Cancer and the Immune System

Deep within us, an arms race rages, an evolutionary arms race which began eons ago, back when we were simple celled organisms. I speak of the immune system’s constant battle to deal with rapidly evolving pathogens, with its own unique strategies for developing highly specific weapons that selectively target these dangerous organisms. What seems curious, upon first inspection, is that if such a system exists to detect dangerous cells in the body, and can eliminate them, why is cancer even a problem?

To understand why this is, we need to consider some basic immunology before applying that knowledge to the challenge of cancer.

A (not so brief) overview of the immune system

Ed Yong, in one of his many excellent pieces on the Atlantic, illustrated the frustration associated with immunology very succinctly with a joke.

“An immunologist and cardiologist are kidnapped. The kidnappers threaten to shoot one of them, but promise to spare whoever has made the greater contribution to humanity. The cardiologist says, “Well, I’ve identified drugs that have saved the lives of millions of people.” Impressed, the kidnappers turn to the immunologist. “What have you done?” they ask. The immunologist says, “The thing is, the immune system is very complicated…” And then, the cardiologist says, “Just shoot me now.””

As with most jokes, this one has a basis in fact; the immune system is pretty complex. Let’s try to break it down, using the example of some nasty bacteria getting into the subcutaneous tissue beneath the skin of my arm.

The first thing to consider is that there are 2 branches of the immune system, the adaptive system, and the innate system. My innate system is the “quick and dirty” fix, detecting the bacteria using special receptors called Toll-like receptors (TLRs) present on macrophages, neutrophils, and Natural Killer (NK) cells. This recognition is very important, because it activates the “effector” components of these system, the parts that directly deal with the virus.

A particularly fun way of thinking about the interplay involved in this is to consider a classic scene from any mafia movie.

The garbagemen working for one mob clear out the trash from another mob’s nightclub, and see that the rival boss is just about to make his exit. They quickly signal the hitmen, who come in, and eliminate the threat to their organisation. In a very similar way, macrophages first detect the bacteria in my skin, and signal to the NK cells that there is a threat, and to be ready to recognise it and kill it, as well as to increase their number (akin to calling for reinforcements!). I’ve tried to illustrate this graphically below, marking the chemical signals involved for anyone who’s interested.

Normally the innate immune system would deal with this bacteria really quickly (within 6 hours or so, quite often), but if this bacteria is a tricky customer, and one that really gets under my skin (excuse the dreadful pun), the adaptive system steps in.

There are certain cells which sample the bacteria, and take up some of its proteins, degrading parts of them. These dendritic cells then present the peptide fragments on their surfaces, while simultaneously being activated directly by the pathogen, through TLR interactions. This allows it to express molecules which it needs to interact with other components of the adaptive system, of note, the B7 molecule.

These activated, antigen presenting, dendritic cells now move down a chemical gradient to reach a lymph node close to my arm. They first present this antigen to a naïve T lymphocyte, and if recognised, the T lymphocyte could become active! However, an important factor to consider here is that we need that B7 molecule on the dendritic cell to activate the T cell fully, and thankfully, this molecule becomes expressed after the dendritic cell is activated by the pathogen via TLR interactions.

This activated T cell now goes on to activate a B lymphocyte, which will form an antibody producing plasma cell. The antibody that it produces will be specific to the pathogen, because the activated T cell could only activate a B lymphocyte that recognises the same antigen as it does. Pretty clever!

Here’s a snapshot of these 3 cell types interacting with one another.

Antibodies are not the only factors at play with the adaptive response, important though they are. We can also produce a type of T cell, the appropriately named killer T cell. This cell recognises foreign cells and kills them directly, by activating the program for cell suicide. The antibodies and killer T cells help the tissue in my arm rein in this bacterial infection, and eliminates it from my body.

Huzzah!

Now that we have a functional idea of how the immune system works, let’s think about cancer.

Cancer

Siddhartha Mukherjee described cancer as a “distorted version of ourselves”; this is quite a telling metaphor, after all, cancer cells are just human cells with a few mutations in the right places. Since these grossly perverted versions of human cells are so unnatural, surely the immune system can deal with them!

One of my teachers in Cambridge proposed an excellent framework for thinking about how the immune system is linked to cancer. He said that the key is to ask 3 important questions.

1. Does the immune system recognise cancer?
2. If it does recognise cancer, does the cancer escape it?
3. Can we use the immune system as a weapon against cancer?

Let’s start with the first question.

Does the immune system recognise cancer?

With the sea of mutations that a cancer cell accumulates, it’s quite likely at some point that some of parts of these mutated proteins might end up getting displayed on the cancer cell’s surface. We call this “neoantigen,” a new antigen formed by the cancer mutations. This is quite common in cancers where we see single DNA base changes, like in melanomas (because of how UV light interacts with particular bases in DNA), or lung cancers.

In addition, because the cancer signalling pathways might activate some stress mechanisms, stress signals might also become expressed on the cancer cells, allowing for recognition by NK cells. Cancer cells also sometimes reduce the expression of carrier glycoproteins which give immune cells a view into their transcriptional activity – these molecules are called class I MHCs. This reduces the risk of being picked up as a cancer cell by T cells, but these molecules also typically prevent NK cells from attacking cells, so by removing them, there is a higher probability of NK recognition, and NK mediated death.

I want to come back to the point about neoantigens, as despite the fact that we have this very clear danger warning for T and B cells to pick up on, it’s unlikely that they will. This is because the T and B cells are typically confined to the lymph, plasma and lymph nodes, unless a clear signal for their movement becomes activated. Therefore, it is unlikely that these cells would simply wander into a tumour environment. Ultimately this is symbolic of a much larger conflict; the need to preserve tolerance against self-cells, and the need to provide surveillance against cancer.

If we opened up tissues to B and T cells, there is a chance that some antigens that B and T cells were not trained to treat as a self-marker will be recognised as foreign, increasing the risk of a catastrophic autoimmune reaction. Although being able to surveil against cancer cells would be great, the risk of autoimmunity is probably too great.

There is some evolutionary logic to this as well; cancer was never a big problem for us until recently, mainly because humans rarely lived long enough to develop this horrific disease. Autoimmune disease was probably more of a concern, so in the evolutionary development of the immune system, there was likely greater pressure to minimise the production of autoreactive B and T cells rather than to increase surveillance against cancer cells.

So to circle back to the first question, it looks like the answer is, “it depends,” but for the most part, it seems unlikely that we would have immune recognition of cancer cells due to the complex distribution of lymphocytes in the body.

If the immune system does recognise cancer, can the cancer escape it?

Now, let’s say we’ve got our ideal situation. A T cell has identified a cancer cell and has received all the signals it needs to become active (an assumption we will get back to in a minute). What are its odds of being able to attack a cancer cell?

The issue is that cancer cells have a bag of tricks to outfox immune cells. For one, they can upregulate an enzyme called IDO in lymphocytes, which, when overly active, results in excessive tryptophan metabolism. When these lymphocytes become starved of tryptophan, they stop proliferating and enter a state of non-responsiveness.

Cancer cells can also create an immunosuppressive environment around them, by secreting anti-inflammatory chemicals like TGF-B and IL-10. The TGF-B acts on T helper lymphocytes, triggering their differentiation into a helper subtype called the iTreg, which is anti-inflammatory. This potentially reduces the risk of activation of immune cells which recognise a cancer cell.

The way in which this TGF-B production can occur is interesting, and merits some further discussion. The signalling pathway that eventually triggers TGF-B production involves a transcription factor called STAT5, which is quite pleiotropic, which is to say, it influences a number of signalling pathways and proteins. This transcription factor is implicated in the expression of cyclin D, the first cyclin in the cell cycle, which possibly explains why constitutively active versions of STAT5 are seen in certain cancers.

On the other hand, it is known that there are immune infiltrates in cancers, comprising mainly of killer T cells, which likely do some good in terms of slowing down tumour progression. In fact, patients with higher killer T cell levels in their tumours tend to have better event free survival rates.

So it sounds like having killer T cells in the mix is key to having an immune cell both detect and attack a cancer. However, it is not quite as simple as this. As I explained in my earlier diagram, T cells require a second signal, the B7- CD28 interaction. The major issue to activating killer T cells in a tumour is that the tumour cells do not generally display the B7 which is needed to effectively stimulate the killer T cell response. Therefore, that ideal scenario I posed at the start of this section is unlikely to come to fruition all the time.

However, one of the times when we might have such a situation developing is in blood cancers. This is for 2 key reasons. The first is that blood cancers are in the blood (forgive the axiom), so this means they exist in the same place as KT cells might. The second is that blood cancers can sometimes express high enough levels of B7 to stimulate a KT cell.

As great as this sounds, this indicates the KT cells do provide some surveillance, albeit what must be incomplete surveillance. If it were complete, leukaemias and lymphomas probably would not exist.

Since the last question is probably the most involved of the trifecta, I’m going to save it for a future post, so I can discuss it in the detail it deserves.