Treating cancer is awful
We can do better.
In May of 1876, an 8-year old boy was admitted to the Hospital for Sick Children in Bristol, England. The child was seen by a Dr. John Ewens, and had presented with a rapidly expanding tumor in the tibia near the knee joint. After examination, it was decided that amputation of the limb above the knee would be the most likely road to survival, and the procedure was executed with all manner of safety and caution which a late-19th century surgical suite could provide. Dr. Ewens deemed the surgery a success (though the records of the size of the recovered tumor were reported lost), and the patient was discharged within a month.
By October 14th of the same year, the boy had returned in worse condition than previously found. He had developed a small mass near the end of the amputated limb, and drew short, difficult breaths. The surgical team considered a second operation, but realized the gravity of the situation and the complexity of the child’s disease, and sent the patient to return home with no treatment available at the time for such a condition — recurring osteosarcoma with likely metastasis to the lung. Dr. Ewens visited the patient several weeks later to find the metastatic lump larger, and the boy’s condition worsening — within a few days it was reported that the patient had succumbed to the cancer.
The report of the above case was published in the British Medical Journal in February of 1878. 100 years later, in 1978, osteosarcomas were still primarily treated with surgery. The value of neoadjuvant chemotherapy (or rounds of chemo prior to tumor excision) had not yet been widely accepted, and oncology teams were only beginning to standardize their cocktail of chemotherapies and their respective dosing schedules. Surgical techniques had improved — they were cleaner, more precise, less painful (thanks to improved analgesics), and involved more advanced diagnostic tools, but complete responses were dismally rare. Today, oncologists still face the same problems — radical surgeries are often required to eliminate a primary tumor, and local recurrence or distal metastasis remain fearfully common.
The problem of curing cancer is simply much more complicated than initially expected, because cancer is not quite as straightforward as many of the other diseases with which we have battled for so long. Bacterial, parasitic, and viral infections — those are relatively easy challenges (with exceptions) due mostly to the fact that they are caused by something which is “non-self” — a foreign invader. Hacking the immune system to detect viruses was Dr. Jenner’s gift to humanity and it was no small present. Likewise, molecules that could disrupt essential bacterial processes with minimal effects on human cellular machinery were not hard to find. In fact, we found some of the first ones by accident. Penicillin and friends (a.k.a. the beta-lactams), isolated by Dr. Fleming, led us to the concept of wonder drugs — small, stable compounds able to be synthesized and distributed at scale, and containing the power to cure human suffering. The impact of these technological breakthroughs was immediately evident — between 1945 and 1955, mortality from tuberculosis dropped from 39.9 per 100,000 to 9.1 per 100,000. That decline was almost entirely due to the discovery of streptomycin by Selman Abraham Waksin. In fact, the early 20th century brought on a remarkable decline in overall mortality due to massive public endeavors to control infectious diseases in general, not just tuberculosis.
At first glance, cancers share some surprising similarities to infectious diseases. They start locally and can spread rapidly through the body if unchecked, they involve the growth of cells or material that can be distinguished from normal tissue based upon structural or biochemical variations, and they are diseases which have plagued humanity for millennia. This differentiates both cancers and infectious diseases from other illnesses which are more subtle, more local, or more recently identified. A number of pathologic conditions have been systematically characterized by biomedical science in modern times, and treatments have been iterated and trialed until management is often nearly identical to a cure. In these cases, identifiable mechanisms have often been rigorously proven to drive the disease, and (while more effective true cures almost certainly exist) the machine of biomedical research has synthesized pharmaceutical or technical solutions with satisfying efficiency.
Certainly drug-cures for cancer must have some merits that have kept physicians and researchers in hot pursuit with undaunted confidence — so why would we pursue them? What are their beneficial qualities, and are those qualities beneficial to patients? In other words, what is motivating this decades long quest where even slight improvements in outcomes have been welcomed triumphantly?
Cancers are inherently different from other diseases — just how different is a question that researchers are only beginning to answer. Malignant tumors hack cellular growth mechanisms to monopolize energy resources and maintain continuous cellular life cycles. Additionally, they fend off or hide from the body’s immune targeting mechanisms that would otherwise destroy unwanted material or invasive cells. While researchers can classify tumor types, and identify similarities in their malfunctioning processes or cellular dependencies, even similarly classified cancers demonstrate a huge amount of variability in growth rate, metastatic tendency, and response to chemotherapy. That frustrating variability complicates the problem of cancer such that one-size-fits-all treatment plans are effectively useless — forcing clinicians and patients into a dangerous game of chicken with deadly drug-toxins in the hopes of the cancer dying first.
Still, there has been some success. Modern medical care for most cancers has dramatically improved since the days of the radical mastectomy and other grotesque violations of human dignity done in the name of medicine. Currently, many myelomas, lymphomas, and leukemias have excellent response rates with the available chemotherapeutic cocktails. But, these are not cures — they are progression free survival curves, probability of event free survival, or 5-year mortality rates. Often, treatment simply prolongs the suffering and postpones the inevitable. The tools which scientists so fondly adore — small molecule drugs and immunologic targeting of unwanted material — have so far proved barely adequate to treat cancer and massively insufficient to cure cancer. It’s interesting to consider what might have occurred had physicians and researchers endlessly explored vaccines (1 — see “Notes” below) as treatments for parasitic or microbial infections such as plasmodium or syphilis instead of happily uncovering antibacterials in such serendipitous fashion. Would we have embarked on an analogous antibacterial moonshot (2), doomed to be as epically catastrophic as the same endeavor for cancer? Unfortunately, the widespread success of antibacterial therapies as means to save humanity from disease collectively conditioned us to believe in the power of a small molecule wonder drug — a magic bullet — and in that faith, drugs have been pursued as treatments for disease, specifically cancers, to unreasonable ends.
Drugs as cures for disease are an incredibly rational and simple concept: they are generally stable, modular chemical structures, which can be synthesized in scalable quantities and delivered to geographic locations which need them. Analysis of small molecule efficacy, while time consuming, tends to be a fairly straightforward process as well. Activity of intended targets can be assessed in model systems and in tissue samples, distribution of the compound throughout the body can be quantified, and metabolic byproducts of the drug can be both predicted in silico and directly analyzed in vivo. Additionally, off-target effects or toxicities can be studied directly (i.e. what effect is anticipated?) or indirectly by monitoring of model systems during testing or early human studies. The entire system is controllable and littered with quantifiable outputs — further; it is extremely conducive to the randomized controlled trial (think double blinding and placebos), a gold standard for scientific methodology in biomedicine and other fields. In sum, drugs are appealing because they are scientifically elegant, but there is more to be said about their attraction.
Any compound with a large market is able to be massively lucrative since most novel, synthetic, chemical structures are patentable (3). Additionally, iterations of similar compounds may be produced (all with unique patents) as treatments for the same disease, drugging the same target. This means that next generation structures with optimized properties can be held back to ensure continued dividends from nearly identical conceptual approaches, and supply chain can control availability and pricing. Thus, drugs are also appealing because they make money. Other approaches to addressing widespread health problems (4) are less appealing because they are more difficult to study and less able to be turned into cash. Of course, monetization does not inherently vilify the concept of the small molecule drug; conversely, inherent difficulty of monetization does not imply noble motives. But it is important to point out the striking appeal of the stereotypical drug compound, and consider whether this really is the most efficient means of addressing a public health problem — or if it is simply the most lucrative means.
To collect the previous points: pharmaceutics are appealing because they have worked for other diseases, because they are conducive to quantitative scientific research, and because they are profitable. Unfortunately, such reasoning does not compose a legitimate rationale for the continued focus on drugs as means to cure cancer. To the contrary, two-thirds of the above points do constitute a great strategy to fight for the improved treatment of cancer, and researchers and physicians are making continuous incremental improvements in cancer care. Unfortunately, treating cancer is not a goal for which medicine should be moon-shooting. Medicine should be trying to cure cancer, or at least prevent its occurrence such that most malignancies are effectively extinct and the rest are infrequent enough to make headlines when they do appear. To that end, it seems reasonable for the biomedical research workforce to step away for a metaphorical coffee break, and really think about their strategy for addressing the problem of human cancer.
— — — — — — — — — — — — — Coffee break — — — — — — — — — — — — —
Humans can be thought of as machines (5), where cancer is an engineering problem. If all cancers which occurred in patients under the age of 25 were excluded, we would have thrown out roughly 1.1% of all cancer cases (6). That is because the largest risk factor for cancer is not sex, race, BMI (7), family history, or drug use. The greatest risk factor for cancer is age. An idea known as the “multiple hit hypothesis” has proven surprisingly reliable: accumulated mutations acquired over time cause progressive failures in the body’s inherent control mechanisms. While you could acquire all of these mutations in just a few minutes by exposing yourself to a heavy dose of gamma radiation, most folks don’t have access to nuclear waste (8), so the mutations are acquired over a life time. Poor diet may add some (particularly for gastric or colorectal cancers), substance abuse may add a couple more (particularly for lung and liver cancers), and family history may tack on a couple more (e.g. the well-known BRCA mutations). But the point to understand is that most of these mutations build up over the course of years or decades, and can become empowered by further mutations, as the human machine ages. Generally, these cancer driving mutations come in the form of failure in mechanisms controlled by genes known as tumor suppressors, or activation of genes known as oncogenes. The safety mechanisms controlling tumor suppressors and oncogenes are primarily identified as the most common causes of human cancers. Safety mechanisms are mechanical problems, and so far the cancer biologists, structural biochemists and molecular pharmacologists among the scientific world have done a pretty good job identifying what’s important in those safety mechanisms, and what those important things do. So, since we’ve started to understand the machine, it seems reasonable to engineer some solutions to the machine’s problems.
In order to grasp how that might happen, it’s important to understand some basic aspects of cancer biology. Cancers are typically driven by sets of mutations which enable further dysregulation of cell functions. Some of those mutations are incredibly common — consistent even in cancers which are very different. Let’s select an exemplary, commonly mutated gene and walk through a potential solution. Human cancers commonly carry mutations in a tumor suppressor protein called P53. Among human cancers, P53 mutation rates vary from 5% to nearly 50% within specific types of cancers. For example, just under 50% of ovarian cancers have acquired mutations in P53, while 17.5% of prostate cancers have acquired similar mutations. Additionally, P53 mutations are not simply correlated with cancer development. Loss of tumor suppressor function from P53 has been rigorously shown to actively drive cancer progression, thus P53 seems like a good thing to fix. This isn’t a random, novel, or unique idea. There are workshops and conferences solely dedicated to understanding P53 structure and function, because P53 is really that big of a deal. Loads of peer-reviewed scientific reports have reported their favorite flavor of P53 directed therapy for a host of cancers. A well-regarded research group from Italy just recently published a review in the FEBS Journal of current research into targeting P53 as a strategy for treatment. Even though P53 targeted therapies would most likely be very effective for treating cancers, this is (again) not what we should want.
We should want to permanently fix P53 such that it does not mutate so frequently. If we were able to do that, we wouldn’t have to worry about P53 targeted therapies for cancers, because P53 would no longer be (frequently) driving cancers. A solution of that nature would be exponentially more effective at reducing cancer mortality, compared to allowing people to get cancer, then offering them the new drugs we made because we knew they were going to be sick. We have powerful genome editing tools at our finger tips; some that have been around for a few years (think TALENS and adenoviral vectors) and some that are effectively fresh out of the box (like Crispr/Cas9). Gene editing means we have power over our biological code, which is more power than any species has ever had over their own evolution. Additionally, we have identified unique faults in our machinery, and many of those faults begin within the basic code (i.e. mutations). If we took a gene like P53 and re-engineered its DNA sequence such that it produced a similarly functioning P53 that was much less susceptible to mutation, our P53 problem would be theoretically solved — note the key word “theoretically”.
That sounds simple on a small scale, but it almost certainly will not be (on a small scale or a large one). This does not mean replacing P53 after it goes bad — that’s almost as short-sighted as just treating cancer after you get it. This means replacing every P53 gene sequence in every human with a new one. It’s just a code update. Additionally, the point here is not really P53. The point is that a strategy like this is an example of a true cure — and, in order for it to be widely impactful, it would probably have to be done in parallel with the re-engineering of other commonly problematic oncogenes or tumor suppressors. Scientists might have already tried this and it may have failed, or maybe the tools for doing this haven’t really been validated well yet, or maybe P53 re-structuring is actually super complicated and tricky. Maybe they’re trying it right now! All of those things could be true, but pursuing that goal is the sort of target that cancer needs. Plenty of projects fail the first time but succeed later on, our tools are going to get better, and we are going to learn more about P53 structures and how to re-build DNA code to be more stable if we are motivated to. Pursuing a massive genetic engineering campaign to decrease the likelihood of cancer by correcting commonly failing machinery might not actually work, but it would at least be a legitimate attempt at an actual solution to some of the largest risk factors for cancer (age, family history). And actual solutions are the things that we want.
These strategies require massive, population-wide editing of the heritable human genome. That might sound crazy to some, but to the 1.68 million people in the United States who will be diagnosed with cancer this year, it sounds a lot better than aggressive surgery, chemo, and radiation right about now. It’s certainly worth trying and talking about. By implementing a public health initiative like this, we would not be risking zombie outbreaks or designer bodies — that kind of sensational and ridiculous headline peddling has always accumulated around similar proposals and is seldom propagated by any semblance of rational opposition. Yes, gene-editing could be a slippery slope, but the existence of a slope to slip on does not preclude the existence of a healthy medium where our tools can be used to massively improve the human condition. We should not abandon all hope in drugs as treatment strategies for cancers or other diseases, neither should we cry foul when medical science fails to deliver perfect solutions for such complex problems. Research scientists and physicians are not wrapped up in some grand conspiracy with the pharma giants. Neither does cancer-preventative genetic engineering spell doom for the pharmaceutical industry. It is simply important to realize the motivations and biases which may influence us to choose to pursue specific solutions over others (or to write opinionated blog posts!). Those thought processes should help guide our opinions when we hear about new research or clinical trials that have buzzwords like “genetic engineering”. Hearing about those things should make us excited for all of the possible futures. We are beginning to have control over our own evolutionary mechanisms, something which no other organism has ever had. How we use that power is critical — our reasoned caution is just as imperative as our unflinching determination in the face of reactionary regressionism. It is hard to imagine a progressive future where the tools at our disposal are simply allowed to lie on the shelf gathering dust for fear of their repercussions. We should, instead, fear the stagnant depression of opportunity wasted in the name of comforting familiarity.
Maybe the real fight against cancer is the fight against our own willingness to settle for goals that are “achievable”, or likely to deliver results quickly, or likely to be financially beneficial. Maybe this is a huge lesson meant to teach us to play the long game, to accept the messiness of risk and failure as inevitable companions on the road to a better future, and to value each other more than we value our own fiscal security. We have a responsibility to collectively choose between accepting known risks or unknown risks — and determine their probability of delivering better outcomes. The known risks are an individual person’s current probability of developing any cancer in their lifetime (>1/3), any man’s chances of prostate cancer (1/7), a woman’s likelihood of breast cancer (1/8), a smokers chances of developing pancreatic cancer (1/32), or an individual’s risk of colon cancer (1/23). The unknown risks are precisely that: at this point nebulous and arbitrary — purely dependent on our own technology, ability, and determination. I think we can do better than 1 in 3. It falls on our collective shoulders, then, to choose to lie contented within our self-imposed limitations or to entertain audacious dreams for something better.
“ Far better it is to dare mighty things, to win glorious triumphs even though checkered by failure, than to rank with those poor spirits who neither enjoy nor suffer much because they live in the gray twilight that knows neither victory nor defeat.” — Theodore Roosevelt
Notes:
1. Vaccines were employed en masse for prevention of viral diseases well before antibacterial drugs were first used.
2. You think I’m joking, but bacterial infections used to kill way more people than cancer does now. It was an enormous problem — and it could be a huge problem again unless we start using antibiotics more responsibly. One thing you can do (yes, you!) is to buy hand soap that is antibiotic-free. That stuff is literally just as effective at cleaning your hands as the soap with antibiotics in it — plus, after you wash your hands you can feel good about yourself because you’ll have just fought bacteria and antibiotic resistance at the same time. #winning
3. Both the chemical structure and the method for synthesis of the structure are often patentable. Structures of natural product compounds are typically not patentable, while their extraction / purification methods, and formulations for delivery often can be. Still, natural products are considered much less lucrative.
4. e.g. addressing obesity and diabetes with diet and exercise.
5. à la Marvin Minsky. For a great read that doesn’t really have anything to do with cancer, see Steven Levy’s story on Mr. Minsky.
6. Those numbers are from Cancer Research UK’s data from 2011–2013.
7. Side note: BMI is a terribly ineffective measure of health, look to body fat % for more accurate assessment.
8. According to Randall Monroe, nuclear reactor pools are surprisingly safe. So, for maximum efficacy, make sure the nuclear waste is outside of the water bath.
