I was recently looking at the torrent of biotechnology IPO’s, and one that stood out to me was Rubius (RUBY) - based on the size of the offering (240M) and the market capitalization (2B at peak). This is a Boston based, venture capital created (by Flagship Pioneering) company. The senior management appears to be experienced and well pedigreed, from top to bottom. The board of directors is headlined by David Epstein, former Novartis Pharmaceuticals Division CEO, and a series of other well-known executives and scientists such as Roger Pomerantz (CEO of MCRB), Noubar Afeyan, and Bob Langer (MIT). The IPO was really large, netting the company over 241M. What was striking was that this company has not even filed an IND, so for professional investors to put this much money down on a unicorn, this had to be a great idea. As a result, I had to go take a look. It turned out not to be what I expected.

Rubius's Main Technology

The entire concept for Rubius is based on using modified red blood cells (RBC) as a therapeutic (see here). Rubius takes hematopoetic stem cells (HSC), then genetically modifies them. They take these modified HSCs and differentiate them to RBC, which express whatever the company inserted. These modified RBCs are the "drug". RBC generally can circulate 90-120 days, so the Rubius idea would be fairly convenient even though it would be an IV infusion.To make this concept happen, you must believe in the idea that there is a universal donor for red blood cell transfusions. Specifically, in order to believe in Rubius, you have to believe that people who are O negative can truly be compatible with all patients such that chronic infusions are possible. To some extent, this concept is true when considering one-time or infrequent transfusions with O negative blood. As a quick reminder, you likely learned in high school that blood is generally typed ABO, which are the major Rh antigens found on the surface of RBCs (red blood cells). If you are interested, you can read about the Rh system here (advanced) or here (basic). One indisputable fact is that proteins in the Rh family can be immunogenic.

The best known members of the Rh system is the classic ABO antigens. You can, of course, be A, B, AB, or O. AB is generally thought to be universal recipient, while O is the universal donor. You probably also learned that the next most important antigen after ABO is Rh-D (Rhesus factor-D). For the purposes of this discussion, reference to Rh positive or negative is shorthand for Rh-D only. For instance, if a unit of blood is referred to as O negative, this is shorthand for O Rh-D negative. O positive is short hand for O Rh-D positive. You can be either Rh-D positive or negative.

It has been known for many, many decades that ABO matching must be done prior to any blood transfusion, since giving someone mismatched ABO blood can be fatal due to pre-existing antibodies in the recipient. However, people who are Rh-D negative generally do not have pre-existing anti-Rh-D antibodies. These anti-Rh-D antibodies should only appear if a person who is Rh-D negative is exposed to Rh-D positive red blood cells (RBC). Situations like this would most commonly occur in a situation such as giving an O Rh-D negative patient a transfusion of O RH-D positive blood or a mother who is O negative but gives birth to an O positive baby. Alloimmunization due to pregnancy is a very well-studied area of medicine, and recent reviews can be found here and here. The gist here is that the introduction of fetal cells into the mother can occur at birth, which can lead to undesirable anti-Rh-D antibodies in the mother. This is undesirable because these anti-Rh-D antibodies can negatively affect future pregnancies, unless there is pharmacologic intervention (discussed below). These anti-Rh-D antibodies will not have impact to the mother (who is Rh-D negative) or the just born baby but rather the next baby, which may be Rh-D positive. Because the Rh-D positive baby will have RBC with Rh-D on the surface (like the first baby) and the maternal anti-Rh-D can cross the placenta, the pregnancy can be terminated due to lysis of the fetal RBC by the anti-Rh-D antibodies. This is call hemolytic disease of the fetus/newborn (referred to as HDFN or HDN). See a relatively recent review on this topic here.

The Rh-D status of the mother and fetus is very important in actual clinical practice. It has long been known that when a Rh-D negative woman conceives a Rh-D positive child (because the partner is Rh-D positive), there is risk that the woman can become Rh-D ANTIBODY positive after the first pregnancy as discussed above. Since the ’60s, Rhogam (human anti-Rh-D antibody) has been used (it is on the list of WHO essential medicines) to eliminate the risk of anti-Rh-D HDN. All pregnant women are screened for Rh-D status, and if there is an incompatibility (baby is Rh positive and mother is Rh negative), doctors administer Rhogam to the MOTHER when the child is born. The Rhogam binds to any circulating fetal blood cells and eliminates them before the mother’s immune system can see them. This keeps the mother free of anti-Rh-D antibodies, even if she is exposed to some of the Rh positive fetal cells. This has been standard of care for decades, so you very rarely see HDN due to Rh-D in the developed world. I found a very good discussion of this (here).

What is less well known to the lay public is that there are a whole family (about 35) of minor antigens in the Rh family which are also potentially immunogenic (see representative papers/presentation one, two). For those really interested, I found an excellent thesis from the Netherlands (here). The ABO and Rh-D proteins are but two members of the Rh family, but they are the most important and best known. It is also known that some of these can be immunogenic and can cause documented HDN (see HDN presentation). Even a cursory scan of the literature and web shows that HDN can at least be caused by many other antibodies to other Rh proteins, such as E, C, c, Kell, Kidd, Duffy, K, and M (see slides). Some of these Rh antigens appear to be less immunogenic, so these situations seem to arise only in women of who receive multiple blood transfusions. However, the inescapable fact is that patients who receive multiple transfusions may potentially produce various anti-Rh antibodies against other members of the Rh family found on the donor but not present in the recipient.

Enter the Rubius Concept

Let’s look at the base concept of Rubius. This company is taking hematopoietic stem cells (HSC) from O negative donors and differentiating them into modified RBC to use as a drug. In their rare disease portfolio, they are taking an enzyme (generally non-human) and using gene therapy to insert it into the HSC. They then manipulate the HSC to differentiate into RBC which contain the foreign enzyme they put in. The advantage here, theoretically, is that you can put an immunogenic protein into the RBC, and the RBC will actually prevent antibodies from getting into the cell and neutralizing the foreign protein. In this way, Rubius believes that they can put phenylalanine ammonia lyase (PAL) into a RBC and administer it to humans. PAL is found in bacteria, yeast, and fungi, but not humans. This protein was modified by Biomarin and is approved to treat PKU (Palynziq)– with a black box warning for risk of serious anaphylaxis (allergic reaction). As expected, this protein can very rapidly degrade phenylalanine, but since it is a non-human protein, it is a difficult drug to use due to the creation of anti-drug antibodies – hence the warning for anaphylaxis. Rubius, theoretically, can put the PAL in a RBC by integrating the PAL gene into the donor HSC. They induce these HSC to mature into RBC which wind up containing the PAL due to the insertion of the gene into the HSC. Thus, even if patient has anti-PAL antibodies in circulation (which they certainly would after some exposure), the anti-PAL antibodies cannot access the enzyme because it is hidden inside a RBC. This is Rubius' program called RTX-134.

The PAL in the RBC can still access the phenylalanine because the Phe can enter the RBC. In this way, on paper, the genius of Rubius can be seen by using gene therapy to create a modified RBC (containing PAL) which can rapidly degrade excess phenylalanine in circulation and give PKU patients a normal life. Also, since RBC can circulate for a long time, this would be a very infrequent administration, probably quarterly or longer. This profile would appear to be more attractive the Biomarin offering. This sounds neat, and probably, this idea would get you an A in your high school biology science fair.

Why I have a problem with this approach

I believe that this idea is tremendously flawed on several levels. I challenge the concept of a universal donor. As I discussed in the beginning, there are a large number of Rh antigens found on RBC. Rubius will only be matching the ABO and Rh-D antigens. There are too many other possible antigens on RBC to possibly consistently match. Women who receive this “drug” will be taking on measurable risk of developing various antibodies against the many Rh minor antigens. Since you cannot predict which, if any, antibodies which might form, a woman could be exposing their future fetus to risk of HDFN even after a single dose.

Somehow, I believe Rubius management must believe that O negative from any person can be given repeatedly to any person with no long term risks. For women of child bearing age, simply giving them O negative RBC from any donor can have a severe impact on their future ability to have children, even with O negative blood. In addition, since Rubius cannot make an infinite quantity of modified RBCs (RBC can only be kept for a certain period since they cannot be frozen), they will need to be constantly making these cells from different donors. This means that the minor antigen mixture will be constantly varying across each donor. I imagine if you get their modified cells enough times, you will become antibody positive for every Rh antigen that the patient is originally negative.

This could result in another unintended consequence. For example, if you have a PKU patient who is Rh-M negative, and you give them several doses (maybe even 1 dose) of M positive “superblood” (arrogantly dubbed by Rubius’ Torben Nissen). It is possible that this patient will convert to anti-M antibody positive. If their subsequent dose of “superblood” is also M positive, I think like in HDN, these anti-M antibodies will bind to the “superblood” and either lyse them right then and there (like in a potential fetus), releasing the payload, or direct the RBC into the splenic/liver garbage disposal in short order. Releasing the immunogenic payload in the blood stream may be dangerous, since Biomarin PAL drug has a black box warning of anaphylaxis (see label). With over 35 Rh antigens, you would think several doses of Nissen’s “superblood” from different donors would convert PKU patients to multiple minor Rh antibody positive, which would make it impossible to use his “superblood” anymore due to a dramatically shortened half-life and/or putting the patient at risk for anaphylaxis, as if they were getting Biomarin’s drug. Palynziq's label states that 100% of treated patients developed anti-drug antibodies (see label). There is no reason to believe that this would not be the same with Rubius.

I am also very critical of the company for choosing PKU. This is a non-lethal disease (see Mayo review here). I’m not questioning that this is disease that has serious issues – patients with this disease have significant dietary restrictions which has a major impact on their quality of life. In addition, non-compliant patients are taking risks against their cognitive ability. However, let us be real, this is not a lethal disease. I believe that Rubius chose this disease because the FDA acceptable phase III endpoint – phenylalanine levels – is easy to measure. Taking a look at RTX-134 from a benefit/risk analysis, I don’t really see this ratio as remotely favorable. For a PKU patient, the benefit could be a much less onerous diet. On the other hand, on the risk side, a female patient could be jeopardize her future pregnancies, and male and female patients would have the risk of anaphylaxis, and likely incur quite a bit of financial toxicity for a term therapy which may be useful for only a short period time (problem since PKU is a long term disease).

Rubius has additional ideas for their use of modified RBCs, but after I thought about their lead rare disease platform, I stopped reading. The problem I am discussing will impact everything they do, since it is based all on their “superblood” multiple transfusion core concept. In my opinion, such an idea would fly in the face of decades of known technology – blood matching. In fact, Rubius is unwittingly, I think, repeating experiments in RBC alloimmunization with all the known consequences (risk of HDN in women) as well as introducing some additional ones – such as anaphylaxis. Rubius is not even in the clinic, but the burden on the company will be to show that this approach can generate a positive benefit/risk ratio. Maybe some of the other indications may be more favorable, but the next indication lined up is refractory gout - hardly life threatening (pipeline here). Lastly, even a cursory look at their concept of producing RBC displaying various antigens is likely to be silly. If it were so important to produce particles/cells which display certain payloads, you could do something like that by simply conjugating the target proteins to beads/liposomes/albumin etc, which would be far easier and cheaper. I think their other non-rare disease programs can be explained by the hypothesis – if you have a hammer, everything looks like a nail phenomenon.

Additional Comments

It baffles me that this company has garnered this much hype and such a high valuation (almost 2B at peak after the IPO) while not even being in the clinic. The issues I am discussing are based on findings from Phillip Levine and Rufus Stetson…who published their data in 1939. I may be wrong, but I remembered from college decades ago that there is no such thing as a universal donor, which triggered my interest in looking at RUBY. After looking at the literature, I think the burden is on the company to show that these issues I listed are not real. However, since they haven’t even submitted an IND, they haven’t gone that far and time will tell.

However, if I am right, I am shocked at the potential lack of critical thinking by the RUBY management team, RUBY board of directors, and institutional investors in RUBY (Fidelity, Janus, Franklin, etc.), the latter who are supposed to have qualified analysts conducting diligence on their investments. We can excuse the "reputable" investment banks (JP Morgan, Morgan Stanley, Jefferies, and Leerink) and sell side analysts (they launched RUBY coverage with titles such as “Rubies are Even “Rarer” Than Diamonds (Yee, Jefferies)” “The Next Wave of Cell Therapy (Harrison, Morgan Stanley), “Bloody Good Approach (Chang, Leerink)”), since we know their interests have nothing to do with bringing a product to market or helping biotech investors generate a return.

I am extremely bullish on biotechnology over the next 10 years. I’m convinced that we will cure more diseases in the next 10 years compared to the last 100. However, companies like RUBY waste capital which should be diverted to ideas that actually have a chance at success and make a contribution to society. I think Flagship Pioneering and other VC companies should consider focusing more on getting real, up to date scientists than hiring big names. The RUBY platform, in theory, engages technology on so many levels – gene therapy, stem cells, cellular therapy, and multiple diseases. The level of difficulty is dramatically higher in a business model like this. In the Flagship zeal to bring in big names, all the people involved didn’t stop to think that maybe you need to get a leader who has more scientific background than a bachelors in pharmacy earned in 1984. The CEO needs to take a lot of the blame here, since he was supposedly a practicing physician in transplant medicine. You would think he would know something about alloimmunization. So maybe that would be the silver lining here, where the worst thing that he can do here is lose investor's money.

I guess the final question is what is RUBY worth? It seems that after the 240M IPO, that RUBY is projected by Jefferies’ Yee to have about 328M on the balance sheet at the end of the year. Given that I believe that the core foundation of the company is flawed, I believe that there is no company. However, if you even value this as a pre-IND company, I don’t know how you could possibly value a company with their most advanced program at IND enabling (RUBY website) worth more than 100M in enterprise value. Rounding down, I’d say the company is worth between 300-400M (call it 350M blended). On November 9 I see that RUBY has a market capitalization of 1.25B. This means there is a little less than 1B in downside, or about 75% (my estimate of cash value of about 3.50-4.50/share) from Friday’s close.