top of page

The NIH’s Letter to Congress Reveals Less Than They Say, and More Than They Think

In their October 20 letter to the House Committee on Oversight and Reform, the National Institutes of Health (NIH) admitted to funding research, conducted by EcoHealth Alliance at the Wuhan Institute of Virology, to study “what factors increase the risk of the next CoV [coronavirus] emerging in people” and “allow animal Coronaviruses to evolve and jump into the human population.” (Quotes from EcoHealth Alliance’s NIH grant proposal, released by The Intercept, will be cited by their online page numbers. The above are from pages 14 and 15, respectively.)

EcoHealth, a U.S. nonprofit headed by Peter Daszak, would conduct this research in concert with Wuhan scientists by using “genomic characterization and isolation of novel CoVs” and analyzing “functional receptor binding domains,” also called cleavage sites (14).

Over the course of the still-active grant period, the NIH write, at least one experiment was performed in which a bat coronavirus was modified, possibly to express a human-specific cleavage site called a furin cleavage site (a rare binding receptor also found on the SARS-CoV-2 spike protein). The furin cleavage site, which suctions onto the ACE2 receptors of human cells, is now thought to be responsible for SARS-CoV-2’s high rate of human infectivity and transmissibility.

As a result of that experiment, the NIH write, “laboratory mice infected with the [modified] coronavirus became sicker than those infected with the [natural] coronavirus”. The modified coronavirus also produced “a one log increase in growth,” enhancing viral reproduction by a rate of 10,000.

In light of such revelations—and obfuscations—Congress should have at least 10,000 questions for the NIH.

Was It a Furin Site?

The furin cleavage site, not mentioned in the NIH letter, is a likely candidate for enhancement for multiple reasons. For one, the NIH write that the modified virus—called a chimeric or recombinant virus, because it fuses genetic material from multiple viruses—was used to test whether “spike proteins from naturally occurring bat coronaviruses circulating in China were capable of binding to the human ACE2 receptor in a mouse model”. Furin cleavage sites target ACE2 receptors; and one of the parent viruses used to create the chimeric virus the letter mentions is unique for its capacity to bind to human cells, specifically via ACE2 receptors.

For another, EcoHealth’s research was meant to analyze various binding domains, both wild-type and mutant, in order to glean insight about how coronaviruses jump from bats to humans. In particular, EcoHealth had proposed to “induce site mutations in S [spike] proteins,” which is also where the SARS-CoV-2 furin cleavage site is located (125).

Finally, the same year that EcoHealth Alliance received their NIH grant, they had drafted a nearly identical DARPA proposal, in which EcoHealth discussed engineering bat coronaviruses with furin cleavage sites in order to make them more transmissible to humans (sources: DRASTIC research [leaked DARPA proposal]; Vanity Fair).

Who or What Is Keeping the NIH From Saying “Gain-of-Function”?

While the NIH claim they determined that EcoHealth’s proposed research “did not fit the definition of research” formerly known as gain-of-function “because these bat coronaviruses had not been shown to infect humans,” it is becoming increasingly difficult to believe them. This is, in part, because the research described in the NIH letter involved humanized mice, which EcoHealth proposed only to use “if viruses [were] identified of significant human infection potential” (124).

Furthermore, the SHC014 virus, one of the bat coronaviruses that the NIH mention being used in the experiment—mere sentences before claiming they did not consider the research in which this virus was deployed “gain-of-function”—is unique precisely for its capacity to infect human cells, as gain-of-function researchers at the University of North Carolina first demonstrated in 2015. EcoHealth’s grant-writers were aware of this; and by the wording of their proposal, they seem to have intended to use the virus for these same infectious properties in the Wuhan.

The NIH letter mentions that (at least) one of the experiments funded and performed at the Wuhan Institute of Virology between 2018 and 2019 enhanced viral pathogenesis ten thousand-fold, making the engineered virus much more contagious and prone to mutation. The NIH frame this as some sort of unexpected accident, but judging by EcoHealth’s DARPA and NIH proposals, it sounds like precisely the sort of result at EcoHealth was aiming—something the NIH would have known when they authorized the grant.

Pathogens, being very simple organisms, already multiply far more rapidly than humans, insects, and other animals. Every time you increase an organism’s rate of reproduction by a factor of k, you also increase its likelihood of mutating by k, all else being equal.

As we’ll see, however, all else might not have been equal: SHC014, the virus known to infect human cells by binding to ACE2 receptors, was also known by EcoHealth researchers to enhance viral mutation dramatically (493).

By the time the NIH received and authorized EcoHealth’s grant, gain-of-function research was legal. By every standard imaginable, every facet of the research design, the experiment described, the aim identified, and the parent viruses used sound like a textbook example of gain-of-function research. What is the NIH afraid will happen if they concede that they knowingly funded gain-of-function research in Wuhan?

Were There More Experiments?

The NIH do not clarify whether the experiment they mention in their letter was the only experiment in which EcoHealth and Wuhan researchers modified bat coronaviruses to bind to human receptors (which the NIH attest is certainly not gain-of-function research, because binding to human receptors is certainly not something that would increase transmissibility to humans).

Indeed, it would be strange if this experiment were the only one of its kind, seeing as one of the main purposes of EcoHealth’s grant was—and remains—to study wild-type and modified bat coronaviruses with human-adapted cleavage sites in order to analyze their transmissibility to humans.

What do we know about the other experiments being conducted by EcoHealth and the Wuhan Institute? If we know nothing, why do we know nothing?

Where are the Missing Sequences?

There’s another, even more important question, however: in nearly every paragraph of their letter, the NIH emphasize that the natural viruses (hint) used in the aforementioned experiment just could not possibly be SARS-CoV-2; since, according to sequencing data, the natural viruses used in that experiment are only 96–97% identical to the SARS-CoV-2 genome. Curiously, the NIH only provide sequencing data for one of these natural parent viruses: WIV-1. Absent is the sequence for SHC014, which researchers combined with WIV-1 to produce an enhanced, chimeric virus—the one whose rate of pathogenesis exceeded that of its parents by ten thousand.

Where is the genomic sequence for SHC014? Where is the sequence for the enhanced, chimeric virus? Where are the genomic sequences for any of that virus’ offspring?

If you read their letter closely, you’ll notice that the NIH do not say whether they or EcoHealth sequenced the viruses produced by their experiments—only that they sequenced the viruses used to initiate one of their experiments. They then fail even to provide the complete pair of parent sequences, attaching the sequence for only one of the naturally occurring viruses in question.

According to that genetic analysis, the NIH write, “the naturally occurring bat coronaviruses used in experiments under the NIH grant… are decades removed from SARS-CoV-2 evolutionarily” (emphasis added).

We know.

This is precisely why no one asked about the naturally occurring bat coronaviruses. We are well aware that naturally occurring bat coronaviruses are decades removed from SARS-CoV-2 evolutionarily. We have been saying that for nearly two years.

What we do not know, and what we would very much like to know, are the genetic sequences of the chimeric virus and its ten thousand times more pathogenic offspring.

That recombinant virus, engineered by Frankensteining together various pieces of multiple viral genomes, is named after its two parents: SHC014 WIV-1. Because we know this virus’ name, we know a bit about it. WIV-1 is the SARS-like coronavirus whose individual sequence the NIH attaches in their letter to Congress. On its own, we learn, WIV-1 is between 96 and 97% genetically identical to SARS-CoV-2.

SHC014, whose sequence is not attached, is a relatively novel virus (or viral category) that EcoHealth researchers have used in experiments for several years. In 2012, an EcoHealth group first isolated this novel SARS-like bat coronavirus and noted something rare and remarkable about it: although SHC014 was a bat coronavirus, it “use[d] the ACE2 receptor [that] SARS-CoV uses to infect human cells” (115).

In fact, SHC014 is the first bat coronavirus ever shown to infect human cells in a laboratory setting. This occurred in 2015, when researchers created a chimeric virus by recombining an SHC014 spike protein with a SARS-like coronavirus. EcoHealth alluded to using the same process to carry out their experiments in their NIH proposal.

EcoHealth had also discovered that “SHC014 is readily divergent from SARS-CoV wildtype [the background strain of known SARS-like coronaviruses]” (493), making it a useful tool for developing more advanced technologies that could test for and sequence other novel coronaviruses (another aim of the proposed research).

“Divergent,” in evolutionary biology, is not a synonym for “different”. “Divergent” means “prone to rapid mutation”. If SHC014 is “readily divergent” from SARS-CoV wildtype, that means its offspring mutate quickly from naturally occurring SARS-like coronaviruses, and they spawn further generations of offspring that also mutate quickly. In other words, the SHC014 offspring’s genome will diverge from the wildtype SARS-CoV genome at a much faster rate—which could bring its offspring closer to SARS-CoV-2 at a much faster rate.

Were either SHC014 or SHC014 WIV-1 ever sequenced? Were SHC014 WIV-1’s offspring sequenced? Is it standard practice to sequence viruses produced by experiments intended to enhance viral transmissibility? If not, why not? If so, why are the NIH not in possession of the SHC014 WIV-1 sequence?

How Quickly Could SHC014 WIV-1 Mutate?

For the sake of argument, however, let’s assume that SHC014 WIV-1 was, like WIV-1, 96–97% identical to the SARS-CoV-2 genome at the time of recombination. The difference between 96% and 100% genetic identity is about the difference between humans and chimpanzees (or, as we’ve recently learned, bonobos), who diverged from a common ancestor approximately six to seven million years ago. In other words, we would have to travel back as many as seven million years to make up for that particular three to four percent variation in our genome.

The NIH point this out, presumably to make us reflect on just how many eons of daylight there are between the naturally occurring bat coronaviruses they studied and the SARS-CoV-2 pathogen. What’s funny is that, if anything, this dampens the “natural origins” theory for SARS-CoV-2. Mutating in the wild, a viral genome can undergo several changes in far less time than a human genome can, although very few mutations are adaptive or genetically significant. (EcoHealth knew this, too, as we’ll soon see.)

When we accelerate that rate of evolution by ten thousand, however, we see far more consequential mutations in far less time. Couple that with a virus known for its ready divergence, and…

Lest we reflect too much, however, the NIH slip in one of public health’s most obvious “tells”: “experts agree.” Specifically, “experts agree that these [natural] viruses”—the natural viruses, even when combined and enhanced to increase pathogenesis ten-thousandfold—“are far too divergent to be the progenitor of SARS-CoV-2.”

If I were questioning the authors of this letter, I would ask to see the names and credentials of these experts.

Of course, there’s nothing wrong with experts agreeing on something. The problem is that the specific phrase “experts agree”—which is the leftward equivalent of former President Trump’s blustering equivocation, “people are saying…”—invariably raises more questions than it attempts to snuff out. How many experts agree? Which experts? In what? Surely not viral pathology. Surely not evolutionary biology. Surely not the natural sciences or natural history. Because presuming an equivalence between human and pathogenic evolution would be misleading, to say the least.

First of all, as already stated, pathogenesis is far more rapid than multicellular organism reproduction. The orders of magnitude at which viruses reproduce and mutate make their evolutionary process impossible to compare with hominid evolution.

For illustrative purposes only, however, we can draw on some basic data. As a word of warning, it is impossible to make assumptions about viral evolution (or even future human evolution) based on one very limited episode of hominid evolution. The intent here is not to pinpoint a specific time frame in which an enhanced virus would have evolved from 96 to 100% genomic identity with SARS-CoV-2. The intent is only to show that the window in which we could reasonably estimate such evolution to occur is infinitesimally smaller than the time frame in which a given population of hominids evolved from 96 to 100% genomic identity with homo sapiens.

The difference is one of millions of years to days.

Humans’ current rate of mutation is estimated at about one substitution for every 60 million nucleotides. This averages to about 100–200 mutations per generation, most of them inconsequential. (The mutations, not the generations. Of course.)

Returning to our genetic kinship with bonobos (which I prefer to chimpanzees, since they are more libertarian), this means that we are seven million years away from four centiles of genomic identity, assuming an average mutation rate of 1:60 million nucleotides.

RNA viruses, on the other hand, mutate at a rate of one substitution for every 1,000 nucleotides.

Mutating at that rate, compensating for a four percent deviation in genomic identity could take days to weeks.

It could also take a year. It’s impossible to tell how many of those 1:1000 nucleotide variations could be significant—but the more of them there are, the more likely significance becomes. This is why, in the wild, “experts” might agree that any of the natural viruses the NIH sequenced are “decades removed from SARS-CoV-2 evolutionarily.”

When you divide those decades by 10,000, however, you get a far more rapid evolutionary timeline. Which, presumably, any expert would know.

This includes the experts at EcoHealth Alliance, who, in their grant proposal to the NIH, wrote: “we expect [mutation] to take small incremental movements in a fitness landscape that may have low fitness valleys [that is, not be conducive to pathogenesis] between a wild-host adapted strain and a human or other host adapted strain.” Viral evolution via laboratory recombination, however, “may be able to take larger leaps across a given fitness landscape” (123).

In plain English, this reads: mutation from a nonhuman-host virus to a human-host virus will be slow in the wild, but faster if we combine viruses that have nonhuman and human-adaptive traits (like ACE2 receptor bonding, as in the case of SHC014).

Indeed, on pp. 122–123 of EcoHealth’s grant application, the authors appear to propose using the recombination process to engineer more rapid evolution than would be found in wild-type viruses.

So, yes: the natural viruses the NIH sequenced are “decades removed from SARS-CoV-2 evolutionarily.” That’s the point. That’s why EcoHealth chose to engineer recombinant viruses, in their own words.

What we need now are the sequences of the engineered viruses and their offspring.

On average, RNA mutations aren’t as significant as DNA mutations; otherwise, we would see extraordinary deviations in SARS-CoV-2 pathogenic sequences taken two months apart. Instead, we’re seeing adaptive mutations at a rate of about four per year.

Four adaptive mutations per year, however, is still quite a bit for a virus. Perhaps the NIH’s letter unwittingly gives us some insight as to why.

Who’s Looking for Answers?

Hypothetically speaking, the 2018–2019 experiments referenced in the letter could well have produced a virus that was four percent divergent from today’s SARS-CoV-2 sequence; and, within a span of days, weeks, or months—certainly by late 2019—that virus could well have evolved into the initial strain of SARS-CoV-2, independent of researchers actually doing anything to modify the virus further. That’s especially the case since the NIH admits that EcoHealth’s experiments in Wuhan involved a novel coronavirus known for rapid genetic divergence and ended up expediting an already exponential pathogenesis rate by 10,000 times.

That’s a lot of pathogenesis: for every pathogen that the original, wild bat coronaviruses produced, the researchers’ recombinant pathogen produced 10,000 new pathogens. This means 10,000 opportunities for mutation for every one opportunity given by the original virus.

That gives lawmakers a rich list of questions to ask and documents to subpoena, should they choose to open an investigation.