Some concepts for an HIV vaccine candidate. Thus far, HIV has presented an insurmountable challenge in vaccinology, but progress is being made constantly.

I’ve observed a lot of confusion regarding the properties of different vaccine classes, especially ascribing effects that are impossible to be due to a vaccine because of the class it belongs to, so in what follows, I present an overview of the different classes of vaccines.

Subunit and Toxoid Vaccines

These vaccines strive primarily to create immunity via antibodies against a particular toxin or protein component produced by the disease-causing agent. Tetanus is a particularly instructive example for describing how these diseases work. Tetanus is a toxin-mediated disease, meaning that the ill effects of the disease are the result of the interactions of the toxin with the body. It’s caused primarily by a toxin produced by the bacterium Clostridium tetani called tetanospasmin. The toxin affects the nerves by destroying the mechanism by which our nerves regulate muscle contractions. Nerves are normally actively working to suppress muscle contraction, and tetanospasmin, once it gets access to the nerves, for instance through a puncture wound, prevents those inhibitory signals from being sent, and muscles have no choice but to begin contracting. This typically occurs with risus sardonicus (an involuntary contraction of the facial muscles into a smile), trismus (lockjaw) and then progresses to opisthotonos (the arching of the back due to muscle contractions) and the toxin eventually makes its way into the central nervous system. The muscle spasms that result are not only extremely painful (in fact there are documented cases of the spasms breaking bones) but if the toxin affects the diaphragm (the large muscle in the chest cavity whose contraction and relaxation controls breathing) death is all but guaranteed. Tetanospasmin is so incredibly potent (second only to botulinum toxin as far as lethal doses go) that a dose less than what your immune system is capable of recognizing can be enough to kill you. We say that this is a sub-immunogenic dose (sub- meaning less than and immunogenic meaning capable of producing an immune response). This means that if you are unfortunate enough to experience tetanus and somehow survive it, you will not gain immunity- because as far as your immune system is concerned, nothing happened. All this havoc happened below the threshold needed for it to do anything. Additionally, the toxin is present in the bacterial spores which are extremely hardy- even though tetanus bacteria are anaerobic (they do not tolerate oxygen well), the spores have incredible tolerance to heat, antiseptics, and oxygen. This is where a vaccine becomes so important. Tetanus toxoid containing vaccines are in fact the only way to gain immunity to tetanus. A toxoid is a toxin (a chemical substance derived from a living thing that exerts poisonous effects, as distinct from a toxicant which is not derived from a living thing but rather is man-made) that has been made to lose its toxic properties while retaining its basic structure. In the case of tetanus, this is accomplished by purifying the toxin and then bathing it in a solution of formaldehyde and lysine (lysine is an amino acid that is found all over your body that helps make up proteins). Before you become alarmed, it’s important to understand why this is done in the way that it is. Formaldehyde is a special molecule because it can do a very neat trick where it can make other molecules freeze as they are in space. It’s sort of like how a bug gets trapped in amber: we can still see the bug and we know it’s there but the bug can’t do anything anymore because it’s trapped in the amber. Tetanospasmin is an enzyme- a substance that catalyzes reactions in biological systems- and it cannot act unless it can move around- formaldehyde stops that process. Of note, your body is constantly producing formaldehyde because it’s a byproduct of a number of processes that are essential for life. The idea is similar for other toxoid vaccines though we can attenuate toxins through other ways like by heating them up (but this doesn’t always work very well because it can change the shape of the toxin too much). Also- the formaldehyde in the vaccine is purified out. In any given vial of DTaP there is less than 0.1 mg (for reference, a grain of rice has a mass of about 25 milligrams) and it is cleared within minutes- not at all a safety concern.

Subunit vaccines are the same concept: we take a single structural component of the pathogen or a few and we use that to immunize. Examples include the pertussis antigens in DTaP and hepatitis B virus surface antigen in the hepatitis B vaccine.

Toxoid vaccines are a really great idea when you have a toxin-mediated disease but unfortunately not all diseases are like that; Lyme disease for instance is notable because it causes all of those problems without any toxin. But, because the diseases in question are toxin mediated they have a lot of potential as vaccines against diseases caused by members of our own microbiome. For example, consider E. coli. It might surprise you to know that everyone has this living in their gut. And most of the time it’s not a problem at all. But unfortunately bacteria like to do this really kinky thing called lateral gene transfer and so bacteria can share DNA that contains instructions on how to make certain toxins. This can make your friendly neighborhood K12 E. coli become an entertoxigenic form that causes diarrhea or a shiga-toxin containing one that causes hemolytic uremic syndrome. This applies to a number of bacteria. For instance, diphtheria’s reservoir is actually humans, in particular the very young ones, but diphtheria causes a toxin-mediated disease and this toxin is carried by a phage which infects the bacteria- so you can have non-pathogenic forms of diphtheria without any issue. They only become a problem if they get the toxin by transduction. By engineering a toxoid vaccine against the toxin, you circumvent the issue. Plus it cuts down on antibiotic use.

As clever as this design is, there is a problem. As it turns out, these subunits and toxoids tend to have pretty poor immunogenicity on their own- they’re bad at activating the immune system to produce reliable memory. There’s just a single antigen within them (or a couple in the case of combined vaccines) whereas a real infection could have your immune system dealing with thousands at the same time (indeed, scraping your knee probably represents a greater challenge for your immune system than the entire vaccine schedule combined). This is largely mitigated through the use of adjuvants- additives to the vaccine that stimulate the immune system. Before we added aluminum adjuvants to DTaP it had very poor immunogenicity, but now it confers immunity to the tetanus component at a rate of virtually 100%, diphtheria at a rate of about 97%, and pertussis at a rate of about 80–85%. Another useful example is Shingrix. Shingrix is a vaccine against shingles which contains the gE protein of varicella and some adjuvants, and it is MUCH more effective than Zostavax- the old shingles vaccine which had live varicella in it (it was basically just a giant dose of the varicella vaccine). With the aid of adjuvants we can faithfully produce strong immune responses, but subunit vaccines tend to produce immunity that lasts for a shorter amount of time than we might hope, so that’s why we need Td boosters every 10 years. This isn’t because the vaccine isn’t well made- subunit vaccines are some of the safest vaccines in existence- but it’s a limitation of the way our immune system works for reasons we don’t fully understand.

Virus-like Particle Vaccines

Broadly, if you’re a virus there are a few tools you must have: genetic information, a means to replicate your genetic information, structural proteins that allow you to invade your target cells, and proteins that initiate the budding off of the cell to go on to infect other cells. In principle, that last function can require as few as 1 protein, and the product of that process is a virus-like particle. This is basically a ball of a viral protein and membrane. The virus-like particle does not have the ability to infect cells- it has no genetic information, no virulence factors- just that protein. So you might wonder how this is different from a subunit vaccine. Well, in a sense, it’s still a subunit vaccine, but it comes specially packaged. The thing about proteins is they fold, and how they fold depends on their environment. The cell membrane is fatty but life occurs in water. So the way a protein folds if it is located inside a membrane is completely different from how it folds outside of it, and this makes a huge difference for your immune system. Antibodies bind based on shape, not sequence. The key example of this is the HPV vaccine, which has the L1 protein to create VLPs from yeast membranes (in the case of Gardasil-9). Like other subunit vaccines, these may require adjuvant help. In the case of Gardasil this is accomplished with amorphous aluminum hydroxysulfate, an aluminum salt.

Polysaccharide and Conjugate Vaccines

Primarily, your immune system responds to protein antigens. Proteins are really useful as antigens because they’re so diverse. There are 20 common proteinogenic amino acids which makes for a really wide alphabet and allows your immune system to zero in on, making them great for distinguishing between self and non-self. But your immune system is capable of responding to polysaccharide too (and lipid and nucleic acids)! Polysaccharides are just chains of sugar molecules. You encounter all sorts of polysaccharides every day. For instance, starch, like in your bread is a polysaccharide made of repeating units of glucose. Cellulose is a polysaccharide that makes up the major component of plant cell walls that’s also made of chains of glucose (but they’re linked together differently and because of that we can’t break it down). Polysaccharides are pretty structurally diverse too and bacteria have unique combinations of sugars in their cell walls that allow the immune system to have a non-protein based way of identifying non-self. The thing is though, the antigen presenting cells of the immune system are not able to present polysaccharide antigens like protein antigens. In responding to polysaccharides we rely very heavily on our B cells- the cells that make antibodies. The B cells are made to produce antibodies specific to the polysaccharide in the vaccine. B cells can respond to other kinds of antigens too though- for instance nucleic acids, and antibodies are more flexible (literally) in what they recognize than T cell receptors. Plus polysaccharide antigens tend to have many repeating units which make them excellent for cross-linking antibodies which allows for the engagement of many cells at the same time producing a robust immune response. There’s a catch though: there is a population of B cells that does this very well called marginal zone B cells which are so named because they reside in the spleen in a region called the marginal zone. Ordinarily B cells can’t make antibodies without help from T cells, but not these guys! They can do it all on their own and are even capable of memory responses (i.e. more robust responses upon re-encountering the antigen). The problem is the marginal zone of the spleen doesn’t develop until about 1–2 years of age, so very young kids don’t respond well to polysaccharide vaccines. We figured out a workaround though: all we have to do is attach the polysaccharides to a protein anchor (i.e conjugate them) which are called conjugate vaccines. The addition of a protein antigen allows us to engage our T cells so that the B cells can get help making the antibodies. Examples of these vaccines include those against meningococcal, Hib, and pneumococcal.

Killed vaccines

These are also called inactivated vaccines. These vaccines are exactly what they sound like. We take the pathogen that causes disease and we… kill it (there are so many different ways to do this that I won’t bother listing them). The antigen retains all the structural features it needs to cause an immune response. Of note, this can be done with either bacteria or viruses. In the past, we used to use DTwP which contained whole cell pertussis, which was the entire Bordetella pertussis bacterium killed inside the vaccine, representing a few thousand antigens (by contrast acellular vaccines have about 5 with some variation depending on the manufacturer). This was actually very effective most of the time, but it had the important downside of being very reactogenic (causing a lot of unpleasant side effects). The nice thing about these is that they don’t require adjuvants. They have all the molecular patterns needed to induce innate immune responses that lead to protective memory (the exception is the adjuvanted flu vaccine which is an inactivated vaccine but because of the changes the immune system undergoes as we age, without adjuvant is not good enough at creating protective responses in the elderly). Plus as everything inside them is dead they can’t cause the disease associated with the pathogen, and they can even be given to those who are immunocompromised. They are not without their drawbacks though. The fact that these vaccines can’t replicate inside you means that any response you make through CD8+ T cells (the killer or cytotoxic T cells that are induced in viral infection to do the purge- possibly the most important cell type for the later phase of a viral infection) is minimal if it exists at all. Plus inactivation can damage the antigens which can reduce their immunogenicity. Examples of these vaccines include influenza vaccines, inactivated polio (the shot), DTwP (used in developing nations), hepatitis A, and rabies.

Live Attenuated Vaccines

From Janeway’s Immunobiology 9th Edition Figure 16.24

Of all the vaccine types these are the ones that seem to cause people the most concern… for some reason. These vaccines do contain a live infectious agent, but it has been adapted to grow really poorly inside your body. Here’s the thing: mutations are acquired randomly, and a small proportion of them will confer an advantage to the survival of the organism in question. But this is very context specific. Imagine for a moment that I took you as you are now and sent you to live in 4th Century Egypt. How well do you think you would do? On the other hand imagine I asked someone from 4th century Egypt to make me a spreadsheet on Excel. The thing is both you and this 4th century person have skills that help you thrive in the environment you are in- but when I abruptly change your environment, suddenly you’re lost. That’s the concept of attenuation. We take the pathogen in question and we grow it under conditions that differ dramatically from physiological ones. This typically means in a cell type it doesn’t usually infect and at a temperature the pathogen isn’t designed to grow in. The pathogen undergoes many generations of replication and then adapts to those conditions. And then with the vaccine I suddenly force it to try to do that under physiological conditions which are completely different. Suddenly that pathogen is the 4th century Egyptian trying to use a computer. Now, there is going to be some replication, and this is important because it allows us to produce CD8+ T cells. But in an immunocompetent (not immunocompromised) individual this pathogen won’t be able to grow to any appreciable extent. The problem is that in an immunocompromised individual, it might. So depending on the extent of the immunocompromise you can’t give these vaccines to immunocompromised people (which is not an issue with all other vaccine types I mentioned so far). Also, like killed vaccines, these do not require adjuvants. But there’s an important downside. In principle, a vaccine should mimic the infection as closely as possible to produce the robust responses of that infection, but this isn’t often feasible with live vaccines. MMR for instance is given by subcutaneous injection (or intramuscular but really subcutaneous is preferred), but measles invades the body through the respiratory tract and possibly the conjunctiva (that point is controversial). On the other hand, this also likely contributes to the fact that MMR doesn’t “shed.” While it can cause a rash, the lesions do not contain significant amounts of measles virus and thus it cannot be passed from person to person- which might not be true if we designed a vaccine that colonizes the respiratory tract as that’s how measles spreads. For eradication campaigns though like what’s happening with polio, it’s actually advantageous to have a vaccine that can spread from one person to the next as it means better coverage. The downside is that enough generations in people means the vaccine antigen can undergo a reversion of virulence (though this has only been documented with the oral polio vaccine, but it is a theoretical concern with any live viral vaccine). Overall though, shedding basically doesn’t happen. The only exceptions really are polio and rotavirus, where the vaccine strain virus can spread through the stool, a few isolated cases of chickenpox from the vaccine, and some cases with Flumist. Examples of live vaccines include Flumist (the nasal spray flu vaccine), MMR, BCG, rotavirus, oral polio, yellow fever, and varicella vaccines.

Viral Vector-based Vaccines

These vaccines sort of combine the concept of a live vaccine with a subunit vaccine. We take a virus, a live one, and we induce it to express vaccine antigens. The immune system will then go on to respond to the vaccine antigens. This is what was done with the Ebola vaccine. The Ebola vaccine uses a VSV (vesicular stomatitis virus) as the vector and it expresses ebola envelope glycoprotein. The downside to these is ensuring that you get the response to the vaccine antigen rather than your viral vectors so this requires a bit of trickery from a manufacturing standpoint (i.e. minimizing your vector’s expression of immunogenic particles that are not the vaccine you want). The major advantage of these is subtle: your vaccine can’t undergo a reversion to virulence that some live vaccines can. If oral polio passes through many generations of unvaccinated individuals, it can go back to its virulent form and cause paralysis (oral polio is used in areas where the virus is endemic). A vector vaccine can’t do that because we exclude the machinery that allows the vaccine pathogen to do these things from its design. And of course, as these are live vaccines they’re not suitable for the immunocompromised. The key example is the Ebola vaccine.

DNA Vaccines

So I’ll start by saying there currently aren’t any DNA vaccines on the schedule. This is largely because they’re not all that effective. The concept is that you have DNA that encodes the immunogenic proteins from the vaccine you want and you inject that into a patient. Basically DNA is really hard to get into a cell so just having naked DNA floating isn’t going to make it get into cells and make your cells imitate that they’re infected and cause an immune response. Typically to make this work you need to use something like an electric shock which causes temporary pores in the cell (electroporation) allowing the DNA to enter. An alternative approach is to coat the DNA onto metal particles and use a gene gun. The concept behind these vaccines is that DNA is a stable molecule so these vaccines should be easy enough to store and therefore be useful for mass immunizations. Unfortunately, the passive uptake of DNA by our cells doesn’t really happen- it very much requires deliberate actions on our part to get it there (think about what it might mean for gene therapy: how easy would it be to fix a genetic defect if we could just make whatever DNA we wanted, inject it into a person and ta-da they’re cured! It doesn’t work like that). However, at this point in time, their efficacy is disappointing so we have no such vaccines widely available yet.

Dendritic Cell Vaccines

These are amazing. So you have a cell in your body called a dendritic cell, which is what’s known as a professional antigen presenting cell. When you have an infection, it’s supposed to eat up the pathogen and infected cells and such and then present the fragments of antigen to T cells to initiate an immune response. The concept of these vaccines is being applied to cancer primarily but also to HIV. In the case of HIV, a portion of the patient’s own dendritic cells are taken and primed with dead HIV. They then go on to initiate an immune response against the HIV. The results are pretty impressive: viral load in this trial dropped 80% and the patients could maintain a state free of viremia (virus in the blood) for almost a year. With cancer vaccines, it’s the same concept. A sample of the patient’s dendritic cells are taken and the tumor is biopsied. The dendritic cells are then activated by tumor specific antigens, and initiate an immune response against the tumor. Overall, these are impressively effective, but they are very expensive and not feasible for most infectious diseases because they take too long to make for mass immunization.

Passive Immunization

Passive immunization is not a vaccine, but rather refers to the transfer of antibodies directly into an individual specific to a given pathogen. During pregnancy, this occurs via transfer of IgG through the placenta into the fetus via the umbilical cord. The major issue with this mode of immunization is in the fact that it’s so transient. Depending on the amount of the immunoglobulin, this protection lasts for at most a few months. With vaccination, you produce specific memory cells which can persist for years under the right conditions and be recalled to produce antibody later on in life with ease. Because no memory cells are made via passive immunization, you also don’t get any memory responses. That said, it does have one major advantage over vaccines: it acts immediately. So for example, it takes some time to mount an immune response to tetanus vaccines (depending on your immunization history) which means that potentially you might be vulnerable if you get a high risk injury so in some cases, you might be given anti-tetanus immunoglobulin with the vaccine. This ensures that you’re protected from the tetanus while your body works to create a memory response. Breastfeeding is another example, but it’s a bit more limited in its usefulness- breastmilk is a source of IgA antibodies which are not the same as IgG antibodies associated with mature immune responses. IgA is very useful for maintaining the barrier integrity at mucosal surfaces which can make it very helpful for preventing various GI bugs, but its usefulness for respiratory infections is more limited. The other major disadvantage is that with the exception of naturally occurring modes of passive immunization, it is generally very expensive. It takes an incredible amount of work to purify a specific antibody from someone’s plasma and so the price is set accordingly. I will also add that in some cases, people have certain immunodeficiencies where they cannot effectively make their own antibodies and thus receive an infusion of polyclonal antibodies every few weeks or so to protect them, which is also a form of passive immunization.

mRNA Vaccines

I made an entire separate post about them here.

I write about vaccines here. You can find me on Twitter @enirenberg and at (where I publish the same content without a paywall)