cannot be distinguished on these sections by anatomy alone, but the stigmata of experimental intervention point to the anterior wall.

In the naïve heart with no arterial ligation, all of the tissue has a normal red appearance with no scar tissue and normal muscle thickness throughout. The rough muscle lining the ventricles with frond-like projections into the ventricular lumen represents normal trabeculations and papillary muscles. On the control and vehicle hearts, the anterior wall is markedly thinned and stains blue instead of red because the muscle has infarcted. The cardiomyocytes have died from ischemic injury and the muscle tissue has been replaced by collagen fibers.

This appearance is characteristic of the scarring that normally occurs when the left anterior descending artery is ligated and the heart muscle infarcts all the way from the subendocardial muscle (adjacent to the ventricle) to the epicardium (surface layer of the heart). We would call this a "transmural" infarct, which is the typical outcome from an acute total obstruction of the left anterior descending artery. The infarct is in the expected location for occlusion of the left anterior descending artery.

The appearance is different on the hearts from treated animals. In treated animal #1, we see a focus of subepicardial scar on the anterolateral wall. The heart muscle is otherwise normal. The appearance of animal heart #1 is very unexpected. The subendocardial muscle is the most susceptible to ischemic injury and appears nearly normal except for possible subtle subendocardial scarring. There is no evidence of a transmural infarction, not even an infarction that was partially repaired. Furthermore, the lateral wall is more affected than the anterior wall. The heart from treated animal #2 demonstrates a focal transmural infarct through the lateral wall of the left ventricle. There is a subendocardial scar along the anterior wall of the left ventricle. Minimal mid-myocardial scarring is present.

(Note: normal mouse coronary artery anatomy is not the same as human coronary artery anatomy.)

The distribution of injury in animal #2 is in the correct territory for injury to the left anterior descending artery, but the lack of transmural injury suggests incomplete ischemia in the LAD territory. The mid-myocardium and especially the subepicardial muscle appears relatively normal in the anterior wall of animal #2's heart.

We can offer two hypotheses for the outcomes of this experiment.

Hypothesis A: ixmyelocel-T almost completely rescues the heart from infarction even when administered two weeks after the injury. In fact, ixmyelocel-T is so effective that nearly all trace of subendocardial and mid-myocardial injury vanished in animal #1.

Hypothesis B: the experiment was flawed in some way.

I like Hypothesis B better. Scarring and wall thinning from a transmural infarct should be well on its way to developing 2 weeks after a major injury such as would be expected from LAD ligation. A simplified (but quite accurate) table of the expected myocardial tissue evolution after infarction in humans is presented in Wikipedia. The natural history of injuries from LAD ligation in mice is well-documented in the literature. LAD ligations followed by histopathologic examination at standardized timepoints was performed by Yang 2002. Tissue death should occur within hours of LAD ligation. Neutrophils infiltrate the injury first, peaking 1-4 days after infarction. Next, macrophages enter the wound, peaking at 4 days but remaining significantly elevated over several weeks. Lymphocytes peak at 7-14 days post-injury. Fibroblasts migrate into the wound starting at 4 days and progressively increase their numbers. At 14 days, approximately two thirds as many fibroblasts are seen compared to 28 days. Per Yang 2002, there is no significant change in the amount of intercellular collagen (which stains blue on trichrome stains) between 2 weeks and 6 weeks.

Stem cell repopulation or changes in the inflammatory response could plausibly be expected to make some change in the infarct's appearance over a 4-week period, but there should still be large evidence of a transmural injury if one existed at the time of ixmyelocel-T administration. After all, the cardiomyoctyes had been dead for two weeks before ixmyelocel-T administration. Wounds do remodel, but a near-complete resorption of collagen and scar over a period of a few weeks with near-perfect regrowth of myocardium is quite unlikely.

On the other hand, it's easy to think of ways in which the experiment could have gone awry: coronary anatomy is highly variable; the ligature around the LAD may not have been tight enough in animals #1 or #2; there may have been compensatory connections between the LAD and other coronary arteries in animals #1 or #2; or animals that had perfect LAD ligations in the treatment group may have been more likely to die and not be included in the study whereas animals with incomplete ligations would have been more likely to survive.

The technique for causing LAD infarcts in mice is not totally trivial. Per Kofidis 2005:

Once the pericardium is opened in a mouse, the left atrium can be seen contracting vigorously. From a lateral approach, one looks for the middle of the free margin of the left atrium. This is the point at which the surgeon usually identifies the LAD and moves distally to the transition from the first to the second third of the vessel course on the surface of the LAD. This can be viewed as the optimal spot for the LAD ligation to obtain a significant infarction of the mentioned magnitude. Ligation in the immediate proximity of the left atrial margin (too proximal) usually causes death of the animal. A ligation further distally will cause a much too small infarction that will not impact left ventricular function sufficiently.

It may be that the images in Aastrom's presentation are hearts from mice with ligations of the LAD that were too distal to cause the large, transmural anterolateral infarcts seen in the control animals.

The last hypothesis that needs to be entertained is that I am totally wrong in my interpretation of the images. I am definitely not a pathologist. However, I look at the radiological sequelae of subendocardial, nontransumural, and transmural infarctions on a routine basis. The education I have received thus far in my career would argue that the images of treated mouse hearts are not consistent with a fully occlusive LAD ligation over a period of two weeks followed by an intervention.

Conclusion

There are a number of much stronger companies in the "regenerative medicine" space such as Athersys (NASDAQ:ATHX) and Neuralstem (NYSEMKT:CUR). In truth, "regenerative medicine" is something of a misnomer: the balance of the evidence suggests that injected stem cells alter the course of disease through immunomodulatory and paracrine effects rather than regenerating new tissues. Aastrom does have some decent (but not irrefutable) data from human clinical trials on ixmyelocel-T in human ischemic cardiomyopathy and of course it is the human data that counts. Unfortunately, in my opinion the preclinical experiments providing the basis for the human trial appear to have been flawed in some way. Rather than finding the results to be unbelievably good, I consider the preclinical experiments to be more along the lines of simply not believable. If the preclinical experiments were flawed, it is less likely that the human trials will turn out well.

I do not believe that Aastrom has been deliberately deceptive. Instead I believe that Aastrom's results may have been confounded by the limitations in their experimental protocols. Of note, the mouse experiments were performed by Medigenix LLC, so Aastrom may not have had complete control over the exact experimental methodology.

I do not bear Aastrom or its investors any animus and in fact wish them the best of fortune in finding new treatments for peripheral vascular disease and ischemic cardiomyopathy. However, Aastrom's common stock is not a safe enough place to attract my personal capital. I believe that the surging stock price has unjustifiably inflated in response to Mr. Napodano's article and will soon return to its pre-article levels.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: If you find any errors in this article, please have the kindness to call them to my attention. This article represents my personal opinions only and does not constitute medical or investment advice. Rely on your own due diligence for all of your investment decisions.

Editor's Note: This article covers one or more stocks trading at less than $1 per share and/or with less than a $100 million market cap. Please be aware of the risks associated with these stocks.