Prometheus Unbound II: Does Prometheus’ Claim Recite a Law of Nature?

By: Robert R. Sachs, Jennifer R. Bush

In our previous post, we explored the philosophical aspects of what is a law of nature. We can now apply that analysis and consider whether Prometheus’ claim recites a law of nature under the variously described criteria. Here is the claim 1 of the ‘623 patent:

1. A method of optimizing therapeutic efficacy for treatment of an immune-mediated gastrointestinal disorder, comprising:

(a) administering a drug providing 6-thioguanine to a subject having said immune-mediated gastrointestinal disorder; and

(b) determining the level of 6-thioguanine in said subject having said immune-mediated gastrointestinal disorder,

wherein the level of 6-thioguanine less than about 230 pmol per 8x10⁸ red blood cells indicates a need to increase the amount of said drug subsequently administered to said subject and

wherein the level of 6-thioguanine greater than about 400 pmol per 8x10⁸ red blood cells indicates a need to decrease the amount of said drug subsequently administered to said subject.

The Supreme Court’s analysis of this claim in its entirety is:

Prometheus’ patents set forth laws of nature—namely, relationships between concentrations of certain metabolites in the blood and the likelihood that a dosage of a thiopurine drug will prove ineffective or cause harm. Claim 1, for example, states that if the levels of 6–TG in the blood (of a patient who has taken a dose of a thiopurine drug) exceed about 400 pmol per 8x10⁸ red blood cells, then the administered dose is likely to produce toxic side effects. While it takes a human action (the administration of a thiopurine drug) to trigger a manifestation of this relation in a particular person, the relation itself exists in principle apart from any human action. The relation is a consequence of the ways in which thiopurine compounds are metabolized by the body—entirely natural processes. And so a patent that simply describes that relation sets forth a natural law.

The Supreme Court assumes that there are laws of nature and concludes that the specific dosage-toxicity relationship claimed is such a law. The structure of the Court’s reasoning is that:

1) there is a natural relationship between the concentration of the metabolite and its effective/toxic dosages;

2) expressing this relationship in if…then form is important because it operates as a prediction of what will happen; and

3) the relationship exists apart from human action.

The Court goes to hold that to be eligible for a patent, something more than merely “applying” a natural law must be claimed. The Court found that Prometheus claimed nothing more than the law itself.

First, let us find the “natural relationship” between 6-TG and effective and toxic doses, which formed the basis for the alleged law of nature. As stated in the ‘623 patent, this relationship is based on two studies of patients who received daily administrations of 6-mercaptopurine (6-MP) or AZA for at least four months. One study had 89 patients and the other had 93 patients. Hepatic, pancreatic and hematological tests for toxicity were obtained every three months. The results of the first study are shown in Table 1 from the ‘623 patent:

In the first study, 58 patients showed specific clinical improvement, "so called responders." Of the responders, 78% (45) had 6-TG levels >225 pmol per 8x10⁸ RBC. In the second study (not shown here), 78% of patients above a median 6-TG of 230 pmol/8x10⁸ RBC were responders. From these findings comes the claim limitation that 6-TG less than “about 230” “indicates a need” to increase the dosage.

Now consider the toxicity level. Table I shows that just six people out of 89 (6.74%) experienced hematologic toxicity; for them, the median 6-TG level was 414 pmol per 8x10⁸ RBC.There is no direct statistical evidence that the concentration should be about 400; that number was most certainly selected by the patent attorney to give some leeway in the patent claim. These six patients are very impressive folks indeed, since they (no doubt unknowingly) form the smallest sample of data upon which a “law of nature” has ever rested.

The inventors of ‘623 patent, Drs. Siedman and Theoret, made no claim that they discovered any law of nature—instead, it was the Supreme Court that gave this imprimatur to the claim. One would think that, if they inventors believed they had discovered something as fundamental as a law of nature, they would have said so.

Applying some of the above criteria shows what Drs. Siedman and Theoret “discovered”—if they discovered anything at all—was an accidental generalization. First, as Table I shows, neither the effective nor the toxic dosages approach anywhere near universality. While the effective dose need not result in a 6-TG level of about 230 in, say, 99% of patients, being effective for less than 80% seems a rather weak condition, especially when 26% of the non-responders had a 6-TG level greater than 225. Indeed, it is easy to see that the claimed “about” could have been shifted one way or the other by 5 or 10 pmol with little real scientific impact. Real “laws of nature” are not so malleable. Similarly, it can be argued that the claimed minimum dosage level could have been set lower, perhaps to 220 pmol per 8x10⁸ RBC. That would have provided superficial support for an assertion that the level was “universal.” Finally, that the confidence level for this finding was p<0.001 simply means there is less than 1 chance in 1000 that this is a random outcome. Thus, there may be a strong correlation, but that does not tell us there is a necessary one.

The toxic dosage is even more suspect, as just six people experienced toxicity in the first study, and the toxic level of about 400 was not even the average, but the median level. With a sample that small, the median can be an unreliable measure of central tendency. Thus, for another group in another study, the results not merely could have, but very likely would have, been different to a noticeable degree. Indeed, this is acknowledged in the Prometheus decision itself: Mayo’s own research led them to pick 450 pmol per 8x10⁸ RBC as the toxic level. Are there then two laws of nature of 6-TG toxicity, Prometheus’ Law and Mayo’s?

Nor does the natural relationship have any sort of necessity. In line with the laws of biology, the particular ability of humans to metabolize 6-MP or AZA into 6-TG at all is a historical accident of evolution. Given that 26% of the patients in the first study metabolized these drugs too poorly to have any therapeutic effect, it is not beyond reason to think we could have evolved to be entirely unable to metabolize 6-MP or AZA. But we do not even need this strong an outcome; we need consider only that humans could certainly have evolved so that the effective or toxic dosages were different from that which was claimed, say both higher by 25%. This demonstrates the counterfactual analysis, in which there is a world like ours in all respects, except that we find that the alleged law of nature does not apply.

The claimed natural law does not explain anything either. There is no causal explanation why the dosages are effective or toxic; there is only a statement of what they are. Once the causal mechanism is discovered, the particular levels may fall away as unnecessary.

Predictive power also marks a law of nature. Here, Prometheus’ claim does provide a testable prediction. Using the data from Table I and Bayes’ Theorem, we can determine the probability that a patient will demonstrate a clinical response to 6-TG, given a concentration above 225 pmol 8x10⁸RBC, is approximately 85%.^[1] That means about 15% of the time patients treated in accordance with the patient's claim will not show any clinical response at 225 pmol 8x10⁸RBC. In that case, what is a doctor to do? As Table I shows, there were seven patients who did not respond and who developed hepatic toxicity; of these, five (71%) had 6-TG levels greater than 225. Thus, if the doctor increases the dosage, there is significant probability that the patient will have a toxic reaction to 6-TG at level well below the claimed “about 400.” Thus, even on its own face, claim 1 does not present a certain path for a successful treatment.

The next criterion above is objectivity, that the claimed law does not depend on human knowledge, beliefs, needs or other pragmatic concerns—in short that like E=mc² or F=ma it is true regardless of any facts about the human condition. Here, the Supreme Court’s assertion of lawhood fails miserably: The claimed law is based on a human need to modify a dosage of drugs for a specific disease—drugs that were invented by humans in the first place. No other generally accepted law of nature in physics, chemistry, etc. has any such intimate connection to an unmistakable subjective human need. Assuming that one believes, like Descartes, in a Creator who fashioned the universe with a fixed set of laws that could be rationally discovered, it seems quite out of place to suppose that the Creator instantiated a specific law for a specific drug for a specific disease—and necessarily would then have to have done so for every possible drug and disease. Einstein famously said that “God does not play dice with the world,”^[2] but it can be equally said that God did not jot down every possible biological correlation that could and would be discovered.

Finally, we come to the “scientific” requirement, that a law-statement is the kind of thing that scientists would consider a law of nature. As we’ve seen above, Mayr tells us that biologists have little care or use for calling statements laws of nature—they get on with their work quite nicely without such a label.

We have sketched out a number of reasons that Prometheus’ claim is not a law of nature any more than (and indeed quite possibly less than) it is a law of nature that robin eggs are greenish-blue. Natural relationships surround us, but being natural does not make them laws of nature. For example, “adult male Chihuahuas are smaller than adult male Great Danes” states a natural relationship which “exists in principle apart from any human action,” but that does not strike us as law of nature. Nor does stating distinctively chemical or biological relationships in conditional form suffice. “If the blood alcohol level is less than about 0.08%, then the central nervous system motor function are substantially impaired,” may be a true generalization of the natural relationship otherwise known as being drunk, but hardly seems lawful—as many a philosopher can attest. ^[3]

The Supreme Court’s recent decision in Association for Molecular Pathology, et al. v. Myriad Genetics, Inc., et al. 569 U.S. ___ (2013) demonstrates that the Court continues to engage in a superficial scientific analysis of patent claims, and is not mindful of the unintended consequences of its logic. The Court ruled that isolated genomic DNA (gDNA) sequences were not patent-eligible, while isolated complementary DNA (cDNA) made from gDNA was patent-eligible. The Court held that isolated gDNA was still a “product of nature” in part because the informational content of the isolated DNA was the same as DNA in the body. For cDNA, the Court held that this was “man-made,” but then ignored the fact that the informational content of cDNA is likewise the same as its gDNA source, since what is excised in the cDNA is the non-coding regions, leaving the relevant coding regions that code for the same protein as the underlying gDNA. Further, the Court overlooked the fact that in many cases, the man-made elements introduced into cDNA are common and routine, the very type of activity that the Court condemned in Prometheus.

Finally, the Court stated that there may be “applications” of isolated gDNA or cDNA that may be patent-eligible. But they did not consider that Prometheus foreclosed an important type of applications: genetic testing based on correlations between given DNA sequences and particular diseases; under Prometheus such “correlations” are simply laws of nature. Thus, we can expect future litigation to assert that cDNA is unpatentable as well.

The Court’s analysis in Prometheus is both unwise and unfortunate. Unfortunate, because the analysis as to what is a law of nature sweeps far too broadly, and thus potentially can invalidate patents in a variety of scientific and technological disciplines because they use specific scientific facts or “natural relations.” Unwise, because the Court’s superficial treatment sets a standard that other courts, as well as the U.S. Patent and Trademark Office, will mimic in implementing the Court’s ruling in future cases and future patent applications.

[1] Where P(A) is the probability of being a responder (58/89), P(B) is the probability of having 6-TG concentration > 225 (53/89), and P(B|A) being the probability of having concentration > 225 given being a responder (45/58). There is insufficient details in the ‘623 patent regarding the second study to do the Bayesian analysis for the recited level of 230 pmol 8x10³ RBC. Since the number of responders having a 6-TG level > 230 pmol 8x10⁸RBC is necessarily less the number of responders having level >225, the analysis in the text is actually conservative, i.e. it overstates the probability of a patient being a responder.

[2] Hermanns, Einstein and the Poet: In Search of the Cosmic Man (1983) Brendan Press: Massachusetts; p. 58.

[3] See, e.g., Bruce, Bruces’ Philosophers Song (1973): Univ. Woolamaloo, Sydney, Australia, as recorded in The Monty Python Matching Tie and Handkerchief, (1973), Charisma/Arista. See, http://www.montypython.net/scripts/bruces.php. See also Hardcastle, Monty Python and Philosophy: Nudge, Nudge, Think, Think! (2006) Carus Publ. Co: Illinios.

*The perspectives expressed in the Bilski Blog, as well as in various sources cited therein from time to time, are those of the respective authors and do not necessarily represent the views of Fenwick & West LLP or its clients.