An inherent defect in the modified RNA (modRNA) instructions for the spike protein in COVID-19 immunizations causes cells to produce “off-target” proteins in addition to the spike. These proteins, which developers either failed to look for or did not report to regulators, cause undesirable immune responses whose long-term effects are unknown.
Developers of the Pfizer/BioNTech BNT162b2 gene therapy immunization product (“vaccine”) overlooked or failed to anticipate a common mistake in the translation of synthetic messenger RNA (mRNA) to protein.
This mistranslation may cause the expression of potentially harmful protein side products. As a result, some people who receive the shots develop immune responses to those protein products — and scientists don’t know what, if any, long-term consequences may ensue.
The paper reporting this effect appeared in Nature in December 2023, three years after the mRNA-based vaccine rollouts and the potential to address these potentially dangerous effects had passed — but not before more than 600 million doses of the mRNA vaccines had been administered in the U.S.
Researchers including first author Thomas Mulroney, Ph.D., a Cambridge University toxicologist, found that N1-methylpseudouridine (NMpU), an artificial instruction inserted into mRNA and other mRNA-based gene therapy products, causes the machinery that translates the gene to the spike protein to “slip” about 10% of the time.
Slips, termed “frameshifts,” cause cells to skip an instruction and generate random, nonsense proteins leading to undefined, unintended immune responses.
“We don’t know what these proteins are or their toxicities, yet there was no sense of urgency in making this discovery known. While peer review can take months for papers to be published, given the implications of these findings, ten months after receiving a manuscript is an awfully long time. Publication should have been expedited, regulatory authorities notified and actions taken.”
The immunogenicity of protein contaminants is a serious issue in biotech-made medicines, particularly for injected or infused products. For example, manufacturers devote considerable resources to eliminating host cell proteins due to their potential to cause dangerous and unpredictable immune reactions in patients.
Host cell proteins and their breakdown products are impurities originating from cells or organisms used to manufacture protein products, but they do not belong in the product. Some have been characterized but many, like frame-shifted proteins, are not.
Lost in translation
mRNA instructions are “written” using a four-letter code comprised of A (adenine), C (cytosine), G (guanine) and U (uricil). We tend to think of genes as stable, as they exist and operate in all our cells throughout our lives, but mRNA is rapidly degraded by enzymes once it enters cells.
To prevent this degradation and to improve spike protein expression, vaccine developers replace uricil with NMpU, an artificial mRNA building block that appears in the Pfizer/BioNTech and Moderna COVID-19 immunization products. This kind of substitution is not possible or relevant to DNA “vector”-based vaccines (e.g. Johnson and Johnson’s) or conventional antigen vaccines (Novavax).
Frame shifts produce new and in most cases previously unknown proteins, as illustrated by this example:
The cat sat on the mat (normal message)
Hec ats ato nth ema t (frame-shifted message leaving out the initial letter, “T”)
The “frame-shifted” English language example is gibberish. However, in genetics, frame-shifted sequences may have meaning in their ability to generate a real protein capable of evoking an immune response.
Suspecting the NMpU replacement was responsible for frame-shifting, Mulroney et al. created pseudouridine-based mRNA instructions for a test frame-shift protein and fed them to commercial laboratory cells. Along with the expected test protein, the cells produced other, unexpected proteins corresponding to a “frameshift,” or loss of one genetic instruction.
Mulroney then vaccinated one group of mice with Pfizer’s BNT162b2 COVID-19 vaccine, another with Astra-Zeneca’s ChAdOx vaccine (which does not use pseudouridine), and a third with a placebo injection. After eight days the immune systems of mice receiving BNT162b2 recognized the frameshift protein, but mice who got the other treatments did not.
This means BNTb162b-injected mice had seen the gibberish protein before, as a result of their vaccination.
Mulroney then repeated this experiment, but instead of mice, he used immune system cells collected from 21 people who had taken the Pfizer vaccine and compared those cells to immune cells from 20 individuals who received the Astra-Zeneca vaccine (which does not use NMpU).
Both groups showed immunity to the spike protein but one-third of the Pfizer group also showed immune recognition of the frameshift protein.
This means the frame shift and production of associated proteins occurs only in subjects receiving BNTb162b, and that humans and mice both develop immunity to it.
Mulroney then designed a novel mRNA sequence to reduce frame-shifting. He found frame-shifting and off-target effects significantly reduced with no effect on production of the intended protein.
Mulroney et al. noted that while frame-shifted proteins in humans could not be linked to adverse events based on his findings — likely because none of his subjects were experiencing them — future mRNA developers should pay attention to possible frame-shifting, particularly for products requiring high or frequent dosing.
They wrote that knowing which regions within the mRNA instructions are responsible for frameshifts enables the design of safer, more effective mRNA products.
Suspecting that these regions involved NMpU, Mulroney et al. created NMpU-based mRNA instructions for a test protein and fed the gene to commercial laboratory cells. Along with the test protein he found several other proteins corresponding to a “frame shift,” or loss of one instruction.”
Wiseman’s take: a high-quality study revealing development flaws
Mulroney’s was a high-quality study, Wiseman said. “The authors are affiliated with the U.K.’s Medical Research Council’s toxicology unit and the National Institute for Health Research, equivalent to the U.S. National Institutes of Health. That such high-quality work spent so long — ten months — in editing suggests that its messaging was heavily worked on.”
Nature received the paper on Jan. 25, accepted it for publication on Oct. 31, and published it on Dec. 6. Wiseman described this timeline as “protracted” given the appearance of off-target proteins that he considers reportable deviations from the expected product.
Also, Nature usually appends a peer-review file to its articles but none appeared with the Mulroney study on its initial publication, although they are there now.
“I think they had more data that they didn’t disclose,” Wisman said. “For example, there was no mention of levels of antibodies to frame-shift proteins. Is it possible the authors of an immunology study didn’t think of measuring antibody levels? That would be really weird because discussions in FDA meetings on the C19 vaccines focussed on B cell/ Antibody responses.”
Mulroney also revealed the failure of Pfizer and regulators to view gibberish proteins as undesirable side products.
Impurities in pharmaceutical products are inevitable because chemical and biochemical manufacturing processes always generate side products. Manufacturers control for possible ill effects by identifying, quantifying and characterizing them according to regulatory guidelines.
Process development scientists spend most of their time tweaking recipes to improve yields and eliminate or reduce impurities, but that is not possible with mRNA products because the “manufacturing plant” is not a brick-and-mortar factory but human cells.
“mRNA vaccines essentially outsource manufacturing to your cells,” Wiseman said. “Sometimes it goes according to plan but other times it doesn’t.”
One Mulroney coauthor, Dr. James Thaventhiran, an immunologist with the MRC Toxicology Unit at the University of Cambridge, acknowledged in a Cambridge University press release that misdirected immunity targeting protein impurities “has huge potential to be harmful.”
However, Mulroney et al. concluded there is no evidence that frameshifted products in humans generated from BNT162b2 vaccination are associated with adverse outcomes.
“This claim is unsubstantiated and disingenuous,” Wiseman said, given that Mulroney was not a controlled trial and the number of vaccinated subjects providing samples was small. “And since no subjects reported vaccine-related side effects their sample is subject to selection bias. A more comprehensive series of studies needs to be done to characterize the toxicological effects of these proteins.”
The failure to recognize the significance of protein contaminants also bucks World Health Organization’s (WHO) mRNA vaccine guidelines, which require manufacturers to provide details on “unexpected ORFs” — “open reading frames” that can include skipped or frameshifted mRNA instructions.
According to the WHO, “The complete annotated sequence identifying all ORFs (including any unexpected ORFs) and all other sequence elements (including their justification for use) should be provided.”
Since Mulroney’s was the first paper to address frameshifted “ORFs” — and three years after BNT162b’s debut — either Pfizer ignored this requirement or FDA did not ask for it.
“The toxicology of these unintended proteins must be studied,” Wiseman said. “The developers should have isolated them, sequenced them, and noted any important similarities between frame-shifted proteins and naturally-occurring peptides and proteins.”
Pfizer executive: We flew the plane ‘while we were still building it’
Wiseman told The Defender that the Mulroney study revealed a failure by Pfizer’s development team and regulators to ask fundamental questions regarding the safety and effectiveness of BNT162b2.
Wiseman said no statement illustrates this more than a comment by the retired head of Pfizer’s vaccine research and development, who told Nature Reviews Drug Discovery, “We got creative — we couldn’t wait for data, we had to do so much ‘at risk.’ We flew the aeroplane while we were still building it.”
“If Mulroney’s team could predict the existence of frameshifted proteins, why were Pfizer’s scientists unable to do so? The same question may be asked of regulators, especially in light of unresolved discrepancies and the specific obligation imposed by the European Medical Agency on BioNTech regarding the identities of the proteins elicited by the vaccine
“We must assume UK regulators, manufacturers, and international regulators, including FDA, knew of these issues many months ago. We await their investigation on the toxicity of off-target proteins, why they were not discovered or reported sooner, what steps FDA has taken to prevent future harms, and whether they plan to inform the public of these findings.”