Vaccines and Autoimmune Disease
The Short Version: Recently I’ve observed an increase in the number of concerns that vaccines for COVID-19 will cause autoimmune disease. This concern is not based in sound science. For a vaccine to cause autoimmune disease would require many multifaceted layers of protection to fail, and historical precedent does not justify these concerns. Even in the very few rare cases in which a vaccine has been associated with an autoimmune disease, the epidemiological evidence generally cannot make a definitive judgment on whether or not the events are causally linked because they are too rare. On the other hand, there are many examples of infections, including COVID-19, provoking new autoimmune diseases, and thus your risk in avoiding a COVID-19 vaccine is actually likely greater because you would be put at higher risk of contracting COVID-19.
Rose NR, Mackay IR. The Autoimmune Diseases, 6th Edition.; 2020. Figure 5.2, summarizing central tolerance in T cells. The fate of a T cell depends on how strongly it is able to bind to presented antigen in the thymus (avidity). Tissue-restricted antigens are presented such that T cells are assessed for responsiveness. The absence of a response from a T cell indicates that T cell is incapable of contributing to host defense, and thus does not receive survival signals (death by neglect). Cells that are slightly responsive to self-antigen can be valuable in host defense, and thus are selected for, which is called positive selection. Cells that respond somewhat strongly undergo functional deviation to become T regulatory cells that suppress immune responses on encounter with that antigen (this is known as agonist selection). If cells respond even more strongly than this, they are likely to produce autoimmune disease if allowed to exit from the thymus and are killed (clonal deletion). Central tolerance results in the death of more than 95% of T cells that undergo it. The key point here is: T cells have to fall into a goldilocks zone with how well they recognize antigen to even be able to survive, and they still get disciplined if they don’t do it well.
Lately I am seeing a huge spike in the number of people concerned that the vaccines for COVID-19/SARS-CoV-2 are going to give them an autoimmune disease. I’d like to talk about that for a bit.
Husebye, E. S., Anderson, M. S., & Kämpe, O. (2018). Autoimmune Polyendocrine Syndromes. The New England Journal of Medicine, 378(12), 1132–1141. Figure 1, demonstrating the manifestations of autoimmune polyendocrine syndromes. APS-1 results entirely from a near-total failure of central tolerance mechanisms. IPEX syndrome, which is more severe in nature, results from a near-total failure of peripheral tolerance mechanisms.
Autoimmunity, Autoreactivity, Autoimmune Disease and the Frameworks of Immunological Function in brief
Firstly, it’s important to be accurate with our terminology:
Autoimmune disease: disease resulting from the aberrant action of the immune system to respond to self-antigen
Autoimmune disease is commonly called the result of an overactive immune system. This is absolutely one surefire path to autoimmune disease (stimulating your immune system to the point of overwhelming self-tolerance mechanisms), but fundamentally the response is more about the wrong target in most cases, rather than overexuberance.
Autoreactivity: the ability of a cell of the immune system to react to self-antigen
Autoimmunity: the state of having autoreactive cells, regardless of whether or not autoimmune disease results (this is NOT the same as autoimmune disease)
Murphy K, Weaver C. Janeway’s Immunobiology. 9th ed. CRC Press; 2016. Table 15.2, showing the numerous suppressive mechanisms that safeguard against autoimmunity.
Self-tolerance is the term for the ability of our immune system to not react strongly to our own self-antigens. In old frameworks of immunological function, the immune system was thought to respond to antigen by distinguishing self from non-self (concepts of self in immunology actually get very granular and almost metaphysical, as the linked review can show you). A series of classical experiments showed that if you inoculated animals with self-antigen, they would not respond against it, but they had no trouble responding to non-self. This model was very useful for understanding, for instance, organ transplants. In the case of an organ transplant, you receive a tissue from another immunologically distinct individual. If it’s from the same species (i.e. another human in your case) it’s known as an allograft. If it’s from another species (which is sometimes done- think heart valves from pigs or cows), it’s known as a xenograft. Sometimes you might get lucky and have an identical twin who can donate tissue to you, it’s said to be a syngeneic, and in general you shouldn’t require immunosuppressants like you would for an allogeneic graft. Further research demonstrated that indeed, cells of the adaptive immune system- B and T cells - undergo complex programming (education) to prevent response to self-antigen in the bone marrow and thymus respectively. This is known as central tolerance (because it occurs in the central, or primary, lymphoid organs). Further evidence that this was important for preventing autoimmune disease came from autoimmune polyendocrine syndrome type 1. In this condition, there is a loss-of-function mutation in a protein called AIRE (autoimmune regulator) which normally causes cells of the thymus to express tissue-restricted antigens (literally antigens that you would find only on certain tissues e.g. thyroid proteins, pancreas proteins, etc), allowing cells to be programmed and selected against strong response (this is called negative selection); some additional details are given in Figure 5.2 from Mackay and Rose above. The result of the lack of functional AIRE is the escape of multiple cells that are primed to respond to self-antigen, resulting in autoimmune disease against multiple organs. B cells do not rely on AIRE for their central tolerance, which instead has more indirect mechanisms. This seems like a pretty good model, but it’s missing a lot. For one thing: we have a microbiota. We have communities of bacteria, fungi, viruses, and even archaea, which are decidedly not self. Except somehow, our immune system tolerates their presence. There are also some important experimental outcomes which cannot be explained simply by the self-nonself framework. For example, the linked study describes how, in humans, T cells were made to tolerate influenza proteins after presentation from immature dendritic cells. Influenza proteins are decidedly not self, yet in this study, they are made to be treated as self. What’s going on?
There is a second layer to self-tolerance. And it is actually more important than central tolerance: peripheral tolerance. Peripheral tolerance refers to the suite of mechanisms available outside the central lymphoid organs to prevent your immune system from attacking your own tissues. It is clearly not enough for your immune system for something to be nonself for it to respond. Enter: the danger model. In the study linked earlier in which the patients developed a tolerance to influenza proteins, there’s a key point: the dendritic cells that were used to present the proteins were immature. Dendritic cells are the link between the innate and adaptive immune system. They are specialized antigen-presenting cells that actively take up antigen from the environment and present them to T cells. Immature dendritic cells cannot stimulate effective immune responses. In fact, they do the opposite. They suppress immune responses. Maturation of dendritic cells requires danger, by which I of course mean, danger signals: something that the immune system can recognize to indicate that indicates damage to the body, and thus the need to respond to a threat. There are many categories of danger signals, but the best characterized are the PAMPs (pathogen-associated molecular patterns) and DAMPs (damage-associated molecular patterns, aka alarmins). Sometimes, pathogens have certain molecules on them that fundamentally should not be inside people, and so these alone are sufficient to trigger immune responses. These are known as PAMPs, and one of the best understood ones is lipopolysaccharide (LPS), or endotoxin. Even a very small amount of endotoxin in the blood is sufficient to cause inflammation severe enough to cause septic shock (technically distributive shock/systemic inflammatory response syndrome, unless the source is a pathogen as opposed to pure endotoxin). On the other hand, some molecules in our own cells have a very specific localization, and if they leave it, it suggests damage. For example, uric acid crystals (the very same ones as in gout) come from our own cells, and trigger painful inflammation, even under sterile conditions. In the course of an infection, pathogens cause damage, liberating all sorts of DAMPs and themselves containing PAMPs. After cues from the innate immune system, dendritic cells will undergo maturation, and begin to express costimulatory molecules (molecules that license immune responses) on their surface as they find a T cell. From there, an immune response can progress, recruiting B cells, and cytotoxic T cells as well.
Already, we have gone through many protective layers against autoimmune disease but they are worth recapitulating.
ElTanbouly MA, Noelle RJ. Rethinking peripheral T cell tolerance: checkpoints across a T cell’s journey. Nat Rev Immunol. Published online 2020. doi:10.1038/s41577-020-00454-2 Figure 1, demonstrating checkpoint mechanisms against autoimmunity in the lifespan of a T cell.
Rosenblum MD, Remedios KA, Abbas AK. Mechanisms of human autoimmunity. J Clin Invest. 2015;125(6):2228-2233 Figure 2: Autoimmune disease does not just happen. It requires a complex series of risk factors and chance events. Firstly, there needs to be a genetic susceptibility to the autoimmunity in question, for instance, as in APS-1, absent functional AIRE. There then needs to be some kind of environmental trigger to take extant autoreactive cells and precipitate an immune response from them, such as an infection. Finally, peripheral tolerance mechanisms have to fail to suppress the resultant tissue damage. This is not something that any arrant immune response is capable of doing. In fact, typically when infections do precipitate autoimmune disease it is due to the presence of structurally similar antigens (molecular mimicry).
Firstly: central tolerance. Central tolerance works to prevent strongly autoreactive cells from escaping into the periphery where they can cause damage. This process is imperfect however. In fact, antibodies specific to self-antigen (autoantibodies) can be found in healthy people who exhibit no signs or symptoms of autoimmune disease (and thus they are clearly not sufficient for the development of autoimmune disease).
From here, we have many peripheral tolerance mechanisms that protect us. For one thing, we have the requirement of danger. It’s not enough to have a strongly self-reactive cell. Something has to provoke the immune response to initiate the cascade of events that produce autoimmune disease. Peripheral tolerance is, however, extremely powerful and more important than central tolerance for preventing autoimmune disease. Autoimmune disease is a breakdown in self-tolerance mechanisms. It requires failures in both central and peripheral tolerance, but a failure in central tolerance is going to be irrelevant if peripheral tolerance mechanisms pick up the slack.
One important mechanism of peripheral tolerance is anergy. If a self-reactive T cell somehow encounters self-antigen that it is specific to outside of the context of danger, it enters a quiescent state of nonresponsiveness called anergy (like if presented by an immature dendritic cell). Anergy can be reversed, however, for example if that T cell also happens to be very effective at responding to a particular pathogen.
In addition to this, there is a suite of regulatory T cells in the body. In the thymus, regulatory T cells are the cells that respond a bit too strongly to self-antigen. They are allowed to exit to the periphery, and if they encounter the antigen they are primed against, produce immunosuppressive factors that protect the tissue from damage from other autoreactive cells. T regulatory cells can also be induced within the periphery (this is called functional deviation), if inflammation grows too great. In addition to producing immunosuppressive factors, T cells can scavenge important survival factors for other cells, like IL-2, keeping them away from other cells, and they can also release proteins that will kill other T cells. (Most) Regulatory T cells express a protein called FoxP3, and without FoxP3, there are no regulatory T cells. This produces a devastating form of autoimmunity called IPEX (immune polyendocrinopathy enteropathy X-linked) syndrome. In this state, virtually any immune response is so huge in magnitude that profound tissue damage results. It is among the only conditions in which all vaccines are contraindicated because the tissue damage is so great from any immune response.
In addition, if T cells get overexcited, they can undergo activation-induced cell death. Depending on the specific antigen, self-antigen in the body is very abundant, so it’s not too hard to see how overstimulation could occur. This results in death of the T cell.
Some antigens are so important that there is a physical barrier between them and the immune system e.g. the central nervous system and the reproductive organs to prevent response.
In every immune response there is a need to balance the clearance of the pathogen with the tissue damage incurred from the immune system itself. After prolonged activity, T cells can become exhausted and have reduced responsiveness. This can be important as a limiting factor to some autoimmune diseases e.g. type 1 diabetes. Additionally, as tissue gets damaged, cryptic antigens can get released, which you may have cells autoreactive towards, and a series of broken checkpoints later, could produce autoimmune disease. Exhaustion also has some more direct relevance in the case of tumors (an excellent argument for why the immune system cannot be made to totally ignore self-antigen: tumors are primarily self-antigen). Exhausted T cells begin to express a protein called PD-1 (programmed death-1) and on contact with PD-L1, the T cells become anergic. Tumors can be driven to express PD-L1 to suppress the actions of the T cells that would kill them.
Directly after exit from the thymus, T cells are said to be quiescent. They have lower levels of metabolic activity and are unprepared to respond to antigen. This prevents premature expansion of T cells as might happen in an immune response.
Tolerance mechanisms focus on T cells because T cells are the architects of the immune response (and also because they are the best studied ones). With the exception of a few populations (and thymus-independent antigens), B cells require T cell help to carry out immune responses. T cells also reciprocally depend on myeloid cell populations (the non-T, B, and NK cells of the immune system) for cytokines to instruct their behavior, but regulatory T cells can suppress these too.
So, with all these layers of protection, how does autoimmune disease result? One big factor is genetic susceptibility. APS-1 and IPEX syndrome pose very obvious examples but some cases are more subtle. For example, people with celiac disease express the HLA-DQ2 allele at very high prevalence. HLA proteins are part of the machinery of antigen presentation and have binding grooves for short protein fragments that instruct the immune system. HLA-DQ2 has a groove shape that is uniquely suited to a particular sequence on gliadin, which makes people uniquely prone to celiac disease. Additionally, there are so many genes related to peripheral tolerance, that problems in any of them can lead to susceptibility in autoimmunity. For example, CTLA-4 is one of the most important co-inhibitory signals that antigen presenting cells can express, potently suppressing T cell responses, and incomplete function of CTLA-4 is associated with many autoimmune disease states, such as multiple sclerosis. Still, even though these factors predispose to autoimmune disease, they are not sufficient to cause it. There still needs to be an environmental trigger. Often, this is an infection, and in particular, caused by pathogens that have antigens with strong structural similarity to self-antigens (this is called molecular mimicry). The tissue damage resulting from an infection, coupled with the presence of structurally similar antigens, can overcome peripheral tolerance mechanisms and induce autoimmune disease. One of the best characterized cases of molecular mimicry is with M protein in group A streptococci. The M protein is excellent at eliciting immune responses, but has a close resemblance to cardiac myosin proteins. This can produce rheumatic fever, which is why doctors prescribe antibiotics so zealously for strep throat.
Can Vaccines Cause Autoimmune Disease?
With some rare exceptions, no. Older vaccines did have the problem of occasionally causing autoimmune disease, but this is mainly related to the fact that the vaccines in question were cultured in real human tissue e.g. rabies vaccines were made in human CNS tissue that contained myelin basic protein. That’s no longer done. In 1976, an influenza vaccine was produced which seemed to cause Guillain Barre Syndrome (GBS). This was still, however, quite rare (though vaccinated people were 4 times more likely to get the disease, the baseline risk of GBS is about 1 per 100,000, meaning the risk became 4 per 100,000), and there has not been any observed increased risk in GBS from any other influenza vaccine (though influenza itself is 17 times more likely to cause GBS than the vaccine- meaning flu vaccines will protect you from GBS). In general, other influenza vaccines have NOT been associated with an increased risk of GBS. There were also infamously, cases of narcolepsy following receipt of Pandemrix, an AS03-adjuvanted influenza vaccine. However, analyses disagree on whether or not the increase in incidence truly differed from background. It was proposed that in genetically susceptible individuals, an excess of nucleoprotein in the vaccine provoked autoimmunity (rather than the adjuvant). I mention this primarily to point out that it is overwhelmingly the choice of antigen that matters most in a vaccine for safety and efficacy, despite how much people fixate on adjuvant formulations in their concerns about vaccine ingredients (I discussed the concern about aluminum salt adjuvants and autoimmune disease here). A great resource for questions is the Children’s Hospital of Philadelphia’s Vaccine Education Center. It summarizes the matter nicely:
Numerous studies have examined many different vaccines. To date, none have consistently been shown to cause autoimmune diseases. In some studies influenza vaccine was shown to cause GBS at a rate of one case per million vaccine recipients. But, this should be viewed in light of the fact that natural influenza infection causes GBS in 17 per million people infected. So, in a sense, influenza vaccine could be viewed as preventing a more common cause of GBS.
The notion that vaccines don’t cause autoimmunity makes sense. Since vaccines don't drive the immune response nearly as vigorously as natural infections do, it is less likely that they would induce autoimmunity. However, scientists continue to study questions related to vaccines as a cause of autoimmunity as they arise.
Multiple epidemiological studies have evaluated whether or not vaccination is associated with an increased risk of autoimmune diseases (the absolutely minimal requirement for beginning to establish a link- multiple additional studies would be required to demonstrate the veracity of that link). None has been found. Next to the immune responses a person will inevitably face, it is extraordinarily unlikely that vaccines would be the trigger over any number of common infections. Consider for instance how many common viruses are associated with type 1 diabetes mellitus in children (and notably rotavirus vaccines, are associated with a reduced risk in a dose-dependent manner, as rotavirus is one of those pathogens, further supporting the point that the vaccines cannot push the immune system to a pathologic state for someone not contraindicated for them e.g. IPEX syndrome).
Ultimately it comes down to this: it would take a huge subversion of multiple protective mechanisms of tolerance for a vaccine to cause autoimmune disease. Vaccines themselves contain attenuated antigens: they are either incapable of replication (and thus there are substantial limits on the quantity of DAMPs they can release), or have a highly reduced capacity to do so (as with live attenuated vaccines). Autoimmune disease is substantially more likely to result from an infectious disease than from any vaccination. In the case of COVID-19, the vaccine candidates that are currently in phase 3 include:
mRNA vaccines, which encoded a single, non-self antigen that has no pathogenic potential on its own despite the discomfort from the adjuvanticity of the vaccine itself
a replication-deficient vector containing the spike protein only
Compare the immune response elicited from these to SARS-CoV-2, which replicates for days asymptomatically through exceptional suppression of the host interferon response, then initiating maladaptive, destructive immune responses driven by chemokines and cytokine storms. Not to mention that there have been new cases of autoimmune disease recorded in relation to COVID-19.
Seems fairly obvious to me which of those options is a bigger risk for new autoimmune disease.
In short, autoimmune disease resulting from licensed vaccines represents an exceptionally rare phenomenon (to the point that in the few instances it is thought to have occurred, the rates are not definitively distinguishable from background) and should not be used to justify withholding a vaccine from patients who will benefit from its protection.
References
Husebye ES, Anderson MS, Kämpe O. Autoimmune Polyendocrine Syndromes. N Engl J Med. 2018;378(12):1132-1141.
Bluestone JA, Anderson M. Tolerance in the age of immunotherapy. N Engl J Med. 2020;383(12):1156-1166.
Takaba H, Takayanagi H. The mechanisms of T cell selection in the thymus. Trends Immunol. 2017;38(11):805-816.
Plotkin SA, Offit PA, DeStefano F, et al. The science of vaccine safety: Summary of meeting at Wellcome Trust. Vaccine. 2020;38(8):1869-1880.
Wraith DC, Goldman M, Lambert P-H. Vaccination and autoimmune disease: what is the evidence? Lancet. 2003;362(9396):1659-1666.
Anderson MS, Su MA. AIRE expands: new roles in immune tolerance and beyond. Nat Rev Immunol. 2016;16(4):247-258.
Rosenblum MD, Remedios KA, Abbas AK. Mechanisms of human autoimmunity. J Clin Invest. 2015;125(6):2228-2233.
Theofilopoulos AN, Kono DH, Baccala R. The multiple pathways to autoimmunity. Nat Immunol. 2017;18(7):716-724.
Park JH, Lee KH, Jeon B, et al. Immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome: A systematic review. Autoimmun Rev. 2020;19(6):102526.
Yamano T, Nedjic J, Hinterberger M, et al. Thymic B cells are licensed to present self antigens for central T cell tolerance induction. Immunity. 2015;42(6):1048-1061.
Nemazee D. Mechanisms of central tolerance for B cells. Nat Rev Immunol. 2017;17(5):281-294.
Rose NR, Mackay IR. The Autoimmune Diseases, 6th Edition.; 2020.
Gonzalez S, González-Rodríguez AP, Suárez-Álvarez B, López-Soto A, Huergo-Zapico L, Lopez-Larrea C. Conceptual aspects of self and nonself discrimination. Self Nonself. 2011;2(1):19-25.
Perkey E, Maillard I. New insights into graft-versus-host disease and graft rejection. Annu Rev Pathol. 2018;13:219-245.
Iberg CA, Jones A, Hawiger D. Dendritic cells as inducers of peripheral tolerance. Trends Immunol. 2017;38(11):793-804.
Dhodapkar MV, Steinman RM, Krasovsky J, Munz C, Bhardwaj N. Antigen-specific inhibition of effector T cell function in humans after injection of immature dendritic cells. J Exp Med. 2001;193(2):233-238.
Eldershaw SA, Sansom DM, Narendran P. Expression and function of the autoimmune regulator (Aire) gene in non-thymic tissue: Peripheral Aire expression and function. Clin Exp Immunol. 2011;163(3):296-308.
Hammer GE, Ma A. Molecular control of steady-state dendritic cell maturation and immune homeostasis. Annu Rev Immunol. 2013;31(1):743-791.
Zindel J, Kubes P. DAMPs, PAMPs, and LAMPs in immunity and sterile inflammation. Annu Rev Pathol. 2020;15(1):493-518.
Schietinger A, Greenberg PD. Tolerance and exhaustion: defining mechanisms of T cell dysfunction. Trends Immunol. 2014;35(2):51-60.
Murphy K, Weaver C. Janeway’s Immunobiology. 9th ed. CRC Press; 2016.
Petersen J, Montserrat V, Mujico JR, et al. T-cell receptor recognition of HLA-DQ2-gliadin complexes associated with celiac disease. Nat Struct Mol Biol. 2014;21(5):480-488.
Cecilio LA, Bonatto MW. The prevalence of HLA DQ2 and DQ8 in patients with celiac disease, in family and in general population. Arq Bras Cir Dig. 2015;28(3):183-185.
Tomer Y. Genetic susceptibility to autoimmune thyroid disease: past, present, and future. Thyroid. 2010;20(7):715-725.
Karabon L, Kosmaczewska A, Bilinska M, et al. The CTLA-4 gene polymorphisms are associated with CTLA-4 protein expression levels in multiple sclerosis patients and with susceptibility to disease. Immunology. 2009;128(1 Suppl):e787-96.
Cunningham MW. Rheumatic fever, autoimmunity, and molecular mimicry: the streptococcal connection. Int Rev Immunol. 2014;33(4):314-329.
Sencer DJ, Millar JD. Reflections on the 1976 swine flu vaccination program. Emerg Infect Dis. 2006;12(1):29-33.
Galeotti C, Bayry J. Autoimmune and inflammatory diseases following COVID-19. Nat Rev Rheumatol. 2020;16(8):413-414.
Kwong JC, Vasa PP, Campitelli MA, et al. Risk of Guillain-Barré syndrome after seasonal influenza vaccination and influenza health-care encounters: a self-controlled study. Lancet Infect Dis. 2013;13(9):769-776.
Ahmed SS, Volkmuth W, Duca J, et al. Antibodies to influenza nucleoprotein cross-react with human hypocretin receptor 2. Sci Transl Med. 2015;7(294):294ra105.
Verstraeten T, Cohet C, Dos Santos G, et al. PandemrixTM and narcolepsy: A critical appraisal of the observational studies. Hum Vaccin Immunother. 2016;12(1):187-193.
Rogers MAM, Basu T, Kim C. Lower incidence rate of type 1 diabetes after receipt of the Rotavirus vaccine in the United States, 2001-2017. Sci Rep. 2019;9(1):7727.
ACIP contraindications guidelines for Immunization. Cdc.gov. Published November 18, 2020. Accessed November 29, 2020. https://www.cdc.gov/vaccines/hcp/acip-recs/general-recs/contraindications.html
The Children’s Hospital of Philadelphia. Vaccines and Autoimmune Diseases. Chop.edu. Accessed November 29, 2020. https://www.chop.edu/centers-programs/vaccine-education-center/vaccines-and-other-conditions/vaccines-and-autoimmune-diseases
Filippi CM, von Herrath MG. Viral trigger for type 1 diabetes: pros and cons. Diabetes. 2008;57(11):2863-2871.
Hardie EM, Kruse-Elliott K. Endotoxic shock. Part II: A review of treatment. J Vet Intern Med. 1990;4(6):306-314.
ElTanbouly MA, Noelle RJ. Rethinking peripheral T cell tolerance: checkpoints across a T cell’s journey. Nat Rev Immunol. Published online 2020. doi:10.1038/s41577-020-00454-2
